The neutron strain scanning technique is a non-destructive and powerful method for the spacially resolved determination of residual stresses in materials and mechanical components. The high penetration power of neutrons allows measurements in small samples and big components using a cube shaped gauge volume. No special sample treatment is necessary for the measurement, i.e. surface polishing essential for X-ray measurements. This makes the method applicable to a large variety of applications in materials science, in engineering and in industry. Some examples are: developing new materials (i.e. composite materials), developing and controlling sample treatments (i.e. hardening, welding etc.), proving the results of finite element calculations. Industrial users get important information to improve the quality and lifetime of components.

The demand for beam time at the D1A strain scanner has been rapidly increasing during the past 2 years and is still growing. The users come from applied research, but also with direct industrial applications. Very active are Italian groups who work on composite materials, quenching and automotive parts, and of course the many user groups from British Universities who work on composites, steel and welding processes. This year we have first proposals from Austrian Universities. The Fraunhofer Gesellschaft für zerstörungsfreie Prüfverfahren, Dresden (Fraunhofer Institute for non-destructive test methods, Dresden, Germany) is very interested in a collaboration with ILL. They have good contacts with industry and other German research groups, and may attract more new German users.
 
 

A proposal for a new strain scanner

The present instrument is not dedicated to stress. It is built on the D1A powder diffractometer and shares the measuring position with powder experiments. Although the beam optics are unique and allow very precise determinations of strains, many improvements are needed. The limited space restricts the allowed size of samples and complicates their alignment. But with more space for bigger samples, better mechanics for sample positioning and an optimised monochromator, this machine will be one of the world's leading instruments in the field of residual stress determination.

Although the strain scanning method is based on the measurement of Bragg peaks, the demands on a strain scanner are very different from those to a powder diffractometer: the resolution need not to be good over a wide 2-theta range but should be optimised in the range of 90^o 2-theta because the gauge volume is best defined here. The use of a powder instrument as basis of the strain scanner is therefore a compromise, and ILL needs a dedicated instrument for the analysis of stress.
 
 

Demands on a neutron strain scanner:

To fulfill the demands of a large variety of applications in materials science, mechanical engineering and industry the instrument has to handle small samples as well as big components. It must be possible to align them quickly and easyily to high accuracy. Optical tools such as lasers, theodolites and closed circuit TV-systems help to do that job efficient. This implies that there is enough space around the instrument to position the alignment tools and to perform the alignment of even very large samples.

When talking about the resolution of a strain scanner, we have to distinguish between lateral resolution and angular resolution. The lateral resolution is determined by the size of the gauge volume and the sharpness with which it is defined by the optical components.

It implies that measurements are undertaken for reflections that lie close to 2-theta=90° because the gauge volume is best defined there. Since not every reflection of a material can be used for stress determination, a large variety of different wavelengths should be available. An instrument with a variable monochromator take off angle is the only convenient solution. For most used metals and alloys, and for composite materials a wavelength range from 1.5 to 3.5A should be available.

The precision of the positioning system is of high importance as well. Inaccuracies of more than 20m m in positioning can lead to fatal errors in the measured strain when there is a stress gradient or a surface. An accuracy of 10m m is recommended for the sample table even for big and heavy samples (up to 100kg).

The angular resolution means two things: the resolution with which neighbouring peaks can be separated and the precision with which the peak position can be determined. The latter depends on the quality of the peak. It is possible to determine a peak position to 1/100 of its full width at half maximum provided that the peak shape can be well described. This is the influence of the monochromator and the other optical components. The peak shape should be non-distorted and symmetric.

The instrument control software is a very important component of the system helping to undertake efficient alignment and measurements. It must be user-friendly to avoid errors and it shall allow online data analysis. Without immediate analysis it is very difficult to find the region of interest in the sample and it is impossible to modify the measuring program if necessary.
 
 

The present strain scanner at D1A

The present set-up contains of a Position Sensitive Detector (PSD) in combination with a radial collimator for the definition of the gauge volume, an xyz-table and a primary slit system. The system is mounted on the two axis powder diffractometer D1A with a fixed monochromator take off angle of 122^o. A vertical focusing Ge-monochromator provides a choice of the following wavelengths: 1.5A, 1.9A, 2.9A. Available sample environment is a 15kN stress rig and a mirror furnace.

Compared to other strain-scanners at monochromatic neutron sources, the ILL instrument defines the gauge volume by using a radial collimator instead of a secondary slit (see picture. The slit mask at the collimator further improves the resolution). This leads to a very sharp defined gauge volume without distorting the peak shape.

The collimator enables the same precision near surfaces as well as deep inside the samples. This is in agreement with the advantage of the neutron strain scanning technique to make non-destructive measurements on large samples. But the limited space at the D1A site restricts the maximum sample size and limits the number of applications.
 
 
 

Details of the proposal

The following section describes the proposed components for a new high precision strain scanner:

Secondary optics:

The radial-collimator/PSD combination as it is used at the present set-up at D1A is the preferred system. The definition of the gauge volume is very precise, and independent of sample properties, which is a very important point for reliable results. The lateral resolution of 0.6 mm is very good and since the collimator is at a fixed position, 150 mm from the reference point, no further alignment is necessary when changing a sample. The performance is the same at the surface as well as deep inside the sample. The background of this system is extremely low and there is no asymmetric distortion of the scattered beam due to the collimator.

The new strain scanner will use this existing system. An extra collimator can be added that defines a bigger gauge volume for faster measurements in big or high absorbing samples when a lower lateral resolution is sufficient.

Primary optics:

The primary optics that defines the cross section of the gauge volume should be held very flexible. Two pairs of motorised slits should define the primary beam cross sections between (0.5 mm)² and 50 mm² to allow measurements in all types of applications. The primary slit system defines the beam divergence to get either highest precision or higher intensity at the sample position.

A vertical focusing monochromator is necessary to get the highest neutron flux possible, but this leads to high divergence in the vertical dimension of the gauge. This reduces the accuracy of vertical scans. Similar to the secondary radial collimator, a primary radial collimator will overcome this problem. This collimator will be removable for horizontal scans when the vertical divergence is of no great importance. A super mirror coating of the collimator blades would increase the transparency. This type of collimator can be developed at ILL.

Sample positioning:

The big variety of samples and applications demands a mechanical system that can load heavy weights and big samples and position them with high precision. The whole system should leave enough space for sample environment like furnaces, stress rig etc.

Interfaces and surfaces are very important parts of a component as the initiation of cracks starts mostly here. The best lateral resolution possible is recommended for this type of analysis. This is not only a question of the optical components but also of the precise positioning of the sample. An uncertainty in the sample position bigger than 20 m m can lead to uncertainties for the measured strain that can lie around 50% or more when a stress-gradient is present or when the gauge volume enters the surface.

The present xy-table does not fulfil the demands because it is not stable enough to prevent tilting the sample during positioning and canít reach the demanded precision of 10um.

An open-frame linear position stage can guarantee a precision of 5m m and is very stable concerning tilt. An offer for such a system is available.

For the omega rotation a precision rail system is proposed. Mounting the sample table on a precision rail circle (diameter approx. 800 mm) allows the construction of a very flat and stable system in terms of torsion and load. It leaves more space for samples and environment than a conventional table and is more stable in terms of torsion. Furthermore it is cheaper than a big conventional omega table.

The detector should be positioned in the same way moving on a 1.60 m diameter rail. Linear encoder ribbons ensure reliable absolute reading of the position.

A New Monochromator:

The present monochromator could be re-used, but eventually a specially design monochromator is needed to provide:

• A variety of wavelengths: variable take off.

 
• Asymmetric cut. - gradient - ultrasonic

Three different possibilities must be checked:

• Asymmetric cut crystals: Asymmetric cut monochromator crystals would scatter the whole width of the incoming beam from the neutron guide into the narrow primary beam defined by the primary optics.

 
• A gradated monochromator could focus a specific wavelength-range

 
• Ultrasonic excited crystals show a reflectivity of 100% for neutrons of a specific wavelength. The neutrons are scattered from an only 100m m deep region below the surface of the crystal. The wavelength-spread can be chosen by varying the amplitude of the exciting ultrasonic wave. This would enable to adjust the angular resolution of the instrument. It is comparable with having a monochromator with a variable mosaicity.

Software

There is plenty of room for development of the present data analysis software. An important feature is the online analysis of data to include a monitor during the experiment to find the region of interest in the sample.

The present closed curcuit TV system, which is used for sample alignment, will be extended by software to permit the direct determination of the sample position on the TV-screen. This will increase the reliability of determining the true position of the gauge volume inside the sample. This is an important but very difficult exercise, as the sample shapes are often not regular. Two TV-cameras must be added to the system. The software will automatically control the motors and regulate their position. In this way it will be possible to position even the largest samples reliably to high precision.

Sample environment

Many samples are textured and since texture is an important parameter for the analysis of stress, a Eulerian-Cradle is essential to allow measurements in the appropriate sample positions. A big kappa-geometry cradle would take large and heavy components, and its open construction would not hinder the optical alignment systems. Such a cradle should be added in future.

A stress rig and a mirror furnace are already available.
 

Conclusion

Building a dedicated instrument at the D1A position would cause many problems. It is doubtful if there is enough room to change the casemate to a variable take off angle without interfering with the D1B monochromator. Furthermore there would be no space to walk around the instrument when it is set to a take off angle of 90 deg. Using the free position behind D1B would solve this problem (see also: technical annex).

The construction of the new high precision strain scanner could start immediately. Many components are already developed and will be re-used. New developments will be the primary optics and eventually the monochromator.

The new strain scanner will be a reliable and precise instrument for a large variety of samples and applications, and will combine all the advantages of the neutron strain scanning technique with the highest flux neutron source: it will clearly be one of the leading instruments in the world.
 
 

The proposed position of the new strain scanner behind D1B, and the present position at D1A. The space available at the D1A position limits the new mechanics, especially when we want to rebuild the casemat to provide a variable take-off angle near 90^o. Only the new position would provide enough space for all types of measurements. But the 9m of neutron guide between D1A and this site had to be replaced because it is too small and does not provide the whole available neutron flux.