[Fast-Neutrons] Fast-Neutrons Digest, Vol 6, Issue 1

Robert Adams adams at lke.mavt.ethz.ch
Fri Dec 13 13:08:06 CET 2019

Hello Michael,

I can offer my thoughts on this topic, for whatever they are worth:

To evaluate the resolution of your radiography system, I don't think a
Siemens star makes much sense because it should be reasonably thick (unlike
e.g. thermal neutrons or X-rays, where a very thin structure can be used
and still obtain a decent attenuation), which means the center of the star
will have some rounded edges depending on how it's manufactured... e.g. if
you make a Siemens star out of some metal or plastic with waterjet,
machining, wire erosion, etc., always the cut-out will have some curvature
on the order of 0.5-1 mm at the inner "corner", on the same order as what I
guess the system will achieve, so it will not give valuable information.
One way around this might be cutting wedges individually, then placing them
together, but then alignment is tricky. There might be ways around this to
get something satisfactory, but in the end I don't see much value above a
good old-fashioned edge measurement. A line pair structure is also a
possibility, but like the wedge probably cannot be satisfactorily machined
monolithically, and does not add a lot of value to an edge in this case in
my opinion.

For an edge, it should only be carefully machined to have properly
parallel/perpendicular faces, but that's pretty easy. Just make sure the
machinist doesn't break or chamfer the corners significantly. Machinable
tungsten alloys are readily available and are a good choice for attenuating
fast neutrons. Could do alu or steel as well to get some variety. Plastic
is also fine but sometimes more annoying to get precise, clean edges and
corners compared to metal. I think on the order of 1 cm or so is
sufficiently thick to get an easily measured amount of attenuation in the
image, without suffering too much from some kind of penumbra blur effect.
To be sure that this penumbra is not affecting you, you can do several
thicknesses (e.g. 5, 10, 20 mm) and check that they agree. The
thickness-direction edge should of course be as parallel as possible with
the beam, and it should be placed close to the detector surface if you want
to characterize only the detector and minimize effects from the source
emitting spot size. You can change the distance to be further away from the
detector surface, and then the difference in result will indicate the
effect of your source emitting spot or collimation or whatever. More
effort, but to check the penumbra effect in more detail you can also put it
on a little rotary stage and take a series of measurements from what you
consider "edge aligned", and +/- few degrees in small steps, then compare
results. If the system is pixelated, the edge should not run palallel to
the pixel edge direction, as that can potentially make the measurement
suggest a resolution better than it actually is, i.e. put it at 10 degrees
from vertical or something like that.

For processing you can get each grayscale value as a function of distance
from the edge (defined by a manually selected line segment), for all pixels
along some length of the edge (e.g. couple cm) and some distance away, then
put that all together on one plot. Then make, for example, an error
function fit to that data. Then take the derivative of that fit to get a
line spread function, and from that the FWHM is an indication of
resolution. Or, process the data a bit further and get an MTF. Or, of
course, the edge data could be processed many other ways.

If you want you can to send me image data and I can try apply some scripts
I already have to it to give you an example of some of what I'm describing.
I expect with some simple edge measurements you can get a robust idea of
how your imaging system is performing.

Best regards,

Robert Adams, Dr. sc.
ETH Zürich
Sonneggstrasse 3
8092 Zürich, Switzerland
Office: +41 (0) 44 632 49 01
Mobile: +41 (0) 78 611 39 39

On Fri, Dec 13, 2019 at 12:01 PM <fast-neutrons-request at neutronsources.org>

> Send Fast-Neutrons mailing list submissions to
>         fast-neutrons at neutronsources.org
> To subscribe or unsubscribe via the World Wide Web, visit
>         https://neutronsources.org/mailman/listinfo/fast-neutrons
> or, via email, send a message with subject or body 'help' to
>         fast-neutrons-request at neutronsources.org
> You can reach the person managing the list at
>         fast-neutrons-owner at neutronsources.org
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of Fast-Neutrons digest..."
> Today's Topics:
>    1. Re: Congrats! Re: First fast neutron      images  at      Phoenix
>       (Michael Taylor)
> ----------------------------------------------------------------------
> Message: 1
> Date: Thu, 12 Dec 2019 20:49:30 +0000
> From: Michael Taylor <Michael.Taylor at phoenixwi.com>
> To: "Lehmann Eberhard (PSI)" <eberhard.lehmann at psi.ch>, Alan
>         <alan.hewat at gmail.com>, "fast-neutrons at neutronsources.org"
>         <fast-neutrons at neutronsources.org>
> Subject: Re: [Fast-Neutrons] Congrats! Re: First fast neutron   images
>         at      Phoenix
> Message-ID:
>         <
> SN6PR04MB4477696A12CFD23A4EC47D2892550 at SN6PR04MB4477.namprd04.prod.outlook.com
> >
> Content-Type: text/plain; charset="windows-1252"
> Hello all,
> Does anyone already have a printed Siemens star they'd be willing to ship
> to me for testing?  Has anyone come up with another method for measuring
> resolution for fast neutron imaging?  We are otherwise going to image an
> edge on different pieces of material: Fe, W, Cu, Al, HDPE to see how they
> compare.
> Best regards,
> Michael
> ________________________________
> From: fast-neutrons-bounces at neutronsources.org <
> fast-neutrons-bounces at neutronsources.org> on behalf of Lehmann Eberhard
> (PSI) <eberhard.lehmann at psi.ch>
> Sent: Monday, November 18, 2019 4:52 AM
> To: Alan <alan.hewat at gmail.com>; fast-neutrons at neutronsources.org <
> fast-neutrons at neutronsources.org>
> Subject: Re: [Fast-Neutrons] Congrats! Re: First fast neutron images at
> Phoenix
> I disagree with Alan about the n-sensitive imaging plates: they are
> made by Fuji on a commercial base adding Gd to the X-ray sensitive
> material. Therefore the efficiency is not bad and much better than
> film methods ?
> Regards
> Eberhard
> __________________________________________
> Paul Scherrer Institut
> Dr. Eberhard H. Lehmann
> WBBA/122
> Forschungsstrasse 111
> 5232 Villigen PSI
> Schweiz
> Telefon: +41 56 310 29 63
> E-Mail: eberhard.lehmann at psi.ch
> Von: fast-neutrons-bounces at neutronsources.org <
> fast-neutrons-bounces at neutronsources.org> Im Auftrag von Alan
> Gesendet: Montag, 18. November 2019 10:43
> An: fast-neutrons at neutronsources.org
> Betreff: Re: [Fast-Neutrons] Congrats! Re: First fast neutron images at
> Phoenix
> Hello Michael, Burkhard and Eberhard.
> Thank you Michael for these nice images. I too am impressed by the
> resolution of your thermal images with an L/D of only 35, and agree that it
> would be good to take thermal images with a camera rather than a neutron
> image plate. The ordinary Fuji n-plates are just x-ray plates with extra
> 6LiF and I think only ~10% efficient. ~20 years ago Fuji made special
> n-plates for ILL that were ~25% efficient but they are no longer available
> (?). You should be able to reduce the exposure well below 20 minutes with a
> camera and a good 6LiF/ZnS thermal neutron scintillator even with only
> 10**4 n.cm-2.s-1. See the images Robert Zboray showed in Sydney from a very
> low flux Triga reactor, one of which I reproduced in my Munich talk.
> It would also be good to compare your fast neutron image with a thermal
> neutron image using the same detector and L/D (with different
> scintillators). With such small objects, that can be put close to the
> scintillator, it would be interesting to see if fast neutrons still have
> some advantage for imaging such defects.
> I also found Burkard's and Eberhard's comments about the best material and
> thickness for a fast neutron Siemens star interesting. It would be good to
> see images with these different resolution objects. I also printed a
> plastic Siemens star, but 40 mm thick, which I have not yet been able to
> test. A final trivial point; please don't use exponent e4 instead of 10**4
> for flux; e has a different meaning for mathematicians.
> Thanks again for sharing. Alan
> ______________________________________________________
>    Dr Alan Hewat, NeutronOptics, Grenoble, FRANCE
>                                from my telephone
> <Alan.Hewat at NeutronOptics.com<mailto:Alan.Hewat at NeutronOptics.com>>
> +33.476.98.41.68
>         http://www.NeutronOptics.com/hewat<
> https://linkprotect.cudasvc.com/url?a=http%3a%2f%2fwww.NeutronOptics.com%2fhewat&c=E,1,nsRpj7Ep0HI2dnJQS9mLWP7NTTRFa_A1rnHdA_hwLcr1LAK7nQ2-Td0N6b_DO9Ja1638KxTpmGVeliGNkSNhx7t1hvZbcROb-wXGKiiSSd4b_OlHxdBS0EA,&typo=1
> >
> ______________________________________________________
> On Mon, 18 Nov 2019, 08:42 Burkhard Schillinger, <
> Burkhard.Schillinger at frm2.tum.de<mailto:Burkhard.Schillinger at frm2.tum.de>>
> wrote:
> Hello Michael,
> congratulations for these nice images!
> For a fast neutron image, that resolution is pretty good. Probably
> better than ours - which shows that a high collimation is also
> important for fast neutrons if you have large samples.
> I am also surprised again by your thermal image at L/D of only 35 -
> but that's what you get when you can put the samples up close to the
> detector.
> For an edge - about everything scatters fast neutrons, but I have
> tried with a 10 mm thick polyethylen pattern (Siemens Star) that was
> 3D printed at our lab. When putting it directly on the detector, it
> was a good measure to test the screen. The 1.5 mm thickness screen was
> much better than the 2.4 mm.
> Not sure what really happens if you take it further away - I assume
> that 10 mm thickness is a good compromise between attenuation and
> scattering blur. You might try 10 mm steel as well, but iron is also a
> diffuse scatterer.
> A perfect edge does not exist.
> Good luck with your new 'toys', and Happy Holidays!
> Burkhard
> > Good day to all,
> >
> > I am happy to announce that we've taken our first fast neutron
> > images at the Phoenix facility in Wisconsin, USA!  Our source is
> > operational now with a source strength of approximately 1.5e12 n/s
> > and an L/D of 450 to achieve a flux at the image plane of
> > approximately 5e5 n/cm^2-s.  Over the next few weeks, we will be
> > increasing our beam current and changing our target to a different
> > material.  We expect to get to full power and have a source strength
> > of 3e13 n/s and a flux at the image plane of approximately 1e7
> > n/cm^2-s.
> >
> > The image attached was taken using a Varex XRD 1621 digital detector
> > array and a PP:ZnS(Cu) screen provided by RCTritec.  The
> > scintillator field of view is 310mm x 310mm, but we plan to use the
> > full field of view of the detector eventually, which is 430mm x
> > 430mm.  The image was taken last night and acquired with 15 frames
> > at 20 seconds each.  The frames were then added and the offsets were
> > applied for background corrections.  The sample is a simulated M982
> > military round.  It is 155mm in diameter.  The outside casing is
> > 1/8" steel and the inside simulant is an HMX equivalent, 6% 6656
> > binder (simulated with 204 epoxy). Chemically it is similar to HMX
> > and RDX but with much less nitrogen.  I have outlined some of the
> > defects of interest that we want to see.
> >
> > We would like to measure the resolution of the system next so I
> > would like to ask if anyone has advice on what material to use as an
> > edge and how thick it should be?  We do not yet have any kind of
> > standard measurement technique for this that I'm aware of, so I
> > would like to know what others are doing.
> >
> > I've also attached a new image we acquired using thermal neutrons of
> > several different military grade .50 caliber ammunition.  We took an
> > X-ray image for comparison and that is shown as well.  The X-ray was
> > done at 350kV but I don't recall the current.  The neutron image was
> > acquired using our thermal neutron generator with heavy water
> > moderator, the L/D was 35 and the flux was approximately 1e4
> > n/cm^2-s.  The exposure time on a neutron sensitive image plate was
> > 20 minutes.
> >
> > I hope everyone is well, it was great to meet so many of you in
> > Garching last month and I wish you all happy holidays as they
> > approach us soon.
> >
> > Best regards,
> > Michael
> >
> > Michael Taylor Ph.D.
> > Neutron Radiography Product Manager
> > Phoenix LLC
> > 2555 Industrial Drive
> > Madison, WI 53717
> > 608-515-3214
> --
> Dr. Burkhard Schillinger
> Technische Universit?t M?nchen - FRM II
> Heinz Maier-Leibnitz Zentrum
> Lichtenbergstr.1
> D-85748 Garching
> Germany
> Tel. +49 89 289-12185
> _______________________________________________
> Fast-Neutrons mailing list
> Fast-Neutrons at neutronsources.org<mailto:Fast-Neutrons at neutronsources.org>
> https://neutronsources.org/mailman/listinfo/fast-neutrons<
> https://linkprotect.cudasvc.com/url?a=https%3a%2f%2fneutronsources.org%2fmailman%2flistinfo%2ffast-neutrons&c=E,1,jwqsB9TTvJz9wzdKEM4CT4zoMezOjLtJUkW95eXQwCfEiAMGKg-OSABSGQ40Ckvu2HD70xqAjFLAhtENigaRLx5bUX2S57JzCNpaiqBOUfbronU,&typo=1
> >
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL: <
> https://neutronsources.org/pipermail/fast-neutrons/attachments/20191212/368a72a1/attachment-0001.html
> >
> ------------------------------
> _______________________________________________
> Fast-Neutrons mailing list
> Fast-Neutrons at neutronsources.org
> https://neutronsources.org/mailman/listinfo/fast-neutrons
> End of Fast-Neutrons Digest, Vol 6, Issue 1
> *******************************************
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.neutronsources.org/pipermail/fast-neutrons/attachments/20191213/ab4cdb77/attachment.htm>

More information about the Fast-Neutrons mailing list