WISH A high-resolution magnetic diffractometer for TS-II

ISIS Second Target Station Project

ISISTGT2/SAC/02/P4_1 WISH

Beamline Name: WISH ISIS Contact: Prof. Paolo G. Radaelli & Dr. Laurent C. Chapon ISIS Facility, Bldg. R3

External Coordinator: Prof Andrew Harrison School of Chemistry The University of Edinburgh Joseph Black Building, The King's Buildings, West Mains Road Edinburgh EH9 3JJ Tel: 0131 650 4745 Fax: 0131 650 4743 e-mail: [email protected]

Rutherford Appleton Laboratory, Chilton Didcot OX11 0QX Tel: 01235-445685 Fax: 01235-445642 e-mail: [email protected], [email protected]

INSTRUMENT OVERVIEW Moderator

Solid Methane, broad side

Incident Wavelengths

1.5 – 15 Å

Single-frame bandwidth

8Å

d-spacing range

0.7 – 50 Å

L1

50m

L2

1-2.5 m

Flight path

Ballistic guide+ guide carousel for focusing options

Choppers

2 disc choppers (50-10Hz)

Detectors Beam size

ZnS Scintillator detectors covering all scattering angles between 10º and 175º. 20 mm x 40mm (unfocussed) to 1 mm x 1mm (super-focussed)

Optimal frequency

10 Hz

Sample/detector tank

Radial Collimator, 2m diameter vacuum tank

Sample environment

All standard equipment + dedicated 15 T cryomagnet

p. 2

ISIS Second Target Station Project

ISISTGT2/SAC/02/P4_1 WISH

p. 3

Science Case: Introduction

Intensity (arb. un.)

High-resolution cold-neutron powder diffraction excels when multiple nearly overlapping Bragg peaks occur at long dspacing. In this case, shear flux is not sufficient to extract all the available information, and much better results can be obtained with a high-resolution diffractometer, even at the cost of losing some flux. The case in point is 0.8 Gem: 80 m illustrated in Fig. 1, where we plot a series of TbBaMn2O6, T=10 K OSIRIS: 3 h diffraction patterns from a sample with a complex 0.6 magnetic structure, collected for comparable Bank 1 amounts of time on several GEM detector banks and on OSIRIS. Clearly, the integrated intensity of the Bank 2 0.4 Bragg peaks on Bank 2 on GEM (2θ~20°), for 1/5 Bank 3 example, is greater than for the OSIRIS back0.2 scattering data, but the latter are much richer in Bank 4 information content due to the higher resolution. Osiris Another noteworthy feature is that the higher 0 resolution enables weak Bragg peaks (e.g., at 5.1 and 5.6 Å) to emerge from the background. The 3 4 5 6 7 8 resolution of GEM and OSIRIS are comparable in d-spacing (Å) backscattering, but the lack of sufficient cold-neutron Fig. 1: Comparison between GEM and OSIRIS data on a flux from the TS-I liquid-methane (L-CH4) complex magnetic structure. The data were collected on a moderator prevents one from employing the GEM TbBaMn2O6 sample at low temperature for 80 minutes and 3 backscattering bank in this range of d-spacing. hours on Gem and OSIRIS, respectively. The example from Fig. 1 already suggests that complex magnetic systems are the prime scientific target for high-resolution cold-neutron powder diffractometry. In complex systems, the magnetic structure has a large number of degrees of freedom (typically the three components of the magnetic moment on several inequivalent atoms), and the d-spacing range available to observe Bragg peaks is limited due to fall-off of the magnetic form factor. In addition, magnetic and nuclear Bragg peaks are often nearly overlapping, especially in the case of large incommensurate magnetic structures. When the magnetic structure is relatively simple but a crystallographic pseudo-symmetry is present, powder averaging of the magnetic structure factor for quasi-degenerate reflections may prevent the determination of the direction of the magnetic moments, and high-resolution data are required to solve the structure. There is an enormous interest to study a variety of systems with these general features. Very often, these compounds are novel phases or do not grow as large crystals, and only powder samples are available. Clearly, it is imperative to offer to the user community the capabilities of studying multi-dimensional phase diagrams as a function of temperature, magnetic field, external pressure, etc. Weak magnetism is another topical subject with demanding technical requirements. Measuring weak Bragg peaks requires high intensity, high resolution (to increase the ratio between signal and intrinsic background) and extremely low extrinsic background. We believe that the quality and quantity of diverse applications for such an instrument in the “conventional” powder diffraction mode are sufficient to justify its construction. Nevertheless, the addition of single-crystal and polarised-beam capabilities, if they came at no expense for the powder diffraction programme, would add a range of truly exciting science to an already excellent portfolio. This option would complement existing and future single-crystal capabilities at ISIS (SXD, PRISMA, HR-SXD LMX), particularly in the domain of high magnetic field diffraction. Scientific requirements for WISH A series of scientific specifications, which are ideally matched to the performance of ISIS-TS-II, naturally emerge from the previous considerations. We propose to construct a state-of-the-art, long-wavelength diffractometer, called WISH†, primarily designed for powder diffraction at long d-spacing on magnetic and large-unit-cell systems, with the option of enabling single-crystal and polarised beam experiments. Instruments such as IRIS and OSIRIS have clearly demonstrated the value and potential of high resolution for magnetic powder diffraction. WISH will capitalise and expand on what has †

WISH is an acronym for “Wide angle In a Single Histogram”. WISH evolved out of a backscattering machine with an extended-angle detector, which could be focussed in a single data set. Lower-angle detectors were added later on to match a wider science portfolio, but a large part of the WISH detector can still be focussed “in a single histogram”.

ISIS Second Target Station Project

ISISTGT2/SAC/02/P4_1 WISH

p. 4

already been achieved, and will make this technique flourish. However, the d-spacing range in backscattering is not sufficient to tackle the largest magnetic structures, so lower angle detectors are needed. Also, background issues need to be carefully considered to work with small moments and small samples - issues that are becoming increasingly important in the emerging field of molecular magnetic systems. Two main considerations were employed to draft the scientific and technical specifications: 1.

WISH must provide a uniquely valuable contribution to the worldwide landscape of magnetism research. To insure this, we have enlisted five leading expert in magnetic neutron diffraction and instrumentation (see list attached) to act as an International Advisory Group for WISH. Because of the moderator characteristics and the low repetition rate of TS-II, WISH would be comparable but not significantly better than the best reactor-based instrument in terms of shear flux on sample. Rather, WISH will be world-leading and unchallenged in the high-resolution (backscattering) end of the long-d-spacing “market”, which is significantly under exploited due to lack of suitable instrumentation.

2.

WISH must cater for the existing needs of the UK community, whilst, at the same time, fostering the development of UK research in magnetism, both in breadth and in depth. In particular, the great majority of UK users studying magnetic system are interested in refining the magnetic and crystal structures simultaneously, and WISH must offer that capability. Great care should be taken in obtaining a manageable resolution function and having it implemented in popular Rietveld codes such as FULLPROF and GSAS. Clearly, there is a need to have advanced sample environment, such as high-field magnets, dilution cryostats and pressure cells integrated in the instrument design from the outset.

WISH science headlines (individual contribution to the scientific case are listed in Appendix I) a.

Ionic transition-metal compounds

These highly topical compounds are typically Mott-Hubbard or small band-gap insulators, which can be driven into a metallic phase either by doping or by an external parameter such as pressure, “chemical” pressure or a magnetic field. The signature of the excitation spectrum of the insulator persists in the metallic phase, giving rise to interesting physics, including superconductivity and colossal magnetoresistance. In all these classes of material, magnetism provides an important dimension to the physical properties, alongside other electronic degrees of freedom such as charges and orbitals. As the complexity of these materials increases, so do the associated magnetic structures. Even when single crystals are available, there is a great interest in studying multi-dimensional phase diagrams, making powder studies essential. The UK has world-leading groups in this thriving field of research, both in physics and chemistry, driven by the dual interest in the interplay between coupled degrees of freedom on a crystal lattice, and the quest for possible application, especially in electronics and computing. WISH is ideally suited to make a very significant impact in this field, since it will enable the study of the magnetic phase diagrams of many of these compounds in unprecedented detail, whilst preserving the ability to refine crystal structures with good accuracy. b.

Magnetism in covalent systems

Recently, there has been a surge of interest in multi-component systems containing covalently bonded light atoms, such as borides, boro-carbides and fullerides. Although the primary research motivation stems from the discovery of superconductivity in many of these compounds, these systems are also known to display a variety of commensurate and incommensurate magnetic structures, often with poor or low-dimensional order. Systems such as CeB6 and DyB2C2 have also been shown or proposed to display quadrupolar ordering, which can be probed by applying a magnetic field. Here, the emphasis is on high flux and low background, all pre-eminent characteristics of WISH, which are particularly crucial in determining the evolution of the lineshapes in the present of low-dimensional ordering. c.

Model systems

There is a continuing interest to study the fundamental properties of magnetic Hamiltonians, both in the classical and quantum regimes, in interesting lattice geometries. Low-dimensional systems, such as chains, ladders and planes with a variety of geometries, continue to be highly topical, and there is a growing interest in frustrated systems, such as pyrochlores and spinels, especially when the spin degrees of freedoms are coupled with charges and orbitals. Inorganic systems exploit well-known “natural” crystal architectures, which usually have a limited degree of tuneability. A strongly emerging field is that of “designer” magnets, which incorporate an increasing proportion of molecular components, giving rise to more complex crystal architectures and, generally, larger cells. A key challenge in this field is to understand, and perhaps control the relation between crystal architecture and magnetic exchange and collective properties, and neutron diffraction provides the most incisive technique for this. In many systems with low-dimensional connectivity, magnetic

ISIS Second Target Station Project

ISISTGT2/SAC/02/P4_1 WISH

p. 5

ordering occurs either in rods or planes, and low background is essential to probe diffuse features. The high resolution, high flux and low background of WISH will make it ideally suited to play a leading role in this research. d.

Metallic magnets

Heavy fermion systems are at the centre of intense theoretical and experimental activity. These strongly correlated systems are dominated by spin fluctuations, which are expected to become critical when tuned by pressure, chemical pressure or magnetic field to a quantum critical point, poised between a magnetic ordered state and a fluctuated state. Interest in this type of system is twofold: it represents a particularly challenging problem in many-body physics, and it might have important consequences for understanding and exploiting other many-body phenomena such as superconductivity. There is also continuing interest to study the evolution of magnetic moments and magnetic structures in itinerant magnets, where the magnetic interaction is modulated through space. The study of the magnetic structure of some of these systems is particularly challenging, because of the small magnetically ordered moments. If equipped with a single-crystal detector, WISH will be ideally suited for these studies. WISH will excel in the study of systems with larger magnetic moments and complex incommensurate structures, both in powder and single crystal form. e.

Magnetic clusters and nanoparticles

In recent years transition metal clusters containing a core of coupled magnetic ions such as Mn12, Fe8, Cr6, separated by organic ligand frameworks have stimulated enormous excitement because of their potential in high-density data storage and quantum computing. Such 'single molecule magnets' represent represent the limit in size for such applications. There is now evidence that some of these compounds may display long-range magnetic order at very low-temperature, which provides the opportunity to study the role played by dipolar interactions, especially in applied field and pressure, whilst monitoring the evolution of the crystal structure. f.

Magnetism under extreme conditions

Application of external pressure and magnetic fields influences magnetism in a variety of ways, either directly (e.g., by modulating the magnetic exchange interactions and the magnetic Hamiltonian), or indirectly (e.g., affecting the stability of the underlying crystal structures). Further, pressure and magnetic field can induce valence or spin state transition both in ionic systems, such as LaCoO3, and in metallic alloys such as YbInCu4. In many cases, spins can be readily re-oriented in a magnetic field, which is often a research topic in itself (for example, in the study of spin-flop transitions). g.

Large-unit-cell structures

The applications of cold-neutron, high-resolution powder diffraction are by no means limited to the field of magnetism. In general, for materials with large unit cells, useful information is only available at long d-spacing, either because high-q data are suppressed by disorder or simply because they are too degenerate to be useful in conventional crystallography. If the crystallographic symmetry is low, even the long d-spacing data are nearly degenerate, and good resolution is required. A dedicated long-wavelengths diffractometer, HRPD-II, is being proposed for TS-II for highly crystalline systems that can benefit from high resolution. However, in many cases, sample-dependent peak broadening is significant, and the instrumental resolution does not need to exceed the WISH resolution (0.1%). In this field, we envisage that a highperformance long-d-spacing diffractometer with sufficient resolution will take a leading role in the study of large inorganic structures, such as zeolites and other framework compounds, and smaller pharmaceutical compounds. Outline Design Specification: Introduction. Powder diffraction has been arguably one of the most striking success stories at pulsed neutron sources. Time-of-flight powder (TOF) diffractometers are clearly very competitive with their constant-wavelength (CW) counterparts in a variety of applications, and there are a number of areas where the two types of machines are distinctly complementary. In general, for a given maximum resolution, CW diffractometers have a higher neutron flux on the sample and a more sharply peaked resolution function near the “take-off” angle, while TOF diffractometers have larger detector solid angles and flatter resolution curves. The latter is achieved by exploiting the polychromatic nature of the pulsed beam at a fixed scattering angle. Backscattering is a natural choice, because it provides the best resolution and is the intrinsic focusing condition for TOF diffractometers, where the resolution becomes largely independent on the beam divergence and sample size. However, other scattering angles are also convenient: for example, background suppression is easiest near 90°, and lower angles provide the widest d-spacing range. In any case, the ability to produce patterns with nearly constant resolution in a wide range of d-spacing is one of the distinct features of TOF diffraction, and has been exploited in the construction of

ISIS Second Target Station Project

ISISTGT2/SAC/02/P4_1 WISH

3

10

WISH

2

10

1

10

0

10

GEM -1

10

-2

10

1

10

Wavelength (Å)

Fig. 2: Spectral flux Σ (flux per 0.1% bandwidth) on GEM and WISH at the sample position. With logarithmic binning, Σ is proportional to the number of counts per bin for an incoherent scatterer.

p. 6

instruments such as HRPD and OSIRIS, which, for some applications, have no CW counterpart. How much the wavelength-dispersive, constant resolution technique can be extended to long d-spacing depends crucially on the flux at long wavelengths. HRPD, which looks at a L-CH4 moderator with thermalised flux peaking around 2 Å, is practically limited to about 4 Å d-spacing in backscattering. OSIRIS looks at a L-H2 moderator with peak flux at 3 Å, and employs strong focusing to enhance the longwavelength flux, yielding “practical” d-spacings of the order of 10 Å. The cold solid CH4 moderator of TS-II, with thermalised flux peaking around 4 Å, is an ideal choice to extend this trend towards even longer d-spacing. Fig. 2 clearly illustrates that GEM and WISH will be complementary, with a crossover point at a wavelength of 2.5 Å. As we have argued, this capability provides a perfect match to the scientific case of the high-resolution magnetic/long d-spacing diffractometer WISH.

Moderator and flightpath. In backscattering, the resolution function of a TOF powder diffractometer is largely determined by the pulse shape. Although, in principle, any line-shape can be modelled and fitted, in practice it is best to choose a de-coupled moderator with a simple pulse shape that can be modelled analytically. Also, one should consider that, for modern diffractometers operating with long wavelengths, the beam can be transported very efficiently at long distances, and the choice of flight-path is related purely to the required single-frame bandwidth. The total flightpath, Ltot, is related to the repetition rate, ν, and the single-frame bandwidth, ∆λ (the maximum bandwidth that can be employed whilst avoiding frame overlap), by the following formula: 3.96 × 103 ∆λ  Å  = ν [Hz ] ⋅ Ltot [m]

In order to maximise the data collection efficiency, it is convenient to limit ∆λ to be a small multiple of the peak-flux wavelength, and obtain longer wavelengths by re-phasing the choppers (rather than reducing the repetition rate). A wider single-frame bandwidth is not convenient, because it extends the frame to a region where the flux is much lower than the peak flux, forcing one either to over-count on the short wavelengths or to under-count on the long ones. We have chosen ∆λ ≈ 1.9 ⋅ λpeak = 7.5 Å , which yields a total flightpath of 53 m. Once established the flightpath, the “ideal” pulse width is a function of the required “maximum” resolution in backscattering. Since magnetic peaks are almost always broadened by domain size effects, extremely high resolution is not required. The “broad-face” of the coupled solid CH4 moderator, with a long-wavelength FWHM of 100-150 µsec, would yield a ∆T/T resolution of