The present invention relates to a method for producing a green body (9a to 9c) for a 3D ceramic and/or metallic body .

The fabrication of 3D objects in terms of ceramic or metallic bodies can be a manifold task due to the

complexity of the design, the difficulties to handle the ingredients, such as radioactive components, and a lot other parameters that might have an impact on the

fabrication . I.e., Paul Scherrer Institut (PSI), CH-5232 Villigen PSI, has a long tradition in developing advanced nuclear fuels. Some major efforts went into the development of actinide and minor actinide containing fuels for

transmutation. One route has been the development of inert matrix fuels, for plutonium transmutation in light water reactors, either as yttria stabilized zirconia pellets [1], or as CerMet pellets [2]. It is almost superfluous to mention that these materials encumber one of the highest complexities in a manufacturing

environment due to the high radioactivity of the actinide metals used in nuclear fuels, especially when they contain minor actinides for transmutation purpose.

The other route was the development of alternative, aqueous production methods, in order to simplify the fabrication, making it ready for remote manipulation. The major effort went into the internal gelation, resulting in sphere-pac fuel. Both, the gelation and fuel

performance of this particular concept are summarized in [3] and [4] . As the pellet fuel is a well experienced concept in commercial light water reactors, and also in many advanced systems, such as sodium cooled fast

reactors, some efforts went also in the direction of developing aqueous fabrication routes, resulting in pellets. Early approaches concentrated on crushing spheres into pellet-shape and heat treat them (see [5] and [6]), resulting in so called hybrid pellets. But also the use of an aqueous direct ceramic shaping methods was already introduced at PSI [7], using the direct

coagulation casting technique. Another effort went in the direction of freeze drying [8] . The closure of the nuclear fuel cycle offers very

promising aspects like an improved uranium resource usage and a major reduction of long lived minor actinides in the final waste. For a comparison of different aspects between the classical once through PWR scenario, the once reuse of Pu option and the fully closed fuel cycle involving fast reactors see Fig. 3 in [9] .

Closing the cycle of course involves reprocessing of spent fuel, and the fabrication of highly active, minor actinide containing fuel. The high activity imposes new production challenges compared to the production of fresh nuclear fuel. Most important:

a) the production must be performed remotely in a

shielded environment (hot-cell) , and

b) any accumulation of fuel educts should be avoided

(because of high equipment contamination and criticality risk) .

The first aspects calls for production equipment with less maintenance need. The latter aspect disfavors powder based production, as dust is volatile and can deposit anywhere in the production environment, the hot-cell.

The solution to these challenges can be a simplified pellet process [10] or particle fuel with much simplified production passes. At PSI, the particle fuel option was researched over decades using the aqueous internal gelation process resulting in fuel pebbles used in the sphere-pac concept. An extensive description of sphere- pac and also the vipac fuel can be found in [3] .

Pellet fuel has heavily been optimized for UO ₂ in Zircaloy cladding and LWR reactors. If considering fast reactors, other fuel matrices, such as carbides and nitrides might become more attractive [11] because of higher metal content and better thermal conductivity. However, the swelling behavior especially of carbides is much higher compared to oxides. Therefore, porosity should be

designed to accommodate the dimensional change with burnup .

The ATF (accident tolerant fuel) initiative is a large, post-Fukushima effort, to reduce the risk of

fuel/cladding failure in case of an accident [12] . Some of the concepts are based on oxidation and high

temperature resistant ceramics. These are by nature brittle, even though pseudo ductility is introduced by using a composite. However, fuel-cladding mechanical interaction should be avoided by designing a large gap. In order to avoid an important temperature step, one concept being suggested is the introduction of porous graphite buffer.

PSI has already implemented a base process for the production of fuel pellets. The internal gelation

technique [13] is well experienced at PSI. As illustrated above, it has been used for decades to produce pebbles for the sphere-pac concept. The main feature of the internal gelation is the heat triggered solidification process. This means that the gel can be formed by heating the feed solution, which was classically performed by hot silicon oil surrounding the droplets of feed solution. In the Swiss CCEM.CH project PINE and MeAWaT [13], [14] and the he European projects ASGARD and PELGRIMM [16], [17], the internal gelation was researched by using microwaves for heating in order to avoid any radiolysis and decay heating influence onto the process and to be able to work with non-cooled solutions [15].

It is therefore the objective of the present invention to provide a method, in particular in nuclear fuel

production, providing an almost dustless fabrication method with the ability to introduce locally varying features like the metal component composition, the enrichment (which is the isotope composition) and the porosity.

This objective is achieved according to the present invention by a method for producing a green body (9a to 9c) for an 3D ceramic and/or metallic body, comprising the steps of:

a) creating a 3D production control model for the green body ( 9a to 9c) ;

b) providing a metal or a mixture of metals and/or a metalloid and/or a non-metal or mixtures thereof in form of at least one aqueous solutions (la, lb, In) , such as a metal nitrate solution; in case of at least two aqueous solutions (la, lb, In) ; said at least two aqueous solutions differ from each other in terms of composition and/or isotope concentration;

c) providing a gelation agent in form of a gelation fluid (2) ;

d) mixing one of the at least one aqueous metal solutions (la, lb, In) with the gelation fluid (2) at a first temperature to form an feed solution mixture wherein the first temperature is chosen to suppress an internal gelation of the feed solution mixture prior to its ej ection;

e) ejecting the feed solution mixture by an inkjet printing process to the green body under construction; f) heating the ejected feed solution mixture on the green body under construction to a second temperature wherein the second temperature is chosen to fix the ejected feed solution mixture on the green body under construction; and

g) repeating the steps e) and f) and optionally d) according to the 3D production control model until a desired form of the green body (9a to 9c) will have been achieved; and

h) optionally heating the green body to a third

temperature under given atmosphere, such as an oxidizing or reducing atmosphere, to achieve the formation of the 3D ceramic or metallic body and/or to partially or fully sinter the green body. In this step also the chemical residuals from the gelation reaction can be driven out of the product . Therefore, the present invention improves the internal gelation process combined with an implementation of the 3D inkjet printing technique for ceramics or metals in order to enable the fabrication of various complex 3D ceramic or metal bodies, such as nuclear fuel pellets and the like. With 3D printing and the possibility to supply different aqueous metal solutions to the mixing step, this is achieved locally; allowing maximal flexibility in 3D design. As Pu and minor actinides are to be introduced in fast reactor fuel for transmutation, the same applies for the local metal composition, which could be optimized with this technique.

It is in many cases very useful if the composition of the green body can be varied as a function of the green body location. According to a preferred embodiment of the present invention, this objective can be achieved when during the repetition of the steps e) and f) the

composition of the mixture varies in terms of composition and/or isotope concentration. It is therefore possible to provide for each new layer of the ejected mixture (or for part of a layer) a different composition since the composition can be easily controlled by the choice of the aqueous metal solution fed actually to the mixing step. A similar approach can be realized with respect to a desired porosity profile of the green body. Hence, a pore builder can be fed in form of a pore forming additive into the mixing step d) .

In order to facilitate a simple handling of the feed solution mixture and to avoid aging/decay problems and/or clogging of the inkjet printing process due to beginning gelation, it can be advantageous when the mixing step d) is performed immediately prior to the ejection of the ink mixture, preferably in proximity to an ink nozzle.

Therefore, the time interval between mixing and ejecting the feed solution mixture is extremely short and can range from some milliseconds to some seconds, such as 200 ms or 5 s or the like.

In order to offer superior conditions for the gelating of the mixture on the mixture target (the green body under construction) , a further preferred embodiment of the present invention may provide that the heating of the ejected mixture on the green body under construction is achieved by heating the green body under construction by any combination of laser, microwave and other heating technique and/or heating of the ejected ink mixture during and/or after depositing the ejected feed solution mixture on the green body under construction by a

mentioned heating technique or any combination of it. In order to construct 3D ceramic or metal bodies having complex form including for example cavities, it is desirable to have a high flexibility during the

deposition of the feed solution mixture. This objective can be achieved when the sample holder can be moved horizontally and vertically and can be rotated around the vertical axis and can be tilted with respect to the vertical axis. Preferred use cases of the inventive method described above are listed in the claims 7 and 8.

The present invention and preferred embodiments thereof are hereinafter described in more detail with reference to the attached drawings which depict in:

Figure 1 schematically a setup for the production of

nuclear fuel pellets; and

Figure 2 schematically a sample holder with some degree of freedom for the movement of the sample holder . Implementation example 1 - Realization of a nuclear fuel pellet with variable composition and porous surface layer

Hereinafter, an implementation example for the production of nuclear fuel pellets with a variable composition and a porous surface layer is explained in more detail. The example describes the production of an advanced nuclear fuel pellet, which varies in its composition and features a soft surface. The aim is to have a well conducting outer pellet region, made out of pure uranium-oxide, and to enrich the inner section in minor actinides, for their transmutation. The porous, soft surface is aiming for a reduction in the pellet-cladding mechanical interaction.

Two metal feed solutions la and lb are prepared, one uranium nitrate solution (deficient uranyl-nitrate solution (ADUN) ) and another nitrate solution containing americium, curium and neptunium. As a variation of that, a further solution can contain plutonium, and/or one or several of the before mentioned minor actinides can be prepared in separate solutions. In the Figure, these metal solutions are represented in a container section 1. A gelling agent 2 is prepared which is the organic compound hexamethylenetetramine, also called HMTA

(Hexamine) , which is acting as a homogeneous

precipitating agent. HMTA causes rapid precipitation of U(VI) . Therefore, in order to prevent a premature

gelation the organic compound urea (CO(NH2)2) is added to complex the uranyl ions [13] .

In a first step, the metal solutions are mixed in a first mixing unit 5, which can be also realized as simple T junction, if only 2 metal solutions are mixed, or, if the number of metal solutions n is higher, in a specially developed blender, with n entries. In the next step, the blended metal nitrate solution is mixed in a second mixing unit 6 to the gelation agent 2. From this time the gelation reaction is starting, especially if the

temperature is kept at ambient conditions, or is even elevated, due to the decay heating of the minor actinides (up to 50°C) . This latter step of in-situ mixing of the metal solution la, lb to the gelation agent 2 has been lately developed in the MeAWaT project at PSI [15] . In the same blending step (in the second mixing unit 6) a pore builder 3, such as graphite, is added which acts as a pore former in the thermal treatment process. As an alternative other additives can be injected here, to achieve modifications to the material. A small

mixing/homogenization device 7 is inserted by having short range windings in the tube. Depending on the design and tube lengths, this step can be optionally omitted. The metal solutions la, lb, the gelation agent 2 and the pore builder 3 are conveyed by corresponding pumps 4.

The ratio of the metal ions (here uranium and minor actinides) can be changed at any time, by adjusting the feeding rate of the corresponding pumps 4. These pumps 4 are realized as high precision HPLC pumps here. The inner diameter of all tubes is chosen to be 0.18 mm. The resolution of composition variations in the final product is given by a ratio between the time resolution of the pumps 4 with the feeding rate and the effective blending volume in the first mixing unit 5. With a simple T junction this is typically the volume with a sphere of the inner tube diameter. All the volume in the tubes after the blending step to the deposition on a sample 9a being supported by a sample holder 9 which is controlled to move vertically according to the progress of the 3D printing process will lead to a delay, which has to be accounted for when programming the feeding rates for the metal solutions. In this application, the typical time after mixing to the gelation agent 2 (in the second mixing unit 6) to the deposition on the sample 9a is in the sub-second range. This means that practically all the gelation reaction will happen after deposition. With these short times, cooling of the feed solutions is not necessary, as only at temperatures approaching 60-100°C the reaction (gelation) time drops into the sub-second range .

As in other 3D printing applications, an ink-jet nozzle 8 is used to deposit the feed solution to the surface of the sample 9a (the feed solution target) . A special version of such a nozzle has been developed, as radiation can degrade piezo crystals. This new nozzle 8 is based on electromagnetic forces.

The nozzle 8 is mounted on a laterally moving stage. The samples 9a, 9b and 9c are placed on a vertically moving sample holder 9, providing together with the nozzle 8 the three-dimensional accessibility. In the application here, where the final product will be a pellet of typically 10 mm diameter and 10 mm height, an array of 10x10 such samples 9a to 9x are produced at the same time. For each sample new layers of feed solution are applied in rings from outside to the center. As the compositions and porosity variations in the pellet are realized in a radial manner, this requires the least variation in the pump-feeding rates. After applying a layer to one sample 9a, the stage moves on to the next one. The newly applied layer is then fixed with a coaxial microwave applicator 10, heating up here actually the surface of the sample 9c. Alternatively other forms of energy applicators can be used. This way the gelation is finalized, and the chemical residuals are driven out the sample 9a. After applying several layers in this way, a high temperature treatment is realized, in order to have calcination and partial sintering. Alternatively, a second heating source is realized with a laser (parallel to the coaxial microwave applicator 10 or replacing it) , which can provide locally very high temperatures and therefore sintering .

In the oxidizing atmosphere, the aimed oxide ceramics will be realized this way. The applied graphite will also oxidize to C02 and leave the wished porosity in the samples 9a, 9b, 9c. By an appropriate supply of the pore builder 3, a gradient in the porosity can be achieved over the sample in at least one dimension.

Implementation example 2 - Realization of a nuclear fuel pellet containing sealed fuel cells

In this example, a fuel pellet is printed containing honeycomb structured SiC boundaries filled with nuclear fuel, such as UC and a porous layer, facing the

boundaries. UC is an attractive fuel matrix, as it features a very good thermal conductivity and a higher content in metal, compared to UO ₂ . Unfortunately, it also features two disadvantages, which are the chemical reactivity in oxidizing surroundings and the higher swelling rate compared to standard fuel.

With the approach presented here, the chemical reactivity is mitigated, by splitting up the fuel in many small cells. Even if a cell boundary fails, the reactive volume is very limited, not affecting the overall integrity of the pellet. With the subdivision into the multiple small cells, also containing the porous outer layer in each cell, the swelling will be accommodated on the cell level, and not on the pellet scale. The pellet therefore remains constant in volume (except for some SiC swelling, which is limited by about 1%) and does therefore not interfere with the cladding. The SiC is very oxidation resistant, and shows excellent high temperature resistance. Therefore, this fuel type could qualify as accident tolerant fuel (ATF) , which is an effort to mitigate the effects of accident scenarios like in Fukushima. The approach has some similarity with the suggestion of including TRISO particles in a SiC pellet. This might even show better accident tolerance, as there are more structural layers being involved.

However, with the resulting low fuel fraction in the pellet, would require a higher enrichment.

The setup chosen to produce these pellets is very similar to the setup shown in Figure 1. The second solution lb in the first example will comprise silicon. There will be no variable mixing at first mixing unit 5, but a digital one. Either the nuclear fuel component la or the silicon solutions lb, such as methyltrichlorosilane, will be applied. When it comes to the deposition to the sample, first the boundaries, finally resulting in SiC, will be printed, and then the porous layer near the boundaries, and finally the fuel will be printed in the cells. The thermal treatment will be performed under reducing atmosphere, resulting in a carbo-thermal reaction. For this reaction graphite will be added in the second mixing unit 6 to the fuel printing solutions. The first gelation reaction will be effected here. In case of the SiC layer, the methyltrichlorosilane will directly be treated thermally which in the reducing atmosphere will result in the SiC ceramic. In case of the porous layer, no graphite will be added, but polystyrene will be applied, by crosslinking of polystyrene chains and carbonization a porous carbon layer will form. As the mixing volumes are very small, all of these solution applications can be performed with the same system/nozzle. Alternatively, several nozzles can be used, to increase the production speed, and simplify the single applicators.

Implementation example 3 - Complex nuclear fuel

structures for very advanced concepts

An example of very advanced nuclear concepts is the fission fragment reactor. Here the kinetic energy of the electrically charged fission fragments is directly transferred into electrical energy. One of the major tasks here is the design and production of the nuclear fuel part, and some surrounding system which

deaccelerates the fission products, either through electric and magnetic fields, or through conductors with different electron densities.

For the production of such an advanced nuclear component, some cavities might have to be maintained. Here an advanced sample holder 12 shown in Fig. 2 will be

applied, allowing the production of any arbitrary shape, including empty volumes. The applied feed solution is then always applied to surfaces with a normal vector in gravity direction. The fixation is then within very short time, as either the existing sample surface temperature is elevated and/or heating is applied though a method already described for the first implementation example (microwave applicator 10, laser or others) . With the method described in Fig. 1, rather flat samples can be produce, with only limited possibility to build macroscopic inner cavities. The reason is the application of the feed solution to the sample surface, which is in a rather inviscid state, as the gelation only starts on the surface. The surface tension of the droplets and the fast gelation dynamics are the only effects which can help to produce craning structures. In most of cases a

macroscopically dense structure with some designed porosity and without overhanging structures will be sufficient .

For some special cases, in Figure 2 a 5-axis sample holder 12 is suggested, which will provide a solution application surface normal, always being parallel to gravity. In detail, the sample holder 12 can be moved in the horizontal plane und can be displaced vertically, too. Further, the sample holder 12 can be rotated around the vertical axis and can be tilted by an angle a with respect to the vertical axis. These movements in total provide a freedom to move in 5 axis. As the gelation will take place very quickly, the sample 9a can continuously be moved to other surfaces, adjusting the angles again. With this method, also horizontal structures can directly be constructed. Figure 2 exemplarily shows the sample 9a having a larger inner cavity 14 which can be easily constructed by the concerted movement of the sample holder 12, the feed solution ejection by the nozzle 8 with its needle 16 and the microwave applicator 10 which allows the immediate gelation of the most recently applied product layer.

As also present for the first and second example, the third example requires a bit more complex 3D production control model for the generation of the green body/sample 9a. This model in detail comprises the detailed data about the ingredients of the ink mixture, the printing control data for the nozzle 8 and the heating data to control the microwave applicator 10. Bibliography

[I] G. Ledergerber, C. Degueldre, P. Heimgartner, M. A. Pouchon, and U. Kasemeyer, "Inert matrix fuel for the utilisation of plutonium, " Prog. Nucl . Energy, vol. 38, no. 3-4, pp. 301-308, 2001.

[2] K. Idemitsu, "Incineration of Excess Plutonium in a CerMet Fuel," PSI, Paul Scherrer Institut, Villigen, Switzerland, TM-43-97-29, Sep. 1997.

[3] M. A. Pouchon, G. Ledergerber, F. Ingold, and K.

Bakker, "3.11 - Sphere-Pac and VIPAC Fuel A2 - Konings, Rudy J.M.," in Comprehensive Nuclear Materials, Oxford: Elsevier, 2012, pp. 275-312.

[4] M. A. Pouchon, L.-A. Nordstrom, and C. Hellwig, "3.25 - Modeling of Sphere-Pac Fuel A2 - Konings, Rudy J.M.," in Comprehensive Nuclear Materials, Oxford:

Elsevier, 2012, pp. 789-817.

[5] G. Ledergerber, "Kugeln als Ausgangsmaterial fiir

Pellets," PSI, Paul Scherrer Institut, Villigen,

Switzerland, TM-43-89-21, Aug. 1989.

[6] G. Ledergerber, F. Ingold, R. W. Stratton, H.-P.

Alder, C. Prunier, D. Warin, and M. Bauer, "Preparation of Transuranium Fuel and Target Materials for the

Transmutation of Actinides by Gel Coconversion, " Nucl. Technol., vol. 114, no. 2, pp. 194-204, May 1996.

[7] M. Streit, F. Ingold, L. J. Gauckler, and J. -P.

Ottaviani, "(Pu,Zr)N Annular Fuel Peiets Shaped By.

Direct Coagulation Casting," J. Nucl. Sci . Technol., vol.

39, no. sup3, pp. 741-744, Nov. 2002.

[8] E. Hart, "Ceramic pellet fabrication by freeze- drying, " PSI, Paul Scherrer Institut, Villigen,

Switzerland, TM-43-02-02, Sep. 2002.

[9] V. Kuznetsov, G. Fesenko, A. Schwenk-Ferrero, A. Andrianov, and I. Kuptsov, "Innovative Nuclear Energy Systems: State-of-the Art Survey on Evaluation and

Aggregation Judgment Measures Applied to Performance Comparison," Energies, vol. 8, no. 5, pp. 3679-3719, Apr. 2015.

[10] K. Asakura, K. Takeuchi, T. Makino, and Y. Kato, "Feasibility Study on a Simplified MOX Pellet Fabrication Process, the Short Process, for Fast Breeder Reactor Fuel," Nucl. Technol., vol. 167, no. 3, pp. 348-361, Sep. 2009.

[II] H.-J. Matzke, Science of advanced LMFBR fuels: solid state physics , chemistry, and technology of carbides , nitrides , and carbonitrides of uranium and plutonium. North-Holland, 1986.

[12] "OECD/NEA Expert Group on Accident Tolerant Fuels for Light Water Reactors (EGATFL) . " [Online]. Available: https://www.oecd-nea.org/science/egatfl/. [Accessed: 19- Aug-2016] .

[13] M. A. Pouchon, "14 - Gelation and other innovative conversion processes for aqueous-based reprocessing and recycling of spent nuclear fuels A2 - Taylor, Robin, " in Reprocessing and Recycling of Spent Nuclear Fuel, Oxford: Woodhead Publishing, 2015, pp. 353-369.

[14] M. A. Pouchon, "PINE - Platform for Innovative

Nuclear Fuels," in CCEM : : Annual Activity Report 2012, Paul Scherrer Institut, Villigen, Switzerland: CCEM.CH, pp. 39-40.

[15] M. A. Pouchon, "MeAWat - Methods of Advanced Waste Treatment," in CCEM:: Annual Activity Report 2014, Paul Scherrer Institut, Villigen, Switzerland: CCEM.CH, pp. 52-54.

[16] "FP7 EU project: Advanced fuels for Generation IV reActors : Reprocessing and Dissolution (ASGARD) (Project reference: 295825)." [Online]. Available:

http : //cordis .europa.eu/project/rcn/100635_en . html .

[Accessed: 18-Aug-2016] .

[17] "FP7 EU project: PELlets versus GRanulates:

Irradiation, Manufacturing & Modelling (PELGRIMM)

(Project reference: 295664)." [Online]. Available:

http : //cordis . europa . eu/proj ect/rcn/101413_en . html .

[Accessed: 18-Aug-2016] .