Lecture Transcript
Nick Vamivakas: Thank you for that nice introduction, Olga, and for the invitation to speak to the group today. Yeah, so I’m a faculty member at the University of Rochester, and in upstate New York, in the United States. And today, I’m going to tell you a bit about the work that we do with NV centers in two different areas of the lab that I have. Before diving into the meat of this talk, I did want to provide an overview of the variety of activities we have going on in my lab. So we do work really, that sits at the intersection of quantum optics, nano-optics, and condensed matter physics. And some of the programs really live at the nexus of these three areas. Other programs we have go maybe sit in two or one of those. And I’m just going to walk you around the slide, giving you bird’s eye view of the different things going on in the lab. If you’re interested in any of these, please get in touch with me and we can talk about them at a later date.I do want to say I know this is a seminar, so I’m happy to keep this very informal. I will hopefully keep the talk to about 40 minutes. But if there’s any questions throughout, please ask. I can’t see the chat, so if someone puts something in the chat, maybe Olga, if you see that you can just can stop me with the question and I can respond then. Great. So starting here on the left, far removed from what I will talk about today, but definitely informed by some of the stuff we’re doing with quantum sensing is we have a big activity in using tools of quantum information science to understand the classical electromagnetic field.
The idea there is that you can express the electromagnetic field or beams of the electromagnetic field as vectors in multi-dimensional vector spaces. When you make that identification, we like to think of the electric field living in a fictitious Hilbert space. But then you can import all of the machinery people use to develop approaches or constraints on quantum systems and quantum information science and use those to help understand and inform us about classical optics. That area has been very fruitful recently.
I will speak a little bit today about in the latter half of the talk, an area that we’ve been working on now for almost a decade, and we refer to this as levitated opto-mechanics. This is optical tweezers in high vacuum. I won’t say more about that, since that’s kind of a teaser for later on in the talk. We have a variety of programs in advanced sensing and scene characterization, I’ll speak a little bit more about that also in the context of NV Center quantum diamond microscope. But we also in this area, do things in near-field optics. Also, we are using tools of quantum metrology and quantum measurement to understand fundamental limits in our ability to extract information from a scene, and then use that information to either estimate parameters or make tests of hypothesis. And those systems involve not necessarily making a direct image of a scene that we’re interested in learning about, but measuring the energy received in ways that are informed by the quantum measurement machinery.
In the space of nanophotonics and condensed matter systems, we are really interested in ways is always saying to use nano structures to modulate or control the flow of electromagnetic energy. We have a big effort in using metamaterials and metasurfaces in both individual optical components, as well as optical system design. Most recently, we demonstrated a component we call a metaform which was a marriage between a metasurface and a free-form optic. The value of that particular component with an optical system is captured in this artist’s rendition of a augmented reality glass. So the curvature or the geometry of the eyeglass is very much understood as a free-form optic, we graphed metasurface or nanophotonic antennas into this special optical component. And this then metaform could serve as the element that fuses the virtual and the real world for the user of these glasses.
And last but not least, we have quite a lot of work in quantum optics and our twist on quantum optics is we like to work with solid-state materials. These could be artificial systems grown through techniques like molecular beam epitaxy, or they could be defects that exist in solid state materials. And I’m of course going to talk quite a bit about one of our favorite defects in solid-state materials, the nitrogen vacancy center.
So today’s talk is going to focus on two efforts we have in the lab. On the left is a photograph of our quantum diamond microscopes. So we have two microscopes that we built. This one is a wide-field imaging system, so in a single shot, it lets us interrogate a fairly wide field of view. You see here is the detector, here’s where the sample sits. And I’ll go through the details of the components that live in that instrument in a slide or two.
A second effort where NV centers and nanodiamonds are particularly interesting to work with, is this is an optical tweezer in high vacuum. So there’s a vacuum chamber. This is another photograph of a nanoparticle that is trapped in this high-vacuum tweezer. And then we dump a little bit of green laser light into it to give the inner photograph a sense of where the particle is just levitating there in space. Now both of these instruments use as their quantum active substrate nanodiamond containing NV centers from Adamas. So all of the data that you see on NV centers throughout the rest of the talk, they’ve come from material that we’ve purchased from Adamas.
First, starting with the diamond microscope. So why go through building something like this? Or what’s the added value? And I see three different strands to why one would want to do this. If you’re someone who’s interested in optical microscopy, optics is good for microscopy, because in many ways it’s a non-destructive approach to learning about an object. And when you marry this with features of an NV center, it gives you opportunities to circumvent the diffraction limit imposed by far-field optical microscopy and spectroscopy, so you’re getting an enhancement in the spatial resolution of your instrument. And you’ll see in one of the example applications where the spatial resolution improvement is pretty valuable. You can also use strategies of quantum control with the, essentially drive, the spin photophysics of an NV center, and that quantum control allows one to achieve a sensitivity to the signal you want to measure that has what we like to refer to as a quantum advantage. So you outperform the sensitivity of a classical sensor when you’re trying to learn about a parameter of interest.
Last, and I don’t think certainly not least, and maybe it isn’t quite as appreciated as the former two is that, the NV center is actuated by a number of different features of its environment. So if you operate this system in the right way, there’s a sensor, you have the opportunity to do multimodal sensing. And I’ll give an example of how we use this ability to actuate different features of the spin photophysics to learn about different properties of the environment. So that’s the why.
So what do we do with this? I’ll give some examples of inspection of a simple semiconductor device. But looking forward, I think we’re starting to work with groups that allow and help them understand more sophisticated devices. There’s opportunity in condensed matter physics, in particular, studying two-dimensional materials that are some sense atomically, or single-layer materials that possess interesting in our example, magnetic properties. And another application area I’ll tell you a bit about is paleomagnetism and geophysics. Maybe talking to a room of experts on this, this can be quick, but you know what’s exciting and without NV centers as we shine green light on them, we get red light out, and the details of the spin Hamiltonian and how that engages the optics provides a really nice platform to learn about the NV center environment if things are done just right.
So that the Hamiltonian is nice and it has a zero-field energy that can shift with temperature. There is a strain splitting that can also result in shifts of both the spin and the excited states of the NV center, and then there is a magnetic field sensitivity in the spin ground states, as well as the excited states that will shift upon application of a magnetic field. So the takeaway here is if you can isolate the ability of this system to engage these different fields be at temperature, some kind of train, magnetic field, there is an opportunity if things are calibrated right to learn about those properties of the environment.
Now, this is the electronic structure of the NV center, it consists of a spin ground state that is a triplet. So we have the zero and plus or minus one. And there are excited states under application of an external field. These states split and the ability to imprint in physics onto the optics happens via the way of the excited states, transition back down to the ground state either radiatively or non-radiatively. If you happen to be in one of the minus or plus one ground states, and you become optically excited up to the plus or minus one manifold, you have a somewhat efficient pathway to decay back down to the ground state non-radiatively instead of radiatively. The net result is the photoluminescence, or light that comes back from the system, this red stuff is slightly less. And so it’s indicated in this curve over here, there are particular frequencies where we were able to put the ground-state population if everything was to start in zero state, either into the minus one or the plus one.
And so if we’ve done that just right, then there was a reduction in the flouresence as to if the system started in the zero state. So if the microwave frequency is such that it doesn’t really drive population in the ground, spin state manifold to the minus or plus one states, then we get some amount of photoluminescence out, we normalize that signal. So it’s one when we’re away from that transition. So the more on the transitions, we see that reduction.
And so this type of curve, where we now measure the red light that comes out under the application of a microwave field that we scan the frequency, doing this for this particular example, with a finite external magnetic field to break the degeneracy. As we sweep that microwave frequency, far away from the ground state spin resonance, we get quite a bit of light out, that light reduces when we have that transition to the minus one to the plus one and then returns when we scan on like way back through.
I’ve also decorated this optically-detected magnetic resonance curve with how things change under the application both of a magnetic field. So we get a shift, as we see here between these two dips, so that the amount of that shift is controlled by the strength of the external magnetic field. And this is somehow how the frequency shifts with the change in that magnetic field. And then also the center is told to us by the Hamiltonian, is going to be determined by the temperature of the system. So if it’s hot, it goes one way, if it’s cool, it goes the other way. So here splitting and center are fingerprints of what’s going on in the environment with regards to both temperature and magnetic field.
And so we take advantage of that feature, or that property in our quantum diamond microscope. A bit more detail, the system is enveloped in three access Helmholtz coils, and what that lets us do is control the magnetic field seen by our sample which sits right in here. The sample itself lives in the middle of this microwave antenna that gets plugged in on top of our three axis piezo stages that let us scan the sample into the focus of our objective. It’s a relatively low NA objective with large working distance, so we have a lot of real estate to move stuff around or put stuff underneath it. That microscope objective serves a dual purpose, it focuses that green light that excites optical transitions in NV center, and then it also delivers the fluorescence light backup to not a point detector, but a CMOS-wide field detector. So we’re able to get about 100 micron-by-100 micron field of view image of our sample with this wide-field imaging system.
This is what things look like when we focus down the green laser light. So within this particular image, no microwave fields are engaging the NV centers. We simply focused on the green light, technically where we’ve taken defocus to the bit, and that you see the rings of the defocus very nicely imprinted into our NV center photoluminescence. And so this is just a spin coded Adamas nanodiamond sample on a substrate, slightly defocused green, and the features of this defocus spot are directly seen in our wide-field image. And as we defocus things further and further to flood illuminate our field of view, we see fluorescence throughout the field of view, stronger and weaker depending on exactly how well the region of the sample is coded.
So now we use this instrument to make a wide-field image. One technical detail was we discovered in doing this, that if we just swept the microwaves and tried to record images, we had some issues with noise, and so there was quite a bit of variation in the signal. As we were doing this, and this masks some of the environmental imprint, we wanted to read out of the spin-modulated photoluminescence. And so this is just some detail of how we dealt with it. The main idea is we take a photoluminescence image with zero-microwave frequency and then an image with microwave frequency that we fix applied, and we take the difference of those two to remove any slow variations in the signal that we’re seeing. And then we record this wide-field image frame-by-frame and take their difference to get the image that we actually do our data processing on.
So here is just an example. This is the fluorescence image of our sample, and now moving across our detector into two directions at every single pixel with one shot. So this is really the added value of this instrument. You don’t have to scan point-by-point to build up the spin resonance data here. We see it every pixel an ODMR signal. And so once we’re able to do that, we’ve started using this instrument to look at a variety of different kinds of samples. This one was as much a less interest, let’s say, in the device that we’re looking at here. And what this is, is there’s some nickel that we’ve coated onto a silicon substrate that we can then run current through. This was as much a calibration tool to convince ourselves we could do not only wide-field magnetic imaging, but also temperature imaging using the responsivity of the NV center electronic structure to these two environmental parameters.
And so what you see here, along the two columns, Column C and Column D, these are magnetic field maps of the device as a function of applied current. These are temperature maps of that same device as a function of applied current. You’ll notice Point 1 and Point 2 in both of these panels are columns. And those just correspond to those particular locations, what we determine the temperature and magnetic field to be as we change that drive current. Now, because the nickel is ferromagnetic, even at zero current, there is some finite magnetic field that we’re seeing, which makes sense. And then of course, as we increase the current, we expect things to both get hot and the magnetic field to increase, and that’s exactly what we see here. So this really was the device or order, the first sample that we looked at where we were able to say okay, we have this tool up and running. And then or now we’ve started to try to use it to look at systems that maybe are a little bit more interesting.
And so the first application of that is in the area of spintronics. And so what we’ve been looking at recently are FGT thin films. And so these are materials that can possess finite magnetization. So these are thin materials that can become ferromagnetic, and we are trying to look at now take advantage of the resolution that we have with our tool to study the spatial distribution of the magnetization in FGT. And so again, here is that sample chuck that we have zoomed in to where the sample sits inside of the microwave stripline waveguide that comes into the space of interest. Here is a microscope image just of the material itself before we coat it with nanodiamonds. Here is that same image, But now the registration instead of a wide-field reflectance map of what we’re looking at. With white light, we’re actually looking at the photoluminescence generated across the device regions where there’s big reductions in photoluminescence, because things are metallic.
Now what we do here is we take a big-field image of the whole device, that a finite bias. So we have those Helmholtz coils, we put them up to 0.54 or 0.5 mT, and then from a single snapshot, we take these pairs of snapshots as a function of microwave frequency, then we can back out the magnetic field in each one of these pixels like we saw on the previous slide. And what’s interesting here is when we now remove that bias field, we’re able to see the region of our FGT thin film that stays magnetized. So now we’re using these kinds of datasets to study within this material. And you’ll see in the next slide, another thin film material, really the spatial structure of how the magnetic domains are forming within these materials. So it’s really giving us again, a sensitivity and a resolution that you might not get with other techniques that could get at the same data.
Here’s a second example in the world of 2D materials where we’re again looking at room temperature ferromagnetism, but now in these very thin monolayer films. Here’s the crystal structure of either tungsten sulfide or molybdenum sulfide. And if you look now from the top view, what we have worked with some collaborators to occasionally replace some of the atoms with iron atoms. So we have iron, tungsten sulfide and iron molybdenum sulfide. And what we are interested in is learning, how does the presence of the iron in these films influence the magnetic properties of the material?
And there’s some predictions that this particular material should become ferromagnetic at room temperature with the right concentration of iron atoms within the material. And so we set out to try to help this team study that effect with our again, our microscope, and the data set that confirm that. We’re two, there is the usual magnetic field versus magnetization plots that one does when you’re studying the magnetic properties of a material. And so for the tungsten sulfide sample, basically, the details of these curves tell you whether it’s paramagnetic or diamagnetic, ferromagnetic. From this curve here, you can conclude that it’s paramagnetic. The opening of this eye or this history, this loop that we see in this plot here tells us that this is certainly ferromagnetic for comparison, or the unbuilt material, HM curves.
And then now in our microscope, we know that we get a shift in the splitting of the spin resonance dips, if there’s a local magnetic field. So in addition to this ensemble measurement that is done to demonstrate the ferromagnetism at room temperature, we have now this local measurement on the sample at one particular spot where we certainly see a shifting in the peaks of the spin resonance, and this is indicative of the magnetic field that is being generated by this federal magnet. In the other material where it wasn’t expected to see any ferromagnetism, we see no shift in the spin dips. And so this is confirmation, in fact that this material, this 2D material doped with iron is giving us ferromagnetism. And this is really the first observation of that affecting this material, and the diamond microscope was part of confirming that this was indeed the case.
So the third application is, so now we were first we’re looking at small things, we’ll keep looking at small things, but it’s to study something big. And so full disclosure is like going to these next few slides, we really bring the hammer to this party, I guess. I’m not a geophysicist, but I have a very smart geophysical collaborator. And so basically, he is interested in so through him, I’m interested in questions about the formation of the Moon. And so remarkably, paleomagnetism gives one away to learn about the formation and evolution of any body, but in this case, the body that’s being studied is the Moon. And so there’s questions, just really basic questions about how the Moon was formed, was it through a gravitational creation process? Was there an impact between two bodies that then gave rise to this third body? And so really, I’ve come to learn, we don’t really know the answer to that question.
And what is surprising is, as I just said, is that magnetic properties at the surface of a body can tell you quite a bit about the origin story of that body. And so just to remind you, our Earth has a magnetic field associated with it. We’ve probably all known that since we were small children, if we ever played with a compass, it’s responsive to that local magnetic field and it helps one get around. Maybe you didn’t know at the time, and I didn’t know till much later that why the heck do we have that magnetic field. And it has to do with what’s going on inside the Earth and actually how it was formed.
And so the reason we have that external magnetic field is that basically, in that formation process of the Earth, it led to a inner molten core that is completely produced or extremely solid iron core, which sits in the middle, and then we have this ionic core of iron and nickel that sits outside of that inner solid core, and because of that, the liquid is able to move. And so there is rotation, and there’s convection motions of this iron-nickel mix that creates current loops. And as soon as you have a current loop that creates for us a magnetic field. And so studying that magnetic field on the surface of the Earth, let’s one learn about history of the Earth, of course, but also details of not quite the inner core, but this outer core of the Earth. And so we call this magnetic field is generated through this convection and rotational motion of this liquid, the Earth’s dynamo.
Now, right now if you were to make a measurement on the surface of the Moon, the way you would do that, as we don’t take our diamond microscope up onto the surface of the Moon, some stuff comes back from the Moon to us. And right now, if you were to measure rocks from the Moon, there is no indication that there is a magnetic field on its surface. Now, just because there is no lunar dynamo right now, it doesn’t mean one never existence, and the existence and the features of this dynamo tell you about, or provide a potential explanation for how the Moon might have formed.
And so what I’m showing here is a plot from this nice review paper on the lunar dynamo. And here is a time axis, and here is the basically the magnetic field, what is called paleointensity and paleomagnetism. And there’s some theories about what the strength of that field could be. And what was interesting in these data points that were plotted here is that if you look back to a fairly long time ago, there seems to be a window when there was a lunar dynamo, and more recently that has gone away. Now, the way that these kinds of measurements work, is you actually take a rock from the surface of the Moon, and inside of this rock, there are magnetic inclusions ones that carry what’s called a remnant magnetization associated with them. And this remnant magnetization gives you a magnetic field that lets you make a statement about the generation process for that magnetic field.
Now, the rocks are dated in one way to figure out this axis, and then other magnetometers are used to try to measure the paleointensity. And so questions related to the dynamo, was it only this window of time? Could we learn something in this window if we had more sensitive sensors? These are all open questions that are people are actively looking at to this day. So of course, there’s an opportunity here, quantum diamond microscope, and this is how I teamed up with my partner at the U of R, John Tarduno. He came to me, just to much better describing what I was just sharing with you. And he said, “Nick, basically, if we can measure small rock with high sensitivity, we have an opportunity to put better resolve but also different data points on this kind of a plot.” And the reason is, the most sensitive magnetometers like a squid magnetometer, are as sensitive as NV center diamond microscope, but the spatial resolution isn’t as good. And so these inclusions are small within the rocks, and here’s an image of a rock. So this is what John brought over to our lab.
This was brought back, this is like the only thing that gets my kids excited when I tell them we look at a rock from Apollo 11. It came back from the Moon in the lab, they can’t quite believe that. But they’re magnetic inclusions that can be identified through other microscopy techniques, that we then want to learn about the magnetic field strength of those magnetic inclusions. And being able to have the spatial resolution to really look at individual included, and not having to average over the ensemble of them, allows you to have a more sensitive and better result measurement.
And so one challenge of doing this is that if magnetization is our memory, we don’t want to do something that can potentially distort or bias or erase that magnetization. And so most diamond microscopes and ours, typically would put through that Helmholtz coil, a DC magnetic field as the background to break that spin degeneracy. But so actually, a solution that has been really nice in these experiments is the nanodiamonds. Remember that middle term in the Hamiltonian, there’s a spin, excuse me a strain differentiation on the spin ground state. So that’s another way the degeneracy is broken in the material.
So by working with strain nanodiamonds, we don’t necessarily need a background magnetic field of the same level to do our quantum diamond microscopy, and so it allows us to look at samples that are sensitive to the environmental magnetization. In addition, we came up with not so sophisticated, but certainly useful way of looking at these rocks, and essentially making PDMS films of nanodiamonds that would allow us to put the film. So we ideally would like to have an array of nanodiamonds separated by a bit more than our diffraction limited spot size that we could then put on the sample and make a magnetic image, and if we could then shift that controllably, that would be fantastic. So we’re working on that right now. What we do is, a severely low density of nanodiamonds within a PDMS film, we can put it onto our sample, lets us look at a sample but then take it off and then put a new one back onto the sample, and it helps just for throughput as we’re making these measurements.
And so this is exciting for us. Our first ODMR images on actually this inclusion right back here. And so we’re able to get ODMR splittings and magnetic field strengths that are reasonable for what one might expect from the type of rock that we’re looking at. So this is just fresh out of the oven, and so I’m hoping there’s going to be a lot more to come with this in the next weeks and months as we continue to look at these samples and work with John on the Sun’s project. Okay, so I see it takes us to about 12:30. We change gears a little bit, and maybe go a bit… I won’t rush through this, but maybe it’d be a bit quicker, so we can be done by quarter of.
Yes. I also know from when I’m attending talks around 40 or 45 minutes is when I start if I can make that long. So, being cognizant of the audience helps. So the next thing I want to talk about is our optical tweezer in high vacuum, either levitated opto-mechanics or when we’re working with nitrogen vacancy centers and nanodiamonds, levitated spin opto-mechanics. The optimum physics part of this is, there is essentially a time average force a dielectric particle field when it’s at the focus of a high NA objective. This plot is capturing how the intensity of that focused field looks, okay? The potential that results from that focus field and the force that the dielectric particle will feel is a result of the gradient of the race.
Now, if you’re an expert in optical tweezers, you know there’s also a scattering force that can come into play, and the scattering force, instead of restoring the particle to the focus will tend to kick the particle along the direction of your tweezing beam here. Because our particles are much smaller than the wavelength of our trapping beam, we can safely neglect that contribution to the force. And so why that is nice is, looking in this again direction is transverse to the beam, propagation direction here. We have a oscillation that, at least, for small displacements from our focal point is going to be very much like a harmonic oscillator. And we can recast now, the spring constants of our levitated nanoparticle in terms of properties of our trap in the polarizability of the particle. This is the power of the trap. This has the numerical aperture and wavelength.
And when you do this, looking, say at one direction, this is the X direction here, things along the optical axis, which could be Z, you see are slightly different than the things that are in the direction transverse to the optical axis. At the end of the day, the motion of this particle in this X direction, if the environmental pressure is just right, becomes that of a harmonic oscillator with fluctuating force that depends on details of the environment, as well as the trapping beam itself. And even possibly, when we start making measurements of the particle via the light scatters away, quantum mechanical back action that acts on the particle.
In our trapping experiments, the size we work with is typically about 100 nanometer diameter, here’s the masses. And then the types of resonance frequencies that we get for these optically levitated particles is in the range of 100 kilohertz. Particle size is important because the wavelength with respect to the particle size determines whether or not we can neglect that scattering force. So if we tried to put, say, a half a micron particle into this single beam, gradient force trap, the particle is just too big and always gets kicked out of the trap so we can grab onto it.
So it’s exciting for us, and it’s something we’ve also been working on for a number of years, is using nanodiamonds. So using dielectric nanoparticles that are also optically active. And what’s interesting for that is the spin is a potential handle to do interesting opto-mechanics experiments with. And the idea is that, I guess I haven’t said, why the heck would one want to make a mechanical oscillator using light in a vacuum? But what’s nice is that the fact that you decouple your mechanical oscillator from any kind of substrate, you have essentially quieted or turned off channels for dissipation and decoherence. So you can end up with a mechanical oscillator in this optical glove that has quality factors that can approach 10 to the 10, 10, to the 11, 10 to the 12. And so these are extremely high-Q mechanical resonators, and it’s just because you don’t have any connection to a substrate.
And if one were now able to take this harmonic oscillator and take energy out of it, and put it into its quantum opto-mechanics transition it from a classical to a quantum harmonic oscillator, and just remind you that the degree of freedom I’m talking about is the center of mass of this particle that’s caught in the trap. And there’s an opportunity to do quantum opto-mechanics where you would have extremely long coherence times and potentially huge mechanical sensitivities that just result because of this quality factor and the fact that it’s so isolated from the environment that it sits in.
And because of that, when you add in the spin degree of freedom, there’s a number of proposals to how one might couple this mechanical motion of the particle, the nanodiamond, in this case, to the spin that’s living inside of it. And this is what actually got me excited about these experiments now, probably eight or nine years ago, that this sort of holy grail experiment we still are slowly making progress on, would be to take this levitated nanodiamond. And I’ll describe in a slide or two how one can actively take energy out of the harmonic oscillator, put that into its quantum mechanical ground state, so the center of mass motion, which is a harmonic oscillator, is into its ground state. And then doing this in the presence of a magnetic field gradient, where you optically modulate the ground state spin of your NV center that lives inside of this nanodiamond. And if you do this just right, you can create exotic quantum mechanical states where the center of mass of the nanodiamond itself is displaced depending on the orientation of the spin.
All right, so you can create these somewhat mesoscopic entangled states between the particles center of mass motion, and the spin lives inside of the nitrogen vacancy center. And so because of that possible feature to make cat states in the system, there was proposals to do extremely sensitive mass spectroscopy. Over here is an example of just doing interferometry with this entangled spin nanodiamond center of mass state. And there’s a variety of other proposals how you might use this to do matter-wave interferometry, and all of these things at the end of the day require being able to generate this state. So it requires learning, knowing a lot about the properties of the NV center of spin, as you try to operate the system at high vacuum and ultra high vacuum. Is it possible to do these in an optical trap? Are there other obstructions that come up that make it not possible? And so this is the path we’re going down right now as we slowly make progress on trying to get to a point where some of these exciting proposals could actually be realized in the laboratory.
The apparatus, the science apparatus where all the magic happens, we have a trap laser that is modulated by an electro-optic modulator. It also goes through an acousto-optic modulator to slightly shift the frequency with respect to the EOM path. Now the EOM path is the laser that creates the potential well the particle sits, so it has a much higher power than the AOM path. The AOM path in this particular experiment is just a probe, so it becomes frequency-shifted, so it doesn’t interfere with the EOM path. And then it also allows us to study the motion of the particle when it passes through the chamber. So we have a collection of detectors on the back end of the system, that just from looking at the scattered light can tell us about the motion of the particle and the focus of our trapping beam.
In addition, don’t forget about the fact that within this nanodiamonds there’s NV center, so we have a green laser line that comes in that will optically actuate the NV center, we have a microwave and spectrum analyzer rigged up into this chamber. So this is kind of a blow-up cartoon of what’s going on in the focus of the high NA trapping objective. And then there’s a channel to collect this fluorescence and either look at the spectrum, do correlation measurements or energy correlation measurements on the light that comes out to learn about the photon statistics of the light that we’re receiving. So there’s a lot of bells and whistles on this. The most important one if you want to take energy out of the system is taking what you know about the position and we built the suite of customer electronics in collaboration with Lukas Novotny’s group that allowed us to measure the position, and from knowing the position, change the details of the EOM trapping path. Remember, we can modulate the power in such a way that we can actually cause dissipation or an optical damping to happen for this levitating nanoparticle.
What a data sets look like, oops, typo there, that should be an I. Yeah, so here’s just looking at one of the channels as a function of time to convince you that yeah, we do see harmonic motion. If we look at the Fourier transforms of the three different position channels that we measure, we see these very nice peaks in the photocurrent spectrum that occur at the frequency of the oscillation. We have three separate harmonic oscillators, because as you might remember, when you focus light in particular, a laser with a high NA objective, you get asymmetric focal volume that is longer along the optical axis, it is narrower in the transverse direction, it is orthogonal to the incoming linear polarization of the life, and then it’s a bit wider than the direction that’s parallel to that linear polarization.
So because the curvature along these three directions is different, it results in an optical spring constant and is slightly different, that then results in frequency differentiation for the three directions of oscillation. So in fact, this optically-levitated nanoparticle is moving in three different directions with frequencies that are all slightly different. So it’s really a multimodal harmonic oscillator. That’s really useful, because what you can do is you can measure any one of these three oscillation directions, and then feed back onto it individually by modulating the power as seen by that particular oscillator. And so that’s exactly the kinds of things that we’re doing. So looking at one of these current frequency or photocurrent spectrums for the X oscillation mode, here in the system we have in Rochester, we’ve taken this down to about 800 millikelvin by just actuating the power along that particular direction in just the right way.
So I think the best way around, maybe is like 1000, if you think in terms of photons for these oscillators, maybe 1000 photons in a given mode. We’ve actually spent quite a bit of time doing other experiments related to like mechanical lasing in this system, but not working with nanodiamonds, working with silica beads. Groups have shown you can in fact take this system all the way down to its quantum mechanical ground state by feeding back on the motion in just the right way. So half of the story is there at least getting to the ground state of the center of mass, and if there is ways to do this with the nanodiamond and bring the spin into the story, there’s lots of exciting science that can be done.
So with the nanodiamonds in the traps, the kind of things we have done is measure the photoluminescence spectrum, done this with nanodiamonds that contains just one NV center. And that’s evidenced in this kind of energy autocorrelation plot. This plot here essentially takes the stream of photons leaving the trap and puts them on a beam splitter. And just ask the questions, if I get a click on Detector 1, what is the likelihood I get a click on Detector 2? Detector 1 is the transmission detector, Detector 2 is the reflection detector. And asking that question when there’s no time delay between when I asked it, if I have a stream of single photons, and that better, there’s going to be zero chance of that event happening because my photon has reflected it can’t transmit or vice versa. And so this is a typical trace that one sees if they want to demonstrate that the light source is giving them single photons.
We just wanted to give into ourselves, we had an NV center and our nanodiamond, and we had one NV center only. And then we’ve done things like optically-detected magnetic resonance on the trap. And there’s some details about doing this in high vacuum being held with a laser beam, instead of a substrate that you have to be mindful of. For example, we’ve noticed there’s a shift in these ODMR spectra, that happens when you first start to pull vacuum. But after we make that pull, it stays pretty much consistent as we continue to do the experiments. And being able to of course, look at the ODMR spectrum as a function of power, we can learn about the temperature of our NV center. So here’s an example of just using that sensitivity of the center of the ODMR spectra to learn about the temperature of our mechanical oscillator and a trap.
And then finally, we’ve maybe kind of recently looking to see if we can coherently control the spins. And so what we’re showing here is from our ODMR data sets, doing something that is time resolved, where we now vary the window of time that we actuate with our microwave to get some Rabi oscillations. And the two different data sets here, the blue and the red are looking at the same and NV center, but what we’ve done is we’ve chopped the trap. There was a bit of a detail, the optical trap is influencing the spin photophysics here, and that’s evidenced by, for example, in this particular [inaudible 00:42:48] contrast is better when we shot the trap, as opposed to let it just continuously hold the nanodiamond. And then in these two configurations, either with it trapped or just letting it sit continuous, we looked at the Rabi oscillations and the light that we get back in the film.
And from that Rabi oscillation, we can extract T2 times. And so when you cross the range of powers that we’re looking, the T2 is stable. That’s not huge right now, but at least it’s consistent with what one would expect if we were to look at like macro bulk type 1b diamonds. So it’s nice to see that and to get this data set. Understanding the spin coherence time is going to be important if we are going to push the envelope on trying to do these kinds of superposition experiments between the center of mass and the spin.
All right, I call upon just a few minutes, past 45 minutes, I appreciate everybody’s attention, and it was really nice to share with you some of the work that we’ve been doing both with our quantum diamond microscope, and in this area, levitated spin opto-mechanics, both enabled by the material that we get from Adamas. And of course, anytime you hear these talks, the person doing the talking needs the work of many folks. I’ve been fortunate, at least in these activities to collaborate with a number of faculty members both at the U of R outside of it. A number of really talented students and postdocs have worked on these experiments and other experiments that we do in the lab. And I’ve been fortunate to have funding from a variety of different sources within both the Department of Defense, as well as the National Science Foundation, and even the U of R has also been very supportive in the work that I’ve been doing. So thank you for the invitation again and all your attention, and I’m happy to take questions if people have any.
Olga Shenderova: And Nick, thank you very much on behalf of the audience for your beautiful talk which is so wide range of topics you touched, it’s unbelievable. So probably while our audience think about their questions or… Actually they started to get already in the chat.
Nick Vamivakas: [inaudible 00:44:58] question, it said: Do magnetic impurities in nanodiamond influence your results? That’s a good question. So I haven’t thought too much about that, but anything that is going to change the magnetic environment of the NV center will influence the result. So, if this magnetic impurity may be a little more explicit, if we were doing something with trying to measure the spin coherence times, and there are magnetic impurities that are fluctuating within the nanodiamond in the vicinity of the NV center, then of course, those fluctuating impurity magnetic fields are going to de-phase the spin physics that we’re observing, and it’s going to influence the results that we have.
To be honest, the bigger problem we have, or the biggest obstruction right now, is that we can’t get to the vacuum levels that are required to do some of the experiments that let you take the center of mass down to the ground state. So we’re not sure if that’s just a feature of… It could be just not a technical issue, it could be a fundamental issue. Just the fact that we have even one impurity in this case, the impurity for us is the thing we want to talk to the NV center. Within that narrow-down, just the fact that, that will absorb some light, and as you start to pull vacuum, you’re turning off dissipation channels for heat. And so at some point, the nanodiamond just can’t withstand the amount of optical energy that’s been converted into heat within the material, and then it can’t persist. But that’s slightly speculative, but we’re trying to figure out how do we even get to temperatures, or excuse me, pressures that are low enough to allow us to start to try to see if we can put it to the ground state for its center of mass.
Yeah, the biggest particles we have trapped have had diameters of 300 nanometers, those weren’t necessarily the nanodiamonds, these are silicone beads, and that’s condition on the fact that we are working with a… I didn’t say the wavelength of our trapping laser is 1064 nanometers. If we were to move to a longer wavelength laser, we could trap larger particles. Now, if we wanted to trap larger particles, because of… You only use one laser, then you just have to have a longer wavelength. Of course, we could trap particles that were bigger in our system, if we were able to turn off the scattering force that I told you we neglect in our experiments. So scattering force pushes along the optical axis in the direction of your laser beam, it’s doing the trapping.
If we wanted to trap, say, a micron particles in our trap, what we would have to do is on the end, on the other side of our focusing objective, so the basically the light comes down, up, as recall, mainly we have to put a mirror. And if we could create a standing wave trap with the 1064, then we effectively cancel out the scattering force, and then the one laser can trap the particle, but now you need to have it propagating along the two different directions.
Olga Shenderova: There is another question from [inaudible 00:48:04]
Nick Vamivakas: Yeah. Do you have the ability to select which… I wish we did. Right now the way we load the trap is we use a nebulizer and you create an aerosol that contains whatever the object we would like to trap, and we just blow it in to the… Kind of a lousy thing to do if you’re trying to pull the vacuum, but we blow it into the vacuum chamber, in the atmosphere, and then collisions with the gas, remove energy. And if that happens within the focal region, then the particle can get stuck in our optical trap. And then like in this image we see right here, you see a bright glow and it would persist, and then we’d start like pumping down the chamber. That’s the operationally how we do it.
There are ways with peers or electrics to de-laminate nanoparticles from a surface and load the trap that way. That approach is very important if you want to go to ultra high vacuum because you don’t want to be spraying stuff in, and you want to have a leeway if you’re losing particles in your trap to reload it. So in vacuum, you could kick a particle up and maybe have a laser beam that comes down every so often to try to be that thing that takes the energy away, like another optical, of course, and then load it into your trap.
When I have people in the lab, the way from where we spray to the size of the spot, it’s like to get the particle loaded in the trap, give it in yards. But it’s like hitting a hole in one on a 500-yard golf hole getting this particle into that trap and you think about the sizes and the distances. The reason we’re able to get holes in one is because we also are taking like a billion or million shots at the hole every time we do a puff of the nebulizer, so.
Speaker 3: Yeah, I’m actually working on a technique for sorting diamonds according to their properties.
Nick Vamivakas: Okay. And so you would be able to… Maybe can you unpack that a little more from your… I’m interested.
Speaker 3: Microfluidic sorting.
Nick Vamivakas: Okay, so properties being size or like what?
Speaker 3: Well, ideally I’d be able to anchor particles in the microfluidics tests, things like perhaps their coherence time, which will take some time, and then release them according to their properties. But that’s kind of a holy grail, it’s far off.
Nick Vamivakas: Yeah, it awesome if we could do things like, of course, as you said, just it would be a holy grail for us too to be able to take exactly the sample we knew we wanted and put it in there, and maybe take it out and then put it back in. But right now, that’s just not something that we’re able to do.
Olga Shenderova: And maybe also I’ll ask related question. Nick, for your limitation experience of putting silica shell on surface of nanodiamonds, wouldn’t be useful if you reduce a refractive index, you can maybe use larger size of diamond and maybe it also can stabilize in the center performance? What I’m curious is that when you put diamond… There are such as groups, right?
Nick Vamivakas: Yeah.
Olga Shenderova: And those are some sp 2 and where environment is changed certain groups start to be even eliminated from diamonds obvious. And there’s now the carboxylic groups on the solvent [inaudible 00:51:38] bonds will start to form some sp 2 which is… And this can be maybe a source of electron can be skewed from this center. So stabilizing this surface with some shells probably might be useful. Do you do anything inside the reaction?
Nick Vamivakas: Yeah, so that’s a good point. And so I didn’t really speak to this in the talk. When we first tried, maybe in 2013 to do these experiments, we thought being a levitation, exponential nanodiamonds, oh, this is going to be easy. We’re just going to put these into the trap, we know how to trap beads, we’re going to build to a number of interesting things because of the NV center in the spin. And as experiments sometimes go, like the first day, we actually started trying. Maybe like the third or fourth time, boom, we caught an NV center. We’re using NV centers that had around 400 or 500 the ensemble nanodiamonds.
And after we saw this one that was bright, and we had photoluminescence, and maybe we turned off a system went to lunch and came back. And then it was weeks, we couldn’t see anything anymore, and weeks and weeks and weeks. So we were not sure what to do. So we actually there was a collaborator, I can’t remember her name, it was a while ago, but we had a paper in nature photonics and we first did this and she’s on this, Eva, and I can’t think of her last name, but we sent them nanodiamonds and then we had them coated in silica. And it just turned out when we sent them the nanodiamonds, we sent them both the versions that had the ensembles and the nanodiamonds that had 1 to 4, just because they were going to work on it. And just once they started doing it, they said it was easy for them to coat a number of different sizes and densities.
And so when those came back, we had hoped that they were going to passivate the surfaces and some of the things you mentioned might improve the ability to trap, improve photoluminescence, and it still turned out that the larger ones didn’t fluoresce. But we said, “Let’s just try the smaller ones.” Because we thought at that point, we’re having an issue collecting photoluminescence signal as much as anything. So we put the smaller being in 1 to 4 in with the coating. And like magic every time we trapped one, we’re seeing photoluminescence. And occasionally, we’re seeing single ones because it just like the randomness and how we were doing it. And so then we thought, okay, great the coating helped but then…
So wrote the paper, and then we just said, “Well, let’s just go back, and we…” Because we never actually tried the one before un-coated just because we thought they were going to be too hard, we wanted to start, it was going to be right in the areas. So I went back and it turned out even those ones that weren’t coated, but had just a few nitrogen vacancies also were routinely fluorescing in the trap. And so at least for the initial experiments, the coating wasn’t making much of a difference, but I do think surface contamination or other non-ideal surface properties, in the long run need to be dealt with if we’re really going to be pushing some of these experiments I was talking about. Because I think that those can be part of a source of maybe even like a local heat sources through absorption that cause problems and going to lower temperature.
As you said, if it’s close to the surface because the crystal size is small, that’s going to influence the spin physics we would like to see. And so I do think shells are a useful way to go, and it was just the way we went to them and then moved away from them and haven’t gone back yet, just because I don’t think the other side of the experiment is quite there yet where the shells are necessary.
Olga Shenderova: Thank you. Do we have more questions? Oh, I can ask him another one. So for your experience of this quantum microscope, can you please elaborate a little bit more on nanodiamond? Is it, you use of photos experiment to sprinkle them over substrate for making of temperature and magnetic fields?
Nick Vamivakas: Yeah.
Olga Shenderova: What sizes do you… And maybe mostly for us is kind of people who try to improve quality of the material you work with, what would be your desirable quantities photos that, what you would like to see from them?
Nick Vamivakas: Yes, we’ve done those experiments with, I have a list, I didn’t actually put it into the presentation. And but we’ve worked with the high-density nanodiamonds in those experiments, and the sizes can be ranged from 100 to a few 100 nanometers. To be honest, and I didn’t say as my talk for where we are, what would be best if you can do the sorting. To just let’s say, the dream substrate would be a sorted one where we knew each the photophysical spin physical properties of the nanodiamonds. And then we also had an array of them where we were able to situate them in space with a spacing that we could dictate. And that would be really useful, and add them in a…
That’s why this PMS stamp is a half-baked way of getting there. But if we had a way to stick something on the rock experiment, especially something we wanted to measure, because we’ve calibrated all of the sensors already, you can move that around and measure different samples. So at the moment, the way we’ve done our measurements, because we’re not really pushing the quantum control and trying to get the highest magnetic field sensitivity that we could have, we’re not so worried about the… It’s not the sample quality that’s limiting us right now, it’s just our ability to prepare the sensor bit of the diamond microscope that if we could do it more regular with a bit more certainty of what was going on at each location, would be really nice.
Olga Shenderova: Okay, thank you. So please let us know if there are more questions. It was very comprehensive talk, very educational. And probably somebody can have more questions in the future.
Speaker 4: I would have another question concerning the optical tweezer that you set up, and it’s about, have you tried different wavelengths in your setup or once, for example?
Nick Vamivakas: I didn’t catch that last part.
Speaker 4: So longer wavelengths, for example, closer to the emission wavelengths of the NV center?
Nick Vamivakas: Yeah. So we have not tried to trap with the optical transition wavelength of the NV center. But an experiment on a sheet of paper, not in practice yet that we’ve talked about, was to stick with the 1064, or even you can do longer. We’ve done slightly longer wavelengths, and we’ve done slightly smaller wavelengths for trapping. But what’s interesting here is, the size of the particle with respect to the trapping wavelength, we would like that to be a large disparity. So actually longer wavelength if you can work with slightly bigger particles. But as long as the particle is like, say the diameter or the radius is about a tenth of the wavelength, then we are in the regime where we can trap with one beam.
And so the experiment on paper that we haven’t tried to do yet is, and I think this is maybe what you’re alluding to or thinking about is, we have a defect or an atom in a solid now. So all of the tools people use, an atom and ion trapping with laser beams can be brought to the bare on the internal structure of the NV center. And so the experiments we wanted to do that we haven’t done yet would be to with a laser that is resonant, let’s say with a zero phone online in the NV center, maybe not a laser, maybe other light sources try to see what is the smallest optical force we could measure.
And the optical force now would be on the optical transmission of the zero phone online. And would there be a way that we could take, say, our laser coherent state and attenuate it down, so on average, it has some maybe few, maybe one photon associated with it? And can we measure the radiation pressure as we send these individual photon wave packets through the levitated nanoparticle? So the center of mass might displace a little bit every time you get an optical force on this resonant transmission within the NV center. So that’s the long-winded kind of [inaudible 01:00:58]. The show is, we had never tried to trap the particle with the optical transmission of the NV center without some other glove being.
Speaker 4: Okay, thanks.
Nick Vamivakas: I guess on the nanodiamond arrays, really the field of view we have is what? 100 micron-by-100 micron. The spot size is about a micron. So probably every three microns on just even a square lattice would be nice. So we’ve talked about… We haven’t done anything, it’s easy to talk, hard to do always, if we had a way to create, what’s that? I’ve been talking for so long, I’ve lost the words I planned to say. Make it a lattice for the laser beam, I don’t know what the right word is, like make an interference pattern.
Olga Shenderova: A mesh making.
Nick Vamivakas: Yes, yeah, like you have some laser interference pattern that creates this array structure that we want, and then do it in like a UV-curable floor or something where we could then optical tweeze nanodiamonds and all of the high-radiance points and cure the thing, turn the lasers off and then go away with it. That was one thing with something like that, but that was just like a… Again, we didn’t really move forward on it because just like a person power always is sometimes limiting until the ideas are pushed forward. But it really is outside of the, a few diffraction limited spot sizes, so there’s no crosstalk between the different nanodiamonds.
And of course, the more precisely that we can position those with respect to some reference point, is going to be valuable information because ultimately, at least in the rock experiment, you’re going to want to do some measurements that lets you really figure out things related, not just to the local magnetic field that you’re measuring, but you might need to make like two different measurements to then do an inverse to what is the details of the magnetization that is sitting within that particular rock. So there’s inverse problem and inverse imaging thing that would have to happen.
Olga Shenderova: And you get the words also a paper by MIT group on distributing the diamonds over electronic device in measuring local magnetic field which is associated with this local electron current, as well as temperature.
Nick Vamivakas: Yeah.
Olga Shenderova: So see, diamonds will just sprinkled over the surface without any metrics associated, it’s kind of very easy way for your rock, local magnetic field on your rock system. Otherwise, this would be a problem just to spread diamonds and-
Nick Vamivakas: So we initially were doing it that way, and we were also on the nickel devices that I talked about that’s how all those were measured. It just cleanliness of the device and being able to… We like to convert on a sensor that, like calibrated sensor that we could then use on multiple… So coming back to the point about it, having the same NV center and every nickel-
Olga Shenderova: Okay.
Nick Vamivakas: Yeah, that’s it. And we’re bummed.
Olga Shenderova: Makes sense.
Nick Vamivakas: We know that paper, we know that MIT group and we know that work. They were a little ahead of us and so we were bummed with it when we saw that because our work with magnetic in temperature-sensing. We needed to come up with a way to differentiate what we had done from what they had done when I’m finally finished, so, yeah. All right.
Olga Shenderova: Okay, thank you very much. I think I don’t see anymore question. And it was actually very, very long discussion. So your talk gave theories, a lot of important questions and high interest in the audience. So, again, you all, thank you for your very inspiring talk today, and we hope it will be more collaborations within different groups.
Nick Vamivakas: I appreciate the invitation, Olga. Nice to see everybody and talk soon.
Olga Shenderova: Okay, okay. Bye-bye everybody. Stay safe, healthy.