In the UK, the Nuclear Industry Radioactive Waste Executive was formed in 1982 and incorporated as a private limited company - United Kingdom Nirex Limited (Nirex) - in 1985. Its mission is to facilitate the final disposal of ILW and certain LLW and to investigate and develop disposal concepts and facilities.

A key objective for both BNFL and Nirex is to ensure that the treatment, immobilisation and packaging of waste, prior to a deep repository becoming operational, will produce packages that will be acceptable for ultimate disposal.

The work carried out by BNFL and Nirex to develop waste packages that can meet the safety requirements during production, storage, transport and ultimate disposal is described below. Included is a summary of existing waste management strategies and waste treatment processes and plants. The limitations of these technologies for heterogeneous wastes are highlighted and the development of alternative complementary processes are discussed. In particular, practical studies to develop and support the disposal of a novel supercompacted sludge waste form are discussed.

Detailed technical studies by BNFL in the early 1980s highlighted the benefits from the early packaging of wastes and subsequent storage of waste in an encapsulated form, as this represented a reduction in the risk level, savings in lifetime costs and reduced dose uptake to operators. As a result of these findings, BNFL developed a strategy to treat ILW by direct encapsulation. Current and future ILW arisings would be treated as they arose with historical wastes retrieved and conditioned when appropriate. The timing for treating historical wastes would depend on the safety of continued storage in an unconditioned form and the availability of conditioning plants at Sellafield.

The adopted strategy is to produce essentially monolithic products that will maintain their integrity during on-site storage and a 50-year operational period in the disposal environment, plus being capable of meeting the requirements for transport to, and final disposal in, the deep geological repository. It has therefore been important for BNFL to work closely with Nirex, the company responsible for developing the transport and disposal systems.

On-going reprocessing operations at Sellafield lead to a number of well-defined wastestreams requiring encapsulation. During the 1980s, processes and plants were developed to treat and condition these wastes based on their physical form.

Solid wastes (such as fuel cladding from decanning operations) were to be placed into stainless steel drums and immobilised using carefully formulated cementitious grouts. Grout infiltration was facilitated by vibration, the whole process being known as vibro-grouting.

Liquid and slurry wastes (such as effluent treatment flocs and pond sludges) were to be homogenised, metered into drums containing a sacrificial mixing paddle, and mixed with specially blended cement powders. This process was referred to as in-drum mixing.

Based on these technologies, the following ILW conditioning plants have been built and are currently operational: The Magnox Encapsulation Plant (MEP) for treating solid magnox swarf wastes; the Waste Encapsulation Plant (WEP) for treating solid fuel hulls wastes, effluent treatment slurries and raffinate purification slurries; and the Waste Packaging and Encapsulation Plant (WPEP) to encapsulate routine arisings of effluent treatment flocs.

The products produced by these plants are cementitious monoliths where the radioactive activity is immobilised by intimate mixing with the cement matrix. Overall containment and handling features are provided by a stainless steel drum.

Extensive research work by both BNFL and Nirex on these types of waste form has been carried out to assess release fractions from such packages during normal operations and accidents. This work has shown releases to be low, with accident releases increasing progressively in a predictable manner with increasing severity of the accident. For example, following a 25 meter drop, particles generated within the drum of less than 100 microns (i.e., of a size capable of being resuspended and dispersed) represent approximately one millionth of the original waste inventory, showing the potentially available source-term for accidental releases to be extremely low. These release data have subsequently been used in safety assessments for a generic disposal facility and has been shown to be consistent with the requirements foreseen as being necessary for the safe transport and disposal of waste.

In addition to operational factors, the effect of these packages on post-closure repository performance has been assessed. Here, the high alkalinity of the cement grout is beneficial, reducing the concentration of key radionuclides by up to five orders of magnitude and suppressing microbial activity (and its associated gas generation). This, combined with the low permeability of the matrix which restricts radionuclide migration and corrosion rates of metals, also means that the long-term disposal requirements foreseen by Nirex could be satisfied by these packages.

Having developed plants and technologies to manage the current and future waste management liabilities, attention was focused on applying the existing plants and now- proven technologies to managing the historical liabilities of wastes that had accumulated and been stored at the site since its initial inception in the 1950s. A number of these historical wastes are well defined and are suitable for routing directly to existing facilities with minimal pre-treatment following retrieval. However, some historical wastes on the Sellafield site consist of a complex mixture of solids and viscous sludges which are not readily compatible with either the vibro-grouting or in-drum mixing processes. Therefore, initial process studies concentrated on retrieval and treatment processes that can segregate the waste mixtures into well-defined sludge and solids streams that are consistent with the feed requirements of the existing plants. One such waste stream was corroded magnox fuel cladding which had arisen from underwater storage of the magnesium alloy cladding for over 20 years. In addition to the cladding itself, the wastestream contains irradiated uranium fuel adhering to the cladding or present as discrete fragments of broken fuel elements, residues furniture such as fuel rod-end pieces, springs etc. Miscellaneous engineering wastes have also been added to the waste storage silos over the years. Consequently, a systematic programme to investigate process options for treating this wastestream was undertaken. The first treatment option considered was based on wet settling. Later, a process based on drying and compaction was also investigated. The next section describes the advantages and disadvantages of the wet settling and drying/compaction processes for treating corroded magnox fuel cladding wastes.

The wet settling concept was devised in 1991 to treat ILW stored at Sellafield. It was developed to segregate ILW into two fractions, one solids based, the other sludge based. These fractions could then be processed using treatment principles and technology described previously.

The process also had advantages in that the technologies adopted were consistent with existing processes operated on site and thus raised few new safety issues. One of the main safety issues associated with this waste stream was the presence of a pyrophoric compound, uranium hydride, a corrosion product of uranium. Conventional practice was to keep all uranium damp to prevent any uranium hydride present from reacting with air and igniting. This requirement was readily achieved in the wet settling process.

Preliminary discussions were held with the Regulators on processing and on-site storage and Nirex with regard to disposal to appraise them of BNFL proposals.

The key stages in the wet settling process were:

As detailed process and engineering design progressed during 1992, a number of major problems were encountered.

Despite an extensive research and development programme focusing on process and product quality concerns, robust solutions to the issues listed above were not forthcoming. By late 1992 a decision was taken to explore other processing solutions.

Early in 1993, following a Value Engineering and Optioneering exercise, the Drypac concept was adopted as the reference process to treat corroded magnox fuel cladding wastestreams.

The Drypac concept consists of a segregation phase similar to the wet settling process to produce sludge/solids fractions. Sludge wastes are then dried to remove free water and compacted to reduce the overall volume of waste for encapsulation. Solid wastes are also compacted. The net result of the process is to concentrate activity in each package. Due to the novelty associated with the process, discussions commenced with Nirex regarding disposal and UK Regulators with regard to on-site safety at a very early stage.

The key stages in the Dryac process are:

• Retrieved wastes are screened to separate bulk solids from sludges.
• The sludge fraction is collected in a receipt vessel and measured amounts of sludge are loaded into sacrificial cans.
• Cans of sludge are placed into ovens to remove free water.
• Dried cans of sludge are compacted to form pucks of waste.
• Bulk solids are size reduced, if required, and loaded into sacrificial cans.
• Cans of solid material are transported directly to the compactor and compacted into pucks.
• Pucks are placed into enhanced drums.
• Drums of waste are then transported to an encapsulation plant for grouting.

A schematic representation of the Drypac process is given in Figure 1.

In comparison to the wet settling process, the Drypac process has many advantages, including:

From a Nirex perspective the concept where the waste is not ultimately grouted but compacted to form an ingot or puck, which is then surrounded by a high integrity grout annulus, raises a number of novel issues when assessing the operational and post-closure implications. These concerns primarily focused upon:

To address the Nirex and regulatory concerns summarised above, a substantial R&D programme has been carried out. This has focused on safety and product quality issues as follows:

Laboratory and full-scale trials have been performed to confirm the behaviour of waste materials in the drying process and enable a safety case to be made for the Drypac process.

• Hydrogen ventilation

The following work has been carried out to provide process/engineering data to the BNFL design teams and Nirex:

In addition, any hydride formed during storage was found to be limited by water availability, and rapid oxidation would be prevented by the robust design of the package and the monolithic nature of the compacted sludge. Hence, the risk from uranium hydride formation on repository safety was judged by Nirex to be low and acceptable. Examples of typical impact test performance are shown in Figure 3.

Figure 4 shows the results from thermal modelling of SDP packages and shows that the presence of the grout annulus affords good protection to the waste in fire accidents.

With respect to post-closure performance, the compacted sludge waste form has been assessed against its potential release activity with time into the repository near-field. The assessments concluded that careful package design and limiting the contents of expansive waste components will prevent packages from breaching for sufficiently long periods to allow short-lived soluble species such as Cs-137 and Sr-90 to decay to insignificant levels. For key longer-lived species such as U-238, the concentration of materials in solution has been assessed to still be low due to the alkaline pH buffering and strong sorption provided by the magnesium hydroxide component of the waste. Therefore, post-closure assessments concluded that this packaging concept is capable of providing acceptable performance for this waste, despite the absence of intimate encapsulation. As a result of this work, construction of the $600 million Sellafield Drypac Plant commenced in 1996 and should become operational in 2003. The suitability of this packaging concept for other historical wastes is currently being investigated.

Based on conducted research and analyses, BNFL has reached the following conclusions: