In situ measurement of austenite grain growth and recrystallization using laser ultrasonics

The development of next generation process models and advanced high-strength steel products for thin slab casting and direct rolling requires quantification of microstructure evolution during thermomechanical processing. Laser ultrasonics is a non-contact in-situ method to record grain growth, recrystallization and phase transformations in metals and alloys. Here, we will present an improved experimental design that facilitates a continuous microstructure measurement through the various stages of simulated hot rolling from reheating to runout table cooling using a Gleeble thermomechanical simulator equipped with a laser ultrasonics for metallurgy (LUMet) system. Austenite grain growth and static recrystallization after hot deformation are quantified based on attenuation of the ultrasound waves whereas austenite decomposition can be recorded with the changes in ultrasound velocity during the phase transformation. Further, the LUMet results for a microalloyed low carbon steel are validated with conventional techniques including optical and electron microscopy as well as double-hit tests. These experimental studies demonstrate the capabilities of laser ultrasonics in the identification of both normal and abnormal grain growth, non-recrystallization temperature, recrystallization, and austenite decomposition kinetics in a single test for a given processing path, as well as its potential for accelerated optimization of process control under industrial rolling conditions.


Introduction
The ever-growing demand for advanced high-strength steels necessitates the development of metallurgical models that correlate process variables with microstructural evolution and structureproperty relationships.For this purpose, vast amounts of data need to be acquired from experimental studies in terms of how microstructure parameters such as grain size and phase fraction respond to different deformation and heat treatment conditions.One common practice is to conduct individual thermomechanical simulations for each metallurgical process, which are then integrated into an overall microstructure model that can be adapted to more complex industrial scenarios [1].Although this can be done using conventional methods such as double-hit tests for the study of softening during recrystallization, dilatometry to measure austenite decomposition as well as optical and/or electron microscopy to quantify austenite and ferrite grain sizes, these techniques are time-consuming and for microscopy destructive thereby losing in situ information that is extremely valuable for monitoring a series of events along an entire processing path [2].In this regard, laser ultrasonics appears to be a promising alternative in which ultrasound waves are generated and detected by laser pulses to obtain quantitative information on microstructure evolution.
Previous studies have shown that the two important properties of ultrasound, namely attenuation and velocity, can be related to the average grain size and elastic properties of materials, respectively.This inspired a wide range of applications in the evaluation of elastic moduli, texture, grain growth, recrystallization, and phase transformation, etc. [3][4][5][6].The present work aims to further advance the experimental design for LUMet measurements of austenite grain growth, recrystallization, and decomposition on selected microalloyed steels such that continuous in situ measurements of microstructure evolution can be performed in conditions that simulate industrial thermo-mechanical schedules.

Materials and experimental procedures
The steel investigated in this work is a Nb-microalloyed low-carbon steel supplied by Algoma Steel with the chemistry shown in table 1.The as-received hot-rolled material was machined to different test pieces and polished to a smooth finish for further testing in a Gleeble 3500 thermomechanical simulator equipped with a LUMet system.Table 1.Key compositions of the investigated steel (in wt%).
For austenite grain growth studies, sheet samples were used with a thickness of 3 mm, a width of 10 mm, and a length of 60 mm.The samples were first reheated to 1300 °C where they were held for 10 s as a solution treatment before quenching to room temperature.Subsequently, they were heated to a range of austenitization temperatures between 1000 °C and 1300 °C at a heating rate of 50 °C/s, followed by an isothermal holding where ultrasound waves were acquired by LUMet.
The measurement of austenite recrystallization is based on the decrease in average grain size due to the formation of recrystallized grains from the deformed microstructure.To improve signal quality, cuboid samples with dimensions of 8×8×12 mm were designed for these tests.After reheating to reach an austenite grain size of 60 μm by grain growth, an axisymmetric hot compression test was conducted at 900-1100 °C at a strain rate of 1 s -1 and a strain of 0.2 or 0.4, after which the sample was held at the deformation temperature for another 5-10 min before cooled to room temperature.Laser ultrasonic measurements were started before the deformation and continued until the end of the test to capture the entire recrystallization process.
To combine these two measurements and extend them to also record austenite decomposition from both recrystallized and/or non-recrystallized grain structures, cylindrical rods were used measuring 120 mm in length and 10 mm in diameter, with a gauge length of 10 mm and a reduced cross-section of 6 mm.Flat surfaces were created at the top and bottom of the gauge section for the generation and detection of ultrasound, which further reduces the sample thickness to 5 mm.A compression up to 0.4 strain and a cooling rate up to 30 °C/s can be attained on these rod samples, allowing the investigation of austenite grain growth, recrystallization and subsequent decomposition in a single test.For austenite decomposition, three strain levels (0, 0.2, 0.4) and two cooling rates (10 °C/s, 30 °C/s) were considered.Ultrasonic velocity, V, was measured during continuous cooling where phase transformation takes place as well as during reheating where the velocity of ferrite can be obtained.In a first approximation, the fraction transformed can be calculated with the lever rule: where V  and V  are the velocities in ferrite and austenite, respectively.To obtain the reference velocity for ferrite (or in general the transformation products) a continuous reheating step to 1150 C was conducted after the CCT treatment.Note that depending on the steel chemistry, the velocity of austenite and ferrite can become very close above the Curie temperature (~750°C), which makes it challenging to measure the phase fractions accurately at these higher temperatures [4,7].In all tests, the acquisition frequency of the LUMet system was tailored based on how fast the microstructure changes and how many laser pulses have to be fired.The single-echo technique was employed using the CTOME software to extract the attenuation and velocity of longitudinal waves [8].For grain size analysis, a waveform with negligible grain scattering was selected as the reference.By comparing the first echo of the current waveform to that of the reference waveform in the frequency domain, the frequency dependence of the attenuation was computed and correlated to austenite grain size based on an established calibration [9].Further, the longitudinal velocity was calculated from the travel distance and arrival time of the first echo.With the cross-correlation between the current echo and the reference echo, the precision of time delay estimation is much better than the sampling time of the digitizer, i.e. 0.008 μs.
To validate LUMet measurements, selected double-hit tests were carried out for recrystallization and dilatometry for austenite decomposition.The fraction of softening was determined from the stress-strain curves such that where σ m is the maximum flow stress in the first hit before unloading, and σ y1 and σ y2 are the yield stresses of the first and second hit, respectively, as obtained with the 0.2% offset method.The fraction transformed was calculated from the dilation measurements with a similar lever rule as shown for ultrasound velocity in equation (1).

Austenite grain growth
Figure 1 shows the effect of temperature on isothermal grain growth of austenite.Above 1150 °C, substantial grain growth can be observed with final grain sizes above 60 μm, whereas 1000 °C and 1050 °C produce a small grain size below 10 μm, approaching the measurement limits of laser ultrasonics.This limited grain growth can be attributed to pinning exerted by fine (Ti,Nb)(C,N) precipitates.Test reproducibility has been confirmed based on at least four tests, except for 1100 °C which shows clear variations between the repeats.According to previous studies [10], this discrepancy indicates the occurrence of abnormal grain growth due to the dissolution of fine precipitates that based on Thermo-Calc calculations is predicted to occur at about 1100 °C.When the grain structures deviate from a unimodal distribution, e.g. in abnormal grain growth, the volume fraction of large grains within the detection volume can vary from test to test, leading to associated variations in the LUMet grain size measurement.

Austenite recrystallization
Based on grain growth studies, isothermal holding at 1150 °C for 7 min produces an austenite grain size of 60 μm which was used as the initial grain size for recrystallization studies.Figure 2 presents the evolution of austenite grain size at various temperatures after applying a 0.4 strain.For all tests, the grain size decreases from 60 μm, i.e. the grain size before deformation, with holding time and especially above 950 °C a grain size of 20 μm is achieved within 15 s.With increasing temperature, the recrystallization rates are significantly accelerated such that at temperatures above 1100 °C the change in grain size occurs within one second.Following the recrystallization process, a slight increase in grain size is observed only at 1100 °C where a grain size of ~25 μm is reached within 2 min.It should also be noted that a one-on-one correlation between the average grain size and the fraction of recrystallization needs to be carefully considered due to the sensitivity of ultrasonic attenuation to grain size distribution.This is particularly important for partial recrystallization where a mixture of large deformed grains and small recrystallized grains is present.
For each temperature, measurements were also carried out at 0.2 strain.As shown in figure 3, 0.2 strain is associated, as expected, with a larger recrystallized grain size and slower recrystallization rates as compared to 0.4 strain tests at the same temperature.
The above LUMet measurements were further validated with the fraction softening measured in double-hit tests, which are also shown in figure 2 and figure 3. Note that the 0.2% offset method does not separate the contributions from recovery and recrystallization but, in a first approximation, the initial 20% of softening can be assumed to be attributed to recovery [2].As the holding time increases, an increase in softening is observed, which is consistent with the decrease in average grain size.Moreover, the minimum grain sizes correspond to softening fractions greater than 90%, both indicating the completion of recrystallization.The larger grain sizes observed for deformation at 900 and 950 C are consistent with the fractional softening of 67 and 74 %, respectively, which suggest that only partial recrystallization has taken place during these tests.

Austenite decomposition
The kinetics of austenite decomposition was studied on both deformed and non-deformed austenite.Due to the relatively slow recrystallization, 900 °C was chosen as the deformation temperature to retain workhardened austenite, followed by continuous cooling at a constant cooling rate.Three examples of these deformation transformation tests are shown in figure 4, which correspond to strains of 0, 0.2 and 0.4, respectively.Identical thermomechanical cycles were applied in laser ultrasonic and dilatometry measurement tests for further comparison.The transformation temperatures (e.g. for 50% transformed) obtained from the two techniques agree within the accuracy of measurements (approximately 10 °C) but there are some differences near the end of transformation especially at 0.2 strain.This might be related to the fact that ultrasonic velocity is sensitive to changes in elastic properties including phases and texture, whereas the latter is not picked up by the dilatometer.Further analysis will be required to investigate the synergistic effect of phase transformation and texture change on the velocity curve and how it depends on the strain level.

Measurement of successive microstructural changes
For more complex thermomechanical paths relevant to industrial processing conditions, it is important to examine how well laser ultrasonics can track the evolution of microstructure in a series of 8th International Conference on Recrystallization and Grain Growth Journal of Physics: Conference Series 2635 (2023) 012039 IOP Publishing doi:10.1088/1742-6596/2635/1/0120396 metallurgical processes.A preliminary test was conducted on a rod sample with the following thermomechanical cycle including three test stages after the solution treatment, as shown in figure 5(a): (i) grain growth at 1150 °C, (ii) deformation with a strain of 0.2 and recrystallization at 1000 °C, and (iii) phase transformation during continuous cooling at 30 °C/s from the recrystallized austenite.The subsequent final reheat step is as aforementioned designed to provide the reference velocity for ferrite.The generation laser was turned on and off to measure the key features of each stage while limiting the surface damage caused by laser pulses.As shown in figure 5(b), the grain size before deformation is around 60 μm, agreeing well with previous grain growth tests.After deformation, the recrystallization kinetics and minimum grain size are similar to those obtained in the separate test with the cuboid sample, which also verifies that laser ultrasonic measurements are independent of sample geometry.Another advantage of the rod sample is the higher amplitude for the first echo due to the reduced thickness resulting in less scatter in grain size analysis compared to the cuboid samples.Figure 5(c) presents the transformation kinetics of the fully recrystallized austenite with a grain size of 30 μm, at 30 °C/s.The temperature of 50% transformed is approximately 675 °C, which is, as expected, between the values obtained by dilatometry for the same cooling rate and austenite grain sizes of 10 μm (690 °C) and 60 μm (630 °C), respectively.

Conclusion
In this work, laser ultrasonics was used to investigate the behavior of austenite grain growth, recrystallization, and decomposition for a Nb-microalloyed steel at a range of temperatures and deformations relevant for hot rolling.For grain growth measurements, test reproducibility is a useful indicator to identify both normal and abnormal grain growth stages.In addition, the effect of temperature and deformation on recrystallization was examined and found to be in agreement with the fraction of softening obtained in double-hit tests.By adopting the faceted rod samples, an improved experimental design was achieved to also measure with laser ultrasonics austenite decomposition in combination with grain growth and recrystallization.In essence these three microstructure features can be recorded in a single test for a given reheat, deformation and cooling condition.These studies demonstrate the efficiency of laser ultrasonics to record the overall microstructure evolution at various thermomechanical conditions, and its potential in measuring a sequence of microstructural changes that can expedite the evaluation of thermomechanical processing routes as well as the development of process models for hot rolling.

Figure 5 .
Figure 5. (a) The thermomechanical cycle tested.(b) The grain size change before and after deformation at 1000 °C with 0.2 strain measured on different samples.Time zero corresponds to 8 s before the deformation.(c) Austenite decomposition kinetics of recrystallized austenite at 30 °C/s.

8th
International Conference on Recrystallization and Grain Growth Journal of Physics: Conference Series 2635 (2023) 012039