Observation of a continuous phase transition in a shape-memory alloy
J. C. Lashley, S. M. Shapiro, B. L. Winn, C. P. Opeil, M. E. Manley,, A. Alatas, W. Ratcliff, T. Park, R. A. Fisher, B. Mihaila, P. Riseborough, E., K. H. Salje, and J. L. Smith

TL;DR
This study investigates the nature of the phase transition in AuZn and Au_{0.52}Zn_{0.48} alloys, revealing a continuous transition in AuZn consistent with mean-field theory and identifying a critical endpoint in the phase diagram.
Contribution
It provides the first evidence of a continuous phase transition in a shape-memory alloy and characterizes the critical endpoint using neutron scattering, specific heat, and pressure measurements.
Findings
AuZn exhibits a continuous phase transition.
The transition in AuZn aligns with mean-field theory predictions.
A critical endpoint is identified at T_M*=2.7K and p*=3.1GPa.
Abstract
Elastic neutron-scattering, inelastic x-ray scattering, specific-heat, and pressure-dependent electrical transport measurements have been made on single crystals of AuZn and Au_{0.52}Zn_{0.48} above and below their martensitic transition temperatures (T_M=64K and 45K, respectively). In each composition, elastic neutron scattering detects new commensurate Bragg peaks (modulation) appearing at Q = (1.33,0.67,0) at temperatures corresponding to each sample's T_M. Although the new Bragg peaks appear in a discontinuous manner in the Au_{0.52}Zn_{0.48} sample, they appear in a continuous manner in AuZn. Surprising us, the temperature dependence of the AuZn Bragg peak intensity and the specific-heat jump near the transition temperature are in favorable accord with a mean-field approximation. A Landau-theory-based fit to the pressure dependence of the transition temperature suggests theā¦
Click any figure to enlarge with its caption.
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5
Figure 6Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Observation of a continuous phase transition in a shape-memory alloy
J. C. Lashley
Los Alamos National Laboratory, Los Alamos, NM 87545, USA
āā
S. M. Shapiro
Brookhaven National Laboratory, Upton, NY 11973, USA
āā
B. L. Winn
Oak Ridge National Laboratory, Oak Ridge, TN 37831, USA
āā
C. P. Opeil
Boston College, Department of Physics, 140 Commonwealth Avenue, Chestnut Hill, MA 02467, USA
āā
M. E. Manley
Lawrence Livermore National Laboratory, Livermore, CA 94550, USA
āā
A. Alatas
Advanced Photon Source, Argonne National Laboratory, Argonne, IL 60439, USA
āā
W. Ratcliff
National Institute of Standards and Technology, Gaithersburg, MD 20899, USA
āā
T. Park
Los Alamos National Laboratory, Los Alamos, NM 87545, USA
Department of Physics, Sungkyunkwan University, Suwon 440-746, Korea
āā
R. A. Fisher
Los Alamos National Laboratory, Los Alamos, NM 87545, USA
āā
B. Mihaila
Los Alamos National Laboratory, Los Alamos, NM 87545, USA
āā
P. Riseborough
Department of Physics, Temple University, Philadelphia, PA 19122, USA
āā
E. K. H. Salje
Department of Earth Sciences, University of Cambridge, Downing Street, Cambridge CB2 3EQ, UK
āā
J. L. Smith
Los Alamos National Laboratory, Los Alamos, NM 87545, USA
Abstract
Elastic neutron-scattering, inelastic x-ray scattering, specific-heat, and pressure-dependent electrical transport measurements have been made on single crystals of AuZn and Au0.52Zn0.48 above and below their martensitic transition temperatures (Ā =Ā 64Ā K and 45Ā K, respectively). In each composition, elastic neutron scattering detects new commensurate Bragg peaks (modulation) appearing at at temperatures corresponding to each sampleās . Although the new Bragg peaks appear in a discontinuous manner in the Au0.52Zn0.48 sample, they appear in a continuous manner in AuZn. Surprising us, the temperature dependence of the AuZn Bragg peak intensity and the specific-heat jump near the transition temperature are in favorable accord with a mean-field approximation. A Landau-theory-based fit to the pressure dependence of the transition temperature suggests the presence of a critical endpoint in the AuZn phase diagram located at Ā =Ā 2.7Ā K and Ā =Ā 3.1Ā GPa, with a quantum saturation temperature Ā =Ā 48.33.7Ā K.
pacs:
81.30.Kf, 71.20.Be,
A class of materials exhibiting martensitic (diffusionless) phase transformations yields properties used in a range of technological applications including implants to increase flow in restricted blood vesselsĀ Duerig , actuators for the treatment of high myopiaĀ Yu , voltage generatorsĀ Suorsa , and orthodontic arch-wiresĀ Jafari . These properties often depend on the history of the material and may allow it to recover its previous shape after deformation, known as the shape-memory effect (SME). It has long been recognized that these transformations are all thermodynamically first order (discontinuous)Ā Albers ; Boyd ; Gash . Special cases arise when the order parameter is coupled to an external field in a complicated way, leading to a weakly first-order transitionĀ Rubini , which is believed to result from a complicated coupling between strain and order parameter fluctuations. Because shape-memory alloys and other multiferroic materials owe their functionality to complicated cross-field responses between two (or more) pairs of conjugate thermodynamic variables, a description of the free-energy landscape and the ability to predict functionality become one and the same.
Phenomenological descriptions of the martensitic transformation pathway, based on reciprocal spaceĀ Bowles and real spaceĀ David geometries, have established rules to determine relative twin and crystallographic orientations between the austenite (high-temperature) and martensite (low-temperature) modifications. Knowledge of the symmetry breaking allows for a definition of the energetic driving force in terms of the difference in free energies between phases. In the Ginzburg-Landau (GL) approach, the free energy, , is expressed as a sum of symmetry invariants. In its simplest form, is approximated by a polynomial expansion in even powers of an order parameter, , as . For martensitic transformations, is taken to be strain or strain coupled to shuffle (displacements involving quasi-static phonons with fractional commensurate wave vectors with uniform shearsĀ Burgers ). The coefficients , and are material parameters to be determined experimentally or by first principle simulations.
For the majority of shear-induced transformations, the GL expansion captures the essential physicsĀ Toledano ; Salje , such as constitutive responseĀ Avadh1 ; Iwata and the occurrence of anti-phase boundariesĀ Cao . When applied to shape-memory alloys, the presence of intervening (premartensitic) phasesĀ Sethna presents difficulties for both experiment and theory as to a unique characterization of the order parameter. Further issues have arisen in materials exhibiting quantum mechanical effects in the band structure and strong electron-phonon couplingĀ Avadh2 . The thermodynamic behavior is best explored when no further transitions occur between TM and 0Ā K. This effect was recently reported for Fe-doped NiTi, where all thermodynamic parameters could be correlated self-consistently within mean field theoryĀ Toni . In order to integrate these cases into GL theory, it is necessary to measure and access each thermodynamic property and infer cross-field couplings of the order parameter. The ability to distinguish subtle differences between first-order and weakly first-order depends on the ability to resolve the behavior of pertinent physical properties.
In this Letter we examine the AuZn system for thermodynamic properties that contribute to the first-order or weakly first-order nature of the free-energy landscape. We measure elastic neutron-scattering, inelastic x-ray scattering, specific-heat, and the pressure dependence of the transition by electrical transport. While we observed classical first-order behavior at 45Ā K for Au0.52Zn0.48, we observed a mean field continuous transformation for AuZn near 64Ā K in the temperature dependence of the Bragg peak intensity, and a Landau step with a anomaly in the excess specific heat.
Single crystals were prepared by fusion of the elements in a Bridgman furnace and were oriented by back-reflection Laue diffraction. The neutron experiments were performed on the BT9 triple-axis spectrometer at the NIST research reactor. Measurements of the phonon dispersion were made on XOR of sector 3 at the Advanced Photon Source, Argonne National LaboratoryĀ Sinn1 ; Sinn2 . Specific-heat measurements were measured using a thermal-relaxation calorimeter by Quantum DesignĀ Lashley . The pressure dependence of the transition temperature was made by a four-terminal ac-transport method in a mechanical pressure cell designed to reach pressures of 3Ā GPa.
Previous investigations have shown that AuZn exhibits a martensitic transition as seen with low-temperature electron microscopyĀ Ted and by recoverable transformation strain (shape-memory effect)Ā Darling . FigureĀ 1 shows the temperature dependence of (a) the excess specific heat and (b) the satellite peak intensity for Au0.52Zn0.48 and AuZn. Although we have measured the specific heat previouslyĀ Ross , it is plotted along with the elastic neutron-scattering data, since it is revealing to compare both sets of measurements that were made on the same samples. We find a latent heat anomaly at 45Ā K for Au0.52Zn0.48. A -anomaly was observed for AuZn, but no latent heat was observed. The specific heat near the transition can be influenced by intrinsic defects, including triple-point and anti-structure defects, both known to be present in the B2 structureĀ Austin ; Gupta . It was therefore necessary to use a microscopic probe to determine whether or not the transition could be a broadened first-order transition. Elastic neutron-scattering measurements indicate that in each composition new commensurate Bragg peaks (modulation) appear at at temperatures corresponding to each sampleās Ā FigureĀ 1(b). Mirroring the specific-heat data, the temperature dependence of the satellite peak intensity shows a rapid jump at 45Ā K in Au0.52Zn0.48, while a continuous variation occurs for AuZn on warming starting at 64Ā K. The satellite peak intensity is proportional to the square of the order parameter, since it leads to the low-temperature rhombohedral phase. The specific-heat data also indicate thermal hysteresis (not shown) of 6Ā K at 45Ā K for Au0.52Zn0.48. The satellite peak intensity for AuZn depicted in Fig.Ā 1(b) was measured by neutrons on heating and cooling. Unlike the specific-heat data Fig.Ā 1(a) we observe a small discontinuity and 1.5Ā K difference in transition temperature on cooling.
FigureĀ 2 shows the satellite peak intensity versus temperature in the vicinity of for AuZn. Near the transition temperature, the satellite peak evolves continuously with increasing temperature, showing a temperature dependence given by where the satellite peak intensity is , the critical temperature is , and the critical exponent isĀ . We obtain TM=64.08Ā K 0.11Ā K and 2=Ā 1.020.08. The value of the critical exponent obtained from this fit (=0.51) is close to the mean-field value of =0.50Ā Binney . One could argue that mean-field exponents were obtained by coincidence given that the critical regime, as defined by the Ginzburg-criterion, is restricted to a range of temperatures too small to be resolved by our experiments, or that the correlation length does not diverge, since it is limited by other extrinsic disorder (i.e., martensite twins) or intrinsic fluctuations. The latter possibility seems more plausible to us, because the inverse correlation length, or width of the Bragg peaks (not shown), increases as the temperature is lowered below .
As the temperature is lowered, the unit cell is modulated in the [110] shear direction. A commensurate shuffle of every third unit cell results in a hexagonal primitive unit cell formed from nine primitive cubic cells of the parent phase. This structure can also be described in terms of its conventional rhombohedral unit cell. In Fig.Ā 3 we show the TA2 phonon-dispersion curves along [110], measured by inelastic x-ray scattering at temperatures of 300Ā K and 200Ā K. The inflection in the phonon frequency near indicates the low shear instability along the TA2[110] branch. This inflection is comparable to an earlier investigationĀ Makita of off-stoichiometric AuZn samples. At 300Ā K, the phonon energy positions have been fit over the energy range -4Ā meVĀ Ā 4Ā meV to a double Lorentzian, as shown in FigureĀ 4. The phonons continuously soften with decreasing temperature to the point where separation becomes difficult below 200Ā K.
FigureĀ 5 (a) shows the change in resistivity between cooling and warming. The data were collected with warming and cooling rates of 0.2 K min*-1*. In the region of there is no hysteresis, to within experimental error. Because of the lack of hysteresis, we elected to measure the pressure dependence of the transition. For a first-order transition, the Clausius-Clapeyron relation, , predicts a linear relationship forĀ versus with the slope . Here, denotes the pressure, the enthalpy, the universal gas constant, and is a reference pressure. FigureĀ 5(b) shows the plot of versus . It appears that there is no observable linear region. Assuming that the entropy does not vary significantly with pressure, the conditions for the applicability of the Clausius-Clapeyron relation are not met. This result provides further evidence that the transition is not first order.
We use a modified GL free energy expression to describe the quantum saturation in the order parameter expected at sufficiently low temperatureĀ Qsat . This saturation effect reflects the departure from classical behavior as the absolute temperature approaches zero and is a direct consequence of the third law of thermodynamics. Following Salje et al.Ā Qsat , we fit the pressure dependence of the transition as \theta_{s}/T_{M}(p)=\coth^{-1}\bigl{[}\coth(\theta_{s}/T_{M}(0))-p/p_{0}\bigr{]}. Here is a phenomenological temperature below which quantum mechanical effects are dominant. We find Ā =Ā 5.60.7Ā GPa and =48.33.7Ā K. The resulting phase diagram predicts a critical endpoint at Ā =Ā 2.7Ā K and Ā =Ā 3.1Ā GPa. FigureĀ 6 depicts the experimental pressure dependence of the transition temperature for pressures up to nearly 3Ā GPa, together with the fit to Saljeās classical regime. We find a classical slope of -11.70.2Ā K GPa*-1*. The difference between the classical and quantum regimes is important for temperatures belowĀ .
In conclusion, we report results of a detailed experimental study of the martensitic transition in AuZn. While the temperature dependence of the satellite peak intensity (proportional to the square of the order parameter) show marked first-order behavior at 45Ā K for Au0.52Zn0.48, we observed a continuous feature well-described by a mean-field exponent (=0.51) for AuZn at 64Ā K. This result is contrary to the established definition of a martensitic transition. Providing supporting evidence to the continuous nature of the phase transition in AuZn are the lack of thermal hysteresis from electrical-resistivity measurements, the -anomaly in the specific heat, and the disagreement of pressure data with the Clausius-Clapeyron relation. Using Saljeās model to describe the quantum saturation effects close to zero absolute temperature, we predict the presence of a critical endpoint in the phase diagram located at Ā =Ā 2.7Ā K and Ā =Ā 3.1Ā GPa, with a phenomenological temperature =48.33.7Ā K for the onset of quantum effects.
Acknowledgements.
This work was performed under the auspices of the United States Department of Energy and Department of Commerce and supported in part by the Trustees of Boston College. Work at Brookhaven is supported by the Office of Science, United States Department of Energy under Contract No. DE-ACO2-98CH10886. The use of the Advanced Photon Source was supported by the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences, under Contract No. DE-AC02-06CH11357.
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1(1) P. W. Duerig, MRS Bulletin 27 , 101 (2002).
- 2(2) I. Yu et al. , Mater. Sci. Eng. A 481-482 , 651 (2008).
- 3(3) I. Suorsa et al. , J. Appl. Phys. 95 , 8054 (2004).
- 4(4) J. Jafari, S. M. Zebarjad, and S. A. Sajjadi, Mater. Sci. Eng. A. 473 , 42 (2008).
- 5(5) R. C. Albers et al. , Comp. and Appl. Math. 23 , 345 (2004).
- 6(6) J. G. Boyd and D. C. Lagoudas, Int. J. Plasticity 12 , 805 (1996).
- 7(7) P. W. Gash, J. Appl. Phys. 54 , 6900 (1983).
- 8(8) S. Rubini and P. Ballone, Phys. Rev. B. 48 , 99, (1993).
