Neutron Spin Filters Based on Polarized Helium 3

Spin-polarized neutrons are useful probes of magnetic matter since both the magnetic scattering of neutrons by unpaired electrons and the scattering of neutrons by nuclei of non-zero spin can be strong functions of the neutron spin state (Refer to reference 1). Unfortunately, the limitations of current polarizing and analyzing devices have substantially restricted the application of this powerful technique. Helium 3 spin-filters have the potential to yield broadband neutron polarizers and analyzers that can be used for cold, thermal, and epithermal neutrons. Such spin filters could make new classes of neutron scattering experiments possible.

For certain applications there are presently no suitable devices. For example, polarization analysis of scattered beams having wide angular divergence is impractical with reflection based devices such as super mirrors because of their inadequate angular acceptance and small angle scattering. At present, we are developing Helium 3 based polarization analyzers for diffuse reflectometry and small angle scattering (S A N S) with neutrons, both of which require analysis of a large divergence beam. Helium 3-based neutron spin filters would also be advantageous on crystal spectrometers since any monochromator or analyzer crystal can be used, rather than just Heusler alloy, greatly increasing the range and flexibility of the instrumentation. Furthermore, polarizing or analyzing thermal and hot neutrons at spallation sources, where the time of flight method is employed, often requires the broadband capability of Helium 3 spin filters. In addition to their utility for neutron scattering, polarized Helium 3 spin filters are also of interest for nuclear and particle physics studies with neutrons, such as determination of weak coupling constants and tests of the Standard Model via accurate measurements of decay correlation coefficients in polarized neutron beta decay (Refer to reference 2).

Neutron spin filters based on polarized Helium 3 rely on the strong spin-dependence of the neutron capture cross section for Helium 3. For a sufficient thickness of 100 percent polarized Helium 3 gas, essentially all of the neutrons with antiparallel spin would be absorbed, while nearly all of the neutrons with parallel spin would be transmitted, resulting in 100 percent neutron polarization and 50 percent transmission. In Figure 1 we show calculated values of the neutron polarization P sub n, and neutron transmission T sub n, for Helium 3 polarization P sub H e = 60 percent, an experimentally achievable value. Since there is a tradeoff between neutron polarization and transmission, we also show P sub n, to the second, times T sub n, which is a useful figure of merit for many experiments. A transmission analyzer can be characterized either by the transmission asymmetry A, or the flipping ratio F. For T sub plus, and T sub minus, defined to be the transmissions for neutrons with polarization parallel and antiparallel to the Helium 3 polarization, respectively, the asymmetry is given by A = open parenthesis, T sub plus, minus T sub minus, close parenthesis, over, open parenthesis T sub plus, plus T sub minus, close parenthesis, and the flipping ratio by F = T sub plus, over T sub minus, = open parenthesis 1 plus A, close parenthesis, over open parenthesis, 1 minus A, close parenthesis. The flipping ratio F is also shown in Figure 1. The asymmetry for a spin filter used as an analyzer is the same as the neutron polarization produced when it is used as a polarizer. For the specific case of cold neutrons (lambda = 0.5 nano meters), a spin filter used as a polarizer with P sub H e  = 60 percent and a pressure-length product of 7 kilo pascal meters would yield P sub n = 90 percent and T sub n = 20 percent, or, when used as an analyzer, would give A = 90 percent and F = 19.
Neutron polarization graph Neutron polarization P sub n, (solid thick line), transmission T sub n, (solid thin line), a figure of merit given by P sub n, to the second, times T sub n, (dashed line), and the flipping ratio F, (dotted line) as a function of the product of Helium 3 pressure in Pascals, (assuming a room temperature cell), cell length in meters and wavelength in nano meters. The calculations are shown for Helium 3 polarization P sub H e = 60 percent. The scale on the right hand y axis is used for the flipping ratio.

We produce polarized Helium 3 gas by two optical pumping methods, spin exchange, which is performed directly at high pressure (0.1 mega pascals to 0.3 mega pascals), and meta stability exchange, in which the gas is polarized at low pressure (approximately equal to 100 pascals) and then compressed (Refer to reference 3). The spin-exchange method is convenient and well matched to continuous operation on a beamline, whereas the metastability-exchange method has a higher polarizing rate. At present, it has been more convenient (for either method) to polarize gas off-line and transport it to the beam line. Maintaining the polarization in the absence of optical pumping requires a homogeneous magnetic field and specially prepared glass cells with slow wall relaxation. We have produced cells with relaxation times as long as one month, dominated by the intrinsic dipole-dipole relaxation in the Helium 3 gas itself (Refer toreference 4). In the future, continuously operating spin filters can be installed directly on the neutron instrument.

Polarization analysis allows for separation of magnetic from nuclear scattering, and also s eparation of coherent from spin-incoherent scattering. In our first demonstration experiment for S A N S, we used a Helium 3 spin filter to extract a small component of spin-incoherent scattering in the presence of strong coherent scattering (Refer to reference 5). We have continued with tests of separating magnetic from nuclear scattering.

Recently we carried out successful tests of a polarized Helium 3 spin filter on the N C N R, N G 1 reflectometer. The first test experiment on a specular reflection was a careful comparison of results obtained with a Helium 3 analyzer to those obtained with the current technique that employs a super mirror analyzer. The test sample was an epitaxial manganese, sub 0.52, palladium, sub 0.48, slash iron, open parenthesis 0 0 1, close parenthesis, by layer, for which chemical ordering and magnetic exchange bias have recently been reported (Refer to reference 6). By varying the magnetic field on this sample, it was possible to test the Helium 3 analyzer under conditions of both significant and negligible spin flip scattering. We employed a compact, magnetically shielded solenoid that was interchanged with the analyzing supermirror on the N G 1 instrument. The solenoid adequately shielded the stray fields from the 0.6 tesla magnet and guide field magnets. The results obtained with the Helium 3 analyzer were identical to those obtained with the super mirror analyzer.

For efficient studies of diffuse scattering on a reflectometer, the Helium 3 analyzer will be combined with a position sensitive detector (P S D). Two issues not tested in the specular experiment are relevant in this case: (1) possible depolarization of neutrons that follow trajectories off the axis of the Helium 3 analyzer, and (2) possible background from the Helium 3 cell. We have established that most of the area of apertures in the magnetic shield is usable, with depolarization only occurring for a small range of extreme trajectories. We have also determined that small angle scattering from the cell is negligible and therefore should not pose an issue for the low neutron fluxes expected in diffuse scattering experiments. This series of tests established the suitability of polarized Helium 3 spin filters for diffuse reflectometry experiments, which we will pursue in the near future.

We are making continual improvements in the polarization and relaxation time of our Helium 3 cells. In the recent reflectometry experiments, the initial Helium 3 polarization was 57  percent, resulting in a neutron flipping ratio of 31, and a transmission for the desired spin state of 24 percent. The relaxation time of the polarized gas was 15 days on the beam line, which was somewhat reduced from the intrinsic value of 24 days for the cell due to a small fractional gradient of 3 times 10 to the minus 4, inverse centi meters, in the magnetic field of the shielded solenoid. Our aim is to increase the Helium 3 polarization to 70 percent and obtain relaxation times at the dipole-dipole limit.


[1] R. M. Moon, T. Riste, and W. C. Koehler, Phys. Rev. A 1 8 1, 9 2 0 9 3 1 (1969).

[2] Fundamental Physics with Pulsed Neutron Beams, Research Triangle Park, North Carolina, June 1st through the 3rd, 2000, edited by C. R. Gould, G. L. Greene, F. Plasil, and W. M. Snow (World Scientific, Singapore, 2001).

[3] T. R. Gentile et al., J. Res. N I S T, 1 0 6, 7 0 9 7 2 9 (2001).

[4] D. R. Rich et al., Applied Physics Letters 8 0, 2 2 1 0 (2002).

[5] T. R. Gentile et al., J. Appl. Crystallog. 3 3, 7 7 1 7 7 4 (2000).

[6] R. F. C. Farrow et al., Applied Physics Letters 8 0, 8 0 8 (2002).

T. R. Gentile and A. K. Thompson
National Institute of Standards and Technology
NIST Center for Neutron Research
100 Bureau Drive, M S 8 4 6 1
Gaithersburg, MD 2 0 8 9 9 8 4 6 1

J. A. Borchers, J. W. Lynn, K. V. O’Donovan, and C. F. Majkrzak
National Institute of Standards and Technology
NIST Center for Neutron Research
100 Bureau Drive, M S 8 5 6 2
Gaithersburg, MD 2 0 8 9 9 8 5 6 2

W. M. Snow and W. C. Chen
Indiana University
Department of Physics
Bloomington, IN 4 7 4 0 5

G. L. Jones
Physics Department
Hamilton College
198 College Hill Road
Clinton, New York 13323

Back to FY2002 HTML main page
Go to nextarticle

To view all symbols correctly, please download Internet Explorer 6 or Netscape 7.1

Last modified 28-April-2004