• No results found

Change of Microhardness in Stoichiometric CuAu

N/A
N/A
Protected

Academic year: 2020

Share "Change of Microhardness in Stoichiometric CuAu"

Copied!
6
0
0

Loading.... (view fulltext now)

Full text

(1)Materials Transactions, Vol. 43, No. 3 (2002) pp. 560 to 565 c 2002 The Japan Institute of Metals. Change of Microhardness in Stoichiometric CuAu Markus Spanl1 , Wolfgang Püschl1 , Boris Sprušil2, ∗ , Jindřich Šachl2 , Vladimı́r Šı́ma2 and Wolfgang Pfeiler1 1 2. Institute of Material Physics, University of Vienna, Strudlhofgasse 4, A-1090 Wien, Austria Department of Metal Physics, Charles University, Ke Karlovu 5, CZ-121 16, Praha 2, Czech Republic. Microhardness of stoichiometric CuAu was measured during isochronal heating starting from the disordered and the ordered state as well as for isochronal cooling from the disordered state. It is shown that a continuous increase of long-range order connected with similar hardness behaviour was observed in all cases of thermal treatment and pre-treatment. In correspondence with earlier resistometric experiments during isochronal temperature variation no hysteresis between heating and cooling was observed. (Received October 9, 2001; Accepted January 28, 2002) Keywords: microhardness, copper-gold intermetallic compound, long-range order, isochronal heating, isochronal cooling, orderdisorder transformation. 1. Introduction There is a renewed interest in the order/disorder phenomena of stoichiometric CuAu, an alloy which shows several interesting features. Below about 410◦ C the originally disordered face-centred cubic alloy shows long-range order (LRO) in two variants: In a shallow region below this temperature (about 385–410◦ C) the orthorhombic long-period antiphase structure CuAuII is found, whereas the tetragonal L10 CuAuIphase is stable at lower temperatures.1) One of the reasons for the recent interest in the ordering processes in the CuAu alloy are new results of differential scanning calorimetry (DSC) and dynamic resistometric measurements, showing the so-called retro-effect.2, 3) For certain thermal pre-treatment of the sample the order-disorder transformation temperature TO/D during heating slightly decreases upon increasing the heating rate, which is in contrast to the expected effect of superheating. In a detailed investigation by means of DSC it turned out that a shift of TO/D to lower temperatures is observed during heating when starting after cooling at low rates.4) This fact is interpreted as arising from a sort of ‘stabilisation’ of the CuAuI phase against complete transition into CuAuII; this is presumed to be due to microstructural changes which help to relieve of accumulated internal stresses, namely twinning and loss of coherence of ordered domains. It is known that ordering into an L10 -structure is accompanied by shape and volume changes of the unit cell during the growth of ordered domains inducing extended lattice distortions.5–8) It was further reported that during annealing within the ordered temperature region a relaxation of lattice distortion occurs which is connected to twinning and dislocation generation. Since there is an ongoing interest in the origin of the retro-effect we look for a correlation between these structural changes and the evolution of LRO. Aside of TEM-observation also microhardness (MH) measurements may give information on this matter. Of course, MH is a complex measure of elasto-plastic properties and will ∗ Professor. Boris Sprušil passed away on August 5th, 2001.. be influenced by far more factors than e.g. a tensile test experiment.9–11) It is known that microhardness measurements produce a plastically deformed zone around the indentations which corresponds to about 15–20% deformation in a tension test, depending on load, loading time, critical resolved shear stress etc.12) It is therefore possible to connect a decrease in MH with an increased mobility and/or concentration of mobile dislocations and a reduction of the internal stress fields. This way the MH data can reflect the influence of LRO changes on the density and mobility of dislocations. Changes in MH may therefore show the temperature region where loss of coherence and twinning appear. There are only relatively old hardness results on CuAu in the literature and even these are not very detailed.11, 12) The aim of the present paper is to report on a recent detailed study of MH during isochronal heating and cooling treatment after different thermal pre-treatment, which has been planned in a form suitable for comparison with other measurements recently done on this alloy by the same research group.3, 4, 13–15) It is hoped to get further information from measurements of acoustic emission16) and surface microscopy.17) 2. Experimental Sample material (chemical analysis: 49.4 ± 0.5 at%Cu, 50.6±0.6 at%Au) was supplied by DEGUSSA via the Institut für Physikalische Chemie, University of Munich, Germany. The samples were rolled to a thickness of about 0.3 mm at room temperature with intermediate and final anneals at 600◦ C. Thermal treatments were carried out in a standard resistance furnace. To prevent surface oxidation and to improve the conditions of thermal equilibration a purified argon atmosphere was used during all anneals. For MH measurements a Zeiss Axioplan optical microscope was used together with a PAAR MHT-4 MH tester. MH values were obtained as an average of 20 widely spaced indentations made on the sample for each thermal annealing step using 42.5 g load with constant indentation rate (10 g/s) and an equal indent time (10 s). Since all MH-measurements were done at room tem-.

(2) Change of Microhardness in Stoichiometric CuAu. perature, the samples were first re-heated to the new annealing temperature by inserting them into the furnace at the readjusted temperature for each consecutive isochronal annealing step (20 min) and then quenched into water for measurement (quenching rate: about 1000 K/s). It has to be noted that this temperature program of quenched isochronal annealing constitutes the main difference with respect to in-situ measurements (e.g. of electrical resistivity or heat capacity). Table 1 gives the thermal treatments of different measuring runs. In order to demonstrate the complex character of the plastic deformation of CuAu during MH-measurement, Atomic Force Microscopy (AFM) observations using Rasterscope 4000 were performed. 3. Results 3.1 Isochronal heating; initial state: disordered, homogenised and quenched In Fig. 1 changes of MH are plotted as a function of annealing temperature as measured for three different samples. Aside of minor differences in the maximum height achieved at about 240◦ C a very good correspondence is observed. Five different ranges (indicated in Fig. 1 by vertical dash-dotted lines) may be distinguished: range I : steep hardness increase until about 160◦ C; range II : plateau or flat maximum in hardness between about 160 and 260◦ C; range III: decrease of hardness until a minimum is reached at 360◦ C;. Table 1 Thermal treatments of different measuring runs. Run. Thermal history. 1. Furnace cooled. 2. Furnace cooling+ isochronal annealing+ 20 min 600◦ C. 3. Initial quench from Tq /◦ C. Heating/ Symbol/Fig. /Cooling ▲/2. —. Heating. 600. Heating. 38 h at 600◦ C. 600. Heating. ▲/1, 7, ■/9. 4. 38 h at. 600◦ C. 600. Heating. +/1, 7. 5. Cooled with 1 K/min. —. Heating. ■/2. 6. 600◦ C. —. Cooling. /2, 8, 9. 38 h at. +1 h at. 450◦ C. ■/1, 7. Fig. 1 Temperature dependence of microhardness of three CuAu samples (see Table 1) isochronally heated from the quenched disordered state. The full line serves as a guide for the eye.. 561. range IV: re-increase of hardness to a maximum at 400◦ C; range V : final decrease of hardness. The change of MH due to a maximum increase of LRO amounts to about 1500 MPa in comparison to the annealed value at 600◦ C. 3.2 Isochronal heating; initial state: ordered, slowly cooled In Fig. 2 MH values versus annealing temperature of two samples are shown for a slowly cooled ordered initial state followed by isochronal annealing at rising temperatures. The upper curve (▲) has been obtained for a furnace-cooled sample 1, whereas the lower curve (■) represents a sample 2, isochronally cooled down at a linearised rate of 1 K/min. It should be noticed that the final hardness value of sample 1 at 600◦ C is higher than that of sample 2. Normalising the asmeasured values to their final, annealed value, however, the difference in overall hardness of both curves vanishes as discussed in subsection 4.2 below. The following features are observed: Starting with a plateau the hardness of sample 1 decreases between 100 and 220◦ C whereas the hardness of sample 2 continuously increases within this temperature interval. 220◦ C: re-increase of MH of sample 1 and decrease of MH of sample 2. 260◦ C: both samples reach a hardness plateau. Above 280◦ C the final decrease of MH starts for both samples. 3.3 Isochronal cooling (reversed isochronal curve); initial state: disordered In Fig. 2 we have also plotted MH values measured on a sample 3 which was first homogenised and then cooled isochronally (reversed isochronal curve, (). This curve has to be read therefore from the right to the left (in the course of experiment). Upon cooling MH first shows a steep increase, down to 390◦ C followed by a plateau in MH. There is a remarkably wide scatter of averaged MH-values just around 400◦ C, where ordering starts. It may be connected with the formation of ordered domains within a still disordered matrix. A drop of MH is observed below 270◦ C, with a subsequent small reincrease of MH at 210◦ C and a plateau below 150◦ C.. Fig. 2 Temperature dependence of microhardness of three CuAu samples (isochronally heated from the ordered state (▲, ■), isochronally cooled (). The full lines serve as guides for the eye..

(3) 562. M. Spanl et al.. Fig. 3 Three-dimensional AFM-picture of a MH-indentation in a disordered CuAu sample. The diagonal of the indentation is of 27 µm, the height of four slopes is of about 0.9 µm above the original surface.. 3.4 AFM observations In Figs. 3, 4, 5 the topology of the sample surface in the vicinity of a MH-indentation is shown. The sample was quenched from 500◦ C into water; i.e. the material was in disordered state. The indentation did not cross a grain boundary, it hit an interior of a single grain. Even in this simple case it is observed that complicated processes are involved in the transport of the material away from the indentation. Near the indentation border the transported material forms four mounds (Fig. 3) sloping away on the outside for a distance comparable to the size of the indentation. These slopes have a fine structure from individual slip traces (Figs. 4(a), (b)) typical for the fcc structure. Beyond the slopes a wavy deformation of the surface is observed, this effect is visible up to the distance of four diagonals from the indentation (Fig. 3). The wavy deformation looks like a static transversal harmonic wave (Fig. 5), the appropriate propagation vector seems not to be in a correspondence with the fcc slip systems. AFM observations in case of ordered samples did not show clear slip traces and other simple behaviour in the vicinity of an indentation. The main reason seems to be the fact that the indentation of an ordered sample hits more than one ordered domain and the appropriate deformation (transport of the material) is influenced by the presence of antiphase boundaries. 4. Discussion 4.1 Changes of LRO during isochronal annealing Figure 6 gives the change of LRO-parameter as calculated by the Rossiter formalism18) from measured values of electrical resistivity during isochronal heating after a disordered state has been quenched in (for details see19) ). The corresponding ranges of MH in Fig. 1 are indicated by vertical dash-dotted lines. It is observed that the steep increase of MH corresponds to a first very steep increase of LRO until a value of order parameter of about 0.5 is reached. Ranges II and III correspond to the continued but slightly flattened increase of LRO to a maximum value of order (about 0.9) at. Fig. 4 (a), (b) Three-dimensional AFM-pictures (a), (b) of slip traces on two segments of opposite slopes from Fig. 3. Horizontal dimensions of segments are 30 µm × 30 µm, typical height of slip steps is of 0.1 µm.. 340◦ C. Ranges IV and V correspond to the decrease of LRO between 340 and 440◦ C. It therefore turns out that isochronal heating of an initially disordered sample first leads to a drastic increase of LRO before ordering practically stops at about 320◦ C and starts to drop at 340◦ C. Complete disordering occurs when crossing the phase transformation temperature TO/D at 410◦ C. Values of the LRO-parameter for isochronal cooling (reversed isochronal curve) are given in addition in Fig. 6 by the dashed line. It is observed that in this case slightly higher values of LRO-parameter are obtained than corresponding to the LRO-maximum during isochronal annealing at rising temperatures. It has to be noted that in the case of isochronal temperature treatment (in contrast to in-situ experiments) an extraordinary small hysteresis of resistivity, if any, is observed.3, 20) 4.2 Comparison of hardness curves For a better comparison the hardness values of Figs. 1 and 2 have been normalised to the respective value in the disordered state. Figures 7 and 8 give these normalised plots of MH versus annealing temperatures for initially disordered state. In.

(4) Change of Microhardness in Stoichiometric CuAu. 563. Fig. 7 Microhardness data from Fig. 1 normalised to the values at 600◦ C.. Fig. 5 Two-dimensional AFM-picture of a wavy surface deformation from Fig. 3. The height profile corresponds to the line across the waves.. Fig. 6 Temperature dependence of long-range order parameter of CuAu as calculated from resistivity data (see the text) for isochronal heating of an initially disordered sample (■) and for isochronal cooling (- - - -).. Fig. 7 the isochronal heating curves are shown, whereas Fig. 8 (sample 3 of Fig. 2) gives a similar plot for isochronal cooling (reversed isochronal curve). In both cases, rising temperatures for Fig. 7 and falling temperatures for Fig. 8, an initial increase of LRO results (Fig. 6). Three corresponding hardness ranges can be detected: MH first increases steeply with LRO (corresponds to range I for heating) then saturates more or less (range II for heating) and then is reduced (range III for heating) during a continuous increase of LRO. For isochronal cooling (Fig. 8) a continuous increase of LRO can be assumed in correspondence with the dashed line in Fig. 6. In case of the sample 1 in Fig. 2 the thermal treatment was another one: the isochronal annealing at rising temperatures started from a highly ordered state after furnace cooling. The appropriate temperature dependence of LRO can. Fig. 8 Microhardness data from Fig. 2, curve 3, normalised to the value at 450◦ C.. Fig. 9 Comparison of normalised isochronal cooling microhardness data from Fig. 8 () and isochronal heating data from Fig. 2, curve 2, normalised to the value at 600◦ C.. be expected to be similar to the dashed curve in Fig. 6 in this case. We therefore can interpret the MH-behaviour of sample 1 as just a reversion of that of sample 3 (Figs. 2 and 8). If the furnace-cooled curve (isochronal heating) is placed upon Fig. 8 (isochronal cooling), both curves in fact nearly coincide as shown in Fig. 9: Isochronal cooling gives the same changes of hardness for each temperature as isochronal heating starting from a highly ordered state. It means that the results of MH during isochronal heating and cooling (and also LRO-parameter in Fig. 6) do not show a noticeable hysteresis within the isochronal temperature interval of 10 K. This is in remarkable contrast to in-situ experiments and will be shortly discussed in 4.5. The small differences between the two heating curves starting from a LRO state (furnace cooling-curve 1, cooling with 1 K/min-curve 2) in Fig. 2 are probably a result of the differ-.

(5) 564. M. Spanl et al.. ent thermal history of both samples. We expect that the MH depends on a domain structure and also on a co-existence of both ordered phases (CuAu I and CuAu II) in the sample. 4.3 Correlation of hardness changes with changes in the degree of LRO It is noted that all normalised curves (heating and cooling) during the period of ordering show a very similar behaviour of steep increase of hardness (range I), hardness plateau or flat maximum (range II) and decrease of hardness (range III). Whereas there is not a linear relationship between MH and LRO, there is yet a marked relationship, which also corresponds to the early results of Köster.11) An explanation of the corresponding MH-ranges I-III may be given as follows:21) Range I: A first increase of MH with beginning LRO falls into a general pattern of restricted dislocation motion due to an ordered state (generation of antiphase faults by moving dislocations). Range II: It is known that the critical resolved shear stress reaches a maximum for intermediate values of the LROparameter. The reason is that superdislocations are formed in ever greater number as LRO increases, reducing thereby effective lattice friction. This may correspond to saturation in MH or a flat hardness maximum. Range III: Starting with a certain ‘critical’ internal stress level, processes like dislocation generation around elastically strained antiphase domains and twinning may be activated, which lead finally to a reduction of the internal stress level. The peak in isochronal heating curves at about 400◦ C, although more or less pronounced, is observed for all measured isochronal heating curves close to the phase transformation temperature TO/D (Figs. 1 and 7). This may be related to a structural change between CuAuI and CuAuII, depending on heating rate and expected close to the phase transformation temperature at about 400◦ C.4) Due to the rather big temperature step of 25 K this is not seen in the hardness curve reported by Köster,11) although a marked shoulder at 390◦ C can be noticed. The small peak at 400◦ C of the present work on the other hand may also be a consequence of changes in elastic properties during the dissolution of LRO as observed by acoustic emission measurements16) possibly caused by drastic changes of Young’s modulus during disordering. 4.4 Hardness decrease in the range of advanced LRO In the literature numerous indications can be found of structural changes (formation of polytwinned plates) in a temperature range of 200–350◦ C. These changes have been observed during isothermal annealing or during continuous heating or cooling always starting from the disordered state.6, 7, 22–24) The observed minimum of MH in the range of already advanced LRO (Fig. 9) near 230◦ C and the decrease of MH in Fig. 7 between 260 and 320◦ C therefore can be attributed to the expected relief of accumulated internal stresses which accompanies the growth of originally coherent ordered domains in the sample. The underlying processes are presumed to be essentially twinning or a change in twin structure reducing thus the barriers for dislocation motion.. These microstructural processes which occur at late stages of ordering reduce the internal energy. This way the CuAuI phase may be stabilised against transformation into CuAuII and give rise to the retro-effect: It is very probable that such processes are a function of time and temperature6, 7) and that the observed microstructure during ordering depends critically on cooling rate. More detailed optical surface microscopy and additional isothermal MH-measurements are planned to further clarify this problem. 4.5 Hysteresis of ordering effects of first order phase transition It is well known that the order-disorder phase transition in CuAu is of the first order and it is for in-situ studies accompanied by a big thermal hysteresis of up to 50–60 K, depending on the rate of temperature change.2, 3, 25, 26) Isochronal measurements of resistivity change on the other hand did not show a hysteresis.19) The width of a hysteresis range, if any, therefore must be smaller than the isochronal temperature interval of 20 K. In a recent detailed investigation using electrical resistivity measurement the isochronal temperature interval was decreased stepwise to 10 K, 5 K, 2 K, again without observing a marked hysteresis effect.20) This missing of a hysteresis during isochronal temperature treatments is now confirmed by the present MH measurements, a finding corresponding with the earlier measurements of Köster.11) Isochronal annealing experiments are of a quasistatic character: The system has time to relax for the isochronal time interval in the direction of a stable microstructural state corresponding to the actual annealing temperature. If this ‘equilibrium’ value actually is reached will be a question of involved relaxation times, annealing temperature and the actual heating process from room temperature to the annealing temperature. On the other hand during an in-situ experiment the microstructural state of the sample is continuously changed. According to the parameters involved nucleation and growth processes may be limited or suppressed at all. A more detailed theoretical and experimental study of this effect is planned. 5. Conclusions (1) Three ranges of MH changes could be detected after different thermal pre-treatment for isochronal heating and cooling during continuous increase of LRO: range I : steep hardness increase range II : saturation of hardness or flat maximum range III: hardness decrease (2) Within the isochronal temperature intervals of 10 K no hysteresis between heating and cooling experiments could be detected. (3) The decrease in MH in the late stages of ordering (range III) hints at essential microstructural changes. These may help to stabilise CuAu I above its thermodynamic stability range and thus be one of the causes of the retro-effect observed in dynamic experiments. Acknowledgements The work has been financially supported by the Austrian ‘Fonds zur Förderung der Wissenschaftlichen Forschung’..

(6) Change of Microhardness in Stoichiometric CuAu. REFERENCES 1) T. B. Massalski ed.: Binary Alloys Phase Diagrams, (ASM International, Materials Park, OH, 1990) pp. 362–363. 2) B. Sprušil, V. Šı́ma, B. Chalupa and B. Smola: Z. Metallk. 84 (1993) 118–123. 3) B. Chalupa, F. Chmelı́k, V. Šı́ma, B. Sprušil, M. Spanl, H. Lang and W. Pfeiler: Mat. Res. Soc. Symp. Proc. 398 (1996) 581–586. 4) M. Spanl, B. Sprušil and W. Pfeiler: Structural intermetallics, ed by Nathal, M. V., et al. (The Minerals, Metals & Materials Society, Warendale, USA, 1997) pp. 83–89. 5) J. L. O’Brien and G. C. Kuczynski: Acta Metall. 7 (1959) 803–806. 6) M. Hirabayashi and S. Weissmann: Acta Metall. 10 (1962) 25–36. 7) V. S. Arunachalam and R. W. Cahn: J. Mater. Sci. 2 (1967) 160–170. 8) A. J. Pedraza and J. Kittl: Acta Metall. 24 (1976) 835–843. 9) L. Nowack: Z. Metallk. 22 (1930) 94–103. 10) U. Dehlinger and L. Graf: Z. Phys. 64 (1930) 359–377. 11) W. Köster: Z. Metallk. 32 (1940) 145–150. 12) E. R. Petty: Measurement of Mechanical Properties, Part 2, ed. by R. F. Bunshah (Interscience Publishers, New York, 1971) pp. 157–221.. 13) 14) 15) 16) 17) 18) 19) 20) 21) 22) 23) 24) 25) 26). 565. B. Sprušil and W. Pfeiler: Intermetallics 5 (1997) 501–505. B. Sprušil and W. Pfeiler: Intermetallics 8 (2000) 81–83. B. Sprušil and B. Chalupa: Mater. Sci. Eng. A 324 (2002) 58–61. P. Mašek, F. Chmelı́k, V. Šı́ma, A. Brinck and H. Neuhäuser: Acta Mater. 47 (1999) 427–434. V. Šı́ma, P. Mašek, F. Chmelı́k, A. Brinck, H. Neuhäuser, B. Sprušil and W. Pfeiler: Japan Inst. Metals, Proc. 12 (1999) 41–44. P. L. Rossiter: The Electrical Resistivity of Metals and Alloys, (Cambridge University Press, Cambridge, 1987) pp. 160–167. H. Lang: Diploma Thesis, (University of Vienna, Vienna, 1994). M. Spanl and W. Pfeiler: Intermetallics, to be published. P. Haasen: Physical Metallurgy, ed. by R. W. Cahn (North-Holland, Physics Publishing, Amsterdam, 1983) pp. 1341–1410. R. Smith and J. S. Bowles: Acta Metall. 15 (1960) 405–415. J. S. Bowles and C. M Wayman: Acta Metall. 27 (1979) 833–839. G. C. Kuczynski, R. F. Hochman and M. Doyama: J. Appl. Phys. 26 (1955) 871–878. V. Šı́ma: J. Chim. Phys. 90 (1993) 451–456. V. Šı́ma: Mater. Sci. Eng. A 324 (2002) 62–67..

(7)

Figure

Table 1Thermal treatments of different measuring runs.
Fig. 3Three-dimensional AFM-picture of a MH-indentation in a disor-dered CuAu sample. The diagonal of the indentation is of 27 µm, theheight of four slopes is of about 0.9 µm above the original surface.
Fig. 8Microhardness data from Fig. 2, curve 3, normalised to the value at450◦C.

References

Related documents

In spite of the growing body of research on obsolescence and the potential benefits of discarding obsolete technologies (Feng et al., 2007), there remains limited understanding of

on past tSllllamiS ( e.g. historical tSllllami catalogue, magnitude­ frequency data) No historical records on past events Scale of analysis Site-specific Local

Diagram Book Chapters Activity 11 Class 3,5 Communication 12 Component 14 Composite structure 13 Deployment 8 Interaction overview 16 Object 6 Package 7 Sequence 4 State machine

The sands at the base are up to 3 ft thick; there is an extensive section of very finely laminated sands, with thicker sand beds at the top.. The P sand was deposited in deepwater,

Figure 2 shows the distribution of residuals in box and whisker of the d and Vm components, by relative height of the compatible systems S2E3, S1E4 and S2E5, which were selected

jhhb hbjl jlk k su suaaaaj j mk mkstr strj j sgr sgrbt bttu turj rj- - sgr sgrbv bvor oro o su suhb hbtk tk o o ho homo mo bm bm um um um umbg bgk k na naus

Applicants should submit passport, Certificate of Eligibility, Certificate of Admission and any other required documents to Japanese Embassy or Consulate in their home