Velocity Selector OperationAn Astrium higher-order wavelength velocity selector has been installed in the reactor beam on BT-7. This velocity selector is designed to efficienty transmit the primary wavelength neutrons while suppressing second order, third order, ... neutrons so that the incident beam from the PG or Cu crystals onto the sample is monochromatic. The suppression of the higher orders not only produces a clearner beam onto the sample but also reduces the activation of the sample.
For diffraction experiments, the best choice is still to use 13.5--14.8 meV neutrons combined with PG filters to suppress higher order wavelength neutrons. These higher orders can be suppressed to any desired level by adding additional PG filter material. For inelastic measurements where variable incident energies are needed, a PG filter which works only at selected energies cannot be used, and this is where the velocity selector is best used. The measurements below demonstrate that the velocity selector reduces λ/2 by than an order of magnitude over an energy range from 11.5 meV to 62.5 meV, suppresses third and higher-orders to negligible values, and reduces background similar to what is obtained using the PG filter in the reactor beam.
A new relatively clean option is to use the velocity selector with PG(004), which greatly increases the energy resolution of the monochromator system. This option is available from an incident energy of about 30.5 meV to 62.5 meV.
Some detailed information follows. ACNS 2018 Velocity Selector Talk (pdf)
Picutre of velocity selector installed in the reactor beam, with the side and top
shielding removed. The 'box' on the right contains the computer controlled
variable apertures, and a 3 cm thick PG filter can be remotely moved into or our
of the beam depending on the experimental configuration required.
Neutron flux ratio for 2nd order wavelength vs. primary wavelength.
There is no significant difference by replacing the 25′ collimation with open -50′ collimation.
With the velocity selector and vertically focused monochromator there is still some λ/2 in the beam. For most situations there is no need to correct the inelastic scattering, but the above data are provided if needed. Corrections for λ/3, λ/4, ... are negligible. Please note that the monitor correction factor provided in DAVE is much larger and is not to be used if the velocity selector is employed.
If the monochromator is vertically flat, there are no corrections needed.
Silicon Diffraction pattern for 14.7 meV (λ = 2.359 Å) under three different conditions. The broad distribution of scattering is from the glass container holding the Si powder:
1) no PG filter in the reactor beam (blue);
2) with the velocity selector translated into the reactor beam (orange),
3) Using the PG filter in the reactor beam (green). At this energy the PG filter works better than the velocity selector.
At higher energies (e.g. 28 meV, 30.5 meV, 35 meV) where the PG filter can be used, the velocity selector performance is superior to the PG filter as indicaed below for Si powder data taken at 28 meV. This is because for PG the transmission of the primary wavelength is substantially reduced and the rejection of higher orders is not as good.
Diffraction scan at 14.7 meV using a Ge(111) single crystal at the sample position. The greatly enhanced sensitvity compared to the Si powder reveals contaminations from aluminum at the sample position. Employing the velocity selector reduces or elimanates higher-order wavelength contaminations, and also greatly reduces the overall background.
Here are the very first inelastic data, taken with and without the velocity selector for comparison. The sample was a 0.15g BiFeO3 single crystal. Söller collimators employed were open-80′-s-80′-120′ and a 58 mm thick PG filter after the sample. Data are normalized to counts/minute. For the 3.5 meV data the background was reduced by 46% while the signal was lowered by 31% with the velocity selector. For the 7 meV data a large contamination is evident due to the Al(111) peak from the sample holder in λ/2, which was completely removed by the velocity selector. The dashed line is an estimate of the actual intensity of the magnon peak.
The figure above shows scans of the velocity selector rotation speed for a few incident energies. The top is for PG (002) and the bottom is for PG (004). The solid black curves are Gaussian fits, with the parameters in the legends
Monitor rate comparison of PG (002) monochromator through the usable Ei range.
Monitor counts versus either energy (left) or wavelength (right) plotted for the velocity selector at various fixed speeds, chosen to correspond to the energies at 14.7, 28, 30.5, and 41 meV where PG filters are commonly used..
Full size beam (Aperture: W: 86* H: 160)
Transmission for λ (%)
Transmission for λ/2 (%)
Table of transmission for primary and 2nd order wavelength neutrons using PG(002). For high energy (Ei > 41 meV), the 2nd order peaks from Si powder are too small to give a reliable fit, and we used the data from the Ge single Crystal to determint the transmission for 2nd order.
For the velocity selector using PG (004) as the primary wavelength, the contaminations are (mostly) from PG (002) and PG (006), which correspond to 2* λ and 2/3* λ. For Ei = 40, for example, these correspond to 10 meV and 90 meV neutrons, respectively. Then the transmission for λ was measured to be 63.6 %, while the transmissions for 2* λ and 2/3 * λ are 2.29 % and 29.1 %, respectively. The flux ratios of 2 * λ and 2/3 * λ compare to λ are then 1.1 % and 12 %.
When Ei = 46 meV, the above numbers change to; transmission for λ: 64.6 %, 2* λ: 1.16 %, and 2/3 * λ: 20.2 %. Flux ratio of 2* λ: 1.13 %, 2/3* λ: 6.77 %
Vertically Flat Monochromator (PG002)
Almost all data collected on BT-7 are taken with a vertically focused monochromator (Vf sagittal). However, rejection of λ/2 is much better with restricted vertical divergence. The figure below compares transmission measurements of the primary (λ) and the 2nd order (λ/2) neutrons with different vertical focusing modes (flat vs sagittal) using PG(002). The results demonstrate that restricting the angular divergence in the vertical diretion greatly reduces the 2nd order neutron (λ/2) transmissions (at the expense of a reduction in the primary intensity) to less than 0.5 % for both Ei = 14.7 and 28 meV ,compared to the vertical sagittal focus mode (~ 7 %). The transmission for primary neutrons (λ) is also improved somewhat, from ~ 65 % (sagittal) to ~77 % (flat).
Figure. Velocity selector transmission measurements using Si powder at Ei = 14.7 meV (top) and 28 meV (bottom), showing the effect of restricting the vertical divergence on the performance of the velocity selector.
The transmission rate for different modes are listed below.
Ei (meV) (open-10´-80´-DD)
Transmission for λ(%)
Transmission for λ/2(%)
Last modified 28-July-2018 by website owner: NCNR (attn: Jeff Lynn)