# Fileset

[2026 Tang Miniature Mechanical Testing to Infer Damage from Accidental and Complex Thermal Exposure for Single Crystal Superalloys.pdf](https://mdr.nims.go.jp/filesets/cdef9fc9-492e-4cf1-8c4f-a2bfd260f780/download)

## Creator

Yuanbo T. Tang, Caspar Schwalbe, Julia Brunthaler, Roger C. Reed, [Satoshi Utada](https://orcid.org/0000-0001-6783-9968)

## Rights

[Creative Commons BY Attribution 4.0 International](https://creativecommons.org/licenses/by/4.0/)

## Other metadata

[Miniature Mechanical Testing to Infer Damage from Accidental and Complex Thermal Exposure for Single Crystal Superalloys](https://mdr.nims.go.jp/datasets/f4aa697e-11f9-4660-9b6f-337c52632b23)

## Fulltext

Miniature Mechanical Testing to Infer Damage from Accidental and Complex Thermal Exposure for Single Crystal SuperalloysTOPICAL COLLECTION: EUROSUPERALLOYS 2026Miniature Mechanical Testing to Infer Damagefrom Accidental and Complex Thermal Exposurefor Single Crystal SuperalloysYUANBO T. TANG, CASPAR SCHWALBE, JULIA BRUNTHALER,ROGER C. REED, and SATOSHI UTADAMechanical testing featuring miniaturized specimens are resource-efficient alternative tostandardized testing that can offer unparalleled insights. In this work, the establishedelectro-thermal mechanical testing procedure was used for evaluating potential damage tomicrostructure given by unexpected thermal/mechanical load. Two conceptualized scenarioswere considered. The first case concerns property deterioration given long-term complexexposure of temperature and stress, using specimens directly extracted from turbine blades. Thesecond case concerns an ‘accidental’ solution treatment at near-incipient melting temperature.Quasi-static and dynamic tests were carried out to pick up potential property change against thereference condition. For the reference microstructure, low cycle fatigue was shown successful inyielding reproducible fatigue life and pick up location-dependent strain hardening responsebetween suction vs. pressure side of a turbine blade. The property difference, given eachscenario, was benchmarked accordingly and rationalized with microscopic evidence.https://doi.org/10.1007/s11661-025-08096-6� The Author(s) 2026I. INTRODUCTIONSUPERALLOY components are built to operateunder extreme loading and thermal profiles. Theirlong-term safe operation is underpinned by high confi-dence life assessment, which requires accurate measure-ments of a range of properties, such as yield strength,low/high cycle fatigue, creep, notch sensitivity, and soon. The qualification testing constitutes a significanteconomic commitment of the alloy and process devel-opment.[1] As driven by the Net Zero 2050 vision, thecurrent sustainability targets in the aerospace sector hasmade gradual change of the technological landscape.[2,3]Evidently, recovery, rejuvenation, and repair are playingan ever greater role in materials technology.[4–6] Along-side with full-scale standardized testing, alternativehigh-throughput testing featuring smaller-scale, evenminiature specimens has gained strong traction thathighlights resource efficiency.[7–9]Various types of mechanical testing featuring minia-turized specimens have emerged in recent years, includ-ing the ones featuring methods based uponindentation,[10] tensile/compression,[11] punching,[12] orelse bending/shearing.[13] Among them, an establishedminiature testing methodology known as elec-tro-thermo-mechanical testing (ETMT) has beenresearched extensively[14] since its inception. The ETMTsystem conducts thermo-mechanical testing within anenvironmental chamber; in addition, it is designed toincorporate electrical resistivity measurements as anovel method for interpretation of microstructuralevolution.[15] Given this unique feature, it has sincebeen utilized creatively for various testing scenarios,such as high-throughput mechanical assessments,[8,16]tracking of recovery, recrystallization,[17] simulation ofdifferential thermal contraction during investment cast-ing,[18] damage assessment for component overheat-ing,[19] and assessment of non-equilibriummicrostructures,[20,21] to name a few.The traditional view on miniaturized testing is thatthe data does not typically go beyond screening andranking purposes—concerns over potential data scat-tering, cross-platform consistency, or else mismatchingYUANBO T. TANG is with the School of Metallurgy andMaterials, University of Birmingham, Elms Road, Birmingham B152SE, UK and also with the Department of Materials, University ofOxford, Parks Road, Oxford OX1 3PH, UK. Contact e-mail:y.t.tang@bham.ac.uk CASPAR SCHWALBE and JULIABRUNTHALER are with the MTU Aero Engines AG, DachauerStr. 665, Munich, Germany. ROGER C. REED is with theDepartment of Materials, University of Oxford. SATOSHI UTADAis with the Department of Materials, University of Oxford and alsowith the Research Centre for Structural Materials, NIMS, 305-0047, 1-2-1 Sengen, Tsukuba, Ibaraki, Japan.Manuscript submitted September 29, 2025; accepted December 22,2025.Article published online January 14, 20261766—VOLUME 57A, MAY 2026 METALLURGICAL AND MATERIALS TRANSACTIONS Ahttp://crossmark.crossref.org/dialog/?doi=10.1007/s11661-025-08096-6&domain=pdfwith full-scale standardized methods. Recent effortshave shown promises in this regard, where the ETMTresults showed exceptional repeatability and consistencyfor quasi-static tensile compared with internationallyrecognized standards—American Society for Testingand Materials (ASTM)[22]—provided the caveats inmeasuring procedures are overcome. On the other hand,the ETMT approach has shown consistently a reducedcreep rupture life compared to ASTM counterpartssuggested by joule heating via electric current.[23]This work will build upon the previously provenETMT methodologies, where we conceptualized twonovel testing scenarios. The case one simulates a seriesof complex thermal and mechanical exposure, which isrepresentative during normal engine operation. The casetwo is to mimick ‘an accidental thermal exposure’during solution treatment. Additionally, we extend theapplicability of testing in two aspects—cyclic loadingbehavior and extraction of location-dependent proper-ties from a thin cross-section in low pressure turbineblade. In each case, we benchmark against the standardmicrostructure as the reference, and we pinpoint andquantify the life debt and property deterioration forintentional thermal and/or mechanical loads to inferpotential damage. Our findings successfully demon-strated pickup of static and dynamic property changesubject to microstructural damage through thermal and/mechanical exposure.II. EXPERIMENTAL METHODOLOGIESA. MaterialsTwo batches of materials were used in the currentstudy for case studies one and two, respectively. Forcase study one, the objective was to directly extractminiaturized specimens from low pressure turbine (LPT)blades with specific microstructure as a function ofextraction location and thermal mechanical history. Theapproach aimed to compare property change fromvirgin blade versus blade after complex cycles of thermaland mechanical exposure. Two LPT blades were selected(Figure 1(a)) and sectioned at the same location (heightand curvature) into sections of c.a. 50 mm � 16 mm � 5mm (h,w,t). The virgin blade was fully heat treated withvirgin microstructure. The second blade was fully heattreated and then subject to a series of complex thermalmechanical exposure on land mimicking in-serviceconditions. Thereafter, the condition is termed as‘pre-exposed.’ The blade was made from SC2000 alloywith its nominal composition shown in Table I.[5]Inverse pole figure (IPF) results show the two materialshave comparable [001] orientation alignment. Specimenswere carefully extracted using wire-based electric dis-charge machining (EDM) with location informationrecorded, i.e., on the suction side or pressure side, seecutting plan in Figure 1(b).Fig. 1—(a) Photo of an example low pressure turbine blade and side view of blade section prior to final machining of miniaturized tests forvirgin condition and after complex thermal mechanical exposure. Corresponding stereographic triangle showing alignment to 001 orientation. (b)Top view of the blade section prior to machining and electrical discharge machining plan showing the location specific extraction of specimens.METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 57A, MAY 2026—1767For case study two, the objective was to conduct LCFtesting and to infer damage of the microstructure due to‘accidental solution treatment.’ The control variable wasthe solution heat treatment temperatures, where anintentional near-incipient melting solution treatmentwas created to simulate ‘accidental treatment’ that isbenchmarked by standard heat treatment. Mar-M-247nickel-based single crystal superalloy was used for thestudy. The nominal composition of the material isshown in Table I. The single crystal material wasreceived in bar form with the longitudinal orientationaligned with [001]. Both bars received had a diameter of22 mm and a length of 90 mm. Two bars were receivedthat came from the same casting, which were thenTable I. Nominal Compositions of Superalloys Used for Current Study in Wt PctNi Cr Co Al Ti Ta W Mo Hf Zr Re C BMar-M-247 Bal 8.4 10.0 5.5 1.0 3.0 10.0 0.7 1.5 0.05 — 0.15 0.015SC2000 Bal 5.0 10.0 5.6 — 8.7 6.0 2.0 0.1 — 3.0 — —Fig. 2—Heat treatment schedule shown schematically for standard heat treatment (a) and near-incipient heat treatment (b). Typical c/c0microstructure for both treatments is shown in (c) and (d), respectively. Specimen geometry used for quasi-static and LCF test (e).1768—VOLUME 57A, MAY 2026 METALLURGICAL AND MATERIALS TRANSACTIONS Asubject to different heat treatment, see a schematicillustration in Figures 2(a), (b). The ThermoCalc simu-lation with TCNI8 database suggests the solidus point is1256 �C.[22] For the standard heat treatment, the solu-tion temperature was achieved by step-wise increasewith 1250�C as maximum temperature according toReference 24. For the near-incipient melting treatment,the solution temperature was achieved by step-wiseincrease with 1290 �C as maximum temperature. Bothsolution treatment was held isothermally for 2 hoursfollowed by double aging treatment according to Ref-erence 25.B. Electro-thermo-Mechanical Testing (ETMT)Mechanical testing was conducted using an Instronelectro-thermo-mechanical testing (ETMT) system. Thespecimens were extracted using wire-based EDM intogeometries shown in Figure 2(e). This geometry wasdifferent from our previous work;[22] it features asmooth transition between the gauge to non-gaugesections with a fillet radius of 20 mm and hence reducingstress concentration during the LCF testing. The samesample geometry was used for all tests including uniaxialtensile. All specimens were grounded using abrasivemedia up to 4000 grits prior to testing.Static tensile tests were conducted at a strain rate of10 �4 s�1 at room temperature for both case one andtwo. Speckle patterns were applied on the specimensurface to allow for tracking of deformation. This isachieved using VHT flame proof high-temperaturepaint, using white primer as the background andsprayed with scattered black paint as tracking speckles.Strain measurement was conducted by Imetrum videoextensometry. The virtual gauge mark was set with anarea of approximately 1 � 1 mm2, and the gauge lengthwas set at 5 mm for RT test. The frame rate was set at 2Hz. 1050 �C tests were conducted on specimens in casestudy two. Heating was delivered by DC current wherethe flat end grips were water cooled.[8] Temperature wasmeasured using type R thermocouple which was spotwelded onto the side of the specimen. A temperaturegradient is therefore established which can be approx-imated as a parabolic profile along the loadingFig. 3—Typical c/gamma0 microstructure for virgin conditions (a), (c) vs. after complex thermal mechanical exposure (b), (d). The extractionlocation, i.e., suction side (a), (b) vs. pressure side (c), (d) for each condition is indicated. Each micrograph also shows a zoomed-in image toillustrate the morphology of c/c0 interfaces.METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 57A, MAY 2026—1769direction.[8,26] The central �3 mm is shown to obtain anequilibrium temperature distribution; thus, the virtualgauge length was set at 3 mm for the high-temperaturetests conducted.[14] In order to account for potentialparasitic voltage, the DC current was reversed bychanging circuit directions. The strain rate conductedwas also 10�4 s�1 and strain was measured by videoextensometry within the central 3 mm for elevatedtesting. Further analysis on strain mapping was con-ducted using MATLAB-based open source software‘Ncorr,’ with detailed procedure provided elsewhere.[27]It is worth noting that the Joule effect heating canpotentially cause a change, usually an acceleration, inthe kinetics of phase transformation, recrystallization ordeformation[23,28] via electroplasticity or electromigra-tion, especially for a long hold. It is, however, unlikelyto impose a significant effect during the relatively shorttests carried out in the present study.[22]Dynamic testing was conducted in load-controlledmode. A strain-controlled fatigue tests was not possiblein the system used as the video extensometry capturerate at 10 Hz was not sufficiently high to be used as acontrol channel. LCF tests were carried out for samplesin case study two under triangular waveform. Timetaken for the maximum and minimum stresses eachtakes 1 s with no dwell, therefore stress frequency was0.5 Hz. The stress ratio (R), i.e., max/min stress, was setas 0.1. The set max stress values were 900, 950, and 980MPa, respectively, yet the applied stress was used as theactual stress level recorded by the load cell. A linearvariable differential transformer (LVDT) was used tomonitor travel between the cross heads.C. Scanning Electron MicroscopeElectron microscopy samples were prepared usingstandard metallography route using serial grinding andpolishing. The final step was finished using 40 nmcolloidal silica suspension for 5 mins in most imagingconditions. c0 size measurements were carried out withsamples subject to electrolytic etching by applying for a3V potential difference in 10 pct H3PO4 solution. c0precipitation imaging was conducted using a ZeissCrossbeam 540 field emission gun scanning electronmicroscope (FEG-SEM). A secondary electron detectorwas used to probe the c/c0 boundary at 15 kV acceler-ation voltage and a probe current of 1 nA.Quantitative analysis of fraction and size distributionof the precipitate was conducted using ImageJ soft-ware.[29] Necessary background smoothing function andbinarization based upon histogram were applied todistinguish c0 precipitates from the matrix. c0 fractionFig. 4—Engineering stress–strain curves for [001] single crystal measured at 1050 �C (a), (b) and room temperature (c), (d). The specimensmeasured were distinguished by the conditions for virgin (a), (c) and after complex thermal mechanical exposure (b), (d), where the sampleextraction location, suction side vs. pressure side, is annotated in the curve, too.1770—VOLUME 57A, MAY 2026 METALLURGICAL AND MATERIALS TRANSACTIONS Aand size analyses were conducted for the dendrite coreregion in all cases using representative area of at least100 lm2. Electron backscattered diffraction (EBSD)mapping was carried out using EDAX DigiView5detector in a Zeiss Gemini300 FEG-SEM system. EBSDpattern quality was recorded as 120 � 160 resolutionwith a step size of 0.1 lm. Collected EBSD indexed datawere postprocessed using EDAX OIM analysisTM 8.Electron channeling contrast imaging (ECCI) wasemployed using a Zeiss Gemini300 FEG-SEM systemat 20 kV acceleration voltage and an angular selectivebackscatter detector with a small stage tilt (~ 5�).III. RESULTS AND DISCUSSIONSA. Case One: Location-Dependent Property ExtractionSubject to Complex Exposure1. MicrostructureTypical microstructure revealing size and distributionof c0 is shown in Figure 3. On the suction side of theblade (convex curvature), the distribution of c0 for virginand complex exposed case is shown in Figures 3(a), (b),where for the pressure side, counterpart (concavecurvature) is shown in Figures 3(c), (d). For the virgincase, the secondary c0 fraction is measured at 77 pct withan average side length of 790 nm (assuming perfectlycubic) for the suction side, whereas for the pressure side,the fraction was 73 pct and average side length was 652nm. The size difference between the two is noticeabledespite not excessive. The location-dependentmicrostructure is considered to originate from thecomplex geometry of the blade which does not experi-ence a homogeneous heat transfer in all locations arisingfrom solidification or subsequent homogenization treat-ment. The different cooling rates from near or above thec0 solvus temperature are known to alter c0 distribu-tion.[30] For the complex exposure case, the suction sidemicrostructure also obtains marginally larger precipitatesize than the pressure side: 582 vs. 524 nm. The volumefraction remained similar at 73 and 72 pct, respectively.The precipitation size reduction in the pre-exposedcase is significant, which is consistently observedregardless of their location found in the blade. Theaverage precipitate size reduced from 790 to 582 nm forthe suction side and reduced from 652 to 524 nm for thepressure side. A plausible explanation is that inversecoarsening has assisted this evolution,[31,32] where largeprecipitates split into smaller ones driven by elasticstresses at the interface as it experiences thermalexposure. Figure 3(b) provides evidence for this, seeannotated arrows, where a large cubic precipitate was inprocess of splitting into smaller ones.The morphology of c0 precipitate in the virginmicrostructure is highly cuboidal with sharp cornersand flat boundaries, which is typical of single crystalsuperalloy with a large lattice misfit.[33] The morphologyhas since changed when that were exposed of thermaland mechanical loads, see Figures 3(b), (d), the cornersof the precipitates show rounded features and non-sharpedges. The rounded features suggest a change, typicallya reduction in the lattice misfit due to localFig. 5—Fractured sample tested at 1050 �C for virgin (a), (b) and pre-exposed (c), (d) conditions. Zoomed-in micrographs of raftedmicrostructure are shown in (b), (d).METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 57A, MAY 2026—1771compositional rearrangement that is facilitated withdefects, such as dislocations at the interface.[34] Furtherexposure, especially at higher temperature (over 1000�C), could promote the formation directional coarsening(rafting).2. Elevated temperature deformation and raftingThe high-temperature deformation is shown inFigures 4(a), (b). For the virgin microstructure, thestress–strain curves seem insensitive to location, wherethe pressure and suction side samples have their curvesoverlapped. Similarly, the pre-exposed specimens alsoobtained comparable stress–strain curves, seeFigure 4(b). All samples showed a similar strain soften-ing behavior postyielding. For avoidance of doubt,similar strain softening is observed for tensile testing atelevated temperatures using conventional furnace withstandard specimen geometries,[10,22] which is attributedto creep strain relaxation instead of overheating due toan early necking—with the aid of DIC, the necking wasfound only at near the rupture point. Despite a slightlylower stress (c.a. 15 MPa) observed in the pressure sidespecimen, it is within the level of measuring uncertaintyat elevated temperatures, especially at a temperatureover 1000 �C where microstructure is highly sensitive totemperature uncertainty. Therefore, we conclude nonotable difference in strength that was observed at1050 �C in between the virgin vs. pre-exposed specimensregardless of sample extraction location. In addition,despite a possibility of a small compositional differencemight incur from the two blades as they were producedfrom different melts. The highly comparable tensilebehavior suggests the potential difference in composi-tion is negligible.Fig. 6—2D strain tensor of a specimen under quasi-static uniaxial tension at five time intervals. Strain revealed in horizontal direction (exx),vertical direction (eyy) and in shear (exy). Bottom left shows the last frame of shear with annotated shear signs.1772—VOLUME 57A, MAY 2026 METALLURGICAL AND MATERIALS TRANSACTIONS AThe post-mortem analysis, however, has shownintriguing insights between virgin and pre-exposedsamples. Figure 5(a), (c) shows the fracture tip ofspecimens of both conditions tested at 1050 �C. Bothrevealed the fracture tip was associated with recrystal-lization, which is expected given a significant fracturestrain of c.a. 20 pct at 1050 �C. Given the presence ofnecking, the actual temperature might be higher thanthis value due to a higher current density at the neckingregion, which further facilitated the recrystallizationevent observed.Furthermore, directional coarsening, the so-calledrafting effect was observed for all samples close to thefracture tip. The c0 phase was originally cuboidal whichnow has become continuous that is perpendicular totensile loading. Although more frequently observed inthe crept samples, it is not surprising it arises forspecimens deformed at a slow strain rate at 1050 �C. Inthe two cases shown, the rafted c0 stripes (dark contrast)are coarsened perpendicular along the tensile axis.The degree of rafting is analyzed using the raftingparameter R to describe the extent of this evolution, asproposed by Ignat et al.[35] For a given area, then raftingparameter R is defined asR ¼ L=2Twhere L is the average length of the precipitate and T isthe average thickness. Thereby, for an idealized unraftedmicrostructure, the R value equals to 0.5. The increaseof R value suggests an increased extent of rafting. In ouranalysis, we have utilized ImageJ software in producingFeret diameters of each isolated c0 strip. For estimation,Feret maximum (or Feret diameter) is taken as L andFeret minimum is taken as T. We conducted theevaluation of 400 lm2 per microstructure. The raftingparameter Rv = 2.5 and Rexp = 3.2 (subscript v forvirgin microstructure and exp for pre-exposed). The c0fraction for the virgin and pre-exposed samples is 0.42and 0.32, respectively, in the rafted region. Bothmicrostructures have experience appreciable rafting,where the exposed sample shows a higher magnitudecompared to the virgin counterpart. The rafting kineticsis facilitated by the reduction of c/c0 coherency and arelaxation of interfacial misfit, which tendency theexposed microstructure has shown even prior to theloading, recap Figure 3. The difference in the c0 fractionof the rafted microstructure might also be attributed topotentially a higher temperature during the final stage ofdeformation when necking has occurred. Nonetheless,the comparable tensile yielding response at 1050 �Csuggests the degree of rafting is unlikely to make asignificant contribution in strengthening response.3. Room-temperature deformation and load dropphenomenonFigure 4(c), (d) shows engineering stress–strain curvestested for both microstructure at room temperature and1050 �C, where sample extraction locations, i.e., suctionvs. pressure side, are annotated. The virgin microstruc-ture, see Figure 4 (c), shows a consistent yield strengthof 837 MPa for both suction and pressure side speci-mens. However, a difference in work hardening behavioris observed. The pressure side specimens consistentlydemonstrated a stronger work hardening response andachieve a higher UTS of 1040 MPa compared to thesuction side (923 MPa). The pre-exposed sample (d), bycontrast, shows higher yield strengths of 884 MPa(suction) and 904 MPa (pressure). An increase in yield islikely contributed by their smaller precipitate size. Onceagain, the pressure side specimen demonstrates a higherUTS of 1021 MPa compared to the suction side (950MPa).An intriguing load drop phenomenon was revealed inall tensile curves at room temperature. As the deforma-tion progresses, there were multiple events of suddendrop of engineering stress which is followed by rapidrecovery. This has been consistently observed for allsamples in case study two as well, in total 10 specimens,therefore unlikely to be isolated events such as sampleslipping. On the macroscopic level, Figure 6 shows the2D strain tensor of the deformation process for pre-ex-posed (suction side) sample at room temperature at fivedifferent strain intervals until global strain reaches 14pct. The exx, eyy , and exy maps illustrate deformationoccurring in horizontal, vertical, and shear, respectively.As a reference, the gauge section is also indicated. As theFig. 7—(a) Example of a speckled specimen under loading (horizontally) and loading profile of 0.5 Hz and R =,0.1. (b) Typical engineeringstress-strain curves for standard and near-incipient heat treatment at room temperature. (c) Fatigue cycle to failure of each heat treatmentcondition plotted against maximum stress applied.METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 57A, MAY 2026—1773Fig. 8—Backscattered electron imaging at two magnifications and electron channeling contrast imaging (ECCI) showing typical dislocationstructures within the c/c0 microstructure in shear bands for standard (a), (c), (e) and near-incipient heat treatment (b), (d), (f).Table II. LCF Cycle to Failure for Standard HT and Near-Incipient HTMaterials Set Stress Measured Stress Cycles to Failure (Nf) Average NfMPa MPaStandard HT 900 884 29421 28730900 881 28039950 932 19376 18211950 930 17046980 969 15539 15695980 969 15851Near-incipient HT 900 887 27882 19150900 888 10417950 938 11480 10662950 935 9844980 967 4951 7160980 969 93691774—VOLUME 57A, MAY 2026 METALLURGICAL AND MATERIALS TRANSACTIONS Adeformation progresses, the near-incipient case shows arelatively flat region where strain hardening is minimal,it is indicative that deformation was initially localizedand then gradually propagated to achieve homogeneity,resembles the ‘Lüders band’ like deformation, despitewithout a distinctive drop in stress.[36,37] This is sup-ported by the DIC evidence as shown in Figure 6, whereplasticity onset was indeed a localized (first frame) to aregion much smaller than the gauge, which thenpropagated homogeneously across the entire gauge.The exy shear, however, shows an interesting feature,where both positive and negative shear strain isobserved at the same sample throughout the deforma-tion. Even when the specimen was initially deformed, asmall shear strain was detected to develop in the gaugedue to lattice rotation. Simultaneously, an opposite signshear is observable at above and below the deformedregion even at small level of deformation. The fixed endsare physical constraints to local lattice rotation, as if thespecimen is ‘fighting the grips.’ As deformation pro-ceeds, the gauge volume will work harden and achieve ahigher yield strength compared to the non-gauge part.For a gradual transition in dog bone shoulder (largefillet radius), such as the case studied here, thenon-gauge part would undergo some deformation.Therefore, as shown in the exy shear map, when twoopposite and concurrent lattice rotations are in action,plastic instability could be promoted and causes rapidelastic recovery due to countering rotation. This isconsidered one potential cause to the macroscopicallyobserved load drop and its immediate recovery. Forsingle crystal materials, load drop phenomenon is oftenreported in much smaller scales, such as compression ofmicro-pillars, and often attributed to dislocation burstsor avalanche.[38,39] It is, however, rarely reported forbulk specimens with a cross-section on the order ofmm2, especially the load drop was significant andconsistently observed. It could be associated with theburst of planar faults within c0 instead of dislocations inthe c channels, those are of higher energy penalty andmore likely to induce a response on the meso/macro-scopic scale. Lastly, it is interesting to note that the loaddrop phenomenon is only observed in the currentsample geometries, and not observed previously forsimilar single crystal superalloys when tested under amuch sharper fillet transition.22 Hence, it is consideredsensitive to the gripping constraints, and the fillet radius.B. Case Two: Accidental Thermal Load1. MicrostructureA typical c/c0 microstructure after standard HT isshown in Figure 2(c). The secondary c0 precipitate ismeasured as 64 pct in volume fraction with veryoccasional occurrence of tertiary c0 found in c channels.The precipitates exhibit a cuboidal morphology, and theaverage side length is 678 nm assuming precipitates areperfectly cubic. The near-incipient HT counterpartshowed a reduced volume fraction of 58 pct in sec-ondary precipitates, where the morphology also changedFig. 9—The BSE SEM near carbides, inverse pole figure map and Kernal average misorientation (KAM) map near fracture tip of LCF samplestested at 980 MPa for standard treated (a) through (c) and near-incipient treated (d) through (f). Histogram of the KAM map is shown in (g).METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 57A, MAY 2026—1775from cuboidal to octodendritic demonstrating signs ofsplitting up. The average c0 size has reduced to 412 nm.On the other hand, tertiary c0 occurrence is much morefrequent in c channels, which is likely formed duringcooling and subsequent aging.2. Shear banding in quasi-static tensileThe engineering stress-strain curves for eachmicrostructure is shown in Figure 7 (b). Both standardand near-incipient HT samples obtained a comparableyield strength of 804 and 806 MPa, respectively. Thework hardening behavior, however, is significantlydifferent between the two. The near-incipient meltingcase showed a plateau of stress up to approximately 10pct engineering strain with a gradual work hardeningafterward until fracture at 25 pct that reached anultimate tensile strength (UTS) of 1013 MPa. Bycontrast, the standard HT case demonstrated a stronghardening response after yielding. The plastic regimeshows a near-linear work hardening until it fractures at18 pct with the UTS of 1204 MPa. In avoidance ofdoubt, the standard HT obtained in the current study isdifferent from our previous study,[22] which resultsindifferent work hardening behavior.Microscopically, the two microstructures have showndifferent tendencies of developing shear banding. Asshown in Figure 8, microscopic shear bands of thefractured specimens subject to quasi-static tensile can befound in the standard microstructure but much lessfrequently compared to the near-incipient HT sample.Examples of higher magnification images are shown inFigures 8(c), (d). It is apparent that shear banding ismore frequently observed in the near-incipient case,which attributes to more sites of strain localization thatwould adversely affect the fatigue behavior. ECCIimaging of some shear bands is shown in Figures 8(e),(f). The bands are shown as intense localization ofdislocations and/or faults with a width that is just belowone micrometer. Although it is challenging to determinethe exact crystal defects using ECCI alone, such asdislocations and planar defects like anti-phase boundary(APB) and different types of stacking faults,[40] theiroccurrence can be assumed given their location accord-ing to the well-established References 41, 42, i.e.,dislocations in c matrix channels and planar faultswithin c0.For the standard HT microstructure in Figure 8(e),dislocations and faults do not show a strong tendencytoward locations – they were frequently found both in cchannels and within c0 precipitates. However, for thenear-incipient HT case, see Figure 8(f), dislocations andfaults are primarily observed within the c channels andless associated within the c0 precipitates. The differentialbehavior agrees with the work hardening rates observed.Shearing stress required of c0 precipitates via stackingfaults and/or APB is considerable, where order harden-ing promotes a stronger work hardening response forthe standard HT case in comparison with piling up ofdislocations at c channels in the near-incipient HT case.Given a distinctive c0 morphology, distribution andfraction between the two cases, it is speculated that alevel of compositional variation in c0 precipitates exists,which the APB energy is sensitive to and therebyaffected the work hardening rates observed.3. Low cycle fatigueThe cycles to failure plot for standard and near-in-cipient HT are shown in Figure 7(c), where the fatiguelife data are shown in Table II. The maximum stressesare plotted as measured rather than set stress. Thestandard HT specimens has shown good reproducibilityof cycles to failure. At each stress level, two tests wereperformed and both specimens exhibited very similarfracture cycles, see Table II. The variation in cycles tofailure between the repeats is not more than 8 pct in theworst case. The trend line plotted suggested a linearrelationship for max stress to the logarithmic of cycles tofailure. The excellent reproducibility of LCF resultssuggests the testing procedure applied is sufficientlyrobust as well as the homogeneity of standard HTmicrostructure is satisfactory.Fig. 10—The LVDT travel of fatigue samples at a stress level from 900 to 980 MPa for the initial 10 cycles for both heat treatments.1776—VOLUME 57A, MAY 2026 METALLURGICAL AND MATERIALS TRANSACTIONS AOn the contrary, the near-incipient HT demonstratesa large level of scatter in cycles to failure. The trend lineis plotted based upon the average values at each stresslevel. At the set stress of 900 MPa, the two specimensfailed at 10417 and 27882 cycles, respectively, differed bya factor of 2.7 time. Similarly at the set stress level of 980MPa, specimens failed at 4951 and 9369 cycles, respec-tively, demonstrate a variation of 1.9 times. Comparedto the standard HT case, not only a large scatter is seenin the LCF life for the near-incipient HT, it also shows amarkedly reduced cycles to failure. For instance, theaverage failure cycle reduced from 28730 to 19149 at thelowest stress level and from 15695 to 7160 at the higheststress level. The fatigue life reduction was 33, 41, and 54pct at max stresses are 900, 950, and 980 MPa,respectively.The reduction of fatigue life is considered caused bythe very different work hardening behaviors comingfrom the two HTs. In the case of a fix load, materialswith higher work hardening rate will be strained to alower strain level during the first cycle. The difference inthe initial plasticity that took place can be considerable.For the given stress-strain curves in Figure 7(b), the firstcycle would produce an engineering strain of �4pct inthe standard HT case, which it would be of �16 pct inthe case of near-incipient HT. The large pre-strainwould have induced large density of dislocations andplanar faults, which might already caused some hardparticles to fracture, such as MC carbides, and therebyincrease the number of fatigue initiation sites. Figure 9shows SEM, IPF and Kernal average misorientation(KAM) map near the fatigue fracture in each case. TheKAM histogram (g) demonstrates that the a massivelyincreased local lattice misorientation observed in thenear-incipient case, attributed to remnant of pre-defor-mation during the first fatigue cycle. This is particularlyintense in the vicinity of carbides. Eventually, theirreversible damage has affected subsequently fatiguepropagation through dislocation and particle fracture,which leads to a considerable fatigue debit in thenear-incipient case.The varied work hardening behavior between the twomicrostructures is considered to originate from the c0size and distribution. The near-incipient HT case has abimodal c0 distribution, and the mean size of secondaryc0 is 412 nm with appreciable tertiary precipitates. Incontrast, the standard HT case has a predominatelyunimodal distribution with only sparsely distributedtertiary c0. The increased work hardening response in theunimodal coarse-c0 has been consistently observedelsewhere, such as in RR1000.[43] The increased Orowanlooping and cross slipping activities around the largerprecipitates are associated with the observed hardeningbehavior.The scatter of LCF is discussed next. Since the workhardening rate in the near-incipient case is low, it ishence more prone to have a substantial deformationduring the first fatigue cycle at a given stress. Tovisualize the pre-deformation of each specimen, thecross-head travel distance is used as a proxy for theinitial ten LCF cycles of all testing conditions as shownin Figure 10. Clearly, for the standard HT case, the levelof initial deformation is highly similar to each other,which gave reproducible LCF life. In contrast, thenear-incipient HT case experienced a vastly differentlevel of plasticity after the initial loading. The scatter inthe initial plasticity suggests a varied level of crackinitiation sites that lead to LCF response. In addition,the markedly different initial deformation manifestsinconsistent work hardening response in each sample,suggesting the microstructure in the near-incipient casewould show a degree of heterogeneity.IV. SUMMARY AND CONCLUSIONIn this work, we highlight the use of ETMT miniaturetesting as an alternative to standardized testing for twoconceptualized scenarios. It has shown potential ingaining insights with excellent consistency in probinglocation-dependent properties from components anddynamic properties. The following specific conclusionscan be drawn.1. Miniaturized specimens extracted from the suctionand pressure side of virgin low pressure turbineblades demonstrated similar yield strength at roomtemperature and 1050 �C. However, the suction sideshows more pronounced work hardening behaviorthan the pressure side at room temperature.2. Specimens after complex thermal–mechanical expo-sure demonstrates a higher yield strength at roomtemperature, which is considered due to their finerprecipitation sizes obtained from inverse coarsen-ing. No significant differences observed forhigh-temperature testing at 1050 �C between thetwo microstructures with/without the exposure,despite a much greater extent of rafting observedin the exposed condition.3. The near-incipient case showed similar yieldstrength compared to the standard case. However,the near-incipient case shows a significantly reducedwork hardening rate and ultimate tensile strength incomparison. This is attributed to change of sec-ondary precipitates morphology from cuboidal tooctodendritic with a reduction in size and volumefraction yet an increased occurrence of tertiaryprecipitates.4. The LCF behavior for the standard heat treatmentshows excellent repeatability from all stress rangestested, suggesting robust consistency of the minia-turized dynamic testing procedure. The near-incip-ient case, however, showed markedly reducedaverage fatigue life and significantly large scatterin fatigue life.METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 57A, MAY 2026—1777ACKNOWLEDGMENTSThe authors are grateful for MTU Aero EnginesAG (Germany) for provision of materials used in thisstudy and for the insightful technical discussions withDr Thomas Göhler.CONFLICT OF INTERESTOn behalf of all authors, the corresponding authorstates that there is no conflict of interest.OPEN ACCESSThis article is licensed under a Creative CommonsAttribution 4.0 International License, which permitsuse, sharing, adaptation, distribution and reproductionin any medium or format, as long as you give appro-priate credit to the original author(s) and the source,provide a link to the Creative Commons licence, andindicate if changes were made. The images or otherthird party material in this article are included in thearticle’s Creative Commons licence, unless indicatedotherwise in a credit line to the material. If material isnot included in the article’s Creative Commons licenceand your intended use is not permitted by statutoryregulation or exceeds the permitted use, you will needto obtain permission directly from the copyrightholder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.REFERENCES1. R. Schafrik and R. Sprague: Key Eng. Mater., 2008, vol. 380, pp.113–34.2. G. McDonald and E. Carichael: 2024 global sustainability inaerospace and defense report, Tech. rep, KPMG International,2024.3. L. Barbara, P. Gomez, G. Parolini, and A. de Quero: Global avi-ation sustainability outlook 2025 white paper, Tech. rep, WorldEconomic Forum, 2025.4. I. Lopez-Galilea, L. Hecker, A. Epishin, D. Bürger, B. Ruttert, P.Thome, S. Weber, and W. Theisen: Metall. Mater. Trans. A, 2023,vol. 54A(5), pp. 1509–25.5. M. Panella, L. Zheng, M. Futoma, C. Schwalbe, D. Eyidi,P. Villechaise, and J. Cormier: Analysis of the recovery potentialthrough repair of key tensile properties for a second generationNi-based SX superalloys exposed to simulated routine serviceconditions, in: Superalloys 2024, (2024).6. D. Reker, R. Sowa, C. Schwalbe, B. Boettger, F. Seidel,M. Panella, K. Moehwald, M. Nicolaus, and W. Tillmann:Improving repair braze gap strength through the development of anovel superalloy filler, in: Superalloys 2024, (2024).7. S. Pahlavanyali, A. Rayment, B. Roebuck, G. Drew, and C. Rae:Int. J. Fatigue, 2008, vol. 30(2), pp. 397–403.8. S. Sulzer, E. Alabort, A. Németh, B. Roebuck, and R. Reed:Metall. Mater. Trans. A, 2018, vol. 49, pp. 4214–35.9. S. Korber, S. Wolff-Goodrich, R. Volkl, and U. Glatzel: Metals,2022, vol. 12(7), p. 1081.10. H. Tammpere, P. McKeown, J. Miller, C. Fang, E. Curtis, M.Gaiser-Porter, M. Burley, J. Campbell, M. Artiles, Y. Tang, S.Utada, R. Reed, and T. Clyne: Adv. Eng. Mater., 2024, vol. 26(21),p. 2301073.11. A. Girard, V. Grolleau, and D. Mohr: Int. J. Solids Struct., 2025,vol. 320, p. 113527.12. S. Soltysiak, M. Selent, S. Roth, M. Abendroth, M. Hoffmann, H.Biermann, and M. Kuna: Mater. Sci. Eng. A, 2014, vol. 613, pp.259–63.13. T. Beerli, C.C. Roth, and D. Mohr: Acta Mater., 2024, vol. 263, p.119539.14. B. Roebuck, D.C. Cox, and R.C. Reed: An innovative device forthe mechnical testing of miniature specimens of superalloys, inSuperalloys 2004. K.A. Green, T.M. Pollock, H. Harada, T.E.Howson, R.C. Reed, J.J. Schirra, and S. Walston, eds., Metals &Materials Society, The Minerals, 2004, pp. 523–28.15. S. Utada, R. Sasaki, R.C. Reed, and Y.T. Tang: Metall. Mater.Trans. A, 2023, vol. 54A(5), pp. 1549–67.16. K.A. Christofidou, H.T. Pang, W. Li, Y. Pardhi, C.N. Jones, N.G.Jones, and H.J. Stone: Microstructural control and optimizationof haynes 282 manufactured through laser powder bed fusion, inSuperalloys 2020. S. Tin, M. Hardy, J. Clews, J. Cormier, Q. Feng,J. Marcin, C. O’Brien, and A. Suzuki, eds., Springer InternationalPublishing, Cham, 2020, pp. 1014–23.17. S. Perry, N. D’Souza, D. Collins, B. Roebuck, and H. Dong:Metall. Mater. Trans. A, 2023, vol. 54A(5), pp. 1582–96.18. Y.T. Tang, N. D’Souza, B. Roebuck, P. Karamched, C. Panwi-sawas, and D.M. Collins: Acta Mater., 2021, vol. 203, p. 116468.19. S. Utada, R. Sasaki, R. Reed, and Y. Tang:Mater. Des., 2022, vol.221, p. 110911.20. J.F.S. Markanday, N. D’Souza, N.L. Church, J.R. Miller, J.J.C.Pitchforth, L.D. Connor, S. Michalik, B. Roebuck, N.G. Jones,K.A. Christofidou, and H.J. Stone: The relationship betweenstrain-age cracking and the evolution of c0 in laser pow-der-bed-fusion processed ni-based superalloys, in Superalloys2024. J. Cormier, I. Edmonds, S. Forsik, P. Kontis, C. O’Connell,T. Smith, A. Suzuki, S. Tin, and J. Zhang, eds., Springer, Cham,pp. 822–35 (2024).21. Y.T. Tang, C. Panwisawas, B.M. Jenkins, J. Liu, Z. Shen, E.Salvati, Y. Gong, J.N. Ghoussoub, S. Michalik, B. Roebuck, P.A.Bagot, S. Lozano-Perez, C.R. Grovenor, M.P. Moody, A.M.Korsunsky, D.M. Collins, and R.C. Reed: Addit. Manuf., 2023,vol. 62, p. 103389.22. Y.T. Tang, C. Schwalbe, M. Futoma, B. Roebuck, S. Utada, andR.C. Reed: Metall. Mater. Trans. A, 2023, vol. 54, pp. 1568–81.23. R. Sasaki, S. Utada, Y. T. Tang, C. A. Nunes, and R. C. Reed:The effect of direct current heating on the creep behaviour ofpolycrystalline ni-based superalloys, in: Superalloys 2024, (2024).24. K. Harris, G. Erickson, and R. Schwer: Mar m 247 derivations -cm 247 lc ds alloy and cmsx single crystal alloys: Properties &performance, in: Superalloys 1984, (1984).25. B. Mansoz, L. Bortoluci Ormastroni, J. Rame, C. Schwalbe,K. Vamsi, P. Caron, J. Cormier, and F. Pettinari-Sturmel: Inter-metallics, 2023, vol. 161, art. no. 107976.26. M. Zupan, M.J. Hayden, C.J. Boehlert, and K.J. Hemker: Exp.Mech., 2001, vol. 41, pp. 242–47.27. J. Blaber, B. Adair, and A. Antoniou: Exp. Mech., 2015, vol. 55,pp. 1105–22.28. A. Nicolaÿ, J.-M. Franchet, N. Bozzolo, and J. Cormier: Metal-lurgical Analysis of Direct Aging Effect on Tensile and CreepProperties in Inconel 718 Forgings, in: J. C. J. C. Q. F. M. H. J.M. A. S. Sammy Tin, Chris O’Brien (Ed.), Superalloys2020, TMS,(2020).29. J. Schindelin, I. Arganda-Carreras, E. Frise, V. Kaynig, M. Lon-gair, T. Pietzsch, S. Preibisch, C. Rueden, S. Saalfeld, B. Schmid,J. Tinevez, D. White, V. Hartenstein, K. Eliceiri, P. Tomancak,and A. Cardona: Nat. Method, 2012, vol. 9, pp. 676–82.1778—VOLUME 57A, MAY 2026 METALLURGICAL AND MATERIALS TRANSACTIONS Ahttp://creativecommons.org/licenses/by/4.0/http://creativecommons.org/licenses/by/4.0/30. R. J. Mitchell, M. Hardy, M. Preuss, and S. Tin: Development ofc0 morphology in a P/M Rotor Disc Alloys During Heat Treat-ment, in: Superalloys 2004, Superalloys 2004, pp. 361–70 (2004).31. C. Su and P. Voorhees: Acta Mater., 1996, vol. 44(5), pp. 1987–99.32. S. Meher, L.K. Aagesen, L.J. Carroll, M.C. Carroll, and T.M.Pollock: Microsc. Microanal., 2016, vol. 22(3), pp. 1258–59.33. R. Ricks, A. Porter, and R. Ecob: Acta Mech., 1983, vol. 31, pp.43–53.34. H. Long, H. Wei, Y. Liu, S. Mao, J. Zhang, S. Xiang, Y. Chen, W.Gui, Q. Li, Z. Zhang, and X. Han: Acta Mater., 2016, vol. 120, pp.95–107.35. M. Ignat, J.-Y. Buffiere, and J. Chaix: Acta Metall. Mater., 1993,vol. 41(3), pp. 855–62.36. W. Mason: The Lüders’ lines on mild steel, in: Proceedings of thePhysical Society of London, Vol. 23, (1910).37. B. Rowlands, C. Rae, and E. Galindo-Nava: Prog. Mater. Sci.,2023, vol. 132, p. 101038.38. S. Oh, M. Legros, D. Kiener, and G. Dehm: Nat. Mater., 2009,vol. 8, pp. 95–100.39. Y. Cui, E. Aydogan, J.G. Gigax, Y. Wang, A. Misra, S.A. Maloy,and N. Li: Acta Mater., 2021, vol. 202, pp. 255–65.40. T. Smith, Y. Rao, Y. Wang, M. Ghazisaeidi, and M. Mills: ActaMater., 2017, vol. 141, pp. 261–72.41. C. Rae and R. Reed: Acta Mater., 2007, vol. 55(3), pp. 1067–81.42. N. Karpstein, M. Lenz, A. Bezold, M. Wu, S. Neumeier, and E.Spiecker: Acta Mater., 2023, vol. 260, p. 119284.43. A. Harte, M. Atkinson, A. Smith, C. Drouven, S. Zaefferer, J.Quinta da Fonseca, and M. Preuss: Acta Mater., 2020, vol. 194,pp. 257–75.Publisher’s Note Springer Nature remains neutral with regard tojurisdictional claims in published maps and institutional affiliations.METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 57A, MAY 2026—1779 Miniature Mechanical Testing to Infer Damage from Accidental and Complex Thermal Exposure for Single Crystal Superalloys Abstract Introduction Experimental Methodologies Materials Electro-thermo-Mechanical Testing (ETMT) Scanning Electron Microscope Results and Discussions Case One: Location-Dependent Property Extraction Subject to Complex Exposure Microstructure Elevated temperature deformation and rafting Room-temperature deformation and load drop phenomenon Case Two: Accidental Thermal Load Microstructure Shear banding in quasi-static tensile Low cycle fatigue Summary and Conclusion Open Access References