Dr. Israel Owens and his team at Sandia National Laboratories have used a crystal smaller than a dime and a laser smaller than a shoebox to safely measure 20 million volts without making physical contact to the electrode.

Tech Briefs: What led you to this idea?

Dr. Owens: How to measure high voltage accurately and safely using a crystal smaller than a dime and a laser smaller than a shoebox: This all started out as kind of a pie in the sky discussion that I had with a few of my colleagues. The problem we were trying to solve was: How does one measure an extremely high voltage — in particular, the kind that we typically generate on our pulse power accelerators at Sandia?

We discussed various approaches and the idea came up of using an electro-optical device that wouldn't interfere with the high energy and the radiation fields in the device. Since it's nonmetallic, it's less prone to interference and noise from the source. We knew going into this that our High-Energy Megavolt Electron Source (Hermes) is pretty much the highest energy gamma ray producer simulator on the planet. So, we saw this as a unique opportunity to resolve a longstanding issue that had been in place for multiple decades. These devices were developed back in the late ‘80s and we still didn't have the capability to measure the voltage directly or accurately. So, that was something that we just kind of brainstormed — tossed ideas around. We finally decided to use an electro-optical device because it wouldn't interfere with the high energy source.

Tech Briefs: Could you describe the setup?

Dr. Owens: There are two main parts to it. What we refer to as the remote part is essentially just the crystal and the laser beam. There are two locations: the control room and the remote area where we position the crystal. We run the laser light to the crystal in the remote location using fiber optics. But only the crystal is in the space where we detect the field. We guide the laser light that comes out of the fiber optic into the vacuum where the electric field is — the field runs through the broad side of the crystal. Then, we collect the light that comes out on the other side of the crystal. That signal is sent back to the control room, where the light intensity is measured with a photodetector. The distance between the crystal and the high voltage cathode is a little bit over 14 centimeters.

Part of our design criteria is that we wanted to use as little material as possible so as not to disturb the field. In our first design, we had a much larger apparatus — we had optical fiber and a bigger crystal setup, and it didn't work out very well. It interfered with the operation of the device — we got high voltage arcs to the sensor. So, we had to think about how to redesign it to avoid the electrical arcing.

Gamma ray generator HERMES III, High-Energy Radiation Megavolt Electron Source, is adjusted for its next shot at Sandia National Laboratories by Chris Kirtley, top, and former Sandia employee JJ Montoya. (Photo: Randy Montoya)

Tech Briefs: Your device measures electrical field strength — how is that related to voltage? Is there some kind of calculation needed?

Dr. Owens: Yes, the electric field is essentially the voltage applied between the anode and the cathode divided by the distance between the two surfaces — in our case, close to 15 centimeters. We neglect the radial spread of the field because we consider it to be somewhat negligible over the length of the centimeter crystal. So that calculation is fairly simple as far as the conversion between the two.

Tech Briefs: How do you calibrate the system?

Dr. Owens: One of the strong selling points of our system is that, in principle, it doesn't require a formal calibration procedure. Since we can rely on electro-optical theory, we're able to model what we would expect based on the known parameters. However, before we take this out to the pulse power accelerator, we do a benchtop lab experiment. This is done at lower electric field strengths to validate our calculations. I think of that as being sort of a calibration. But we've always been a little bit careful about the terminology because we think that one of the advantages of the system is that, technically, it doesn't require a calibration. So, it is a calibration in the sense that we look at the lower-strength fields and make sure that they match the theory. We get that sort of validation before taking the device out into what we call the field environment, which is one of the pulse power accelerators.

Tech Briefs: So, are you saying that the ratio between the actual voltage in megavolts and the millivolt signal is a constant?

Dr. Owens: Yes, the measurement is linear — when we see the signal on our oscilloscope, we know that it’s a direct relation — it's in the units that we want to measure. Since they're both voltages, it’s a linear transfer function between the two. It ends up that tens of millivolts on the scope translates into the megavolts that we measure on the accelerator — it's a constant, and it's linear. We've emphasized that in the paper because the other available techniques involve derivative responses.

Tech Briefs: What are some of the pulse parameters?

Dr. Owens: I can compare and contrast our benchtop experiment versus the field experiment. In our benchtop lab, we have much lower fields, about 5 KV per centimeter, but extremely narrow pulse widths — less than 2.5 nanoseconds. The system we work with can easily see the defined timing structure in that pulse. In the field, it's kind of the opposite; we have a much larger field, but the pulses are about 15 to 20 times wider than the pulses we see in the benchtop — they’re around 30 nanoseconds in width, but still quite narrow. They're very high energy and in comparison, relatively narrow.

Tech Briefs: So, you’re reading the peak pulse voltage?

Dr. Owens: We're reading the peak voltage as well as the time dependent waveform. In our group, the researchers are just as interested in the specifics of the waveform as they are in the actual peak value. Both of those parameters are very important.

Tech Briefs: Can you give a simple explanation of how the pulses are generated?

Dr. Owens: It starts out with a bank of capacitors in what’s called a Marx generator, being charged in parallel up to a very high energy. And then there's an auto switch that puts them all into series, which generates a high voltage. Then, the high voltage in our Hermes accelerator goes through a series of multiple pulse-shaping sections that start out very wide — probably milliseconds — and as the electromagnetic wave travels towards the endpoint device, it goes through a series of sections that do a pulse compression on it. All the different design elements are geared to making the pulse narrower, so that by the time you get to the end point, you've got this nice, clean 30-nanosecond pulse versus the much wider, hundreds of microseconds, or millisecond, pulse that starts out at the capacitor bank.

This laser-illuminated crystal, less than a half-inch long, is supported by a free-standing retaining structure with no physical connection to Sandia National Laboratories’ HERMES accelerator cathode. In the actual experiment, the light is initially extinguished by crossed polarizers. When the accelerator fires, the polarized light is rotated so that it leaks through the second polarizer. The leaked amount is directly proportional to the electric field. (Photo: Israel Owens)

Tech Briefs: How do you do the pulse shaping?

Dr. Owens: There's a series of elaborate sections that have giant water capacitors. There are also different types of transmission lines with characteristic impedances and lengths that allow for the compression of the pulse. It ends in kind of an inductive linear adder. It’s what they call a magnetically insulated transmission line in a series of cavities that inductively transfer the power. It all gets added up in a single rod at the end of the device. There are literally dozens of different types of sections that do the pulse shaping. The pulse shaping is designed by looking at the length of time for the pulse to travel through a particular section versus its characteristic electrical impedance. If one imagines the system like a variable dimension coaxial cable, it's changing its shape and its geometry as it goes along, and that in turn causes a change to the shape of the waveform.

Tech Briefs: How does performance of your measurement technique compare to other methods?

Dr. Owens: There are several other methods, but the ones that are most likely related are referred to as Vdots and Bdots. One of the disadvantages of these types of devices when measuring electric and magnetic fields is that they are electrically based, what we refer to as metallic-based components. Although they have some limited functionality, they are not really a good match with these high-energy systems. That is because when the system fires, there's so much electromagnetic interference that it interacts directly with the device itself — they create their own spurious currents that are a source of noise. That’s one of the big disadvantages: you have to do a calibration to figure out how it's going to work. Then, that device is taken into a high energy environment that's different from the calibration lab and subject to increased levels of noise and dynamic changes to the instrument’s electrical impedance properties. Depending on how high you go up in energy, it gets to a point where you simply can't use them because there is too much noise on the line and it's acting like an antenna that's radiating. In contrast, with our device, since it's a dielectric — essentially plastic — there is much less interaction or interference from the electromagnetic source.

Tech Briefs: Do you have any ballpark idea of the accuracy of your measurement?

Dr. Owens: As far as accuracy, precision, and resolution, we're only limited by the resolution of the instruments that we're using. We're using fairly high-speed photodetectors, and that's the last bottleneck as far as our resolving power. But what I can say about our experiment is we're probably orders of magnitude above the kind of minimum value that we would measure in an experiment. For example, we're measuring up to peak signals that are over a volt, whereas with this system we could probably measure down to about a millivolt or so of resolution. It gets a little bit more challenging once we get below that because of the oscilloscopes and detectors and the other components that contribute to some of the inherent background noise. I would say we have several orders of magnitude of resolution below what we need for our measurements.

Tech Briefs: What are other potential applications for this measurement system?

Dr. Owens: Yes, in fact, I've been talking to one of my senior managers about that because we both noticed that even though this was demonstrated on a very high energy accelerator. In fact, the device in some ways might even be more useful for lower-energy applications.

We can imagine scenarios where the crystal could be placed far off in some remote location and interrogated remotely by a laser to obtain the field and voltage information. The voltage measurement would be a little bit more challenging, but certainly in instances where one would want to interrogate an electric field, a version of our device could be used. One would be able to monitor whatever it is they're looking at as it’s evolving in time with the electric field and get fairly accurate and precise measurements.

There is interest from quite a few researchers who work with pulse power accelerators and who have contacted me and would like to use the device for their experiments. And then there have been others who are working in areas such as lightning research and a few other applications of interest who have contacted me with ideas, so there's been quite a bit of interest.

For example, I would think that utility companies might be interested, because it would give them a high voltage standoff capability and they would be able to obtain fairly accurate and precise results about the electric field, from which they can infer voltage in certain applications. In particular, the power industry is interested in identifying voltage transients in the AC high voltage power grid, and this device would have the capability to measure those transient signals.

Edward Brown is Associate Editor.