2/6/97


Perhaps someone can explain what I saw earlier today while examining a

piece of archeological glass by SEM. When examining the sample at 10, 5

and 3 kV, I saw a distorted reflection of the secondary electron detector

at the surface of the sample. The sample was attached to an aluminum stub

using double-sided adhesive tape, and was not coated. The SEM is a

Cambridge stereoscan 100. I did not observe any reflected image at higher

kV.

Any thoughts?

James Martin

Williamstown Art Conservation Center

James.S.Martin@williams.edu


James,

You have encountered one of the neater artifacts you can accomplish

with a SEM. What happened was that while imaging at 10 keV you

established a fairly uniform charge on the surface of the glass. Then

when you dropped down in energy, the incident electrons were elastically

reflected by this uniform charge field and bounced backwards without

ever hitting the specimen. In other words, the uniform charge field

created a very nice "mirror" which reflected the electron beam towards

your secondary detector so that's what you got an image of. It is also

very typical to get a nice image of your final lens pole piece. The

field created by this type of charging tends to be hemispherical in

shape, so you get a kind of "fisheye" lens effect which at low mag

allows you to get a "panoramic" image of the inside of your specimen

chamber.

It is very easy to duplicate this phenomenon. I have found that a piece

of smooth polysterene works quite nicely (I used a divider from a

plastic parts bin). Simply charge the surface by scanning at low mag

for a while at a high voltage and fairly high spot size and then switch

to a lower keV and you will see the mirroring effect. A saphire bead

also works very well. The better the insulator, the longer the effect

lasts. You can actually get very nice images of the inards of your SEM.

Take a few pictures and see if your microscopist friends can figure out

how you did this!

This is really quite a fascinating effect which can really "blow your

mind" if you stumble across it accidentally without knowing that such a

thing is possible. This is such a neat effect that it just seems that

there SHOULD be some good use for it -- Alas -- I don't know of any,

other than to amuse yourself and your friends.

Fred Schamber

fhscham@sgi.net


Some of the jolly service guys from Philips have performed a similar trick

with uncoated styrofoam without knowing why it worked. They first

bombarded the foam at 25 kV, then turned the acc. voltage down to

3kV or so. Here's my interpretation:

The glass, or styrofoam or whatever, charges up with electrons due to lack

of grounding. As more primary electron bombard the sample they begin to

be repelled by the like charge that has built up within the sample.

When you use low kV the primary electrons are not able to penetrate

the cloud of electrons around your sample, and are repelled by it. These

electrons begin to hit the detector (what you saw), the final

lens (what I saw with the styrofoam), or whatever else in the chamber,

eliciting secondary electrons, and forming an image. What you see

probably depends on the geometry of the chamber. We were able to look

right up the final lens and see the aperture. At higher kVs you

don't see the effect. Kinda cool, huh? Take a picture. Enjoy. Coat the

sample and remove the coating after viewing.

I thunk this up myself - does this sound right?

Aloha,

Tina

http://www.pbrc.hawaii.edu/bemf/microangela

Tina (Weatherby) Carvalho tina@pbrc.hawaii.edu

Biological Electron Microscope Facility (808) 956-6251

University of Hawaii at Manoa http://www.pbrc.hawaii.edu/bemf


I've seen the same when trying to image a minute marine snail that was

poorly adfixed to the stub. I presumed that it was charging SO MUCH

that the Primary Electrons were being completely repelled from the

specimen back to the roof of the chamber. It looked great - a normal

background with a shell-shaped "mirror" showing the ceiling of the

chamber!

Geoff Avern

Microscopy Labs

Australian Museum

Sydney, Australia

geoffa@amsg.austmus.gov.au


Hello, fellow microscopists!

Fred Schamber and Tina Carvalho described a little gizmo which we sell.

It's a lucite sphere mounted on carbon which will produce a reflected image.

We call it a "Lumisphere".

Best regards,

Steven E. Slap, Vice-President

Energy Beam Sciences, Inc.

Adding Brilliance To Your Vision

ebs@ebsciences.com

http://www.ebsciences.com/


Fred et al. Actually, this effect does have one

useful application: if you suspect something wrong

inside the chamber, such as a loose wire, or a cold

stage tubing gone amuck, you can check it out this

way without having to open up the chamber to atmos.

That's the only use I've found; any others?

Damian Neuberger

Baxter International

neuberd@baxter.com


I've seen a couple of responses attributing this effect to reflected

electrons. Although I've never observed these "reflection" images (we

religiously coat our glass samples, although I understand why one may not want

to coat an art relic), I think I have a better explanation. The surface of the

non-conducting sample is acting as part of a capacitor... the other part being

the inside of the chamber. What you are imaging is a surface charge set up by

this capacitance, and not reflected electrons. Electrons deflected completely

away from the sample are unlikely to image much of anything (try running the

EDS simultaneously with this effect...if the electrons are reflected there

should be a huge bremstraalung peak, and nothing else.

Jeff Fortner

jeff_fortner@qmgate.anl.gov


Tina,

Got it in one. Somebody (SPI? EDS?) used to sell a "specimen

chamber inspection" stub that was basically half a marble that did this.

Charged some outrageous price for it.

Phil

P.S. I hope my "not the obvious" was taken as it was meant.

Philip Oshel

Station A

PO Box 5037

Champaign, IL 61825-5037

(217)244-3145 days

(217)355-3145 evenings

oshel@ux1.cso.uiuc.edu


Tsk, Tsk.

One can also go to any well stocked hardware store and buy a nylon

lock nut, the type that has a hemispherical surface closing off one

side.

Damian Neuberger

neuberd@baxter.com


Gotta disagree with you on this one Jeff,

The idea of a surface image charge is interesting, but doesn't

correspond to the characteristics of the phenomenon. Let me state

several distinctive characteristics:

1. You can focus these images just like an ordinary specimen image.

2. The topography of the reflecting "mirror" specimen disappears.

3. One can image and differentiate objects which are at a uniform

ground potential.

4. One can image objects which are quite some distance away.

5. The effect requires one to first "charge up" the surface at a higher

beam voltage.

Let me illustrate by my first encounter with this effect. I was imaging

on a swatch of nylon fabric at 1 keV when a local thunderstorm knocked

the power out. When power came back on, the SEM came back online at 30

keV and after tuning up my filament settings, I attempted to return to

imaging my nylon material at 1 keV (not very bright in retrospect, but

hey, it was 3 a.m. and I wasn't too coherent). Instead of seeing the

fibrous structures I had seen earlier (at the same relatively low mag) I

was seeing unexpected unfocused contours -- some rounded, some linear --

vaguely familiar, but I couldn't put my finger on it. Fiddling with the

controls, I noticed that by focusing to longer working distances, the

shapes became sharper -- but still in no way recognizable as the

material under the beam (I had meanwhile peeked through the glass

viewport and verified that the nylon sample really WAS what was under

the beam). As the image came into sharper focus, I suddenly realized

that what I was seeing was a "fisheye" image of the entire inside of the

SEM chamber -- polepiece, detectors, stage, and other internal

mechanisms. For a moment I felt like I had entered some sort of

"Twilight Zone"!

After I figured out what was going on I became enamored with the effect

for a time and perfected my technique -- graduating to a saphire bead

with which I could make truly detailed and relatively distortion-free

images of the chamber interior. I could easily zoom in on and image

parts of the chamber which were 12-15 inches from the specimen (this was

a very large chamber).

The point is that although the nylon material had an obviously irregular

topography, it produced a quite uniform "mirroring" field. This is not

surprising, since the electric potential of an insulator charged up by

30 keV electrons will deflect a 1 keV electron at some distance from the

surface. At this distance the contributions from the various surface

charge sites integrate into a quite uniform equipotential surface which

acts as a nice smooth mirror for the incident electrons. The incident

electrons "bounce" off the equipotential surface according to the normal

laws of optical reflection and the electron trajectories behave exactly

as if one introduced a mirror in the path such that the scanning beam

now scans the objects visible in the mirror -- the interior of the

specimen chamber. It takes only a very weak field to attract the

produced secondaries to the detector so a usable image is produced

(though typically weaker, given the longer collection distances.)

If a capacitive "image charge" were responsible, one would need to have

a very regular capacitor surface to retain an intelligible image --

clearly not the case with the nylon swatch. Secondly, creation of such

an imaging charge would require that the features being imaged would

need to be producing strong variations of local field at the capacitor

surface -- not the case when I could image features of the chamber which

were all at ground potential and at considerable distance. Finally, an

image charge would not lend itself to focusing.

I can see the merits of your hypothesis, especially when the original

question involved seeing the secondary detector (which is at an elevated

potential) on a glass surface (presumably smooth). I can imagine a weak

image charge being produced on the glass via the proximity of the SED

field. But this would be evidenced by a rather subtle and smooth

modulation of the normal image contrast (remember that the SED

collection field at the specimen is necessarily weak and uniform, else

the incident beam would also be badly deflected). In the case of the

effect I have been talking about, the topography of the specimen is

REPLACED with a highly detailed reflected image -- exactly as if a

mirror were inserted above the specimen.

I don't recall ever attempting to look at the x-ray spectrum produced.

I wouldn't expect to see anything much since the electrons are striking

objects which are out of the line of sight of the EDS detector. A pure

brehmstrahlung spectrum should be produced, as you suggest, and this is

an interesting idea.

A final note -- in my earlier posting I commented that I knew of no good

use for this effect. In fact, I did once use it to locate a breakdown

across an interior insulator. It is also a good way of noting which

interior features of the chamber are most strongly producing secondary

electron "background" as would normally occur from electron

backscattering onto the chamber walls. But its best use IMHO is still

its considerable potential for amusement!

Fred Schamber

RJ Lee Instruments Ltd.

fhscham@sgi.net


Yes, one would hope and expect to see few x-rays then. However, some are

detected and they are not all bremsstrahlung:

Unless fast electrons reach the atoms in the sample they cannot generate

bremsstrahlung in the sample. The electrons reflected by a strong enough

electrostatic field produced by an electrically isolated, electrically

charged specimen can generate characteristic x-rays as well as

bremsstrahlung from all specimen chamber materials which are visible in the

mirror image of the chamber.

If an abnormal background shape is observed in x-ray spectra recorded in the

"mirror" condition, a significant source is the reflected electrons

themselves -- either entering the x-ray collimator and generating

detectable x-rays there or even penetrating the detector window and

reaching the detector crystal itself (especially if there is a thin window

or if there is no magnetic "electron trap" in the collimator). BTW, this

same effect can be observed clearly in TEM or STEM if the collimator's

internal geometry is bad and the EM is operated with a very low or zero

magnetic lens field at the specimen, as is commonly found in low mag mode.

Best wishes,

Brian

Brian W Robertson Office 402 472 8308

Associate Professor Lab 402 472 8762

Department of Mechanical Engineering and FAX 402 472 1465

Center for Materials Research and Analysis,

University of Nebraska-Lincoln

255 Walter Scott Engineering Center

Lincoln NE 68588-0656 USA

bwr@unlinfo.unl.edu


HI

To image the inside of my SEM, I first stick a piece of PTFE or a round glass

coverslip onto a stub and then glue a small ball bearing (3mm) onto the centre.

This gives a much better view of the chamber. You can try using larger ball

bearings, also to give a weird effect try sticking two small ball bearings together.

Also try switching on your back scatter detector once you have 'charged' up the

stub, you can visualise the sectors - (if + the sector is white and if - the sector is

black)

Kevin Mackenzie

Tillydrone E.M. Unit

University of Aberdeen

Tillydrone Avenue

Aberdeen

AB9 2NT

SCOTLAND

Tel 01224-272847

Fax 01224-272396

web site: http://www.abdn.ac.uk/~nhi691/

nhi691@abdn.ac.uk


[Return to Tips & Tricks Menu]