Quantum Computing Work Experience – Week 2

Today concludes another fantastic week working at Imperial. This post will probably be a lot less massive than the previous one owing to time constraints (I’ve just come back from several exhausting hours of rock climbing at the Westway and Google Calendar tells me I have my driving theory test at some point in the near future).

Firstly, some pics to support stuff from my last post.

The entire table looks like this - I can understand why everything needs to be realigned and tweaked every 15 minutes for the experiment to work! The two blue lasers are I think the main cooling lasers, pumping from the ground to the high energy levels of Ca+.

This is a used Cu O-ring (actually called a gasket) - you can easily see where the knife edge bit into the Cu making a vacuum seal

This is a used Cu O-ring (actually called a gasket) - you can easily see where the knife edge bit into the Cu making a vacuum seal

A top view of the laser setup. You can see the diffraction grating (with an arrow drawn on it) and the connections to the piezo behind it. Click to embiggen.

A top view of the laser setup. You can see the diffraction grating (with an arrow drawn on it) and the connections to the piezo behind it. Click to embiggen.

Equipment

Tantalum Oven

This is the equipment used to produce neutral atoms which are to be ionised.

The oven is suspended between the two electrodes by a Ta wire. The Cu foil is there for a test run of the oven - if it works we should be able to see a spattering of grey Ca on the Cu.

The oven is suspended between the two electrodes by a Ta wire. The Cu foil is there for a test run of the oven - if it works we should be able to see a spattering of grey Ca on the Cu.

This is the plate on which the entire experiment (ion trap, oven and all) will sit. It will get inserted into one of the holes in the central 6-way cross can.

This is the plate on which the entire experiment (ion trap, oven and all) will sit. It will get inserted into one of the holes in the central 6-way cross can.

When I first heard they were going to use an oven, for some reason I imagined some sort of miniature baking oven that somehow emits atoms when turned on! The actual oven is actually a tiny 1cm long tube of tantalum (Ta), sealed at one end by essentially squashing the end, with a tiny hole in the middle of the tube. Ca shavings are stuffed into the open end before the oven is closed, again by squashing. The whole oven is attached by a piece of Ta wire to two electrical contacts across which a potential is applied. The Ta conducts current and heats up, acting as a heating filament. The Ca heats up and the most energetic atoms spit out of the hole (the process is basically evaporating the Ca at very low pressure and high temperature).

I asked why Ta is used – presumably Tungsten (W) has virtually the same properties in that it heats up when current flows through it, and since W is the metal of choice in light bulbs, presumably it’s cheaper? Apparently W can indeed be used; for such applications the criteria for metals are that they are UHV-suitable (don’t trap/adsorb other molecules/atoms on/to their surface which subsequently outgas, ruining Ultra-High Vacuums) and won’t melt at high temperatures. However Ta is normally more suitable than W because it’s more malleable (whereas W is very springy) and can be easily spot-welded (to stick the wire onto the oven). However thoriated W is better than Ta as an electron source since the thorium gives it a much lower workfunction, allowing more electrons to pop out for the same energy input.

Vacuum Pumps

In my previous post about this work experience I omitted some detail on the pumping that I learnt this week. As it turns out, the actual pumping requires three pumps. The first is a roughing pump, to get the pressure down to a very rough vacuum (~10-2 mbar). Here they were using a rotary vane pump:

This image was nicked from wiki

This image was nicked from wiki

Essentially as the off-centre internal cylinder turns, the vanes get longer / shorter accordingly such that the pressure at the input is always getting lower and at the output it’s always getting higher, forcing the air out of the output. The principle is essentially PV conservation.

The second stage is a turbo pump which is basically an electrical version of the intake fan of a jet engine. It spins extremely quickly (so quickly that it requires a low pressure to operate lest it smash itself to pieces) and the idea is that it spins so quickly that any molecule that hits a spinning blade hits the part of it such that it gets kicked outwards, away from the vacuum. This gets the pressure down to about 10-5 mbar. The final stage of course the ion pump.

When air is pumped out the can sits in an oven - the idea is to heat up anything that can outgas while pumping, making it outgas more, thus getting rid of outgassable stuff

When air is pumped out the can sits in an oven - the idea is to heat up anything that can outgas while pumping, making it outgas more, thus getting rid of 'outgassable' stuff

Wavelength Tuning using Iodine

There are several ways of tuning wavelength (I wrote something about the cavity method – setting up a standing wave – in my previous post), but I found this way of doing it particularly interesting. Like all other elements, iodine has a certain absorption spectrum, a feature used in star spectroscopy to determine elemental composition. But instead of doing what astronomers do (measure wavelengths to identify elements), here we were using a known map of iodine’s spectrum to tune the wavelength: light shining through the iodine has a certain attenuation which is dependent on the wavelength (owing to electron energy levels). By shifting the wavelength around using a piezoelectric it is possible to obtain a local iodine absorption spectrum (wavelength against intensity). By comparing this local spectrum with an ‘atlas’ – iodine’s spectrum for a large range of wavelengths, it is possible to locate the local spectrum within this atlas, thus identifying the wavelength. Apparently a narrow band of local spectrum is sufficient to identify a unique location in the atlas: there are no ‘repeats’. Whether this is non-repeating property is specific to iodine (hence its use) I’m not sure; the isotope used is radioactive so there must be some really good reason to want to use it!

Techniques and Procedures

Saturated Absorption Spectroscopy

All atoms radiate photons. However in a cloud of atoms these photons are affected by the Doppler shift owing to the random movement of the atoms, and a graph of frequency against intensity (basically a spectrum) shows an underlying distribution for this radiation. However the interesting bit of spectroscopy occurs on the surface of this curve, in the form of ‘ripples’ on the underlying curve’s surface. While a human can normally see the ripples roughly by eye, the underlying curve gets in the way of accuracy.

The solution is a method of somehow obtaining the underlying distribution without the ripples using lasers and subtracting this curve from the spectrum, resulting in a graph of just the ripples. I’m still clueless as to precisely how this works / is performed since I didn’t personally bear witness to the process (I heard something about matching lorentz curves to points but that was probably more to do with analysis of the ripples rather than the process of saturated absorption spectroscopy) so it looks like some wiki-ing is called for.

Walking the beam

This isn’t some physicist’s attempt to be a pirate and getting the words muddled; it’s actually a rather clever (though extremely time-consuming) method of aligning a laser beam. Bascially the ideal situation is a laser beam passes precisely through two points. This is very difficult to achieve with just one stand so a setup with mirrors is necessary. Here are several different failed attempts at diagram-ifying the thing:


The dotted lines were added by me to show where the beam will go

The dotted lines were added by me to show where the beam will go

At each of the two points the beam needs to go through an adjustable iris is placed (think circular doors in sci-fi films), and mirrors alpha and beta (making up the periscope) can be adjusted so the beam’s height (h) and angle of elevation (e) can be adjusted more or less independently. Then the following two-step process is iterated until the beam is almost exactly where it needs to be.

1. Open B completely, close A so it becomes a tiny hole, and adjust the laser so it goes through A using mirror alpha
2. Open A, close B so it becomes a tiny hole, and adjust the laser until it goes through B using mirror beta.

Illustrations of the steps are as follows:

For some reason it reminded me of numerical analysis / Newton Raphson type things – constantly optimising and getting closer and closer to perfection yet never reaching it. GL’s cobweb illustration of numerical analysis seemed particularly similar to this situation. Anyways while I quite like how it works, walking the beam does start to lose its novelty after doing it for a couple of hours…

Scanning Tunnelling Microscopy

Danny also explained some awesome stuff on this and how it works. Basically the idea of STM is to use quantum tunnelling calculations to make a map of a surface. A probe is held (say at +5V) very near a surface (grounded), and owing to quantum tunnelling, a certain current flows between the probe and the surface. This current is proportional to exp(-l) (or something like that) so it is possible to measure l to a very high degree of accuracy. As the probe is scanned across the surface, a matrix of measurements of l against (x,y) can be created, thus mapping the topology of the surface. This mapping can in fact be so accurate that it can pinpoint individual atoms sticking out from / adsorbed to an otherwise flat surface.

Conclusion

The last two weeks have been nothing short of awesome. I’ve learnt (and sometimes noted down) many new things on every one of the last ten working days and I’ve recounted here and in last week’s post merely a handful of the more interesting bits and pieces. I even solved an apparently insurmountable practical problem thus moving the entire scientific community forwards! I mean … I came up with a (pretty good) solution to unscrewing a stuck nut… Many thanks to Danny Segal for giving me such a wonderful opportunity.

There is one thing I still can’t work out though:

Advertisements

One Response to Quantum Computing Work Experience – Week 2

  1. […] lots of good theory in practical experiments (cf. my Imperial work experience – week 1, week 2), so it’s all […]

Leave a Reply

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

WordPress.com Logo

You are commenting using your WordPress.com 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 )

Google+ photo

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

Connecting to %s

%d bloggers like this: