Radiation detection: There’s an app for that

Abstract

The nuclear accident at Fukushima, which led to the release of radioactive fallout over densely populated areas, highlights a broader problem: What is the best way for the public to detect and monitor radiation levels in urban areas before, during, and after a major radiological event? The authors describe a system that would incorporate inexpensive radiation detectors in mobile phones equipped with the capability of continuously transmitting data to a central server. This server would correlate data from a large number of detectors carried by people moving randomly in an urban environment, locating fallout hot spots and illicit radiological sources carried by terrorists. The authors discuss both technical aspects of the system and some practical issues involved in implementing it, including cost.

How can society detect radiation that is potentially harmful to the public? That deceptively simple question has, until recently, brought forth a laboratory expert’s answer: Get a survey meter and the training in health physics necessary to use it. But that response makes less and less sense as technology becomes ever more ubiquitous. In the United States, for example, 83 percent of adults have a cell phone, 42 percent of which are smart phones.

Making this connection in 2006, we—the authors—proposed that mobile telephones could be equipped with low-cost gamma detectors as a means for dealing with large-scale radiological catastrophes. At the time, the suggestion was met with a mixture of curiosity and skepticism. Government officials could not imagine how such a system would be useful; industrial partners struggled to see how they could make money from it.

Now, in the wake of the Fukushima disaster, which spread long-lived radioactive isotopes like cesium-137 across wide areas, the potential value of such a proposal seems obvious. The system we’ve patented ¹ works by equipping mobile phones with inexpensive gamma-ray detection sensors. As seas of cell phone-carrying people move through urban areas, those with radiation-detecting phones would send digital packets containing information on gamma-ray counts to computer servers. If enough people used these phones, they could quickly sweep large percentages of the ground space of a city multiple times.

The data set created with a cell phone-based system of radiation detectors would create a two-dimensional map of radiation baselines—not a single point, as is the case for a traditional detector. Once this baseline surface of radiation-count statistics is established, it could include—and correct for—most natural and human-made background radiation. Only deviations from that baseline level would become reason for further investigation.

Turning gamma detection into an Android or iPhone app used simultaneously on millions of mobile devices could address multiple public safety and security threats in a way that limited numbers of survey meters cannot. A hand-held survey meter is essentially zero-dimensional; it measures radiation right here, the point where it is. Such a single-point measurement does not give the normal background radiation at the spot, and without that historical measurement, it is essentially impossible to conclude whether the radiation being measured is in fact abnormally high. Attempts to interdict the transportation of illicit sources of radioactive material or track fallout hot spots after a nuclear accident with one or a few detectors are, therefore, very unlikely to succeed. In 2007, the US Homeland Security Department and area authorities proposed a detection system for New York City that would use rings of radiation detectors. Such a one-dimensional fence is clearly better than a zero-dimensional fence. But even a one-dimensional approach raises the question of whether the fences are close enough together.

In a two-dimensional network, however, millions of mobile telephone users would sweep over the populated area of a city, providing many measurements and improved statistics for background radiation. Random walk models indicate that in a city such as Indianapolis (which had a 2010 population of 829,718 or roughly 2,294 people per square mile) a complete measurement of the populated surface of the city might occur on the order of every 24 hours. In Chiba, Japan, similar statistics might be acquired every hour due to the higher population density. The principle of the mobile phone-based system is straightforward: The higher the effective detector density, the better and more rapid the identification of radiological threats, whether they be terrorists’ weapons or nuclear-accident fallout.

Radiation detection for the masses

In 1895, Wilhelm Röntgen was experimenting with vacuum tubes and noticed a pale green light on a small cardboard screen painted with barium platinocyanide. The barium platinocyanide was being excited by something emitted from a nearby cathode ray tube, even though the tube was encased in a light-tight covering, and then de-exciting by emitting light, or fluorescing. After confirming the fluorescence through other experimentation, Röntgen called the mysterious, penetrating emissions from the cathode ray tube x-rays, and their discovery led to his Nobel Prize in Physics in 1901.

Since that time, many different classes of detectors have been developed to measure various forms of ionizing radiation; ² the detectors vary in size, price, functionality, and scope, ³ depending on the application, from a 9-square-millimeter thermo-luminescent dosimeter crystal used to track the radiation exposure of nuclear industry workers to the Super-Kamiokande neutrino detector, a 50,000-cubic-meter chamber holding 50,000 tons of ultra-pure water. Survey meters are used in industries where exposure to radiation is possible or probable, ensuring that workers’ exposures are within legal limits. But these instruments can be quite expensive (a high-quality survey meter costs more than $1,000), and they require semi-annual calibrations that make them impractical for use by the general public.

Radiation detection for a cell phone era

The cell phone-based radiation-detection system we have developed applies general gamma-ray astronomy techniques to the data set obtained from a vast swath of mobile phones. Just as a computational astronomer uses statistical analysis to distinguish a star in a star-field photo containing noise in each background pixel, our system similarly places detection pixels in the image plane (that is, mobile phones on a city map). Using image-processing techniques employed by astronomers and computing engineers, the system makes illicit radiation appear as bright “stars” on a map; fallout hot spots might appear as “pools” or more diffuse regions of higher-than-normal radioactivity.

Such a system seems particularly well-suited to help manage two types of disaster: nuclear accidents and radiological terrorism. In both cases, a region’s vulnerability is based on the density of its population: the larger the population, the greater the risks—and the greater number of cell phones. The more cell phones, the more sensitive a region’s mobile phone detection network will be to the presence of radiation, helping government officials and the general public respond on the basis of facts rather than fear.

The catastrophe at the Fukushima Daiichi Nuclear Power Station clearly illustrates the potential value of radiation detection by mobile phone network. Had the Fukushima region been populated with mobile telephones equipped with real-time gamma-reporting capabilities, information about the fallout cloud that escaped from the stricken nuclear reactors there could have been displayed, as it settled, in real-time two-dimensional video. The first hint of its presence would have shown up as a significant deviation from previously measured background radiation levels. One of the important reminders of the Fukushima accident is that the fallout drifts are carried by winds that are essentially a turbulent fluid. The random settling of fallout from this turbulence creates many hot spots, which a mobile phone-based system could help pinpoint.

Mobile phone gamma-detection networks would also be of obvious use in protecting against radiological terrorism. Such a system could easily detect an individual carrying a source that differs significantly from background radiation levels (let’s say by five standard deviations) as he walked from his porch to the car, drove down Main Street, entered a parking garage, and continued to register as 5-sigma all afternoon. In fact, in 2008, we were contacted discreetly by a contractor and host government when a high-activity source that could have been used in a radiological dispersal device was stolen. The officials of the country were embarrassed and wanted the perpetrators apprehended quickly. We told the officials that, given the location from which the source was stolen, had our system been in place, we may have been able to detect not only where the perpetrators went, but also what escape route they took.

One of the most significant problems after a radiological event is determining actual exposure to individuals, so those who need it can get medical treatment. Presently, this is done with various bio-assay techniques, which are time- and resource-intensive and in large events nearly impossible to perform in an effective way on all potential victims. It would be quite beneficial to have a detector present when the event occurred. The Biomedical Advanced Research and Development Agency, which is tasked with determining how to measure individual radiation doses in a post-incident environment, agrees: The agency has expressed interest in adopting our system.

The initial cost of implementing a cell phone-based radiation-detection system would be quite reasonable, in terms of governmental expenditures on defense and homeland security, and it would likely go down significantly over time due to economies of scale and improvements in technology. The first generation of the system we envision would place a scintillator (i.e., material that luminesces when excited by radiation) and a photo-diode detector, along with signal-conditioning and -counting chips, on the circuit board of a cellular telephone. In low volumes, the total cost of this additional hardware would be on the order of $100 per unit, but if the system were expanded to the overall cell phone market, costs would likely fall dramatically.

Given the societal benefit such a system would provide, tax credits or rebates would certainly seem to be in order. If economies of scale reduced the cost per unit to $50, then 100 million cell phones with radiation-detection equipment could be deployed free of charge to American citizens for an aggregate federal subsidy of $5 billion. Although this represents a significant investment, it could be spread over time and easily accommodated within a multitrillion-dollar federal budget that includes more than $600 billion in defense spending and some $55 billion in homeland security expenditures.

Technical challenges to the concept

The accident at the Fukushima Daiichi Nuclear Power Station shows that mitigation of radiological events is necessary; technology makes it feasible. But a cell phone-based radiation-detection system is obviously dependent on the existence of a working mobile phone network. As a result of the massive 2011 earthquake and tsunami that decimated significant portions of the northern third of Japan, power outages and cell phone service disruptions were widespread, and there is the possibility that our proposed system would itself have been inoperable. Temporary backup power and cell networks—increasingly used in disasters, such as Hurricane Katrina and the earthquake in Haiti—could have alleviated these problems.

Some critics have suggested that a mobile phone-based gamma-detection system would produce a huge number of false positives, warning of contamination where none exists and possibly causing unnecessary panic. This suggestion is inaccurate, because it misrepresents how the system would work. The cell phone detectors in the system we propose do not act as single-point, zero-dimensional detectors, but instead as data contributors to an entire, statistically smoothed surface. Detection is not performed by any single device but by software that operates on a multitude of readings from a map-region and constantly compared with a radiological baseline of that region. There are no vigilantes or panic buttons, only statistical inference from many measurements.

Because a detection of unusual radioactivity would be measured against a baseline, the granite steps at the train station, for example, would never cause a false positive. But what about the cancer patient returning from an iodine treatment at the nuclear medicine clinic? In our system, a nurse would simply ask the patient, “May I see your phone for a minute?” A tag—similar to the IFF (identification, friend, or foe) transponders used by aircraft—would then be added to the phone, identifying the newly radioactive patient as a “friendly,” or legitimate, source. By identifying legal nuclear sources through such tagging, harmless bright spots can be edited out of the database. Or, to put it in the modern vernacular: When dealing with false positives, there’s an app for that.

Another criticism of a mobile phone-based detection system is the possible level of noise in the data. More specifically, the concern is that with such a massive amount of data to track, evaluate, and store, there is the possibility that noisy or useless data will make characterizing new signals difficult. Of course, until a large number of measurements have been made at each square meter in a city, the data in the system will be noisier and, therefore, have less power. Given the short time scales over which suitable statistics are generated as more mobile users carry gamma detectors, the noisy period would be fairly short-lived. Once 30 measurements of a square meter have been made, the signal from that location is no longer of low value. It becomes a history against which subsequent measurements can be compared. Hot spots from fallout, contaminated water flowing from a sewer, or an illicit radiological source would not only show up as different from established background radiation levels; they would also have geometric differences from one another. And being able to track the movement of a radiological source gives society what it presently lacks in the face of a disaster—a fighting chance to see it coming.

Critics will correctly argue that some fallout hot spots (or terrorist weapons) might require that a detector be within centimeters of the effective source to obtain a reading. It turns out, however, that such proximity is a regular occurrence in urban populations. We have modeled the locations of random sources and human beings in city environments and found that the mean free distance between a randomly placed source and a cluster of people, in a high-density population, is sufficiently small such that much can be accomplished in an otherwise difficult detection scenario requiring detectors to be close to sources. Threat probability and detection probability go hand in hand if everyone is carrying gamma detection in their phone.

To date, the strongest opposition to a mobile phone-based gamma-detection system typically comes from civil libertarians. Would such a system not mean that every phone is tracked by the government, they complain, and that the government would know where an individual is at all times? But why should a radiation-detection system receive the brunt of the blame for a circumstance that already exists? Mobile phone companies track the location of cell phones now when location services are turned on (and probably even when they aren’t), and the geolocation-reporting capabilities of cell phones are unrelated to nuclear detection. Our proposal is to simply do something useful with the already-existing location data—something that could save lives and even entire cities.

A networked future of radiation safety

As new networks of mobile phone radiation detectors become widespread, the hardware cost will come down and the sophistication of the technology will increase. If those trends are shaped by the proper data-protocol standards, old phones and newer ones can work together to create increasingly more useful radiation-mapping layers. Eventually, the movement of illicit nuclear material will be made difficult, not through the actions of a central authority, but through tens of millions of interactions in a network. With networked physical detection, networked software-detection analysis, and multiple networked responders, response time and quality will improve. Such networks can reduce the risk of radiological terrorism by making the highest-value and most vulnerable targets—that is, the densest cities—the hardest to penetrate. As a result, in the wake of a radiological event, the ability to locate hot spots, quickly determine appropriate medical treatment, and direct safe population movements would be greatly improved.

Mobile phone nuclear-detection systems would serve the public interest in many ways, before and after accidental and intentional radiological events, and should become a normal part of civil defense. Such systems are analogous to good insurance—you don’t need it until something happens, but then you are very happy to have purchased full coverage.