362 IEEE TRANSACTIONS ON NUCLEAR SCIENCE, VOL. 54, NO. 2, APRIL 2007
Neutron Collimation With Microchannel Plates:
Calibration of Existing Technology and
Near Future Possibilities
Anton S. Tremsin, Daniel S. Hussey, R. Gregory Downing, W. Bruce Feller, David F. R. Mildner,
David L. Jacobson, Muhammad Arif, and Oswald H. W. Siegmund
Abstract—A new type of high performance and compact neu-
tron collimator can be manufactured from Gd- or B-doped mi-
crochannel plates (MCPs). Structures only a few mm thick have
very narrow rocking curves and high out-of-angle rejection ratios,
as observed previously with a cold neutron beam. We present the
results of measurements with a collimated (L/D ratio )
thermal neutron beam. MCP collimators doped with 3 mole % of
Gd O as well as doped with 10 mole % of BO were cal-
ibrated for transmission versus tilt angle. The MCPs used in this
study were only 0.6 and 0.8 mm thick with m circular pores
on 11.5 m centers. All the measured rocking curves agree well
with the theoretically predicted performance. Both experimental
and modeling results indicate that very efficient MCP collimators
(with wide rocking curves and a rejection ratio exceeding
) can be built with the existing technology.
The possibility to manufacture collimators with very large L/D
ratios exceeding 1000:1 is also discussed for the case of unetched
MCPs. The peak transmission of such devices with very sharp
rocking curves will be limited to % by the transmission of
the undoped glass. Application of MCP collimators for scatter
rejection in neutron radiography is also considered in terms of
possible image distortions, which are shown to occur only for
the systems with detector spatial resolution better than 20 m
FWHM.
Index Terms—Collimators, neutron optics, neutron scattering.
I. INTRODUCTION
I
N most neutron scattering experiments the angular spread of
the neutron beam is defined by the quality of the neutron col-
limator. Soller slit collimators, comprising an array of absorbing
films (e.g., Gd layer) separated by neutron-transparent spacers
(e.g., Si wafers) [1], [2] as well as honeycomb-like packed struc-
tures [3] are widely used to shape the neutron beam. In order to
achieve a high degree of collimation the thickness of the spacers
should be optimized relative to the length of the structure and the
Manuscript received June 1, 2006; revised December 18, 2006.
A. S. Tremsin and O. H. W. Siegmund are with the Space Sciences Labo-
ratory, University of California, Berkeley, Berkeley, CA 94720 USA (e-mail:
ast@ssl.berkeley.edu; ossy@ssl.berkeley.edu).
D. S. Hussey, R. G. Downing, D. F. R. Mildner, D. L. Jacobson. and M.
Arif are with the National Institute of Standards and Technology, Gaithersburg,
MD 20899 USA (e-mail: daniel.hussey@nist.gov; gregory.downing@nist.gov;
david.mildner@nist.gov; david.jacobson@nist.gov; muhammad.arif@nist.
gov).
W. B. Feller is with NOVA Scientific Inc., Sturbridge, MA 01566 USA
(e-mail: bfeller@novascientific.com).
Color versions of one or more of the figures in this paper are available online
at http://ieeexplore.ieee.org.
Digital Object Identifier 10.1109/TNS.2007.891080
thickness of absorbing coatings. In other words the L/D aspect
ratio of the collimating structure (the ratio of its length to the
thickness of transmitting channel) has to be large for efficient
collimators.
Compact collimators currently produced use stacked arrays
of thin single-crystal silicon wafers ( m thick) with ab-
sorbing Gd coatings. Our theoretical predictions supported by
preliminary experiments with cold neutrons indicate that MCP
collimators can have substantially better efficiencies than ex-
isting structures [4], [5]. Current MCP technology allows for
manufacturing of the structures with length to diameter ratios
of up to 250:1 with active areas up to 10 10 cm . The diam-
eter of MCP pores can be m, making these collimators
very compact. Moreover, the geometry of the pores can be con-
trolled in order to achieve the optimal collimation in the two di-
rections separately. In addition to beam shaping, these compact
structures can be used for the rejection of scattered neutrons,
which degrade the quality of images in high resolution neutron
radiography.
Recently, modification of MCP base glass materials through
incorporation of high neutron cross-section elements such as
boron and gadolinium, has resulted in MCPs capable of direct
imaging detection [6]–[9] as well as collimation of neutrons
[4], [5]. The glass mixture in these novel MCPs contains neu-
tron absorbing gadolinium or boron atoms, while the rest of the
manufacturing procedures remain the same. Neutron sensing
microchannel plates are currently being produced for high-res-
olution neutron counting (with spatial resolution as small as
m and temporal resolution of s). Our theoret-
ical studies indicate that these detectors should indeed have very
high detection efficiency for thermal and cold neutrons [10] and
achieve high spatial resolution. The spatial resolution of such
detectors for UV and X-ray photons has already reached the 10
m level [11], and we believe the progress in MCP photon de-
tection can be extended to neutron imaging with spatial resolu-
tion of m and below.
Our first experimental measurements of Gd-doped MCP col-
limation performed with a cold neutron beam, proved the ef-
ficiency of these compact structures [4]. The results of these
measurements also support the validity of our theoretical effi-
ciency calculations, which can be used to predict the ultimate
performance of MCP collimators using existing manufacturing
technology. In this paper we extend our experimental studies of
MCP collimators to a thermal neutron beam and compare the
results of these measurements with predicted performance. We
0018-9499/$25.00 © 2007 IEEE