WAND²—A versatile wide angle neutron powder/single crystal diffractometer

The Wide Angle Neutron Diffractometer Squared (WAND²) instrument is the latest iteration of a neutron diffractometer with a position sensitive detector in the thermal beam room at the High-Flux-Isotope Reactor (HFIR). The original WAND (Wide Angle Neutron Diffractometer) instrument was built jointly by the Oak Ridge National Laboratory (ORNL) and the Japan Atomic Energy Agency (JAEA) in the 1980s. Over the years, the instrument has evolved during different upgrade programs. The last upgrade program has been finished in the beginning of 2018, and the instrument, now called WAND², is looking forward to becoming one of the most significant instruments in the HFIR beam room. An overview of the history of WAND and its latest upgrade is given in Ref. 1. WAND² is located on the radial thermal horizontal beam tube HB-2 on the HB-2C position. The high thermal neutron flux of HFIR enables WAND² to perform experiments on small samples and/or short time scales. The 2-dimensional position sensitive area detector with high efficiency and nearly 2 × 10⁶ pixels over 120° horizontally and 15° vertically allows the measurement of a large portion of reciprocal space for single crystal measurements and the collection of a larger fraction of the Debye-Scherrer cone for powder measurements. The wide selection of the sample environment, the open access to the sample area, and the event mode data collection allow the realization of many experimental conditions, making WAND² an instrument well suited for non-standard experiments.

WAND² is one of the four instruments on the horizontal beam port number 2 at HFIR. The technical layout of the instrument is shown in Fig. 1. The HB-2 beam tube is radial from the core with a broad flux spectrum. Cooled sapphire filters remove the high energy neutrons. WAND² has a hot-pressed Ge-monochromator optimized for the 113 reflection with 15 pieces in a venetian blind arrangement (mirrored at the center) to focus the beam vertically. The fixed take-off angle is 51.5° which yields a wavelength of λ = 1.482 Å. Due to the diamond-type structure of Ge, the beam does not contain any higher order contamination. An optional additional wavelength is 0.946 Å from the 115 reflection. After the shutter, a beam monitor allows normalization on the incident neutron flux. The beam monitor sits on an optical rail which extends to the sample table. The optical rail allows us to mount beam shaping apertures, collimators, and a ³He spin filter for half-polarized measurements. Close to the sample is a motorized compact slit package from JJXray. The sample table has sample rotation, x-y translation, and goniometer stages. Maximum load capacity on the standard goniometer stage is 200 kg which is the weight of HFIRs 5T cryomagnet with cryogenic liquids. If a heavier sample environment is required, the goniometer stage can be removed and a maximum of 340 kg can be loaded on the rotation stage. The instrument has been spaced to accommodate the large 10T cryomagnet with a diameter of 600 mm. The standard goniometer stage allows ±10 mm movement on the x-y table. Goniometer tilts are restricted by the sample environment. A small κ-ϕ for texture measurements is available.

Between the sample and detector is a fine oscillating radial collimator with a uniform collimator angle of 1.22°. The collimator angle has been chosen to achieve a cut-off diameter of 45 mm for 2θ = 38.68°, the (111) reflection from Al. 45 mm is the can diameter of the ultra-low ³He/⁴He dilution fridge insert used at HFIR. Both the collimator and detector are on a common detector table. The table can be moved in 2θ by 30° and in height by 100 mm. This allows us to maximize the coverage of the detector. The detector is a 2-dimensional position sensitive (2D-PSD) Brookhaven National Laboratory 120-degree detector (BNL120) originally built for the Los Alamos National Laboratory.^2,3 Currently, there exists one identical detector on the WOMBAT instrument⁴ at the Australian Nuclear Science and Technology Organisation (ANSTO). The coverage of the detector is 120° in 2θ and 15° in μ vertically. The detector is built out of 8 identical, seamless stitched segments. On each segment, the X-axis cathode consists of 120 50 μm wires with 1/16 in. pitch. The Y-cathode consists of 129 Cu-strips on a 1/16 in. Rogers 4000 laminate board. Over the whole detector, this sums up to 123 840 points. However, the sub-wireresolution of 0.4 mm or 0.031 25° both horizontally (3840 pixels) and vertically (512 pixels) is achieved by using digital centroid finding electronics.¹⁰ The detector and collimator are shielded by pieces of epoxy resin mixed with B₄C. This type of shielding has proven very successful and is optimized for the operation of the instrument in the HFIR beam room. The high flux combined with the low background enables either very fast data acquisition or the measurement of very small samples and very weak signals.

For the instrument control, WAND² uses the Experimental Physics and Industrial Control System (EPICS) with Control System Studio (CSS) as a graphical user interface. This approach follows the strategy of the Neutron Sciences Directorate to convert all neutron scattering instruments to a common user interface. It allows users to transport experience with the user interface from one instrument to another within ORNL and allows instrument scientists more easily to fulfill the role of local contact at different instruments. Similarly, the data reduction and analysis uses the MANTID package^5,6 as a facility-wide standard.

WAND² is a high-flux medium resolution powder diffractometer with a large area detector. It is especially suited to investigate phase transitions and magnetic structures under extreme conditions. The resolution limits structure refinements to higher symmetric lattices, but the good signal to noise ratio allows us to measure also small samples and detect weak signals, for instance, from ions with a small magnetic moment. A powder pattern from Si is shown in Fig. 2. The 2D pattern can be individually masked and/or binned along the Debye Scherrer cones to analyze texture in materials. For a standard powder sample, the whole 2D pattern is normalized, efficiency corrected, and integrated into a 1D-plot which can be refined with a program like FullProf.⁷ The normalization and error calculation uses the method presented in Ref. 11. The refinement of the 2*20 min plot yields a conventional R-factor R_wp = 10.4 and a Bragg R-factor for the pattern of 2.45. The refinement of the 1s collection yields R_wp = 29.2 and a Bragg R-factor for the pattern of 11.6.

For single crystal diffraction, the sample is preferably aligned within a scattering plane although data can be taken first and the alignment can be calculated from the data afterwards. This might be, for instance, a strategy when a single crystal has moved in a pressure cell under load application. In the following, an aligned crystal in the HK0 geometry is considered as an example. The alignment process can be used to determine the UB-Matrix which is saved in the data files. The UB Matrix can be either refined or newly determined afterwards again. The sample is rotated around the L-axis in steps of typically 0.1°. Depending on the symmetry of the lattice, a reduced angle can be sufficient, but normally 180° are covered. The data are then normalized and combined in MANTID to plot the slice views of the reciprocal space. One example of these slice views is given in Fig. 3. The sample is a rhombohedral Y-hexaferrite BaSrCo₂Fe₁₁AlO₂₂, oriented in the HK0 geometry. The top left slice shows the HK0 plane with a slice thickness in L of ∼ 0.14 r.l.u. As can be seen in the perpendicular slices H-HL (bottom left) and HHL (bottom right), magnetic reflections can be found on non-integer L positions. At 200 K, BaSrCo₂Fe₁₁AlO₂₂ has two magnetic phases: one with an incommensurate propagation vector and one with a commensurate propagation vector.⁸ The strength of a 2D-PSD is obvious in the perpendicular plots (lower panel), where the out-of-plane coverage allows the measurement of a significant large enough portion of reciprocal space for an analysis without the need to re-orient the sample. In this particular case, the c-lattice constant is very large (43.3 Å) which yields coverage in L of ±3.8 r.l.u. In systems with smaller (perpendicular) lattice constants, the detector can be vertically translated by 100 mm to use the full 15° coverage. In practice, due to the vertical halfwidth of the reflections, the useful coverage is up to μ = 12°. The relation sin μ = λ/c can be used to estimate the out-of-plane coverage. Using the 12° and λ = 1.486 Å yields a lattice constant of 5.74 Å, which means that for a hypothetical structure with a perpendicular lattice constant of this size, both HK0 and HK1 (and everything in between) can be measured in one orientation on WAND².

WAND² collects data in event mode, which means that each neutron is individually assigned a 2D position and time stamped. The setup allows us to synchronize neutron events with two time domains, as depicted in Fig. 4. The events from the detector go through the original BNL electronics which put out an x-y encoded event. Detector events and fast metadata are streamed through an optical connection into the “Concentration Card” developed at ORNL where data are combined and time stamped. From there, the data are sent through one optical connection to the data acquisition system. The time stamping can use “time” from two sources. One source can be directly connected to the concentration card via the Optical Distribution Board (ODB). For instance, an electric pulse generator is used to apply a periodic electric field on a sample. The pulses are also fed into the ODB and are used to synchronize the neutron events with the electric pulse. Using this method, a time stamping accuracy of 100 ns can be achieved. Alternatively, a time signal is sent from a server giving an absolute time. This time can be recorded with the sample environment and allows synchronizing of the event and parameter with an accuracy of a few ms. The event mode allows us to filter the data after the measurement. Typically, a measurement is started and then an external parameter is varied. This allows us to quickly determine transition temperatures (or other outer parameter), and in many cases, data can be binned above and below the transition, saving an additional measurement. This method also has the additional bonus that the intrinsic of the sample can be studied, for instance, an ongoing change in the lattice parameters after the thermometer shows a stable temperature indicates poor thermal contact to the sample, or as another example, the critical magnetic field is higher than that previously determined from laboratory measurement, due to the different demagnetization fields of the different samples. One example of the temperature filtering is shown in Fig. 5. At one sample position, several magnetic peaks can be measured in the detector. The temperature was ramped from 100 K to 500 K with 0.3 K/min, and data were collected. Afterwards, temperature filtering was applied in 10 K bins. Individual regions of interest (ROIs) corresponding to individual peaks have then been plotted. Up to 320 K, little variation is seen in both peaks, but above that, one magnetic phase vanishes, while the spectral weight is transferred to another phase which also undergoes a transition into a different magnetic structure with a different propagation vector.

The new WAND² instrument at the HB-2C beamport at HFIR combines a high flux with a high-resolution, high efficiency 2D-PSD. This allows fast measurements to the study kinetics of phase transitions or the measurement of very small signals. The wide choice in the sample environment and the open access to the sample area make WAND² a preferable choice for novel techniques in, for instance, the study of lipid bilayers.⁹ Overall, WAND² is comparable to other world class diffractometers using a 2D-PSD and is a unique instrument in the diffraction suite in North America.

This material is based on the work supported by the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences, under Contract No. DE-AC05-00OR22725. This research used resources at the High Flux Isotope Reactor, a DOE Office of Science User Facility operated by the Oak Ridge National Laboratory. WAND has been and WAND² is operated jointly by ORNL and the Japan Atomic Energy Agency under the US-Japan Cooperative Program on Neutron Scattering. Taro Nakajima and Kazuhisa Kakurai advised on the BaSrCo₂Fe₁₁AlO₂₂ single crystal measurement during commissioning.