During discussion with the ILL, a potential contamination of lambda/2 was mentioned with Germanium crystals that were ‘generated’ at the ILL. Using the test crystals we borrowed from the ILL (orientation (115)) we performed some time-of-flight (TOF) experiments to be able to check the wavelength of the neutrons reflected from the (335) orientation. The ‘diamond’ crystal structure of Germanium forbids a lambda/2 reflection: the next possible harmonic can only be (9 9 15)!
The Ge115 crystal was tested on the ‘ROG’ neutron reflectometer here at the RID. Placed in a 2theta=150deg orientation, the crystal reflects neutrons of 1.67AA (as expected), but also show some contamination of approx 2.2AA neutrons (several percent in intensity compared to the (335) reflection). This corresponds with the (224) reflection of Germanium. The claimed lambda/2 contamination cannot be seen on this instrument, because the incoming beam has little intensity at the wavelength (1.67/2 AA). The width (in lambda) of the reflection curve in the figure below are caused by the opening time of the TOF-pulse chopper (4%). No explanation yet for the (224) contamination.
The contamination of (224) in (335) cannot be explained by the geometry of the (amateur) setup: in the figure below the divergence of the incoming beam is so low, that any neutron of 2.2AA that reflects on the crystal (pink line) should do so outside of the opening angle of the detector, indicated by the thick black lines. The tau-vectors in this reciprocal space representation of the crystal are shown for (335 black) and (224 pink). The reflected 1.67AA neutrons all fall well within this opening angle of the detector, despite the increase divergence of that reflected beam. There should be no (224) in this (335) direction!
During the long silence since the last post we have been testing components and designing the instrument further. The detector we are going to build is based on scintillator techniques as used at ISIS, combined with smart electronics to discriminate (unwanted) gammas. In the picture below you see the wavelengh-shifting fiber on the first design drawings of the detector. Some 1000 of these fine fibers should be packed into a detector together with the scintillator material and optically isolated from neighbouring fibers.
In the months to come we will try to think of a manageable way to construct such a large dectector out of these fine fibers. Beside the mechanical challenges, the neutrons and gammas dictate that the should be a minimum amount of material in the flight way of the neutrons.
The unwanted radiation of the reactor core has to be absorbed to shield people and the instrument from this radiation. For people the instrument should not introduce more than 1uSv/h above the background level in the reactor hal. For the instrument the requirement is a good-as-possible signal to noise (S/N) ratio (for instance 0.001). To verify that the current setup is feasible shielding-wise from these requirements, MCNP calculation have been performed on this instrument. This includes the core, the reactor shielding and the instrument shielding (with the current level of detail).
The figure below shows the logarithmic levels of the neutron flux according to these calculations. A standard iron-PE-boron mixture has been used in these calculations. The picture below shows the beam from the reactor coming from the left. The germanium crystal monochromator (not displayed) reflects the wanted part of the neutron beam to the sample (in blue circle). The neutrons scattered from the sample are reflected to (green) detector. The detector is situated in an acceptable background flux of 1n/s/cm2. Collimation will improve this level of the background.
Fast, epi-thermal neutron, and gamma flux levels are higher but not unacceptable.
The ILL furnished test crystals of February 24th are being tested on our neutron reflectometer. After speaking to the ILL monochromator group this summer during the ECNS conference in Prague, the idea came to test these crystals for the alledged lambda/2 contamination which would contribute some 4% to the reflected intensity.
This nice image is born under the swiping finger of Jeroen.
The advantage of doing this in TOF mode is that the specific 2-theta configuration can be installed and tested for the whole thermal spectrum in one go. Above you see the detector tube horizontally above the valve through which the beam impinges on the crystal (in front). The crystal reflects the neutrons to the detector, Soon we realized that the intensity of the incoming beam at lambda/2=0.8AA is basically nihil, but there’s some interesting weak signal at higher d-spacings (see below). Even more unlikely than lambda/2 !
As can be seen directly in this graph, the tailored mosaic spread is rather non-Gaussian. The rising intensity wings left/right is due to division by the monitor spectrum (incoming spectrum).
The setup is being improved at the moment to avoid unwanted (spurious) reflection to contribute to the detected intensity. Whether these low-wavelength features are real, should be clarified with the improved setup.
The shielding simulations up to now have been done with the R1 beam tube of the reactor. There’s a possibility to use the L1 beam line, which should have similar performance. Now that we have generated a more realistic model of those beam tubes, it was time to run some simulations with it. What can be seen in the figure here is that the thermal neutron intensity in the two beam lines is indeed approximately the same, but the epi-thermal and fast neutron part of the spectrum is considerably lower! In short, due to the core configuration, the L1 beam line acts some-what as a tangential beam line, which is precisely what we need for the diffractometer. This will improve our signal/noise ratio significantly. More on that soon.
We have done some more testing on the scintillator detectors developed by ISIS, UK these last few days. The aim was a “stress-test for the scintillator bank”, something quite popular these days. The previous tests (April) showed that the efficiency of the detector was some 80% of a standard 3-He detector (used to be approx 20% before these developments). We then also tried to determine the gamma sensitivity of the scintillator detector, which is/was ‘the other concern’ compared to 3-He detectors. We never seemed to manage to produce enough gammas to actually quantify that number.
This time we had the help of our radio protection service to provide us with some gamma sources that we would use to illuminate the detector while it was counting neutrons. Due to some technical and regulatory restrictions we have only determined the gamma sensitivity to be well below 1e-5 and probably below 1e-6. The reason for that is that you need so many gammas to get the damn detector out of balance. Eventually, taking the detector out of the neutron beam and putting the gamma source right in front of the detection surface, we got a hint of some ‘double counting’. The gamma sensitivity (in the no-neutron case) is approx 1e-8.
There’s still some analysis going on, but with these results so far, we can already conclude that this ISIS detector is very good alternative to an expensive 3-He detector! We now aim to have (a part of) the final detector ready for testing on our beam line within a year.
Thanks to Jeroen for the initiative! Thanks to the ISIS team for the collaboration!
The scintillator detector as designed by the detector group of ISIS, UK has proven to be a valuable alternative to the ‘classical’ 3He detector. This was shown by the tests we did (2 May 2011) at our TOF reflectometer. As mentioned in that post, the gamma discrimination abilities need to be tested more thoroughly. We foresee those tests in June, where we will try to disturb the registration of TOF neutrons by the detector, by simultaneously overloading it with gammas from an external source.
The ISIS detector group has visited our facility to perform some tests with us on their new scintillator detectors. Scintillator detectors are usually more gamma sensitive and less neutron sensitive than the standard 3-He detectors. But since the enormous increase in the price of 3-He, more efforts have been put into the development of scintillator based detectors. The ISIS team came to the RID for tests and demonstration of clear-fibre and wavelength-shifting fibre detectors and a whole set of electronics to evaluate the output of these detectors.
During a three-day test on our reflectometer, we have compared these detectors with the standard 3-He detector of the reflectometer. The preliminary results of these tests show that the neutron detection efficiency is some 70-80% of this 3-He detector and the gamma sensitivity (without optimization of the pulse analysis of the scintillator electronics) is between 20 and 50% higher than the 3-He tube. More involved tests and optimizations need to be done for the gamma-sensitivity. A more severe testing would be counting neutrons in a strong(er) gamma background.
The photo above shows the installation of one of the detectors (the aluminium box) inside the sample-chamber of our neutron-reflectometer at RID. The neutron beam impinges from the left on the scintillators inside the box (not visible) and the light generated by these scintillator crystals is amplified by the photon-multiplier-tubes (PMTs) which are read out by the electronics (not visible). The pulse-shape- and coincidence analysis of these electronics determine whether a neutron or a gamma is detected. The detectors were tested on the reflectometer, so that we could use the time-of-flight option and diafragms of this instrument.
Posted in general, specs, timeline
Tagged detector, diffraction, isis, neutron, npm2, reflectometer, rog, rrr, scattering, scintillator, tests, tnw, tudelft
The first components are in: perfect single crystals lent to us kindly by the ILL. The Copper ones are made in-house at the ILL and the Germanium ones are available commercially. But then, they are so perfect that they reflect too few neutrons, so the trick is to make them less perfect. At the ILL they are pressed at high temperature to introduce a controlled ‘micro-cracking’. Like that the whole block will consist of many small perfect crystallites that have slight misorientations with respect to the original orientation. This mosaic spread increases the intensity by allowing a larger wavelength band to be reflected. Unfortunately this also increases the divergence of the reflected beam. Typical mosaic spreads are less than a half a degree, although for instance with graphite crystals you can have 2 or 3 degrees.
Posted in general, specs
Tagged components, crystals, cu, diffraction, ge, ill, neutron, npm2, rrr, scattering, tests, tnw, tudelft
With the detailed planning of the development project of the neutron diffractometer, we also need to know more precisely what the components will cost.
The two most demanding items on the list of instrument components are the neutron detector that needs a high spatial resolution and a monochromator with the right crystalite mosaic.
In a short visit to the ILL we have discussed the possibilities to develop a detector and a monochromator for our needs. On our side, we now need to see what part of that development can be done in-house at the university. This should then clarify how much budget we have to reserve for these components. The collaboration on the development has to be formally signed by both institutes.
Posted in general, specs, timeline
Tagged components, crystals, cu, diffraction, ge, ill, neutron, npm2, rrr, scattering, shopping, tests, tudelft