Cygnus—A New Radiographic Diagnostic for Subcritical Experiments

J.R. Smith, R.L. Carlson (P-22), R.D. Fulton (P-23), J.R. Chavez, P.A. Ortega, R.G. O'Rear, R.J. Quicksilver (DX-3), B. Anderson, D.J. Henderson, C.V. Mitton, R. Owens (BN), S. Cordova, J.E. Maenchen, I. Molina, D. Nelson, E. Ormond (SNL)
Excerpted from LA-14202-PR

Introduction

The subcritical experiment (SCE) program was initiated after the 1992 moratorium on underground nuclear testing in support of stockpile stewardship.¹ The dynamic material properties of plutonium are a major topic of exploration for the SCE program. In order to provide for a multilayered containment of plutonium, the SCEs are executed in the U1a underground tunnel complex at the Nevada Test Site (NTS). We developed Cygnus, a new radiographic x-ray source, for diagnostic support of the SCE program at NTS. We took the name Cygnus from the binary star Cygnus X-1, located in the constellation Cygnus, which is a strong x-ray source. For the special conditions related to this application, we emphasize design in areas affecting machine placement, remote operation, and reliability.

Typically, SCEs have been limited to surface diagnostics such as interferometry and pyrometry. The interferometry system commonly used on SCEs is called VISAR (velocity interferometer system for any reflector). We developed x-ray radiography to complement the existing surface diagnostics, provide a more extensive spatial view (albeit temporally limited), and provide internal (penetrating) measurements. The Stallion series of SCEs consists of four shots: Vito, Rocco, Mario, and Armando.² Armando was the initial experiment for Cygnus radiography. The Rocco, Mario, and Armando tests use identical physics packages, permitting the use of Armando radiographic results as a confirmation of VISAR measurements. The main x-ray source requirements for an SCE involve spot size, intensity, penetration, and duration. The Cygnus source meets the following design specifications: ~ 1 mm diameter, 4 Rads dose at a distance of 1 m, ~ 2.25 MeV endpoint energy, and < 100 ns pulse length. Two Cygnus sources (Cygnus 1, Cygnus 2) were fielded on Armando providing two radiographic views separated in space by 60° and in time by 2 μs.

Cygnus Design and Layout

A multiorganizational team (LANL, Bechtel Nevada, Sandia National Laboratories, Naval Research Laboratory, and Titan/Pulse Sciences Division [Titan/PSD]) was formed to design a prototype x-ray machine in response to the programmatic need for fielding a radiographic diagnostic on SCEs. The design concept is based on three criteria:

1. it must fulfill the radiographic source requirements as previously stated;
2. it must accommodate the U1a tunnel environment in terms of footprint, operations, and safety; and
3. it must be reliable.

To satisfy the first criterion, the key is use of the rod-pinch diode that is a low-impedance, high-dose diode.³ This approach yields a smaller source size in comparison with other diode types. We met the footprint component of the second criterion by using a modular pulsed-power design where the energy storage unit (Marx) is independently located with respect to the inductive voltage adder (IVA), which are both connected by a coaxial water-transmission line. We achieved the reliability criterion by using proven pulsed power elements, extraordinarily conservative mechanical and pulsed-power designs, and an extensive test program.

The overall Cygnus layout in Figure 1 shows the two sources as fielded at NTS/U1a. Proceeding downstream, the major elements are: Marx generator, pulse-forming line, water-transmission line, IVA, and rod-pinch diode. Not shown is the high-voltage trigger generator used to initiate the Marx pulse. Titan/PSD integrated these pulsed-power elements into a design for the Cygnus system.⁴

Cygnus Test Plan and Deployment

Cygnus Prototype Tests at LANL. The “demonstration test” for Cygnus was performed at LANL. This test involves assembly of a prototype machine and testing via a series of shots designed to prove that the Cygnus design would meet or exceed x-ray source requirements (diameter, dose, endpoint energy, and pulse length), as well as operational requirements (reliability, reproducibility). A total of 118 protype test shots were completed.

Shot analysis shows design specifications for the source were met or exceeded.⁵ Successfully demonstrating that the prototype machine meets all design criteria and provides an acceptable level of reliability, we proceeded with the project along the following paths:

1. Promotion of the existing machine from prototype to field status (i.e., a source suitable for use on a SCE shot) and designation as Cygnus 1. Relocation of Cygnus 1 to a LANL firing site (R-306) for extended testing.
2. Approval of fabrication and testing of a second source (Cygnus 2) at Titan/PSD.

Cygnus 1 Field Tests at LANL. Cygnus 1 was moved to the LANL firing site for the extended testing phase. At this site, a structure was built to replicate underground installation at U1a for the express purpose of certifying the instrument in a realistic environment. Matching R-306 and U1a features include: wall and ceiling dimensions; bulkhead wall construction; electrical-utility placement; and installation of a camera house, camera monitoring system, and a remote control system. We completed 202 shots at R-306.

The R-306 tests show that Cygnus 1 was suitable for installation and operation in the U1a environment, did not produce levels of radiation adverse for operation of other planned diagnostics or systems, and produced images with the desired spatial quality and intensity. Therefore, Cygnus 1 was approved for underground installation at U1a.

Cygnus 2 Construction at Titan/PSD. Titan/PSD was contracted to deliver a second machine, Cygnus 2, which is similar to Cygnus 1 with the exception of a longer coaxial transmission line. Assembly and testing of Cygnus 2 paralleled with testing of Cygnus 1 at R-306. Testing at PSD consisted of 410 shots. Cygnus 2’s performance during the acceptance tests was similar to that of Cygnus 1, and the machine was ultimately approved for installation at U1a.⁶

Cygnus 1/Cygnus 2 Deployment at NTS/U1a. Figure 1 shows Cygnus 1 and 2 in the U1a.05 drift. Note that placing the Marx tanks end-to-end and the IVAs side-to-side accommodates the tunnel dimensions. The rod-pinch diodes are oriented with a 60° included angle. A bulkhead containment wall separates the zero room, which houses the experiment, from the Cygnus machine area. The test object is contained in a 3 ft diam vessel, thereby permitting reuse of the zero room. The camera system is protected in a camera room that has 1 in. thick steel walls. Tungsten collimators are placed in the bulkhead wall and camera room wall to reduce image “noise” from scattered x-rays. Both Cygnus machines may be monitored and controlled either from a screen room located in the .05A drift alcove, or from a diagnostic trailer on the surface. A total of 237 shots, the sum of shots from both machines, were fired at U1a.

Table 1 gives a breakdown of the number of shots completed for each facility as well as for each source. The total shot number (967) is an indication of the considerable effort given in this endeavor.

Cygnus Reliability is a Key Operational Element

The risk inherent in the execution of a SCE is that a high-stakes package is expended in a single event where there is no reprieve from equipment failure. In this respect, an SCE is similar to a NASA rocket launch. On command, Cygnus must reliably deliver x-rays at a specified time. Because Cygnus radiography is the primary diagnostic for Armando, and therefore a key metric for its success, pressure to guarantee high reliability was intense. We were especially concerned with two catastrophic failure modes for Cygnus that would result in losing the radiographic image. Just as in the rocket-launch scenario, a critical time in the SCE countdown specifies a “point of no return” where the event cannot be stopped. Both Cygnus failure modes correspond to malfunctions that occur after this critical time. The first failure mode is due to either a trigger generator or Marx generator prefire, resulting in premature radiation. The second failure mode is a no-fire situation, where the Marx fails to breakdown on command due to either a trigger failure or Marx failure, resulting in no radiation.

The three provisions below were implemented to enhance the probability for success.

Readiness. Historically, the Cygnus trigger generator and Marx generator components have exhibited a tendency for failure. In order to check these components, as well as the control and data acquisition systems, we instituted a preliminary shot procedure. This involves firing a dual-resistive load shot approximately 15 to 30 minutes before the planned radiography shot. This procedure is very effective in discovering preshot personnel error or equipment failure and in assuring system readiness.

Prefire Protection. Although we thought that trigger-generator problems previously discovered during prototype tests were fixed, initially there were frequent trigger-generator prefires on both Cygnus machines at U1a. As a result, we sent the trigger generators to LANL for extensive evaluation, testing, and modification to remedy the problem. We implemented an additional level of precaution to prevent prefire: we turned on the trigger generator high-voltage power supply just before a shot (~ 10 s). This method minimized the exposure to prefire and worked well even with trigger generators prone to prefire. Dirty spark gaps, as was discovered early in the program, encourage Marx generator prefire. Spark-gap purging is included as a command in the control software. It is done liberally (i.e., with significant gas flow) and often. Several Marx generator prefires occurred on Cygnus 1 just before the confirmatory shot. After the confirmatory shot, an inspection revealed the cause as several broken trigger resistors. After replacing the resistors, and during the period between the confirmatory shot (#100) and the Armando shot (#144), neither Cygnus machine experienced trigger-generator or Marx-generator prefires.

No-Fire Protection. Marx no-fires are typically caused by spark gap overpressure. Too much pressure leads to a no-fire, while too little pressure yields prefires. The Cygnus Marx pressures were set to favor the prefire case because there are no timely warnings for a no-fire, while the prefire has a good probability of occurrence before the critical time where the shot can be “saved.” We observed zero no-fire events on either machine.

Cygnus Reproducibility is a Key Operational Element

Shot-to-shot reproducibility of x-ray parameters is an important demonstration that source quality is consistent and that good radiography performance will likely be delivered on the Armando shot. It is also important because reproducible shots are required to baseline other diagnostic systems (e.g., to produce a repeatable electromagnetic interference background) and adjust the camera imaging system (e.g., alignment, intensity).

Dose results for U1a shots are shown in Figure 2. The gaps in the data primarily represent nonradiation shots, either the Marx resistive load or large area diode test modes. The low-dose shots, most predominantly seen on Cygnus 2, were caused by alignment or vacuum problems in the diode. These problems had been fully addressed by the day of the confirmation shot.

The average dose for all rod-pinch shots in the time period from the confirmation shot to the Armando shot is: 4.4 ± 0.5 Rad (Cygnus 1) and 4.1 ± 0.4 Rad (Cygnus 2). These measurements are standardized for a source-detector distance of 1 m, and for attenuation through 1 in. thick aluminum. Endpoint voltage is an indication of source penetration. The endpoint voltage for the same database as cited above is: 2.1 ± 0.1 MV (Cygnus 1) and 2.3 ± 0.1 MV (Cygnus 2). Reproducibility in Cygnus source intensity and penetration, shown by the standard deviation in dose and endpoint voltage above, easily satisfies the desired shot-to-shot performance.

Conclusion—Cygnus Succeeds on the Armando SCE

The Armando shot was executed on May 25, 2004. The efforts of more than three years of planning and testing are encompassed in Cygnus performance on this single shot. It is, therefore, appropriate to conclude by presenting Armando results. Diode voltage, diode current, and x-ray waveforms (Figure 3) are key indicators of Cygnus performance. By visual comparison of the Cygnus 1 and 2 waveforms, it is evident that both machines produced markedly similar results. Several quantitative measurements of Cygnus performance on the Armando shot are given in Table 2. The measurements are placed in three categories representing different facets of machine output: Cygnus diode, electrons, and x-rays. On the Armando shot, Cygnus 1 (Cygnus 2) satisfied source requirements as follows: ~ 1mm x-ray source diameter, 4.5 (4.2) Rad dose, 2.1 (2.2) MeV endpoint energy, and 47 (48) ns pulse length. Additionally, the desired timing with a pulse separation of 2 μs was delivered. The success of Cygnus at U1a is a historical milestone that demonstrates the feasibility of fielding large and complex diagnostic systems downhole. Its success paves the way for Cygnus participation in future SCEs as well as promotes consideration of even larger radiographic facilities at U1a.

References

1. L.R. Veeser et al., “Subcritical plutonium experiments at the Nevada Test Site,” 1997–1998 Physics Division Progress Report, Los Alamos National Laboratory report LA-13606-PR (May 1999) pp. 94–101.
2. D. Fulton, M.D. Wilke, and N.S.P. King, “Dynamic material studies in subcritical experiments: Rocco, Mario, Vito, and Armando,” Physics Division Activity Report, Los Alamos National Laboratory report LA-14112-PR (February 2003) pp. 23–26.
3. G. Cooperstein et al., “Theoretical modeling and experimental characterization of a rod-pinch diode,” Physics of Plasmas 8, 4618–4636 (2001).
4. D. Weidenheimer et al., “Design of a driver for the Cygnus x-ray source,” Proceedings of the 13th IEEE International Pulsed Power Conference, R. Reinovsky, M. Newton, Eds. (IEEE, Piscataway, New Jersey, 2001), Vol. (1) pp. 591–595.
5. J.R. Smith et al., “Performance of the Cygnus x-ray source,” Proceedings of the 14th International Conference on High Power Particle Beams, TA. Melhorn, M.A Sweeney, Eds. (AIP, Melville, New York, 2002), Vol. (1), pp. 135–138.
6. V. Carboni et al., “Pulse power performance of the Cygnus 1 and 2 radiographic sources,” Proceedings of the 14th IEEE International Pulsed Power Conference, M. Giesselmann, A. Neuber, Eds. (IEEE, Piscataway, New Jersey, 2003), Vol. (2) pp. 905–908.

Acknowledgment

We gratefully acknowledge support for this work from many colleagues in Physics and Dynamic Experimentation Divisions at LANL, and from colleagues at Bechtel Nevada, Sandia National Laboratories, Titan/PSD, and Naval Research Laboratory. Special recognition is given for Nevada Operations and Bechtel Nevada support at NTS which was instrumental in fielding Cygnus at U1a. This work was sponsored by the U.S. DOE.

For further information, contact John R. Smith, 505-665-8546, smith@lanl.gov.