PolRef

PolRef is used for reflectivity studies using a polarised neutron beam. The technique is primarily used to study magnetic thin film materials (such as metals, oxides, multi-layers and superlattices) that serve as the basis of a large proportion of electronic technology.

Neutrons have a magnetic moment and precess around a given magnetic field direction (similar to protons in NMR). The neutron beam can be prepared in either a ‘spin up’ (neutron polarisation aligned to the guide field) or ‘spin down’ (anti-aligned) configuration. The refractive index of a magnetic sample has an additional magnetic component, which either adds or subtracts from the non-magnetic refractive index. By comparing the ‘spin up’ reflectivity to the ‘spin down’ reflectivity the magnetic refractive index depth profile can be determined. This allows the layered materials to be studied, and the results analysed, leading to developments that pave the way for these systems to be ultimately turned into electronic devices.

Recommended and optimal sample dimensions.

SAMPLE: Recommended and optimal sample dimensions (area)

For condensed matter experiments it is important that samples be as large as possible, while still maintaining flatness and thickness uniformity across the area of the sample. Furthermore, consideration of the sample environment that is to be used is essential, as this will provide the upper limit on sample dimensions.

The size of the sample is important, since in comparison to an x-ray source, the flux of a neutron source is much lower. Consequently, this requires that samples be large to maximize the use of the available flux.

The average size of sample that is used on POLREF for hard condensed matter investigations is approximately between 10 x 10 mm and 20mm x 20mm. This provides a decent count time and reasonable max Q range, while not being so large that extra efforts (such as thicker substrates or more comprehensive sample growth) are required to maintain thickness uniformity and sample flatness.

General sample size recommendations:

The largest sample sizes are restricted by the sample environment (see sections below). The average and minimum recommended sample sizes are as follows:

• Average size measured on POLREF: 10 x 10 mm to 20 x 20 mm
• Minimum recommended size: 5 mm (long) x 5 mm (wide). Be warned that both the count time will increase and max Q reachable will decrease significantly. Smaller samples have been measured, but this required very careful sample preparation and experiment design to work with the high-count times, and large reduction in Q range.

If you have no other choice but to use very small samples (below the recommended minimum size), please enquire with one of the instrument contacts prior to proposal submission so that they can advise on experiment feasibility.

PolRef electromagnet magnet only:

• Largest recommended size: 100 mm (long) x 40 mm (wide) – while in principle we can measure longer samples than this, it becomes diminishing returns and can add measurement complexity.

PolRef flow Cryostat 3.5 K base temperature (electromagnet is always included):

• Maximum: 40 mm x 28 mm is the maximum size that can fit in the cryostat due to the sample mount and gives the best results.

Variox Cryostat 1.5 K base temperature (electromagnet always included):

• Maximum: 20 x 20 mm – due to the sample mount.

HTS110 80 mm bore 3 T superconducting magnet only:

• Maximum: 25 mm x 25 mm – limited by the sample holders and the bore of the magnet. Furthermore, samples need to be mounted to the sample holders – typically using GE varnish. Samples can and do move when measuring and changing fields.

SAMPLE: Thickness

PolRef can measure film thicknesses up to approximately 4000 Å (dQ/Q = 1%) and still resolve fringes. This upper limit is challenging, and normally PolRef can easily measure maximum thicknesses of 2000 Å. The lower limit on thin samples than can still obtain a measurable signal is about 10 Å. This can be circumvented by the use of multilayers to boost the reflectivity.

The average sample thickness is usually between 0 -1500 Å (dQ/Q = 2% to 4%) for condensed matter systems.

Before submitting a proposal or coming to ISIS, it is always worth reaching out to the local contact about the best sample dimensions for the experiment in question.

SAMPLE: Roughness, off-specular scattering and curvature

Roughness: Surface quality is paramount. PolRef can deal with roughness on the atomic scale of up to and around 200 Å RMS (approximate maximum).

Generally, if a sample is visibly rough to the naked eye, then it is very unlikely a good specular reflection can be achieved. This will manifest itself often as a rainbow-coloured, oil slick-like reflection when the sample is held up to a light and tilted slightly. In extreme cases, it will look like scrunched up tin foil – this certainly will not work.

Off-specular scattering: There are exceptions to this, for instance deliberately patterned media can give rise to in-plane Bragg peaks. Again, if you are not sure please contact the beamline scientist or assigned local contact.

In off-specular scattering mode, PolRef can resolve features on the order of 0.25 micrometres to approximately 40-50 micrometres, the best result being obtained for features on the order of about 2-4 microns in size. It is also important that the surface coverage is high, otherwise the count times become very large, especially if polarised neutrons are being used (on the order of a few days). Again, it is best to contact the instrument scientist or assigned local contact about this.

Curvature: Large area samples often suffer from this. Convex or concave samples act as focusing/defocusing mirrors; however, the focal point will almost never be at the detector (Well so far anyway). In the past, samples have been checked for this using a profilometer (expensive) or simply by shining a laser pen (cheap) onto the surface and measuring the reflected spot size (at a suitable distance) as compared to a reflected spot from a piece of thick silicon wafer, or any other reflective surface that is almost atomically flat. Any sample that produces a spot that is a factor of 2 wider or more than a flat silicon wafer (these tend to be the flattest thing most easily/cheaply available for purchase) might not work too well.

An on-beamline solution to this is to make a smaller area sample out of the big area sample, often by cutting the sample up. Often, due to the growth, the central regions of samples are flatter than the outer regions, but this can be heart breaking to do, so check beforehand.

The problem of curved samples due to the stress of the film can often solved by simply growing on thicker substrates. A lot of polymer experiments are done on 5 or 10mm-thick silicon substrate blocks for this reason.

SAMPLE: Substrates and reference layers

Substrates: There are a few things to remember concerning substrates. Some materials are not as friendly as others as substrate materials.

Generally, it is best to avoid double sided polished substrates, unless the substrate material in question has a significant neutron absorption cross-section, e.g. borated glass. This is because, if the substrate is relatively thin (in the order of mm or less), reflections from both the surface and back surface of the substrate can reach the detector in a similar location – the resultant reflectivity becomes a superposition of the two surfaces, which has to be accounted for in data analysis; adding unnecessary parameters and complicating the fit in comparison to using a single sided polished substrate.

Structural transitions with temperature can cause issues. The classic case is STO substrates, which are renowned for the structural transition they undergo around 120 K (It has two further transitions below the primary 120 K transition; this makes experiments below 120 K very challenging). This substantially reduces the reflectivity of the sample below these temperatures, due to a significant roughening of the surface as the system moves through the transition. These transitions can also change the structure/magnetic properties of thin films, sometimes irreversibly, when cooled down below the transitions, which may or may not be what you want to happen. There are alternatives, like LSAT, that have similar lattice properties – although these substrates can often be rougher than STO; substrate screening prior to growth is essential. If you need temperatures below 120 K and can avoid using STO, we recommend you do so.

MgO and other oxide substrates can vary in quality – even within the same batch! Therefore, as mentioned above, it is essential substrates are screened prior to growth – by using transverse off-specular rocking curves and traditional XRR the degree of twinning and substrate broadening can be quantified. Contact the instrument team if you are unsure about any aspect of this.

Sapphire substrates can also cause issues. If you are using D₂O in your experiment the contrast between sapphire D₂O is very small. Aligning the system is going to be very hard and you will need extra time in your experimental plan.

Cleaning substrates can also be problematic. For instance, if you are using substrates for solid\liquid work, you also need to think about the substrate cleaning procedure and a suitable method to verify that the cleaning has worked. ISIS can provide various cleaning methods, from ozone cleaning to etching; however, you must contact your beamline scientist or assigned local contact.

Reference layers: There has been an increase in biological experiments using Au, Cr, Fe and Ni reference layers, to name a few, to provide extra contrast for biological or chemistry scattering experiments. Often, these reference layers are not sputtered by the groups that are using them, and the quality can vary enormously.

It is wise to check this before coming to ISIS at your home institution; however, we do have a lab-based X-ray machine for doing this offline, if you cannot get access at your home institution, but demand at ISIS is very high. Please contact the instrument scientist or assigned local contact about using this machine beforehand to check your substrates and reference layers. This should be done prior to use on the beamline. This can save a lot of lost beamtime and experiments.

It has come to our attention that the magnetic reference samples often require some magnetic characterisation. Please enquire with the instrument contacts for more information about this.

Soft matter Experiments:

Liquid and biological experiments have standard sample sizes of 200 mm (long) x 80 mm (wide) due to the sample Teflon troughs provided. We can supply dimensions and designs if required. Some user groups build their own troughs.

Polymer samples:

Polymer films require large areas as well but must be flat to the eye in-order to get good reflectivity.

Successful experiments have used polymer systems spin coated onto thick silicon blocks about 25 to 30 mm in diameter or bigger. Again, the bigger the surface area, the better for neutron reflectivity measurements, but more care is required to stop bending and surface curvature.

The thick nature of the films can result in stress on the substrate. In order to avoid curvature of the film, thick substrates need to be used 5 to 10 mm thick. This is not always the case though.

If two or more polymers are to be used, it is important that at least one polymer is deuterated. This is due to the fact that most polymer compounds have almost identical contrasts, making it very hard to distinguish between them with neutrons, unless deuterated material is used for at least one of them – deuterated polystyrene being a classic example. There is an ISIS Deuteration facility, which can be inquired about at the user office and through the ISIS proposal system.

Characterisation Measurements

It is highly recommended, but optional, that some basic characterisation measurements are made before coming to the beamline. It is understood that this may not be possible in the case of certain samples, systems and time constraints.

Hard condensed matter (HCM) experiments: (minimum)

It is suggested, at a minimum, that a Cu Kα X-ray or equivalent structural X-ray be taken beforehand. Both specular and off-specular curves time permitting.

If the experiment is polarised, it is also recommended that the equivalent magnetometry also be performed. (MOKE/VSM/SQUID).

XRR, XRD, SQUID and VSM are available at ISIS. Again, demand is very high, and it is advisable to do this at your home institution.

With all of these, it is much easier to plan a successful HCM experiment as you will have a rough idea of where in phase space to look – e.g. what fields, temperatures etc. to use. As count times can be on the order of days, this is invaluable in avoiding mistakes.

This data is also invaluable in fitting the neutron data as it allows you to tie down variables such as the total moment of a sample, layer thickness and roughness. This makes the data analysis easier and more rigorous.

Other helpful characterisation measurements:

Other characterisation techniques that are of great help in neutron experiment planning and data fitting are:

MFM/AFM: Can provide top surface roughness as well as magnetic domain size for estimating in-plane magnetic features. FFT of AFM/MFM images is also useful for validating off-specular data.

TEM: TEM images are brilliant for fitting data. If done well, you can directly observe the SLD variation in the images allowing you to set up a fitting model more accurately. They are also brilliant justification to referees for why a model has been set up. There is a user facility at RAL that can provide this. Please see the Diamond ePSIC website for details. Transport data – resistivity/TMR/GMR/etc.: Either as a function of field or temperature. This allows us to set fields and temperatures correctly if we are looking for transitions, such as superconducting phase change, and are again very useful for planning a successful experiment – if your experiment involves in-situ transport measurements, then pre-characterising is essential.

These techniques are but a few of the many available. It is worth reaching out to your local contact as some of these can be provided at ISIS.

If you have these measurements done beforehand, they are ideal for papers/thesis write-ups/experimental reports/follow up proposals.

Soft matter experiments:

For soft matter experiments, there are a vast range of characterisation techniques available, some of which are listed below. It is understood that different groups have their own way of working and techniques. We strongly suggest characterising the property under study and optimising your experimental conditions prior to beamtime. Below is a list of surface characterisation techniques often employed in soft matter experiments to provide complimentary information to interfacial structure described by neutron reflectometry data. Some of these techniques are available at ISIS.

Atomic force microscopy: Widely used microscopy technique for examining structures at the solid/air or solid/liquid interfaces. (Available at ISIS)

Brewster angle microscopy/ imaging ellipsometry (BAM): Microscopy techniques that have high z-axis resolution and are generally employed to gain in-plane information on the structure of interfacial films. (Available at ISIS).

Surface infrared spectroscopy: Infrared techniques, such as Attenuated Total Reflectance (ATR) and external reflection (sometimes known as IRRAS), allow for the chemical composition of interfacial films to be examined and can be very effective in examining interactions at interfaces. (Available at ISIS).

Surface pressure measurements: Users working at the air/liquid interface are likely to measure surface pressure prior to or during NR measurements. It is highly desirable that users also conduct these measurements prior to beamtime. (Available at ISIS).

Surface plasmon resonance: A widely used technique to monitor interactions at the solid/liquid interface. Most universities will have access to this apparatus.

Transmission electron microscopy: An excellent imaging technique, although sample preparation for interfacial films can be challenging.

Quartz calorimetry: Surface calorimetry technique, provides similar information to surface plasmon resonance, although quartz calorimetry instruments are often orders of magnitude cheaper to obtain.

Please note that if you require help with sample preparation that you are advised to contact the ISIS Support Laboratories. This is a facility is available to assist users with sample preparation/optimisation measurements prior to ISIS beamtime.

If you have these measurements done beforehand, they are ideal for papers/thesis write ups/experimental reports/follow up proposals etc.

1. Sample point goniometer capable of moving 700 kgs in 6 degrees of freedom.
2. GMW resistive magnet (0.7 T 100 mm gap, or 1 T for 50 mm gap).
3. Continuous flow cryostat (3.5 – 300 K) – mounted within the GMW magnet (±0.7 T Max field).
4. Variox Cryostat (1.6-300 K) – mounted within the GMW magnet (±0.7 T Max field).
5. Vacuum Furnace (300 – 500 K) – mounted within the GMW magnet (±0.7 T Max field).
6. HTS110 3 T magnet with horizontal and axial field up to ±3 T, with a minimum field of approx. ±300 Oe with polarised neutrons.
7. Multi-sample changer – up to 6 for air/solid measurements in GMW and HTS at room temperature.
8. 2-position sample changer in the continuous flow cryostat.
9. In-situ 4-point probe and 2-point probe for applying high voltages (+/-1000 V) available in furnace and cryostat.
10. Helmholtz coil set for very low field work sub 30 Oe down to 1 Oe.
11. Custom sample environment can be developed in collaboration with the ISIS Sample Environment team. Please contact the instrument scientists to discuss.

PolRef Data Reduction is handled by the Mantid data reduction software, using Jupyter notebooks – please contact the instrument team for more information.

PolRef Data Fitting: The PolRef instrument team support Refl1d as our primary fitting package, but the team also have experience in GenX. More widely, the NR instrument team supports and develops RasCAL for soft matter experiments – all software is available on Ada (previously IDAaaS). Requires ISIS user login:

• Refl1d – See read the docs, and Github pages. NR/PNR/PA fitting, including support for the Felcher effect. Bayesian analysis included in the fitting package and is our recommended package for HCM experiments on PolRef.
• GenX: Simultaneous X-ray and Neutron (NR/PNR/PA) fitting. Can fit Soft X-rays, Cu K alpha X-rays, NR, PNR and PA.
• RasCAL: Optimised for biological data sets. Included Bayesian data analysis and nested sampling routines.

• Neutron scattering lengths and cross sections (NIST)
• Scattering length density calculator (NIST)
• Open Reflectometry Standards Organisation (ORSO)