Plasma Spraying of a Microwave Absorber Coating for an RF Dummy Load

coatings

Article Plasma Spraying of a Microwave Absorber Coating for an RF Dummy Load

Andreas Killinger 1, Gerd Gantenbein 2 , Stefan Illy 2, Tobias Ruess 2, Jörg Weggen 2 and Venancio Martinez-Garcia 1,*

1 Institute for Manufacturing Technologies of Ceramic Components and Composites (IMTCCC), University of Stuttgart, Allmandring 7b, 70569 Stuttgart, Germany; [email protected] 2 Institute for Pulsed Power and Microwave Technology (IHM), Karlsruhe Institute of Technology, Hermann-von-Helmholtz-Platz 1, 76344 Eggenstein-Leopoldshafen, Germany; [email protected] (G.G.); [email protected] (S.I.); [email protected] (T.R.); [email protected] (J.W.) * Correspondence: [email protected]

Abstract: The European fusion reactor research facility, called International Thermonuclear Experi- mental Reactor (ITER), is one of the most challenging projects that involves design and testing of hundreds of separately designed reactor elements and peripheric modules. One of the core elements involved in plasma heating are gyrotrons. They are used as a microwave source in electron–cyclotron resonance heating systems (ECRH) for variable injection of RF power into the plasma ring. In this work, the development and application of an alumina-titania 60/40 mixed oxide ceramic absorber coating on a copper cylinder is described. The cylinder is part of a dummy load used in gyrotron testing and its purpose is to absorb microwave radiation generated by gyrotrons during testing phase. The coating is applied by means of atmospheric plasma spraying (APS). The absorber coating is Citation: Killinger, A.; deposited on the inner diameter of a one-meter cylindrical tube. To ensure homogeneous radiation Gantenbein, G.; Illy, S.; Ruess, T.; Weggen, J.; Martinez-Garcia, V. absorption when the incoming microwave beam is repeatedly scattered along the inner tube surface, Plasma Spraying of a Microwave the coating shows a varying thickness as a function of the tube length. By this it is ensured that the Absorber Coating for an RF Dummy thermal power is distributed homogeneously on the entire inner tube surface. This paper describes a Load. Coatings 2021, 11, 801. modeling approach of the coating thickness distribution, the manufacturing concept for the internal https://doi.org/10.3390/ plasma spray coating and the coating characterization with regard to coating microstructure and coatings11070801 microwave absorption characteristics.

Academic Editor: Robert B. Heimann Keywords: microwave; absorber coating; Al2O3/TiO2; APS; 170 GHz; gyrotron; fusion reactor

Received: 17 May 2021 Accepted: 25 June 2021 Published: 2 July 2021 1. Introduction Energy production by means of fusion technology is one of the most ambitious projects Publisher’s Note: MDPI stays neutral that mankind initiated in the last century. The world’s largest research facility so far, which with regard to jurisdictional claims in published maps and institutional affil- is still under construction, is ITER in Cadarache in southern France. Thirty-five nations are iations. collaborating to erect a facility that hopefully proves the feasibility of fusion as a large-scale and carbon-free source of energy [1]. It is based on the so-called tokamak concept that uses a magnetic confinement to heat, compress and hold the plasma in place. Many other research facilities are operated in parallel by international consortia throughout the world. In Germany this is Wendelstein 7-X (W7-X) in Greifswald (stellarator concept) and ASDEX Copyright: © 2021 by the authors. Upgrade in Garching (tokamak concept). The Joint European Torus (JET) in Oxfordshire Licensee MDPI, Basel, Switzerland. in UK is another important European research facility. USA, China, Japan, Korea and This article is an open access article distributed under the terms and India are planning or already running their own research centers. However, besides the conditions of the Creative Commons magnetic confinement principle (tokamak and stellarator), there exist other fusion concepts Attribution (CC BY) license (https:// like inertial confinement that are being developed in parallel. creativecommons.org/licenses/by/ 4.0/).

Coatings 2021, 11, 801. https://doi.org/10.3390/coatings11070801 https://www.mdpi.com/journal/coatings Coatings 2021, 11, 801 2 of 14

Fusion technology involves thousands of components that have to be designed and integrated into an overall functionable system. The core components comprise the vacuum torus vessel and several types of superconducting magnetic coils to heat the plasma and keep it in place. Other important components allow for injection of microwaves for plasma heating, as well as components for plasma diagnostics. Electron–cyclotron resonance heating (ECRH) is one of the main concepts for variable microwave heating of the plasma. Gyrotrons are the main source for the microwave power and are capable of operating in the radio frequency (RF) range of 100–200 GHz with an output power of 1–2 MW. In the case of ITER, they are operated at a frequency of 170 GHz running at a continuous power of 1 MW. In total, 24 gyrotrons will be installed at ITER in the first phase. A major step in the development and commissioning phase of gyrotrons is extensive testing and conditioning of the device. Stable and reliable RF loads are required to drive the gyrotron to full performance. We report on an RF load with a cylinder, covered with RF absorbing material as the main element. The development of absorber coatings used in fusion reactor components began in the last decade of the last century. In order to efficiently work as a microwave absorber in the relevant wave length, ceramic coatings with thickness values in a typical range of 100 to 200 µm are required. Resonant coatings in this thickness range can exceed absorption values of 90% for 140 GHz microwave stray radiation [2]. The most important requirements of the ceramic materials are a high dielectric constant, a reasonable high heat conductivity, high resistance against electromagnetic and thermal loads and a good substrate adhesion [2]. Moreover, a high vacuum compatibility is required to minimize outgassing effects induced by the desorption of volatile substances. Nearly all components in fusion reactors that are situated close to the plasma are operated under high or ultrahigh vacuum conditions. Regarding the requested properties and coating thickness range, thermal spraying in general is the technology of choice to apply oxide and some non-oxide coatings. Plasma spraying as the industrial standard, either operated in vacuum (VPS) or under atmospheric conditions (APS), has proven to be a suitable technique to apply ceramic coatings on various metal components (Cu, Mo, Al, Co, carbon, etc.) for this purpose. Most used absorber materials at present are chromium oxide (Cr2O3), mixed alumina-titania (Al2O3/TiO2) sprayed using APS and boron carbide (B4C) sprayed using VPS. Depending of the component, to which the coating is applied, one of these ceramic materials may be favored. For thermally highly loaded devices that are operated close to the plasma ring, an absorber coating with a higher thermal conductivity may be of advantage. As shown in [3], chromium oxide has some benefits in direct comparison to other oxides like alumina-titania or zirconia based coatings. It also exhibits a higher melting point and can withstand high thermal loads without showing any alterations in its microstructure. An in depth study of the absorption behavior of different APS sprayed oxides ceramic coatings for 140 GHz stray radiation was worked out in [4]. The application in mind was the coating of water-cooled baffle shields that protect cryo-pumps and prevent them from heating up in the microwave stray field of the W7-X fusion reactor in Greifswald. For this purpose, an effective absorber coating was searched. Aluminium oxide, chromium oxide, titanium oxide, zirconium oxide and different mixtures of alumina-titania were evaluated and compared. The study revealed that mixed alumina-titania coatings achieved the highest absorption values of over 90%. However, within different ratios (97/3; 87/13; 60/40), different maximum absorption values were observed, with a 50/50 mixture showing the highest values. Moreover, spray angle and grain size distribution of the initial spray powder also had a tremendous effect of the maximum absorption value. The development of gyrotrons for W7-X as well as ITER has induced some more efforts to develop absorber coatings, especially for testbed dummy loads, as published in [5]. The paper describes the basic concept for a dummy load design with a mixed aluminum titanium oxide absorption coating. A prototype with a steel tube body was tested successfully. It can be regarded as the technical precursor of the copper tube used in this work. In [6], plasma Coatings 2021, 11, x FOR PEER REVIEW 3 of 14

Coatings 2021, 11, 801 3 of 14

work. In [6], plasma sprayed B4C and Cr2O3 as single coatings and mixed bilayer of Cr2O3·B4C were evaluated for high power loads going up to 1 MW/m2. The paper discusses sprayed B C and Cr O as single coatings and mixed bilayer of Cr O ·B C were evaluated advantages4 and drawbacks2 3 of the different materials and emphasizes2 3 the4 importance of a for high power loads going up to 1 MW/m2. The paper discusses advantages and drawbacks high thermal diffusivity which is best for B4C, acceptable for Cr2O3 but somewhat limited of the different materials and emphasizes the importance of a high thermal diffusivity which for mixed alumina-titania. VPS also shows superior results if directly compared to APS is best for B C, acceptable for Cr O but somewhat limited for mixed alumina-titania. VPS due to the lower4 porosity of the2 coatings.3 Since then, more work has been published for also shows superior results if directly compared to APS due to the lower porosity of the further improvements of high-power dummy load design capable of collecting radiation coatings. Since then, more work has been published for further improvements of high-power power pulses in the range 1–2 MW/m2. Several research groups from Japan [7,8] presently dummy load design capable of collecting radiation power pulses in the range 1–2 MW/m2. working at ITER and JT60SA on this subject. Similar effort is still going on to coat bolom- Several research groups from Japan [7,8] presently working at ITER and JT60SA on this subject. Similareters and effort other is highly still going loaded on todevices coat bolometers that are operated and other close highly to the loaded plasma devices ring for that diag- are operatednostic purposes, close to theas described plasma ring in for[9] from diagnostic a research purposes, group as working described at in W7-X. [9] from a research group working at W7-X. 2. Radio Frequency Technical Requirements and Design Principles 2. RadioThe Frequencymain design Technical goals for Requirements the high-power and microwave Design Principles load under consideration can be summarizedThe main design as follows: goals for the high-power microwave load under consideration can be• summarizedThe load should as follows: be capable of absorbing at least 1.5 MW of microwave power at fre- • Thequency load levels should in the be range capable 140 of GHz absorbing to 170 GHz. at least 1.5 MW of microwave power at • frequencyThe back-reflected levels in power the range from 140 the GHz load to has 170 to GHz. be as small as possible. • The back-reflecteddistribution of powerthe microwave from the power load has on to the be absorbing as small as area possible. of the load should be • Theas homogenous distribution as of possible the microwave to avoid power the formation on the absorbing of thermal area hotspots. of the load should be • asThe homogenous mechanical asdesign possible of the to load avoid should the formation be robust of and thermal it has hotspots. to be easy to cool. • TheIf absorbing mechanical coating design is applied, of the load both should the absorbing be robust material and it has and to its be connection easy to cool. to the • Ifunderlying absorbing structure coating is should applied, withstand both the excessi absorbingve heating material and and electric its connection arcing events. to the • underlyingThe overall structure costs for should components withstand and excessive manufacturing heating should and electric be in arcing a reasonable events. • Theframe. overall costs for components and manufacturing should be in a reasonable frame. Figure1 1 shows shows the the general general concept concept we we used used for for the the design design of of the the load: load: It It is is based based on on a symmetric cylindrical structure, where thethe microwavemicrowave beambeam isis axiallyaxially injectedinjected throughthrough an input port from oneone side,side, scatteredscattered byby aa conicalconical mirrormirror onon thethe oppositeopposite sitesite andand fromfrom therethere distributeddistributed toto thethe coatedcoated cylindricalcylindrical surfacesurface by multiple reflections.reflections. Tapering ofof thethe absorbing ceramic layer on the cylindercylinder allowsallows anan adaptionadaption andand balancingbalancing ofof thethe powerpower density on the structure to avoidavoid hotspots,hotspots, especiallyespecially in the casecase ofof thethe firstfirst impactimpact ofof thethe scattered microwavemicrowave beambeam [[5,10].5,10]. This requirementrequirement impliesimplies thatthat thethe thicknessthickness ofof thethe layerlayer increasesincreases from thethe mirrormirror sideside toto thethe inputinput areaarea accordingaccording toto thethe decreasingdecreasing microwavemicrowave power density.density. Finally, Finally, the the required required local local thickness thickness of theof the layer layer is determined is determined by microwave by micro- absorptionwave absorption measurements measurements (see Section (see Section 4.3). 4.3).

Figure 1. Concept of tube-shaped absorber featuring a microwave absorption layer layer with with a a varying varying coating thickness.thickness.

Coatings 2021, 11, x FOR PEER REVIEW 4 of 14 Coatings 2021, 11, 801 4 of 14

For the design of the load we used a relatively simple “ballistic” ray-tracing code, Coatings 2021, 11, x FOR PEER REVIEW For the design of the load we used a relatively simple “ballistic” ray-tracing4 of 14 code, which even does not take into account the wave diffraction. It simply propagates a larger which even does not take into account the wave diffraction. It simply propagates a larger number of single rays through the geometry, resampling an injected Gaussian beam. At number of single rays through the geometry, resampling an injected Gaussian beam. At each reflection point, the given power of the ray is reduced according to the assumed eachFor reflection the design point, of the load given we power used a of relatively the ray issimple reduced “ballistic” according ray-tracing to the assumedcode, absorption coefficient of the respective material until the power of the ray is non-signifi- absorptionwhich even coefficientdoes not take of the into respective account the material wave diffraction. until the power It simply of the propagates ray is non-significant a larger cant or the ray hasornumber left the the ray of structure hassingle left rays the through through structure the through geometry input port. the, resampling input Figure port. 2 an exemplarily Figure injected2 exemplarilyGaussian indicates beam. indicates At how such a singlehoweach ray suchreflection propagates a single point, ray through the propagates given the power geometry, through of the the ray losing geometry, is reduced its power losing according after its power tomultiple the after assumed multiple reflections. Besidereflections.absorption the power coefficient Beside density the of power thedistribu respective densitytion distribution materion theal untilinternal on the power internalwalls ofof walls the the ray ofload, theis non-signifi- load, the the code code also gives informationalsocant givesor the informationray about has left the the about back-reflected structure the back-reflected through power the input power (by port.integrating (by Figure integrating 2 theexemplarily theremaining remaining indicates power power of all raysofhow leaving all such rays thea leaving single load ray the through loadpropagates through the inputthrough the inputport). the port).geometry, losing its power after multiple reflections. Beside the power density distribution on the internal walls of the load, the code also gives information about the back-reflected power (by integrating the remaining power of all rays leaving the load through the input port).

Figure 2.FigureTrace of2. aTrace single of ray a single starting ray at starting the input at position the input (on-axis) position with (on-axis) a very shallow with a angle veryand shallow propagating angle and through Figure 2. Trace of a single ray starting at the input position (on-axis) with a very shallow angle and the proposedpropagating load geometry. through The the loss proposed of power isload qualitatively geometry. indicated The loss by of the power color (starting is qualitatively from red toindicated magenta, by refer to propagating through the proposed load geometry. The loss of power is qualitatively indicated by the color (starting from red to magenta, refer to caption on the right side). caption on the right side). the color (starting from red to magenta, refer to caption on the right side). While the described simple numerical tool has its limitations, it is sufficient to design While the describedand optimizeWhile simple the thedescribed numerical internal simple shape tool numerical of has the its load, toollimitations, has especially its limitations, it theis sufficient scattering it is sufficient to conical design to design mirror. In and optimize theaddition,and internal optimize it shape allows the internal of tuning the shapeload, the variable ofespecially the load, thickness especiallythe scattering of the the absorbing scattering conical ceramic conical mirror. layermirror. In to In avoid addition, it allowsexcessiveaddition, tuning it values theallows variable of tuning the power thicthe knessvariable density of thic onthekness the absorbing wall, of the taking absorbing ceramic into account ceramic layer theto layer avoid measurements to avoid excessive values ofexcessiveof the the absorbers power values density of described the power on inthe density Section wall, on 4.3taking the. Figure wall, into3 taking shows account into the accountthe expected measurements the load measurements profile on the of the absorbers described in Section 4.3. Figure 3 shows the expected load profile on the of the absorbers cylindricaldescribed partin Section in the case 4.3. of Fi 1.5gure MW 3 injectedshows the microwave expected power load at profile a frequency on the of 140 GHz cylindrical part in the case of 1.5 MW injected microwave power at a frequency of 140 in the case of the already optimized design. Each incident beam at the input of the load is cylindrical part inGHz the in case the case of 1.5of the MW already injected optimized microwave design. Eachpower incident at a frequencybeam at the ofinput 140 of the assigned a certain power, when this beam hits the circumferential surface of the load, the GHz in the case loadof the is assignedalready aoptimized certain power, desi whengn. Each this beamincident hits thebeam circumferential at the input surface of the of the power is reduced by the absorption coefficient defined at this point. Adding up the losses of load is assigned load,a certain the power power, is reduced when bythis the beam absorption hits the coefficient circumferential defined at thissurface point. of Adding the up the rays gives the power per area. Since none of the peaks exceeds 125 W/cm2, overheating the losses of the rays gives the power per area. Since none of the peaks exceeds 125 W/cm², load, the power problemsis reduced of by the the thermally absorption well conductingcoefficient water-cooled defined at this copper-chrome point. Adding structure up are not overheating problems of the thermally well conducting water-cooled copper-chrome the losses of the raysexpected gives [5 the]. The power calculated per area back-reflected. Since none microwave of the peaks power exceeds leaving 125 the W/cm², load has a value overheating problemsstructure of are the not thermallyexpected [5]. well The calculatconductinged back-reflected water-cooled microwave copper-chrome power leaving the ofload 2.4% has and a value is therefore of 2.4% and easily is therefore manageable. easily manageable. structure are not expected [5]. The calculated back-reflected microwave power leaving the load has a value of 2.4% and is therefore easily manageable.

FigureFigure 3. ExpectedExpected power power density density on on the the cylindrical cylindrical section section of the of the load load in the in thecase case of 1.5 of MW 1.5 MWinjected injected powerpower at a frequency of of 140 140 GHz. GHz.

Figure 3. Expected power density on the cylindrical section of the load in the case of 1.5 MW injected power at a frequency of 140 GHz.

Coatings 2021, 11, x FOR PEER REVIEW 5 of 14 Coatings 2021, 11, 801 5 of 14

3. Materials and Methods 3. Materials and Methods 3.1.3.1. Copper Copper Collector Collector TheThe copper copper collector collector consists consists of a tubeof a manufacturedtube manufactured by centrifugal by centrifugal casting ofcasting a CuCr1- of a CCuCr1-C GZ alloy GZ (DIN alloy 2.1292), (DIN with 2.1292), the with dimensions the dime summarizednsions summarized in Table1 in. PhotographsTable 1. Photographs of the tubeof the during tube preparationduring prepar areation shown are in shown Figure in4. Figure Error! Reference source not found.4.

TableTable 1. 1.Approximate Approximate tube tube dimensions. dimensions. Dimension Value Unit Dimension Value Unit Inner diameter 500 mm InnerCoating diameter length 500 1200 mm mm CoatingWall thickness length 1200 10 mm mm Wall thickness 10 mm Component weight 185 kg Component weight 185 kg

(a) (b) (c)

FigureFigure 4. Tube4. Tube prior prior to coating to coating operation, operation, mounted mounted on a rotationalon a rotati supportonal support specially specially constructed constructed for this purposefor this (purposea). Before (a). Before sand blasting (b) and after (c). sand blasting (b) and after (c).

TheThe state-of-the-art state-of-the-art for for internal internal coatings coatings of of cylinders cylinders are are miniaturized miniaturized rotating rotating spray spray torches,torches, in in which which part part is is ﬁxed fixed during during thethe coatingcoating processprocess andand the cylindrical surface surface is iscoated coated by by an an axial axial moving moving and and simultaneously simultaneously rotating rotating torch torch [11]. [11 This]. This technique technique is espe- is especiallycially applicable applicable for for coating coating of ofsmall small diameter diameter cylinders, cylinders, such such as as cylinder cylinder liner liner surfaces surfaces in inengine engine blocks blocks with with a a diameter diameter of of typically typically 90 90 mm. mm. However, However, for for larger larger diameters, diameters, the the ro- rotationtation of of the the tubular tubular part part with with a synchronized a synchronized linear linear movement movement of the of torch the torch is more is more mean- meaningful.ingful. In this In case, this case,a handling a handling system system is used is to used position to position the part the exactly part exactly under underrotation rotationwith the with centerline the centerline of the cylindrical of the cylindrical surface surface [12]. This [12 ].approach This approach allows allowsfor using for a using stand- a standardard plasma plasma torch torchsetup setupoperated operated at optimal at optimal spray spraydistance distance (~150 (~150mm) where mm) where the highest the highestparticle particle velocities velocities can be canachieved be achieved [13]. This [13 ].way, This a way,dense a densecoating coating microstructure microstructure and op- andtimal optimal coating coating adhesion adhesion can be can achieved. be achieved. To rotate To rotate the copper the copper collector, collector, a hand-ling a hand-ling system systemwas designed was designed and constructed, and constructed, transmitting transmitting the therotation rotation with with a belt a belt drive, drive, as as shown shown in inFigure Figure 5a.5a. The The torch torch is is mounted mounted to toan an extensio extensionn of ofthe the robot robot system system that that allows allows for enter- for enteringing the thecylinder cylinder and and moving moving in an in axial an axial dire directionction synchronously synchronously to obtain to obtain a 2 mm a 2 mmoffset offset per cylinder revolution, as shown in Figure5b. per cylinder revolution, as shown in Figure 5b.

Coatings 2021, 11, 801 6 of 14 Coatings 2021, 11, x FOR PEER REVIEW 6 of 14

(a) (b)

FigureFigure 5. 5.Tube Tube prior prior to to coating coating operation operation mounted mounted on on a a rotational rotational support support specially specially constructed constructed for for this this purpose. purpose. AAbelt belt drivedrive was was designed designed to to set set the the tube tube into into rotation rotation during during coating coating (a). (a The). The plasma plasma torch torch is mounted is mounted on a on robot-guided a robot-guided cantilever cantilever for internal operation. Plasma torch mount shown in more detail in (b). for internal operation. Plasma torch mount shown in more detail in (b).

3.2.3.2. PowderPowder FeedstockFeedstock InIn aa previousprevious work,work, differentdifferent oxideoxide coatingscoatings werewere analyzed,analyzed, revealingrevealing thatthat coatingcoating thickness,thickness, grain grain size size and and material material composition compositio stronglyn strongly influence influence the the system system absorption absorption [4]. It[4]. turned It turned out that out Althat2O 3Aland2O3 TiO and2 mixedTiO2 mixed powders powders in different in different compositions compositions led to higher led to absorptionhigher absorption values compared values compared to pure oxide to pure ceramic oxide materials, ceramic withmaterials, an increasing with an absorption increasing capabilityabsorption with capability increasing with contentincreasing of TiOcontent2. All of the TiO spray2. All the powders spray hadpowders the same had the particle same sizeparticle distribution size distribution of −20 + of 5 −µ20m + (prevalent 5 µm (prevalent mesh size mesh notation size notation based onbased ASTM on ASTM B214 sieve B214 analysis,sieve analysis, referring referring to a particle to a particle size distribution size distribution of D10 of= 5D and10 = 5 D and90 = D 20,90 = respectively). 20, respectively). The coatingThe coating thickness thickness was fixedwas fixed to 150 toµ 150m [ 4µm]. [4]. TheThe selected spray powder powder by by consid considerationeration of ofprevious previous works works is an is Al an2O Al3/TiO2O3/TiO2 60/402 60/40mixture mixture for APS for coatings APS coatings supplied supplied by Oerlikon by Oerlikon Metco Metco AG (Pfäffikon, AG (Pfäffikon, Switzerland), Switzerland), with withthe characteristics the characteristics summarized summarized in Table in Table 2. 2.

TableTable 2. 2.Oxide Oxide ceramic ceramic Spray Spray Powder. Powder. Given Given are are manufacturer’s manufacturer’s data. data.

Manufacturer Designation Composition Particle Particle SizeSize ( µ(µm)m) 60%60% AlAlO2O,3, D10 == 55µ µmm OerlikonOerlikon MetcoMetco AG AG Amdry 62506250 2 3 40%40% TiOTiO2 2 D90 == 3535µ µmm

3.3.3.3. CoatingCoating DepositionDeposition andand CharacterizationCharacterization AtmosphericAtmospheric plasmaplasma sprayingspraying waswas performedperformed usingusing aa standardstandard F6F6 typetype plasmaplasma gungun fromfrom GTVGTV GmbH GmbH (Luckenbach, (Luckenbach, Germany). Germany). The The spray spray torch torch was was guided guided by a bysix-axis a six-axis robot robotsystem system type typeRX130B RX130B from from Stäubli Stäubli International International AG AG(Pfäffikon, (Pfäfﬁkon, Switzerland). Switzerland). Plasma Plasma pa- parametersrameters were were adjusted adjusted in in order order to to achieve achieve a a well well adherent adherent and and dense coating structure. PowderPowder feedfeed rate rate was was adjusted adjusted to to achieve achieve a suitablea suitable coating coating thickness thickness in the in rangethe range of 15 ofµ m15 perµm sprayper spray pass usingpass using a PF962 a PF962 powder powder feeder feeder from GTVfrom GmbHGTV GmbH (Luckenbach, (Luckenbach, Germany). Ger- Themany). coating The thicknesscoating thickness per past per is adjusted past is ad byjusted modifying by modifying the amount the ofamount powder of injectedpowder duringinjected the during process the adjustingprocess adjusting the rotation the rota speedtion of speed the feed of the disc. feed The disc. optimized The optimized spray parametersspray parameters that were that ﬁnally were finally applied applied for coating for coating operation operation are summarized are summarized in Table in3 Table. 3. Surface activation of the inner tube surface has been done by bead blasting with glass beads (300 µm) and degreased using acetone before and after the bead blasting process in orderTable to3. Optimized enhance the spray mechanical plasma spray adhesion parameter of the set coating used in to coating the substrate. operation The of the considerably tube. soft material had to be treated carefully to avoid implantation of the bead particles and therefor thisParameter operation was performedSymbol using a reducedValue air pressure of 2 bar.Unit A surface Plasma gun current I 500 A

Coatings 2021, 11, 801 7 of 14

roughness with Ra 5.3 µm and Rz 40 µm was achieved. Surface roughness was measured using a contact proﬁlometer type Perthometer Concept from Mahr GmbH (Göttingen, Germany), Rz and Ra values were determined according to DIN EN ISO 4287.

Table 3. Optimized spray plasma spray parameter set used in coating operation of the tube.

Parameter Symbol Value Unit Plasma gun current I 500 A Plasma gun voltage V 75 V Plasma gun power P 37,5 kW . VArgon 44 l/min Plasma gas mixture . V 10 l/min . H2 Powder feed gas VArgon 8 l/min Spray distance h 130 mm Torch transversal speed 1 v 500 mm/s Meander offset b 3 mm Rotational speed ω 20 r.p.m Cooling nozzles distance d 65–70 mm 1 Torch transversal speed measured relative to the moving substrate surface.

The particle size distribution of the Al2O3/TiO2 60/40 mixture powder was determined by means of laser diffraction particle size analysis in a Mastersizer S from Malvern Panalytical Ltd. (Malvern, United Kingdom) in order to ensure the powder specifications. The material was investigated by scanning electronic microscopy (SEM); the sample was sputter-coated with gold and examined in a LEO VP 438 from Leo Elektronenmikroskopie GmbH (Oberkochen, Germany). Prior to spraying on the tube component, testing of the spray process was performed on flat samples that were mounted inside a dummy tube with the same diameter to be coated under similar conditions as the absorber coating. This way, a comparable spray process situation like that on the real component is guaranteed. Test samples of CuCrZr copper alloy (formerly CuCrZr), with discs geometry of 124 mm diameter and 8 mm thickness, were coated and used for coating characterization regarding absorption. Coating characterization was done by metallographic cross section preparation and optical micrograph images of the coatings were performed using a Leica MEF4M (Leica GmbH, Wetzlar, Germany) as well as SEM micrographs. The coating thickness is determined in the copper collector by non-destructible methods, using a portable contact system DUALSCOPE® FMP100 coating thickness gauge from Helmut-Fischer GmbH (Germany), which uses the eddy current testing method (DIN EN ISO 2360) for non-ferrous metals. Temperature measurements during the coating process were performed using an IR pyrometer type Optris CTlaser LTF-SF-CB8 from Micro-Epsilon Messtechnik GmbH & Co. KG. (Ortenburg, Germany). The bond strength was measured according to the pull-off method ISO 14916 with a measuring device on a universal testing machine Zwick Z100 from ZwickRoell GmbH & Co. KG (Ulm, Germany). HTK ULTRA BOND® from HTK Hamburg GmbH (Hamburg, Germany) was used as adhesive and pins with a diameter of 14.2 mm were used as tension rods. Precipi- tation heat treatment of the glue took place at 190 ◦C for 35 min. The test speed was set to 0.5 mm/min. The phase composition of APS coatings and spray powder was investigated using the Advance D8 diffractometer from Bruker Corporation (Billerica, MA, USA) with a Cu Kα beam and a graphite monochromator. Microwave absorption measurements were performed on flat sample coatings as- sprayed with different thicknesses from 30 µm to 150 µm at 140 GHz and 170 GHz. A schematic view of the reflection measurement setup is shown Figure6[ 14]. A millimeter (mm)-wave is transmitted by a corrugated circular horn antenna. The mm-wave is focused on the dielectric coated samples using two mirrors. The reflected signal is separated from the forwarding transmission path using a Mylar foil at 45◦ angle. The reflected signal is received by a corrugated circular horn antenna and is evaluated by a Coatings 2021, 11, x FOR PEER REVIEW 8 of 14

Coatings 2021, 11, x 801 FOR PEER REVIEW 8 8of of 14 on the dielectric coated samples using two mirrors. The reflected signal is separated from the forwarding transmission path using a Mylar foil at 45° angle. The reflected signal is receivedonperformance the dielectric by networka coatedcorrugated analyzersamples circular (PNA).using two Thehorn mi frequencyrrors. ante Thenna rangereflected and of is the sign evaluated PNAal is isseparated limited by a upfrom performance to net- workthe26.5 forwarding GHz analyzer and is transmission increased(PNA). The using path frequencyextension using a Mylar modules,range foil of at covering the45° angle.PNA the Theis frequency limited reflected band up signal to from 26.5is GHz and is increasedreceived140–220 GHz.by using a corrugated extension circular modules, horn ante coveringnna and isthe evaluated frequency by a bandperformance from 140–220net- GHz. work analyzer (PNA). The frequency range of the PNA is limited up to 26.5 GHz and is increased using extension modules, covering the frequency band from 140–220 GHz.

FigureFigureFigure 6. 6.6. SchematicSchematic Schematic view view view of of the the of mm-wave mm-wavethe mm-wave absorption absorption absorption measurement measurement measurement setup for coated setup disks for [14]. [14 coated]. disks [14]. 4.4. Results Results 4. Results 4.1.4.1. Powder Powder Characterization Characterization 4.1. PowderTheThe particle particle Characterization size size distribution distribution analysis analysis of of the the APS APS powder powder for for quality quality control control is is sum- sum- marizedmarized in in Figure Figure 77a.a. The grain size distributi distributionon curve exhibits micrometer-sized particlesparticles withThe D10 =particle 8 µm, D50 size = 17 distributionµm and D90 = analysis 33 µm. The of particle the APS size powder distribution for represents quality acontrol is sum- marizedwith D10 =in 8 Figureµm, D50 7a. = 17 The µm grain and D90 size = 33distributi µm. The onparticle curve size exhibits distribution micrometer-sized represents particles amonomodal monomodal distribution. distribution. ThereThere waswas nono signiﬁcantsignificant ﬁnefine fractionfraction nornor agglomerates in the withdelivereddelivered D10 powder. powder.= 8 µm, The D50 analysis = 17 µmof the and SEM D90 images = 33 is µm.in in accordance accordance The particle with the size results distribution of the represents laser granulometry, showing irregularly shaped particles, as can be seen from Figure7b. alaser monomodal granulometry, distribution. showing irre Theregularly was shaped no particles,significant as can fine be fractionseen from norFigure agglomerates 7b. in the delivered powder. The analysis of the SEM images is in accordance with the results of the laser granulometry, showing irregularly shaped particles, as can be seen from Figure 7b.

(a) (b)

FigureFigure 7. 7. LaserLaser granulometry granulometry ( (aa)) and and scanning scanning electron electron microscope microscope image image ( (b)) of of the the alumina-titania alumina-titania spray powder.

4.2.4.2. Structural Structural and and Microstructural Microstructural Properties Properties of of the the Deposited Deposited Coatings on Flat Samples

(aInIn) a a first first step, step, flat flat copper copper discs discs (CuCr1Zr (CuCr1Zr copper alloy; 124 mm ( bin ) diameter and 8 mm inin thickness) thickness) were were spray-coated employingemploying a meander movement. These These samples samples were Figure 7. Laser granulometrythenthen used (a) and forfor scanning evaluationevaluation electron and and coating coating microscope characterization characterization image regarding (b regarding) of the their alumina-titania their absorption absorption ability spray ability as powder. a asfunction a function of varying of varying thickness thickness starting starting from 30fromµm 30 to µm 150 toµm 150 (i.e., µm 30, (i.e., 45, 60,30, 75,45, 100, 60, 12575, 100, and 4.2.125150 andµStructuralm), 150 obtaining µm), and obtaining the Microstructural same coatingthe same characteristics, coating Properties characteristics, suchof the as Deposited microstructure such as Coatingsmicrostructure and porosity. on Flat and Samples porosity.Inside the tube, environmental conditions are strongly affected by plasma gas and coolingInsideIn a gas first the flows. step,tube, These environmentalflat strongly copper differ discs conditions from (CuCr1Zr a corresponding are strongly copper affected outsidealloy; by124 coating plasma mm process. in gas diameter and To and 8 mm incoolingmimic thickness) thegas coating flows. were These conditions spray-coated strongly that differ predominate employin from a corresponding insideg a themeander copper outside tube,movement. coating a second process. seriesThese To of samples were thenmimiccopper used the disks coating for were evaluation conditions then mounted thatand andpredominate coating coated insidecharacterization inside a dummy the copper steel regardingtube, tube a refer second totheir Figure series absorption8 ofa ability with the same dimensions as the final copper tube, as shown in Figure8b. As described ascopper a function disks were of thenvarying mounted thickness and coated starting inside from a dummy 30 µm steel to tube 150 refer µm to (i.e., Figure 30, 8a 45, 60, 75, 100, withbelow, the the same optimal dimensions thickness as range the final to coat copper the final tube, copper as shown tube in was Figure selected 8b. fromAs described 50 µm to 125below, and the 150 optimal µm), thickness obtaining range the to samecoat th ecoating final copper characteristics, tube was selected such from as 50microstructure µm and porosity. Inside the tube, environmental conditions are strongly affected by plasma gas and cooling gas flows. These strongly differ from a corresponding outside coating process. To mimic the coating conditions that predominate inside the copper tube, a second series of copper disks were then mounted and coated inside a dummy steel tube refer to Figure 8a with the same dimensions as the final copper tube, as shown in Figure 8b. As described below, the optimal thickness range to coat the final copper tube was selected from 50 µm

Coatings 2021, 11, 801 9 of 14 Coatings 2021, 11, x FOR PEER REVIEW 9 of 14

100 µm (i.e., 50, 60, 75, 90 and 100 µm), according to the simulations presented in Section2 to 100 µm (i.e., 50, 60, 75, 90 and 100 µm), according to the simulations presented in Section and the MW absorption measurements on Section 4.3. 2 and the MW absorption measurements on Section 4.3.

(a) (b) Figure 8. TestTest samples mounted in dummy steel tube prior to plasma spray coating operation ( a). Planar Planar test test sample sample after plasma spray deposition of the oxide ceramic (b).). See more details inin text.text.

TheseThese samples samples were were also characterized by by means of microscopic image analysis of coatingcoating cross cross sections sections and by by the the eddy eddy current current testing testing method. Figure Figure 99 showsshows anan exampleexample ofof the the coating coating microstructure microstructure that was achieved under internal coating conditions with a 100 µmµm coatingcoating thickness.thickness. TheThe coppercopper particles particles that that can can be be found found within within coating coating are are due due to topreparative preparative restraints restraints during during cross cross section section polishing polishing treatment. treatment. In the In opticalthe optical microscope micro- scopeimage, image, shown shown in Figure in Figure9a, and 9a, SEM and images,SEM images, shown shown in Figure in Figure9b, dark 9b, dark areas areas represent repre- sentAl2O Al3 rich2O3 rich phases, phases, whereas whereas bright bright areas areas represent represent TiO2 TiOrich2 phasesrich phases andstoichiometric and stoichiometric solid solidsolution solution of Al 2ofO 3Aland2O3 TiOand2 TiO[15].2 [15].

(a) (b)

Figure 9. Microscopic imagesimages ofof coatingcoating cross-sectionscross-sections in in optical optical microscope microscope (a ()a and) and SEM SEM (b ).(b Some). Some copper copper particles particles can can be be found within the ceramic coating near the interface. This happens due to the polishing process during cross-section befound found within within the ceramicthe ceramic coating coating near the near interface. the interface. This happens This happens due to the due polishing to the processpolishing during process cross-section during cross-section preparation. preparation. X-ray diffraction (XRD) patterns acquired on the as-deposited coatings and spray powder,X-ray shown diffraction in Figure (XRD) 10 ,patterns lead to the acquired conclusion on the that as-deposited the coatings coatings deposited and by spray APS powder, shown in Figure 10, lead to the conclusion that the coatings deposited by APS powder,have phase shown composition in Figure of10,γ lead–Al2 Oto3 ,the aluminium conclusion titanate that the (Al coatings2TiO5) anddeposited rutile TiOby APS2 as havemain phase constituents. composition It is well of γ known–Al2O3, that,aluminium due to thetitanate high (Al cooling2TiO5) rates, and rutile a major TiO fraction2 as main of constituents. It is well known that, due to the high cooling rates, a major fraction of α– constituents.α–Al2O3 of the It is initial well phaseknown is that, converted due toγ –Althe 2highO3 during cooling the rates, spray a process.major fraction Aluminium of α– 2 3 2 3 Altitanate2O3 of (Althe2 TiOinitial5) isphase also formedis converted during γ–Al the2O spray3 during coating the process.spray process. A smaller Aluminium quantity ti- of tanateα–Al2 O(Al3 and2TiO anatase5) is also TiOformed2 is also during visible the inspray the coat patterning process. [16]. The A bond smaller strength quantity analysis of α– Al2O3 and anatase TiO2 is also visible in the pattern [16]. The bond strength analysis shows

Coatings 2021, 11, x FOR PEER REVIEW 10 of 14

values of 9.67 ± 0.83 MPa with adhesive fracture mode, which are consistent with the results obtained in other works [4,17].

Coatings 2021, 11, x FOR PEER REVIEW 10 of 14

Coatings 2021, 11, 801 10 of 14

values of 9.67 ± 0.83 MPa with adhesive fracture mode, which are consistent with the re- shows values of 9.67 ± 0.83 MPa with adhesive fracture mode, which are consistent with sults obtained inthe other results works obtained [4,17]. in other works [4,17].

Figure 10. XRD patterns of the as-sprayed APS Al2O3/TiO2 60/40 coating (upper) and spray powder (lower).

4.3. Microwave Absorption Measurements on Flat Samples The microwave absorption measurement results of the coated samples with Al2O3/TiO2 60/40 at 140 GHz and 170 GHz are shown in Figure 11. In addition, previous measurements with a mixture of Al2O3/TiO2 87/13 at 140 GHz are depicted [4]. In each measurement, the resonance is clearly visible at which an absorption maximum is reached. The resonance occurs when the thickness 𝑡 corresponds to 𝜆/4, where 𝜆 is the wavelength in the medium, given by Equation (1):

𝜆 𝜆 𝑐 𝑡= = = . (1) 4 4⋅√𝜖 4⋅𝑓 ⋅ √𝜖 Figure 10. XRD patterns of the as-sprayed APS Al2O3/TiO2 60/40 coating (upper) and spray powder (lower). Figure 10. XRD patternsThe of theresonance as-sprayed condition APS Al2O is3/TiO dependent2 60/40 coating on the (upper) frequency and spray 𝑓 and powder the relative permit- (lower). 4.3. Microwave Absorption Measurements on Flat Samples tivity 𝜖. In particular, the amount of TiO2 affects the dielectric properties of the composite The microwave absorption measurement results of the coated samples with Al O /TiO because of its very high relative permittivity of 𝜖 ≈ 169 at 140 2GHz3 [18].2 The relative 4.3. Microwave Absorption60/40 at 140 Measurements GHz and 170 GHz on are Flat shown Samples in Figure 11. In addition, previous measurements permittivity of Al2O3/TiO2 60/40 is calculated to be ϵr ≈ 17.5, in contrast to ϵr ≈ 14.8 of the The microwavewith a mixture absorption of Al2O 3/TiOmeasurement2 87/13 at 140 results GHz are depictedof the [4coated]. In each samples measurement, with the resonance87/13 mixture. is clearly Changing visible at which the relative an absorption permitti maximumvity of is the reached. mixture The resonance affects the oc- resonance con- Al2O3/TiO2 60/40 at 140 GHz and 170 GHz are shown in Figure 11. In addition, previous cursdition, when as the shown thickness in tFigurecorresponds 11, which to λm/ 4directly, where λ mchangesis the wavelength the absorption in the medium, of the coating thick- measurements withgivennesses. bya mixture EquationIn addition, (1):of Al the2O3 /TiOabsorption2 87/13 ofat the140 coatingGHz are at depicted170 GHz [4].is higher In each compared to 140 λ λ c measurement, theGHz resonance up to a coating is clearly thicknesst = visim ble =of ≈at √1100 which µm.= an0√ absorption. maximum (1)is 4 4 · e 4 · f · e reached. The resonance occurs when the thickness 𝑡 rcorrespondsr to 𝜆/4, where 𝜆 is the wavelength in the medium, given by Equation (1):

𝜆 𝜆 𝑐 𝑡= = = . (1) 4 4⋅√𝜖 4⋅𝑓 ⋅ √𝜖 The resonance condition is dependent on the frequency 𝑓 and the relative permittivity 𝜖. In particular, the amount of TiO2 affects the dielectric properties of the composite because of its very high relative permittivity of 𝜖 ≈ 169 at 140 GHz [18]. The relative permittivity of Al2O3/TiO2 60/40 is calculated to be ϵr ≈ 17.5, in contrast to ϵr ≈ 14.8 of the 87/13 mixture. Changing the relative permittivity of the mixture affects the resonance condition, as shown in Figure 11, which directly changes the absorption of the coating thicknesses. In addition, the absorption of the coating at 170 GHz is higher compared to 140 GHz up to a coating thickness of ≈ 110 µm.

2 3 2 FigureFigure 11. 11.Comparison Comparison of theof absorptionthe absorption measurements measurements of Al2O3 of/TiO Al2 O87/13/TiO at 14087/13 GHz at (red140 GHz (red curve) curve)[4] and [4] Al and2O Al3/TiO2O3/TiO2 60/402 60/40 at 140 at 140 GHz GHz (blue (blue curve) and and at at 170 170 GHz GHz (black (black curve). curve).

Figure 11. Comparison of the absorption measurements of Al2O3/TiO2 87/13 at 140 GHz (red curve) [4] and Al2O3/TiO2 60/40 at 140 GHz (blue curve) and at 170 GHz (black curve).

Coatings 2021, 11, 801 11 of 14

The resonance condition is dependent on the frequency f and the relative permittivity e . In particular, the amount of TiO affects the dielectric properties of the composite Coatings 2021, 11, x FOR PEER REVIEW r 2 11 of 14 because of its very high relative permittivity of er ≈ 169 at 140 GHz [18]. The relative permittivity of Al2O3/TiO2 60/40 is calculated to be er ≈ 17.5, in contrast to er ≈ 14.8 of the 87/13 mixture. Changing the relative permittivity of the mixture affects the resonance 4.4. Coating of the Coppercondition, Dummy as shown Load in Figure 11, which directly changes the absorption of the coating thicknesses. In addition, the absorption of the coating at 170 GHz is higher compared to The copper dummy140 GHz load up to was a coating coated thickness in differ of ≈ent110 stepsµm. to obtain the specified coating thickness distribution in different lengths from 50 µm to 100 µm (i.e., 50, 60, 75, 90 and 4.4. Coating of the Copper Dummy Load 100 µm). The thickness distribution relative to the position in the collector was determined The copper dummy load was coated in different steps to obtain the specified coating ® non-destructively thicknessusing the distribution DUALSCOPE in different FMP100 lengths from from Helmut 50 µm to Fischer 100 µm (i.e.,GmbH 50, 60,(Sindelf- 75, 90 and ingen, Germany) coating100 µm). thickness The thickness gauge. distribution The obtained relative to thecoating position distribution in the collector is wasmeasured determined in increments of 20non-destructively mm. For each using measuring the DUALSCOPE position® FMP100 in the from axial Helmut direction, Fischer four GmbH points (Sindelfin- circumferentially gen,distributed Germany) are coating determined thickness gauge.to calculate The obtained the mean coating value distribution and isstandard measured in increments of 20 mm. For each measuring position in the axial direction, four points circum- deviation, as shownferentially in Figure distributed 12. The are maximum determined tocoating calculate thickness the mean valuetolerances and standard for the deviation, application are specifiedas shown in ±10 in Figure µm. 12To. Therepresent maximum the coating good thickness stability tolerances of the coating for the application process are and the high accuracyspecified of the in ±coating10 µm. Tothickness, represent the the goodtolerance stability limits of the shown coating are process specified and the in high ±5 µm. accuracy of the coating thickness, the tolerance limits shown are specified in ±5 µm.

FigureFigure 12. Coating 12. Coating thickness thickness distribution distribution measurement measurement (blue line) of the (blue coated line) copper of collector;the coated also copper shown are collector; desired valuesalso of shown coating thicknessare desired (red values line) and of tolerance coating limits thickness (black (r dotteded line) lines). and tolerance limits (black dotted lines).

4.5. Results from High Power Tests with a 1 MW Gyrotron 4.5. Results from High Power Tests with a 1 MW Gyrotron The load was tested with a 1 MW 170 GHz gyrotron in short and long pulses up to The load wasquasi tested continuous with a wave1 MW operation. 170 GHz The gyrotron experiments in were short performed and long at the pulses KIT gyrotron up to test quasi continuous wavestand. Foroperation. full power The operation expe (~1riments MW) the were maximum performed pulse length at the is limited KIT gyrotron to 180 s due to test stand. For fulllimitations power operation in the power (~1 supply MW) of the gyrotron,maximum for lower pulse power length (~500 is kW) limited longer to pulses 180 are possible. The applied flow rate for the examples given below was 700–750 L/min. s due to limitations in theFigure power 13a shows supply the temperatureof the gyrotron, behavior for of lower the cooling power water (~500 during kW) a pulse longer of 180 s pulses are possible.duration The applied and an average flow powerrate for deposition the examples of 800 kW. given In the ﬁrstbelow part andwas after 700–750 the end of L/min. the pulse an oscillation of the temperature is visible. In order to have correct calorimetry Figure 13a showsboth oscillationsthe temperature have to bebehavior taken into of account. the cooling This iswater due to during the closed-loop a pulse cooling of system and is not related to the operation of the gyrotron or the load. Due to optimization 180 s duration and an average power deposition of 800 kW. In the first part and after the of the operating parameters, the gyrotron power increases during the pulse. This allows to end of the pulse an oscillation of the temperature is visible. In order to have correct calorimetry both oscillations have to be taken into account. This is due to the closed-loop cooling system and is not related to the operation of the gyrotron or the load. Due to optimization of the operating parameters, the gyrotron power increases during the pulse. This allows to calculate the instantaneous power at the end of the pulse from the temperature value to be more than 900 kW, with still increasing tendency. Very long pulses with a duration of up to 26 min were performed with an RF power of approximately 500 kW. Figure 13b shows a 1572 s pulse, where operating parameters were not modified during the pulse. It is visible that after 200 s the water temperature is stable, indicating a constant output power of the gyrotron. In general, the behavior of the load was very good during the experiments. Arcing in the load, which is a frequent problem in such operations, was very much limited and did not slow down the experimental progress. Visual inspection of the load after the first experimental phase did not reveal any defects.

Coatings 2021, 11, 801 12 of 14

Coatings 2021, 11, x FOR PEER REVIEW 12 of 14 calculate the instantaneous power at the end of the pulse from the temperature value to be more than 900 kW, with still increasing tendency.

(a) (b)

FigureFigure 13. 13. CoolingCooling water water temperature duringduring aa shotshot with with 180 180 s pulses pulse duration duration (a ).(a Cooling). Cooling water water temperature temperature during during a shot a shotwith with 1571 1571 s pulse s pulse duration duration (b). (b).

5. DiscussionVery long pulses with a duration of up to 26 min were performed with an RF power of approximatelySeveral challenges 500 kW.had Figureto be overcome 13b shows when a 1572 applying s pulse, the where internal operating absorber parameters coating were not modified during the pulse. It is visible that after 200 s the water temperature is on the dummy load. Realization of a variable coating thickness as a function of the tube stable, indicating a constant output power of the gyrotron. length needs some preliminary studies to learn more about the spray conditions that con- In general, the behavior of the load was very good during the experiments. Arcing trol deposition rate and efficiency as well as coating morphology, i.e., porosity and phase in the load, which is a frequent problem in such operations, was very much limited and composition inside the tube. These properties are not necessarily the same if compared to did not slow down the experimental progress. Visual inspection of the load after the first coatings produced on small samples under regular conditions. Besides, the manufactur- experimental phase did not reveal any defects. ing of an internal coating especially demands a sophisticated temperature guidance by using5. Discussion air lances close to the plasma torch for the internal cooling and additional air knives for outside cooling of the tube surface. Several challenges had to be overcome when applying the internal absorber coating on For calibration purposes, test samples have to be manufactured, preferably showing the dummy load. Realization of a variable coating thickness as a function of the tube length comparable coating properties. With these data in hand, a correlation function of thick- needs some preliminary studies to learn more about the spray conditions that control ness and absorption values can then be derived. deposition rate and efficiency as well as coating morphology, i.e., porosity and phase compositionWe therefore inside decided the tube. to fabricate These properties the flat samples are not necessarilyinside a steel the tube same to mimic if compared similar to processcoatings conditions. produced The on small cooling samples concept under had regularto be adapted, conditions. as the Besides, heat flow the of manufacturing the massive copperof an internal tube is much coating more especially effective demands than it is a the sophisticated case in the temperature separated flat guidance copper samples by using mountedair lances inside close the to the steel plasma tube. torchDuring for the the coat internaling ofcooling flat samples, and additional the dummy air steel knives tube for reachedoutside a cooling maximum of the temperature tube surface. of between 120 and 170 °C and the copper samples tem- peraturesFor calibrationbetween 100 purposes, and 130 °C. test During samples the have coating to be of manufactured, the copper tube, preferably the tube reached showing acomparable maximum temperature coating properties. of between With these50 and data 60 in°C. hand, In principle, a correlation this has function to be oftaken thickness into accountand absorption when directly values comparing can then be deposition derived. efficiencies and resulting coating thickness duringWe the therefore plasma decided spray process. to fabricate On thethe flat other samples hand, inside the aoverall steel tubetemperature to mimic similarlevels reachedprocess in conditions. the process The were cooling quite concept low (100 had to to 120 be °C adapted, and 50 as to the 60 heat°C, respectively) flow of the massive there- forecopper we do tube not is expect much moreto see effective strong differences than it is the in casecoating in the thickness. separated flat copper samples mountedFinally, inside for implementation the steel tube. Duringof a variable the coating coating of thickness flat samples, in the the copper dummy tube, steel several tube approachesreached a maximumare possible. temperature A continuous of betweenchange of 120 the andcoating 170 thickness◦C and the values copper could samples basi- callytemperatures be realized between by a successive 100 and adjustment 130 ◦C. During of the the transfer coating speed of the of the copper plasma tube, torch the dur- tube ingreached the coating a maximum process. temperature Indeed, this ofwould between have 50made and necessary 60 ◦C. In several principle, trials this on hasdummy to be tubestaken to into exactly account work when out directlyand adjust comparing the torch deposition and tube efficiencies kinematic andto meet resulting the claimed coating thicknessthickness distribution during the plasmafunction. spray In the process. present On case the this other was hand, not feasible the overall due temperature to existing costlevels and reached time restrictions. in the process The were stepwise quite app lowlication (100 to of 120 layered◦C and coatings 50 to 60 therefore◦C, respectively) repre- sentstherefore a good we compromise do not expect between to see strong economic differences viability in and coating technical thickness. requirements. More- over, this approach is scalable and adjustable to almost any size and thickness gradations.

Coatings 2021, 11, 801 13 of 14

Finally, for implementation of a variable coating thickness in the copper tube, several approaches are possible. A continuous change of the coating thickness values could basi- cally be realized by a successive adjustment of the transfer speed of the plasma torch during the coating process. Indeed, this would have made necessary several trials on dummy tubes to exactly work out and adjust the torch and tube kinematic to meet the claimed thickness distribution function. In the present case this was not feasible due to existing cost and time restrictions. The stepwise application of layered coatings therefore represents a good compromise between economic viability and technical requirements. Moreover, this approach is scalable and adjustable to almost any size and thickness gradations.

6. Conclusions The application of an absorber coating with varying thickness into a copper dummy load was successfully conducted in three steps. First, the estimation of the optimal thickness distribution of the coating as a function of the tube length was calculated from raytracing simulation of the microwave beam. In a second step, planar copper samples where coated in varying thickness to achieve the absorption as a function of the coating thickness for plasma sprayed Al2O3/TiO2 60/40. From this calibration, the coating thickness distribution in the tube could then be calculated. It turned out, that a stepwise alteration of the thickness was sufﬁcient for the task and was much easier to realize in the manufacturing process. Careful simulations of the power loading of a high-power RF absorber load with a ray-tracing code have shown that the device can handle up to 1.5 MW without exceeding the maximum allowable power density limits. Several low-power microwave measurements at different frequencies were performed in order to define the optimum absorption and thickness of the layer to get a preferably homogeneous power distribution. These measurements verified the optimum thickness variation along the axis in the order of 50 to 100 um. In the first phase of experiments, the load was operated successfully with a power in the order of 800–900 kW (limited by the available gyrotron). Up to now no limitations of the load have been observed. In the near future it is planned to use this load with a 1.5 MW gyrotron.

Author Contributions: Conceptualization, V.M.-G., A.K., and G.G. conceived and designed the experiments; V.M.-G. made the samples, V.M.-G., S.I., and T.R. performed the experiments; V.M.-G., A.K., S.I., T.R., and J.W. analyzed the data; V.M.-G. technical construction for coatings; V.M.-G., A.K., and G.G. wrote the paper. All authors critically reviewed the content and approved final version for publication. All authors have read and agreed to the published version of the manuscript. Funding: This research received no external funding. Institutional Review Board Statement: Not applicable. Informed Consent Statement: Not applicable. Acknowledgments: We would like to thank Andreas Eigenthaler from the Institute for Manufactur- ing Technologies of Ceramic Components and Composites in Stuttgart for his help with the technical construction and preparation of the samples. We thank Willi Schwan from IFKB at University of Stuttgart for his support in sample preparation and microscope analysis. Conflicts of Interest: The authors declare no conflict of interest.

References 1. Bigot, B. ITER: A unique international collaboration to harness the power of the stars. Comptes Rendus Phys. 2017, 18, 367–371. [CrossRef] 2. Wacker, R.; Leuterer, F.; Wagner, D.; Hailer, H.; Kasparek, W. Characterization of absorber materials for high-power millimetre waves. In Proceedings of the Twenty Seventh International Conference on Infrared and Millimeter Waves, San Diego, CA, USA, 26 September 2002; pp. 159–160. 3. Spinicchia, N.; Angella, G.; Benocci, R.; Bruschi, A.; Cremona, A.; Gittini, G.; Nardone, A.; Signorelli, E.; Vassallo, E. Study of plasma sprayed ceramic coatings for high power density microwave loads. Surf. Coat. Technol. 2005, 200, 1151–1154. [CrossRef] 4. Floristán, M.; Müller, P.; Gebhardt, A.; Killinger, A.; Gadow, R.; Cardella, A.; Li, C.; Stadler, R.; Zangl, G.; Hirsch, M.; et al. Development and testing of 140GHz absorber coatings for the water baffle of W7-X cryopumps. Fusion Eng. Des. 2011, 86, 1847–1850. [CrossRef] Coatings 2021, 11, 801 14 of 14

5. Schmid, M.; Erckmann, V.; Gantenbein, G.; Illy, S.; Kern, S.; Lievin, C.; Samartsev, A.; Schlaich, A.; Rzesnicki, T.; Thumm, M. Technical developments at the KIT gyrotron test facility. Fusion Eng. Des. 2011, 86, 518–521. [CrossRef] 6. Bin, W.; Bruschi, A.; Cirant, S.; Muzzini, V.; Simonetto, A.; Spinicchia, N.; Angella, G.; Dell’Era, F.; Gantenbein, G.; Leonhardt, W.; et al. Absorbing coatings for high power millimeter-wave devices and matched loads. Fusion Eng. Des. 2013, 88, 2510–2514. [CrossRef] 7. Ioki, K.; Hiranai, S.; Moriyama, S.; Tanaka, S. Development of a dummy load and waveguide components for 1 MW CW gyrotron. Fusion Eng. Des. 2016, 109–111, 951–955. [CrossRef] 8. Igami, H.; Kubo, S.; Shimozuma, T.; Yoshimura, Y.; Takahashi, H.; Tsujimura, T.I. Development of a quick-response microwave bolometer for the stray radiation measurement in LHD. In Proceedings of the 2018 43rd International Conference on Infrared, Millimeter, and Terahertz Waves (IRMMW-THz), Nagoya, Japan, 9–14 September 2018; pp. 1–2. 9. Marsen, S.; Corre, Y.; Laqua, H.P.; Moncada, V.; Moseev, D.; Niemann, H.; Preynas, M.; Stange, T.; The W7-X Team. First results from protective ECRH diagnostics for Wendelstein 7-X. Nucl. Fusion 2017, 57, 1–7. [CrossRef] 10. Ives, L.; Mizuhara, M.; Kobayashi, T.; Moriyama, S.; Collins, G.; Borchard, P.; Neilson, J. Design and operation of a 2-MW CW RF load for gyrotrons. IEEE Trans. Electron Devices 2014, 61, 1800–1805. [CrossRef] 11. Barbezat, G. The state of the art of the internal plasma spraying on cylinder bore in AlSi cast alloys. Int. J. Automot. Technol. 2001, 2, 47–52. 12. Killinger, A.; Gadow, R.; Buchmann, M.; López, D. Method and Device for Coating the Interior of Hollow Areas by Thermal Injection. PCT Patent WO 2004/005575 A2, 14 January 2004. 13. Gadow, R.; López, D.; Candel, A. Ceramic and Cermet coatings for cylinder liner in ultralightweight engines-novel processing and manufacturing of ceramic layer composites. Adv. Sci. Tech. 2006, 45, 1330–1335. [CrossRef] 14. Ruess, T.; Gantenbein, G.; Ioannidis, Z.; Rzesnicki, T.; Wagner, D.; Thumm, M.; Jelonnek, J. Frequency measurement techniques for megawatt-class gyrotrons. Gruyter J. 2021. under review. 15. Zhou, J.; Sun, K.; Huang, S.; Cai, W.; Wei, Y.; Meng, L.; Hu, Z.; Li, W. Fabrication and property evaluation of the Al2O3-TiO2 composite coatings prepared by plasma spray. Coatings 2020, 10, 1122. [CrossRef] 16. Ahn, J.; Hwang, B.; Song, E.P.; Lee, S.; Kim, N.J. Correlation of microstructure and wear resistance of Al2O3-TiO2 coatings plasma sprayed with nanopowders. Met. Mater. Trans. A 2006, 37, 1851–1861. [CrossRef] 17. Vicent, M.; Bannier, E.; Benavente, R.; Salvador, M.; Molina, T.; Moreno, R.; Sánchez, E. Inﬂuence of the feedstock characteristics on the microstructure and properties of Al2O3–TiO2 plasma-sprayed coatings. Surf. Coat. Technol. 2013, 220, 74–79. [CrossRef] 18. Schermer, R.T.; Stievater, T.H. Millimeter-wave dielectric properties of highly refractive single crystals characterized by waveguide cavity resonance. IEEE Trans. Microw. Theory Tech. 2019, 67, 1078–1087. [CrossRef]