Scanning Electron Microscope

I have always wanted to use a Scanning Electron Microscope (SEM) and my chance came last week at Cambridge University.  What a thrill.  I received an optical microscope for Christmas one year as a child and had endless fun and excitement with it, but the SEM is a whole new ball game and has a resolution down to 1.4 nm.   That’s a thousandth of a micron or a millionth of a millimeter. Hard to imagine.  To add to my excitement, it also had a focused ion beam (FIB) microscope adjacent to it that we used.  I didn’t even know what they were until a few days before we used it.  Its similar to the SEM microscope but fires ion instead of electrons which are much bigger and heavier so can inflict more damage.  When the current is turned up, you can machine or mill the sample in a very controlled and accurate way.  We also injected gas molecules (organic platinum) between the sample and the FIB.  If the current is just right, it won’t machine the surface but rather split the platinum from the organic part of the molecules and deposit them on the surface of the sample.  Because you have the SEM pointing at the same place you can look at what you have done.  I want one!

SEM with FIB at 54 degrees

You don’t think of a microscope firing a current, but that’s exactly what a current is, a flow of electrons.  To get an idea of how many electrons are being fired at the sample; A Coulomb = 1 Amp per second.  The charge of a single electron is around 1.6 x 10-19 coulombs so there are 1 / 1.6 x 10-19 = 6.25 x 1018 electrons per second for 1 Amp.  We have a typical probe current of 200 pA so there will be 6.25 x 1018 x 200 x 10-12 = 1250 million electrons per second or 1250 electrons every μs bombarding the sample.  Just shows how unimaginably small they are and how many of everything must move before we can even notice it.  Of course, the FIB also fires a current, just not the negative electrons rather the positive part of the atoms that are left once some electrons have been stripped.  They usually use Gallium ions because it melts below 30°C.  Its solid at room temperature but melts in your hand, unlike treats (now called M&M’s).  A proton weighs 1.6727 x 10-24, a neutron about the same at 1.6750 x 10-24 and an electron 9.110 x 10-28 so a proton is 1837 times heavier than an electron.  Gallium normally has 31 protons and normally 39 neutrons in its nucleus.  It therefore normally has 31 electrons in orbitals around the nucleus but if the electrons are stripped from the outermost shell (valance) it will become a little smaller and be positively charged.  Because only one electron is in the outermost shell of gallium it will have exactly the same charge as an electron but be positive rather than negative (1.6 x 10-19 coulombs).  So, each ion in the FIB will carry the same charge as each electron in the SEM but weigh about 1837 x 70 = 128,590 times as much.  So, you can imagine how much more momentum and damage a beam of ions can do to the sample surface compared to electrons.  That’s the same difference as an average man (70 kg) compared to a freight train with an engine and 63 loaded rail cars.

We started by measuring the features on a DMD (Digital Micromirror Device) using the SEM.

Image of DMD mirrors

The mirrors are about 13 μm square with a Ø1.5 μm hole and the gaps between the mirrors are about 1.1 μm.  So surprisingly we could get an array of 100 x 100 or 10,000 mirrors in a square of 1.3 mm on the side.

Well a DMD chip has several hundred thousand microscopic mirrors arranged in a rectangular array on its surface; each of which correspond to a pixel in the image displayed.  Each mirror can be individually rotated by around ±10° which represents an on or off state.  When on, light from the projector is reflected off the mirror, through a lens onto the screen as a bright pixel.  When off, the light is directed to a heatsink, so the same pixel appears dark.  To produce grey, the mirror is toggled on and off very quickly by pulse width modulation, so the shade of grey corresponds to the ratio of time on to time off.  To produce colour, three coloured projectors and three mirrors are required for each pixel and the ratio of on to off for each colour determines the pixel colour brightness.  The mirrors are made from aluminium and are mounted on a yoke which is connected to support posts by torsion hinges.  Because of the small scale, hinge fatigue does not usually cause a problem.

We did a few more experiments, carrying the acceleration voltage and the probe current etc, and then changed the sample for an EPROM and set up the FIB.  If you set the distance an M stage tilt correctly, its possible to be able to view the same part of the sample with the SEM and the FIB without changing the focus or moving the sample.  This is called the Eucentric point.  Below is the same image using both devices.Compare

TrapYou can see that the SEM is viewing the sample at an angle but there is a tilt correction mode if you like.  The resolution and contrast of the SEM is so much better than with the FIB.  Next we drew a trapezium on the sample that we were to mill out by setting the accelerating voltage and current just right we got a really nice cut of 3 microns deep.D200pSEM

We found the optimum current to be around 200 pA.  Its amazing how accurate and easy it is to machine the sample; this image is taken at 4000 times magnification.  The maximum optical lens you can use is about 1000 times.  This is just run of the mill, its capable of far greater magnification if you like.  When looking at biological samples, they have to be stained with a rigorous process prior to viewing so that the different tissue types stand out and can be seen.  When looking at anything else, you just put it in and focus; and the images are so much better than the ones I could get on my optical microscope as well.

GISOur next step was to inject a gas between the FIB and sample when it was scanning.  The gas we used was made from platinum organic molecules.  If everything is set correctly it will split the gas and deposit the platinum part on the sample while the organic part is vented away.  This time we drew a small rectangle on the sample as it takes a while.

Platinum deposit on EEPROM

We then decided to do the same cut, half on the deposit and half off.108

We noticed some funny furry deposit at the front of the cut so zoomed in for a closer look.Fur

This view was taken at 8200 time magnification and we could have gone in much further.  We then moved onto our last part where we analysed the surface at different places to do a spectral analysis and find out which elements were present.

Map Sum Spectrum
Spectrum of area around the cut out and deposit
Summary of elements at various places

The redeposit column was taken on the furry deposit which was found to be made up of platinum, silicon and carbon.  Basically, everything that was blasted away when the cut was made.  Its interesting that there is so much carbon around; I think the only place that could come from is the organic part of the gas which should have been vented away.  Not quite sure how it gets vented away anyway as everything happens in a vacuum and to vent away you need a lower pressure than where it is.  Well, you don’t get much lower pressure than in a vacuum.

Image inside vacuum chamber

Sorry its been so long since my last post but I have been so busy I cant tell you.  We have lectures all next week at Cranfield then the following week at Cambridge and then no more!  So once I have written up the reports for those I will only be doing my thesis so should have more time to catch up with everything else that has happened.  I will get there in the end.


One thought on “Scanning Electron Microscope

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s