DC Resistance Measurements in Multi‐Layer Additively Manufactured Yttrium Barium Copper Oxide Components

Additive manufacturing can offer new methods to shape materials that are difficult or too brittle in nature to process traditionally. In this work the optimization of a linseed oil and ICP solvent based solution with 90 wt% solid loading of the high temperature superconducting material yttrium barium copper oxide (YBCO) for additive manufacturing is demonstrated. The post sintered, printed components are analysed using a combination of X‐ray diffraction, scanning electron microscopy (SEM), magnetometry, and DC transport measurements. It is shown for the first time, for multilayer additively manufactured YBCO, direct DC transport measurements that indicate a current carrying percolative path with an upper transition temperature of 86.5 K.


Introduction
The superconducting phenomenon manifests in certain materials when sufficiently cooled below its critical temperature, T c , that are then able to carry an electrical current with zero resistance.It was first discovered by Onnes in 1911 when the electrical resistance of mercury was observed to decrease rapidly at 4.2 K. [1] Since then, a variety of materials have been discovered with increasing critical temperatures.More recently, Durrell et al. [2] summarised a road map for future applications of bulk superconductors, which fall into three categories: flux shielding, flux pinning, and flux trapping applications.Examples include: portable systems for high field magnet systems in medical devices, ultra light superconducting rotating machines, and magnetic shielding applications.These applications are currently limited by the method of cooling the material with either liquid nitrogen (T c < 77 K) or a combination of liquid nitrogen and helium (T c < 77 K).
Yttrium barium copper oxide (YBCO) was the first superconducting material with a transition temperature above that of liquid nitrogen at 93 K. [3] As a result, it is a relatively inexpensive means of obtaining superconductivity.Its popularity originates from this high superconducting transition along with high critical current densities and critical fields of up to of up to 8 MA cm À2 and 250 T respectively, [4,5] with industrial applications that could include the transport of current over long distances or for use in electrical machines. [6]owever, as a ceramic, the brittle nature of YBCO limits its use in applications in bulk form where complex geometries are required, such as high field superconducting magnets.Traditionally, YBCO components are produced through tape casting, thin and thick film manufacturing, and lithography techniques. [7,8][11] Whilst these methods are limited to simple geometric shapes, the geometric complexity of bulk YBCO has been improved with slip casting, [12] which has the disadvantage of requiring bespoke dies for each shape (and the expense associated with this).Additive manufacturing (AM), in contrast, has the potential to produce bulk structures with increased complexity without the need for dies, casts, or presses.
There is a sparse published literature on the AM of superconductors.This is likely due to the common applications of superconductors involving wires and coils, therefore, the manufacturing methods to produce tapes, wires, and coils have been sufficient. [13]One application that could benefit from AM is resonant cavities, which are used as particle accelerators for nuclear physics, high-energy physics, materials science, and the life sciences. [14]For example, titanium aluminium vanadium resonators have been additively manufactured through selective laser melting. [15]This process was ideal to produce highly dense metal structures and achieve a superconducting transition of 4.5 K, at the higher end of the stated literature range of 1.3-6.3K. [15] Other superconducting materials such as an aluminium silicon alloy (Al-12Si) [16] and niobium have also been successfully produced through AM.Niobium structures were produced with a relative density greater than 99% and a yield and tensile strength representative of wrought reactor grade niobium.The metallic superconductors mentioned were used due to the relative ease of processing metals though AM.
YBCO has previously been processed through the powder bed fusion and material extrusion AM techniques, whereas Ag-YBCO has been processed through the selective laser sintering technique. [17,18]The powder bed consisted of the YBCO precursor powders: yttrium III oxide (Y 2 O 3 ), barium carbonate (BaCO 3 ), and copper II oxide (CuO), produced through the citrate sol-gel technique. [17,18]A laser processed the materials and resulting structures were then annealed in oxygen for 6 h.Due to the aggressive heating of the laser, large quantities of impurities were formed.As YBCO is denser than the three precursor powders used in this reaction, the solid component also has reduced volume, which results in reduced accuracy of the printed parts and increased frequency of microcrack formations.
[21][22] In one study, strontium titanate and YBCO were additively processed to form gradiometers to be used in conjunction with SQUID instruments. [19]The ceramic materials were mixed in 1,2-propandiol or in polyethylene glycol 200 and a T c of 84 K was recorded.Published images hint towards a low solid loading and poor deposition accuracy with only one layer printed and it was suggested that the low T c was due to a low mass density and a high disorder of the YBCO crystal phase.
Wei et al. [20] investigated, multi layered structures with various YBCO pastes being produced with the precursor powders Y 2 O 3 , BaCO 3 , and CuO.These were mixed with polyethylene glycol, polyvinyl alcohol, solsperse 20 000, and deionised water.Components were then printed with a 72.5 wt% solid loading paste and fired at 940 °C.In this example, samples showed excellent superconducting properties with a T c of 92 K and a diamagnetic signal with a mass magnetisation of À10 emug À1 at 4.2 K.However, poor geometric resolution alongside thick layer inconsistencies were observed as well as deformation of the printed component after the sintering process.Wei [21] then went on to using a YBCO powder that was then formulated into a paste to form two binder systems, which were used: firstly, polyvinyl butyral, ethanol, PEG-400, and solsperse 20 000 and secondly dextrin, PEG-400, solsperse 20 000, and deionised water.The ethanol-based solution was able to print with a solid loading of 84 wt% YBCO powder.This showed a T c of 90 K and a diamagnetic signal with a mass magnetisation of À30 emug À1 at 4.2 K.Although poor resolution was still seen in the layer height and extrusion width, images showed reduced deformation after sintering.
More recently, YBCO pastes have also been formulated with a mixture of polystyrene, o-xylene, dichloromethane solution and surfactant (99/1 wt%). [22]YBCO powder was mixed with these organic binders at concentrations of 2.4 and 6.7 gml À1 .The quantities of organic compounds were not stated so a solid loading can only be estimate at 70 and 86 wt%, respectively.Samples sintered at 950 °C for 15 h had a T c of 92 K and a diamagnetic signal with a mass magnetisation of between À10 and À15 emug À1 at 1.6 K.The critical current density was calculated from magnetometry to be 66-155 kAcm À2 at 1.5 K.A higher resolution and lightweight geometries of YBCO components were produced through direct ink writing. [23]A paste consisting of the precursor powders Y 2 O 3 , BaCO 3 , and CuO in an organic mixture of carboxymethyl cellulose sodium, epoxidised soybean oil, and water was prepared.At solid loading of 60 wt% the paste appears to have been printed through a 410 μm nozzle, an improvement over previous literature.Rheology data showed the paste had a shear thinning behaviour, however, supplementary videos to the article indicated a poor thixotropic response with the paste slumping after printing.The transition temperature was measured to be 91 K, with a diamagnetic signal that had a mass magnetisation of 2.3-6.6 emug À1 and a calculated critical current densities between 1807 and 5562 Acm À2 .
In this work, we demonstrate the use of additive manufacturing to print YBCO components where the solid carrying medium was reversed engineered from pigment carrying technology into two organic components: linseed oil and ICP solvent at a 13:9 mixture.This paste was shown to be able to reliably print while carrying a solid loading of 90 wt%, higher than that of current literature.This simple but effective recipe is desirable as it is industry ready with inexpensive off the shelf components.We printed with a nozzle with a 254 μm diameter, comparable to the current standard of material extrusion processes.We also demonstrate zero resistance below 86.5 K: direct evidence of a superconducting path in an additively manufacture component.

Powder Preparation
YBCO feedstock powder was prepared by the solid-state method by combining yttrium III oxide (Y 2 O 3 ), barium carbonate (BaCO 3 ), and copper II oxide (CuO) (Sigma-Aldrich, 99.99% trace metals basis) precursor powders according to the following [24] Y After weighing the precursor powders to the required amount to form the Y123 phase of YBCO, they were mixed and milled at 300 rpm in a planetary ball mill for 24 h.The powder was then calcined at 900 °C for a further 48 h with intermittent pestle and mortar grinding stages every 12 h to improve homogeneity of the powder.The resultant YBCO powder was milled again at 300 rpm in the planetary ball mill for a further 24 h then passed through 200, 100 and 38 μm mesh sieves to control the particle size.
X-ray diffraction (XRD) measurements were obtained using a Bruker D2 Phaser benchtop powder X-ray analyser.Particle size distribution analysis was performed on a Malvern Instrument's Mastersizer 3000 with Aero S accessory, where YBCO powder was fed through the system at 4 bar to break up any agglomerations.

Paste Preparation
A YBCO loaded paste was produced using the prepared powder combined with an organic binder solution of 13:9 weight ratio of linseed oil and ICP solvent.An investigation was carried out to determine the highest possible solid loading.Paste mixtures of 70-92.5 wt% were formulated and each were mixed at 3000 rpm for 30 s in a Synergy Devices speed mixer.
Rheology data was obtained using the TA Instruments Rheolyst AR-100 with a 10 mm acrylic parallel plate geometry.A continuous flow curve was recorded over an increasing shear rate from 0 to 100 s À1 .Using a three step viscometric process the thixotropy of the paste can be shown with the time taken to rebuild the viscosity after experiencing a high shear rate.A low shear rate of 1 s À1 is applied followed by a high shear rate of 100 s À1 , lastly the low shear rate is applied again, the viscosity measured throughout the three regimes.Viscoelastic measurements were also taken over an oscillating shear stress from 0.1 to 10 000 Pa while the frequency was maintained at 1 Hz.
Components were printed using the Hyrel Hydra system 30M 3D printer with a 254 μm nozzle diameter and a layer height of 200 μm.Printing speeds were kept to 5 mm s À1 .Post-processing of components consisted of the following in air: drying at 80 °C for 12 h, binder removal at 550 °C for 12 h, sintering at 900 °C between 24 and 120 h.Finally annealing at 900 °C in an oxygen rich environment for 48 h.

Print Characterisation
Scanning electron microscopy (SEM) images were taken using a JEOL 7100.Vibrating sample magnetometry (VSM) and DC transport measurements were performed on a cryogenic cryogen free measurement system (CFMS).Magnetisation of the YBCO samples were recorded as a 6-quadrant loop ((1) 0 to þ5 T, (2) þ5 to 0 T, (3) 0 to À5 T, (4) À5 to 0 T, (5) 0 to þ5 T, and (6) þ5 to 0 T) at 4 K.Using the critical Bean model [25,26] and derived into the standard form using Kim's model [27] the critical current density can be calculated as follows Where a and b are the dimensions of the cross-section perpendicular to the applied magnetic field.In this case dimension a must be greater than b.This can be formulated for the case where a = b to return Where R is the effective radius of the cylinder.In both cases ΔM is the difference in the volume magnetisation between the increasing and decreasing field (see Figure 4a) and R is the effective radius of the sample.For this formula ΔM is in the units of Ac m À1 (1 emuc m À3 = 10 Ac m À1 ). [28,29]lectric transport measurements were obtained from 60 to 100 K heating at a rate of 0.2 K min À1 with an applied current, I, of 0.01 mA, chosen to maximise the signal to noise ratio whilst minimising sample heating.The resistivity was calculated from the measured voltage, V, using the sample length between voltage leads (L x ) and cross-sectional area, A as Sample surfaces were prepared using an abrasive paper prior to attaching electric contacts.

Powder and Paste
Figure 1 shows the characterisation of the resultant YBCO powder that was prepared as outlined in the experimental section.Different reaction temperatures were explored which concluded that 900 °C was the optimal temperature returning the most uniform YBCO. [24]Figure 1a shows the XRD spectra of the YBCO powder prepared by the solid-state route, where the characteristic (013) peak at 2θ = 32.4°and the (103)/(110) peaks between 2θ = 32.7°-32.8°canbe seen.The powder was a combination of five separately prepared batches, which likely had minute differences in their stoichiometry, particle size, and strain, resulting in a shift and broadening of the observed diffraction peaks.As the structure of YBCO crystal is easily altered by oxygenation and stoichiometry, further broadening of the peaks due to disorder within the mixed powder can be expected.This resulted in overlap of the expected (013) and ( 110)/(103) peaks at 32.4°and 32.7°, respectively.
The sieveing process outlined in the experimental section resulted in a particle size distribution shown in Figure 1b with a range from 0.1 to 40 μm (D10, D50, and D90 statistics were 0.363, 3.33, and 16 μm, respectively).Particle size and distribution effects the rheological properties due how the particulates interact with one another and whether the larger particulate will agglomerate and clog the nozzles.The large shoulder seen in this data indicates that the powder was multi-modal, containing two separate distributions.Further refining of the sieving process could be completed to reduce this multi modal distribution, however this would likely result in an altered rheology due to a change in particle interactions. [30]igh quality additively manufactured structures have a high density of desired materials with ideally no porosity, unless engineered to do so.However, the direct ink writing process requires a medium to carry the material for selective material extrusion and in this case we engineered a paste to achieve the highest solid loading possible while maintaining the rheological properties to be able to be printed through the chosen additive manufacturing process.Figure 2 shows the characterisation and optimisation process of the solid loaded pastes to be used in the direct ink writing process.
Figure 2a shows viscometric (rotational) measurements for paste mixtures of 70-92.5 wt% of YBCO in a 13:9 mixture of linseed oil:ICP solvent.Viscometric analysis of the pastes indicates that they all display shear thinning behavioura declining viscosity with increasing shear ratewhich is a requirement for direct ink writing where the paste should flow out of the syringe on demand when a force (the plunger) is applied to the paste.Here, we see that increasing solid loading from 70-90 wt% increased the viscosity due to the increased amount of solid particulates interacting with one another.The 92.5% shows that the solid loading is too high and the paste is unstable rapidly breaking down as the shear rate is increased.
Figure 2b shows a three step viscometric test that simulates the printing process: 1 -Low shear regime,the paste sits at rest within the syringe; 2 -shear regime plunger is depressed applying a high shearing force; and 3 -Low shear regime, the paste sits at rest in the build volume.For high quality printing it is necessary that the paste recovers the high viscosity after the shearing force is removed.This allows the material to remain in place after extrusion and also carry the weight of the following layers printed on top.
Figure 2c shows viscoelastic (oscillatory) behaviour of the 90 wt% solid loading paste.This initial plateau shown with increasing oscillatory stress is the linear viscoelastic region.
Here the storage modulus G 0 is larger than the loss modulus G 00 .This is known as a solid like behaviour meaning that at these stresses the material will not experience a flow.Past this point, G 0 drops and the moduli cross and G 00 is now larger than G 0 .At these stresses the paste is behaving as a liquid and experiencing flow.This is behaviour ideal for direct ink writing as the pastes will flow under a high enough shearing force but retain a solid behaviour when at rest. Figure 2d i-iii shows components printed with 85 wt% (i), 89 wt% (ii), and 90 wt% (iii) pastes.At 85 wt% the printed component had very little shape retention forming an elongated undefined shape with little reminiscence of the desired rectangular bar.At 89 wt there is improved shape retention although it had slumped under its own weight.At 90 wt% showed the printed component experienced little slumping and maintained the desired printed dimensions.

Printed Components
The printing parameters were empirically chosen to achieve the highest resolution possible while maintaining low failure rate due to nozzle clogging and printed components colliding with the nozzle.The loaded pastes printed reliably with a 254 μm nozzle; although 150 and 100 μm nozzles were achievable, frequent clogging occurred causing prints to fail.A layer height of 200 μm was chosen, which is approximately 80% of the extruded width, commonly used for material extrusion processes.3D-modelled components were sliced using the CURA slicing software.A print speed of 5 mm s À1 was used for all movements.A single perimeter was used and a 100% "zigzag" infill was used to avoid the nozzle travelling over previous printed areas to avoid collisions.The loaded paste was printed into 2 Â 2 Â 8 mm rectangular bars for magnetometry and DC transport measurements and to demonstrate stability of the printing and post-sintering process, including ease of printing, volume loss, and resultant superconducting transition.
Figure 3 shows photographs and SEM images of the printed and sintered components.Figure 3a demonstrates the capabilities of the paste: with a decrease in printed feature size to that previously reported, this paste achieved successful and reliable printing with a 254 μm nozzle and 200 μm layer height.Figure 3c shows the cross section of the printed components where the extruded paths have left spaces between one another.This shows that the paste behaved as a solid like material after printing and not continuing to flow and retaining is extruded shape.The larger void that were unable to close during the sintering process are likely due to artefacts from the tool path of the 3D printer.As the nozzle opening is circular, any change in direction by the nozzle will leave a radius on in the our edge of the extruded path, thus leaving a considerable void.This is a common material extrusion techniques and can be addressed with slicing parameters such as overlap with the outer perimeter.
Figure 3d shows the porosity measurements determined through the water saturation method.Samples showed a total porosity between 20-30%.These values are lower than expected as the extruded paste had a 90 wt% solid loading, which corresponds to a volume percentage of 59%, which would suggest a maximum porosity of 41% if the components retained their pre-sintered geometry.However, shrinkage occurred during the sintering process as the particulates consolidated.Printed 2 Â 2 Â 8 mm bars resulted in 1.8 mm, wide, 1.6 mm high, and 7.1 mm long bars therefore shrinking 10-15% during the post-processing.Overall, the samples show similar dimensions, therefore the shrinkage was likely to have happened within the first 24 h of the sintering treatment.The sintering process has had an effect on densifiying the samples with the closed porosity decreasing as they are sintered for longer.The open porosity shows an increase likely due to the out gassing of the internal pores, despite this the overall porosity dropped as sintering time was increased.This densification process increases the percolative path for the superconducting current, it is therefore expected that the resistivity would reduce as the sample was sintered.Figure 3e,f shows further examples of the printing capabilities of complex shapes.Figure 3e shows an array of split ring antennas while Figure 3f shows a space-filling Hilbert curve.These are demonstrations that AM has the potential for applications other than 3D manufacturing.These patterned were processed using the same instrumentation as the bulk samples.

Superconducting Measurements
Figure 4 shows the magnetic and electric characterisation of sintered and annealed printed components.In Figure 4a 6-quadrant magnetization measurements are shown, which exhibit the expected type II superconducting behaviour: a linear diamagnetic response up to H c1 , observed as a negative magnetization as the magnetic field is increased (virgin curve), after which the sample enters a mixed phase state as the density of magnetic vortices increases with continued increasing field.As the field is decreased from its maximum value (in this case 5 T), these magnetic vortices are 'locked in' resulting in a positive net magnetization.This behaviour is repeated in each quadrant of the plot.Only by heating the material above is critical temperature quenching the trapped field lines can we see the virgin curve again.
Analysis of critical current densities have been determined from magnetisation data using the extended Bean model. [22,31,32] normalising the measurements to the mass or volume this method is no longer a characterisation of the sample quality but the quality of the superconducting material.In addition to this, although a demonstration of superconductivity is shown, this does not prove that additively manufactured YBCO components can carry an electric current.Resistive data is absent in the present literature for multi-layer additively produced components.This may be due to either AM being unable to produce electrically superconducting components across the bulk, or that the measurement was unavailable in the research groups investigating it.The current literature also does not explicitly state the nozzle sizes used and therefore it is difficult to estimate the resolution and minimum features that are possible.
The sample that was sintered for 120 h and annealed in an oxygen-rich environment showed the strongest diamagnetic response with a lower critical field H c1 = 33 mT.The breadth of these initial virgin curves is due the samples being polycrystalline with many defects.Figure 4b shows the critical current density, J c , determined from magnetometry data using the extended Bean model (Equation 1).Only the unannealed and annealed samples that were sintered for 120 h are shown for clarity.The sample with the longest thermal treatment (120 h of sintering and 48 h annealing) showed comparable critical current densities to that similarly prepared YBCO 24.Although, our value of 3.6 kA cm À2 is far below the results reported by Mendes (66-150 kA cm À2 ) for similarly prepared YBCO recorded at 4 K. Preliminary transport measurements showed that samples had a highly resistive non superconducting coating, which surface etching before contacts were added resolved.Figure 4c shows the DC transport measurements of the four samples, all of which indicate some transition, and with the exception of the 24 h annealed sample (Figure 4c inset) to zero resistance and a finite resistivity at room temperature.This demonstrates the achievement of a percolative path in the 3D printed multilayer components.
The resistive coating was explored through XRD to determine its origin.Figure 5 shows XRD data of the of the 3D printed sample after sintering for 120 h.Barium fluoride peaks were seen at 2θ of 24.9°, 41.2°and 48.8°formed by the (111), (220), and (113) planes.As barium fluoride was exclusively found on the surfaces of the printed samples, a component of the atmosphere in the furnace during the sintering and annealing stages may have acted as a catalyst for the formation and growth.It is likely due to this impurity that the critical current and transition temperature is lower than the previously reported literature.Barium fluoride being an insulating material it would have not contributed to the electrical conductivity therefore creating a more resistive sample.In addition to this, and the yttrium barium and copper were added in stoichiometric ratios so migration of barium atoms to the surface could have resulted in additional phases containing the left-over yttrium and copper from barium vacancies, resulting in a lowered transition temperature.Furthermore, magnetisation data that is normalised to the mass or the volume of the assumes that the whole sample is contributing to the magnetic signal.Therefore, the barium fluoride is contributing to the lower magnetic signal, thus, lower critical current density.
Analysis of additively produced YBCO only shows magnetic data in the literature.Comparing the magnetometry (Figure 4a) and resistivity data (Figure 4c), of the 24 h sintered samples shows that if a sample displays a typical type II superconducting diamagnetic response this does not necessarily indicate that it can carry a superconducting current.It could now be explained that this may be due to the samples that were printed not showing the zero-resistivity phenomenon.The 120 h sintered sample shows a transition fully into the superconducting state with a residual resistivity of 10 À9 Ω m.Overall, the general trend was that the longer the sample was sintered, the lower the resistivity and higher the superconducting transition temperature.This could be due to the growth of the grains producing fewer grain boundaries and pores that may inhibit the electrical path.Also, the extended sintering time promotes oxygenation even in ambient atmosphere.Therefore, by sintering for longer times the superconducting properties of additively manufactured YBCO components are improved.Annealing in oxygen further reduces the resistivity.This is likely due to the promotion of cooper pairs in the copper planes due to the lack of oxygen

Conclusion
In this work we demonstrate for the first-time electric measurement of additively manufactured multi-layered YBCO components that confirms superconducting current can be passed through a multilayer additively manufactured object.A costefficient paste with simple recipe was reliably produced with a solid state reacted YBCO powder.The paste formulated with inexpensive off the shelf components enabled the highest solid loading of 90 wt% with the rheological properties still enabling use of 254 μm diameter nozzles for high resolution prints.Components printed demonstrated both the Meissner effect through VSM and the characteristic drop to zero resistivity below the critical temperature of 86.5 K.Although this does not match the maximum of 93 K for YBCO it is still possible to reach with liquid nitrogen and could likely be further improved with modifications to post processing.The ability to produce finer components with superconducting behaviour that can pass a current (i.e., with a percolative electronic path) now opens the technology for applications requiring selectively placed superconducting material.Other than producing whole components the technology may be used to repair traditional made components where cracks may have appeared or have become damaged.Future work may go into refining this process to reduce the impurities, increase sample density which would in turn improve the superconducting properties to that comparable to bulk YBCO.This investigation into the direct ink writing process shows potential for other magnetic and electrical material traditionally difficult to process with an ultimate goal of multi material printing of composite structures with complete geometric freedom.

Figure 1 .
Figure 1.Characterisation of the YBCO powder prepared through the solid state reaction.a) XRD with the YBCO peaks highlighted.b) Particle size distribution.

Figure 2 .
Figure 2. Characterisation results of the pastes with a 70-92.5 wt% solid loading.a) Viscometric flow curve.b) Three step thixotropy flow test.c) Viscoelastic measurements of the Storage (G 0 ) and Loss (G 00 ) moduli for the 90 wt% paste d) Photograph of components printed with the i) 85 wt%, ii) 89 wt%, and iii) 90 wt% pastes.

Figure 3 .
Figure 3. Images of the printed YBCO components.a) SEM image of the cross-section.b) SEM image of the side profile.c) Photograph of the printed bars.d) Porosity measurements of the sintered YBCO components.e) Photograph of printed split ring antenna array.f ) Photograph of a printed Hilbert space-filling curve antenna.

Figure 4 .
Figure 4. Characterisation of the printed and sintered YBCO components.a) Mass magnetisation calculated from magnetometry data taken at 4 K. b) Critical current density calculated from the critical Bean model at 4 K. c) DC transport measurements.
This work was supported via the EPSRC Centre for Doctoral Training in Additive Manufacturing (EP/L01534X/1).For the purpose of open access, the author has applied a Creative Commons Attribution (CC BY) licence to any Author Accepted Manuscript version arising.Supporting data will be made available via the Loughborough data repository under https://doi.org/10.17028/rd.lboro.22220674.The authors acknowledge use of the facilities and the assistance of K. Yendall in the Loughborough Materials Characterisation Centre.

Figure 5 .
Figure 5. XRD spectra for the printed yttrium barium carbonate component after the post-processing treatment.