Assessment of cleaner process options: a case study from petroleum refining.

Weston, Neil ; Clift, Roland ; Basson, Lauren 等

INTRODUCTION

The Paradigm Shift in Chemical Processing

Over roughly the last two decades, there has been a shift in both thinking and practice away from the "clean-up" or "end-of-pipe" technology approach which had replaced the "dilute-and-disperse" approach to emissions from process plants, to approaches to process selection, design and operation which have the objective of avoiding rather than managing emissions and wastes (e.g. Clift and Longley, 1995; Allen and Rosselot, 1997; Sikdar et al., 1999). In North America, this new paradigm is generally termed Pollution Prevention, defined as "source reduction and other practices that reduce or eliminate the creation of pollutants through increased efficiency in the use of raw materials, energy, water or other resources; or protection of natural resources by conservation" (Sikdar et al., 1999). In Europe, the term Clean Technology is more commonly applied, defined as "a means of providing a human benefit which, overall, uses less resources and causes less damage than alternative means with which it is economically competitive" (Clift, 1995). In this context, "overall" means over the whole life cycle of a product or a process; that is, over the whole supply chain of materials and energy from the "cradle" of primary resource to the "grave" of long-term managed disposal. This attention to complete life cycles is also implicit in Pollution Prevention and other conceptual approaches such as Cleaner Production (see Clift, 2001).

The shift in assessing environmental performance away from the individual process to a broader system boundary encompassing complete life cycles requires, inter alia, new approaches to identifying best available technology (BAT) for any product, process or service (e.g. Geldermann et al., 1999; Nicholas et al., 2000). The parameters describing environmental performance form a set of categorically distinct environmental impacts; any attempt to reduce them to a single metric such as damage cost is unsatisfactory for a number of reasons, one of which is that aggregation to a single objective hides essential information about the origins and nature of the impacts. Consideration of complete life cycles may require trade-offs to be recognized between impacts arising in different parts of the life cycle. In turn, this means that defining the boundary delineating the product or process system is critical. To deal systematically with trade-offs between different environmental impacts and economic costs, process selection, design and operation must use multi-objective optimization (see e.g. Clift, 2006) and decision-structuring and analysis approaches appropriate for complex contexts (see e.g. Basson and Petrie, 2007). This complication and complexity mean that simple decision rules may not be universally applicable and may even be misleading. This applies even to simple waste hierarchies--"Prevention, Minimization, Recycling, Disposal" (e.g. Finnveden et al., 2005) or "Elimination, Source Reduction, Recycling, Treatment, Disposal" (e.g. Crittenden and Kolaczkowski, 1992; Balm et al., 1997)--which are contested for some wastes; the debate over whether paper waste should be recycled as a recoverable material or burned as a renewable biofuel is a well-known example (e.g. Ekvall and Finnveden, 2000; Arena et al., 2004; Hart et al., 2005).

Three levels of change have been recognized in modifying a process to produce a particular product by a cleaner route (Allen, 1997):

1. Minimize arisings of waste and effluent and consumption of energy within an essentially unchanged process;

2. Modify the process or technology to use the same materials but more efficiently and with less energy;

3. Redesign the process completely and change input materials.

Substitution of the product by another way of delivering the same function is a more radical step (Clift, 2001), not considered in this paper. Process change at level l is essentially incremental, corresponding to changes in process configuration or operation which are low-risk and for which the return is readily calculable. Level 2 changes are usually matters of routine process engineering; some financial risk may be involved, so that return on capital is of concern, but there is limited uncertainty. Changes at level 3 usually involve significant capital investment, with uncertainty over financial performance and possibly over the environmental impacts of changing the supply chain of materials and energy.

These levels of change are also associated with different barriers and drivers, and therefore require different approaches to decision-making (e.g. Basson, 2004; Mitchell et al., 2004). Incremental, level 1 changes are matters of routine process evolution. Level 2 changes can be assessed in terms of simple parameters such as payback time or return on capital but these approaches must be adapted to include consideration of environmental impacts unless these impacts are "internalized" into the economic calculations, for example, as emission charges or traded emission permits. Changes at level 3 are more contentious and risky, so that more sophisticated approaches to decision-making are required (see e.g. Wrisberg and Udo de Haes, 2002). The barriers to adoption of new cleaner technologies have been the subject of several careful studies (e.g. Christie et al., 1995; Clayton et al., 1999) which dispel the neo-classical economic myth that a technology which meets demand more cheaply and complies with regulations will necessarily be adopted. A general conclusion is that technological developments do not proceed by smooth linear progression; rather, "lock-in" to past technologies and investments constrains the paths which are available for future developments (e.g. Arthur, 1989; Stirling, 1998). Occasionally major technological shifts do occur, as the development of ammonia processes illustrates, but the more usual pattern of development is by changes at levels 1 and 2--incremental changes or small steps, but guided by the longer term and larger scale changes driven by changes in public expectations as well as market conditions.

Sustainability Considerations

The agenda which is embodied in the term "sustainable development" requires assessment of the social benefits of a product or service as well as its economic and environmental performance (see e.g. IChemE, 2002; Mitchell et al., 2004). Although established models of sustainability group technological and economic factors together, and it is accepted that many technological hurdles may be overcome with sufficient economic resources, technical and economic considerations are subdivided here for pragmatic reasons. The following considerations are relevant in assessing the cleaner process options discussed below. There are almost always trade-offs to be made between different aspects of performance, and this is one of the principal reasons why sustainability concerns introduce the need for structured decision-making processes (Elghali et al., 2008).

Technological

An important consideration in assessing feasibility is time taken to implement the change. This is especially true in the process industries where shutdowns must be scheduled to allow intrusive modifications to be carried out. Any modification requiring a shutdown must therefore be assessed in terms of its complexity so that a decision can be made as to whether it is a viable for completion within any forthcoming scheduled shutdown. Furthermore, reliability of any process modification is an important concern: a business is unlikely to implement a process change unless there is a high degree of confidence in the ability of the modified process to perform its function reliably. Modification time and reliability are treated here as technological concerns.

Economic

Economic considerations address the financial implications of carrying out a proposed modification. Unless a modification is forced by external factors (such as new regulations), a business will require some financial incentive for carrying out the work. Economic metrics used for assessing economic performance are routine: for example, net present value (NPV), which accounts for all the savings over a given timescale that will result from implementing the modification, and payback, which calculates the length of time required to recover the cost of installing the modification. These metrics show in principle that any size project is economically feasible if the cost savings from making the modification justify the expenditure.

Environmental

Environmental performance covers both use of materials or energy and environmental emissions to land, water or air. Qualitative considerations can guide identification of the most significant factors in assessing process options. However, full consideration of options requires quantitative analysis of the whole supply chain of materials and energy (see The Paradigm Shift in Chemical Processing Section above) to ensure that process changes do not merely displace environmental impacts to some other place in the life cycle.

Social

It is also necessary to consider the impact of process changes on different social groups affected, including local residents, the workforce and the general public. Performing this high level qualitative assessment of social factors indicates the extent to which different social groups are affected by the options being considered and thereby aids decision makers in determining the appropriate level of involvement in a deliberative decision-making process.

System Boundaries--There Is no such Place as "Away"

The discussion above introduced the importance of approaching environmental performance on a life cycle basis and of attention to strict definition of system boundaries. The appropriate tool for such analysis is Life Cycle Assessment (LCA). While the intention here is not to go into the details of LCA (1), it is necessary to introduce some of the considerations in drawing system boundaries for LCA. The system must include the full supply chain from primary resource extraction through to waste management: wastes cannot simply be "sent away". It is useful to make a pragmatic distinction, illustrated in Figure 1, between (Clift et al., 1999, 2000):

1. Foreground: the set of processes whose selection or mode of operation is affected directly by the change being considered;

2. Background: all other processes which interact with the Foreground, usually by supplying or receiving materials or energy.

The total Life Cycle Inventory for the product produced by the system in Figure 1 consists of:

1. Direct Burdens emitted from the Foreground processes

2. plus Indirect Burdens arising from the supply chains of materials and energy provided to the Foreground

3. minus Avoided Burdens associated with activities in the Background displaced by materials or energy recovered from the Foreground processes.

This distinction was originally introduced to help in compilation of the Life Cycle Inventory (Nord, 1995). Foreground processes are identifiable and can be represented by plant-specific primary data whereas Background processes are not specifically identified and are therefore represented by generic average data. If a transfer between the Foreground and Background takes place through a homogeneous market, then the precise source of the materials or energy is unknown and so the supply chain is necessarily in the Background.

[FIGURE 1 OMITTED]

The distinction between Foreground and Background systems requires careful consideration when assessing process changes at the different levels outlined above; treating emissions from known processes as direct emissions is of particular value in determining the specific consequences of modifications to the system. (2) Processes whose selection or design may be changed substantively must be treated as part of the Foreground system, because the nature of the environmental impacts associated with the process may be changed. Where only the amount of material (or energy) sent to or output from a process is affected, the scale but not the nature of the environmental impacts is changed; these processes can be considered part of the Background system. However, if the exchange is with a specific identifiable process, then that process can be represented by specific data; in effect, it can then be treated as part of the Foreground. As a specific example considered in the case study set out in this paper, treating waste on-site rather than exporting it for management elsewhere means that part of the waste treatment moves from the Background into the Foreground.

Aim and Outline of Paper

This paper describes developments in a specific process at a UK oil refinery, to provide an illustration of the methodological and practical challenges associated with the introduction of cleaner technologies. The section Existing Alkylation Process and Waste Management below introduces the process--alkylation--concentrating on the management of fluoridic waste arising from HF used as catalyst. The ensuing section Waste Reduction and Management Section introduces possible technological changes which could avoid, reduce or enable partial recycling of the waste, identifying the barriers and drivers associated with each option and focusing particularly on the way in which the system boundaries must be drawn to show the trade-offs to be considered in assessing them and examines how quantitative LCA should be carried out to compare the full environmental implications of the alternative changes. Discussion and Conclusions Sections discuss the more general conclusions to be drawn from this case study.

EXISTING ALKYLATION PROCESS AND WASTE MANAGEMENT

Alkylation is an operation in petroleum refining which produces a high value blend for transport fuels, alkylate, with a high octane rating; it contains no aromatic compounds or olefins and is "essentially free of sulphur and other impurities" (Hommeltoft, 2003). As fuel specifications have become more stringent and additives, primarily lead compounds, have been phased out, alkylate has become an increasingly important refinery product (Dunham, 2005). Alkylate is formed by the following overall liquid-phase reaction; see Albright (2003) for a fuller discussion.

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII.]

The detailed mechanism is still under study, especially amongst researchers seeking to establish alternative catalysts (e.g. Hommeltoft, 2003; Kane and Romanow, 2003).

An acidic catalyst is used to obviate the use of high temperature and pressure. Two processes are in use worldwide, using respectively hydrogen fluoride (HF) and sulphuric acid as catalyst (Wood et al., 2001; Nieto et al., 2007); globally, total outputs from the two processes are almost equal (DuPont STRATCO, 2006). There are relatively few configurations of alkylation units, the primary differences being between the major licensors: UOP and ConocoPhillips (previously Phillips) for HF units, and DuPont/STRATCO and ExxonMobil for [H.sub.2]S[O.sub.4] units.

Hydrofluoric Acid (HF) Alkylation

The refinery which is the focus of this case study uses the HF process. As a highly reactive compound, HF brings significant safety hazards that must be carefully managed. Losses of fluorine through various process streams exacerbate the safety hazards as they require increased deliveries of fresh HF and removal of spent fluoridic compounds which are wastes and must be treated prior to disposal. This section gives an overview of the alkylation unit to establish the main process characteristics and objectives and to identify the causes of fluorine losses. Waste Management Systems Section considers the current strategies at the refinery to deal with the fluorine-containing wastes, including the additional material and energy resources consumed by the waste treatment systems and the eventual form of the final waste.

[FIGURE 2 OMITTED]

Alkylation process

The system under consideration here is a Phillips-type HF alkylation unit, shown schematically in Figure 2; it has been in operation for nearly 30 years. The principal inputs to the process are liquid olefins and iso-butane (iC4). The olefins are distilled (A) to remove water and impurities, primarily ethane, which form by-products in the alkylation reaction while the iC4 is dried (B). The fresh olefin and iso-butane streams are mixed with recycled iso-butane and then contacted with HF acid. The reaction is exothermic and occurs instantaneously in the riser entering the acid settler (D); the mixture contains excess iso-butane so that the olefins react to completion. The hydrocarbon reaction products and HF catalyst are immiscible; they are allowed to settle and separate in the acid settler (D). The heavier acid layer contains free HF, water, organic fluorides (produced by incomplete reaction) and a group of heavy hydrocarbon by-products known as conjunct polymers or, more commonly, acid soluble oils (ASO) (Miron and Lee, 1963; Albright et al., 1988; Berenblyum et al., 2002). Make-up anhydrous HF (<0.01 wt% water) from the batch HF storage drum (C) is mixed with this stream (3); it is cooled and all but a small proportion is recycled to the reaction loop.

Controlling acid strength is important in operation of the alkylation unit for a number of reasons discussed below; therefore an acid regeneration loop is required to remove impurities and return clean acid. A small proportion (<0.1%) of the bottom stream from the acid settler (D) is partially vaporized before entering the acid re-run column (E) ; the liquid, comprising about 10% of this stream, contains ASO, water and HF. A hot stream of iso-butane vapour enters towards the bottom of the re-run column to strip HF from the liquid for return as a vapour to the acid settler (D). Water, some of the heavier organic fluorides and the ASO pass to the bottom of the acid re-run column (E) and are drained from the process. The acid re-run column is normally operated to maintain 1-2 wt% water in the acid loop; this level favours a high octane rating of the alkylate product (Hutson and Hays, 1977) while higher water concentration can lead to accelerated corrosion. Due to inefficient stripping and the formation of an azeotrope at 38.26 wt% HF in water (Munter et al., 1947), a significant part of the fluorine losses are in the form of HF in the acid re-run bottoms; this is considered in greater detail below.

The liquid hydrocarbon stream from the top of the acid settler (D) to the downstream separation columns (F & J) consists of alkylate, propane, n-butane, un-reacted iso-butane and some organic fluorides. Part goes to the depropaniser (F) to separate propane from iso- and normal butane. Propane vapour leaves the top of the depropaniser for further separation (G) and subsequent treatment. A two-phase side-stream from the depropaniser, containing a high proportion of iso-butane, is condensed and recycled to the reaction stage while a proportion is vaporized and sent to the acid re-run column. The liquid bottom stream from the depropaniser, mainly normal butane, and the rest of the top stream from the acid settler pass to the iso-stripper (J) where iso- and normal butane are separated from the alkylate, which leaves the bottom of the column as liquid product. This contains some (in the order of ppm) organic fluorides but is not defluorinated. (4) Iso-butane vapour from the top of the iso-stripper is condensed and joins the recycle stream from the depropaniser (F). A small two-phase side stream from the iso-stripper, containing primarily normal butane, goes for further treatment.

The normal butane and propane product streams require treatment to remove organic fluorides and trace HE This is carried out from vapour in the defluorinators (H and K) and from the liquid phase in the potassium hydroxide treaters (I and L). In each defluorinator, a bed of activated alumina is used to decompose the organic fluorides into olefins and free HF, which reacts with the alumina to form aluminium trifluoride (Al[F.sub.3]) (Blachman, 2000); this constitutes one of the main fluoridic wastes. In the KOH treaters, residual traces of HF react with solid KOH to yield water and potassium fluoride (KF) which is a further fluoridic waste. Units H, I, and K have two columns in parallel to enable one bed to be replaced while the other is on-line. Depending on specification, some of the n-butane may be added to the alkylate product, by-passing the KOH geater (L). A single column is used for L, replenished while the n-butane by-passes this treatment.

Not shown in Figure 2 but vital to the safe operation of the unit (and relevant to later discussion regarding fluorine losses) is the acid relief system whose primary function is to receive process fluids when disturbances cause an overpressure in the system. It is also used when a vessel is vented to create an inert atmosphere for safety reasons; this occurs immediately prior to refilling the acid feed storage drum (C), with smaller quantities vented when a defluorinator or KOH greater is replenished. Fluids entering the relief system are lost from the process, neutralized, and burned at the flare.

Fluorine losses

Fluorine is lost from the process as organic fluorides which leave with the product streams and as HF (plus traces of organic fluorides) in the bottom stream from the acid re-run column, limited by the water/HF azeotrope. In total, approximately 0.2-0.9 kg of HF are lost per cubic metre of alkylate produced. (5) Table 1 summarizes the routes by which fluorine is lost from the process with indicative figures for the contribution of each route. The main process parameters that lead to the losses will be introduced briefly here. The eventual fate of the losses is discussed in Waste Management Systems Section.

Re-run column bottoms. As described above, the re-run column maintains acid quality by removing water and acid soluble oils (ASO) using a hot stream of iso-butane vapour to strip HF and return it to the process. Approximately half the lost fluorine leaves in the stream from the bottom of this column (see Table 1). The three main factors determining this loss are:

i. water entering the alkylation process;

ii. ASO formation; and

iii. incomplete stripping in the re-run column.

The amount of water leaving the process via the re-run column and removing HF as an azeotropic mixture is equal to that entering the process from the dryers (A & B). A certain level of water in the process (~1-2 wt%) is favourable to achieving a high octane rating of the alkylate product (see above); this level is therefore a fixed operating parameter, controlled by the operation of the re-run column.

The polymerization reaction forming ASO is complex and unavoidable. As for water, some ASO in the process (~4-5 wt%) is favourable to a high octane rating of the alkylate and so operators seek to maintain it at this level. ASO represents the largest component of the fluids leaving the re-run column bottoms and the rate of its formation determines the rate at which fluids are drained from the re-run column; therefore the greater the rate of ASO formation, the greater the HF losses in the re-run column bottoms. ASO formation can be reduced by removing impurities in the olefin feed, a process which is already carried out at the refinery.

The stripping efficiency of the re-run column, measured by the proportion of HF recovered from the re-run column feed, is a function of its design, which determines phase-contacting efficiency, and operating conditions. Increasing temperature improves recovery but the operating temperature is limited by the allowable working temperature of the equipment.

Losses via product streams. The next most significant fluorine losses (see Table 1) occur in the form of HF and organic fluorides in the propane and butane product streams. Organic fluorides, which cause much the larger loss, are present in these streams because they have similar volatilities to the products; they are only present as a few ppm but this is significant because the product flows are large. Organic fluorides are formed in the alkylation reactions and require further contact with acid to progress the reaction towards the desired products. The availability of "acid sites" (i.e., protons) in the reaction solution, determined by the acid strength, is therefore a key influence on the residual concentration of organic fluorides. The acid strength in operation is determined by a number of factors including alkylate yield, quality and the discharge rates that can be effectively handled by the acid re-run column.

Losses to the relief system. Major disturbances that could lead to overpressure are avoided as a matter of course on any process unit since they create safety hazards and losses of valuable material. Minor disturbances, which cause momentary overpressures, occur more frequently; their causes can be complex so that they are difficult to avoid completely. Fluorine losses to the relief system, which are less significant than the other losses (see Table 1), occur whenever the HF storage drum is purged with nitrogen to remove HF to make the vessel safe for manually refilling with fresh HF, with smaller losses associated with replenishing the beds of alumina and KOH. Therefore losses to the relief system depend primarily on the total rate of HF loss and make-up.

[FIGURE 3 OMITTED]

Waste Management Systems

Three waste management systems are used to treat the fluorine losses:

* defluorinators: treat organic fluorides in propane and butane product streams;

* KOH treaters: treat free HF in propane and butane product streams;

* neutralization with sodium hydroxide (NaOH): treats all HF losses from the re-run column and relief system.

Referring to Table 1, approximately 25 % of the losses are treated by the defluorinators and KOH treaters while approximately 55 are neutralized by NaOH; the remaining 20% is lost as organic fluorides in the alkylate product. This paper concentrates on the wastes treated by NaOH because they account for more than half the losses, are the most problematic for the refinery and offer the greatest potential for process improvements.

Defluorinators

Once approximately 60% of the defluorinator bed has been converted to Al[F.sub.3], it is considered 'spent' meaning that it is no longer effective at removing organic fluorides. The bed must then be removed and replaced with fresh [Al.sub.2][O.sub.3]. Treatment of the spent defluorinator bed presents a waste management problem. This problem has been addressed by seeking uses for the Al[F.sub.3]. A partnership has developed with a company producing aluminium electrolytically from alumina ([Al.sub.2][O.sub.3]); Al[F.sub.3] is added to their process as a flux. The partner company collects the spent defluorinator beds and replenishes the defluorinator with fresh [Al.sub.2][O.sub.3].

KOH treaters

Water produced by the reaction between KOH and HF dissolves the other reaction product, KF, with some KOH. The liquid collects as a slurry at the bottom of the KOH treaters; it is drained off and sent for treatment at the refinery's waste-water treatment plant.

Neutralization with sodium hydroxide

The third treatment route for wastes from the alkylation unit is also the most problematic in terms of the volume of waste generated, additional materials required, manual operations involved and cost to the refinery. The current management system for wastes from the acid re-run column and relief system is shown schematically in Figure 3. These wastes are first treated at the refinery in separate batch wash vessels by neutralization with NaOH to produce sodium fluoride (NaF) and water. For the acid-rerun bottoms, the process also causes Acid Soluble Oils (ASO) to form a separate phase, removed from the top of the vessel for use elsewhere on the refinery as a fuel oil. NaOH is delivered as 50 wt% solution in water but is diluted to 4.5 wt%, using local river water, to prevent precipitation of NaF in the process (Science Lab, 2005). The two streams of spent fluoridic caustic (SFC) are blended for transport.

SFC is a hazardous and odorous slurry requiring further treatment before it can be disposed to landfill. It contains approximately 95 % water, NaF, unreacted NaOH and a variety of other impurities including mercaptans. Currently, it is transported approximately 400 km to a specialist waste management company where it is converted into a form suitable for final disposal. It is reacted with hydrogen peroxide ([H.sub.2][O.sub.2]) over an iron catalyst to remove odours, primarily mercaptans, and then mixed with other waste streams to precipitate the NaF along with other waste components. The precipitate is sent to controlled land-fill while the supernatant water is sufficiently clean to be discharged to sewer.

WASTE REDUCTION AND MANAGEMENT

Preliminary Identification of Alternatives Table 2 summarizes alternatives identified so far at the refinery to avoid, reduce or manage fluoridic wastes, classified by the levels of change introduced in The Paradigm Shift in Chemical Processing Section. The objective of each possible change is also summarized in Table 2, categorized according to the appropriate tier in the waste hierarchy (see The Paradigm Shift in Chemical Processing Section). It is notable that the level of the change in the waste hierarchy does not correlate with level in Allen's (1997) categorization of process changes; the waste hierarchy is useful in identifying possible process changes whereas the level of change provides a guide to the analysis needed to assess the change. The different process modifications are also associated with different drivers and barriers (see Appendix A).

[FIGURE 4 OMITTED]

Figure 4 summarizes level 1 changes, that is, incremental changes to reduce waste arisings by (a) changing operating parameters to reduce by-product formation or (b) concentrating the waste prior to transport to off-site treatment. Level 2 changes, that is, process modifications, are shown in Figure 5. As for the level 1 changes, these may be designed to reduce waste formation (c--equipment modification to improve the efficiency of separation in the re-run column) and to achieve more efficient management of the waste (d--regeneration of caustic and production of calcium fluoride as a saleable product). Figure G shows possible changes at level 3, that is substantial changes to the process or inputs to (e) regenerate the HF catalyst, (f) replace NaOH by KOH to reduce the volume of waste exported, and (g) eliminate HF completely by changing to a different catalyst.

[FIGURE 5 OMITTED]

Qualitative analysis (Table 3) shows that alternative (c) is clearly attractive; in fact, this modification is scheduled for implementation (see below). For the other possible modifications, quantitative evaluation is necessary considering the implications of these changes for the whole material life cycle, not restricted to refinery operations, that is, using Life Cycle Assessment. However, as introduced in the section on Sustainability Considerations above, this requires careful distinction between the Foreground and Background (sub) systems. Figures 4 to G show the Foreground and Background systems needed for this evaluation.

For both waste reduction and waste management, the higher level changes generally require broader Foreground systems. For changes at level 1, the Foreground is confined to the specific operation modified. Shifting from level 1 to 2 moves waste management into the Foreground--waste management is viewed as an additional process step, to be considered as part of the "design space" for process optimization. Some of these steps (e.g. acid regeneration, off-site by-product processing) may be operated by parties other than the refinery. In these cases, life-cycle based optimization would require discussion and possibly active collaboration with these parties, but may provide opportunities for economic, environmental and social benefits (see Introduction). This type of analysis helps to identify the stakeholders who will need to be considered and possibly involved in the decision, thereby paving the way for a systematic deliberative decision making process, such as that presented by Basson and Petrie (2007). Furthermore, identification of the key issues associated with each alternative can act as a precursor to risk assessments, such as the process environmental risk assessment (PERA) approach presented by Sharratt and Choong (2002).

[FIGURE 6 OMITTED]

We now consider these options in more detail to examine what decisions can be made based upon the qualitative assessment summarized in Table 2 and Figures 4 to G and the issues raised for the refinery in selecting a preferred alternative. In particular, the following are considered:

* What decisions can be made based upon the initial qualitative assessment;

* What dilemmas are presented within and between each dimension of sustainability.

Level 1 Changes

Superficially, incremental changes or additions to the process should be the most straightforward ways to reduce waste, not requiring a shutdown. Assessment of such changes must take a life cycle perspective but it suffices to limit the Foreground to the process itself (see Figure 4). However, process design and operation are usually already optimized within this restricted system boundary, so that incremental changes may require the most detailed quantitative analysis and justification.

Operational changes (a)

The factors affecting operation were outlined in Existing Alkylation Process and Waste Management above. Making adjustments to the operating conditions may appear to be the simplest way of reducing HF losses. However, the alkylation unit was designed to produce alkylate and has been operated to do this safely and at minimal cost. Environmental impacts are not normally included in optimizing plant operation, except where they are "internalized" through emission taxes, traded permits or feedstock costs. It is possible to use process modelling to make the trade-offs between economic and environmental performance explicit (e.g. Clift, 2006) but this analysis has yet to be carried out in this case.

Waste concentration (b)

Some of the impacts associated with off-site waste management arise from road transport of the waste; the waste, resource use and emissions are obvious, while the disruption associated with heavy traffic movements can dominate the social impacts. Reducing the volume of waste to be transported is therefore a driver to concentrate the spent fluoridic caustic (SFC) on-site by evaporation, to form slurry or perhaps even a dry solid. Given that the acid re-run bottoms is already neutralized in batches (see Neutralization with Sodium Hydroxide above), this modification should be possible without shutting down the alkylation process itself. However, this would required installation of new processing equipment (for which space may not be available) and also increase energy demand. Thus, although a rather simple process change in concept, it may be complex to evaluate and difficult or even impossible to implement.

Level 2 Changes

Design changes to alkylation unit (c)

A significant driver was that a shutdown was coming up, making internal changes to the alkylation unit feasible. This introduced a time constraint: it was not feasible to perform a full assessment of all the alternatives in time for the turnaround, so priority was given to alternatives which depended upon the unit being shutdown. Although this decision was made at an early stage, it is well supported by the qualitative analysis. Tackling the problem where it originates, the alkylation unit, offers the opportunity to reduce HF losses, thereby reducing both HF consumption and the downstream treatment needed.

HF losses from the re-run column can be reduced by modifying the column internals to improve stripping efficiency. The internals currently in use consist of four baffles or "shed-decks": sloped trays extending from the edge to the centre of the vessel causing the liquid to cascade towards the bottom of the vessel. Such phase contacting devices are relatively inefficient; they are used with fluids which tend to "foul" but this is not the case here.

The alternative internals include horizontal trays with fixed valves (no moving parts) and a "chimney tray" to encourage disengagement of gases from the acid feed; these are more efficient than the baffles currently used. Detailed design calculations showed that the improved stripping efficiency should reduce HF losses from the re-run column by about 32% (and overall HF losses by ~16%, losses from the re-run column accounting for approximately 50% of the losses). Savings from waste treatment and input materials mean that this project should achieve payback in less than 1 year.

Re-use of waste (d)

Moving along the waste hierarchy, re-use of the waste ("open-loop recycling", following LCA parlance) for another beneficial purpose or direct recycling within the process ("closed-loop recycling") should be considered after source reduction. Given the way the process operates, option (d) could in principle be implemented without a process shutdown. However, it differs from the level 1 options in the way LCA must be applied: converting the waste into an input to another process requires the waste processing to be brought into the Foreground with the Background extended to include the other process, as shown in Figure 5, to allow for the Avoided Burdens (see Sustainability Considerations Section).

In this specific case, there is a possibility of reacting the SFC with calcium hydroxide to regenerate the caustic soda, for example using the "HARD TAU reaction (Veolia, 2008),

2NaF + Ca[(OH).sub.2] [right arrow] Ca[F.sub.2] [down arrow] + 2NaOH

Calcium fluoride, precipitated by this reaction, could be used as a flux in metallurgical operations (Kirk and Othmer, 1994). Therefore economic credits would arise from any sales of this material, along with environmental "credits" corresponding to the Avoided Burdens; that is, the production of fresh calcium fluoride displaced. Imports of caustic soda should also be lower, reducing the Indirect Environmental Burdens (see Figure 1). Like the use of aluminium oxide and production of aluminium fluoride in the product defluorinators (see Defluorinators Section) this would be an example of Industrial Symbiosis: mutually beneficial cooperation between different businesses to reduce net use of materials and energy (see e.g. Clift, 2001).

The viability of re-using the waste in this way depends not just on economic assessment but also on technical viability: whether impurities--primarily those, including arsenic, entering with HF catalyst--would build up in the process or whether they would leave with the calcium fluoride. (6)

Level 3 Changes

Level 3 changes, involving substantial redesign of the process and/or changes in inputs, typically require the system assessed to be much broader (see Figure 6) so that the process itself ceases to be the sole focus for evaluation of technical, economic, environmental, and social performance.

Regeneration of HF catalyst (e)

Option (e) includes option (d) but would treat the precipitated calcium fluoride with acid to regenerate HF for recycling to the process as alkylation catalyst. Again, given that catalyst introduction and waste treatment are batch operations, this could in principle be implemented without a shutdown.

Qualitative considerations show obvious environmental and social benefits from this change, including those associated with greatly reduced transportation of HE Whether the change is technically viable and whether the trade-offs of capital and operating cost against improved environmental and social performance justify the change will need to be assessed by the same kind of process modelling as option (a) but with the extended system boundary shown in Figure 6 (e). This also illustrates a case where the economic measures can change over time: increasing landfill costs and charges will progressively favour this kind of process modification.

Neutralization with KOH (f)

A further possibility is to neutralize the acid re-run column bottoms with potassium hydroxide rather than sodium hydroxide. This has the benefit of greatly reducing the volume of waste to be transported and processed off-site, because KF has much higher solubility in water than NaF so that the resulting spent fluoridic waste can be much more concentrated. Compared to the other options, it is simple to implement. (7) It also introduces the possibility of common treatment of KF along with the waste from the existing KOH treaters. However, KOH is a less widely available commodity than NaOH--it is more expensive, and the Indirect Burdens associated with production and supply may differ from those associated with conventional caustic soda. Both the Foreground and Background for LCA must therefore be expanded, as shown in Figure 6 (f).

Change of catalyst (g)

A more radical change would be to use a different catalyst altogether. Currently the only proven alternative is sulphuric acid (see Existing Alkylation Process and Waste Management Section) and the trade-offs are such that there is no general preference for one process over the other (Chapin et al., 1985; Dunham, 2005). Therefore this options remains for longer-term consideration, probably dependent on commercialization of one of the less hazardous processes which are being developed.

Assessment of any new process will need to use the system shown in Figure 6(g). Catalyst production is treated as part of the Background because both HF and sulphuric acid are bulk commodities so that any change in use for this alkylation plant would represent no more than a marginal change in demand. This approach is common practice in both "attributional" and "consequential" LCAs (see Baumann and Tillman, 2004).

DISCUSSION

Decision Structuring

The case study shows that decisions over process modifications to improve sustainability (or, more accurately, to reduce unsustainability) must inevitably be taken in the face of significant uncertainties: over future technological developments, over costs (including the costs of waste treatment and disposal), over environmental impacts and their relative significance and over the significance attached to social impacts by different groups (including the media and general public). Companies like that operating the refinery which provides the case study for this paper already have procedures in place to make decisions over process changes, particularly where they involve significant capital expenditure. However, decisions involving trade-offs between incommensurable "costs" and "benefits" with considerable uncertainties require approaches to decision structuring which are unfamiliar in the process sector (Mitchell et al., 2004; Basson and Petrie, 2007; Elghali et al., 2008). The case study introduced here is being used as one of the vehicles to explore the use of new decision processes incorporating results from LCA.

Life Cycle Assessment

The use of LCA to evaluate the environmental performance of processes is becoming more routine. However, LCAs are notoriously time-consuming, requiring much detailed data collection. The pragmatic distinction between Foreground and Background sub-systems (Figure 1) is intended to provide a rational approach to limiting the amount of detailed work needed. The purpose of the case study discussed in this paper is to show how the LCA system boundaries must be tailored to the process and modifications being considered. In general, higher level changes require the sub-systems to be broader, but this must be considered on a case-by-case basis. Furthermore, the level of detail required in the process analysis does not necessarily correspond to the breadth required of the LCA: option (a), for example, needs the simplest LCA but a detailed process model.

Existing software packages for LCA are generally designed for application to products and services, rather than to chemical processes. There is therefore a recognized need for LCA packages which can be linked readily to process simulation packages. The case study introduced here shows why the Foreground and Background will have to be defined by the user even when such LCA tools become generally available. They must also be designed to show uncertainties explicitly, a feature which will differ from conventional process simulation. The next stages in this work will entail life cycle assessment of some of the options in Tables 2 and 3, both to assess their merits and trade-offs and to inform the use of LCA in the company operating the refinery.

CONCLUSIONS

Assessment of the sustainability of processes requires a range of considerations--environmental and social as well as technical and economic--which go beyond those familiar to chemical engineers. The case explored in this paper, cleaner production approaches to reduce fluorine use and waste in the alkylation process at a specific UK oil refinery, illustrates the need to broaden the issues and trade-offs considered in assessing processing and waste management options. Furthermore, it is necessary to extend the boundaries of the system analyzed beyond the immediate process to consider the complete life cycles of materials and energy, but the way in which this is best done depends on the level and nature of the change considered. Allen's (1997) classification of process changes provides a useful framework, and the case study shows how it can be used to guide definition of system boundaries for life cycle assessment of the environmental consequences of a process change. However, the level of process change is not related to the level of objective in the waste management hierarchy and is not a guide to the level of details and complexity needed in process analysis. Furthermore, the (financial) scale of a project is not a guide to the extent or significance of its environmental consequences, underlining the need for systematic sustainability assessment.

ACKNOWLEDGEMENTS

Neil Weston is registered on an Engineering Doctorate Programme at the University of Surrey; financial support from the UK Engineering and Physical Sciences Research Council (EPSRC) is gratefully acknowledged.

Manuscript received January 9, 2008; revised manuscript received February 18, 2008; accepted for publication February 22, 2008.

REFERENCES

Albright, L. F., "Alkylation of Isobutane With C3-C5 Olefins to Produce High Quality Gasolines: Physiochemical Sequence of Events," Ind. Eng. Chem. Res. 42, 4283-4289 (2003).

Albright, L. F., M. A. Spalding, C. G. Kopser and R. E. Eckert, "Alkylation of Isobutane With C4 Olefins. 2. Production and Characterisation of Conjunct Polymers," Ind. Eng. Chem. Res. 27, 386-391 (1988).

Allen, G., "Financial Investment and Environmental Performance," Phil. Trans. Roy. Soc. 355, 1467-1475 (1997).

Allen, D. T. and K. S. Rosselot, "Pollution Prevention for Chemical Processes," John Wiley and Sons, New York (1997).

Arena, U., M. L. Mastellone, F. Perugini and R. Clift, "Environmental Assessment of Paper Waste Management Options by Means of LCA Methodology," IEC Res. 43, 5702-5714 (2004).

Arthur, W B., "Competing Technologies, Increasing Returns, and Lock-In by Historical Events," Econ. J. 99, 116-131 (1989).

Bahu, R., B. Crittenden and J. O'Hara, "Management of Process Industry Waste," Institution of Chemical Engineers, Rugby (1997).

Basson, L., "Content, Compensation and Uncertainty in Environmental Decision Making," Ph.D. thesis. University of Sydney (2004).

Basson, L. and J. Petrie, "A Critical Systems Approach to Decision Support for Process Engineering," Comput. Chem. Eng. 31, 876-888 (2007).

Baumann, H. and A.-M. Tillman, "The Hitch-Hiker's Guide to LCA," Studentlitteratur, Lund (2004).

Berenblyum, A. S., L. V Ovsyannikova, E. A. Katsman, J. Zavilla, S. I. Hommeltoft and Y. Z. Karasev, "Acid Soluble Oil, by Product Formed in Isobutane Alkylation with Alkene in the Presence of Trifluoro Methane Sulfonic Acid. Part I Acid Soluble Oil Composition and its Poisoning Effect," Appl. Catal. A 232, 51-58 (2002).

Blachman, M., "Defluorinator Catalysts Evaluated," Hydrocarbon Engineering (2000).

Chapin, L. E., G. C. Liolios and T. M. Robertson, "Why Alkylation-HF or [H.sub.2]S[O.sub.4]?," Hydrocarb. Process. 64, 67-71 (1985).

Christie, L, H. Rolfe and R. Legard, "Cleaner Production for Industry," Policy Studies Institute, London (1995).

Clayton, A. M. H., G. Spinardi and R. Williams, "Policies for Cleaner Technology--A New Agenda for Government and Industry," Earthscan Publications Ltd, London (1999).

Clift, R., "Clean Technology--An Introduction," J. Chem. Technol. Biotechnol. 62, 321-326 (1995).

Clift, R., "Clean Technology and Industrial Ecology," in: "Pollution--Causes, Effects and Control," R. M. Harrison, Ed., Royal Society of Chemistry, London (2001), pp. 411-444.

Clift, R., "Sustainable Development and its Implications for Chemical Engineering," Chem. Eng. Sci. 61, 4179-4187 (2006).

Clift, R. and A. J. Longlet' "Introduction to Clean Technology," in "Clean Technology and the Environment," R. C. Kirkwood and A. J. Longley, Eds., Blackie Academic and Professional, London (1995), pp.174-198.

Clift, R., R. Frischknecht, G. Huppes, A.-M. Tillman and B. Weidema, "SETAC-Europe News," 10, 14-15 (1999).

Clift, R., A. Doig and G. Finnveden, "The Application of Life Cycle Assessment to Integrated Solid Waste Management: Part 1--Methodology," Trans. IChemE (Process Saf. Environ. Prot.) 78, 279-287 (2000).

Consoli, F., D. Allen, I. Bousted, J. Fava, W Franklin, A. A. Jensen, N. de Oude, R. Parrish, R. Perriman, D. Postlethwaite, B. Quay, J. Seguin and B. Vigon, "Guidelines for Life-Cycle Assessment: A 'Code of Practice'," SETAC Brussels and Pensacola (1993).

Crittenden, B. D. and S. T. Kolaczkowski, "Waste Minimization Guide," Institution of Chemical Engineers, Rugby (1992).

Dunham, D., "Upgrade Alkylation for Refining Environment," Hydrocarb. Process. 84, 93-98 (2005).

DuPont STRATCO Refinery Solutions, "HF Alkylation Course," (2006).

Eastman, A. D., B. B. Randolph, W P. Moore and R. L. Heald, "Consider Online Monitoring of HF Acid When Optimizing Alkylation Operations," Hydrocarb. Process. 80, 95-100 (2001).

Ekvall, T. and G. Finnveden, "The Application of Life Cycle Assessment to Integrated Solid Waste Management: Part 2--Perspectives on Energy and Material Recovery from Paper," Trans. IChemE (Process Saf. Environ. Prot.) 78, 288-296 (2000).

Elghali, L., R. Clift, K. Begg and S. McLaren, "Decision Support Methodology for Complex Contexts," Eng. Sustainability (in press).

Finnveden, G., J. Johansson, P. Lind and A. Moberg, "Life Cycle Assessment of Energy from Solid Waste--Part 1: General Methodology and Results," J. Cleaner Prod. 13, 213-229 (2005).

Geldermann, J., C. Jahn, T. Spengler and O. Rentz, "Proposal for an Integrated Approach for the Assessment of Cross-Media Aspects Relevant for the Determination of 'Best Available Techniques' BAT in the European Union," Int. J. LCA 4, 94-106 (1999).

Hart, A., R. Clift, S. Riddlestone and J. Buntin, "Use of Life Cycle Assessment to Develop Industrial Ecologies--A Case Study: Graphics Paper," Trans. IChemE (Process Saf. Environ. Prot.) 83, 359-363 (2005).

Hommeltoft, S. L, "Fixed-Bed Alkylation Using SLP-Type Catalyst in a Chromatographic-Type Reactor Concept," Ind. Eng. Chem. Res 42, 5526-5534 (2003).

Hutson, T. Jr. and G. E. Hays, "Industrial and Laboratory Alkylations," American Chemical Society, Washington, D.C. (1977).

IChemE, "The Sustainability Metrics. Sustainable Development Progress Metrics," Institution of Chemical Engineers, Rugby (2002).

James, H. G. and G. E. Handwerk, "Petroleum Refining: Technology and Economics," Fourth Edition, Marcel Dekker, Inc, New York (2001).

Kane, L. and S. Romanow, "New Catalyst Boosts Gasoline Yields," Hydrocarb. Process. 82, 32 (2003).

Kirk, R. E. and D. F. Othmer, "Encyclopedia of Chemical Technology," John Wiley & Sons Chichester (1994).

Miron, S. R. and J. Lee, "Molecular Structure of Conjunct Polymers," J. Chem. Eng. Data 8, 150-160 (1963).

Mitchell, C. A., A. L. Carew and R. Clift, "The Role of the Professional Engineer and Scientist in Sustainable Development," in "Sustainable Development in Practice: Case Studies for Engineers and Scientists," A. Azapagic, S. Perdan and R. Clift, Eds., John Wiley & Sons, Chichester (2004), pp. 29-55.

Science Lab, Material Safety Data Sheet (MSDS) for Sodium Fluoride, (2005), Available at: http://www.sciencelab.com/xMSDS-Sodium fluoride-9927595.

Munter, P. A., O. T Aepli and R. A. Kossatz, "Hydrofluoric Acid-Water and Hydrofluoric Acid-Hydrofluorosilic Acid-Water," Ind. Eng. Chem. 39, (1947).

Nicholas, M. J., R. Clift, A. Azapagic, F. Walker and D. E. Porter, "Determination of 'Best Available Technology' for Integrated Pollution Prevention and Control: A Life Cycle Approach," Trans. IChemE, B. 78, 193-203 (2000).

Nieto, O., M. Nino, R. Martinez and R. G. Tailleur, "Simulation of a Spouted Bed Reactor for Solid Catalyst Alkylation," Fuel 86, 1313-1324 (2007).

Nord, "Nordic Guidelines on Life-Cycle Assessment," Nordic Council of Ministers Report Nord 1995:20 (1995).

Rebitzer, G., T. Ekvall, R. Frischnecht, D. Hunkeler, G. Norris, T. Rydberg, W.-P. Schmidt, S. Suh, B. P. Weidema and D. W Pennington, "Life Cycle Assessment Part 1: Framework, Goal and Scope Definition, Inventory Analysis, and Applications," Environ. Int. 30, 701-720 (2004).

Sharratt, P. N. and P. M. Choong, "A Life-Cycle Framework to Analyse Business Risk in Process Industry Projects," J. Cleaner Prod. 10, 479-493 (2002).

Sikdar, S. K., J. Drahos and E. Drioli, "Tools for Pollution Prevention," in "Tools and Methods for Pollution Prevention," S. K. Sikdar and U. Diweka, Eds., Kluwer Academic Publishers, Dordrecht (1999), pp. 1-4.

Stirling, A., "On the Economics and Analysis of Diversity," Electronic Working Paper Series: Paper No. 28, Science Policy Research Unit (SPRU), University of Sussex, Brighton, UK (1998).

Veolia, http://www.veoliawaterna.com/technology/hardtac.htm, (accessed 14 January) (2008).

Wood, C. B., J. V. Laan and L. W Shoemaker, "Consider Improved Liquid Alkylation Catalysts," Hydrocarb. Process. 80 (2001).

Wrisberg, N. and H. A. Udo de Haes, Eds., "Analytical Tools for Environmental Design and Management in a Systems Perspective," Kluwer Academic Publishers, Dordrecht (2002).

(1) For a complete exposition of LCA and its uses see, for example, Baumann and Tillman (2004).

(2) This kind of assessment is sometimes termed a "consequential" LCA; see, for example, Rebitzer et al. (2004).

(3) Some impurities, including SOz, [H.sub.2]S[O.sub.4], [H.sub.2]Si[F.sub.6], and As, enter with the HF but are not discussed in this paper.

(4) Alkylate is defluorinated when used in aviation fuel, but this is not carried out at this particular refinery.

(5) Converted from Eastman et al. (2001) who state that HF losses between 0.08 and 0.3 lb/bbl of alkylate have been observed on operating units.

(6) Arsenic levels are sufficiently low that they are unlikely to render the calcium fluoride unusable, but this must also be verified.

(7) It could also be followed by further processing of the waste as in options (d) and (e), but in that case the principal advantage of substituting NaOH by KOH is no longer present.

Neil Weston, [1,2] Roland Clift, [1] * Lauren Basson, Andrew Pouton [2] and Neil White [2]

[1.] Centre for Environmental Strategy, University of Surrey, Guildford GU2 7XH, United Kingdom

[2.] Chevron Limited, Pembroke Plant, Pembrokeshire SA71 SSJ, United Kingdom

* Author to whom correspondence may be addressed. E-mail address: rcliftaa sarrey.ac.ak Can. J. Chem. Eng. 86:302-315, 2008

Table 1. Routes by which fluorine is lost from the process

Source of fluorine How it is lost Unit (Fi- Approx. pro-
loss gure 2) portion (%)

HF in azeotrope with Extracted via re-run E 5
water column bottoms
Free HF Entrained in re-run E 45
 column bottoms
Free HF and organic Extracted from H & I 10
fluorides propane product
Free HF & organic Extracted from K & L 15
fluorides butane product
Organic fluorides Entrained in J 20
 alkylate product
Free HF and organic Relief system (Distri- 5
fluorides buted)

Table 2. Waste reduction and management alternatives and level process
of change required

Alternative Description

(a) Make operational changes Change operational parameters to reduce
 generation of organic fluorides, ASO
 generation and water fed to the unit
(b) Remove water Evaporate water from neutralized
 solution to reduce volume of material
 sent for waste disposal
(c) Make design changes (to Make structural changes to the acid
alkylation unit) regeneration section of the alkylation
 unit to improve stripping efficiency in
 the re-run column
(d) Feedstock to another React fluoridic waste with limestone to
process form calcium fluoride (Ca[F.sub.2]) as a
 co-product. Also may enable recycling of
 caustic
(e) Regenerate HF As (d) but with (Ca[F.sub.2]) co-product
 acidified to regenerate HF for use in
 the process or elsewhere
(f) Alternative neutraliza- Replace sodium hydroxide with potassium
tion substance (KOH) hydroxide, which produces a fluoride
 (KF) with greater solubility in water
 therefore allowing a greater
 concentration to be used resulting in a
 lower waste volume
(g) Alternative catalyst Avoid generation of fluoridic wastes by
 removing HF from the unit and replacing
 with an alternative catalyst
 (Hommeltoft, 2003; Nieto et al., 2007)

Alternative Objective Level of change

(a) Make operational changes Reduction 1

(b) Remove water Improved 1
 treatment
(c) Make design changes (to Reduction 2
alkylation unit)
(d) Feedstock to another Recycling 2
process
(e) Regenerate HF Recycling 3

(f) Alternative neutraliza- Reduction 3
tion substance (KOH)
(g) Alternative catalyst Elimination 3

Table 3. Drivers and barriers for process alternatives

Alternative Technological Economic Environmental
 Drivers Barriers Drivers

(a) Make Easiest change Have to Reducing
operational to make. Offers operate within waste means
changes potential to design limits less waste
 save on all of plant so disposal
 types of large costs--as these
 fluoridic improvements costs are
 wastes may not be gradually
 achievable increasing
 then maybe
 this will
 become
 profitable

(b) Remove Simple Need to find Reduces waste
water technology. space for disposal costs
 Can be carried additional
 out while unit equipment
 is in operation

(c) Make Possible Intrusive Reducing
design modifications modifications waste means
changes (to vary in their require the less waste
alkylation complexity but unit to be disposal costs,
unit) they are all shutdown and chemical
 proven in costs. Also
 industry and reduces scope
 feasibly for any
 implemented downstream
 waste
 treatment
 strategies
 therefore
 reducing their
 cost

(d) Feedstock Does not Requires Eliminates
to another involve Ca[F.sub.2] to waste disposal
process additional be high enough costs and may
 treatment quality for provide a
 equipment further use small sales
 revenue

(e) Regene- Proven HF Save money
rate HF technology is regeneration on HF imports
 commercially is something and potentially
 ready. Off-site that would not eliminate
 facilities "normally" be waste disposal
 exist which done at an oil costs. Using an
 already carry refinery. May off-site
 out this pro- not be large contractor
 cess (e.g. HF enough flow to would mean
 manufacturers). make on-site no capital
 Potential to facility costs of
 combine flows feasible. regeneration
 with other Relies upon equipment
 refiners in a being able to
 common achieve good
 regeneration quality
 process. Ca[F.sub.2]
 Modification
 may be carried
 out while unit
 in operation

(f) Alterna- Simple Need to Compared to
tive neutra- modification establish all previously
lization already in hazards of mentioned
substance commercial modifying alternatives
(KOH) use, which equipment this is a
 offers large while unit is relatively
 reduction in in operation cheap way to
 waste volume. reduce waste
 Potentially
 carried out
 while unit is
 in operation.
 Possible
 precursor to
 other
 alternatives

(g) Alterna- Less Only sulphuric Savings will
tive catalyst hazardous acid ([H.sub.2] result from
 technologies S[O.sub.4]) has reduced
 are being been proven material
 developed and over long term consumption
 some are industrial use and waste
 already being and this brings disposal. Other
 commercially its own hazards. benefits, such
 applied Chapin et al. as improved
 (Hommeltoft, (1985) suggest octane rating
 2003; Nieto that although of the
 et al., 2007). [H.sub.2]S alkylate, may
 Some [O.sub.4] has occur as a
 modifications benefits in result of the
 are more easily terms of safety, modification
 installed as an the choice therefore
 additive to HF between HF providing
 alkylation and [H.sub.2]S further (and
 processes [O.sub.4] is perhaps more
 (Wood et al., dependent on significant)
 2001) many factors financial
 many of which incentive
 are refinery
 process and
 site specific.
 For a
 discussion on
 the factors
 affecting the
 choice of
 catalyst and
 the relative
 merits of the
 two catalysts
 see, for
 example,
 Chapin et al.
 (1985) and
 Dunham
 (2005)

Alternative Social
 Barriers Drivers Barriers

(a) Make Current Reducing waste at Any affect on
operational operations the source reduces alkylate
changes generally set HF, caustic and production
 up to optimize water efficiency or
 profitability of consumption and quality will
 the unit amount of waste lead to greater
 therefore any transport and demand for
 changes are disposal feed products,
 likely to which would
 increase costs offset
 or reduce environmental
 value of benefits of
 product waste
 reduction

(b) Remove Would require Transport of dry Produces a
water capital solids reduces risk waste water
 expenditure of release. stream.
 which must be Reduced transport Increases
 justified by emissions energy
 cost savings demands
 which may
 offset reduced
 emissions
 from transport.
 Still requires
 disposal of
 hazardous
 waste to
 landfill

(c) Make Some more Reducing waste at Life-cycle
design complex the source reduces emissions
changes (to modifications HF, caustic and associated
alkylation may be water with the
unit) expensive consumption and fabrication of
 compared to amount of waste new
 the savings transport and equipment
 achieved. disposal
 Delays
 restarting the
 unit damages
 financial gains

(d) Feedstock No further Avoids sending Importing
to another economic waste to landfill, limestone may
process expense to and reduces increase
 water removal emissions due to overall
 technologies Ca[F.sub.2] emissions
 (described production
 above) elsewhere

(e) Regene- Building an Reducing caustic Importing
rate HF on-site facility imports (by limestone may
 would incur recycling) may increase
 capital costs. offset increased overall
 Cost of emissions from emissions.
 sulphuric acid limestone imports. Leads to the
 imports Closes the loop generation of
 for HF losses. gypsum,
 Eliminates which is
 hazardous waste already in
 products. Using abundant
 off-site supply and
 regeneration could therefore
 process raises become a
 possibility of waste product.
 partnerships with Introducing
 other refineries [H.sub.2]S
 achieving [O.sub.4]
 synergistic brings new
 benefits environmental
 hazards

(f) Alterna- Would require Reduced waste so Source of KOH
tive neutra- capital less transport and may have a
lization expenditure chemical large impact on
substance which must be demands. Creates life-cycle
(KOH) justified by possibility for a emissions
 cost savings combined
 treatment route
 with HF losses to
 KOH treaters

(g) Alterna- Redesigning Complete Alternative
tive catalyst the alkylation elimination of the catalysts may
 unit to operate life-cycle result in other
 with an emissions related types of waste,
 alternative to the current which may be
 catalyst would treatment route more hazardous
 incur large for fluoridic or greater in
 costs, which losses. Also volume. Large
 would not be remove the risks scale
 recovered by of environmental modifications
 the benefits of impacts from HF to the alkyla-
 waste release tion unit would
 prevention result in
 alone associated
 emissions and
 current
 equipment
 would have to
 be
 decommissioned

Alternative Drivers Barriers

(a) Make Reducing waste at Refinery
operational source reduces would not
changes manual handling make any
 operations and modifications
 transport of that might
 hazardous increase safety
 material risks of the
 unit

(b) Remove Transport of dry May introduce
water solids is less laborious
 hazardous and solids removal
 therefore more task
 acceptable to the
 general public

(c) Make Reducing waste at Refinery
design source reduces would not
changes (to manual handling make any
alkylation operations and modifications
unit) transport of that might
 hazardous increase safety
 material risks of the
 unit

(d) Feedstock Does not require Same as for
to another additional water removal
process transport or option
 manual handling
 operations of
 hazardous
 materials

(e) Regene- An onsite facility Introduces
rate HF would reduce HF another highly
 transport and hazardous
 manual handling substance
 operations (sulphuric
 acid) to the
 site requiring
 more manual
 handling
 operations and
 transport

(f) Alterna- Reduce manual Effectively an
tive neutra- handling end-of-pipe
lization operations and solution which
substance transport of fresh does not
(KOH) and spent caustic reduce HF
 therefore reduced losses from the
 safety risks and unit and
 road use therefore
 transport of HF
 would remain
 the same

(g) Alterna- Replacement with Replacing with
tive catalyst a less hazardous [H.sub.2]S[O.sub.4]
 catalyst would be would increase acid
 favourable to consumption
 plant safety and therefore
 would receive leading to
 support from the increased
 general public transport of
 hazardous
 substances by
 road and
 through local
 towns. Despite
 its hazards
 there have
 been relatively
 few incidents
 resulting from
 exposure to HF
 (Wood et al.,
 2001)