Committee for Risk Assessment (RAC) Committee for Socio-economic Analysis (SEAC)

Transcription

1 Committee for Risk Assessment (RAC) Committee for Socio-economic Analysis (SEAC) Opinion on an Application for Authorisation for the use of chromium trioxide in a catalyst for the dehydrogenation of propane to propene ECHA/RAC/SEAC: Opinion N AFA-O /F Consolidated version Date: 01/06/2017 Annankatu 18, P.O. Box 400, FI Helsinki, Finland Tel Fax echa.europa.eu

2 Consolidated version of the Opinion of the Committee for Risk Assessment and Opinion of the Committee for Socio-economic Analysis on an Application for Authorisation Having regard to Regulation (EC) No 1907/2006 of the European Parliament and of the Council 18 December 2006 concerning the Registration, Evaluation, Authorisation and Restriction of Chemicals (the REACH Regulation), and in particular Chapter 2 of Title VII thereof, the Committee for Risk Assessment (RAC) and the Committee for Socio-economic Analysis (SEAC) have adopted their opinions in accordance with Article 64(4)(a) and (b) respectively of the REACH Regulation with regard to an application for authorisation for: Chemical name: chromium trioxide EC No.: CAS No.: for the following use: Use of chromium trioxide in a catalyst for the dehydrogenation of propane to propene Intrinsic property referred to in Annex XIV: Applicant: Article 57 (a), (b) of the REACH Regulation Clariant Produkte (Deutschland) GmbH Reference number: Rapporteur, appointed by the RAC: Co-rapporteur, appointed by the RAC: Rapporteur, appointed by the SEAC: Co-rapporteur, appointed by the SEAC: Susana VIEGAS Sonja KAPELARI Simon COGEN Ivars BERGS This document compiles the opinions adopted by RAC and SEAC. Annankatu 18, P.O. Box 400, FI Helsinki, Finland Tel Fax echa.europa.eu

3 PROCESS FOR ADOPTION OF THE OPINIONS On 21/03/2016 Clariant Produkte (Deutschland) GmbH submitted an application for authorisation including information as stipulated in Articles 62(4) and 62(5) of the REACH Regulation. On 01/11/2016 ECHA received the required fee in accordance with Fee Regulation (EC) No 340/2008. The broad information on uses of the application was made publicly available at on 09/11/2016. Interested parties were invited to submit comments and contributions by 09/01/2017. No comments were received from interested parties during the public consultation in accordance with Article 64(2). The draft opinions of RAC and SEAC take into account the responses of the applicant to the requests that the SEAC made according to Article 64(3) on additional information on possible alternative substances or technologies. The draft opinions of RAC and SEAC were sent to the applicant on 10/05/2017. On 01/06/2017 the applicant informed ECHA that they did not wish to comment on the opinions. The draft opinions of RAC and SEAC were therefore considered as final on 01/06/2017. ADOPTION OF THE OPINION OF RAC The draft opinion of RAC The draft opinion of RAC, which assesses the risk to human health arising from the use of the substance including the appropriateness and effectiveness of the risk management measures as described in the application and, if relevant, an assessment of the risks arising from possible alternatives was reached in accordance with Article 64(4)(a) of the REACH Regulation on 15/03/2017. The draft opinion of RAC was agreed by consensus. The opinion of RAC Based on the aforementioned draft opinion and in the absence of comments from the applicant, the opinion of RAC was adopted as final on 01/06/2017. ADOPTION OF THE OPINION OF SEAC The draft opinion of SEAC The draft opinion of SEAC, which assesses the socio-economic factors and the availability, suitability and technical and economic feasibility of alternatives associated with the use of the substance as described in the application was reached in accordance with Article 64(4)(b) of the REACH Regulation on 16/03/2017. The draft opinion of SEAC was agreed by consensus. The opinion of SEAC Based on the aforementioned draft opinion and in the absence of comments from the applicant, the opinion of SEAC was adopted as final on 01/06/

4 THE OPINION OF RAC The application included the necessary information specified in Article 62 of the REACH Regulation that is relevant to the Committee s remit. RAC has formulated its opinion on: the risks arising from the use applied for, the appropriateness and effectiveness of the risk management measures described, the assessment of the risks related to the alternatives as documented in the application, as well as other available information. RAC confirmed that it is not possible to determine a DNEL for the carcinogenic properties of the substance in accordance with Annex I of the REACH Regulation. RAC confirmed that the operational conditions and risk management measures described in the application limit the risk, provided that they are adhered to as described in the application. THE OPINION OF SEAC The application included the necessary information specified in Article 62 of the REACH Regulation that is relevant to the Committee s remit. SEAC has formulated its opinion on: the socio-economic factors and the availability, suitability and technical and economic feasibility of alternatives associated with the use of the substance as documented in the application, as well as other available information. SEAC took note of RAC s confirmation that it is not possible to determine a DNEL for the carcinogenic properties of the substance in accordance with Annex I of the REACH Regulation. SEAC confirmed that there appear not to be suitable alternatives in terms of their technical and economic feasibility for the applicant. SEAC considered that the applicant's assessment of: (a) the potential socio-economic benefits of the use, (b) the potential adverse effects to human health of the use and (c) the comparison of the two is based on acceptable methodology for socio-economic analysis. Therefore, SEAC did not raise any reservations that would change the validity of the applicant s conclusion that overall benefits of the use outweigh the risk to human health, whilst taking account of any uncertainties in the assessment. 3

5 SUGGESTED CONDITIONS AND MONITORING ARRANGEMENTS Conditions -- Monitoring arrangements -- REVIEW Taking into account the information provided in the application for authorisation prepared by the applicant the duration of the review period for the use is recommended to be 12 years. 4

6 JUSTIFICATIONS The justifications for the opinion are as follows: 1. The substance was included in Annex XIV due to the following property/properties: Carcinogenic (Article 57(a)) Mutagenic (Article 57(b)) Toxic to reproduction (Article 57(c)) Persistent, bioaccumulative and toxic (Article 57(d)) Very persistent and very bioaccumulative (Article 57(e)) Other properties in accordance with Article 57(f): 2. Is the substance a threshold substance? YES NO Justification: Chromium trioxide has a harmonised classification as Carc. 1A (H350) and Muta. 1B (H340) according to Classification, Labelling and Packaging (CLP) Regulation, (EC) 1272/2008. Based on studies which show its genotoxic potential, the Risk Assessment Committee (RAC) has concluded that chromium trioxide should be considered as non-threshold substance with respect to risk characterisation for carcinogenic effect of hexavalent chromium (reference to the studies examined are included in the RAC document RAC/27/2013/06 Rev. 1, agreed in RAC-27). 3. Hazard assessment. Are appropriate reference values used? Justification: RAC has established a reference dose-response relationship for carcinogenicity of hexavalent chromium (RAC/27/2013/06 Rev.1 Final) which was used by the applicant. The molecular entity that drives the carcinogenicity of chromium trioxide (CrO3) is the Cr(VI) ion, which is released when the substances solubilise and dissociate. Cr(VI) causes lung tumours in humans and animals by the inhalation route and tumours of the gastrointestinal tract in animals by the oral route. These are both local, site-ofcontact tumours. There is no evidence that Cr(VI) causes tumours elsewhere in the body. Dose-response relationships were derived by linear extrapolation. Extrapolating outside the range of observation inevitably introduces uncertainties. As the mechanistic evidence is suggestive of non-linearity, it is acknowledged that the excess risks in the low exposure range might be an overestimate. 5

7 In the socio-economic analysis (SEA) the remaining human health risks are evaluated based on the dose-response relationship for carcinogenicity of hexavalent chromium (RAC27/2013/06 Rev.1). Are all appropriate and relevant endpoints addressed in the application? All endpoints identified in the Annex XIV entry are addressed in the application. 4. Exposure assessment. To what extent is the exposure from the use described? Description: Short description of the use This application for authorisation relates to the use of a catalyst which contains Cr(VI) due to its previous surface treatment with chromic acid outside the EU. This catalyst is needed in a specific process to dehydrogenate propane to propene. Catofin is imported in the European Economic Area by the applicant, Clariant (Deutschland) Produkte GmbH based in Frankfurt am Main, Germany. The dehydrogenation process, however, takes place at Borealis Kallo, Belgium. Borealis demand for the catalyst Catofin is < 500 tonnes every three years. According to the applicant, this amount is expected to remain stable until The largest part of the quantity used is inside the dehydrogenation reactors, about 20% is kept in storage. The fresh catalyst used in the process contains less than 10 tonnes Cr(VI). In the process Cr(VI) is reduced to Cr(III) during the reaction to produce propylene. Cr(III) is then oxidised to Cr (VI) during the regeneration process. The catalyst pellets are mixed with alumina grains. Every 3 years, the catalyst / alumina mix needs to be replaced by a fresh batch due to a loss of catalytic activity. A partial exchange of the catalyst is needed after one to two years, halfway between two full exchanges. During the catalytic cycle Cr(VI) and Cr(III) are converted from one to the other. 6

8 Below it is reported a simplified diagram with the overview of the uses: According to the applicant, the manufacturing process of propene ensures that no Cr(VI) is contained in the final product. Exposure scenario The use is described in a single exposure scenario, concerning an industrial use at a single site as follows: The use of chromium trioxide in a catalyst for the dehydrogenation of propane to propene. The exposure scenario is comprised of nine Worker Contributing Scenarios (WCS) and one Environmental Contributing Scenario (ECS). According to the applicant, the exposure scenario includes all relevant processes and tasks associated with the use of Cr(VI) that could result in either environmental or human exposure. Worker exposure According to the applicant, the potential for exposure occurs primarily during the loading of the catalyst in the dehydrogenation reactor, the mixing of the catalyst and the unloading of the spent product from the reactor. The unloading and loading of the whole catalyst mix is done during a full shutdown of the production site under closely monitored circumstances. The complete shutdown is needed to perform maintenance operations on the installation. Between two complete production shutdowns, there is a minor shutdown of the installation, mainly to replace the top layer of the catalyst in the reactors, which is called top-layer-exchange. As the complete shutdown occurs every three years (for +/- 18 days working days), the minor shutdown in between also takes place every three years (for +/- 12 days working time). The unloading and the loading process of the catalyst mix is performed every three years and consists of four steps: Unloading of the spent catalyst mix from the reactors; Separating spent catalyst pellets from the alumina grains (the latter are reused); 7

9 Mixing of fresh catalyst pellets with the (re-used) alumina grains; Loading of the fresh catalyst mix into the reactors. The top-layer exchange is of shorter duration than the entire exchange as only a fraction of the catalyst (the top-layer) is unloaded. Neither separating of spent catalyst pellets nor mixing of fresh catalyst take place during the activity. The catalyst for the top-layer exchange is taken from the storage which is prepared during the full shutdown of the production (see above). The reaction of dehydrogenation of the propane for the manufacture of the propene takes place in closed reactors where the catalyst is used. At the same time in other closed reactors the catalyst is regenerated and then reduced. According to the applicant, exposure is controlled and minimised by technical and organisational risk management measures (RMM). If exposure cannot be avoided personal protective equipment (PPE) has to be used (see table 1). Table 1: Summary of WCS, technical and organisational RMM and PPE and operational conditions presented in the use WCS Technical RMM Organisational RMM PPE* Duration/ frequency 0 of tasks WCS 1 Mixing of Catofin catalyst (PROC 26) Mixing activity performed in closed shelter with controlled access. Mixing installation is closed except conveyors leading to mixing zone. Conveyers are enclosed in container located in shelter. LEV with filter present inside shelter. Storage bins for mixed material located outside container, but connected with closed piping to the container. Controlled access to restricted area. Strict inventory management and tracking of Crcontaining product. Decontamination procedure and facilities. Specific worker training on carcinogenic substances and on correct use of PPEs. Supervision during entire activity. Disposable dust filter FFP3 with APF (UK) of 20. Tyvek coverall and standard construction safety gloves. Chemical resistant nitrile gloves are used in case of possible direct contact with the substance, e.g. when a worker enters the mixing container. Duration: 14 days / 3 years; 9 hours per a 12 hours shift Frequency: days / year WCS 2 Loading of the reactor with Catofin catalyst (PROC 26) Reactor is closed vessel with limited number of openings. LEV with filters creates net airflow into reactor via top manholes. Controlled access to restricted area. Strict inventory management and tracking of Crcontaining product. Decontamination procedure and facilities. Demand valve breathing apparatus with APF (UK) of 1,000. Tyvek coverall and chemical resistant nitrile gloves. Duration: days / 3 years; 4 hours per a 12 hours shift Specific worker training on 8

10 carcinogenic substances and on correct use of PPEs. Supervision during entire activity. Frequency: days / year WCS 3 Use of the Catofin catalyst in a fully closed system (PROC 1) Reactor is closed during use. Leak tightness regularly checked. Entrainment of catalyst is prevented. Proper functioning of installation monitored continuously. Specific worker training on carcinogenic substances and their presence on site. Supervision. Continuous use; No worker exposure. WCS 4 Unloading of the spent Catofin catalyst from the reactor (PROC 26) Vacuum technology creates net airflow into reactor. Vacuum cyclone filter avoids release of dust to atmosphere. Controlled access to restricted area. Strict inventory management and tracking of Crcontaining product. Decontamination procedure and facilities. Specific worker training on carcinogenic substances and on correct use of PPEs. Disposable dust filter FFP3 with APF (UK) of 20. Tyvek coverall and chemical resistant nitrile gloves. Duration: days / 3 years; 9 hours per a 12 hours shift Frequency: days / year Supervision during entire activity. WCS 5 Lab analysis (PROC 15) Exempted from authorisation. WCS 6 Separation of the Catofin catalyst (PROC 26) Separation activity performed in closed shelter with controlled access. Separation installation is enclosed in container located in shelter. LEV with filter present inside shelter. Storage bins for separated material located outside container, but connected with closed piping to the container. Controlled access to restricted area. Strict inventory management and tracking of Crcontaining product. Decontamination procedure and facilities. Specific worker training on carcinogenic substances and on correct use of PPEs. Supervision during entire activity. Disposable dust filter FFP3 with APF (UK) of 20. Tyvek coverall and standard construction safety gloves. Chemical resistant nitrile gloves are used in case of possible direct contact with the substance, e.g. when a worker enters the separation container. Duration: 12 days / 3 years; 9 hours per a 12 hours shift Frequency: days / year WCS 7 Specific worker training on carcinogenic substances and on Disposable dust filter FFP3 with APF (UK) of 20. Duration: 10 days / 3 years; 9

11 Cleaning of equipment (PROC 28) correct use of PPEs. Supervision during entire activity. Tyvek coverall and standard construction safety gloves. 6 hours per an 8 hours shift Frequency: days / year WCS 8 Supporting far-field activities (PROC 26) Far-field exposure only. Supporting activities related to mixing and separation are performed in a closed shelter with controlled access. Mixing and separation installations are enclosed in container located in a shelter. Controlled access to restricted area. Decontamination procedure and facilities. Specific worker training on carcinogenic substances and on correct use of PPEs. Supervision during entire activity. Disposable dust filter FFP3 with APF (UK) of 20. Tyvek coverall and standard construction safety gloves. Duration: 52 days / 3 years; 9 hours per a 12 hours shift Frequency: 11.2 days / year Supporting activities related to loading and unloading are performed outdoors. WCS 9 Storage of the Catofin catalyst in fully closed system (PROC 1) Substance is fully contained. Specific worker training on carcinogenic substances and their presence on site. No worker exposure. * APF of PPE = Assigned protection factor = an indicator of effectiveness: APF 20 = 95% effectiveness (100% / 20 = 5% level of residual risk), APF 1,000 = 99,9% effectiveness (100% / 1,000 = 0.1% level of residual risk). 0 The frequency of exposure is calculated as follows: In a 12 hours shift regime, workers typically work three shifts in one week and four shifts in the next week, which means seven working days over a 14 days period. 1 The mixing / separation operation takes 14 days in total and the operation is performed once every three years. Therefore, the worker exposure frequency is seven days per three years or 2.3 days per year. 2 The loading takes place twice every three years: once during full catalyst exchange (9 days), and once during top layer exchange (5 days). Therefore, the frequency of exposure of a worker is max. six days during the full catalyst exchange period and four days during the top layer exchange period, or ten days over three years. This corresponds to 3.3 days per year. 3 The unloading operation takes place twice every three years: once during full catalyst exchange (8 days), and once during top layer exchange (4 days). Therefore, the frequency of exposure of a worker is max. five days during the full catalyst exchange period and four days during the top layer exchange period, or nine days over three years. This corresponds to 3 days per year. 10

12 4 The cleaning operation takes place twice every three years: once during full catalyst exchange, and once during top layer exchange. The cleaning operation takes ten days in total the frequency of exposure of a worker is 10 days per three years or 3.3 days per year. The applicant provided a detailed description on the workers involved in each of the exposure scenarios. They distinguished clearly between workers who have direct contact to Cr(VI) and workers who are present in the contaminated area during the catalyst exchange operations. These workers do not perform activities for which direct contact with the catalyst is expected. The applicant identified their tasks as far-field activities (see WCS 8). The frequency of exposure for these workers is calculated as the sum of the exposure frequencies for the subsequent operations: mixing (2.3 days / year), loading (3.3. days / year), unloading (3.3 days / year) and separation (2.3 days / year). This results in a total frequency of 11.2 days / year. RAC notes that sampling tasks on the substance for laboratory use were not considered in the exposure assessment. On RAC s request the applicant clarified that sampling is performed only on the spent catalyst on every exchange (top-layer-exchange or complete exchange) of the catalyst. That means sampling occurs every 1.5 years from inside the reactor. The tasks can be considered as covered by the exposure assessment of WCS 4. Only a minimum amount is required to enable the envisaged analyses. The samples are transported to the laboratory in double closed containers (a closed box inside another closed box). Exposure estimation methodology: Inhalation exposure: The assessment of inhalation exposure provided by the applicant is based on qualitative assessment and on the results of air monitoring campaigns undertaken with static and / or personal sampling. The measurements include simultaneous determination of inhalable and respirable dust fractions. The applicant provided several monitoring results for WCS 1, WCS 2, WCS 4, WCS 6 and WCS 8 on the inhalable and respirable fraction of Cr(VI). The applicant pointed out that for WCS 1, WCS 2 and WCS 4 measurement results were obtained above the limit of detection (LoD) while the measured concentrations for WCS 6 and WCS 8 were below the LoD. The absolute LoD was for both types of air measurements (inhalable and respirable fraction) mg per sample. The sampling period varied from 91 minutes to 591 minutes and was related to the duration of the tasks performed (see below). For WCS 1 and WCS 2 the highest measurement result of several measurements on the inhalable fraction of Cr(VI) was selected for risk characterisation. Only for WCS 4 the highest result of the respirable fraction was taken into account as the inhalable fraction showed values below the LoD. It has to be noted that two sampling devices were used (to guarantee the collection of the respirable and inhalable fraction on Cr(VI)), that is why slightly different volumes were collected. As the collected volume on the respirable fraction was a little bit bigger than on the inhaled fraction the measured results on the respirable fraction were above the LoD. Besides, the applicant pointed out that there were some measurements for WCS 1 including also tasks of WCS 6. This relates to the highest measured value for WCS 1 taken forward for risk characterisation. For the purpose of this assessment, however, the applicant assumed that the main source of exposure to Cr(VI) occurred during the mixing activity (WCS 1) which is substantiated by the fact that no Cr(VI) was detected in any of the personal measurements performed during separation (WCS 6). All of the 11

13 10 measurements results provided showed values below the LoD. If an activity was interrupted by a break, the sampling device was taken off the worker and re-installed after the break. The static measurements, however, were not interrupted during breaks. Therefore the results of the personal measurements reflect the average exposure concentration of Cr(VI) over the entire activity while the results of the static measurements which also was performed reflect the Cr(VI) concentration over the entire shift (activity and breaks). No air monitoring data were provided for four WCS: For WCS 5 the exemption for scientific research and development is claimed. For WCS 3 and WCS 9, the assessment on inhalation exposure is based on a qualitative assessment only. For WCS 7, the applicant used the same value for the risk assessment as for the unloading of the reactor (WCS 4). According to the applicant, this approach is justified as a vacuum truck is used for both, the unloading operation and the cleaning and the volume of catalyst mix handled during unloading is many orders of magnitude larger than the volume of dust collected during vacuum cleaning. For WCS 3 (production process) and for WCS 9 (storage of the catalyst) the potential exposure is negligible, according to the applicant due to closed process / system. In addition to the information related to air monitoring, there are also biomonitoring data. The workers of Borealis as well as the contract workers are under biomonitoring surveillance during all activities in the catalyst exchange process. The average urinary concentration of total chrome in the latest exchange process in 2015 of 18 samples was 0.33 µg Cr/L (creatinine corrected). This value is similar to the typical urinary chrome concentration expected for a non-exposed population. As no contextual information was gathered to this monitoring campaign, the applicant stated in the trialogue and in writing after the trialogue that in the future, more contextual information will be collected during the biomonitoring campaigns and that blank samples will as well be taken. 12

14 Table 2: Inhalation exposure WCS Method of assessment Exposure concentration (µg/m 3 ) Exposure value (µg/m 3 ) corrected for RPE Exposure value (µg/m 3 ) corrected for RPE*, duration and frequency WCS 1 (PROC 26) Measured data (10 personal measurements) * WCS 2 (PROC 26) Measured data (9 personal measurements) ** WCS 3 (PROC 1) WCS 4 (PROC 26) WCS 5 (PROC 15) WCS 6 (PROC 26) WCS 7 (PROC 28) WCS 8 (PROC 26) WCS 9 (PROC 26) Qualitative assessment Measured data (8 personal measurements) Measured data (10 personal measurements) Measured data (the same as for WCS 4) Measured data (34 personal measurements) Qualitative assessment * Exempted from authorisation < 0.10 < 0.005* * < 0.08 < 0.004* * RPE is considered with 95% effectiveness. ** RPE is considered with 99.9% effectiveness. Dermal exposure: The applicant has not assessed dermal exposure in accordance with the RAC reference document which states that there are no data to indicate that dermal exposure to Cr(VI) compounds presents a potential cancer risk to humans (RAC27/2013/06 Rev. 1). 13

15 Combined exposure: The applicant considered combined exposure due to the fact that the catalyst exchange procedure consists of several operations that are performed sequentially. Based on the measured exposure levels per activity and taking into account the duration and frequency of each single activity, the applicant provided a combined risk assessment for two different groups of workers. Group A includes workers with direct contact to the catalyst and group B is comprised of workers that are present in the contaminated area (see section 6). Uncertainties related to the exposure assessment: RAC further acknowledges that the applicant used the inhalable fraction of Cr(VI) for the exposure (risk) assessment for WCS 1, WCS 2, WCS 6 and WCS 8 which is a worst case approach as the inhalable fraction consists of respirable and non-respirable particles. However, RAC notes that the exposure estimation for WCS 7 might be rather conservative as it is based on measurement data of WCS 4 (unloading of the reactor) where the volume of catalyst mix handled is many orders of magnitude higher than the volume of dust collected during cleaning. During the trialogue, the applicant stated that they will include WCS 7 in future monitoring campaigns which are planned during every catalyst handling. For WCS 8, the applicant assumed that the operations are performed subsequently resulting in an overestimation of the number of shifts in the provided exposure (risk) assessment. The actual number of shifts performed by a worker is smaller. Environmental releases / Indirect exposure to general population (humans via the environment) The applicant considered that Use of reactive processing aid at industrial site (no inclusion into or onto article), ERC 6b, is the most appropriate environmental release category. However, the applicant decided to base the release factors on plant specific measurement information and not on the default release factors for ERC 6b. Estimation of releases According to the applicant, there are strict emission control measures in place to avoid emissions to all environmental compartments. The presence of Cr(VI) in air and water is caused by the entrainment of catalyst dust by the gas flow. In one case the entrainment occurs by air and in the second case by dry steam. Release to air The air used for regeneration contains a low level of Cr(VI). This is the result of entrainment of catalyst dust caused by the high gas (air) velocities in the reactor. A high air volume is needed for quickly evacuating the heat generated by the burning of the coke. If the air volume was too low it would cause hot spots in the reactor that would damage the catalyst. The air and reduction gas from the production is burned in a heat recovery boiler from which the fluegas is collected in a stack. The Cr(VI) concentration from one measurement, performed on 14 th January 2016 during normal operating conditions, in the exhaust of the stack, ca. 8 m below the top, was used to calculate the annual 14

16 emission to air and the corresponding release factor. In the trialogue, the applicant explained that since 14 th January 2016 regular Cr(VI) measurements have been included in the analysis programme at the Kallo site. Since the submission of the application, a further 10 measurements are available for The measured exposure concentrations, provided on RAC s request, showed that in one case the result was four times the average, variability that was surprising for the applicant since the process is very stable (e.g. flow rate, temperatures, pressures). As at the time of construction of the application for authorisation only the measurement from January 2016 was available, the calculations of the environmental exposure were performed with this value. So the variability of data does not seem to underestimate the environmental exposure. However, the applicant will still try to find out the reasons for the variability as they have not been found yet. Besides the applicant pointed out that this type of measurements are very difficult to perform due to the very high air flow rate. Release to water In the past, total chromium was monitored four times a year at the final disposal, immediately prior to the release to the river surface water (according to the Belgian regulatory framework with a limit of detection (LoD) of mg/l. For the purpose of the dossier, an additional Cr(VI) measurement (LoD of mg/l) was performed, which was located more upstream. As the result of this Cr(VI) measurement was below detection limit, it was decided not to perform any additional measurements on this location, but only continue measuring at the final disposal point (effluent of the WWTP). Since 2016 the measurements at the final disposal point include Cr(VI) measurements. This will also still be done in the future measuring campaigns. Furthermore, additional samples will be taken during turn-overs on a weekly basis. The monitoring results support the assumption that releases to wastewater are limited and occur only as a result of entrainment in the steam injected in the reactor at the end of the dehydrogenation phase. In this phase chromium is normally only present as Cr(III). Entrainment of catalyst dust within the steam is minimised by the use of a support layer of alumina balls under the catalyst bed. However, any entrained Cr(VI) in the catalyst dust would be captured in the steam condensate and evacuated through the wastewater treatment system. The applicant calculated the release factor of Cr(VI) to wastewater by estimating the total load of Cr(VI) in the steam using the concentration of Cr(VI) measured in the air (see above) applied to the steam flow. The applicant assumed that the entrainment per gas volume of dry steam is the same as for air. This can be considered a very precautionary approach as the velocity of steam is much lower than that of the air in the reactor. According to the applicant, releases to soil are excluded and are therefore considered to be negligible. If any spill to soil or the drainage system occurs, the potential contaminated soil is sent to appropriate waste handler under the supervision of a soil expert. The same waste management practice is done with any contaminated sludge and water used for cleaning as well as with all consumables (e.g. filters / bags / plastics) in contact with Cr(VI) containing substances. 15

17 Table 3: Releases to the environment Release Release factor / release rate Release* per year Release estimation method Air Release factor: < 0.5% Local release rate: < kg/day < 8.7 kg Release based on measured site-specific data (airflow rate and concentration of Cr(VI) in the exhaust air). Water Release factor: < 0.005% Local release rate: < kg/day < kg Release is estimated using the measured Cr(VI) concentration in the air applied to the steam flow. Soil * The release is based on 347 days per year. Exposure estimation methodology: The applicant provided an assessment of indirect exposure to humans via the environment at local and regional scale based on EUSES modelling. Table 4: Summary of indirect exposure to humans via the environment Protection target General population Inhalation (µg/m 3 ) General population Oral (µg/kg bw/day)* Exposure estimate, EUSES, local scale * The applicant did not consider the oral route for risk assessment. The applicant noted that in general the inhalation and oral exposure route might exist for exposure of general population. However, as the site is located in an industrial area with no citizens living within a radius of 1 km, the applicant s exposure assessment for the general population comprises an inhalation exposure assessment for the 1820 workers in this area. Exposure due to the consumption of contaminated food and drinking water has been waived on the assumption that food and drinking water of the workers is not locally sourced. RAC notes that using the Cr(VI) concentration in air estimated at a 100 m distance from the emission source is likely to overestimate exposure for many of the individuals working within a 1 km radius of the point source. However, it may not overestimate exposure for all workers, particularly if they work closer than 100 metres from the point source. As such, RAC does not necessarily concur that this is a worst case assessment for indirectly exposed workers. RAC acknowledges the applicant s method to refine the default release factors of ERC 6b by using measured site-specific data for the releases to air and for using these data 16

18 to calculate also the releases to wastewater. The release factor to air based on plant specific measurement information (< 0.5%) is greater than the default release factor to air for ERC 6b (0.10%) whereas the applicant s estimated release factor to water (< 0.005%) is much smaller than the default value (5%). It is not possible to corroborate the applicant s estimate of wastewater releases using the available wastewater monitoring data as the LoD reported for Cr(VI) is currently too high. Borealis is aware that a lower LoD for determining Cr(VI) in water has been described in literature, however, the exact LoD depends on the water matrix, resulting in the LoD presented. Uncertainties related to the assessment of exposure to humans via the environment: RAC acknowledges that assessment of indirect exposure to humans via the environment using default assumptions via EUSES is conservative, particularly at the local scale and could lead to an overestimation of risk (and number of statistical cancer cases). RAC notes that the applicant did not assess general population exposure per se, but instead assessed the indirect exposure of workers within a 1 km radius of the point source. Whilst RAC considers that such an assessment is useful, it notes that there is no prescribed minimum distance from a site that determines if a general population assessment should be undertaken (or not) and that the exposure to Cr(VI) occurring at a distance from the site consistent with the nearest residential areas could have been estimated by the applicant (or the default concentration applied to the general population). However, RAC considers that the indirect exposure assessment presented by the applicant is suitable for impact assessment and that the absence of a true general population exposure assessment does not introduce significant uncertainty in this application on the basis that (i) oral exposure is not generally considered to be significant for Cr(VI) because of the rapid transformation of Cr(VI) to Cr(III) in the environment and (ii) the use of the air concentration 100 metres from the point source is likely to balance out the influence of a potentially larger exposed population size further away (>1 km) from the site. RAC notes that the annual emission to wastewater is calculated by extrapolation of the measured concentration of Cr(VI) in air while the Cr(VI) was not detectable (LoD < mg/l) in wastewater itself. Therefore the exposure assessment can be considered as a conservative approach. Conclusion RAC considers that for both workers exposure and indirect exposure of humans via the environment: - The description of use provided allows conclusions related to exposure situations to be drawn. - The methodology used and the information provided, related to exposure resulting from the use applied for, is considered to be sufficient to use in the risk assessment and in the risk characterisation. - Concerning workers exposure via inhalation, the applicant provided measured data for six out of seven WCS. - Estimates of releases to air and wastewater follow an appropriate approach. Overall, the uncertainties identified are considered to be relatively minor and do not invalidate the applicant s exposure assessment. 17

19 5. If considered a threshold substance, has adequate control been demonstrated? YES NO NOT RELEVANT, NON THRESHOLD SUBSTANCE Justification: 6. If adequate control is not demonstrated, are the operational conditions and risk management measures described in the application appropriate and effective in limiting the risk? YES NO Evaluation of the Risk Management Measures Regarding the RMMs in place, RAC concludes that the highly automated petrochemical production process with a high degree on line measurements and several control loops to keep the process conditions automatically within control limits as well as measures to minimise entrainment of catalyst dust with the injected air do assure the minimisation of any emissions of the substance during the production process. Furthermore, the technical (e.g. gravitational loading with LEV extraction from reactor bottom, dedicated gravity and pneumatic piping, unloading with vacuum tracks, filter in exhaust) and organisational measures (e.g. the restricted access to this area to personnel with proper permit and training on carcinogenic substances and correct use of PPE, the installation of a decontamination zone including a clean room, a washing room and a process room) during the unloading and loading activities of the catalyst and the obligation to wear adequate PPE also seem to be appropriate in limiting the risks. RAC considers that the implemented RMMs and OCs as described in the application are appropriate and effective in limiting the risks to workers and the general population, provided that they are adhered to. Justification: Risk characterisation The applicant has estimated cancer risk according to the RAC reference dose response relationship for carcinogenicity of hexavalent chromium (RAC 27/2013/06 Rev. 1, agreed at RAC 27). The applicant has conservatively assumed that all inhaled chromium trioxide particles are in respirable range and contribute to the lung cancer risk. 18

20 Worker Based on exposure for 40 years (8 hours/day, 5 days/week), the excess lifetime lung cancer mortality risk according to the RAC reference dose response relationship is per µg Cr(VI)/m 3. The exposure assessment was mainly based on measured (WCS 1, WCS 2, WCS 4, WCS 6, WCS 7 and WCS 8) data. However, for WCS 3 (production process) and WCS 9 (storage of the catalyst) a qualitative assessment was undertaken. Table 5: Excess risk estimates for 40 years exposure for workers calculated by the applicant Inhalation route WCS / PROC Corrected exposure estimates (µg/m 3 ) Excess lung cancer risk WCS 1 / PROC WCS 2 / PROC WCS 3 / PROC WCS 4 / PROC WCS 5 / PROC 15 Exempted from authorisation. WCS 6 / PROC WCS 7 / PROC WCS 8 / PROC WCS 9 / PROC

21 Table 6: Excess risk estimates for combined exposure for workers calculated by the applicant Excess lung cancer risk WCS / PROC Group A Group B WCS 1 / PROC WCS 2 / PROC WCS 4 / PROC WCS 6 / PROC WCS 7 / PROC WCS 8 / PROC * Total * The Excess risk for the supporting activity as described in WCS 8 is However, this is based on a frequency of 11.2 days / year and an exposure duration of 9 hours per shift and is considered for workers of group B. The excess risk for supporting far-field activities for workers of group A is corrected for a shorter duration (max. 5 hours / day) and lower frequency (3.3 days / year) corresponding to the frequency determined in WCS 2. RAC agrees with the methodology used by the applicant to calculate the risk and with the results obtained. RAC considers that the applicant s risk assessment is acceptable as for the main WCS a robust dataset of measurements is provided and the remaining uncertainties with regard to the exposure assessment are considered to be minor by RAC. Indirect exposure to humans via the environment The applicant conservatively assumed that all inhaled chromium trioxide particles are in respirable range and contribute to the lung cancer risk. Thus, an excess life-time lung cancer risk is per μg Cr(VI)/m 3 for 70 years of exposure (24 hours/day, 7 day/week). The applicant provided the assessment of indirect exposure to humans via the environment at the local scale based on EUSES modelling and the population of indirectly exposed workers within a 1 km radius of the point source. As discussed in section 4 1, RAC considers that this approach is acceptable in this application. 1 The applicant noted that in general the inhalation and oral exposure route might exist for exposure of general population. However, as the site is located in an industrial area with no citizens living within a radius of 1 km, the applicant s exposure assessment for the general population comprises an inhalation exposure assessment for the 1820 workers in this area. This assumption has been used when calculating the excess lung cancer risk presented in the Table 7. 20

22 Table 7: Excess risk estimates for indirectly exposed workers calculated by the applicant Protection target Exposure estimate, EUSES, local scale Excess lung cancer risk General population Inhalation (µg/m 3 ) Conclusion - The RMMs are appropriate and effective in limiting the risks to workers and the general population. - RAC considers the methodology used for cancer risk calculation to be appropriate and that the estimates of excess lung cancer risk for directly and indirectly exposed workers are reliable and allow health impact assessment for both workers and the true general population. - RAC agrees with the applicant s approach not to consider the regional scale for the exposure assessment for humans via the environment. Overall, the uncertainty related to the assessment of the risk for workers and humans via the environment are considered to be low. 7. Justification of the suitability and availability of alternatives 7.1 To what extent is the technical and economic feasibility of alternatives described and compared with the Annex XIV substance? Description: Summary of the analysis of alternatives undertaken by the applicant In this application by Clariant, CrO3 makes up less than 5% weight of the total amount of a proprietary catalyst called CATOFIN. This catalyst is specifically designed to be used in conjunction with the CB&I Lummus technology Borealis currently uses at their Kallo site for the dehydrogenation of propane to propene. The applicant solely acts as an importer of the CATOFIN catalyst for the use applied for. It is actually the applicant s downstream user Borealis who uses this catalyst in the dehydrogenation of propane. The resulting propene (as well as additional imported propene 2 ) is then predominately polymerized to polypropylene 3 which is used in a wide array of applications (including automotive, medical and food applications). Less than 3 tonnes per year of Chromium Trioxide (CT) which, as said earlier, makes up less than 5% wt of the total amount of CATOFIN - is purchased by Borealis to replace the deactivated catalyst. 2 With the currently used 240kt/year Lummus plant, additional import of polymer-grade propene is necessary to feed into the polypropylene plant at the Kallo site. In other words, Borealis own production is not sufficient to meet demand. 3 Internally by Borealis at the same site. 21

23 In their clear and relatively comprehensive Analysis of Alternatives (AoA), which is tightly linked to the SEA in this case, the applicant provides SEAC with an overview of their activities and especially those of its downstream user Borealis. They also elaborate on the global propene market, the different production methods 4 and their market share, as well as the importance of polypropylene for which propene is the precursor monomer. After this more general discussion the applicant goes into more specifics on the dehydrogenation process, the key requirements for the catalyst and research done by Clariant to find an alternative. This means that the focus of the applicant s R&D discussion is placed squarely on the currently used CB&I Lummus technology at the Borealis plant, which is understandable. However, when discussing alternatives and alternative technologies beyond drop-in alternatives, Clariant continues to adopt this narrow viewpoint to assess technical and economic feasibility. Since this is an upstream application, albeit with only one downstream user, SEAC consider that the focus should have been broader. Nevertheless, the applicant provided SEAC with enough information, both in the initial application and in the responses to our questions, on all the possible types of alternatives to assess the technical and economic feasibility from Borealis viewpoint. During the opinion-making process SEAC became aware of a press release from September 2016 (see annex) which states that Borealis is performing a feasibility study for an on-purpose propane dehydrogenation plant 5 (PDH plant) to be used in addition to the current Lummus production unit. This additional PDH plant would also be located at the Kallo site and make use of Honeywell UOP s Oleflex technology (platinum-based catalyst) instead of CB&I s Lummus technology. SEAC asked for more information regarding this study since it undoubtedly affects the committee s assessment of feasibility (as well as that of the SEA) and will also play a role in the determination of the review period. The applicant did not provide SEAC with preliminary results on this study (which was/is due to be finished in July 2017). It was however once again made clear by the applicant that the reasoning behind this new plant is to respond to (expected) growth in the polypropylene market, the decrease in supply of propene from steam crackers and refineries as well as to become less reliant on the import of propene 6. As previously mentioned, the applicant has indicated that it is the intention of Borealis to use this plant in parallel with the Lummus technology to ensure a stable and sufficient supply of propene in light of the aforementioned market prospects/evolution. The fact that Borealis is currently investing heavily in order to extend the longevity of Lummus plant seems to support the need for this source of supply. During the trialogue the applicant, as well as Borealis, provided further information regarding this issue and addressed satisfactorily the questions that were raised because of the press release. Conclusion Overall the AoA was considered to be clear and relatively comprehensive by SEAC. While the Committee has valid reservations on the viewpoint taken by the applicant, ample 4 Of which the CB&I Lummus technology is only one. Another, for example, is the Honeywell UOP Oleflex process (which uses platinum-based instead of chromium-based catalyst). 5 The press release claims that the planned plant would be the largest in the world with a targeted annual production capacity of 740 kt, which is more than triple the capacity of the current propane dehydrogenation plant. 6 See section 7.2 for a discussion on the issues with importing propene. 22

24 information was present for SEAC to assess the technical and economic feasibility of the discussed alternatives and come to a firm conclusion. 7.2 Are the alternatives technically and economically feasible before the sunset date? YES NO Justification: It is important to note again that Borealis is the company that actually uses CT and that SEAC s assessment of the AoA therefore fully takes on their viewpoint and not only that of the applicant, Clariant. Furthermore, there is only one possible scenario for Clariant in case of nonauthorisation, to wit the cessation of import and supply of the Catofin catalyst, and the impacts of that scenario are largely felt outside the EEA (for the applicant). This strengthens SEAC s decision to focus mainly on Borealis since their viewpoint becomes of relevance to the AoA and the subsequent choice of the non-use scenario from which the socio-economic impacts are derived. The applicant discussed 3 types of alternative which cover 10 alternatives in total. A. DROP-IN ALTERNATIVES 7 1. Cr-based catalyst with CrO3 concentration < 0.1% Clariant, together with Borealis, is trying to decrease the amount of Cr(VI) present in the Catofin catalyst and is having some success with this. It is however clear, both to the applicant and SEAC, that this alternative can in no way be seen as a suitable alternative as Cr(VI) would still be generated and emitted, even though the exposure to Cr(VI) would be reduced. 2. Vanadium-based dehydrogenation catalyst This type of catalyst has been tested by Clariant on laboratory scale only and was shown to produce lower conversion rates (higher number of passes before propane is fully converted) and have a shorter lifetime (more plant downtime to replace the catalyst). Both of those characteristics clearly negatively impact the economics of the Borealis site. The applicant claims that the margin for Borealis polypropylene production would decrease by several tens of million euros per year. Based on the provided calculations, SEAC can agree to this. This does not mean that the process (and consequent production of polypropylene) would not be profitable anymore, only less so. SEAC s opinion on the unsuitability of the alternative does however not change because of this since at the moment a Vanadium-based catalyst cannot be considered a technically feasible alternative on an industrial scale. 3. Platinum-based Catalyst The same remarks can be made here as for the Vanadium-based catalyst. The only difference is that laboratory tests performed by Clariant have shown that this alternative performs even worse. It is however not clear to SEAC if the technical performance of 7 For use in the currently used Lummus process for dehydrogenation of propane. 23

25 the catalyst would make the process (and consequent production of polypropylene) completely unprofitable and in this case also economically infeasible. The catalyst used in an Oleflex dehydrogenation plant (see also later) is based on platinum, but was not considered technically feasible as such for use in the Lummus process due to differences in engineering 8. Clariant has clearly indicated during the trialogue that over the next years it, in close cooperation with Borealis, will continue research to develop a chromium-free drop-in alternative for the Lummus plant in Kallo. This is considered to be the most advantageous option for Borealis since this means that the capacity for an independent propene supply would remain at an optimal level. It is furthermore less costly, although more time-consuming, option than scrapping the Lummus plant and building a second Oleflex plant in its stead (which would engender costs of several hundred million euro 9 ). SEAC agrees with this reasoning and finds this option also preferable from society s perspective. B. ALTERNATIVE CATALYSTS (PRODUCTION INSTALLATION) 1. Oleflex dehydrogenation process Oleflex is an alternative technology for the on-purpose dehydrogenation of propane and, as stated earlier, uses a platinum-based catalyst. This technology is the main competitor of the Lummus process and they have roughly an equal market share. Since the Oleflex process is already widely used it is absolutely clear to SEAC that this technology is technically feasible. Adopting it would however entail the building of an entirely new plant, the costs of which were roughly estimated at several hundred million of euros (assuming the same production capacity of 240 kt/year as the current plant). Considering the fact that Borealis is/was carrying out a feasibility study for the construction of an Oleflex plant with a capacity of 740 kt/year and strongly considering the investment 10, SEAC finds that this dehydrogenation process can become economically feasible already in the very near future, if looked at in isolation and not considering Borealis specific situation. However, SEAC also took into consideration the market prospects and linked concerns regarding continuity of supply. Based on information provided by the applicant, the new 740 kt/year Oleflex plant 11 is designed to cover a widening supply-demand gap due to the expected growth of the polypropylene market and the decreasing supply of propene from traditional sources (steam crackers and refineries). As such, this 740 kt/year plant cannot be seen as an alternative to the current Lummus plant since that propene production unit would still need to be operated in parallel with the planned Oleflex plant. To provide a replacement for the current 240 kt/year propene production and the additional propene import (needed to meet the present demand), further capacity would need to be built. According to the applicant this would, due to technical reasons, require a construction of a separate plant. Although technically possible, this would entail further investments 12. As a result, SEAC finds that scrapping the Lummus plant and 8 Catalyst pellet shape and size are different for both technologies and have a huge effect on its operation. 9 On top of the costs for the 740 kt/year plant which is due to the changes in the propene market a necessity for Borealis to keep doing business. 10 Due to the expected growth in the polypropylene market and the decrease in supply of propene. 11 Which is as large as technically possible for an Oleflex plant. 12 As much as double the investments for the planned 740 kt/year Oleflex plant. 24

26 building additional capacity in the form of a second 13 Oleflex plant in its stead can in no way be seen as an economically feasible alternative at the sunset date. 2. Krupps-Uhde Star process The applicant indicates that this is a relatively new technology for the on-purpose dehydrogenation of propane. Only one such installation is in operation globally and the applicant therefore does not consider it to be fully mature. While SEAC can understand the hesitation for considering a less proven technology, especially compared to Oleflex, it cannot be considered technically infeasible. The reported cost for the one plant already in operation seem to suggest that they are in the same range as the Oleflex plant and as such this alternative technology can become economically feasible as well in the near future for the construction of a new plant. However, as discussed above, SEAC also took into consideration the market prospects and linked concerns regarding continuity of supply and concludes therefore that scrapping the Lummus plant and building additional capacity, i.e. a second 9 plant in its stead, would not be economically feasible by the sunset date. C. ALTERNATIVE SOURCES OF PROPENE Three of the alternatives in this category namely catalytic metathesis, Fluidized Catalytic Cracking (FCC) and a steam cracker are inextricably linked to the gas and oil industry and are as such far away from Borealis core business in Belgium (onpurpose production of propene and subsequent polymerization to polypropylene). Catalytic metathesis for example only makes economic sense when integrated into a steam cracker and FCC when part of a crude oil refinery. In addition to that, both FCC and a steam cracker are big sources of propene, but it is only a by-product in those processes and often not even near the quality necessary for use in Borealis polypropylene plant. SEAC therefore finds the above-discussed alternatives not to be economically feasible. Two final alternatives, biosourced propene and the purchase/import of propene, are said to pose problems of availability as well as economic feasibility. Biosourced propene is produced using sugar cane of which a huge amount is needed to feed the 240 kt/year plant that Borealis has now. In the EU there is insufficient renewable sources available for the size of the plant without affecting availability of land for food or other resources. Furthermore, the applicant indicates that the investment cost for such an installation might be about 100% higher than with other on-purpose technologies. The applicant did not provide concrete evidence for the cost side of this alternative, but to SEAC it does seem likely that availability is a serious issue especially considering the fact that Borealis has plans to significantly up their propene production in the next couple of years. Importing propene as an alternative poses similar issues of availability, especially considering the way the propene market is evolving. It is unclear if a sufficient and sufficiently reliable supply will be available over an extended period of time. What is however clear to SEAC based on the arguments provided by the applicant, is that importing propene would cut significantly into the profit margins of Borealis. While maybe not completely economically infeasible, SEAC does not find this to be a suitable alternative. International trade is said to be limited, usually done locally (e.g. Japan and 13 In other words, in addition to the 740 kt/year plant which is considered to be necessary because of the evolution of the propene market. 25

27 South Korea export to China) and often reflects imbalances in supply and demand. The rising popularity of integrated on-purpose propene production is a major consequence of this. Conclusion From the applicant s clear and relatively comprehensive assessment of suitability it is evident to SEAC that most of the discussed alternatives will not be technically and/or economically feasible at the sunset date. For some of the alternatives issues of availability are of serious concern. Two alternative technologies for the on-purpose production of propene, the Oleflex and Krupps-Uhde Star process, are however already used industrially and are therefore considered to be technically feasible from Borealis viewpoint. During the opinionmaking Borealis was investigating the feasibility of erecting a 740 kt/year Oleflex plant at their Kallo site. This leads SEAC to believe that economic feasibility might not be an issue in the very near future, if looked at in isolation and not considering Borealis specific situation. However, SEAC also took into consideration the market prospects and linked concerns regarding continuity of supply. Scrapping the Lummus plant and building additional capacity, i.e. a second 8 Oleflex plant, in its stead would not be an economically feasible alternative by the sunset date. 7.3 To what extent are the risks of alternatives described and compared with the Annex XIV substance? Description: Clariant Produkte GmbH described the efforts made to identify alternative substances, technologies and alternative sources of propene. The applicant provided a table including six substances (e.g. chromium (III), molybdenum trioxide, platinum, tin, divanadium pentaoxide, digallium trioxide), with the EC number, CAS registry number and the classification according to the entry in Annex VI, CLP Regulation and / or a short note on the notified classification according to CLP criteria. According to the applicant, all these substances would lead to risk reduction as the substances are not listed in Annex XIV, REACH. The applicant also stated that purchase and importation of propene would decrease the risks (to human and environment) as no Cr(VI) has to be used. Regarding potential alternative technologies (e.g. catalytic metathesis, fluidised catalytic cracking, biosourced technologies), the applicant stated that they have no access to information about these processes. Therefore they did not provide any conclusion on risk reduction capacity. 26

28 7.4 Would the available information on alternatives appear to suggest that substitution with alternatives would lead to overall reduction of risk? YES NO NOT APPLICABLE Justification: The applicant listed the hazards of six alternative substances needed for either a modified technology or an alternative technology but has neither presented any exposure scenarios nor any detailed risk assessment. Although human hazards posed by the presented alternative substances seem to be less dangerous than those posed by Cr(VI), the judgement on this issue is rather difficult due to the spare information provided. Besides, the applicant was not able to provide any information about risk reduction capacity of some of the alternative technologies as they do not have any access to this information. Conclusion RAC is not able to conclude on the risk reduction capacity of the described alternative substances respectively due to insufficient information on this issue. 7.5 If alternatives are suitable (i.e. technically, economically feasible and lead to overall reduction of risk), are they available before the sunset date? YES NO NOT RELEVANT Justification: 8. For non-threshold substances, or if adequate control was not demonstrated, have the benefits of continued use been adequately demonstrated to exceed the risks of continued use? YES NO NOT RELEVANT, THRESHOLD SUBSTANCE Justification: The applicant evaluates one potential non-use scenario and provides evidence that the benefits of continued use outweigh the associated risk. 27

29 Additional statistical cancer cases estimated by RAC The estimated number of additional statistical cancer cases has been calculated using the excess risk value presented in section 6 and the estimation of the number of exposed people (146 at the production site in Kallo, Belgium) provided by the applicant. It reflects the expected statistical number of cancer cases for an exposure over the working life of workers. The applicant considered in the health impact assessment exposure of the workers in nearby plants. The number of such workers is No ELRs were determined (also monetised values of those risks) for the general population as there, according to the applicant, are no residents within a radius of 1 km of the emission source. Inhalation and oral exposure at the regional level was not considered to have any significant impact and was therefore not further taken into account. Table 8: Estimated additional statistical cancer cases for workers directly and indirectly exposed (40 years of exposure) WCS Excess cancer risk Number of workers performing the activity Estimated statistical cancer cases (inhalation exposure) WCS 1 / PROC WCS 2 / PROC WCS 4 / PROC WCS 6 / PROC WCS 7 / PROC WCS 8 / PROC Total Indirectly exposed workers as a proxy for the general population , Assessment of Impacts The assessment of impacts which has been undertaken by the applicant includes a quantitative monetary assessment of the societal impacts associated with the non-use scenario (i.e. assuming authorisation is not granted). The assessment of impacts is based on impacts occurring within the EU and which are incremental to the baseline situation, these impacts being defined in terms of a non-use scenario in which the plant in Kallo would be scrapped and a new Propane dehydrogenation (PDH) propene production facility, using a non-chromium catalyst, next to the current plant, would be built. 28

30 The assessment of economic impacts undertaken by the applicant is based on a wellestablished benefit-cost methodology. SEAC cannot, however, fully agree with several assumptions made and values used since they lead to an overestimation (scarring cost) or underestimation (lost profit margin during importation of propene and necessary investment cost) of aforementioned externalities, although it shall be stressed that this does not affect final conclusions of SEAC. Costs of continued use (risks) The applicant carried out a quantitative human health impact assessment based on the estimated excess cancer risk of workers and the general population arising from the exposure to Cr(VI) from the use of chromium trioxide. Lung cancer has been identified as the main health endpoint associated with exposure to Cr(VI). Exposure values were calculated based on the use of 2.7 tonnes/year of chromium trioxide. The dose-response relationship published by ECHA was used in the applicant's assessment, assuming workers are exposed during 8 hours per working day over a work life of 40 years, and permanent exposure of 70 years for the general population. The applicant did not adjust the resulting risks to reflect the 12-year review period requested and produced worst-case estimates of additional statistical cancer cases among on-site workers ( cases for 12-years period) and of ( cases for a 12-year period) additional statistical cancer cases among the general population (1,820 workers in nearby plants). Using the values derived from Valuing Selected Health Impact of Chemicals: Summary of the Results and a Critical Review of the ECHA Study (ECHA, February 2016), the applicant used only upper bound of a value per fatal lung cancer cost (3.69 million EUR; this and all other values have been inflated to 2015 figures using an EU area GDP deflator). The average mortality rate for lung cancer was presumed to be 84.6% (source - survival statistics for lung cancer in Belgium). Morbidity risks were also evaluated setting the value of case at 432,800 EUR. Based on the above values, the applicant monetised the health impacts over a 20-year period (ignoring latency periods) and arrived at a net present value of 32,000 for both on-site workers and the general population. SEAC in general concurs with the approach followed and the values used. However, SEAC recalculated the monetised health impacts using the values from RAC s excess risk assessment on the estimated additional statistical cancer cases (table 8 above) arriving at a total human health cost of 23,000 calculated for an adjusted 12-year period. Benefits of continued use (cost of non-use scenario) The applicant evaluated a potential non-use scenario (NUS) whereby the applicant would be the stopping of importation and supply of the Catofin catalyst to Borealis in Europe. Manufacture of Catofin is located outside of the European Economic Area (EEA). The applicant states that, in order to secure the long-term supply of propene to its polypropylene plant at Kallo, it is assumed that its client - Borealis would construct a new PDH propene production facility, using a non-chromium catalyst, next to the current Lummus one, which it would scrap. 29

31 Based on the applicant s explanations, SEAC finds it credible that the applicant would not incur significant losses within EEA while its client - Borealis would be forced to construct a new PDH propene production facility. The applicant identified several possible negative economic impacts for Borealis related to the NUS: - New PDH Plant (data source publicly available information on similar projects throughout the world); - Lost profit margin during importation (market data provided by ICIS); - Unemployment; - Rehiring costs; - Scarring cost. Also, positive economic effects were identified: - Reservation wage; - Asset disposal. Negative economic impacts (NPV) associated with a non-granted authorisation as estimated by the applicant are 451 million. In response to SEAC s questions and during the trialogue the applicant provided more detailed information with regard to: - Distributional effects providing answer on what part of propene would be supplied from the outside of EEA if authorisation is not granted; - Risk on the possibility to be left without any propene and consequent loss in case if authorisation is not granted; - Borealis plans on building world-scale (740 kt/year) propane dehydrogenation plant that would be located at the existing Borealis production site in Kallo; - Borealis estimate of future market balance - demand and supply and subsequent needs for additional production capacity SEAC considers the total welfare impact as assessed by the applicant to be: - slightly overestimated regarding scarring cost since Borealis most likely will rehire the same personnel as soon as it will be economically and assumption that employees would earn less is not clearly justified; - somewhat underestimated regarding lost profit margin during importation. The applicant has used average polypropylene margin under different propene supply chains for the time period from 2009 to 2016 while it can clearly be observed that applying values for the time period 2014 to 2016 potential lost margin during importation may be up to 45% higher than reflected on the SEA. In response to the SEAC s question the applicant answered that Borealis would not expect any propene to be sourced from within the EEA. The most likely source locations are North America and the Far East. The applicant also stressed numerous times during the opinion development process that such imports would be extremely difficult and risky logistically and financially, and Borealis would only attempt it because the alternative would be the immediate closure of the Kallo polypropylene plant. Whilst some uncertainty about the investment cost remains, SEAC is confident that the estimate provided by the applicant is representative and based on reliable and publicly 30

32 available information sources. SEAC believes that construction cost of new PDH propene production facility in Kallo as indicated in SEA may be even somewhat underestimated. Additional calculations made by SEAC indicate that the monetised risk and benefits of continued use to the applicant and Borealis are realistically assessed by the applicant. SEAC highlights that its reassessment does neither attempt to include the potential loss of profits incurred by the customers and suppliers of Borealis nor other wider economic impacts and is therefore likely to underestimate the welfare impacts of the non-use scenario. The applicant just states that due to several reasons these impacts are expected to be limited. Conclusion The applicant s analysis of the benefits of continued use (i.e. the costs of non-use) is based on the costs due to construction of the new PDH propene production facility in Kallo, lost margin during importation of propene (profits are transferred outside EEA) and redundancies resulting from the closure of the Borealis plant for the time period of three to four year. The applicant s assessment of the monetised risks to human health amounts to 32,000 for 20-year period. SEAC recalculated the monetised health impacts using the values from RAC s excess risk assessment on the estimated additional statistical cancer cases and arrived at a total human health cost of 23,000 calculated for a 12-year period. Based on the applicant s assessment, the benefits of continued use appear much larger than the associated risks to human health. In particular, the applicant reports that the net benefits of continued use are 451 million. SEAC in principle agree with the calculations made by the applicant and additional calculations made by the SEAC support the numbers provided by the applicant. SEAC considers that the implied benefit-cost ratio of more than 19,000:1 clearly and with a sufficient margin of error demonstrates that the benefits outweigh the risks of continued use. 9. Do you propose additional conditions or monitoring arrangements YES NO Description for additional conditions and monitoring arrangements for the authorisation: None Description of conditions and monitoring arrangements for review reports: None Considering that the implemented risk management measures and existing operational conditions are considered appropriate in limiting the exposures and the risk, neither additional monitoring arrangements nor conditions on review reports are considered necessary. 31

33 10. Proposed review period: Normal (7 years) Long (12 years) Short (. _years) Other: Justification: In identifying the review period SEAC took note of the following considerations: RAC s advice: RAC gave no advice on the length of the review period. Other socio economic considerations The applicant considers that their AfA provides sufficient justification for a review period of 12 years. SEAC can agree with this assessment. Borealis is the company that actually uses CT and therefore SEAC s assessment of the AfA mainly takes on their viewpoint. By focussing on Borealis, a reassessment of the criteria for a long review period becomes necessary. Borealis is currently doing a feasibility study on the construction of a 740 kt/year Oleflex plant at their Kallo site. The current aim is to start producing, in parallel with the Lummus technology, in 2021 at 80% capacity and 4-5 years later at full capacity. Borealis has stated during the trialogue that there is not a short term intention for shutting down the Lummus plant itself considering the need for sufficient supply of propene (see also section 7.2). Clariant, however, clearly indicated during the same trialogue that over the next years it, in close cooperation with Borealis, will continue research to develop a chromium-free drop-in alternative for the Lummus plant in Kallo. This is considered to be a less costly, although more timeconsuming, option than scrapping the Lummus plant and building a second Oleflex plant in its stead. SEAC agrees with this reasoning and finds this option also preferable from society s perspective. SEAC assessed the criteria for a long review period in relation to this application while taking on the viewpoint of Borealis. The costs of using the alternatives are very high and very unlikely to change in the next decade SEAC considers the recurring costs to be similar and affordable, at least for the Oleflex dehydrogenation plant. This was also confirmed by Borealis during the trialogue. The building of a new plant is a high one-off investment cost, however the applicant finds it plausible due to the expected growth for polypropylene. Additionally, there is also decrease in propene output due to steam crackers and refineries shifting to alternate input sources. The consequent shift to more on-purpose propene production is therefore considered to be a necessity. This is exemplified by Borealis strong intent 32

34 to build an Oleflex plant more than 3 times bigger than the current Lummus one, however, to fulfill the projected demand for polypropylene the applicant states that production at the currently used Lummus plant needs to continue in parallel with the new Oleflex plant. Recent investments made by Borealis in order to extend longevity of Lummus plant seem to support this view. Scrapping the currently used plant and building a second Oleflex plant to replace it would cost several hundred million euros (on top of the planned 740 kt/year plant!). According to SEAC, a more beneficial scenario from society s perspective would be to let Clariant and Borealis find a suitable drop-in alternative for the Lummus plant (see section 7.2 A). However, such an alternative does not exist at the moment, and developing it might take 10+ years according to the applicant. As such, SEAC finds that this criterion is met. SEAC does wish to note that the projections for the future need of propene (and consequent demand in polypropylene) cover only a 10-year period. SEAC did not find this to be problematic when deciding on the review period. The applicant can demonstrate that research and development efforts already made, or just started, did not lead to the development of an alternative that could be available within the normal review period. As stated multiple times already, alternatives are available and have a high market penetration. For example, the Oleflex technology has an equal market share as the Lummus technology. This criteria is therefore not met according to SEAC. The possible alternatives would require specific legislative measures under the relevant legislative area in order to ensure safety of use (including acquiring the necessary certificates for using the alternative). This criterion is not relevant according to SEAC. The remaining risks are low and the socio-economic benefits are high, and there is clear evidence that this situation is not likely to change in the next decade. SEAC considers this criterion to be met. The applicant s impact assessment was considered by SEAC to provide robust conclusions in this respect. Weighing all the above considerations SEAC finds a long review period justified and consequently recommends a review period of 12 years. 11. Did the Applicant provide comments to the draft final opinion? YES NO 11a. Action/s taken resulting from the analysis of the Applicant s comments: YES NO NOT APPLICABLE 33

35 Annex. Press release by Borealis Media Release Vienna, Austria 22 September 2016 Borealis to study feasibility of a new, world-scale propane dehydrogenation plant in Belgium Borealis, a leading provider of innovative solutions in the fields of polyolefins, base chemicals and fertilizers, will study the feasibility of a new, world-scale propane dehydrogenation (PDH) plant. The plant would be located at the existing Borealis production site in Kallo, Belgium. The feasibility study will be carried out over the next nine months. The final investment decision is expected to be taken in the third quarter of 2018, while the potential start-up of the plant is scheduled for the second half of The new PDH plant would have a targeted annual production capacity of 740 Kiloton per calendar year, making it one of the largest and most efficient facilities in the world. The Borealis Kallo location has been chosen due to its excellent logistical position and its experience in propylene production and handling. Borealis has selected Honeywell UOP s Oleflex technology for the new plant. This technology is widely used and is a reliable and sustainable choice for on-purpose propylene production. A new PDH plant of this scale would be a significant investment for Borealis in Europe. It would strengthen our long term commitment to be the innovative polypropylene and propylene supplier that is meeting the needs of our customers today and in the future, says Markku Korvenranta, Borealis Executive Vice President, Base Chemicals. During the coming quarters we will be engaging with the value chain partners and authorities to work out the commercial and operational details of the project. In Europe propylene demand is increasing while the supply growth from steam crackers and refineries is slowing down. With the market tightening, an on-purpose propylene investment is needed to ensure a reliable platform for continuous, long-term growth in propylene and its derivatives in Europe, explains Thomas Van De Velde, Borealis Vice President, Hydrocarbons & Energy 34

36 Borealis Hydrocarbons & Energy Flexible, integrated steam crackers as well as the PDH plant and their associated production units, form the backbone of the Borealis olefins and polyolefins portfolio in Europe. Borealis sources basic feedstocks such as propane, naphtha, butane and ethane from the oil and gas industry and converts these into ethylene, propylene and other high value hydrocarbons. Note to editors In an on-purpose propylene production route, propane is selectively dehydrogenated, which means that hydrogen is removed from the molecule in a chemical process, to become propylene. Propylene is also a by-product from the steam cracking of liquid feedstocks such as naphtha or liquefied petroleum gas (LPG), and from off-gases produced in refineries. Photo: Borealis production location in Kallo, Belgium Borealis END For further information please contact: Kerstin Artenberg Vice President HR & Communications Tel. +43 (0) (Vienna, Austria) kerstin.artenberg@borealisgroup.com Borealis is a leading provider of innovative solutions in the fields of polyolefins, base chemicals and fertilizers. With headquarters in Vienna, Austria, the company currently has around 6,500 employees and operates in over 120 countries. Borealis generated EUR 7.7 billion in sales revenue and a net profit of EUR 988 million in The International Petroleum Investment Company (IPIC) of Abu Dhabi owns 64% of the company, the remaining 36% belonging to OMV, an international, integrated oil and gas company based in Vienna. Borealis provides services and products to customers around the world in collaboration with Borouge, a joint venture with the Abu Dhabi National Oil Company (ADNOC). 35