The martensitic phase transition in Ni2MnGa, fundamental to its shape memory behaviour, can be described by two successive 110 type shears leading to 36 possible different orientations for the axes of the pseudo-tetragonal martensitic phase. The distribution and orientation of the domains formed on cooling Ni2MnGa into the martensitic phase has been studied using single-crystal neutron diffraction with a multi-detector. The number of domains actually occurring was rather low and was reduced by residual strain. Thermal cycling through the phase transition was found to further reduce the number of domains occurring, which may stabilize after several cycles. Uniaxial stress or a magnetic field applied parallel to [001] is able to switch domains whose pseudo-tetragonal axis is [100] or [010] to ones of type [001]. The results suggest that plastic deformation in the martensitic phase takes place by twinning (change of domain), rather than by slip and that the shape memory property arises from the fixed orientation relationship between the martensitic twins and the high temperature cubic axes.

High resolution neutron powder diffraction and single crystal measurements on the ferromagnetic shape memory compound Ni2MnGa have been carried out. They enabled the sequence of transformations which take place when the unstressed, stoichiometric compound is cooled from 400 to 20 K to be established. For the first time the crystallographic structure of each of the phases which occur has been determined. At 400 K the compound has the cubic L21 structure, and orders ferromagnetically at TC ≈ 365 K. On cooling below ∼ 260 K a super-structure, characterized by tripling of the repeat in one of the ⟨110⟩cubic directions, forms. This phase, known as the pre-martensitic phase, persists down to the structural phase transition at TM ≈ 200 K and can be described by an orthorhombic unit cell with lattice parameters aortho = 1/√2acubic, bortho = 3/√2acubic, cortho = acubic and space group Pnnm. Below TM the compound has a related orthorhombic super-cell with bortho ≈ 7/√2acubic, which can be described within the same space group. The new modulation appears abruptly at TM and remains stable down to at least 20 K.

The influence of an applied magnetic field on the antiferromagnetic transition temperature of the localised metallic antiferromagnet Tb2AgIn has been studied. Magnetisation measurements show that as the applied field strength increases the transition temperature is shifted to lower temperatures with a B2 dependence as expected on the grounds of symmetry. On the basis of measurements made in static fields of up to 5.5T, it was estimated that a field of 7.7T would be sufficient to suppress antiferromagnetic order at T=0K. This prediction has been confirmed using pulsed high field magnetisation and neutron diffraction measurements in an applied field. The magnetic isotherms have a complex field dependence producing an entropy change ΔS which becomes negative above the Néel temperature ∼58K. The temperature at which the entropy change passes through zero provides a more precise value for the Néel temperature than can be obtained from Arrott plots. Significant magnetic entropy persists above 58K and specific heat measurements show that the Rln(13) expected for Tb3+ is only recovered at ∼2.5TN.

The effect of heat treatment on the structural and magnetic properties of Pt3V has been investigated by neutron diffraction and SQUID magnetometry on single crystals. The results indicate that the ordered Cu3Au (L12) structure can be stabilised at low temperatures. However, both the atomically ordered and disordered crystals remain essentially paramagnetic down to at least 2K. Polarised neutron measurements in a field of 4.6T yielded at 5K a moment per Pt3V formula unit of 0.009(2)μB for the atomically ordered sample, which was a factor of three larger than that determined for the disordered sample but more than two orders of magnitude smaller than that predicted by band theory.

Polarized neutron scattering has been used to determine the changes in the distribution of unpaired electrons which take place in the martensitic transition in Ni2MnGa. Ni2MnGa is a ferromagnetic Heusler alloy which undergoes a reversible transition at about 220 K from a high temperature cubic phase to a low temperature tetragonal one. It has been suggested, on the basis of band structure calculations, that the structural phase transition is driven by a band Jahn-Teller distortion involving redistribution of electrons between 3d sub-bands of different symmetries. The results of the neutron scattering experiments show that the transition from the cubic to the tetragonal phase is accompanied by a transfer of magnetic moment from Mn to Ni. The unpaired electrons in the cubic phase have overall eg symmetry. In the tetragonal phase, the degeneracy of the eg and t(2g) bands is raised and the unpaired electrons are redistributed in such a way that the subbands based on orbitals extending towards the c-axis are preferentially occupied. Although the experimental moments differ in detail from those expected from band structure calculations, the change in symmetry of the magnetization distribution is consistent with a band Jahn-Teller origin for the phase transition.

Specific heat measurements are reported for the alloy series Pd2GdxLa1−x In for 0⩽x⩽1.0, and also for Pd2LuIn and Cu2GdIn. A well defined anomaly associated with the onset of the magnetically ordered state is observed in the specific heat for Pd2GdIn at T=9K. As x decreases the anomaly becomes broader and the maximum reduces to a lower temperature. Measurements on the non-magnetic compounds Pd2LaIn and Pd2LuIn were made to isolate the magnetic contribution to the heat capacity of the Gd compounds. Such an analysis reveals a significant magnetic contribution remaining at temperatures far above TN (>2TN) for the compounds where x=1.0, 0.75 and 0.50. The total magnetic entropy at 20K is well in excess of the theoretical value of xRln(8) for these compounds. Above 20K magnetisation measurements on Pd2GdIn show a Curie–Weiss behaviour and the absence of short range magnetic order, whilst the specific heat of Pd2GdIn is still enhanced at 30K. Above 20K the enhanced specific heat varies linearly with temperature, suggesting a γ-value of 202mJ/K2mol. Subtraction of such a linear term over the entire temperature range yields a magnetic entropy more consistent with that expected for Gd. For the S-state element Gd, the existing theories put forward to explain such an enhanced γ-value on the basis of heavy fermion behaviour are not applicable. The possibility of an additional conduction electron contribution to the magnetic specific heat is supported by measurements on Cu2GdIn, although the magnitude of the enhancement in Pd2GdIn is difficult to reconcile with this approach.

Specific heat and magnetization measurements on the hard superconductor confirm the presence of a structural phase transformation at 117 K and a superconducting transition temperature of 9 K. Above 117 K the compound has the cubic Laves phase C15 structure. High-resolution neutron powder diffraction measurements reveal that below 117 K there is a mixed phase structure consisting predominantly of an orthorhombic component together with a cubic untransformed part. Both of these phases superconduct and their critical temperatures are similar, if not identical. In the normal state the magnetic properties are consistent with those expected for a Pauli paramagnet in which the Fermi level lies just below a high peak in the density of states. These conclusions are supported by band-structure calculations carried out using the linear muffin-tin orbital (LMTO) approximation.

The antiferromagnetic structure of the intermetallic compound has been determined from neutron polarimetric measurements and refined by combining these data with integrated intensity measurements. The structure was found to be non-collinear with the U moments confined to the a - b plane. The moments of U atoms in each of the two sets of sixfold sites are arranged hexagonally with rotations of between them and the two sets are rotated with respect to one another by . The third (twofold) set of U atoms has no ordered moment. These conclusions are in disagreement with a previous determination of the structure from powder data which gave a collinear structure with moments parallel to the c axis. Magnetization measurements made on single crystals in the temperature range 300 - 2 K can be understood in terms of a transition to a non-collinear easy plane antiferromagnetic structure stable below 22 K. Polarized neutron measurements have been used to determine the contribution of each of the U sites to the susceptibility between 22 and 2 K. These show that of the two sixfold U sites, that with the smaller ordered moment contributes more than half of the total susceptibility. The twofold site, which is characterized by a small U - U separation, makes the smallest contribution.

Heat capacity and magnetisation measurements are reported on polycrystalline samples of Pd2Gd x La1−x In. For x=1, a Curie-Weiss susceptibility is indicated down to approximately 15K with a magnetically ordered state at temperatures below 9K. Anomalies are seen in specific heat measurements with the magnetic transition temperature and magnetic entropy decreasing with decreasing x. Subtraction of the lattice and electronic contributions to specific heat for the x=1 compound indicates a magnetic entropy of Rln(7.87) in good agreement with theory. Above magnetic ordering temperatures the specific heat remains high with an enhanced electronic heat capacity coefficient. γ=(250+/−20) mJ mol−1 K−2 for x=1.

An investigation is reported of the magnetic ordering in Pd2DySn using magnetisation, specific heat and neutron scattering. A first order magnetic phase transition is identified at T=5K leading to an antiferromagnetic ordering below. An alternative magnetic order is suggested for temperatures between 5K and 12K.

Magnetisation and susceptibility measurements on intermetallic compounds based on the Heusler composition Pd(2)TiZ where Z is Al, In or Sn are reported.

I have read the article "Units and unity in magnetism" by John Crangle and Mike Gibbs (November p31) and the subsequent correspondence (January p19) with interest – and a strong sense of history repeating itself! Your correspondents make a number of interesting points, but no one has yet touched on what seems to me to be the fundamental reason why there have been problems with the basic descriptions of magnetic phenomena, particularly since SI was introduced.

A ternary intermetallic compound containing elements which themselves do not order magnetically has been found to have a spontaneous magnetisation. The compound, Pd2TiSn, is crystallographically ordered in the L21 structure with a lattice parameter of 6.38 Å. Magnetisation measurements indicate a ferromagnetically ordered moment per formula unit of σ = 0.005μB which essentially does not change with temperature over the range of 2 to 300 K. Over this temperature range, magnetic hysteresis was observed, confirming the existence of a ferromagnetic state. Arrott plots reveal that the isotherms in the high field region are characterised by straight lines which, over the temperature range investigated, converge approximately to a single point on the σ2 axis. Such behaviour can be accounted for if the coefficients in the free energy expansion giving rise to Arrott plots have a similar temperature dependence. A description of these observations is given in terms of a band model.

Single-crystal magnetisation measurements are reported for the intermetallic compound U14Au51. Two magnetic phase transitions are identified at low temperatures. The results are compared with measurements carried out on powder samples.

Inelastic neutron scattering experiments have been carried out on rare earth based Heusler alloys. Results are reported for Pd2HoIn and Pd2YbIn. The crystal field (CF) level scheme is determined and the results are compared to bulk measurements of the magnetisation in the temperature range 2 K to room temperature and in applied magnetic fields of up to 5 T. A comparison is carried out for the CF schemes between the In and Sn based series.

Superweak ferromagnetism in transition metal compounds is characterised by the combination of unusually large transition temperatures and small magnetic moments. Using a simple microscopic model it is argued that superweak ferromagnetism is more general phenomenon which characterises the magnetic behaviour of transition metal compounds with the lighter transition elements.

An investigation of the effects of heat treatment on the magnetic properties of Pd2MnIn is reported. The results are compared to theory in a simplified model employing an RKKY interaction between the Mn atoms.

When we measure any quantity it is important to define the units that we use. If other workers are to use the results of our measurements it is essential that they know how to convert our units into their preferred units. It is convenient if they use the same units, especially if everyone uses the same set of units, universally defined and agreed. This is not the case in magnetism as, for example, the two feature articles in this issue of Physics World illustrate. The article on giant magnetoresistance (p34) uses oersteds for the magnetic field, whilst the author writing about nanocrystalline magnets prefers tesla (p40).

Powder neutron diffraction measurements have been undertaken on Pd2TiIn over the temperature range 4-300 K. Previous measurements of the bulk magnetic properties indicated a magnetic transition in the neighbourhood of 110 K. The neutron measurements reveal a structural phase transition at 92 K. The transition appeared to be of first order, and no evidence was found of any long-range magnetic order below 92 K. However, the existence of ferrimagnetic order, involving the Pd and Ti atoms, could not be ruled out. The results are discussed in terms of the stability of the magnetic moment and the presence of spin fluctuations.