Introduction

The Radionuclide Aerosol Sampler/Analyzer (RASA) is an automated aerosol collection and analysis system designed by Pacific Northwest National Laboratory (PNNL) in the 1990s [1–3], and is deployed in several locations around the world in support of the International Monitoring System (IMS) specified in the Comprehensive Nuclear-Test-Ban Treaty (CTBT). The RASA operates unattended, iterating samples through a three-step process on a 24-h interval. The RASA is novel in that it functions in a continuous sample feed, allowing for continuous, automatic collection, packaging, and radiation measurement of discrete 24-h samples. The conceptual design of the RASA is shown in Fig. 1.

Fig. 1
figure 1

Schematic of the original RASA sample path. Inside the sample head (left), the six filter strips collect particulate from a ~1,000 m3/h air flow. The filters are sealed together in encapsulating mylar strips upon exiting the sample head and are held in an interim decay position for 24 h before advancing into the detector enclosure. IMS stations have an additional roller turn to lengthen the path between sampling head and detector to prevent shine into the detector cavity during the decay step

Full size image

General Dynamics (GD) acts as the equipment provider and oversees the operations and maintenance (O&M) responsibilities for the IMS radionuclide monitoring systems owned and/or maintained by the United States—specifically, 11 aerosol systems and 4 xenon systems. As such, they are in a unique position to gather and evaluate performance data and assess the operational impact of both equipment and O&M issues as well as anomalous events. In their equipment provider role, GD is responsible for identifying, distributing, and implementing technology upgrades to the RASA.

On 11 March 2011, a 9.0 magnitude earthquake sent a massive tsunami toward the eastern coast of Japan, resulting in power loss and cooling failures at the Daiichi nuclear power plants in Fukushima prefecture. A majority of IMS RASA stations successfully collected and measured aerosols, including all eleven of the U.S. RASAs. A number of lessons and courses of research were learned from RASA performance, station operators, and the responsible nations and the CTBT organization (CTBTO)—all ultimately pointing to potential improvements and upgrades to the RASA. Recent and ongoing upgrades and considerations include a new System Instrumentation Unit (SIU), research into current and future cooling alternatives, and sample processing and handling options.

While some of the upgrades and much of the research discussed in this paper are specific to the eleven RASAs operated by the United States, it is anticipated that these will also be considered for use in the other IMS RASAs.

Overview of upgrades

During its 15+ year operational life, the RASA has received a modest number of updates, most in the vein of minor engineering improvements, such as a new blower and blower motor, changes to the filter media and encapsulating tape, shielding modifications, and various changes to electrical components and frame/structure. Many of these changes have required associated software improvements as well. Recently, however, the obsolescence of many electrical components and failure or discontinuation of commercial subsystems has brought to focus the need to evaluate the maintainability and sustainability of the individual subsystems and the RASA as an overall system.

System instrumentation unit

The SIU is the central control hardware of the RASA, directing subsystem operation and signal and power distribution. A complete re-design of the SIU was accomplished between 2011 and 2012. As stated previously, the age of internal components and resultant difficulty in finding and procuring direct replacements has grown into a major obstacle to the continued maintenance and repair of the original SIU. With regular maintenance visits performed annually and the cost of unscheduled maintenance or repair extremely high for many of the remote and hard-to-access RASA locations (e.g., Wake, Midway, Easter Islands in the Pacific, Antarctica), modernization became the provident choice for continued sustainment. Requirements for a new SIU focused on ease of maintenance and repair, long-term availability of components and avenues for improvement taken from GD’s extensive experience. The most significant and impactful design change is in the move from direct point-to-point wiring to printed circuit boards. Experience and best practices also led to the segregation of power and signal lines into separate circuit boards. The resulting improvement in internal organization is evident in the side-by-side comparison of the old and new SIUs shown in Fig. 2. Painfully illustrative of the need for simplicity (as well as improved manufacturing control) is the single wire color throughout the original commercial unit. The new circuit board format eliminates the internal tangle of wires, provides for quick visual confirmation of signal status via board-mount LEDs, a series of troubleshooting switches, and high-density connectors. Originally envisioned as movable jumpers, the small three-position switches—similar in appearance to a potentiometer—provide the same debugging/troubleshooting function, but eliminate the potential loss of jumpers. Continuing the simplicity theme, the switches all follow one configuration: as-wired, active, and non-active as shown in Fig. 3. The new design also features new I/O hardware, namely Opto 22 SNAP PAC control, input, and output modules.

Fig. 2
figure 2

The original SIU on the left is full of point-to-point wiring. The newly re-designed SIU on the right demonstrates the marked decrease in wire clutter, most notably on the upper “signal” side. Also visible is an integral troubleshooting area (a). Each signal has an LED indicating active signal status and a three-position switch for use in debugging and troubleshooting. The new SIU has a centralized I/O assembly (b), compared to the original SIU which passed signals to an external secondary I/O box (about the size of a standard 2U rack box)

Full size image
Fig. 3
figure 3

Schematic of SIU three-position signal switches

Full size image

Detector cooling

While many radionuclide IMS stations use liquid nitrogen to cool their high purity germanium (HPGe) detectors, the RASA has employed a mechanical cooler since its inception to maintain the low temperature needed for HPGe. This is primarily because of the unavailability or prohibitive cost of liquid nitrogen at the remote locations for which the RASA was designed. Failures related to the mechanical coolers comprise a significant portion of missing IMS data availability for aerosol systems.

General Dynamics, in concert with the Provisional Technical Secretariat (PTS) of the CTBTO, has accumulated >100 instrument years of operating experience with the X-Cooler models produced by ORTEC. A significant number of spontaneous temperature excursions were observed among the IMS RASA stations. The state-of-health data from the radionuclide stations were examined with particular emphasis on the detector crystal temperature.

One especially troublesome observation is the spontaneous temperature excursion, i.e., warming, of the detector. A significant excursion is defined as a temperature increase of greater than 30 °C above the normal baseline temperature (typically −185 °C) that is not due to loss of cooler power. Of the several spontaneous temperature excursions recorded, none were related to power outage, vacuum problems, or other apparent problems and the effect to mission capability was no more than a few days.

Both GD and PTS have aggressively pursued alternative cooling technologies to solve the frequent cooler failure and spontaneous detector warm-up problems. Recent work has been done to customize COTS electrical cooling options to the RASA [4, 5]. Other potential solutions for cooling via modified COTS cooling has been shown for manual IMS stations [6, 7] that have shown promising extension to the RASA. Particular near-term research for alternative RASA cooling is focused on a compact liquid nitrogen generator. PTS has tested the ELAN2 manufactured by MMR in California at two RN stations, RN47 at Kaitaia, New Zealand and RN26 at Nadi, Fiji and a third unit at PTS headquarters. PNNL modification of the ELAN2 specific to mate with the RASA is in the early stages.

While new COTS coolers and liquid nitrogen generators look promising, an obvious path toward long-term sustainability is the consideration and evaluation of detectors that do not require external cooling, recognizing that without novel geometries, arrays or new detector advancements, decreased resolution is the likely compromise.

Integrity of detector measurements

The RASA may need additional engineering modifications to shield the detector from ambient radioactive material. Shortly after the Fukushima event, released particulate was observed at a station before the sample on which that particulate should have been collected had advanced into the detector chamber [8]. As an international group of experts watched IMS particulate data in the days after the initial ejection of radioactive material from the affected Fukushima reactors, they began to see some unexpected activity in preliminary spectra at the RASA in Takasaki, Japan (JPP38) about 2 days after the initial release (recorded on ~15 March 2011). When the sample upon which some of that first material was collected was counted, a discrepancy was noticed in a sample measured previously; namely, that elevated activity displaying similar composition to that expected from Fukushima was observed during the count of a sample collected before the actual particulate was collected on a RASA sample. After reexamining the RASA data it was concluded that the observed activity was the result of air infiltration inside the detector housing.

Recommendations

The technology in the RASA, save enhancements made by General Dynamics and the recent PNNL upgrade to control hardware, is mid-1990s era top of the line, and as such, is not only outdated, but no longer representative of the best possible collection and detection capabilities. Notwithstanding, the RASA performs consistently and effectively to detect low-level atmospheric nuclear explosions. It is the potential for detecting underground, and potentially evasive, nuclear explosions that requires careful investigation and development of next-generation technology. The detection limit of the RASA could be improved by more efficient sampling methods, e.g., electrostatic collection, optimized filter media (resulting in improved collection efficiency, smaller filter size, etc.), and by enhanced detector construction [9]. These improvements, specifically the addition of a second low-background detector, are expected to improve RASA sensitivity by a factor of 10–70. Other high sensitivity-to-cost ratio improvements include active anti-cosmic veto and anti-Compton detectors.

The three 24 h steps equate to a 72-h wait time for full processing of a sample; the earliest possible nuclear data is a preliminary count sent to the International Data Center (IDC) in Vienna, Austria 2 h into the count, about 50 h after the start of sampling. This inherent delay is for good reason; namely, to allow for the decay of short-lived isotopes. Nevertheless, two simple suggestions arose from discussions around the perceived need for faster data: (1) modification of the sampling protocol to allow for an alternative sampling state, e.g., an emergency mode, in which a shortened sampling and/or radioactive count could be initiated, either by a push from the IDC or based on an elevated local radiation measurement, and (2) the addition of a simple, low-resolution detector, e.g., a commercial NaI/CsI quick-read detector, mounted at the immediate rear of the RASA sampling head. A conceptual drawing of an installed quick-read detector is shown in Fig. 4. The benefits of a local quick-read detector include realtime measurement of the sample radioactivity, potential feedback into an altered sampling scheme, and safety information for station operators. A potential third application of an alternative sampling scheme was observed at the Takasaki IMS station. The JPP38 station operators, upon the first observation of high-activity samples, elected to decrease the flow rate of their RASA to approx 300 m3/h in the days following (approx 30 % of normal operation) [10]. The ability to shorten sampling times (or reduce flow) could be readily accomplished by temporarily modifying sampling parameters in the control software based on realtime count rates.

Fig. 4
figure 4

Concept of low-resolution quick-look detector mounted to rear of RASA sampling head. The view is from the rear of the RASA and includes a the sampling head, b the blower, c the air outlet, and d the rear of the detector enclosure. A CsI/NaI detector e is proposed to provide real time measurement of activity inside the sample head. The activity measurements may be used to enhance health monitoring, pinpoint arrival times for atmospheric transport modeling, or potential alternative sampling modes. The detector is called out for clarity

Full size image

Physical contamination prevention for internal parts specific to sampling and counting activities should be implemented. The RASA is well-shielded against external gamma rays, but the shielding is not airtight and there is a large opening at the top of the lead through which the filter assembly enters and exits the detector enclosure. A cross-section of an electronic drawing of the detector assembly is shown in Fig. 5. The opening is evident at the top. Appropriate measures considered include dust cover for filter rolls, dust mitigation changes by sealing or covering the detector and/or detector enclosure, and eliminating the large empty air volume (>4 l) within the lead shield.

Fig. 5
figure 5

Cross-section of the RASA detector and shield assembly. Circular parts A and B are respectively thin copper (~1.5 mm) for shielding and an aluminum roller (1.6 mm thick) upon which the filter tracks around the detector. The opening at top and empty volume within the lead enclosure are clearly seen from this view

Full size image

In the face of shorter collection times and the occasional desire for faster data, the opportunity exists to enhance measurement capabilities while mitigating the health and delay concerns. Consider the case in which a plume of high activity material passes a station. As a quick-read detector records an uptick of activity in the sampling head, the control software takes note of the increase and resets the sampling time parameter to a shorter period. This limits the total activity of the sample and fast-forwards the sample to subsequent processing steps. Assume the sampling time was changed to only 6 h. In this circumstance, it is a certainty that interested parties will hope for a measurement as soon as possible. Normal operation dictates a 24-h decay between the end of sampling and start of counting. In this case, where data is wanted soon after collection, an alternative detector capable of measuring the sample while short-lived radon progeny exists among the isotopes of interest would be particularly useful. An ideal detector would increase sensitivity to temporally-subsequent events, eliminate the need for the current 24-h decay period, and require little to no external cooling. Appropriate candidate materials and configurations are being considered.

Additional lessons relevant to RASA operation and useful future upgrades taken from collecting and processing samples following the Fukushima reactor event are numerous. One perhaps most addressable lesson is the need for capacity to obtain faster results. While manual collection/sampling procedures are easily modified, such intervention requires on-site and available personnel to exchange filters. In the days of the excitement and confusion after the first radioactive releases from Fukushima, there was almost a desperation for more information. Additionally, activity measurements at some stations were higher than typically considered for nuclear explosion monitoring and at times exceeded the upper dynamic range of the systems. Radiation protection measures should be considered at radionuclide station as standard operational procedures.

Conclusions

The recent success of a new SIU and porting of the original control software from QNX to a modern Linux architecture demonstrate the utility of incremental upgrades and improvements to maintenance, repair, and sustainment. Lessons learned over the life of the RASA thus far, as well as those taken from the 2011 Fukushima nuclear power plant event, underscore the need for not only failure or breakage-related issues, but to look forward toward the potential for significant gains in sensitivity and improved operational oversight, such as health monitoring for local station operators.

As the operational lifetime of the RASA lengthens, the need for incremental upgrades, component and subsystem replacement, and obsolescence management will continue to be issues requiring attention. It is foreseeable that a point of sustainability plateau and long-term maintenance will be reached. At that point, the station operators and maintainers can expect another 10+ year period free of major component failures and obsolescence. A potential complication to this incremental upgrade path is the protracted, and most importantly, total upgrade cost versus that of engineering and deploying a new, more sustainable, and more efficient and sensitive collection and detection system. The PTS has recently established a technology foresight website for relevant parties, but a more in-depth technical review and development forum should be set up to facilitate discussion between station operators, maintenance and logistics contractors, PTS, national laboratories, etc., regarding priorities for upgrades and promising future technologies. Support for novel collection and nuclear measurement methods and engineering approaches is strongly encouraged.