In some regions, the traditional

Towards a zero gasoline production refinery: part 1 Integrating products from the steam cracker, aromatics complex and FCC unit to produce petrochemicals without gasoline may offer more attractive margins BLASIS STAMATERIS Foster Wheeler In some regions, the traditional export markets for gasolinefocused refineries are diminishing. In this environment, investments in new refining projects (or in significant refinery upgrades) need to be optimised to fulfil market requirements. An opportunity exists to integrate the value chain from crude oil processing in refineries to the production of petrochemicals in line with market requirements. The objective of this article is to present alternative refinery configurations that are able to process relatively heavy crudes, producing middle distillates, petrochemicals and aromatics, without producing any gasoline at all. The article will demonstrate how, by relying on well-proven refining process technologies, configurations could be adapted in an existing FCC-based refinery or can be applied to a grassroots project. It will show how to integrate products from the steam cracker, aromatics complex and FCC unit while rationalising investments. Part 1 of the article presents ways of upgrading streams produced in refineries that have traditionally been oriented towards the production of transportation fuels to produce petrochemicals, which may offer more attractive margins, allow diversification of product slates and reduce the impact of refined product market volatility. Streams from the FCC unit offer building blocks towards propylene and aromatics production. Part 2, to be published in a forthcoming issue of PTQ, will show the impact of these processes on the overall product slate of the refinery for different refinery configurations. FCC unit: not just a gasoline-making machine In addition to gasoline, the refinery FCC process also produces one-third of the global propylene supply. yields from the FCC unit when operating in petrochemical mode can be at least 10 wt% of the feed. In order to maximise propylene production in the FCC unit, the following technical aspects need to be considered: 1 Feed quality: hydrogen content of the feedstock strongly correlates to the propylene yield Increased riser outlet temperature (ROT) and high catalyst-to-oil ratio: increased severity yields higher conversion and a higher propylene yield Catalyst system: use of ZSM-5 zeolite, which converts the C 7 olefins into light olefins with high catalyst Micro Activity Testing (MAT) activity Hydrocarbon partial pressure in the reactor: shifts the reaction equilibrium to favour low molecular weight olefins. This is achieved though a low operating pressure and the addition of steam, and should be balanced with increased plant costs due to larger vessel requirements. is produced in the FCC unit by cracking of olefinic to lighter olefins. The cracking reactions that take place in the initial reaction step on the lower section of the riser with the feed and hot regenerated catalyst are endothermic. The catalyst supplies the necessary heat to reaction temperature. The riser is no more than a straight pipe, the diameter of which is set to provide the feed with a certain residence time. Other converting reactions occur in a later step in the middle or upper section of the riser. In the sequential reactions, olefins that are initially produced from cracking reactions are consumed by subsequent secondary reactions yielding iso-paraffins and/or aromatics. The reactions producing light olefins are controlled by an equilibrium mechanism and thermodynamics limit the propylene production from the FCC unit. Table 1 summarises the FCC unit s operating conditions as compared with steam cracking. As Table 1 shows, cracking in the FCC/DCC units occurs at moderate temperatures compared with steam cracking, which makes Parameters Units FCC unit Deep catalytic cracking Steam cracking Residence time Seconds 1-3 10-16 0.1-0.2 Catalyst/oil wt/wt 5-10 9-15 - Steam wt% of the feed 1-10 10-30 30-80 Reaction temp C 510-550 550-590 760-870 Pressure KPa 15-30 10-20 15 Table 1 FCC unit operating conditions as compared with steam cracking 2 www.eptq.com PTQ Q3 2013 11

Products FCC, wt% Steam cracking, wt% Ethylene 1.5-4.5 25-30.8 12-17 17-14 s 14.4-17 About 12 s 65.4-51.2 11.3-4.7 Table 2 Yield comparison of cracking in FCC units versus steam cracking 4,5 the process very efficient from an energy consumption perspective. Coke that is deposited in the catalyst is burnt to regenerate the catalyst and provides the heat required by the cracking reaction. Also, since the feedstocks employed, such as gas oils and residues, are relatively cheap compared to steam cracking (traditionally fed in Europe), the process can be economically attractive. In addition, the DCC unit requires very clean (desulphurised, low metals content, low Conradson carbon) feedstocks with a high hydrogen content. The increase in the ROT brings increased production of olefinic liquefied petroleum gas (), dry gas and coke. Ethylene produced in the FCC unit (typical yield <2 wt% of the feed) could be recovered instead of using it in the fuel gas system. The addition of ZSM-5 zeolite, with its characteristic pore size, which provides shape selectivity by limiting access to the interior of the catalyst to mostly linear non-branched paraffin and olefin molecules, gears the resulting equilibrium distribution of the C 3 and lighter olefins towards propylene, the olefin product with higher yields. Ethylene is also produced, but its yield is largely dependent on reaction conditions, as mentioned above. The reaction chemistry and the use of ZSM-5 catalyst favour conversion of the olefinic molecules in the C 7 -C 10 range to olefinic. This depletes the catalytic of olefins, which, along with the fractionation of the light catalytic (LCN) typically composed of C 5 -C 6 molecules, contributes to the high aromatics content of the heavy catalytic another synergy that can be exploited for aromatics production. 3 Processing of light catalytic (LCN) The direct cracking of C 5 -C 6 -type molecules contained in the LCN to form light olefins requires initial dehydrogenation to form olefins that can then be cracked through olefin reaction pathways and require high severity (temperatures of about 650 C) and a high catalystto-oil ratio. So, because of the selectivity of ZSM-5 zeolites to crack larger molecules, the cracking of the lighter molecules is regarded more as a thermal cracking process. Coke make when cracking LCN is low. Therefore, in order to achieve a significant increase in propylene production, large amounts of LCN need to be processed. One option is to process the LCN into the same riser as the main feedstock. In this case, the LCN would be injected slightly below the main feedstock for it to detect a high catalyst-to-oil ratio. This option, although relatively low in cost, is not recommended, as the relatively large amounts of required to significantly increase the propylene yield (say by about 2%) would cool down the catalyst 1 Butadiene saturation Mixed s Selective hydrogenation 1 s processing into FCC. Direct recycle vs. oligomerisation. Simplicity vs. selectivity Mixed s from refinery and petrochemicals Butadiene extraction Raffinate-1 Total hydrogenation Steam cracker feed/ Butadiene MTBE decomposition Isobutylene Raffinate-2 Butene-1 fractionation Butene-1 Co-monomer for polyethylene production Butene-2 fractionation Metathesis feed Figure 1 s processing options 12 PTQ Q3 2013 www.eptq.com

and the lower temperature would not allow complete vaporisation of the main feedstock, leading to unnecessary coke formation and deposition in the feed zone. Another option is to use a separate riser, where the temperature and catalyst-to-oil ratio can be optimised and where the processing of the LCN does not interfere with the cracking of the main feedstock. In this case, the increase in propylene yield is still relatively small (about 2 wt% on top of the one achieved with the main feedstock), with low conversion per pass, a high gas yield and a great portion of the olefins produced being further converted due to secondary reactions into aromatics. Another option to dispose of the LCN is to send it along with the straight-run light to a steam cracker. Table 2 compares the yields that can be obtained when cracking on a steam cracker versus in a separate riser on a FCC unit. The trends are similar in both cases. As the severity increases, the ethylene yield increases; the propylene yield in the steam cracker decreases with increased severity, whereas it increases in the FCC unit. yields remain about the same, but the yield decreases with increased severity, hence the once-through conversion of on the steam cracker is higher than on the FCC unit. From a yields perspective, disposing of the LCN through the steam cracker versus a separate riser in the FCC could yield better economic returns, but this needs to be further investigated for each specific case. Processing s There are several commercially proven ways to upgrade s produced from the FCC unit and steam crackers. 6 Figure 1 illustrates some of the options. Although the butadiene content of s from the steam cracker is significantly higher than that of the FCC unit, the potential to recover butadiene is lost when the light recovery section of the steam cracker and FCC unit are integrated, because of the dilution effect that the s from the FCC unit have on the combined s stream. The main options for the production of on-purpose propylene are: Direct recycle of s to a separate FCC riser Via metathesis Oligomerisation combined with recycle of the oligomerate to be processed in a separate riser of the FCC unit. The direct recycle of the s cut involves a secondary high-severity riser parallel to the main riser designed to upgrade the s into ethylene, propylene and catalytic. Some fuel gas is also produced. The products from both risers merge at the reactor outlet and travel as a common stream to the main fractionator. An advantage of this process is that selective hydrotreatment for the diene conversion of the s stream as well as the removal of 14 PTQ Q3 2013 www.eptq.com

Mid/heavy PyGas Light reformate C 5 Raffinate Aromatics / gasoline HT CCR Medium reformate C 6 /C 7 Reformate splitter Mixed aromatics Benzene/ toluene splitter C 7 C 8 Light ends Benzene Toluene Reformate stabiliser Raffinate Heavy reformate C 8 C 7 Toluene column C 8 Xylene loop C 8 Mixed xylene Purify Trans alkylation Benzene mixed xylene Paraxylene Light ends FCC HCN HT Aromatics extraction Xylene column C 9 C 9 C 10 C 11 C 6 CCR HT HCN Continuous catalytic reformer Hydrotreater Heavy catalytic Heavy aromatics Figure 2 Configuration of the aromatics plant impurities is not required. Also, as with the cracking of the LCN, the aromatics content of the catalytic increases significantly compared with the catalytic produced from the cracking of vacuum gas oil (VGO) and residue feedstocks, which helps increase the production of benzene, toluene and xylenes (BTX). 7 From the investment perspective, it is a relatively low-cost option, but the cracking towards propylene is not selective. In the metathesis case, which is essentially an equilibrium disproportionation reaction between two olefins, the n-butenes react with ethylene to produce mainly propylene and other byproducts. The vapour phase reactions take place in a single fixed-bed reactor. The per-pass conversion is greater than 60%, with overall selectivity to propylene exceeding 90% when the feed is rich in 2-butene. This is a simple process to upgrade the value of n-butenes to high-value propylene. The paraffins pass through the system as inerts and, once recovered as non-reactive light materials, can be sent along with heavier -type products to a steam cracker. In order to maximise propylene yields, the s need to be selectively hydrogenated to convert the butadiene into butenes and to isomerise the isobutylene and 1-butene into 2-butenes. The isobutylenes isomerisation reaction is equilibrium limited with a conversion of isobutylenes of about 62% and selectivity to n-butenes of about 90%. To maximise the propylene yield, the remaining isobutylene should be removed from the metathesis feed either through fractionation or methyl tertiary butyl ether (MTBE) decomposition (if pure isobutylene is to be produced) because the isobutylenes react with the available n-butenes competing with ethylene in the production of propylene. In the recycle of oligomerate, the idea behind the process is that oligomer cracking, with a catalyst that has ZSM-5 zeolite additive, is more selective, resulting in higher propylene yields when recycled to extinction than the recycle and cracking of s in a separate FCC riser. 8 The oligomerisation process can be applied to the s from the steam cracker and FCC units. The reaction steps are: Convert s olefins into C 8 -C 12 oligomers in a poly unit Recycle the oligomers into a separate riser in the FCC unit to selectively crack them into propylene. The separate riser allows adjustment of the operating conditions to maximise propylene yield through a high reaction temperature and high catalyst-to-oil ratio. As an alternative, since the poly unit could be run to produce 100% gasoline-type material or 30% gasoline/70% distillates, this option offers the flexibility to increase the production of kerosene-type material, which, after hydrotreating to saturate the olefins, would have a relatively low freezing point (-60 C) and high smoke point (>33 mm). With this option, the increase in propylene yield would be 3-5 percentage points over the propylene yield provided by cracking the heavy feed into the main riser, depending on the amount of oligomers recycled. www.eptq.com PTQ Q3 2013 15

The incremental capital investment required by the poly unit is moderate and its economic justification competes with the direct recycle of the s to a separate riser. However, the process is selective towards the production of propylene mainly because of the operating conditions and catalyst additive (ZSM-5) used to crack the oligomers. Configuration of the aromatics plant The configuration of the aromatics complex, which is used to convert and pyrolysis gasoline (pygas) into BTX, includes the following process technologies: Continuous catalytic reformer (CCR) for the production of aromatics from at high severity Extractive distillation for the recovery of benzene and toluene Paraxylene purification for the recovery of paraxylene by continuous adsorptive separation from a mixed xylenes stream C 8 aromatics isomerisation for the isomerisation of xylenes and the conversion of ethylbenzene Transalkylation for the conversion of toluene and heavy aromatics to xylenes and benzene. In order to exploit potential synergies with the refinery and steam cracker, the heavy FCC, which is highly aromatic and naphthenic, as well as the pygas, are attractive feedstocks for aromatics production. Coupling aromatics production with a high olefin yield FCC maximises the value added from the FCC unit; with a aromatics content of >55 wt%, FCC gasoline becomes a desirable aromatics feedstock. On the other hand, pygas composition varies widely with the type of feedstock being cracked in an ethylene plant. Light steam cracker feedstocks tend to produce a pygas that is rich in benzene but contains almost no C 8 aromatics. Substantial amounts of C 8 aromatics are found only in pygas from ethylene plants cracking feedstocks. All pygas contains significant amounts of sulphur, nitrogen and dienes that must be removed by two-stage hydrotreating before being processed in an aromatics complex. Figure 2 shows the configuration of the aromatics plant. The salient features of the configuration of the aromatics complex are: Hydrotreat the straight-run to remove sulphur and nitrogen compounds and send it to a CCR, where paraffins and naphthenes are converted to aromatics Stabilise the reformate product and send it along with the pygas to a reformate splitter column. The C 7 fraction from the overhead is sent to the extractive distillation unit for extraction of benzene and toluene. The C 8 fraction from the bottom of the reformate splitter is clay-treated and then sent directly to the xylene recovery section of the complex Hydrotreat the heavy catalytic to remove impurities and send to a separate aromatics extraction unit. The extract is combined with extract from the reformate and sent to the BTX fractionation section to recover high-purity benzene and toluene products Send the raffinate from the heavy catalytic aromatics extraction unit to the straight-run to remove traces of extraction solvent and convert the naphthenes into aromatics in the CCR, whereas the raffinate from the reformate aromatics extraction unit is used as feedstock for the steam cracker plant The rest is a conventional aromatics plant, which includes a C 8 aromatics isomerisation unit, paraxylene purification unit, a heavy aromatics column and transalkylation unit, where toluene is blended with C 9 and C 10 aromatics (A9-) from the overhead of the heavy aromatics column and processed for the production of additional xylenes and benzene, which are recovered in the BTX fractionation section. The incorporation of a transalkylation unit into the aromatics complex allows the C 9 aromatics of the heavier straight-run/catalytic with an end point of above 170 C to be converted into additional xylenes. Overall, the steam cracker efficiently converts light paraffins and rejects aromatics and unconverted naphthenes in pygas, whereas an aromatics unit efficiently converts naphthenes and efficiently recovers aromatics, but rejects light paraffins in fuel gas, light ends and raffinate streams, allowing exploitation of the composition of the streams to maximise the production of valuable products while minimising the investment cost. Integration of refining and petrochemicals The drivers for integration between refining and petrochemical facilities have been extensively discussed elsewhere. 9 While olefins require further integration into polyolefins or other olefin derivatives, since they are not readily transportable, the aromatics are directly marketable. The key is to have refining units, steam crackers and aromatics complexes on the same site. From the revenue perspective, having integrated refining and petrochemicals complexes minimises the impact of the volatility of crude oil price, the primary driver of petrochemical costs and prices. 10 From the pricing perspective, it is observed that: Price differentials between and gasoline are relatively narrow, compared with and aromatics and and ethylene or propylene The spread between these price differentials has been increasing over time. Many new refinery projects, including some of those under development in China and the Middle East, include integrated refinery and petrochemicals complexes, addressing growth needs in the Asia Pacific, Middle East and European markets. The other aspect that needs to be considered for refinery/petrochemical integration to be successful is the yields versus capital required, since it affects the revenues and the profitability of the facility and, in the case of the steam cracker, the yields are heavily influenced by feedstock selection. Figure 3 shows the yields for steam cracker feeds versus FCC and CCR for petrochemicals. It shows that cracking gives a far wider range of products than gas-based steam cracking. From the 16 PTQ Q3 2013 www.eptq.com

Ethylene Butenes n-c 2 (ethane) n-c 3 (propane) n- (normal butane) i- (ISO butane) Light AGO (atmospheric gas oil) FCC (fluid catalytic cracker) CCR (continuous catalytic reformer) BTX 0 10 20 30 40 50 60 70 80 90 Figure 3 Yields for steam cracker feeds vs FCC and CCR for petrochemicals capital investment perspective, cracking is also far more complex. However, it makes sense when integrated with refineries that produce aromatics, because of the synergies that can be exploited with other facilities within the refinery, such as: Hydrogen produced in the steam cracker in the refinery, the CCR and in steam reforming can be used to supply the requirements of the hydrotreating units The raffinate stream produced in the aromatics complex is a perfect steam cracker feedstock: light material, low in aromatics, high in paraffins The pygas produced in the steam cracker can be processed in the aromatics complex to recover BTX components. So, liquid-based steam crackers can be economically competitive because of the credits obtained with the byproduct slate that the -based cracker produces, which maximise the revenues. Another aspect of integration that leads to lower investments but needs careful consideration during design is the use of combined light ends recovery facilities between the steam cracker and the FCC unit. In terms of investment, there are different levels of integration. For instance, the minimum would be to send off-gas from the FCC unit to Ethane Steam cracker C 3 Ethylene Hydrogen fuel shared utilities Light ends Petrochemical FCC Wet gas Unsaturated gas plant MIxed s to complex Crude VGO FCC gasoline C 5 Refinery Gasoline Fuel oil Figure 4 Integration opportunities to maximise olefins production 18 PTQ Q3 2013 www.eptq.com

the steam cracker to recover ethylene or use a common propylene-propane splitter for full integration with only the front end of the FCC unit (consisting of reactor, regenerator and main fractionation) with overhead sent to the wet gas compressor, whereas the olefins unit water quench tower goes to the cracker gas compressor sharing the product recovery systems. 11 Figure 4 shows examples of streams that are produced in refineries that can be used on steam crackers such as and/or the use of common facilities for the recovery of ethylene and propylene. In summary, refinery and petrochemical integration: Allows the upgrade of low-value streams to high-value products Minimises the cost of petrochemical feedstocks, since they are readily available from the refinery Provides stability over the value creation chain by diversifying the product slate, which dampens cyclic profitability impact Reduces hydrogen production in steam reformers by recovering the hydrogen produced by the steam cracker and catalytic reformer Optimises capital, operating costs and resources through shared infrastructure for utilities supply, off-sites (tankage allows transfer of refinery intermediate products to petrochemicals, common flare, wastewater treating facilities) and infrastructure (buildings, laboratory), leading to lower investments. The second part of this article will demonstrate that the full integrated scheme leads to significant savings in investment and operating costs, but has a lot of design challenges to guarantee the operability of all process units. FCC in a secondary riser of the FCC unit for maximum propylene production, Fuel Processing Technology, 2008 (89), 864-873. 5 Ethylene, Chemsystems PERP program, PERP 08/09-5, Sept 2009. 6 Kantorowicz S, processing options to upgrade steam cracker and FCC streams, 2nd Asian Petrochemicals Technology Conference, Korea, 7-8 May 2002. 7 Niccum P K, et al, KBR catalytic olefins technologies provide refinery/petrochemical balance, 25th JPI Petroleum Refining Conference, Recent Progress in Petroleum Process Technology, 26-27 Oct 2010, Tokyo. 8 Dupraz C, (R)FCC product flexibility with FlexEne, ARTC 2012, Bangkok. 9 Allen A, Refinery/petrochemical integration: past, present and look into the future, Offshore World, 29 Dec 2007-Jan 2008, 29-34. 10 Scott A, et al, Using microeconomics to guide investments in petrochemicals, McKinsey on Chemicals, No 4, Spring 2012, 47. 11 Dharia D, et al, Catalytic cracking for integration of refinery and steam cracker, Advances in Fluid Catalytic Cracking, CRC Press, 2010, 119-126. Blasis Stamateris is Downstream Business Consultant in the Business Solutions Group of Foster Wheeler, Reading, UK. He has over 25 years experience in the oil refining and upgrading business, and holds a degree in chemical engineering. References 1 Knight J, Mehlberg R, Maximise propylene from your FCC unit, Hydrocarbon Processing, Sep 2011, 91-95. 2 Kayode Coker A, Modelling of Chemical Kinetics and Reactor Design, Gulf Professional Publishing, 237. 3 Bedell M, Ruziska P A, Steffen T R, On purpose propylene from olefinic streams, Tulane Engineering Forum, Sept 2003, 3-4. 4 Wang G, C Xu, Jinsen G, Study of cracking www.eptq.com PTQ Q3 2013 19

Towards a zero gasoline production refinery: part 2 Refinery configurations can suit various processing objectives, such as variations to the propylene-to-ethylene ratio and production of middle distillates and aromatics BLASIS STAMATERIS Foster Wheeler, UK DAN GILLIS Foster Wheeler, USA The objective of this article is to present alternative refinery configurations that are able to process relatively heavy crudes, producing middle distillates, petrochemicals and aromatics, without producing any gasoline at all. By relying on well-proven refining process technologies, the configurations could be adapted to an existing fluidised catalytic cracker (FCC)-based refinery or used in a grassroots project. The article illustrates options for integrating products from the steam cracker, aromatics complex and FCC unit while rationalising investments. Part 1 (PTQ, Q3 2013) presented ways of upgrading streams produced in refineries that have traditionally been oriented towards the production of transportation fuels to petrochemicals. These solutions may offer more attractive returns and allow for diversification of a refinery s product slate. The possibilities offered by the streams from the FCC unit in terms of propylene production and aromatics production were also illustrated. The benefits of refinery and petrochemical integration were highlighted, in particular, diversifying the product slate while optimising capital costs, operating costs and resources through fully integrated production and shared infrastructure for utilities supply and offsites. This second part outlines the impact of these processes on the Crude oil Crude distillation unit Atmospheric residue ARDS SR Middle distillate splitter Heavy Light Steam cracker Benzene Aromatics complex Paraxylene Fluidised Raffinate catalytic cracker Pygas Pygas Pygas and catalytic Ethylene Hydrotreated residue C 2 C 3 S processing Residue fluidised catalytic cracker Diluted crude oil S Light cycle oil To sales Ethylene To fuel oil sales Figure 1 Refinery configuration with ARDS RFCC www.eptq.com PTQ Q4 2013 17

Crude oil Crude distillation unit Atmospheric residue SR Light vacuum gas oil Vacuum distillation Heavy unit vacuum gas oil Vacuum residue Middle distillate Hydrocracker Vacuum residue desulphurisation splitter diesel Hydrotreater residue Heavy Light C 2 C 3 Steam cracker S Residue fluidised catalytic cracker Benzene Aromatics complex Paraxylene Fluidised Raffinate catalytic cracker Pygas Pygas Pygas and catalytic S processing Ethylene diesel To sales Ethylene Diluted crude oil Light cycle oil To fuel oil sales Figure 2 Refinery configuration with HCU VRDS RFCC product slate of the integrated refinery/petrochemical complex for different configurations. Refinery configurations A wide range of refinery configurations for the production of petrochemicals, aromatics and transportation fuels has been studied. Example material balances, developed with a linear programming model, are presented to illustrate ways to vary the product slate (C 3 =/ C 2 = ratio, aromatics, middle distillates) and to demonstrate the effect of s processing on the production of propylene while optimising capital investment. The basis for the cases presented here are: Crude processing: 15 million t/y, about 300 000 b/d Middle Eastern-type crudes of around 28 API; capacities of process units and/or product slates can be adjusted to process heavier crudes (17-22 API) Complex configuration with ARDS RFCC This configuration is typically used for maximising gasoline and, more recently, propylene yields from atmospheric residues. For heavy sour residues, such as those being considered in this study, pretreating the residue fluidised catalytic cracker (RFCC) feed in an atmospheric residue desulphurisation (ARDS) unit is required to maximise conversion to the desired products. Figure 1 shows the schematic of a refinery configuration, where an ARDS unit is followed by the RFCC unit. The key feedstock properties that impact the performance of an RFCC unit are: Conradson carbon residue (CCR): CCR is the main feed quality indicator that affects the heat balance of the reaction section, in particular the regenerator temperature, which can also affect the performance of the catalyst (high temperatures produce catalyst sintering with a negative impact on yield pattern). The higher the CCR content, the higher the coke make and the higher the regenerator temperature. For residues with a CCR content under 8 wt%, the RFCC unit could be equipped with catalyst coolers to control the regenerator temperature Metals (mainly nickel and vanadium): these metals, especially vanadium, reduce catalyst activity. Additionally, nickel will nonselectively crack the feed to undesirable light fractions. The higher the metals content in the feed, the higher the catalyst make-up to maintain the high MAT activity required to maximise the propylene yield. (Note: Micro Activity Test (MAT), defined by ASTM Procedure D-3907, is widely used to determine the relative activity and selectivity of FCC catalysts for conversion of a standard feedstock.) Hydrogen content: This affects the RFCC feed s conversion and 18 PTQ Q4 2013 www.eptq.com

yield selectivity, in particular the propylene yield. The higher the hydrogen content, the higher the propylene yield. For the types of crude oil that have been considered here, the typical CCR content of atmospheric residues is about 12-14 wt%, and metals are 80-90 wtppm. Thus, pretreatment of the feed is required. The ARDS unit is a specially designed that pretreats the feed to the RFCC unit to reduce contaminants including sulphur, nitrogen, organometallic metals and CCR. Additionally, it increases the hydrogen content of the atmospheric residue, thus improving the crackability and selectivity of the residue. For a typical ARDS HDS rate of 90%, the CCR reduction is about 65-70% and the metals reduction (HDM rate) is about 85-90 wt%. The HDT RFCC feed will then have a CCR content of less than 6 wt% and metals content of less than 10 wtppm. Due to the feed rate, nature of the feed, and processing objectives, a relatively high pressure and low space velocity design is required compared to other hydroprocessing technologies. Even so, the cycle length of the ARDS unit is typically only one year. For the case evaluated and ARDS operating objectives, at least two reactor trains, each with three to five very large and thick-walled reactors, will be required. Consequently, both the initial investment and on-going catalyst replacement cost will be much higher than for most hydroprocessing units. The other key features of this configuration to maximise the production of petrochemicals are: Process straight-run light along with light catalytic (LCN) in the steam cracker Process heavy straight-run in a catalytic reforming unit (CRU) to produce reformate Process heavy catalytic (HCN), pygas and reformate through the aromatics complex to produce paraxylene and benzene The s with relatively high olefins content are either sold or recycled to the second riser of the FCC (directly or through oligomerisation) to maximise propylene production Products from the RFCC unit and steam cracker are combined in a single light ends recovery section to produce ethylene and propylene, minimising the capital investment Hydrogen from the CRU and steam cracker is supplemented by further hydrogen production from steam reforming to be used in the s. Complex configuration with HCU VRDS RFCC The intention of this refinery configuration is to increase the production of middle distillate at the expense of petrochemicals. Relative to the first configuration of ARDS RFCC, the LVGO fraction of the crude is routed to a VGO hydrocracker (HCU) to shift the yields in the desired direction. Figure 2 shows a schematic of the refinery. The key features of this refinery configuration are: A vacuum tower is installed to produce a diesel-type cut to be processed in the distillate, a LVGO cut to be processed in a two-stage, almost full-conversion hydrocracker unit oriented towards the production of middle distillates The HVGO is combined with vacuum residue to be pretreated in the vacuum residue desulphurisation (VRDS) unit before processing it in the RFCC unit Depending on the quality of the combined vacuum residue and HVGO feed to the VRDS unit, a slip stream of vacuum residue could be taken and blended with diluents such as RFCC decanted oil (DCO) and light cycle oil (LCO) to produce some fuel oil. The VRDS unit would then be able to reduce the CCR to the levels required by the RFCC feed, especially when the CCR content of the feed to the VRDS unit exceeds 24 wt%. Complex configuration with SDA DCU HCU RFCC The refinery configurations illustrated above, where residue s reduce the feed 20 PTQ Q4 2013 www.eptq.com

contaminants to levels required for the RFCC unit to operate in an economic manner, while increasing the hydrogen content of the feed to maximise high-value products, tend to be capital intensive and have high operating costs because of: The difficulty of removing contaminants such as organometallic and nitrogen compounds in the heavier asphaltenic compounds contained within vacuum residues. This results in relatively high pressures (160-200 bar(g)) and relatively low space velocity (0.1-0.2 hr -1 ) The need to replace large volumes of catalyst in the hydrotreating unit (typically once a year) The RFCC unit requires a two-stage regenerator, catalyst cooler(s) and CO boiler, adding to the cost of the unit A high catalyst make-up on the RFCC unit to keep a high MAT activity to favour the propylene yield Additional investments required in tankage to store hydrotreated feed to continue running the RFCC while the ARDS/VRDS catalyst is changed in each reaction train because of the comparatively longer RFCC cycle length versus ARDS or VRDS units. An alternative solution would be to have a refinery configuration with unit processes capable of separating the components of the vacuum residue into two fractions: one more easily hydrotreated, which would then be processed in a more conventional FCC unit, and another where the impurities such as polynuclear aromatics (PNA), metals, CCR and asphaltenes could be concentrated then processed in a non-catalytic thermal conversion process such as a delayed coking unit (DCU). Solvent deasphalting (SDA) is an attractive solution for this type of application, as it separates residues by molecular type. An SDA unit is a robust, relatively low-cost residue separation process that uses an aliphatic solvent (typical light paraffinic solvents with carbon chains of three to five) to separate the more valuable oils and resins from the aromatic and asphaltenic components of its residue feedstock. Contaminant in DAO, % Sulphur CCR Nickel Vanadium Figure 3 DAO contaminants versus DAO yield The paraffinic solvent precipitates the more polar, higher molecular weight components, such as resins and asphaltenes, typically called pitch, from the higher-quality, relatively low contaminant and higher hydrogen content components called deasphalted oil (DAO). The heavier the solvent, the higher the DAO yield and the higher the level of contaminants it will contain. This process has relatively low capital and operating costs, as it operates at relatively low pressures and temperatures, and no catalysts are used. A well-designed SDA unit, such as the UOP/Foster Wheeler Solvent Deasphalting Process, is highly reliable and will typically have very long run lengths between planned shutdowns for inspections and general maintenance. Figure 3 shows the impact of DAO yield on the percentage of feed contaminants in the DAO. For most residues and from a hydrotreating processing perspective, the contaminants in the DAO are relatively low, up to about 70 wt% of DAO, vol% Option: VGO Light DAO Heavy DAO VR % of VR N/A 35-50 50-75 100 Feed CCR, wt% 0-1 5-10 10-15 18-28 Feed C 7 insolubles, wt% 0 <0.02 <0.05 5-30 Feed C 5 insolubles, wt% 0 <0.1 <0.3 20-30 % HDS 95 90-95 85-92 85-90 LHSV, hr -1 0.75-1.5 0.3-0.5 0.2-0.4 0.1-0.2 Pressure, bar(g) 70-100 100-125 120-140 160-200 Catalyst cycle, months 24 24 18-24 12 Table 1 Residue hydrotreating options DAO yield. Although not shown in Figure 3, even at DAO yields of 70 wt%, the DAO will be relatively free of asphaltenes, which are a significant factor in setting the s severity. When an SDA is integrated with a DCU, the combination allows DAO from the SDA unit and heavy coker gas oil (HCGO) from the DCU to be sent along with the VGO from the vacuum distillation unit to the FCC feed hydrotreating unit and subsequently to a conventional FCC unit. Foster Wheeler has taken this concept and optimised the integration of SDA with delayed coking (see Figure 5). In this type of flow scheme, the various streams are routed to where they are most effectively processed. Another synergy that can be exploited with this configuration is to eliminate the production of slurry oil from the FCC unit by processing it in the DCU. This also has a side benefit of decreasing the coker feed s viscosity, which would be relatively high if only the SDA pitch was sent to www.eptq.com PTQ Q4 2013 21

SDA HL coking SDA ML coking 20% 25% 50 7 2 3 70 7 6 14 DAO VGO Distillate Coking 30% 29 29 14 0 5 10 15 20 25 30 35 ML = Medium lift 50% Coke, wt% HL = High lift 70% 0 20 40 60 80 100 Liquid products, wt% Figure 4 Optimised SDA DCU integration the DCU s heater. Likewise, unconverted oil from the VGO hydrocracker can be selectively separated in the SDA unit. This option would typically be used when DAO is processed in a VGO hydrocracker. Due to the improved quality of the feed to the, its severity will be much lower in comparison to the residue hydrotreating configuration operations. Table 1 shows the relative ranges of process conditions when comparing residues and VGO hydrotreating options. The benefits of this process configuration are illustrated in Figure 4; the combined liquid product yields increase as the SDA lift increases. (DAO yield is increased.) Also, the coke yield decreases as the SDA lift is increased. Although there are two residue processing units with SDA DCU, the overall capital cost is similar to the cost of the DCU alone due to the relatively low cost of a SDA Coke Coke drums Heater SDA pitch multiring aromatics compared to a DCU, and because the capacity of the DCU is lower when combined with the SDA unit versus a standalone DCU. Figure 6 shows a schematic of a refinery configuration that uses SDA DCU instead of VRDS RFCC to upgrade residues, eliminating the operational/catalyst management challenges of a VRDS unit and a complex reactor/regenerator section associated with an RFCC unit. The key features of this refinery configuration are: A vacuum tower is installed to produce a diesel-type cut to be processed in the distillate, and a LVGO cut to be processed in a two-stage, almost full-conversion hydrocracker unit oriented towards the production of middle distillates. As an alternate case, the impact of a partial-conversion HCU, whose unconverted oil is fed directly into the FCC unit, is also presented The HVGO is combined with Coker fractionator Heavy HCGO Coker feed Heavy coker gasoil DAO SDA extractor Residue feed Gas Light coker gasoil HDT/FCC Hydrocracking HCU unconverted oil (UCO) FCC slurry oil (CSO) Figure 5 Integration between SDA and DCU: liquid yields versus coke production 1 DAO and the HCGO to be pretreated in a conventional VGO HDT unit before its processing in the FCC unit The vacuum residue is processed in the SDA unit, and the pitch combined with the FCC slurry is processed in the DCU If the facility is located where low-cost natural gas is not available, the coke instead of being exported could be used to generate steam/electricity in circulating fluidised bed boilers (CFBB). Processing of mixed streams: metathesis Metathesis is essentially used as a way to adjust the propylene-toethylene product ratio of the refinery, thus upgrading the value of butenes to high-value propylene, improving economics and consuming ethylene in addition to butenes. In order to maximise propylene production, the key to processing olefinic streams through metathesis is to prepare the feed such that it is rich in 2-butenes, because every molecule of ethylene that reacts with a molecule of 2-butene gives two molecules of propylene, whereas all of the other olefins give only one molecule of propylene and other by-products, competing with 2-butene for the ethylene available. The challenge is that the components have close boiling points and are difficult to separate through fractionation (see Table 2). Therefore, other means have to be used to prepare the rich 2-butene stream for metathesis. Several processing routes have been proposed to upgrade the olefinic streams. 2 The processing steps to be taken to include 22 PTQ Q4 2013 www.eptq.com

Crude oil Crude distillation unit Atmospheric residue Vacuum residue Vacuum distillation Heavy unit vacuum gas oil Solvent deasphalting delayed coking unit Coke Circulating fluidised bed boilers SR Light vacuum gas oil Light coker gas oil Heavy coker gas oil Deasphalted oil Middle distillate Hydrocracker Vacuum gas oil splitter diesel Hydrotreated feed Heavy Light C 2 C 3 Steam cracker Fluidised catalytic cracker Diluted crude oil S Raffinate Pygas S Light cycle oil Aromatics complex Fluidised catalytic cracker Pygas and catalytic processing Pygas Benzene Paraxylene Ethylene diesel To sales Ethylene To fuel oil sales Power Steam Figure 6 Refinery configuration with SDA DCU metathesis in the refinery configurations are shown in Figures 1, 2 and 5: Extract butadiene from steam cracker s and selectively hydrogenate the raffinate-1 or selectively hydrogenate the whole olefinic stream to reduce the butadiene content to less than 10 wtppm. In this case, the catalyst system and operating conditions of the selective hydrotreating unit are adjusted to isomerise 1-butene to 2-butene Remove iso-butylene through MTBE decomposition: in this case, separation by reaction takes advantage of the fact that only the iso-butylene reacts with alcohols to produce ethers, which can be converted back to high-purity iso-butylene. This step could also be required if high-purity 1-butene is desired (for instance, as comonomer for the production of polyethylene) Removal of 1-butene and lighter components: this process consists of two super fractionators. First, the stream is fed to the first column to remove as a bottoms product the n-butane along with the bulk of 2-butene and some 1-butene. Second, from the top, the bulk of 1-butene and other light components are taken to the second column, where iso-butane is removed with some 1-butene and Component Normal boiling point, C Iso-butane) -11.7 1-butene -6.3 Isobutylene (2=-methyl propene) -6.9 1-3 butadiene -4.4 n-butane -0.5 Trans-2-butene 0.9 Cis-2-butene 3.7 Table 2 Normal boiling point of olefins sent to the steam cracker. The bottom product is high-purity 1-butene If high-purity iso-butylene and/ or 1-butene are not required, only the first super fractionator would be required to produce the feed to the metathesis unit. For the rest of the s, there is potential to further increase the production of ethylene and mainly propylene through direct cracking or oligomerisation, followed by cracking in a separate riser in the (R)FCC unit. Results Table 3 summarises the material balances produced for each of the configurations shown in Figures 1, 2 and 5. The results show that, regardless of the refinery configuration, when deep conversion units are considered and under certain operating conditions designed to maximise the production of petro- www.eptq.com PTQ Q4 2013 25

Product ARDS ARDS RFCC HCU VRDS HCU HCU SDA/DCU RFCC with RFCC SDA/DCU DAO HDT metathesis DAO HDT/FCC FCC partial conversion HCU metathesis s, KTA 1151 974 792 625 551 Gasoline, KBPD 0 0 0 0 0 Jet fuel, KBPD 15.6 15.6 23.5 36.2 27.4, KBPD 128.8 128.8 147.3 131.6 128 Fuel oil, KBPD 7.2 7.2 6.1 1.3 1.3 Ethylene, KTA 1123 907 1081 1106 975, KTA 1100 1558 838 800 1312 Paraxylene, KTA 1600 1587 1573 1754 1830 Benzene, KTA 490 501 509 563 582 Coke 348 353 Table 3 Material balances for different complex configurations chemicals, world-scale production of ethylene, propylene and aromatics with crude processing of 15 million t/y about 300 000 b/d can be produced. Choosing between the ARDS RFCC, VRDS RFCC and SDA DCU configurations is not easy, as all can be adjusted to produce similar amounts of distillates, petrochemicals and aromatics. For the refinery configuration with ARDS RFCC, both units are relatively large. The capacity of ARDS and RFCC are 121 000 b/d and 108 000 b/d, respectively. Two cases were developed: one where all of the s are sold and another one to show the impact a metathesis-based olefins conversion unit (OCU) has on the production of ethylene and propylene. As can be observed, in the first case, both ethylene and propylene production exceeds 1.1 million t/y. In the second case, where an OCU has been incorporated, propylene production is increased by about 40% at the expense of ethylene. In both cases, there is potential to further increase the production of ethylene and propylene through direct cracking or oligomerisation of the s, followed by cracking in a separate riser in the RFCC unit. Paraxylene production is world scale and is likely to require two trains of paraxylene recovery. The refinery configuration with VRDS RFCC shows the contribution the HCU brings to the increase in middle distillates production, compared with the ARDS RFCC case, at the expense of propylene production, which decreases mainly because of the reduced RFCC unit capacity (in this case, about 65 000 b/d). For the refinery configuration with SDA DCU, two cases are presented to show the flexibility this configuration offers to adjust the product slate to suit market requirements: a configuration oriented to maximising middle distillate production with a full-conversion HCU, and another where the unconverted bottoms of a 70% conversion HCU are fed to the FCC unit and a metathesis unit is incorporated to maximise the propylene production. Both SDA DCU cases show [US$ million] ARDSRFCC VRDSRFCC SDADCU ARDS/VRDS 601 465 - RFC44 318 - HCU - 371 371 VGO HT - - 235 FCC - - 169 SDADCU - - 360 Total 1045 1154 1135 Table 4 Order of magnitude total installed cost (TIC) estimates higher middle distillates production than the ARDS RFCC configuration can achieve. In the first case, propylene production decreases mainly due to reduced FCC capacity, which in this case was about 59 000 b/d as compared to the ARDS RFCC configuration. The flexibility of the SDA DCU configuration in terms of product slate is shown in the last case. Where the partial-conversion HCU is feeding the unconverted bottoms to the FCC unit, the capacity of the unit increases to 76 000 b/d. This, plus the metathesis, allows for an increase in propylene production. In relation to paraxylene production, since the amount of straight-run material is the same for all cases, the contribution of the produced in mid- and deep conversion should be analysed to understand the variations in paraxylene production. It should be noted that besides (R)FCC capacity, the feed quality also has a significant influence on the catalytic yield. An RFCC unit processing a full range of treated residue feedstock with a relatively high content of resins and asphaltenes in the feed, as in the ARDS or VRDS RFCC configuration, cannot be expected to produce the same yields as those that can be obtained in an FCC unit with a VGO DAO deeply hydrotreated feed. The other key differentiator is the quantity and aromatics content of the heavy (R)FCC. It would be highly aromatic and naphthenic because of the high severity and the use of additives required to maximise propylene production. The aromatics content of the heavy catalytic produced from the RFCC would be slightly higher than that produced from an FCC unit under similar operating conditions. Fuel oil production is significantly reduced, as expected. Coke can be either exported or used on site to produce steam/electricity in a circulating fluidised-bed boiler. Impact of capital investment costs With regards to capital investment costs, Table 4 summarises the order 26 PTQ Q4 2013 www.eptq.com

of magnitude total installed cost (TIC) estimates that have been prepared, based on Foster Wheeler s in-house cost estimating database. As can be seen, the investments for the upgrading section of the refinery configurations are of the same order of magnitude. With similar product slates and investments, the differences in economic indicators between the refinery configurations will not be significant enough to clearly favour one configuration over another. Therefore, when it comes to the decision of which refinery configuration to choose, other factors besides the product slate and investments, such as operating complexity and associated reliability of the units, should be taken into consideration. Ultimately, the selected option is based on an individual refiner s objectives. The SDA DCU options represent well-proven technologies. If there is a suitable outlet for the coke produced, this is a viable and competitive configuration. Certainly, if the objective is to maximise revenues with proven technology, in a reliable environment, the configuration that has been developed by Foster Wheeler utilising its SDA DCU combination offers a competitive advantage. How to make the right decision The integration of refining and petrochemical facilities offers several opportunities for adding value to refinery streams by exploiting synergies between different process units while rationalising investments that can positively impact project economics. Different refinery configurations can be tailor-made to suit a range of processing objectives, including variations to the propylene-toethylene ratio, and the production of middle distillates and aromatics. Capacities of the process units can be adjusted to accommodate different crude oils. The concepts outlined are applicable to grassroots projects and potentially to the revamp of existing facilities. The zero gasoline production refinery is a practical and viable reality. References 1 Gillis D, Unique opportunities with proven technologies to maximise residue conversion & refinery margins, Asia Technology forum, Bangkok, 10-11 Oct 2012. 2 Edwards S M, et al, Relative economics of mixed s processing routes, PTQ, Q1 1998, 1. Blasis Stamateris is Downstream Business Consultant in the Business Solutions Group of Foster Wheeler, UK, participating in refinery configuration studies, feasibility studies and FEED work for grassroots refineries, CTL facilities and refinery upgrades. A graduate in chemical engineering, he has over 25 years experience in oil refining and upgrading. Email: blasis_stamateris@fwuk.fwc.com Dan Gillis is Director, Refining for Foster Wheeler, US. He provides technical direction and applications development for Foster Wheeler s proprietary heavy oil technology. He is an engineering graduate of the University of Saskatchewan, Canada. Email: dan_gillis@fwhou.fwc.com www.eptq.com PTQ Q4 2013 27