In situ neutron scattering is a powerful tool to reveal materials atomic structural response such as phase transformation, lattice straining, and texture under external stimuli. The advent of a high flux neutron source such as the Spallation Neutron Source (SNS) allows fast measurement in even non-equilibrium conditions, i.e., phase transformation in steels. However, the commercial fast heating apparatus such as commercial physical simulation equipment is not designed for in situ neutron scattering, which limits its application to in situ materials research by using neutrons. Here we present a resistive heating gas enclosure loadframe (RHEGAL) for non-equilibrium phase transformation studies by using in situ neutron scattering, which takes the advantage of high flux neutron sources like SNS. RHEGAL enables fast resistive heating of metal samples to 1200 °C at a rate up to 60 °C/s in an inert atmosphere. It provides both horizontal and vertical positions for scattering optimization. The mechanical loading capability also allows in situ high temperature tension above the oxidation temperature limit. The optimized translucent neutron scattering window by silicon allows both reflection and transmission measurements, making this equipment applicable for neutron diffraction, small angle scattering, and imaging. To demonstrate the fast heating capability, the phase transformations of an example of advanced high strength steel heated at 3 °C/s and 30 °C/s were measured with the VULCAN engineering diffractometer, and the different phase transformation kinetics by neutron diffraction were presented.
The advanced neutron sources such as the Spallation Neutron Source (SNS) provide unprecedented capabilities including high flux neutrons, novel neutron detector technologies, and time event neutron data acquisitions (DAQs)1 for enabling time resolved studies in materials.2,3 The in situ explorations by neutrons enable rapid scientific findings in comparison with ex situ characterizations, thus making in situ neutron scattering a powerful tool for materials research. In situ neutron scatterings are now widely used in engineering material characterizations such as phase transformation, precipitations, intergranular lattice strain, twining, or texture under individual or combined external stimuli such as temperature, pressure, force, and electrical field.4–21 Most studies are performed when materials are under equilibrium or quasi-static conditions, and our understanding of materials under non-equilibrium condition is lacking. For example, the widely used advanced high strength steels in automotive industry are carefully designed by controlling the thermomechanical process which involves fast heating and cooling, which is far from equilibrium.11,12 In laboratories, a commercial made thermal-mechanical simulator is used to simulate material thermal and mechanical processes during a manufacturing process, and the macroscopic response of stress, dilatometry, etc., during heat treatment can be extracted for materials study and new materials development. Neutron beamlines often can host large sample environments, but the commercial simulators are not optimized for neutron scattering due to the large dimensions which make them not possible to be embedded directly to a neutron instrument. In this paper, an in-house device, resistive heating gas enclosure loadframe (RHEGAL), is presented to enable in situ neutron scattering while materials are under non-equilibrium process. The overall design, resistive heating capability, rapid data acquisition, and performance by measuring phase transformation of an advanced high strength steel (AHSS) under fast heating up to 30 °C/s are presented.
RHEGAL enables experimenters to simultaneously apply variable heating and loading stimuli in a controlled gas atmosphere on a neutron scattering instrument while preventing sample oxidation at high temperatures. Uniform resistive heating can be rapidly (up to 60 °C/s) applied on conductive materials up to 1200 °C by using a high DC power supply. The mechanical load is applied up to 4450 N (1000 lbf) in tension or compression through a mechanical gear load path by a controlled stepper drive. The heating and loading chains are contained inside a gas enclosure made of Al6061 aluminum alloys with neutron friendly quartz windows. RHEGAL can be mounted in either vertical or horizontal orientation for different types of instruments or experimental configurations. When a mechanical load is required, it could be placed horizontally with the loading direction in the bisector direction of the angle between the incident beam and 90° diffraction bank. This allows an instrument like VULCAN6 to measure different directional strain components from its different detector banks. For powder diffraction measurement, small angle neutron scattering (SANS), or transmission imaging, RHEGAL could be placed in vertical configuration, where it could rotate around samples’ longitudinal direction for different measurement purposes, such as with the sample normal along the incident beam for SANS or imaging measurement.
A. Overall design
RHEGAL’s main design requirements are rapid sample heating, thermomechanical loading, controlled atmosphere to avoid oxidation, synchronized grip movements to maintain static centroid of the sample scattering volume, a neutron friendly scattering window with least absorption, and being portable to different instrument suites. To satisfy the requirements, the main concept adopted here is by designing a fast-resistive heating device with loading capability inside a controlled atmosphere gas enclosure. In Fig. 1, RHEGAL is shown as an airtight aluminum box with quartz windows which have least neutron absorption. RHEGAL is about 813 mm long, 559 mm wide, and 160 mm thick and weighs 146 kgs. An access door with the quartz window is locked airtight by the top two quick turntable locking latches with a rubber sealing strip between the door and the enclosure box. Thanks to its compact size, RHEGAL is portable and can be positioned either horizontally or vertically for different experiments. For details, main design aspects are described below, which are mechanical chain design, resistive fast heating by pulsed DC current, heating control and fast DAQ, and gas enclosure for controlled atmosphere.
B. Mechanical drive system
To accommodate different sample lengths and with additional mechanical loading capabilities, a mechanical drive system is designed, as shown in Fig. 2(a). The load/displacement is generated by a stepper motor (Animatics SmartMotor SM2316D-PLS2-ETH) which drives a planetary reduction unit. The planetary reduction unit drives a set of worm gears which rotate a set of acme screws; thus, it has a combined ratio of 600. The acme screws load a set of acme nuts which generate the load on a set of steel beams. The steel beam on the left in Fig. 2(a) transfers the load through a loadcell (InterfaceForce 1100) and into the clamp plate. The clamp plate transfers the load through a set of ceramic standoffs or screws depending on whether it is a compression or tension load which then goes through the left sample clamp/grip to the sample. The load path is thus as follows: the stepper motor, the planetary reduction unit, the worm gear, the acme screws/nuts, the left steel beam, the loadcell, the clamp plate, the ceramic standoffs, and the left sample clamp. The steel beam on the right transfers the load directly through the ceramic standoffs and then through the right sample clamp/grip to the sample. It is noted that both beams move simultaneously against each other, thanks to this design, which allows static centroid of the sample scattering volume. A 4450 N (1000 lbf) load could be supplied by the beams. When the clamp is in tension, the load path is from the beam through the 10-24 UNC-2A screws into the load cell through the 10 M × 1.5 screws into the mounting plate and through the 1/4 20-UNC-2A screws and into the clamp, as seen in Fig. 2(b). While the loadcell measures the mechanical force, the mechanical displacements are calculated based on the stepper motor revolution by its encoder, gear thread pitches, and the corresponding gear reduction ratio. The local sample strain/displacement can be measured by attaching a commercial extensometer to the sample. The transducer signals from the loadcell and the extensometer can be recorded by a National Instruments CompactRio system with strain/bridge input modules. Because the mechanical loading/displacement is via a large gear reduction ratio, the mechanical drive system is not designed for dynamic or fast (cyclic) loading.
C. Gas enclosure for controlled atmosphere
The RHEGAL gas enclosure will be under slight positive pressure with full flow of the nitrogen or argon gases via a 0.66 mm diameter orifice to prevent the sample oxidation at a high temperature. The supply of regulated atmosphere can be provided by a regulated plant N2 line, and the flow is controlled with a flowmeter. After the sample being mounted, the internal air can be purged out of the enclosure via an exhaust valve. An oxygen sensor is used to monitor the oxygen level reaching a minimum before heating. The stress analysis under pressure has been performed on the RHEGAL gas enclosure by finite element analysis. The stress distribution is shown in Fig. 3. With 0.014 MPa (2 psi) internal pressure, the maximum tensile stress (9.94 MPa) is concentrated at the middle of the quartz windows. The quartz windows are designed to be 6.4 mm thick to hold 0.034 MPa (5 psi). Meanwhile, overpressure of RHEGAL is protected by a dedicated pressure relief valve set at 0.014 MPa (2 psi). It will prevent personnel working at the RHEGAL enclosure from harm when the RHEGAL exhaust valve is closed or the back-pressure regulator fails to shut. When changing the sample, the internal pressure is relieved by opening the exhaust valve, and then the access door can be unlocked by turning the two locking latches for a quick sample change.
D. Resistive heating by pulsed DC current
Resistive heating could provide uniform heating to the sample volume rapidly in comparison with other heating such as conduction heating which heats the sample from surface. For high temperatures over the oxidation point, it provides a relatively easier setup to create a compact gas enclosure. Furthermore, it can be conveniently coupled with a mechanical load as for this case. As shown in Fig. 2, the ceramic standoffs between the clamps/grips and the mounting plate provide the electrical insulation. Samples are mounted on the flatten grips with set screws to provide good electrical contact. A close view of the conductive grips with “blue” water cooling lines is shown in Fig. 2(b). The grips are made of copper with high conductivity. The cooling water is supplied by using a water chiller. Resistive heating to the sample is provided by passing pulsed high DC current with low voltages. The pulsed DC current is supplied by an Ametek 20 V/500 A DC power supply and limited to 300 A due to circuit feed. The reason the pulsed mode DC heating is chosen is to avoid artificial temperature DAQ via welded thermocouples (TCs). The local current field near thermocouples would generate low voltages between the thermocouple (TC) electrodes; thus, it will create a deviation of the temperature readings. To acquire accurately the change in millivolt potentials only due to sample temperatures, the TC DAQ should only be conducted when its current is off duty. Thus, the pulsed mode DC heating is required. The details of how the DAQ is implemented are shown in Sec. II E.
E. LabView control and fast DAQ
The main control and DAQ of RHEGAL includes movement of the grips and loading transducers (displacement and force) and resistive heating (power supply control and temperature readings). The control and DAQ of grip movement and the loading transducers are adopted from in-house LabView codes developed for a mechanical loadframe. The details on the LabView loadframe control and DAQ can be found in Refs. 22 and 23. The resistive heating control and DAQ panel shown in Fig. 4 was newly developed in LabView to allow aforementioned pulsed high DC current heating while taking temperature data when the heating is off duty. The user operation level with a clean interface is shown in Fig. 4(a). The left plot shows the target temperature (overlapped with the control temperature) in a white solid line and the feedback of the real power output in a red line as scaled to fit in the plot. The small plot window shows the selected TC temperature. Figure 4(b) shows the detailed control level with RHEGAL heating and DAQ settings. The control flow and communication diagram is shown in Fig. 5. The National Instruments CompactRio with a field-programmable gate array (FPGA) module provides a rapid closed loop control by either heating rate or heating power control to fast heating. In heating rate control, the target temperature setpoint and heating rate are set in the user interface, and the dynamic pulsed power outputs are adjusted by reading the temperature from control TC while maintaining the desired heating rate. The power supply is capable of tunable control rates from 3 to ∼1000 Hz at harmonics of the neutron beam pulse rate (60 Hz nominal). A short off duty can be synchronized with the beam for the quick DAQ. For example, the neutron pulses operate at 60 Hz (maximum neutron pulse time width at the sample position is then <16.7 ms), and we can set up the short off duty time of 1 ms in between each pulse to allow acquiring temperature data. Given high sampling rates (75 S/s) from CompactRio, more temperature data points can be measured to obtain a running average in a short time window. When the instrument choppers run at different frequencies, corresponding adjustment can be made through the control panel to ensure acquiring TC readings, while neutron pulses are not striking samples. The heater power supply voltage and current outputs are acquired as well with the TC readings and saved with UTC timestamp for post-synchronization with time-of-flight neutron data whose timestamp is also registered with respect to UTC. An over-temperature TC limit is set for monitoring overheating of the sample; when it happens, the CompactRio triggers the interlock box to power down the power supply.
III. FAST HEATING TEST AND PERFORMANCE
We presented a practical use of RHEGAL in a neutron beam by conducting an in situ neutron diffraction study of non-equilibrium phase transformation of an advanced high strength steel (AHSS). Part of the scientific findings of this work have been published elsewhere11,12 in comparing two grades of AHSS, and here we refer one case for the demonstration of RHEGAL. The sample shown here is a dual phase material (DP980) with ferrite (bcc) and martensite (bct) phases. The details of the materials can be found in the published work.11 The in situ fast heating experiment was conducted by using the VULCAN instrument, an engineering materials diffractometer specialized for structural materials research under external stimuli. The details of the instrument layout can be found in Ref. 6. Figure 6(a) shows that RHEGAL was mounted on the VULCAN instrument vertically to maximize the material in the neutron beam during phase transformation measurement. As shown in Fig. 6(b), the sample’s longitudinal direction is along the vertical direction, and RHEGAL was rotated to align the sample’s transverse direction to be 45° off the incident beam, which allows diffracted neutrons to exit through the quartz windows to both the ±90° detector banks of the instrument. The sample volume was defined by the incident beam opening of 17 mm (H) × 5 mm (W) and steel strip’s thickness of 2 mm, resulting in ∼240 mm3 (1.4 × 5 mm × 2 mm × 17 mm) sample volume in the beam, as also shown in the schematic in Fig. 6(b). The VULCAN instrument was at 30 Hz chopper spinning setting with a measurable lattice spacing from 0.5 to 2.4 Å. The high intensity beam mode was chosen to maximize the beam flux at samples.
Different sample designs are considered for achieving uniform temperature distribution inside the scattering volume at different heating rates because at a relative slow heating rate, due to the heat loss at the gripped areas, a large temperature gradient may be expected in the sample scattering volume. Here, for example, two sample designs were adopted for different heating rates, as shown in Fig. 7(a). The flat straight sample design of 2 × 8 mm2 steel strip provides relative uniform heating in the scattering volume, thanks to the high heating rate. However, for the slow heating, the small notches (1.5 mm deep and 8 mm wide) have to be cut at ∼56 mm from the scattering center, which reduces the heat loss by compensating heat from higher local resistance at the notches as well as smaller heat conduction to the water-cooled grips. Each of the designs needs to allow less than 3 °C temperature variations inside the neutron gauge volume during heating based on the measurable lattice resolution by diffraction (for iron, 3 °C will induce ∼36 μm strains due to thermal expansion, which is less than 50 μm strains, a typical error from diffraction measurement). Multiple TCs were welded to the samples to monitor temperature gradients. A snapshot when a sample was above 1000 °C during heating to 1050 °C at 30 °C/s rate is shown in Fig. 7(b). As shown as an example in Fig. 8, the heating data during heating and cooling of different TCs are presented. A zoomed-in view in 3 s shows the differences in Fig. 8(b), and the temperature differences during heating between the central TC and the other offset TCs (10 mm and 15 mm apart from the central TC) are negligible. Only 2 °C variations were found within ±10 mm from the central TC which satisfies <3 °C within total 17 mm height gauge volume. For the 3 °C/s heating rate, similar performance was achieved by introducing the notched design.
During fast heating, it becomes challenging to obtain an analyzable diffraction pattern at conventional neutron sources due to low neutron flux. We demonstrated in Ref. 11 that 1 s pattern is sufficient for heating rate at 3 °C/s by one single sample heating run. A high temperature austenite phase transformation during heating and its transformation back to the ferrite/martensite phase during cooling were observed. The 1 s pattern averaged over 3 °C is statistically enough for Rietveld full pattern refinement to reveal the lattice d spacing and phase volume fraction change over time.11 However, when being heated at 30 °C/s, to achieve averaged data over 3 °C, 0.1 s data slice has to be made, and from a single sample run, it will not provide enough neutron statistics. To obtain similar statistics of 1 s data at 3 °C/s and avoid non-repeatable phase transformation dynamics due to the change in microstructures, a stroboscopic approach24 was adopted by repeating heating of original material samples under the same conditions over 10 times. The time event neutron data were then sliced at 0.1 s time interval, and all 0.1 s neutron data with the same temperature range from the 10-sample runs were synchronized by the temperatures recorded by the fast acquisitions and summed in the VDRIVE software.25 The neutron diffraction contour plot of the 0.1 s data binning is demonstrated in Fig. 9, which clearly shows phase transformations with reducing austenite phase and rising ferrite/martensite phases. The synchronized data in detector bank 1 were used to conduct Rietveld refinement in GSAS software.26,27 The phase transformation volume fraction and lattice parameter changes in the DP980 alloys heated at 3 °C/s and 30 °C/s were shown in Figure 4 of Ref. 11. The heating rate effect is reflected by the shift of volume fraction over higher and broader temperature ranges. The heating rates also show effects on the lattice transition temperature with shifting to a higher temperature regime at the higher rate. The details of the studied alloy and other alloys far from equilibrium can be found in Refs. 11 and 12.
The developed RHEGAL device allows non-equilibrium study of materials under rapid heating to high temperatures by taking the advantages of the high flux and time event data acquisition at SNS. The portable design makes RHEGAL available for different neutron scattering instruments for physical phenomena observations at different length scales by using diffraction, small angle scattering, and imaging. Time resolved measurements including stroboscopic data acquisition and reduction provide high temporal resolution to reveal the materials behavior under the rapid stimuli. The overall advancing neutron instrumentation including a high flux neutron source, sample environment, data acquisition, and reduction makes neutrons to be a powerful tool in materials research in both fundamental and applied research.
This work was supported by the Spallation Neutron Source (SNS), Oak Ridge National Laboratory (ORNL), supported by the U.S. Department of Energy, Basic Energy Sciences, Scientific User Facilities Division. RHEGAL hardware and data reduction were supported by the Laboratory Directed Research and Development projects of ORNL.
This manuscript has been authored by UT-Battelle, LLC under Contract No. DE-AC05-00OR22725 with the U.S. Department of Energy. The United States Government retains and the publisher, by accepting the article for publication, acknowledges that the United States Government retains a non-exclusive, paid-up, irrevocable, worldwide license to publish or reproduce the published form of this manuscript, or allow others to do so, for United States Government purposes. The Department of Energy will provide public access to these results of federally sponsored research in accordance with the DOE Public Access Plan (http://energy.gov/downloads/doe-public-access-plan).