Methods of Ex Situ and In Situ Investigations of Structural Transformations: The Case of Crystallization of Metallic Glasses

Summary

Here, we present a protocol to describe ex situ and in situ investigations of structural transformations in metallic glasses. We employed nuclear-based analytical methods which inspect hyperfine interactions. We demonstrate the applicability of Mössbauer spectrometry and nuclear forward scattering of synchrotron radiation during temperature-driven experiments.

Abstract

We demonstrate the use of two nuclear-based analytical methods that can follow the modifications of microstructural arrangement of iron-based metallic glasses (MGs). Despite their amorphous nature, the identification of hyperfine interactions unveils faint structural modifications. For this purpose, we have employed two techniques that utilize nuclear resonance among nuclear levels of a stable ⁵⁷Fe isotope, namely Mössbauer spectrometry and nuclear forward scattering (NFS) of synchrotron radiation. The effects of heat treatment upon (Fe_2.85Co₁)₇₇Mo₈Cu₁B₁₄ MG are discussed using the results of ex situ and in situ experiments, respectively. As both methods are sensitive to hyperfine interactions, information on structural arrangement as well as on magnetic microstructure is readily available. Mössbauer spectrometry performed ex situ describes how the structural arrangement and magnetic microstructure appears at room temperature after the annealing under certain conditions (temperature, time), and thus this technique inspects steady states. On the other hand, NFS data are recorded in situ during dynamically changing temperature and NFS examines transient states. The use of both techniques provides complementary information. In general, they can be applied to any suitable system in which it is important to know its steady state but also transient states.

Introduction

Iron-based MGs prepared by rapid quenching of a melt represent industrially attractive materials with numerous practical applications¹. Especially since their magnetic properties are often superior to conventional (poly)crystalline alloys^2,3. To better benefit from their advantageous parameters, their response to elevated temperatures should be known. With increasing temperature, the amorphous structure relaxes and, finally, the crystallization starts. In some types of MGs, this can lead to the deterioration of their magnetic parameters and, consequently, poorer performance. There are, however, several families of iron-based MGs with special compositions^4,5,6,7 in which the newly formed crystalline grains are very fine, typically below about 30 nm in size. The nanocrystals stabilize the structure and thus, preserve acceptable magnetic parameters over a wide temperature range^8,9. These are the so-called nanocrystalline alloys (NCA).

The long-term performance reliability of MGs, especially under elevated temperatures and/or tough conditions (ionizing radiation, corrosion, etc.) demands thorough knowledge of their behavior and individual physical parameters. Because MGs are amorphous, the assortment of analytical techniques that are suitable for their characterization is rather limited. For example, diffraction methods provide broad and featureless reflections that can be used only for the verification of amorphicity.

It is noteworthy that several, usually indirect methods exist that provide fast and non-destructive characterization of MGs (e.g., magnetostrictive delay line sensing principle). This method provides fast characterization of structural and stress states including the presence of inhomogeneities. It was advantageously applied to fast and non-destructive characterization along the whole length of MG ribbons^10,11.

More detailed insight into disordered structural arrangement can be achieved via hyperfine interactions that sensitively reflect the local atomic arrangement of the resonant atoms. Moreover, variations in topological and chemical short-range order can be revealed. In this respect, the methods like nuclear magnetic resonance (NMR) spectrometry and/or Mössbauer spectrometry, both performed on ⁵⁷Fe nuclei, should be considered^12,13. While the former method provides response exclusively to magnetic dipole hyperfine interactions, the latter is sensitive also to the electric quadrupole interactions. Thus, Mössbauer spectrometry makes simultaneously available information on both the structural arrangement and magnetic states of the resonant iron nuclei¹⁴.

Nevertheless, to achieve reasonable statistics, the acquisition of a Mössbauer spectrum usually takes several hours. This restriction should be considered especially when temperature-dependent experiments are envisaged. Elevated temperature that is applied during the experiment causes structural modifications in the investigated MGs¹⁵. Consequently, only ex situ experiments performed at room temperature upon samples that were first annealed at certain temperature and then returned to ambient conditions provide reliable results.

The evolution of MG structures during heat treatment is routinely studied by analytical techniques which enable rapid data acquisition as for example X-ray diffraction of synchrotron radiation (DSR), differential scanning calorimetry (DSC), or magnetic measurements. Though in situ experiments are possible, the obtained information concerns either structural (DSR, DSC) or magnetic (magnetic data) features. However, in the case of DSC (and magnetic measurements) the identification of the type of (nano)grains that emerge during the crystallization is not possible. On the other hand, DSR data do not indicate the magnetic states of the investigated system. A solution to this situation is a technique that makes use of hyperfine interactions: NFS of synchrotron radiation¹⁶. It belongs to a group of methods that exploits nuclear resonant scattering processes¹⁷. Due to extremely high brilliance of radiation obtained from the third generation of synchrotrons, temperature NFS experiments under in situ conditions became feasible^{18,19,20,21,22,23}.

Both Mössbauer spectrometry and NFS are governed by the same physical principles related to nuclear resonance among energy levels of ⁵⁷Fe nuclei. Nevertheless, while the former scans hyperfine interactions in the energy domain, the latter provides interferograms in the time domain. In this way, the results obtained from both methods are equivalent and complementary. To evaluate the NFS data, a reasonable physical model must be established. This challenging task can be accomplished by the help of Mössbauer spectrometry which provides the first estimate. Complementarity between these two methods means that in situ NFS inspects transient states and Mössbauer spectrometry reflects the stable states, viz. the initial and/or the final state of a material studied ex situ.

This article describes in detail selected applications of these two less common methods of nuclear resonances: here, we apply them to the investigation of structural modifications that occur in a (Fe_2.85Co₁)₇₇Mo₈Cu₁B₁₄ MG exposed to heat treatment. We hope that this article attracts the interest of researchers to use these techniques for the investigation of similar phenomena and eventually with different types of materials.

Protocol

1. Preparation of a MG

NOTE: To demonstrate a broad range of diagnostic capabilities of NFS in combination with Mössbauer spectrometry, an appropriate MG composition was designed, namely (Fe₃Co₁)₇₆Mo₈Cu₁B₁₅ (at.%). This system shows the magnetic transition from the ferromagnetic to paramagnetic state below the onset of crystallization. Moreover, crystallites that emerge during the first crystallization step form bcc-Fe,Co phase. Because cobalt replaces iron in some atomic positions of the bcc lattice, deviations in the respective hyperfine interactions occur.

1. Preparation of the melt
   NOTE: Mössbauer spectrometry and NFS scan the local atomic arrangements via hyperfine interactions of ⁵⁷Fe nuclei that are present in the investigated samples. Natural abundance of this stable nuclide among all Fe isotopes is, however only 2.19%. To decrease the acquisition time of in situ NFS experiments, the relative content of the ⁵⁷Fe isotope should be increased to about 50%.
   1. Take a quartz glass crucible (cylindrical shape with a diameter of 15 mm), cover its inner walls with boron nitride to avoid possible contamination of the content by Si from the walls, and insert 0.4050 g of highly enriched ⁵⁷Fe (~ 95%) and 0.5267 g of standard electrolytic pure iron (purity 99.95%) to this crucible. The total mass of the mixture is of 0.9317 g and ensures isotopic enrichment of ca. 50% ⁵⁷Fe.
      NOTE: Because of the high price of the stable ⁵⁷Fe isotope, optimize its amount to the lowest possible mass. Approximately 500 mg of ⁵⁷Fe should be enough to ensure the overall weight of the melt to approximately 1.5 g. This is the low technological limit of the preparation equipment.
   2. Add 0.3245 g of electrolytic Co (99.85%), 0.0184g of Cu (99.8%), 0.2222 g of Mo (99.95%) and 0.0470 g of crystalline B (99.95%) into the same quartz glass crucible. The total mass of the mixture is of 1.5438 g and the intended composition of the powder is (Fe₃Co₁)₇₆Mo₈Cu₁B₁₅.
   3. Melt the obtained mixture of standard electrolytic materials by inductive heating in a quartz glass crucible under protective Argon (4N8) atmosphere to avoid oxidation, and use a radiofrequency field of 90-120 kHz.
      NOTE: The radio frequency field ensures mixing of the individual components in the crucible. Their mixing proceeds further by the help of eddy currents when a melt is formed. Allow sufficient time to melt the powder mixture and form a liquid. Visual inspection is sufficient, there is no need to measure the temperature of the resulting liquid.
   4. Remove the obtained small ingot from the crucible. Visually check the occurrence of any traces of slag spots on its surface. If present, remove them by mechanical polishing.
2. Preparation of the ribbon-shaped sample
   1. Use an apparatus for planar flow casting. An example of such a device is shown in Figure 1.
      NOTE: The melt inside a quartz tube is expelled by Ar flow onto a quenching wheel which rotates in the air. There is no need for special atmospheric conditions under which the quenching wheel is operated (e.g., vacuum or inert gas environment) for this composition of the melt.
   2. Because of small weight of the ingot (~1.5 g), choose a quartz tube with a nozzle that has a round orifice of 0.8 mm in diameter. Put the ingot inside and melt it using inductive heating. Keep the temperature of the melt at 1,280-1,295 °C.
   3. Adjust the surface speed of the cooling wheel to 40 m/s.
   4. Cast the melt upon the rotating quenching wheel under ambient conditions, i.e., in the air.
      NOTE: The resulting ribbon is about 1.5-2 mm wide, 25-27 µm thick, and 5 m long. The air side of the ribbon, which was exposed during the production to the surrounding air atmosphere, is optically shiny (glossy) while the opposite wheel side, which was in direct contact with the quenching wheel, is mat (dull). These subtle ribbon qualities result from the low mass of the melt. Thus, it is important to verify the final chemical composition of the produced as-quenched ribbon because of the low input masses of the individual elements.
3. Verification of the final chemical composition of the ribbon
   1. Prepare several (up to five) short pieces of the ribbon, each having a mass of about 0.70 mg. Chose them from different parts of the produced ribbon along its length.
   2. Dissolve every single piece of ribbon in 1 mL of concentrated (67%) HNO₃ acid and fill with water to reach 50 mL total volume of the solution.
   3. Determine the content of Mo and B by optical emission spectrometry with inductively coupled plasma (OES-ICP). Use the method of external calibration as provided in the instrument's manual. Record the signals at the following wavelengths: Mo at 203.844 nm and 204.598 nm, and B at 249.773 nm.
   4. Determine the content of Fe, Co, and Cu by flame atomic absorption spectrometry (F-AAS). Use the method of external calibration as provided in the instrument's manual, and select these wavelengths: Fe at 248.3 nm, Co at 240.7 nm, and Cu at 324.7 nm.
4. Structural characterization of the produced ribbons
   1. Check the amorphous nature of the produced ribbons by performing X-ray diffraction (XRD) in Bragg-Brentano geometry; use the Cu anode with a wavelength of 0.154056 nm, record the diffraction pattern from 20-100° of 2Θ with an angular step of 0.05° and acquisition time of 20 s for one point.
      NOTE: The XRD diffractogram of an amorphous sample is characterized by broad reflection peaks like those shown in Figure 2. No narrow lines that indicate the presence of crystallites should be present.
   2. Prepare small pieces of the produced ribbons with a total mass of about 3-5 mg and place them into a graphite crucible of a DSC equipment.
      NOTE: Small pieces of about 2 mm in length can be cut off from the ribbon by scissors.
   3. Perform the DSC experiment with a temperature ramp of 10 K/min in a temperature range of 50-700 °C under Ar atmosphere.
   4. Determine the temperature of the onset of crystallization T_x1, which is taken at the kink of the most pronounced peak on the DSC curve.
      NOTE: The temperature of the onset of crystallization T_x1 is indicated in Figure 3 by an arrow.
   5. Chose five temperatures of annealing that cover both the pre-crystallization and crystallization regions on the DSC for further ex situ annealing.
      NOTE: In our case, appropriate temperatures are 370, 410, 450, 510, and 550 °C as shown in Figure 3.
5. Ex situ annealing
   1. Prepare five groups of ~ 7 cm long pieces (the total length) of the as-quenched ribbon. The individual ribbons should be at least 1 cm long.
   2. For ex situ annealing, use a furnace (Figure 4). Set up the destination temperature and wait for 15 min for its stabilization.
      NOTE: The furnace design ensures minimal onset times for isothermal annealing. This furnace consists of two parts: upper and lower round massive nickel-plated copper blocks that act as a temperature homogenizer. Kanthal A strips heat up the blocks with high dynamics of temperature regulation and stabilization. The destination temperature is that determined in step 1.4.5.
   3. Insert pieces of the ribbon into the evacuated and thermally stabilized zone. To do this, open a 7-10 mm gap between the two blocks and slide the ribbons directly into the center of the heated zone.
   4. Close the gap immediately. In this way, the temperature of the sample achieves the furnace temperature in less than 5 s within a 0.1 K difference.
   5. Perform the annealing at 370, 410, 450, 510, and 550 °C for 30 min under vacuum to prevent surface oxidation.
   6. After annealing, remove the heated ribbons and place them on a cold substrate inside the vacuum system. This ensures fast cooling of the samples to room temperature.
      NOTE: Thermal treatment of the as-quenched ribbons induces structural changes that eventually lead to crystallization of the originally amorphous material.

2. Methods of Investigation

1. Mössbauer spectrometry
   NOTE: The use of iron enriched to about 50% in ⁵⁷Fe for production of the studied MG ensures sufficiently short acquisition times for in situ NFS experiments. On the other hand, the effective thickness of the ribbons is significantly increased. This poses issues related to extremely high broadening of the absorption Mössbauer spectral lines recorded in a conventional transmission geometry experiment. That is why surface sensitive techniques of Mössbauer spectrometry should be considered. Namely, Conversion Electron Mössbauer Spectrometry (CEMS) and Conversion X-ray Mössbauer Spectrometry (CXMS) can be applied. While CEMS scans subsurface regions to the depth of about 200 nm, CXMS provides information from deeper areas that extend down to about 5-10 µm.
   1. Prepare the samples for CEMS/CXMS experiments; use 6-8 pieces of ~ 1 cm long ribbons for one sample.
   2. Attach the ribbons side-by-side to an aluminum holder to form a compact area of about 1 x 1 cm²; use adhesive tape over the ends of the ribbons; all ribbons must be placed with their air sides upwards.
      NOTE: Ensure that there is an electric contact between the ribbons and the holder and that the central part of the sample (about 8 x 10 mm²) is clean from any surface contamination, e.g., remains of the adhesive tape.
   3. Insert the aluminum holder with the sample into the CEMS/CXMS detector.
   4. Prior to the measurement, thoroughly wash the inner detector volume with a stream of the detection gas to expel all residual air. Allow 10-15 min to accomplish this procedure.
   5. Adjust the gas flow through the detector by a needle valve to 3 mL/min.
   6. Connect a high voltage to the detector: a typical value is about 1.2 kV for CEMS and about 200 V higher for CXMS.
   7. Record the CEMS and CXMS Mössbauer spectra using a constant acceleration spectrometer equipped with a ⁵⁷Co/Rh radioactive source. Operate the spectrometer with a gas detector at room temperature according to the manual.
   8. Accomplish the detection of conversion electrons and X-rays by a gas detector filled with a He+CH₄ and Ar+CH₄ gas mixture, respectively. Keep the amount of CH₄ at 10% in both cases.
   9. Repeat steps 2.1.2 to 2.1.8 for the wheel side of the investigated ribbons.
   10. Perform velocity calibration¹⁴ of the apparatus by using a thin (12.5 µm) α-Fe foil.
   11. Evaluate the CEMS/CXMS spectra; quote the obtained isomer shift values with respect to a room temperature Mössbauer spectrum of the calibration α-Fe foil.
       NOTE: The obtained Mössbauer spectra can be evaluated by any suitable fitting code, for example by the Confit software²⁴.
2. NFS
   1. Accomplish the NFS experiments using a suitable nuclear resonance beamline at a synchrotron. A possible option: ID 18 at the European Synchrotron Radiation Facility (ESRF) in Grenoble, France.²⁵
   2. Tune the energy of the photon beam to 14.413 keV with a bandwidth of ~ 1 meV.
   3. Place an approximately 6 mm long ribbon of the investigated MG in a vacuum furnace.
   4. Record the NFS time-domain patterns during continuous heating of the sample to a temperature of up to 700 °C with a ramp of 10 K/min. Use 1-min time intervals for acquisition of experimental data during the entire in situ annealing process.
      NOTE: The transmission geometry of the NFS experiment ensures that information on hyperfine interactions is obtained from the sample's bulk.
   5. Evaluate the NFS experimental data using a suitable software (e.g., www.nrixs.com).
      NOTE: During one in situ experiment, typically up to 100 NFS time-domain patterns are recorded. During their evaluation by the CONUSS software package^26,27, consider the application of a special free software called Hubert that can evaluate such enormous data quantities in a semi-automatic mode²⁸.

Results

The XRD pattern in Figure 2 exhibits broad featureless diffraction peaks. The observed reflections demonstrate that the produced ribbon of the (Fe_2.85Co₁)₇₇Mo₈Cu₁B₁₄ MG is XRD amorphous.

Due to its sensitivity, XRD has some limitations in unveiling surface crystallization. The presence of crystallites amounting to less than about 2-3%...

Discussion

Ex situ Mössbauer effect experiments describe a steady situation which is encountered in the investigated MG after the applied heat treatment. Each spectrum was collected for a duration of several hours at room temperature. Thus, the evolution of the originally amorphous structure was followed as a function of annealing conditions. Because Mössbauer spectrometry is sensitive to hyperfine interactions acting upon the resonant nuclei, faint details of structural and/or magnetic modifications induced by e...

Materials

Name	Company	Catalog Number	Comments
stable isotope, 57Fe	Isoflex USA	iron-57	metallic form
standard eletrolytic Fe, 99.95 %	Sigma Aldrich (Merck)	1.03819	fine powder
electrolytic Co, 99.85 %	Sigma Aldrich (Merck)	1.12211	fine powder
electrolytic Cu, 99.8 %	Sigma Aldrich (Merck)	1.02703	fine powder
electrolytic Mo, 99.95 %	Sigma Aldrich (Merck)	1.12254	fine powder
crystalline B, 99.95 %	Sigma Aldrich (Merck)	266620	crystalline
calibration foil for Mössbauer spectrometry, bcc-Fe	GoodFellow	564-385-23	foil 0.0125 mm, purity 99.85 %
HNO3 acid, ANALPURE Ultra	Analytika Praha, Czech Republic	UAc0061a	concentration 67 %, volume 500 mL
spectrometer for atomic absorption spectrometry	Perkin Elmer 1100, Germany
spectrometer for optical emmission spectrometry with inductively coupled plasma	Jobin Yvon 70 Plus, France
X-ray diffractometer	Bruker D8 Advance, USA
differential scanning calorimeter	Perkin Elmer DSC 7, Germany