# Fileset

[Adv Eng Mater - 2024 - Tammpere - Profilometry‐Based Indentation Plastometry at High Temperature.pdf](https://mdr.nims.go.jp/filesets/6f589923-3640-4550-8e44-119a87158eff/download)

## Creator

Hannes Tammpere, [Phil McKeown](https://orcid.org/0009-0000-2289-1316), [James Miller](https://orcid.org/0000-0001-7063-3215), [Chizhou Fang](https://orcid.org/0000-0003-2601-2993), Emily Curtis, [Marcus Gaiser‐Porter](https://orcid.org/0009-0003-5809-3748), [Max Burley](https://orcid.org/0000-0002-8102-4105), [James Campbell](https://orcid.org/0000-0001-9158-1418), [Maria Artiles](https://orcid.org/0000-0003-4681-5806), [Yuanbo Tang](https://orcid.org/0000-0002-9667-7846), [Satesh Utada](https://orcid.org/0000-0001-6783-9968), Roger Reed, [Trevor Clyne](https://orcid.org/0000-0003-2163-1840)

## Rights

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

## Other metadata

[Profilometry‐Based Indentation Plastometry at High Temperature](https://mdr.nims.go.jp/datasets/f67c41fe-e4a4-49fd-b88f-72e1ed18612a)

## Fulltext

Profilometry‐Based Indentation Plastometry at High TemperatureProfilometry-Based Indentation Plastometry at HighTemperatureHannes Tammpere, Phil McKeown, James Miller, Chizhou Fang, Emily Curtis,Marcus Gaiser-Porter, Max Burley, James Campbell, Maria Artiles, Yuanbo Tang,Satesh Utada, Roger Reed, and Trevor Clyne*1. IntroductionThe PIP (profilometry-based indentation plastometry) procedureis now a well-established mechanical testing technique. It hasevolved from more than two decades of research aimed at pro-cedures for inferring stress–strain relationships from outcomesof instrumented indentation tests. Much early work involved useof nanoindenters, with the instrumentation often relatingprimarily to the acquisition of load–displacement data and aprogressive focus on reducing the scaleof the indentation. The most popular meth-odology was based initially[1–7] on directconversion (using empirical relationships)of the load to an equivalent stress and thedisplacement to an equivalent strain—thisis sometimes referred to as the instru-mented indentation technique. Thisapproach is attractively simple, but it even-tually became clear[8] that such conversionscannot capture the complexity of the evolv-ing stress and strain fields during indenta-tion and do not lead to reliable stress–strainrelationships being obtainable in a univer-sal way.The procedure that has evolved into acommercially viable methodology is thatof iterative finite element method (FEM)simulation of the indentation,[9–15] converg-ing on the stress–strain relationship(captured in a constitutive law) that givesoptimal agreement between measured and modeled outcomes.It has also been shown[16–19] that there are important advantagesto using the indent profile, rather than the load–displacementcurve, as the target outcome. This methodology is conceptuallysimple and rigorous, but the development of commercially viablestand-alone products has required optimization of control andsimulation software, linked to equipment for automated creationand characterization of suitable indents. A further importantadvance has been the recognition that if scale-independentH. Tammpere, P. McKeown, J. Miller, C. Fang, M. Gaiser-Porter, M. Burley,J. Campbell, M. Artiles, T. ClynePlastometrex Ltd204 Science Park, Milton Road, Cambridge CB4 0GZ, UKE-mail: twc10@cam.ac.ukThe ORCID identification number(s) for the author(s) of this articlecan be found under https://doi.org/10.1002/adem.202301073.© 2024 The Author(s). Advanced Engineering Materials published byWiley-VCH GmbH. This is an open access article under the terms ofthe Creative Commons Attribution License, which permits use,distribution and reproduction in any medium, provided the originalwork is properly cited.DOI: 10.1002/adem.202301073E. Curtis, T. ClyneDepartment of Materials ScienceUniversity of Cambridge27 Charles Babbage Road, Cambridge CB3 0FS, UKY. Tang, R. ReedSchool of Metallurgy and MaterialsUniversity of BirminghamElms Road, Birmingham B15 2SE, UKS. UtadaResearch Centre for Structural MaterialsNIMS305-0047 1-2-1 Sengen, Tsukuba Ibaraki, JapanR. ReedDepartment of MaterialsUniversity of OxfordParks Road, Oxford OX1 3PH, UKThis is a first report on profilometry-based indentation plastometry (PIP) at hightemperature (HT), covering both thermal characterization and issues forobtaining stress–strain curves. The heating system has a relatively low thermalinertia, reaching 800 °C within about 10 min, while both indentation (≈20 s) andcooling (≈20 min) are also quick. This capability is useful in terms of limitingexposure of the sample to prolonged periods at HT, and hence avoiding theformation of thick oxide layers (which can affect indent profiles and henceinferred stress–strain curves). There is good general consistency between stress–strain curves obtained via HT-PIP and those from tensile testing. However, thepossibility of creep (time-dependent deformation) affecting the outcomes (of bothtypes of test), particularly at higher temperatures, should be borne in mind. Creephas a characteristic effect on tensile curves, which can often be confirmed andinvestigated by changing the imposed strain rate. It can also be revealed bycarrying out the HT-PIP testing with different penetration velocities or bymonitoring the shape of the load–displacement plot.RESEARCH ARTICLEwww.aem-journal.comAdv. Eng. Mater. 2024, 2301073 2301073 (1 of 15) © 2024 The Author(s). Advanced Engineering Materials published by Wiley-VCH GmbHmailto:twc10@cam.ac.ukhttps://doi.org/10.1002/adem.202301073http://creativecommons.org/licenses/by/4.0/http://www.aem-journal.comhttp://crossmark.crossref.org/dialog/?doi=10.1002%2Fadem.202301073&domain=pdf&date_stamp=2024-08-28stress–strain relationships are being sought, then the volumebeing deformed must be large enough to be representative ofthe bulk, which usually translates into it containing a relativelylarge number of grains. In most cases, nanoindenters are notsuitable for this. There are also certain other requirements, suchas a need[13,20] to create plastic strains in the range of up to at leasta few tens of %.Recent work has covered application of the PIP technique to awide range of materials and effects. These include study ofwelds,[21] pipelines,[22] additively manufactured components,[23,24]metal matrix composites,[25] effects of residual stress,[26] case-hardened layers,[27] very hard metals,[28] and porous metals.[29]There is also a review article[30] that covers various aspects ofthe methodology. Consultation of this article is recommendedbecause many details of the way that the modeling is carriedout are covered there. These include use of the Voce equation[31]as the constitutive law of choice. The advantages of PIP (particu-larly in comparison with tensile testing), which partly concernease of sample preparation and speed of testing, are arguably evenmore significant for characterization of plasticity at high temper-ature (HT). Theremay be benefits inminimizing the period spentat HT, to reduce microstructural changes, and this is likely to beeasier with the relatively small, flat samples used in PIP testing.While the relatively small size of PIP samples allows use of ahot stage with low thermal inertia, the fact that the tested regionis not very small—it is typically about 1mm in lateral extent and200 μm deep—ensures that the problems of thermal drift andinstability that are commonly encountered when using nano-indenters at HT (or even at room temperature[32]) can be avoided.As for room temperature operation, the “intermediate” scale ofPIP has the potential to allow speed and simplicity of testing incombination with a rigor comparable to that of tensile testing.Nevertheless, elevated temperature certainly introduces newtechnical challenges for PIP testing. This article describes thedevelopment of a commercial HT PIP (HT-PIP) setup andpresents data related to its operation.2. Materials and Test Procedures2.1. MaterialsThe testing program involved a total of six materials (see Table 1),which gives the compositions and suppliers. The DS Cu was onlyused in thermal trials (representative of metals with high thermalconductivity), but the others were all subjected to mechanicaltesting (using both tensile and HT-PIP tests). They exhibit a widerange of thermomechanical characteristics.2.2. HT Tensile TestingMost of the tensile testing was carried out in Cambridge using anInstron 3369 loading frame, with a Severn Thermal SolutionsFurnace. The tensile samples were cylindrical, with threadedends. The reduced section length was 29mm, with a diameterof 5mm, and the clip gauge length was 25mm. The ceramicknife-edges of the clip gauge come through a small slot in thefurnace wall, while three thermocouples (to be touching the sam-ple within the gauge length) are introduced from the other side ofthe furnace. Samples were heated at 10 Kmin�1 for the SS310and S690QL samples and 20 Kmin�1 for the C-300. In both cases,there was a 20min soak prior to the start of the test. The standarddisplacement rate during loading was 10mmmin�1, correspond-ing to a strain rate of about 7� 10�3 s�1.For testing of the Waspaloy and Al-2618, undertaken atOxford, an Instron 8862 loading frame was used, with the sametype of furnace. These samples were also cylindrical, with a gaugediameter of 3.5mm and a gauge length of 6 mm. Straining wasmonitored via an Imetrum video extensometry system, withviewing through a quartz window in the furnace wall. Heatingwas at 25 Kmin�1 with a 10min soak prior to the start of thetest. The displacement rate during loading was 3.6 mmmin�1,corresponding to a strain rate of about 10�2 s�1. The standardstrain rates used in the two laboratories were therefore verysimilar.2.3. HT-PIP Testing2.3.1. Hot StageThe hot stage, for integration into a standard benchtop plastom-eter, is depicted schematically in Figure 1. The glass–ceramicdisks have a very low thermal conductivity (K≈ 1W m�1 K�1).The air (K≈ 0.025W m�1 K�1) gap also presents a large thermalbarrier, with the air jet ensuring that there is no recirculatoryconvective heat flow between the load cell platen and the locationplate. The silicon nitride (ball) is not a good thermal conductorTable 1. Codes, nominal compositions, and suppliers of the metals used. SS310 is a (high-Cr) austenitic stainless steel, S690QL is a “high-speed steel”(HSS), Al-2618 is an age-hardening Al alloy, Waspaloy is a Ni-based superalloy, C-300 is a maraging steel, and DS Cu is dispersion strengthened copper(containing alumina particles).Code Composition [wt%] SupplierC Si Mn Co Mo Cr Ni S P Cu Fe Mg Al TiSS310 0.25 1.5 2.0 – – 25 20 0.03 0.04 0.5 bal. – – – NeonickelS690QL 0.20 0.8 1.7 – 0.7 1.5 2.0 0.001 0.012 0.5 bal. – – – Brown-MacFarlaneAl-2618 – 0.2 0.02 – – – 1.1 – – 2.4 1.1 1.5 bal. 0.05 Dynamic MetalsWaspaloy 0.04 0.03 0.03 12.3 3.8 19.3 bal. 0.001 0.003 0.02 1.5 – 1.3 3.0 NeonickelC-300 0.01 0.04 0.02 9.3 4.8 0.13 18.2 0.001 0.005 0.07 bal. – 0.1 0.7 Dynamic MetalsDS Cu – – – – – – – – – bal. – – 7.0 Al2O3 – Gadi Co. Inc.www.advancedsciencenews.com www.aem-journal.comAdv. Eng. Mater. 2024, 2301073 2301073 (2 of 15) © 2024 The Author(s). Advanced Engineering Materials published by Wiley-VCH GmbH 15272648, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adem.202301073 by National Institute For, Wiley Online Library on [23/09/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.aem-journal.comand the sectional area for heat flow through it from the sample issmall. The thermal mass of the indenter housing is large com-pared with the net upward heat flux. The hot zone is thus quitewell confined.The stage has four main components—the heating assemblyitself (which locates onto the platen), a plate that bolts onto thebaseplate (with cooling fans, linear variable displacement trans-ducer, and connector block attached), the indenter housingassembly (which is bolted to the crosshead), and a box containingthe electronic control system. Assembly and disassembly arevery quick and easy—it can be carried out in about 10min.The plastometer is simply plugged into a standard single phase(220/240 V) mains socket. The power requirement is for atheoretical peak of around 2 kW, although the measured peakis about 1.5 kW (during the initial transient when the cartridgeheaters are cold and have a low electrical resistance). Steady-statepower consumption at the highest operating temperatures isabout 0.5 kW.2.3.2. Thermal Fields and Control of Test TemperatureAn important issue with relatively compact systems is that ofcontrolling the thermal field so that the temperature of thesample—or in this case of the part of the sample being tested—can be preselected. For routine testing, it is not always possible tolocate a thermocouple within (a drilled hole in) a sample and thereare also problems in ensuring that a thermocouple is in good ther-mal contact with the free surface of the sample. The approachadopted here is to use information about the temperature withinthe heating block (close to the bottom of the sample) and to inferthe sample test temperature from this. A target temperature canthus be set for the governing thermocouple (closest to the top ofthe heating block), based on the required test temperature andknowledge of the axial thermal profile in and around the sample.This calculation involves both the sample thickness and its ther-mal conductivity, which are input by the user.Once a steady state has been reached (and neglecting lateralheat flow—the heater block is well insulated, although there maybe some losses from the sample), the thermal profile across thesample will be as shown schematically in Figure 2. A heat trans-fer coefficient (hi) characterizes the thermal conductancebetween heating block and sample. This is expected to have arelatively high value (low thermal resistance) because both sur-faces are flat and a (small) load is applied to the indenter duringheating. However, experimental measurements were needed toestablish whether it could be taken to be infinite (no thermalresistance, with ΔTi= 0).Thermal conductivities are required for both block and sam-ple. Fortunately, thermal conductivities of metals are fairly insen-sitive to microstructure and in general can be specified in asimilar way to that in which stiffness is handled during PIPtesting—i.e., just identifying the base metal or alloy type shouldbe sufficient to allow a suitable value to be ascribed—see below.Figure 1. The HT-PIP stage, located within a plastometer: a) a schematic sectional depiction and b) a 3D visualization.Figure 2. Schematic depiction of the axial thermal profile across the sam-ple and the top part of the heating block, once a steady state has beenestablished. (Thermocouples were located within samples only fortrials—this is not done during routine HT-PIP testing).www.advancedsciencenews.com www.aem-journal.comAdv. Eng. Mater. 2024, 2301073 2301073 (3 of 15) © 2024 The Author(s). Advanced Engineering Materials published by Wiley-VCH GmbH 15272648, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adem.202301073 by National Institute For, Wiley Online Library on [23/09/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.aem-journal.comMoreover, while the presence of impurities and relatively low sol-ute levels can influence (reduce) the conductivity of highly con-ducting metals such as copper, silver, and aluminum, the exactvalue is of little significance if it is high—the thermal gradients inthe sample will in these cases be low.Nevertheless, it is important to ascribe reasonably accuratevalues, particularly for (common) low conductivity metals suchas Fe- and Ni-based alloys. Furthermore, these often exhibit asignificant dependence on temperature. A representative set ofvalues (with temperature dependence) is tabulated in Table 2,with an indication of their source. These data are plotted againsttemperature in Figure 3. Two broad groups can be identified.Values for the main transition metals used for engineering pur-poses (Fe, Ni, and Ti) fall in the range of 10–50Wm�1 K�1 (oftenwith some dependence on temperature), whereas alloys based onother common metals (Al, Cu, Ag, etc.), most of which have thehighest symmetry (fcc) crystal structure, tend to exhibit highervalues (≈100–400W m�1 K�1), with the temperature having rel-atively little effect. (Ranges are therefore shown in Table 2, butthey are plotted as single values in Figure 3: the important pointhere is simply that as these materials are highly conductive, theexact value is of little significance in the present context.) Interms of practical methodology, expressions in the table canbe implemented via a pull-down menu, with the user just choos-ing the broad alloy type (as is currently done during implemen-tation of PIP for stiffness values).2.3.3. Indentation ProcedureOnce the sample is on the heating block, a test temperature isselected. The low thermal inertia of the system ensures that heat-ing (and cooling) rates are quite rapid, with a typical time to reachthe maximum test temperature being of the order of 10min—some details are given below. The indentation itself is normallycarried out with a constant penetration velocity of 10 μm s�1, so ittypically takes about 20 s (although a few tests were carried outwith slower and faster rates of penetration, to check for any timedependence—i.e., creep effects). A small load (≈35 N) is appliedto the system throughout the heating period. This is designed tobring the sample and indenter into close thermal contact, ensuringthat the temperature in this vicinity is relatively uniform (at the tar-get level). After completion of the indentation, the furnace isswitched off (but the fans remain on). The time needed for the sam-ple to cool sufficiently for the profilometry to be carried out—i.e., toreach about 50 °C—depends on the test temperature, but is lessthan 30min. Typically, each test was repeated 2 or 3 times. In gen-eral, the reproducibility of measured indent profiles was high.2.3.4. Profilometry and Convergence on Optimized Stress–StrainRelationshipAs noted in the Introduction, the FEM model used to simulatethe indentation procedure and its iterative use to convergeon the optimal solution are fully described in previouspublications.[12,13,19,20] This solution takes the form of a best-fitset of Voce parameter values, so that an optimized true stress–trueplastic strain relationship is obtained. This can be used for variouspurposes, including the simple one of converting it (via analyticalequations) to the nominal stress–nominal strain relationship thatis the immediate outcome of a tensile test (up to the peak, whichusually corresponds to the onset of necking). It can also be used inother FEMmodels, for example, ones that describe the behavior ofcomponents under service conditions.The standard version of the FEM model itself is 2D (radiallysymmetric) and is implemented via commercial code supplied byCOMSOL. The mesh contains about 5000 volume elements, allsecond-order quadrilateral and/or triangular. Convergence isachieved via a fitting parameter (Sred) that characterizes the misfitbetween measured and modeled residual indent profiles. Thisparameter, which is a dimensionless, positive number, has avalue of zero for perfect fit, but for optimized agreement it mighttypically have a value of about 10�4. Details of the convergenceprocedure followed in parameter space are supplied in previousTable 2. Thermal conductivity data from the literature, including anysignificant dependence on temperature (T in °C).Metal Conductivity, K [W m�1 K�1] SourceAustenitic stainless steel (310) 12þ 0.015 T [51]Ferritic stainless steel (HT9) 25 [52]Medium C steel (0.5%) 55 – 0.031 T [53]High C steel (1%) 40 – 0.012 T [53]Ni superalloy (Inconel 625) 11þ 0.016 T [54]Cast Iron 40 [55]Titanium (alloy) 21 [56]Aluminum (alloy) ≈100–200 [57]Copper (alloy) ≈200–400 [58]Figure 3. Plot of the thermal conductivity data presented in Table 2.The lines for the Al and Cu alloys are simply for guidance—in the highconductivity regime, the exact value is of little importance for currentpurposes.www.advancedsciencenews.com www.aem-journal.comAdv. Eng. Mater. 2024, 2301073 2301073 (4 of 15) © 2024 The Author(s). Advanced Engineering Materials published by Wiley-VCH GmbH 15272648, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adem.202301073 by National Institute For, Wiley Online Library on [23/09/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.aem-journal.compublications. This operation is no different for HT-PIP than forstandard room temperature operation.A point to note, however, concerning HT-PIP operation is thatthe stiffness (Young’s modulus), which is an input parameter forthe modeling, may differ from the corresponding value at ambi-ent temperature. Some published measurements may be open toquestion, partly because of the possibility of creep deformation,but there is work[33] based on theoretical (interatomic bondstrength) calculations that is expected to be reliable. In general,the dependence on temperature is relatively weak, but significantchanges might in some cases be expected over the HT-PIP test-ing range. For example,[33] the stiffness of Ni falls from around200 GPa at room temperature to 150 GPa at 800 °C (T/Tf≈ 0.62).The outcome of PIP testing does not have a high sensitivity to thestiffness, but a change such as this may need to be taken intoaccount.Overall, there are naturally certain ways in which the HT-PIPtest procedure differs from that at room temperature, althoughnone of them are very fundamental. The logic path for HT-PIPtesting is depicted in Figure 4. The initial part involves specifyingcertain properties of the sample (at the test temperature). A smallamount of lubricating grease is put on the sample before heatingis started. This ensures that the friction coefficient is in a suitablerange (relatively small) and also helps in reducing oxide forma-tion. It is removed after cooling to room temperature (before theprofilometry is carried out). As described in previous publica-tions, a stylus profilometer is used, with a resolution of around�1 μm. Scans are obtained in several different directions (allpassing through the central axis of indentation), partly to checkfor the possible presence of anisotropy in the sample. Heating tothe test temperature is under software control, as is the indenta-tion, cooling, and profilometry. Inferring of stress–strain curvesis thus essentially no different from the procedure at room tem-perature, although potential effects of creep should be noted–seeSection 4.2.3. HT-PIP Testing Characteristics3.1. Control over the Thermal FieldAn example set of thermal histories is shown in Figure 5. Forthese experiments, in addition to the thermocouples in the heat-ing block, one was located in a hole drilled into the sample fromthe side (as shown in Figure 2), with its end located on the axis ofthe system, close to the free surface. The test temperature is typ-ically reached in nomore than about 10min. There is good agree-ment between measured and modeled histories for the samplesurface temperature—details of the model are given below.Several thermal trials of this type were carried out, with dif-ferent sample materials, thicknesses, and target temperatures,and steady-state readings of the three thermocouples wererecorded. Some results are shown in Table 3, with the distancesrelating to the thermocouple locations indicated in Figure 2. Theheat flux values shown were obtained from the two measuredtemperatures in the heating block, the distance between themand the thermal conductivity of the block material for the tem-perature range concerned (Table 2), using:Q ¼ ðTB2 � TB1ÞL21KB (1)The interfacial heat transfer coefficient, hi, can be estimatedfrom the heat flux (provided the temperature drop across theinterface ΔTi is known), using the simple relationship:Figure 4. Logic path for HT-PIP testing.www.advancedsciencenews.com www.aem-journal.comAdv. Eng. Mater. 2024, 2301073 2301073 (5 of 15) © 2024 The Author(s). Advanced Engineering Materials published by Wiley-VCH GmbH 15272648, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adem.202301073 by National Institute For, Wiley Online Library on [23/09/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.aem-journal.comhi ¼QΔTi(2)This interfacial temperature drop, which is equal to (TBi – TSi),can be estimated via extrapolation to the interface of thethermal gradients in block and sample (recognizing that theheat flux is the same in both, provided that a steady state hasbeen established and that heat flow is predominantly axial).This leads toTBi ¼ TB1 � ðTB2 � TB1ÞL1iL21� �(3)TSi ¼ TS þ ðTB2 � TB1ÞLSL21� �KBKS� �(4)The heat fluxes in Table 3 are in the range 5–200 kW m�2,rising as the temperature is increased and being higher forthicker samples. The inferred values of ΔTi are all low, althoughthey rise with increasing temperature (and heat flux).Figure 5. A set of thermal histories, for a 3mm thick SS310 sample being heated to 800 °C.Table 3. Steady-state temperatures from a series of heating trials, for the HSS (S690QL) and the Al alloy (2618), all with L21= L1i= 2 mm.Metal LS [mm] Temperatures [°C] Q [kW m�2] ΔTi [°C] LS [mm] Temperatures [°C] Q [kW m�2] ΔTi [°C]TB2 TB1 TS TB2 TB1 TSHSS 4.25 101 100 96 7 2.2 8.3 106 104 97 14 2.1HSS 4.25 208 204 195 30 1.6 8.3 217 212 196 38 2.7HSS 4.25 423 415 391 72 7 8.3 329 322 296 58 5.8HSS 4.25 533 525 488 78 19 8.3 437 426 389 99 2.6HSS 4.25 646 634 596 126 10 8.3 561 547 492 136 7.6HSS 4.25 753 741 696 135 15 8.3 668 651 581 178 7.8HSS 8.3 803 785 699 202 15Al alloy 3.4 103 102 98 7 2.7 – – – – – –Al alloy 3.4 212 209 201 22 4 – – – – – –Al alloy 3.4 324 318 305 49 5 – – – – – –Al alloy 3.4 428 420 402 72 7 – – – – – –www.advancedsciencenews.com www.aem-journal.comAdv. Eng. Mater. 2024, 2301073 2301073 (6 of 15) © 2024 The Author(s). Advanced Engineering Materials published by Wiley-VCH GmbH 15272648, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adem.202301073 by National Institute For, Wiley Online Library on [23/09/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.aem-journal.comThese values can be used to estimate hi, which is expected to beapproximately constant (because it just depends on the physicalcontact conditions at the interface, which are not expected to varymuch). Figure 6 is a plot of Q against ΔTi, using the values inTable 3. While there is inevitably some scatter, there is at leastan approximately linear trend, with the gradient (hi) having a(high) value of about 12 kW m�2 K�1. It is unsurprising thatthe interfacial thermal contact is good, leading to relatively smalltemperature drops, but it is worthwhile to take them intoaccount.The required temperature “correction increment” can be writ-ten as the sum of the three contributions:TB1 � TS ¼ ðTB1 � TB2ÞL1iL21þ Qhiþ ðTB1 � TB2ÞLsL21KBKs� �∴TB1 � TS ¼ QL1iKBþ 1hiþ LSKS� � (5)The units here are kWm�2 forQ, mm for L, Wm�1 K�1 for K,and kW m�2 K�1 for hi. (Use of these units for L, Q, and hi elim-inates the need for conversion factors of 103, etc.)The value of Q to use in this equation could be obtained fromEquation (1), but that would require real-time monitoring ofTB2—complicating the algorithm considerably. Recognizing thatthe approximate relation between the target temperature and theheat flux is universal (for this particular setup), it can be esti-mated from data such as those in Table 3. It can be expressed:Q ≈ 30TS250� �0.9 LS3� �0.9kWm�2 (6)where TS is in °C and LS is in mm. This is evidently an empirical(approximate) equation, which is specific to the thermal charac-teristics of this setup, but this is all that can be done withoutdetailed (FEM) modeling, which is impractical in the presentcase. Figure 7 shows plots of the relationship represented byEquation (5) and (6), for selected cases. Agreement between thesemodeled curves and the experimental data is good. There is alsogood agreement between corresponding measured and modeledheat fluxes. It is therefore expected that this procedure can beFigure 6. Plot of the data for Q and ΔTi in Table 3, showing how an esti-mate is obtained for the interfacial heat transfer coefficient, hi.Figure 7. Comparison between modeled (Equation (5) and (6)) and measured values of the incremental temperature correction, as a function of sampletemperature.www.advancedsciencenews.com www.aem-journal.comAdv. Eng. Mater. 2024, 2301073 2301073 (7 of 15) © 2024 The Author(s). Advanced Engineering Materials published by Wiley-VCH GmbH 15272648, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adem.202301073 by National Institute For, Wiley Online Library on [23/09/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.aem-journal.comused to control sample temperatures with an acceptable degree ofaccuracy.The picture that emerges from these thermal measurementsis that this difference in temperature (TB1 – TS) may be relativelysmall, but is likely to be of the order of several tens of °C andcould be as high as 100 °C—becoming larger for higher test tem-peratures, thicker samples, and materials with lower thermalconductivities. Such differences clearly cannot be neglected, sothis “correction” is essential (although it is probably unrealisticto aim for an accuracy of better than about �10 °C). The proce-dure is thus that once the target temperature has been selected,the value of Q is obtained from Equation (6) and the necessarygoverning thermocouple temperature then found fromEquation (5), using values of LS and KS input by the user—the latter currently being available as an option via a menu ofmaterials. (The values of L1i and KB, and also hi, are knownand fixed.)This temperature correction will be of greater significance formetals in the “lower conductivity” category. Not only is the incre-ment raised by the lower conductivity (and hence the higher ther-mal gradient), but metals in this category are in general morelikely to be tested at high temperatures—they tend to have bothhigher melting temperatures and better microstructural thermalstability. It is also worth noting that the thermal correction, andthe error likely to be associated with it, will be lower with thinnersamples, particularly for low conductivity metals. The minimumthickness required to ensure that the sample is effectively“semi-infinite” during indentation is around 3mm (with the1mm radius indenter), so a thickness of, say, 3–5mm may bea good choice in many cases. It may also be helpful toensure that both the bottom of the sample and the top of theheater block are flat and clean, so that the value of hi is in theexpected range.3.2. Control of Oxidation and Other Chemical ReactionsA potential concern during HT-PIP testing relates to chemicalreactions being stimulated, before and/or during the test. Twotypes of reaction are possible—sample/indenter and sample/atmosphere. A common sample/indenter reaction involves for-mation of metal carbides, particularly with diamond indenters.Not all metals are strong carbide-formers, but many of themare. In fact, diamond indenters cannot be used above about300–400 °C because of erosive oxidation and/or reaction withthe sample.[34,35] However, Si3N4 indenters are much less likelyto react with metallic samples,[36] even at very HT, so this is not aconcern for HT-PIP. There are some metals that can formnitrides more stable than Si3N4, but even those tend to react onlyquite slowly, unless the temperature is very high. Silicon nitridecan undergo oxidation, but this only becomes significant at tem-peratures well above the range of interest here.[37,38]However, sample/atmosphere reaction is always a possibility.Virtually all metals (except gold) have a surface oxide film.Oxidation is always thermodynamically favored (in air) and thick-ening of an oxide layer can only be avoided by being kineticallyvery slow. The growth rate of an oxide film normally rises as thetemperature is increased (despite the associated decrease in thethermodynamic driving force). The apparently obvious way toinhibit growth is to bring the oxygen partial pressure down toa low level. However, it may need to be very low (ppm range)to reduce the rate of oxidation substantially. This is basicallybecause the rate-determining step is usually the diffusion of ionsthrough the oxide, rather than the rate of adsorption of oxygen atthe free surface. There is certainly the possibility of attempting toshroud the sample with an atmosphere having a very low oxygenpartial pressure, or possibly with one that would create a reduc-ing environment—for example, one containing hydrogen.In practice, however, there are difficulties, particularly in anenclosure on the HT-PIP heating stage, in which high thermalgradients, and hence strong convection currents, are created.Details are provided in a separate article,[39] which is focusedon the issue of oxidation during HT-PIP testing and the potentialfor employing counter-measures.While such oxide films (perhaps having a thickness of theorder of several tens of microns) are of little significance duringtensile testing (of a sample with a diameter of several mm), theycould affect HT-PIP testing, in which the response of the near-surface region is of considerable importance. Typical pile-upheights are of the order of a few tens of microns, or sometimeseven less, so having a surface layer (with completely differentmechanical properties) with a thickness of more than a fewmicrons is a potential concern. However, the low thermal inertiaof the HT-PIP setup is advantageous because it allows the time atHT to be limited to just a few minutes. Furthermore, the lubri-cant that is placed on the sample before heating (and cleaned offbefore the profilometry) helps to reduce the rate of oxidation.At least in most cases, these measures allow the thickness ofany oxide layers to be kept relatively low. (If the effect of a heattreatment is being investigated, then it is recommended that thisshould be carried out separately and the surface then regroundbefore being tested using HT-PIP.)4. HT Mechanical Properties4.1. Indent ProfilesFive materials were tested, both in tension and via HT-PIP, over arange of temperature. All of the materials tested were confirmed(via PIP testing) to be isotropic and homogeneous—the indentswere radially symmetric and the same outcomes were obtainedfrom different locations in the as-received materials. In mostcases, Voce sets were found that gave excellent agreementbetween measured and modeled indent profiles, although therewere some limitations to this at high temperatures—see below.Examples are shown in Figure 8, which are for the SS310 alloy.The procedures for obtaining the model results are fullydescribed in previous publications.[12,13,19,20]4.2. Stress–Strain Curves4.2.1. Stainless SteelNominal stress–strain curves for SS310 are shown in Figure 9.There is good agreement between tensile and PIP, with all casesup to 700 °C being within �10%. This would be acceptable atroom temperature, and there are clearly more potential sourceswww.advancedsciencenews.com www.aem-journal.comAdv. Eng. Mater. 2024, 2301073 2301073 (8 of 15) © 2024 The Author(s). Advanced Engineering Materials published by Wiley-VCH GmbH 15272648, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adem.202301073 by National Institute For, Wiley Online Library on [23/09/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.aem-journal.comof discrepancy with HT testing, so this level of agreement isencouraging. Following the standard PIP procedure, nominalstress–strain curves are shown only up to the onset of necking(expected to occur at the peak in the plot). These are derived fromthe true curve, using the standard analytical equations. The ten-sile plots, on the other hand, are shown up to the point ofrupture, so that the “ductility” (nominal strain at failure) canbe seen. In fact, this value is not of any intrinsic significance(because it depends on the dimensions of the test-piece), butthese portions of the tensile curves are shown for completeness.It is, however, notable that there is significant disagreementbetween PIP and tensile for the 800 °C case. This is thoughtto be associated with creep deformation starting to have a signif-icant effect. This is investigated in detail below.The serrations in the tensile curve for 600 °C were observed inrepeat runs and appear to be real. The most likely explanation forthem is that they arise from “Suzuki locking,” caused by segre-gation of substitutional solute—presumably Cr and/or Ni in thiscase—to stacking faults between dissociated pairs of partial dis-locations. This is a weaker effect than the more familiar pinningby interstitials (such as carbon), which occurs in a much lowertemperature range (because they have much higher mobility, aswell as binding more strongly to dislocations). Nevertheless,Suzuki serrations can be noticeable and similar types of curveto the one observed here, occurring in similar temperatureranges, have been reported a number of times for fccsystems,[40–44] together with evidence that the pinning doesindeed result from substitutional solute segregating to stackingfaults between dissociated partials. Such serrations cannot, how-ever, be reliably detected in load–displacement data acquired dur-ing indentation (or in residual indent profiles) because thedisplacement at any given load is affected by the local responseof regions that have reached a wide range of different strains.This “integration” removes the sensitivity to effects occurringthroughout the sample at a specific strain (as in a tensile test).A similar limitation applies to initial strain bursting (“Lüdersbands”) and transitional yielding.Figure 8. Comparisons between modeled and measured PIP indent pro-files, for the SS310, after testing at two different temperatures. (Theapplied loads were 2.53 kN at 200 °C and 1.518 kN at 600 °C.).Figure 9. Nominal stress–strain curves for SS310, from HotPIP and tensile testing.www.advancedsciencenews.com www.aem-journal.comAdv. Eng. Mater. 2024, 2301073 2301073 (9 of 15) © 2024 The Author(s). Advanced Engineering Materials published by Wiley-VCH GmbH 15272648, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adem.202301073 by National Institute For, Wiley Online Library on [23/09/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.aem-journal.com4.2.2. High-Speed SteelCorresponding results for S690QL are shown in Figure 10. Thereis some minor strain bursting at ambient, which disappears athigher temperature. Again, such effects cannot be capturedvia PIP testing, but its overall significance is minimal. For thismaterial, there is considerable softening between 400 and 600 °C.This change involves going well above the temperingtemperature and is associated with coarsening of fine carbidesand reconstruction of fine bainitic and twinning structures.These trends—i.e., being harder than the SS310 at room temper-ature, but ending up softer at around 600 °C—are thus consistentwith their microstructural characteristics.It is, however, clear from Figure 10 that, while the agreementis in general quite good up to 400 °C, substantial discrepanciesarise at higher temperatures, with the strain to the apparent onsetof necking (peak in the nominal plot) being much lower for thetensile test, followed by a progressive drop in the stress level. Thishas been investigated in detail for the highest temperature(700 °C). The tensile testing was carried out at three differentstrain rates (differing by a factor of about 30 between highestand lowest rates). It should be noted, however, that none of theserates (lying in the approximate range of 10�2 to 10�3 s�1) areoutside of what is conventionally regarded as the quasistaticregime (in which no dependence of plastic deformation charac-teristics on strain rate is normally expected). The sharp differen-ces between the three tensile curves (with the apparent yieldstress dropping by a factor of about 50%) clearly indicate thatcreep effects are highly significant (despite the duration of thetest covering only periods of up to about 5–10min). It shouldbe noted in this context that it is primary creep that is likelyto be dominant here and this is often much faster than secondary(steady state) creep, particularly at stress levels close to the yieldstress. Creep during a tensile test can certainly lead to a reductionin (nominal) stress with increasing strain, apparent well beforenecking has become established—this would often be referred toas stress relaxation (under strain control conditions). In fact, forthese tensile tests, the stress fell to low levels before final neckingand rupture occurred. It may be concluded that creep is having astrong effect on the tensile outcomes over the complete range ofstrain rate covered in these tests.The HT-PIP testing at 700 °C was also carried out at different“strain rates” and the PIP-derived stress–strain curves inFigure 10 are shown for two different penetration velocities—the standard one of 10 μm s�1 and a fast one of 50 μm s�1.The difference between the two confirms that time-dependenteffects can also arise during these tests. In fact, a slow testwas also carried out, at 1 μm s�1, but it was not possible to obtaina good fit to the resultant profile using a Voce plasticity curve.The indent profiles for the three cases are shown inFigure 11, where it can be seen that the slow profile was certainlybeing strongly affected by creep—note the much lower loadneeded to reach the target penetration range (less than half thatfor the fast case). It is unsurprising that the profile from what wasprobably creep-dominated indentation could not be well capturedusing a plasticity relationship. (It may also be noted that the aver-age strain rate during these HT-PIP tests, weighted by theamount of plastic strain being induced at different strain rates,Figure 10. Nominal stress–strain curves for S690QL, from HT-PIP and tensile testing. At 700 °C, tensile tests were carried out with “slow” (7� 10�4 s�1)and “fast” (2� 10�2 s�1) strain rates, in addition to the default rate (7� 10�3 s�1), while the standard PIP testing, with a penetration velocity of10 μm s�1, was supplemented by a fast one at 50 μm s�1.www.advancedsciencenews.com www.aem-journal.comAdv. Eng. Mater. 2024, 2301073 2301073 (10 of 15) © 2024 The Author(s). Advanced Engineering Materials published by Wiley-VCH GmbH 15272648, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adem.202301073 by National Institute For, Wiley Online Library on [23/09/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.aem-journal.comwas calculated via FEM simulation to be �2.5� 10�2, 5� 10�3,and 5� 10�4 s�1, respectively, for these fast, standard, and slowcases: these values cover a similar range to those in the tensiletests).As a further confirmation that the creep-affected tensilestress–strain curves are quite different from those representingpure plasticity, two interrupted tests were carried out, up to astrain of 10%. This was done, using the standard strain rateof 7� 10�3 s�1, at temperatures of 200 and 700 °C. It can be seenfrom Figure 10 that, in both cases, this strain is well past the peakin the plot. Conventional interpretation of such curves wouldtherefore envisage that a sample strained to this level should con-tain a well-defined neck, with virtually all of the recorded strain-ing in this regime being due to its development. Figure 12 is aphotograph of the two samples obtained in this way. While thesample tested at 200 °C does indeed contain a neck, which hasdeveloped to approximately the degree that might have beenexpected at this strain, there is no trace of a neck in the sampletested at 700 °C. It is clear that, for this case, the progressive post-peak fall in the nominal stress is due to a creep relaxation effect,rather than to the development of a neck.As indicated in the CORSICA logic path of Figure 4, a check isautomatically made on whether creep is likely to be having a sig-nificant effect on the outcome (of PIP testing, although, for casesin which the effect during PIP is noticeable, it is also likely thattensile stress–strain curves will be creep-affected). This check ismade by monitoring the shape of the load–displacement plot(during penetration at the standard rate of 10 μm s�1). Underpurely plastic conditions, this plot is expected to have a positivecurvature—so that, if it is fitted to a power law relationship, thevalue of the exponent is greater than unity. A contribution fromcreep, however, progressively introduces extra displacement (thatwill be greater for slower penetration rates). If this contribution issubstantial, then the curvature is likely to become negative, cor-responding to a fitted power law exponent having a value belowunity. A linear relationship, with an exponent value of unity(constant gradient), can be taken to be broadly indicative ofthe transition regime.A warning is issued if such a condition is being met (as indi-cated in the flow chart of Figure 4). Examples are shown inFigure 13 of load–displacement curves for cases (temperatures)in which the effects of creep are negligible or strong. (These twocases are the ones for which photos are shown in Figure 12).Figure 14 is a summary plot of n values obtained in this wayfor the Al-2618 alloy, over a range of temperature—details ofthe stress–strain curves obtained for this alloy are given inFigure 11. Measured indent profiles after HT-PIP testing of S690QL at700 °C, carried out with “slow,” “standard,” and “fast” indenter penetra-tion velocities. The loads needed to penetrate to the depths shown areindicated.Figure 12. Photograph of two samples, which were obtained by interrupt-ing tensile tests after a strain of 10%, with the tests being carried out attemperatures of 200 and 700 °C.Figure 13. Measured load–displacement plots obtained during indenterpenetration at the standard velocity of 10 μm s�1, for PIP testing ofS690QL samples at temperatures of 200 and 700 °C. Also shown arepower law curves, with best-fit exponent values.www.advancedsciencenews.com www.aem-journal.comAdv. Eng. Mater. 2024, 2301073 2301073 (11 of 15) © 2024 The Author(s). Advanced Engineering Materials published by Wiley-VCH GmbH 15272648, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adem.202301073 by National Institute For, Wiley Online Library on [23/09/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.aem-journal.comSection 4.2.4. There is clearly scope for using this value (acquiredin real time during standard indentation) to decide whether a“creep warning” message is needed.For standard PIP testing, the inferred stress–strain curves areconstrained to be Voce plots (converted to nominal form), sothere is no scope for replicating the stress relaxation effects cre-ated by creep in the tensile tests. It could be argued that the PIPoutcome for a given average strain rate could be regarded as aplasticity relationship that takes account of creep, in a similarway to a tensile curve being tagged with an associated strain rate.However, in practice this is similarly unsatisfactory, quite apartfrom the possibility that it may not be possible to obtain a Vocecurve giving good fit to the measured profile. Outcomes will, inpractice, depend on the sensitivity of the creep rate to the stress(commonly in the form of an exponent)—particularly for PIPtesting, in which there is a wide range of local stress levels.However, there may be potential for using PIP outcomesobtained with a range of penetration velocities to infer both plas-ticity and creep characteristics, using an extension of the inverseFEM approach applied to plasticity in the standard procedure.Something similar could also be done for tensile test outcomes.There will certainly be many cases for which creep effects aresmall or negligible, but these outcomes for S690QL at 700 °C(and even at 500 °C) illustrate clearly that there is potential forassociated changes to be dramatic, such that making compari-sons between tensile and PIP-derived curves must be approachedwith considerable care. Monitoring of the load–displacement plotduring PIP is a good way of checking on creep contributions,although repeating tensile tests with different strain rates mayalso be advisable. (This should be standard procedure for HT ten-sile testing, irrespective of whether PIP testing is being done.)Ideally, both the pure plasticity and the pure creep characteristicsof a metal (at a particular temperature) should be captured,although the nature of the equations used is different in thetwo cases (and primary creep will need to be included in theformulation). Once this information has been obtained, it shouldbe possible to model any given loading scenario (in which bothoccur simultaneously).4.2.3. Ni-Based SuperalloyCorresponding results for the Waspaloy are shown in Figure 15.This is a material that tends to retain its strength up to relativelyHT and the changes in the stress–strain curve over the rangestudied are not so great. In fact, there is very little changebetween 300 and 700 °C, with good agreement between tensileand HT-PIP outcomes. There is apparently no significant contri-bution from creep over this temperature range. This is broadlyconsistent with known information[45,46] about its creepcharacteristics.4.2.4. Age-Hardening Al AlloyThe Al-2618 alloy can be used to highlight some of the issuesrelated to microstructural stability when testing at HT. For thisalloy[47,48] the ageing temperature is about 190–200 °C. This ishigher than for most Al-based age-hardening alloys and indeedit is specifically designed to be suitable for HT applications.Nevertheless, mechanical testing at or above this temperaturerange is likely to involve thermal histories that will cause changesin the properties at the test temperature, via over-ageing prior tothe test. The tests themselves were completed in less than a min-ute, so it is unlikely that much over-ageing took place during thisperiod. However, the prior thermal histories are likely to have aneffect, so the same treatment (a heating rate of 20 K min�1, fol-lowed by a 20min soak at the test temperature) was used for bothtypes of test, to ensure that the over-ageing effect was the same ineach case. It should also be noted that, when materials becomesoft in this way, there is in general greater potential for creep toaffect tensile tests, as shown in Figure 10.Figure 14. Plot of measured values of the power law exponent against testtemperature, derived from load–displacement data obtained during PIPtesting of Al-2618, with a penetration velocity of 10 μm s�1.Figure 15. Nominal stress–strain curves for Waspaloy, from HT-PIP andtensile testing.www.advancedsciencenews.com www.aem-journal.comAdv. Eng. Mater. 2024, 2301073 2301073 (12 of 15) © 2024 The Author(s). Advanced Engineering Materials published by Wiley-VCH GmbH 15272648, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adem.202301073 by National Institute For, Wiley Online Library on [23/09/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.aem-journal.comFigure 16. Nominal stress–strain curves for Al-2618, fromHT-PIP and tensile testing. At 300 °C, tensile tests were carried out with fast (2� 10�2 s�1) andslow (7� 10�4 s�1) strain rates, in addition to the default rate (7� 10�3 s�1), while the PIP tests were carried out with fast (50 μm s�1), default(10 μms�1), and slow (1 μm s�1) penetration velocities.Figure 17. Nominal stress–strain curves for C-300, from HT-PIP and tensile testing.www.advancedsciencenews.com www.aem-journal.comAdv. Eng. Mater. 2024, 2301073 2301073 (13 of 15) © 2024 The Author(s). Advanced Engineering Materials published by Wiley-VCH GmbH 15272648, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adem.202301073 by National Institute For, Wiley Online Library on [23/09/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.aem-journal.comA comparison is shown in Figure 16 between tensile and PIP-derived curves for the Al alloy. The curves do indicate that, fortemperatures of 150 °C and above, there is little or no work hard-ening. (This is certainly expected in this temperature range forthis alloy.) The peaks are therefore expected to occur at lowstrains, even if there is no creep. The curves for 200 and300 °C do have peaks at low strains, but it is likely that these testswere affected by creep—particularly at 300 °C. The obvious wayto check on this is to carry out the tensile test with a range ofstrain rates—as has been done for the 300 °C case in Figure 16.A similar effect is observed to that in Figure 10 for 700 °C,although the differences are less dramatic—the apparent fallin “yield stress” between fastest and slowest strain rates beingabout 25%, compared with 50% for the S690QL alloy at700 °C. This again leaves the problem that there is in fact nowell-defined yield stress (and also no meaningful UTS)—or atleast there is no way of accessing these values via stress–straincurves that have been affected by creep in this way. It can also beseen in Figure 16 that different PIP curves were obtained withdifferent penetration velocities, but again the shapes of PIP andtensile curves are different (and neither represent the pure plas-tic characteristics). The PIP outcomes may still give some indi-cation of the tensile “plateau” levels—at least for cases in whichcreep is not completely dominating the outcome, but it is impor-tant to recognize that neither PIP nor tensile plots for this tem-perature are “genuine” stress–strain curves.4.2.5. Maraging SteelFinally, certain points can be noted for the C-300, which has anageing temperature[49] of around 400–500 °C. Its HT strength islargely due to the presence of fine Ni3Ti needle-like precipitatesthat are strongly resistant to coarsening in this temperaturerange.[50] Above 500 °C, the strength does drop sharply asover-ageing becomes rapid. The same prior thermal historieswere again used for the two types of test. The curves are shownin Figure 17. Up to 500 °C there is good agreement between PIPand tensile curves, with very little work hardening and a yieldstress that does not fall much over this temperature range.At 600 °C, however, the tensile curve is starting to become“creep-affected,” with the onset of necking no longer coincidingwith the peak in the curve. In this particular case, the yield stressobtained via PIP testing is close to the stress level at the peak inthe tensile curve, but the subsequent shape is not captured. Thedeviations will certainly be greater at higher temperatures.5. ConclusionsThe following conclusions may be noted: 1) The viability of car-rying out HT-PIP testing (to obtain stress–strain relationships) atelevated temperatures (up to 800 °C) has been demonstrated anda commercial setup is described. 2) The thermal characteristicsof the setup have been investigated. It has a relatively low thermalinertia, such that the heating period is typically no more thanabout 10min and the time spent close to the test temperatureis just a couple of minutes or so. 3) The temperature at thetop of the sample, where indentation is carried out, is relatedto that of the governing thermocouple, located within the heatingblock. Heat flow analysis has been used to establish this relation-ship, which depends on the thickness and conductivity of thesample, as well as the target temperature. This temperature dif-ference (between that of the governing thermocouple and that atthe top of the sample) is typically a few tens of °C, but it could beas high as 100 °C in some cases. 4) In general, good levels ofagreement are observed between stress–strain curves obtainedvia HT-PIP and from conventional tensile testing. However, par-ticularly at very high temperatures, the stress–strain curvesobtained via both tensile and HT-PIP testing may be significantlyaffected by creep (despite the test durations being short in bothcases). This can be investigated experimentally by carrying outthe tests with different strain rates (by changing the tensile dis-placement rate or the HT-PIP indenter penetration velocity).Significant differences in the outcomes for different strain ratesindicate that these curves are “creep-affected.” Furthermore,automated monitoring of the shape of the load–displacementplot can be used to detect whether this is the case. In such cases,use of the standard PIP procedure to capture the progressivecreep-driven stress relaxation that is likely to occur during a ten-sile test is not possible.AcknowledgementsFinancial support for T.W.C. is acknowledged from EPSRC, via (grant no.EP/I038691/1). Relevant support has also been received from theLeverhulme Trust, in the form of an International Network grant (IN-2016-004) and an Emeritus Fellowship (EM/2019-038/4). In addition,an ongoing Innovate UK grant (project no. 10006185) covers work in thisarea and JEC is in receipt of a Future Leaders grant from InnovateUK (MR/W01338X/1), which is focused on development of the PIPtechnique.Conflict of InterestThe authors declare no conflict of interest.Data Availability StatementThe data that support the findings of this study are available from thecorresponding author upon reasonable request.Keywordscreep, high temperatures, indentation plastometry, stress–strain curves,tensile testingReceived: July 14, 2023Revised: July 29, 2024Published online:[1] B. X. Xu, X. Chen, J. Mater. Res. 2010, 25, 2297.[2] A. H. Mahmoudi, S. H. Nourbakhsh, R. Amali, J. Test. Eval. 2012, 40,211.[3] R. O. Oviasuyi, R. J. Klassen, J. Nucl. Mater. 2013, 432, 28.[4] C. Yu, Y. H. Feng, R. Yang, G. J. Peng, Z. K. Lu, T. H. Zhang, J. Mater.Res. 2014, 29, 1095.[5] S. Pathak, S. R. Kalidindi, Mater. Sci. Eng. R 2015, 91, 1.www.advancedsciencenews.com www.aem-journal.comAdv. Eng. Mater. 2024, 2301073 2301073 (14 of 15) © 2024 The Author(s). Advanced Engineering Materials published by Wiley-VCH GmbH 15272648, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adem.202301073 by National Institute For, Wiley Online Library on [23/09/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.aem-journal.com[6] J. S. Weaver, S. R. Kalidindi, Mater. Des. 2016, 111, 463.[7] K. Jeong, H. Lee, O. M. Kwon, J. Jung, D. Kwon, H. N. Han, Mater.Des. 2020, 196, 109104.[8] J. E. Campbell, H. Zhang, M. Burley, M. Gee, A. T. Fry, J. Dean,T. W. Clyne, Adv. Eng. Mater. 2021, 23, 2001496.[9] C. Heinrich, A. M. Waas, A. S. Wineman, Int. J. Solids Struct. 2009,46, 364.[10] J. Dean, J. M. Wheeler, T. W. Clyne, Acta Mater. 2010, 58, 3613.[11] D. K. Patel, S. R. Kalidindi, Acta Mater. 2016, 112, 295.[12] J. Dean, T. W. Clyne, Mech. Mater. 2017, 105, 112.[13] J. E. Campbell, R. P. Thompson, J. Dean, T. W. Clyne, Mech. Mater.2018, 124, 118.[14] L. Meng, P. Breitkopf, B. Raghavan, G. Mauvoisin, O. Bartier,X. Hernot, Int. J. Mater. Form. 2019, 12, 587.[15] H. Xue, J. X. He, W. N. Jia, J. L. Zhang, S. Wang, S. Zhang, H. L. Yang,Z. Wang, Structures 2020, 28, 2752.[16] J. Lee, C. Lee, B. Kim, Mater. Des. 2009, 30, 3395.[17] W. Z. Yao, C. E. Krill, B. Albinski, H. C. Schneider, J. H. You, J. Mater.Sci. 2014, 49, 3705.[18] M. Z. Wang, J. J. Wu, Y. Hui, Z. K. Zhang, X. P. Zhan, R. C. Guo,Mater.Sci. Eng. A 2017, 679, 143.[19] J. E. Campbell, R. P. Thompson, J. Dean, T. W. Clyne, Acta Mater.2019, 168, 87.[20] T. W. Clyne, J. E. Campbell, Testing of the Plastic Deformation of Metals,Cambridge University Press, Cambridge, U.K. 2021.[21] W. Gu, J. E. Campbell, Y. T. Tang, H. Safaie, R. Johnston, Y. Gu,C. Pleydell-Pearce, M. Burley, J. Dean, T. W. Clyne, Adv. Eng.Mater. 2022, 24, 2101645.[22] M. Warwick, H. Vaka, C. Z. Fang, J. Campbell, J. Dean, T. W. Clyne,Steel Res. Int. 2023, 94, 2200920.[23] Y. T. Tang, J. E. Campbell, M. Burley, J. Dean, R. C. Reed, T. W. Clyne,Materialia 2021, 15, 101017.[24] T. J. F. Southern, J. E. Campbell, K. I. Kourousis, B. Mooney,Y. T. Tang, T. W. Clyne, Steel Res. Int. 2023, 94, 2200881.[25] R. Reiff-Musgrove, M. Gaiser-Porter, W. Gu, J. E. Campbell, P. Lewis,A. Frehn, A. D. Tarrant, Y. T. Tang, M. Burley, T. W. Clyne, Adv. Eng.Mater. 2023, 25, 2201479.[26] M. Burley, J. E. Campbell, R. Reiff-Musgrove, J. Dean, T. W. Clyne,Adv. Eng. Mater. 2021, 23, 2001478.[27] S. Ooi, R. Reiff-Musgrove, M. Gaiser-Porter, M. Steinbacher, I. Griffin,J. E. Campbell, M. Burley, C. M. Warwick, H. Vaka, T. W. Clyne,Adv. Eng. Mater. 2022, 25, 2201512.[28] J. E. Campbell, M. Gaiser-Porter, W. Gu, S. Ooi, M. Burley, J. Dean,T. W. Clyne, Adv. Eng. Mater. 2022, 24, 2101398.[29] R. Reiff-Musgrove, W. Gu, J. E. Campbell, J. Reidy, A. Bose,A. Chitrapur, Y. T. Tang, M. Burley, T. W. Clyne, Adv. Eng. Mater.2022, 24, 2200642.[30] T. W. Clyne, J. E. Campbell, M. Burley, J. Dean, Adv. Eng. Mater. 2021,23, 21004037.[31] E. Voce, J. Inst. Met. 1948, 74, 537.[32] S. Verma, P. Sarkar, P. Pant, J. Mater. Res. 2021, 36, 3436.[33] R. Rahemi, D. Y. Li, Scr. Mater. 2015, 99, 41.[34] J. M. Wheeler, D. E. J. Armstrong, W. Heinz, R. Schwaiger, Curr. Opin.Solid State Mater. Sci. 2015, 19, 354.[35] J. M. Wheeler, R. A. Oliver, T. W. Clyne, Diamond Related Mater. 2010,19, 1348.[36] E. H. Andrews, J. Mater. Sci. 1966, 1, 377.[37] M. Maeda, K. Nakamura, M. Yamada, J. Mater. Sci. 1990, 25, 3790.[38] T. Narushima, T. Goto, T. Hirai, Y. Iguchi,Mater. Trans. JIM 1997, 38,821.[39] P. J. McKeown, M. Gaiser-Porter, J. R. Miller, Y. T. Tang,J. E. Campbell, T. W. Clyne, Mater. Sci. Eng. A 2024, in preparation.[40] H. Saka, Philos. Mag. A 1983, 47, 131.[41] T. Kamino, Y. Ueki, H. Hamajima, K. Sasaki, K. Kuroda, H. Saka,Philos. Mag. Lett. 1992, 66, 27.[42] Y. Kaneko, K. Kaneko, A. Nohara, H. Saka, Philos. Mag. A 1995, 71, 399.[43] A. Chiba, M. S. Kim, Mater. Trans. 2001, 42, 2112.[44] Z. W. Jiang, L. L. Zhu, L. X. Yu, B. A. Sun, Y. Cao, Y. H. Zhao, Y. Zhang,Mat. Sci. Eng. A 2021, 820, 141575.[45] B. Wilshire, P. J. Scharning, Mat. Sci. Tech. 2009, 25, 242.[46] M. Whittaker, W. Harrison, C. Deen, C. Rae, S. Williams, Materials2017, 10, 61.[47] I. N. A. Oguocha, S. Yannacopoulos,Mater. Sci. Eng. A 1997, 231, 25.[48] J. H. Wang, D. Q. Yi, X. P. Su, F. C. Yin,Mater. Charact. 2008, 59, 965.[49] W. Sha, Scr. Mater. 2000, 42, 549.[50] F. Zhu, Y. F. Yin, R. G. Faulkner, Mater. Sci. Technol. 2011, 27, 395.[51] J. Blumm, A. Lindemann, B. Niedrig, R. Campbell, Int. J.Thermophysics 2007, 28, 674.[52] L. Leibowitz, R. A. Blomquist, Int. J. Thermophysics 1988, 9, 873.[53] M. J. Peet, H. S. Hasan, H. Bhadeshia, Int. J. Heat Mass Transf. 2011,54, 2602.[54] E. Kaschnitz, L. Kaschnitz, S. Heugenhauser, Int. J. Thermophysics2019, 40, 27.[55] D. Holmgren, Int. J. Cast Met. Res. 2005, 18, 331.[56] H. A. Abdel-Aal, M. Nouari, M. El Mansori, Tribol. Int. 2009, 42, 359.[57] P. Olafsson, R. Sandstrom, A. Karlsson, J. Mater. Sci. 1997, 32, 4383.[58] G. A. Dubskii, K. N. Vdovin, A. A. Nefed’ev, L. G. Egorova, RussianMetall. 2018, 2018, 252.www.advancedsciencenews.com www.aem-journal.comAdv. Eng. Mater. 2024, 2301073 2301073 (15 of 15) © 2024 The Author(s). Advanced Engineering Materials published by Wiley-VCH GmbH 15272648, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adem.202301073 by National Institute For, Wiley Online Library on [23/09/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.aem-journal.com Profilometry-Based Indentation Plastometry at High Temperature 1. Introduction 2. Materials and Test Procedures 2.1. Materials 2.2. HT Tensile Testing 2.3. HT-PIP Testing 2.3.1. Hot Stage 2.3.2. Thermal Fields and Control of Test Temperature 2.3.3. Indentation Procedure 2.3.4. Profilometry and Convergence on Optimized Stress-Strain Relationship 3. HT-PIP Testing Characteristics 3.1. Control over the Thermal Field 3.2. Control of Oxidation and Other Chemical Reactions 4. HT Mechanical Properties 4.1. Indent Profiles 4.2. Stress-Strain Curves 4.2.1. Stainless Steel 4.2.2. High-Speed Steel 4.2.3. Ni-Based Superalloy 4.2.4. Age-Hardening Al Alloy 4.2.5. Maraging Steel 5. Conclusions