[00:00:05]
>> It's a pleasure to have Chandra Rahman with us today Chandra got his bachelor's degree in electrical electronics engineering from Cal Tech and then a master's degree in the same subject from the University of Michigan go blue. And then his Ph d. in applied physics also from Michigan after a post-doc at mit he came to Georgia Tech in 2001 where he's currently an associate professor of physics.

[00:00:34]
He was elected a fellow of a.p.s. 2013 I think and and also one of the interesting things I saw in his bio was that he took a leave of absence during his time here at Georgia Tech to work for 2 and a half years a company a 0 cents and then brought back what he learned to his lab here as academic lab so it's interesting thing anyway it's a great to have him and I'll turn this over to him Larry.

[00:01:10]
Who can hear me can you hear me say it better so that's. When the power that are that good can you guys hear me you folks hear me in the back Ok All right good. Ok well thanks for the kind introduction it's it's a pleasure to be here I've as David said I've been to Georgia Tech since 2001 but this is the 1st time I'm giving seminar at i.b.m. so and that's reflective of the fact that for the 1st time in my research career I'm actually doing things that might be of interest to 2 people here so I'm very excited about that in the last couple of years we've started up some research areas that I think will be interesting to explore some possible collaboration's And and so I hope I can convince you that some of these things are very interesting and the world of atoms is one which might be of interest to 2 engineers particularly Ok great so.

[00:02:18]
The title of my talk is quantum sensing with atoms and quantum is a buzzword that really popular nowadays and. There's something called the National quantum initiative there's been a lot of interest hopefully there will be some funding as well that goes along with it. New funding in the area of trying to do.

[00:02:41]
Quantum based sensing and quantum based computation you probably have heard something about and I'll tell you a little bit about about this topic of quantum sensing using atoms neutral atoms this are sort of a pictorial representation of that I'd also like to take the opportunity to thank the National Science Foundation over in our in the Air Force Our which has supported the work that I'll be telling you about today Ok great.

[00:03:10]
So here's just a picture of my. Group at the moment and I'd like to thank some of the students that have been involved this research and our research areas are basically divided into 2 in my group there's one part which does research on ultra cold atoms so these are Adams very close to absolute 0 they're trapped this is a sample of 10 to the 6 atoms suspended in a vacuum chamber and the motivation for this research is to try to simulate comments matter systems are more complex systems the low this is a gas.

[00:03:42]
And the idea is to examine things like quantum magnetism many body quantum dynamics quantum entanglement in a system that is very well controlled and very precisely know all the parameters and hopefully that will shut some insights into complex phenomena such as high temperature superconductivity and maybe one it one day might have some applications there's another part of my research which is more directly application related and that's an attempt to try to being at Atomic instruments into the realm of micro fabrication to try to make miniature devices.

[00:04:21]
Chip scale atomic clocks atom interferometers and eventually to do inertial sensing using using atoms on a chip and so this is an area I would be telling you so this will be the 1st part of my talk will do on this area and depending on the time I tell you a little bit about what we're doing with trickled out of those.

[00:04:47]
So just to thank some of our other g.t. collaborators we also started a collaboration with these group here in d.c. e and his students have been helping out my students to build some nice devices we have a couple of papers that we have written together based on the research I'll be telling you here and we have a patent application as well and we've also started a project with the group also an e.c. e. and then a for tonics that's just getting started.

[00:05:17]
Ok so let me back up and. Start this talk by saying what are atomic instruments really. You may have heard about atomic clocks but what are the advantages for building an instrument using atoms so the key advantages are that atoms are a highly accurate way of keeping track of time they're very precise and they're very repeatable so if I build So this is of a graph from this paper from the NIST group that built 2 atomic clocks based on terbium atoms we call them a terbium one and it terbium 2 Now if you go out and buy 2 clocks they're never going to give you the same time for ever However these 2 clocks are made of the identical atom they're both the terbium atoms and so there's a very high degree of correlation between the 2 and this this repeatability means that you can build these clocks all over the place and they're all going to read the same time and that's a very important.

[00:06:28]
Advantage However this is a tabletop instrument so what advantages or infer can conferred by solid set instruments they're highly integral so you can put a whole bunch of things together you can manufacture them at low cost these are very expensive things to put together. And they have to be done by hand at present and so there's no real prospect for mass manufacture and it's also difficult to make these atomic instruments low power there's a lot of lasers there's a lot of power supplies there's a lot of things that go into this that are bulky and consume a lot of power but for example a solid state.

[00:07:10]
Sensor a clock might just be an atom inside of a or a vacancy inside of a piece of crystal in matter and it can be very low power you can put all the electronics very close to it and I think that this is a huge advantage of the solid state but atom still have the accuracy and repeatability and so if there's a way to move atoms to chip scale perhaps we can get some of the benefits of solid state by having access to for example being able to integrate all all the all the controls around the this the sensor and we can do that without a mess because after all the individual atoms that are in these clocks are themselves very tiny so there's no ultimate limit to how small you can make them it's really an engineering challenge to try to bring all that to the small scale engineering in a Science Challenge Ok so I was told that there are many students in the audience here so I want to back up and say Well from a fundamental point of view what's going on why are Adam so good atoms are so good at keeping time for example because of one thing they have discrete energy levels that are unique to the atom.

[00:08:27]
So the energy states of one rubidium atom This is the hyper fine structure of the rubidium atom with one particular isotope the Rubidium 87 isotope has has a so-called hyper fine structure but I prefer the structure is very weakly coupled to the outside world only through accidental magnetic fields and then even though the energies depend on this magnetic field acts as a symbol for the magnetic field I didn't draw this graph I borrowed it from Steve Jefferts whose.

[00:08:58]
Atomic clocks scientists at at NIST but there are states that are very very strongly sensitive to the magnetic field but there are some that are very weakly sensitive to magnetic fields and those are the ones that are used to make an atomic clock in fact they're used to make the atomic clocks that are used in g.p.s. based ations So how good is that clock Well the number of digits known is very high and if you if you have an oscillator that's slave to this transition you can keep time very accurately the other thing that I mentioned similar to the a terbium clocks that I showed you in the earlier slide is there's no variability from one atom to the next except insofar as it might have a weak dependence on the environment that it's in but that can be hopefully something controlled very carefully it's very stable over time and highly accurate so you can make a precise clock but even if it's not all that precise it's extremely accurate.

[00:10:02]
So what kind of precision instruments are there after all the hype refined structure of rubidium has been known for many decades so people have known that you could build very precise atomic clocks. What what type of clocks are there out there in fact there are many instruments that you can buy So here's an atomic clock that's based clock that's very precise made by a company called a silicon or it's it was a competitor to this h p.

[00:10:30]
50 sorry 4071 a which was a mainstay of. Naval navigation for a long time and so this this clock is something that sits on a rack mount and it has a little atomic beam inside of it unfortunately not so little atomic beam actually a big atomic beam and this this clocks are very accurate so they can have a precision of 10 to the minus 14 absolute accuracy of 10 of the minus 12 but they're expensive so if you want one of these clocks it cost tens of thousands of dollars you can buy one for 10 k. and e-bay but if you want to buy new then it's going to be more expensive than that there's also no prospect for volume manufacturing so each one of these has to be assembled carefully it's not something that can be stamped out in a large number and they're not portable This is rackmount and heavy So where do we go from here how does one make things smaller Well atomic physicists like myself are good at building things that fit on an entire optical table so an optical table that's 5 feet by 12 feet this one from a Ph d. thesis from Stanford University and this is an atom interferometer gyroscope So there's an atomic beam inside of this steel.

[00:11:51]
Can hear and it can be used because the atoms are moving in that vacuum and they are sensitive to inertia it can be used to build a very sensitive gyroscope in fact this ph d. thesis demonstrated micro degree perused our sensitivities for a gyroscope minimum sensitivities which is navigation grade and below and about but unfortunately it's on a tabletop size and very difficult to make it smaller so the point of this is that portability and cost solutions would really enable applications in navigation geolocation etc So finding ways to make these instruments portable and disseminate being able to disseminate them widely would go a long way to words practical applications.

[00:12:42]
So people have been working on making chip scale devices for atoms and so I wanted to review some of that So what are people doing in the world of atomic physics trying to make things smaller and more portable so there's this falls into 2 categories there are devices using cold atoms so these are atoms that have been laser cooled and brought to temperatures close to absolute 0 I'll tell you a little bit about that towards the end of the talk in my section on.

[00:13:11]
Quantum simulation but the idea is that cold out of the can be actually brought to a chip and then trapped above the ship surface using some wires and the wires you can run a current through the wire and then it can find the atoms because of their magnetic moment and then you can send laser beams in and probe the atoms This was actually an application where you can actually because the atoms are magnetically sensitive you can make little magnetic maps of the surface that are pretty sensitive.

[00:13:42]
Just using the atoms however there are lots of issues with this kind of technology. One of the main issues is that although the coherence times are very long so you can have coherence times for the sample that are seconds that's much higher than in any solid state system and a moderate cetera but the problems with this are duty cycle issues so you can't really measure the atoms all the time you have to spend some time cooling them down and trapping them and waiting for them to get into the right frame of mind so that you can use them for your sensor application Ok and atoms are a little finicky they like to you know they take some time to settle in and that's a problem if you want to measure something that's dynamically changing then if you have a duty cycle you're going to miss out on some of the dynamics and it will introduce extra noise.

[00:14:33]
And so that is one of the main issues plus the complexity of the apparatus that's needed in order to trap and cool the atoms is significant and a lot of progress needs to be made on making that more more portable. On the other hand Adams can easily come for free in a vapor I can go out and buy some rubidium atoms for 50 bucks and there I have many many many atoms but I can use so I can put some and in a little they Purcell this is a memes fabricated they Purcell's came from a a review by John Kitching Scroope John Kitching himself at NIST on the state of such sensors this was used to build a chip scale atomic clock and also magnetometers are built this way little memes fabricated vapor cells and you heat them up and then you have a vapor of atoms and you can use that vapor of atoms the advantage of that is that it's very much simpler technology than the cold out of the world however the disadvantage is that you still don't have very long coherence times like you do in the cold atom world so my group we looked at this paradigm and we started to say Well is there a way to get some of the benefits of the simplicity of this hot atom systems but still retain some of the coherence properties of the cold atom world so come on get the benefits of both worlds in some way.

[00:16:00]
And so we started to talk to some of our the c.e. colleagues over here and we explored some new ideas and that's what I think is great about Georgia Tech is that. As a scientist you can go out and talk to some engineers and they'll tell you some cool things that you can do they can also set you straight and tell you well that's ridiculously hard so we need that balance but the idea that we've sort of been playing around with this a vapor cell is a box and the atoms are moving in all directions and eventually they encounter the walls and when they encounter the walls that's a problem because the atoms lose their memory sometimes after a certain number of collisions with the wall and so it's not the best thing for making a very precise instrument you really want the atoms to be isolated and not go anywhere so this is vacuum inside this wall except for the atoms and so we thought well what if our atoms were all travelling in the same direction so this is a vapor so the atoms are traveling in all directions the atoms are all traveling in the same direction that's called an atomic beam.

[00:17:05]
And then at least as far as the atoms don't reach the end of the sensor they're not encountering anything at all so they won't bounce off the walls and they have very few chances to bounce off each other either because they're just making one transit and so could this be a useful device for making a sensor or a clock and the answer is yes so Adam's don't touch the wall the other important thing is that Doppler broadening is eliminated so that here the atoms velocity is in all directions but here the velocity vector is pointing in this direction so if I sent a laser beam perpendicular to that direction I would have 0 Doppler shift.

[00:17:50]
So Doppler broadening is eliminated so you can do precise spectroscopy certainly. And you can also build. A clock using this method you can interrogate the atoms here and then at some later time interrogate the atoms here but the time is not real time it's just the transit time of the atom so in fact this is a continuous laser beam this is a continuous laser beam and the atoms.

[00:18:18]
Can go from one to the other and the 2nd laser beam detects the coherence that was created by the 1st laser beam this is a well known technique called Ramsey spectroscopy and the single particle coherence can lead to a clock because the atoms internal energy levels evolve from one Ramsey zone to the next Ok there are possibilities that maybe one could even exploit.

[00:18:45]
Collective properties of the atoms but even if you didn't do that and you just had single particle coherence this type of clock is useful because the atoms don't encounter the wall in between the 2 zones and because of that is potentially a way to make a very accurate sensor so that's a a an application that we're currently pursuing.

[00:19:07]
Ok so. So what are the challenges in miniaturization So I told you about a miniature vapor cell so that's actually a fairly mature technology now so you can start out with a vapor cell that's a centimeter in size so this is a little. Glass cell in which there's a little bit of rubidium that's been put inside or cesium alkaline metals rubidium and cesium are typically useful for making clocks and magnetometers and then they can be shrunk down to the millimeter scale by using mems fabrication So this is a piece of silicon there's some patterns that I've been through and then there's glass bonded to both sides with a little bit of rubidium inside and you can shrink things down to the millimeter scale using this kind of technique and and so that's that's pretty powerful so the problem is with atomic beams is that they're long and so you're not starting with something that's a centimeter and trying to make it a meet millimeter we're starting something with something that's a meter in size Ok similar to what I showed you on that 1st slide with the atom interferometer gyroscope and it has to be a meter in size this is a picture of an atomic beam makes apparatus in our laboratory that we use for laser cooling and to take that from a meter sized down to something really small in the millimeter size seems very challenging at 1st so what are the principles the physical principles that make it a metre why does it have to be a metre Why couldn't I have started with something smaller and it turns out that there's good scientific reasons why it's that size and you have to if you want to try to the message is that if you want to try to miniaturize some thing you have to understand the physical principles that require it to be that size that it started so here is a Tomic beam So how did I get all my atoms to travel in the same direction when they started out in a in a vapor in which they were travelling in all directions so this is the cleverest dumb trick in the world you just take a box and you put a hole in it.

[00:21:10]
And as long as there's vacuum everywhere then the atoms will slowly if use out of this little hole and they'll if use out with a cosigner distribution so the atoms that travel at a large angle are suppressed by a the cosigned if they are that sort of 90 degrees there's nothing but it's still a fairly broad distribution on axis and there's 100 percent probability and then I simply put another aperture down stream and I get rid of all of these atoms that are travelling at an angle so then I have a beam of atoms was velocity vector is more or less in the forward direction and I filtered out all of the other atoms however that's Ok because I have a lot of atoms to start with and I can afford to do that to make the clock Ok so those rack mounted clocks are doing exactly that they're filtering out the atoms.

[00:22:01]
Ok but why why does it have to be the size it is it's because simply because of the geometry if you're trying to do this filtering and I want to pass atoms through the 1st nozzle and then have them be filtered through the 2nd nozzle and I want to reasonable divergence saying go and if I want my source size to be not too small then it is say if it's 20 mil if it's 2 millimeters in diameter and I want 10000000 radian of collimation angle here then it has to be 20 centimeters long and that's the way that device operates so that's what makes it that size it's not just simply because of of habit or anything like that it's just really the way that this device is designed.

[00:22:49]
Ok so can we fix that yes we can fix that there's been a solution to that that's been known for a very long time it's called a capillary array and so the way it's done is to draw a bunch of glass tubes and then you shrink down the size of each tube so it's very small so I have the same aspect ratio in terms of the length of the device to the size of each puir but I didn't have to make my whole device 25 microns I can have a whole array of them so I still get a reasonable flux of atoms coming out in order to make a sensor I need to have a large number of atoms so I can detect something Ok so this capillary arrays great it's it's an array of tubes and has a large aspect ratio so the beam that comes out is collimated So let's try to understand the principle behind that Ok so I have a box full of atoms here and I don't Ok and this looks like it's really just a cartoon drawing but in fact it's a money Carlo simulation and there's a tube attached to this box the atoms bounce around inside the tube and eventually come out but there's more of them that tend to come out without hitting the 2 walls at all.

[00:23:57]
As long as there is not a 100 in Amec flow inside the tube you'll get a beam of atoms coming out on axis and you still get some out of coming out to the side but the atoms that were coming out to the side actually that are bouncing off the walls of the tube actually have a good probability of coming back to the source and not leaving so the flat flux on axis is about the same as you would have for cosigned emitter but you have more of a directed beam Nonetheless if you look at the angular distribution of the flux coming out here that angular distribution extends to all angles and in fact if you look at the for this type of aspect ratio if you look at the atoms coming out with outside of the.

[00:24:42]
So if you look at the flux between 0 degrees and the half angle let's say that half with that half maximum of the Sangar distribution the area under that is only one percent of the total flux so even though you've produced a nice collimated beam you still spring out a lot of.

[00:25:03]
Flux to the sides so that the off axis flux is reduced due to all collisions It's a simple device to make but the beam purity is not very good you still have a large offset axis contribution. So if you're making a miniature sensor you can certainly start with a device that's miniature like this however there's still a lot of.

[00:25:24]
Extra vapor that's produced and that could cause a problem for building a precise instrument because that vapor interrogates your is interrogated by your laser beam as well Ok so. So we started to talk to our electrical engineering colleagues about this and I have and I fire chaos he and I sat down one day and just started to think about what we could do with silicon fabrication could we do something to make atomic beams using silicon and we came up with this concept of trying to make a plainer device so the idea here is that they instead of using a capillary array we would use tubes made out of silicon so this is a device that they fabricated for us it's a silicon collimate or it has little holes into silicon and another wafers bonded on top and the cross-section of these tubes is 100 micron by $100.00 microns so it's pretty small certainly from the point of view of atomic beam devices it would be it would be entirely appropriate for us and so we made it a set of column headers that had 100 micron cross-section and 3 millimeter channel length and so we you know you had asked what's the point of doing this when the capillary arrays exist the point of doing this is that these features a defined in photo lithography So they're not drawn in glass tubes here you can just patent it to be whatever you want it's completely flexible and you can make feature sizes that are microns or even possibly smaller and here's the key it's really easy to integrate this with other things so for example optoelectronics is something where you could then imagine that the atoms coming out of one of these tubes interacts with some opto electronic device Ok And it's right there everything can be fabricated together.

[00:27:14]
That's the vision. So I love briefly tell you about the fabrication process but only because I've learned about it I didn't actually do it. So this is the way it's done apparently you start out with silicon and you do deep. After oxide grout So there's an oxide growth layer that's put on and then it's at and then we take another way for and put a gold layer on top and then flip it on top and bond the 2 together using a silicon gold eutectic bond so there may be other ways to do this bonding that we're exploring right now and thinking about but then this whole structure becomes sealed and then it's dice to millimeter lengths so in now we have this device that's only 3 millimeters in size.

[00:28:06]
So here are some of the micro channels you can make so we can we don't need to necessarily have what you can do with drawn glass capillaries straight micro channels but we can make micro channels with different wall sizes so here's a 10 micron wall with 100 micron width we're able to make that and this is a cross-section of that before the top way for was put on that shows that there really straight and nice so they did a great job with making these so this device has 90 percent open fraction for the atoms which is great because we don't lose any flux by having walls really.

[00:28:42]
The other the thing in this is something you can do using lithography that you can do with my machining Ok you can make channels that focus in some place with the atoms with all focus to some point in space you can make 2 beams that have 2 different.

[00:28:56]
Angles of divergence with respect to each other that might be useful for some applications. So we made a bunch of these and then how do we test them so we test them by putting them inside a copper tube on the end of a copper tube and there's rubidium inside the copper tube and the whole thing sits inside a test chamber that's at 10 to the minus 6 Tor vacuum that vacuum level is is good enough for us to it's not it's certainly easy to get this vacuum with this kind of test chamber but it does not need to actually be that high probably even Milla Tor $200.00 micro Tor would be would be adequate.

[00:29:38]
But this is just a convenient a vacuum level and then we can interrogate the atoms inside this test chamber. So how do we detect the atoms and how do we interrogate them to make sure I'm. Ok. So we use fluorescents detection so we start with a diode laser That's at $789.00 a meters that's resonant with the rubidium atoms and then we shine that light directly into the chamber but it's perpendicular to the direction of the atomic beam so the light interacts with all the atoms regardless of their velocity as I mentioned earlier there's no Doppler shift and then we simply collect the fluorescents using a photodiode So this is collected fluorescents on a camera that shows you the output of each of these channels So either these are 20 micro channels here and every single one of them skipping is operating so the atoms come out with a longitudinal velocity of around 300 meters per 2nd and we're only sensitive to their transverse velocity that is the the beam is moving to the right but we're sensitive to their velocity in the direction perpendicular to the beam.

[00:30:48]
And you can collect this on a photodiode you have to be a little bit careful about using low noise current amplification which we learned how to do and this is pretty good. So what does the spectrum look like so now I scan my diode laser and I scan it over about one gigahertz of frequency range and I hit this resonance where the.

[00:31:15]
Of the deed to line of rubidium and within that resonance there are several peaks that correspond to different high profile and levels in the excited state. Within one peak the say the central peak that's the strongest one the the the line has a with that is narrow but has a long tail as you can see and that long tail is precisely the long tail of the angular distribution of the collar matter it's the velocity of atoms at very large angles that are emitted because they are diffusely scattering off the walls and we can see that in the spectrum directly after the red curve is just saturated absorption that's used as a calibration marker that shows you that there's a rubidium peak so this is from a rubidium vapor cell.

[00:32:03]
So we start to get data like this and we said hey collimate is working there actually atoms coming out of there in a narrow beam. So what can we do that's different what can we do that's new that you can't do in other respects and here's what we can do we can build a so I showed you this ordinary color mating device that produces a beam and as well as the halo that's surrounds it.

[00:32:32]
But on chip we can make a cascade of these to use and that's very hard to do using Glass collimated because you have to align them all but here they're South aligned by the lithography process and all we have to do is at some gaps so we have gaps in the column meter that's a little bit more tricky because it's a 2 dimensional device you have to achieve it and then you have to make sure that the rubidium vapor can escape from those gaps and so my student Charlie who worked on this was really very good with his hands and and aligning things.

[00:33:03]
You know so I should say that you know as a as a physics lab we're used to doing things that are a centimeter size but he got good at making things millimeter size and even smaller with his hands of course in the clean room it's easy to do that but in a.

[00:33:21]
When you're building something by hand it's tricky and the way he was able to do you take these. These column headers that had a gap etched into them and align them in such a way that the atoms could escape from the gaps and now here's what happens so now if an atom is defuses off of the walls it has a chance to be pumped out in one of these gaps but the atoms that were travelling down the tube did not have any.

[00:33:47]
Then this this tube looks the same as this year is no different it's only the atoms that are off axis that are affected so the on axis beam brightness is the same however the off axis atoms are all gone now. Hey that's because they've been pumped out within the source and so this device has never been built before and we built it in our lab and.

[00:34:11]
The way we actually gaps in the 1st iteration we took that same silicon die and and Jeremy Yang dice to use the deicing blade to just cut out the top. The top wafer so you can see the channels underneath it but he just hatched gaps in the in the way for.

[00:34:34]
The atoms could escape in more recent. Iterations we've also used laser cutting to remove the to make the gaps but I know that some of you are engineering students I was an electrical engineering student as an undergraduate it's nice to think of things in terms of circuits so here's a circuit model of a column 8 or so the circuit you can think of as a resistor here the atoms enter at one end and leave at the other that's like a current flow and the resistance is really the drop in the pressure as you go through and the difference between I single call a mater and a cascaded column either is that there are resistances and so you can develop this kind of shunt resistance model that shows you there's actually an exponential suppression of the flux as is you increase the number of stages.

[00:35:25]
so that's a nice model way of thinking about it but the odd axis fluxus not captured by this it simply the ballistic transport of adam's through the to you so that's not affected so what happens when you build one of these so the blue curve was the data i just showed you a few minutes ago on the collimate or the red curve is the data from the cascaded collimate are the one that has the gaps a hatched and you can see that the peak hide is almost the same in fact because of the way we make the measurement it obscures the fact that the fluxes actually exactly the same and that's because of the the fact that the the laser beam that's interrogating the adams has the finite size however the off axis fluxes completely surprised so you can really see clearly this this hype refined peak hear this one here and this one here but there's no contribution from the wings anymore ok and so the if you measure the full with asked maximum of these 2 column maters the cascaded column mater and the ordinary column mater they're not too different within a factor of 2 or so i've each other but the wings are suppressed greatly so you can't tell that really by looking at this graph but if you earn But if you plot it on a log scale and here's where the low noise current application was necessary you really see that there's a big suppression of a factor of 40 or so.

[00:36:50]
Actually the this ratio is actually more like 70 but the integrated suppression is about 40 And so this suppression was was all that with the same. So the on axis flux is roughly the same but the off axis flux is suppressed by a factor of 70 and that's because of the.

[00:37:11]
That's because of the gaps that we have into the column 8 or so although this device had these 2 devices have the same length there are only 3 millimeters in size. The collimation purity is much higher in terms of the socks off axis flux not to achieve the same thing using a macroscopic atomic beam would be about 20 centimeters in length so we did all that within 3 millimeters of length so we're very excited by this we thought this is really the possibility for making miniature devices because you can do all the stuff on the on the micro scale.

[00:37:46]
So we also have a nice patent application on this that's currently going through. As well as a paper in Nature Communications that appeared last year so what have we done since then. Before I get to that I just want to mention that you can deconvolve this and it gives you the velocity distribution of the atoms it's not actually that easy to predict that from 1st principles but we have mana Carlo methods that can be used to obtain it but the deconvolution from the spectrum matches up with this mana Carlo simulations quite well and gives you a half at half maximum of 4 meters per 2nd.

[00:38:24]
What about throughput so does this work as well as as some other methods might be and the answer is yes so. He can get up to 10 of the 10 if you tend to the 10 atoms per 2nd per 2 which is a large number of atoms you can do a lot with that.

[00:38:42]
And the throughput the total throughput for the cascaded collar mater and the single stage called a meter are consistent with what you would expect from an if use of model Ok so since then we've done some longer term tests as well to determine the reliability of this column either so we built one of these and put it in a little chamber and left it there for 6 months Ok so it's been running for 6 months and during that time we've changed the temperature every so often but we measure the flux at regular intervals and the measured throughput is actually pretty constant over time and then we can periodical the at the places that are marked by the arrows take images of the of the column meters with the camera and see that all the different channels are operational and during the 6 months we haven't seen even one of these channels get clogged so I'm very excited about this because it's suggests that this could potentially be a device that gets put into service for potentially for years and I'm very excited about it all the this this testing we've done at different temperatures means that it's equivalent to sort of 100 degree c. operation of 4 for 2 years unfortunately we didn't have the ability to keep the test going for longer but I would like to be able to do that in the future.

[00:40:06]
Ok. So. Let's see how am I doing on time and probably running a little bit low on time so I think I'm going to skip this part and just tell you that another aspect of of this kind of micro scale manipulation of the atoms is that you can position them very close to Manna structures and I mentioned that one could imagine single atoms crossing through a resonator and photonic resonator and one could imagine a strong coupling between the single atom and a single photon that's in this resonator that's the limit we'd like to get to the idea being that this is the sort of thing that people have done with cold atoms but here maybe one could do it using thermal atoms and then this could be a quantum device that could be scaled to large numbers.

[00:40:56]
So that's a something we're working on in collaboration with the group he's built some resonators for us here is some 1st resonators that he built they haven't yet been tested they're kind of the optical characterization is in process but these are resonators that are $100.00 microns in length and we would bring our collimate are very close to these to these rez rays resonators and then atoms would cross through the resonator and interact with the single photons So that's the idea it's an experiment that's kind of in process and look forward to more results on that in the future so I was going to talk a little bit about the.

[00:41:33]
About the called out of work but I see that I'm. I'm running quite a bit behind so I think I'll skip ahead and just conclude at this point. If you want to hear a little bit about called atoms please come back and talk to me afterwards I have plenty of slides and this is an experiment on vector solid tons that we just submitted a paper on but let me just conclude by saying.

[00:42:00]
By saying that it's. I just wanted to sort of present some of the opportunities for collaboration that areas where I see that there might be a fruitful collaboration between them tonic physicists and and people who participate in i.e. n.. I think I've told you a little bit about integrating precision atomic clocks with Microelectronics and so this is sort of the beginning of that Ok so what other things could be imagined maybe there's new types of sensing applications that we have not yet come up with but atoms are great sensors they can sense electric fields magnetic fields are there some new types of applications that one might like to invention and then we can try to combine them together and see if it works fabrication methods we started with silicon maybe there's other things out there maybe there are other materials that might be might be useful or novel fabrication methods as well.

[00:42:58]
So our I showed you the diode laser that we use for spectroscopy of the rubidium and for making doing the sensing in detection that diode laser sits in a completely different so look I'm sorry 35 semiconductor optical cavity that's located in a different place and comes over an optical fiber would be great if that laser actually were on the same chip as the the one that we're that where atoms are so chip scale laser development and integration at 700 meters would be really really great as well as some infrared wavelengths as well for optical trapping or guarding and in general I'd say that anything that makes atomic systems better more reliable and robust and low cost is something I am interested in so if you have great ideas about that I'm all ears so with that I'll just conclude and thank you for your attention.

[00:43:52]
Briefly let me tell you we also have a so as a as part of the quantum efforts at Georgia Tech we started a new group called pick atomic molecular and photonic instruments on chip for quantum sensing that involves a whole bunch of faculty at Georgia Tech that are already interested in this area including some folks.

[00:44:11]
And we have some collaborators outside we just got a conceptualization grant to to planning grant to write a proposal to build an institute based on integrating atomic and molecular and photonic devices with with the solid state so look forward to more on that in the future and I'll just thank you for your attention.

[00:44:45]
That. You. See. Why that. Yeah it does but the vacuum does not have to be so high it does have to be a moderate vacuum but my understanding is that probably you know 100 Micra toward to military might be enough. To be able to. That's right. You might want to.

[00:45:18]
Keep it that's right yeah and so but that kind of fact vacuum packaging has been developed so there are miniature getters that you can use for maintaining moderate vacuum so that basically the vacuum has to be such that the mean free path for collisions is longer than the size of the device.

[00:45:45]
You need to be in the molecular version. Well. I did. Things. All I can right so. You know. Better so for rubidium and alkaline metals you have to operate them at elevated temperatures it has to be $100.00 c. is typically kind of necessary because otherwise you'll just condense the active in some.

[00:46:17]
Way that's a great question it would be great to have really good thermal stability you need to have good few thermal gradients. And Ok so if you have some ideas about that that's horrible man that would be you don't go again in the uniform there are a lot of my critic going to suffer from the same problem that you're in the hot spot.

[00:46:39]
Over the life I think the one who he may be so the woman to do a lot of these type of work and it will be my good run but I want to see it with what we're. So this is more on the packaging and or all of the cheap and packages last week you didn't pack it in there will but.

[00:47:00]
These days when war this is this is a hyperbolic if you can read it they go out in the fall that for example you look at your microphone. They go up another way for a level a microphone movies you have all. They have to create a lot of interest and if you go to great local effect or even the.

[00:47:23]
Temperature and the face of the city he's everywhere because we'll be here while they work at the same day right so it's rather late it's a microscopic hopefully dusty just adopt a program that. Is exactly doing this that. They didn't even get in face change on the feet and the love you.

[00:47:48]
Need them that the way you can work incarcerate that you will not permit it and we are to go with the back will you could you could not physically when they get it will be a great is. The no death. Yeah yeah so temperature stability is important for the column it is but for the clock.

[00:48:15]
And particular for the gyroscope applications where we have multiple interrogation zones having temperature stability there could be very important because it directly translates into gyro bio stability that's what I would go good to go for you would never do that your out of brought this all. Into the loft.

[00:48:40]
Of your act will be thought that you bought the record with the radio the last yes. Really the whole there are good. But liberation. Let's take it from room to let them talk.