A New Era of Nuclear Criticality Experiments: The First 10 Years of Radiation Test Object Operations at NCERC

1

Abstract

The work presented in this paper focuses on the first 10 years (2011–2020) of radiation test object (RTO) operations at the National Criticality Experiments Research Center. RTOs are subcritical configurations of special nuclear material that are built by hand. These configurations are utilized for benchmark experiments, detector testing/characterization, and training. An overview of the types of measurements used in RTO operations is given as well as a history of RTO operations at Los Alamos National Laboratory from 1944–2011.

Keywords:

• Criticality experiments
• gamma detectors
• National Criticality Experiments Research Center
• neutron multiplication
• radiation test object
• special nuclear material
• subcritical benchmark
• subcritical measurements

I. INTRODUCTION

Radiation test objects (RTOs), in the context of this work, are subcritical configurations of special nuclear material (SNM) that are built by hand. The configurations vary in SNM type, mass, form, and geometry, resulting in a wide range of subcritical neutron multiplication (from near 1 to about 20). These configurations often include moderator and/or reflector materials. RTOs have primarily been constructed for the purposes of training and radiation measurements and to provide information for the criticality safety community. RTOs are utilized for research and development (R&D) because their geometry, material characteristics, and radioactive emissions provide surrogates for real-world items of concern or interest.

This work will focus on RTO operations performed at the National Criticality Experiments Research Center¹ (NCERC), which resides within the Device Assembly Facility (DAF) at the Nevada National Security Site. Prior to the startup of NCERC, Los Alamos performed RTO operations at the Los Alamos Critical Experiments Facility (LACEF) in Technical Area 18 (TA-18) from 1946–2006. Note that many names were used throughout the years to refer to the location (TA-18, Pajarito Site, and Pajarito University to name a few), but this work will refer to all of these operations during this time period as “LACEF.” In 2005, the U.S. Department of Energy (DOE) finalized the decision to move LACEF capabilities to Nevada. Efforts to upgrade and relocate the critical experiments capability to Nevada began. For relocation of the TA-18 nuclear material inventory, the pace quickened to characterize the inventory and to make disposition plans. RTO measurements began in the DAF in 2007 under Los Alamos National Laboratory’s Nuclear Material Handling and Measurement (NMHM) operations. Table I shows the timeline of these capabilities from 1944 to the present.

TABLE I History of Los Alamos RTO Operations

Years	Location
1944–1945	Omega Canyon
1946–2004	LACEF
2004–2006	DAF NMHM startup
2007–2011	DAF
2011–present	NCERC

This paper will start by describing operations conducted from 1944–2011 in Sec. II. Section III describes the types of measurements typically performed during RTO operations. Section IV focuses on the first 10 years of RTO operations at NCERC (2011–2020). The operations described include benchmark experiments (Sec. IV.A), detector testing and characterization (Sec. IV.B), university consortia measurements (Sec. IV.C), neutron diagnosed subcritical experiments (NDSEs) (Sec. IV.D), and training classes (Sec. IV.E). Last, conclusions and future work are discussed in Sec. V.

II. RTO HISTORY AT THE LOS ALAMOS CRITICAL EXPERIMENTS FACILITY

Understanding the critical mass of nuclear material was of great importance during the Manhattan Project. In order to address this, integral experiments were performed at Los Alamos. Of primary interest were critical mass data on highly enriched uranium (HEU) and weapons-grade plutonium (WG Pu). These early experiments^2–4 were the first measurements ever performed on large (kilogram) quantities of HEU and plutonium metal and are similar to many of the RTO experiments discussed in this work. On August 21, 1945, shortly after the end of World War II, Harry Daghlian was conducting hands-on experiments with a 6.2-kg sphere of $\delta$-phase WG Pu reflected by tungsten carbide bricks.⁵ During the experiment, a brick slipped from his hand onto the assembly, resulting in a criticality accident that yielded a lethal dose of radiation to Daghlian. This resulted in moving these operations from Omega Canyon in Los Alamos to TA-18, a more remote area of the laboratory. A second criticality accident occurred on May 21 of the following year using the same plutonium core reflected by beryllium.⁵ These two accidents led to a requirement that any systems that have a reasonable chance of reaching criticality must be controlled remotely,⁶ ultimately leading to construction of several critical assembly machines and the establishment of LACEF (Ref. 7).

While many critical experiments occurred in the following years, hands-on measurements with subcritical systems were still performed at LACEF. These hands-on measurements are all labeled as RTO measurements for the purposes of this paper, although the term was not yet being used at the time. The RTO and critical assembly operations complement one another. Both utilize the same SNM, personnel, detection systems, and measurement methods. For this reason, some of the experiments presented in this work may look similar to, and share SNM with, experiments on the Planet,⁸ Comet,⁹ Flattop,¹⁰ and Godiva IV (Ref. 11) critical assembly machines.

LACEF had a very large collection of unique SNM items, which included hundreds of kilograms of uranium and plutonium. All of this material was moved to NCERC and continues to be used to support the capability to create and measure RTOs. This work will focus on four particular SNM items made of one of the following elements: uranium, neptunium, and plutonium, which are among the “most popular” items due to the unique characteristics described below.

The Thor core consists of three pieces of $\delta$-phase WG Pu that have a combined mass of approximately 9.5 kg (Ref. 12). The Thor core was originally used in a Comet experiment in which it was reflected with thorium,⁹ hence the name. Combinations of Thor core pieces were measured in a well counter in the early 1980s to demonstrate what is now known as the Hansen-Dowdy formalism,¹³ a neutron noise method that is still in use. The Thor core is extremely useful for RTO measurements because one can use combinations of the three pieces (along with 0.5-in.-diameter plutonium cylinders that can be inserted in the center piece) to achieve a wide range of plutonium masses (which will result in a large range of system multiplication). This same material has been used for measurements at NCERC as described in Sec. IV.B.

The Beryllium Reflected Plutonium (BeRP) ball (shown in Fig. 1) was cast and clad in stainless steel in 1980. The BeRP ball consists of an $\alpha$-phase WG Pu sphere with a reported mass of 4483.884 g and density of 19.6039 g/cm${ }^3$ (Ref. 14). This SNM item was originally used in a Planet experiment in which it was reflected by beryllium,¹⁵ hence the name. The majority of RTO operations with the BeRP ball since that time do not involve beryllium reflection. The BeRP ball is a very useful piece of SNM because it is a large mass of spherical plutonium but is still below the water-reflected critical mass for $\alpha$-phase WG Pu (Ref. 16). The multiplication [total multiplication near 4.5 (Ref. 14)] and the neutron emission [near 1e6 n/s (Ref. 14)] of the BeRP ball are both fairly high, which is useful for detector testing and validation.

Fig. 1. (a) Photograph of the BeRP ball. (b) Schematic showing the BeRP ball cladding. (c) Setup of the nickel-reflected subcritical benchmark.⁵⁴

In addition to using the BeRP ball in several benchmark experiments at NCERC (described in Sec. IV.A), the BeRP ball was also used for many detector validation measurements at LACEF (Ref. 17). Objectives of these measurements included training, technology functionality, and algorithm validation and verification. Between 2007 and 2011, the BeRP ball was used for several measurements at the DAF under NMHM operations. These included experiments reflected with polyethylene (shown in Fig. 2) (Refs. 18, 19, and 20), Lucite,^18,21 and nickel.^18,22 These experiments had a variety of goals, but most focused on neutron noise measurements. Many simulations of the BeRP ball were also performed in this time frame,^23,24 including simulations of specific experiments conducted at the DAF (Ref. 25). Furthermore, these measurements and simulations were used with sensitivity methods, often to help inform nuclear data, with a focus on ${ }^{239}$Pu $\bar{\nu }$ (Refs. 26 through 30). Experiments with the BeRP ball at NCERC are described in Secs. IV.A, IV.B, and IV.C.

Fig. 2. Experiment with the BeRP ball reflected by polyethylene, performed in 2009 at DAF (Ref. 19)

In the mid-1990s, the Rocky Flats (RF) metal uranium shells were relocated from the former Rocky Flats Critical Experiments Facility to Los Alamos for use at LACEF. The RF shells are nesting hemishells of HEU metal, each approximately 1/8 in. thick.^31,32 There are 80 hemishells in total, resulting in well over a critical mass of bare HEU. The RF shells are useful for RTO operations due to the ability to create a large number of combinations of shells, allowing for varying mass and geometry. These hemishells have been used on Planet in experiments with the BeRP ball³³ and the ${ }^{237}$Np sphere.³⁴ In the 1990s and early 2000s, many RTOs were assembled with the RF hemishells. Various configurations were assembled and measured with ${ }^3$He detection systems. These included passive measurements with no source as well as active measurements with driving sources of ${ }^{252}$Cf, ($\alpha$,n), photoneutrons (using a Betatron), and 14-MeV neutrons (using a neutron generator).³⁵ These measurements supported active diagnostic technique development efforts. Many measurements included a Linitron 2000 (pulsed bremsstrahlung with 10-, 8-, and 6-MeV endpoint energies). Various neutron noise methods were analyzed using these RF measurements.^36–41 Simulations were then performed and compared to measured data.^42–44 Experiments with the RF hemishells at NCERC are described in Secs. IV.C and IV.D.

A sphere of ${ }^{237}$Np was cast at Los Alamos in 2001 (Ref. 45) in order to investigate the critical mass of ${ }^{237}$Np. The ${ }^{237}$Np sphere has a mass of 6070.4 g and is surrounded by tungsten and nickel to reduce the gamma dose rate. Neptunium-237 has many unique properties, which makes this item useful for RTO operations. The critical mass experiment with the ${ }^{237}$Np sphere was performed on the Planet critical assembly machine in which it was surrounded by RF hemishells.^34,46,47 To support this experiment, subcritical multiplication measurements were performed in advance at LACEF to provide approach-to-critical information.⁴⁸ Experiments with the ${ }^{237}$Np sphere at NCERC are described in Sec. IV.A.

III. TYPES OF MEASUREMENTS

Many applications benefit from the use of SNM for R&D purposes. Examples include nonproliferation, treaty verification, safeguards, emergency response, and criticality safety. The reason that RTOs are so useful is that they can be utilized to build configurations of relevance to those applications. Measurements performed with RTOs are generally focused on radiation or physical characteristics of a configuration. This includes gamma emission, neutron emission, or physical characteristics determined through radiography. Both passive and active measurements are performed on RTO configurations. Detection systems include both commercial off-the-shelf (COTS) systems and those developed by national laboratories and/or universities.

Gamma measurements can have multiple objectives. The most common objective is to investigate gamma spectra for radioactive nuclide identification using primary gamma rays and other features such as (n,$\gamma$) interactions. These spectra can also be used for detector characterization. Other output from radioisotope identification devices (RIIDs) and gamma imagers such as isotope identification and spatial information can be used to evaluate onboard identification and localization algorithms. Gamma detection systems of many types are used for RTO operations including organic scintillators (OSs), inorganic scintillators (ISs), and semiconductors. Last, gamma dose rates are often measured using standard health physics instrumentation. Gamma measurements are also performed for material control and accountability purposes (to help quantify the nuclear material). One example of a gamma detection system used for gamma spectrum analysis is a liquid nitrogen-cooled COTS system shown in Fig. 3.

Fig. 3. High-purity germanium detection system

Neutron measurements also have multiple objectives. Systems are used to infer the total neutron emission or dose rate from a configuration. Neutron noise analysis methods are also used, which take advantage of the fact that multiple neutrons can be created from a single fission event at the same time.^49,50 Neutron noise methods are performed to infer the system multiplication and/or effective multiplication factor $k_{eff}$. Neutron imaging is an objective for some measurements. Neutron detector systems generally include proportional counters or scintillators. One system that estimates the neutron emission rate is the Shielded Neutron Assay Probe (SNAP) detector shown in Fig. 4. One example of a ${ }^3$He system that is often used for neutron noise measurements is the MC-15 detector (sometimes referred to as the NoMAD detector),¹⁴ shown in Fig. 5. An example of an OS system used for neutron noise measurements is shown in Fig. 6.

Fig. 4. SNAP neutron detection system

Fig. 5. MC-15 multiplicity counter

Fig. 6. Organic scintillator system

Both passive and active neutron measurements can provide useful information on measured configurations. Active interrogation (AI) is often performed with a deuterium-tritium (D-T) or deuterium-deuterium neutron generator. In addition, photon sources may be used as well and provide neutrons through ($\gamma$,n) interactions. Active interrogation can provide additional information since the time of the introduction of the interrogating source particles can be utilized. This provides benefits for neutron multiplicity, imaging, and estimation of material isotopics. Active interrogation can be paired with any of the same detector types used in passive measurements. Examples of AI measurements at NCERC are shown in Fig. 7.

Fig. 7. Examples of AI using neutron generators

Radiography is an active measurement that typically uses X-ray generators. These devices usually use an electrical current to generate electrons and subsequently X-rays through bremsstrahlung. A detection system (which could be film or digital panels) is placed in line with the beam angle of the X-ray generator. Typically, an object (such as an RTO configuration) is located in between the X-ray generator and the detection medium. Through selection of energy and duration, varying amounts of X-ray penetration will occur through the object resulting in varying exposure to the detection system. For RTO operations, it is typical to perform both low-energy exposures (hundreds of kilo-electron-volts) and high-energy exposures (mega-electron-volts). An example of radiography at NCERC is shown in Fig. 8.

Fig. 8. (a) Setup for a radiograph of the BeRP ball using a Betatron. (b) Resulting radiograph of the BeRP ball

IV. RTO OPERATIONS AT NCERC

At NCERC, four components are required for successful RTO operations:

1. SNM: Hundreds of kilograms of plutonium and HEU are present at NCERC. Our team thinks of innovative ways to utilize existing material for new RTOs while also securing additional SNM from throughout the DOE complex.

2. Nonfissile inventory: From both historical and recent projects, NCERC has a large collection of nonfissile materials. These are largely reflector and/or moderator materials. This inventory includes various materials such as metals including depleted uranium (DU), beryllium, tungsten, nickel, copper, steel, and aluminum in addition to plastics such as high-density polyethylene (HDPE) and Lucite. While the geometries vary depending on application, we have a large collection of hemishells, as shown in Fig. 9, the majority of which nest with one another. This collection has continued to grow throughout the 2011–2020 time period focused on in this work.

3. Diagnostic systems: Our team has COTS and laboratory-developed neutron detectors (examples in Figs. 4, 5, and 6), gamma detectors (example in Fig. 3), radiography equipment (example in Fig. 8), sources, and neutron generators (example in Fig. 7). Other organizations also bring in their own equipment, which could range from early prototypes to COTS systems.

4. Expertise: Our team includes experts in neutron detection, gamma detection, radiography, and criticality safety. This helps ensure not only that the RTO configurations can be safely assembled but also that participating organizations obtain quality data.

Fig. 9. Examples of nonfissile hemishells: (a) DU, (b) copper and HDPE, and (c) nickel

Figure 10 shows the number of operational days in which RTOs were built and measured at NCERC for each year, starting in August 2011. It should be noted that 2011 was not a full year of operations and therefore had fewer builds than subsequent years. COVID-19 greatly impacted our program in 2020, which ultimately resulted in fewer operations than in previous years. As seen in Fig. 10, RTO operations can be classified as either measurements or classes. On average, 23% of operational days during these 10 years were in support of classes while 77% were for measurements. In total, over 900 RTO configurations have been assembled in the first 10 years at NCERC.

Fig. 10. Number of operational days by year

Fig. 11. The SCR$\alpha$P experiment. (a) Partial assembly. (b) Assembled configuration with NoMAD detectors

Fig. 12. The NeSO experiment. (a) Partial assembly. (b) Assembled configuration with NoMAD detectors

Fig. 13. Example of a COTS handheld gamma detector being used

Fig. 14. Photograph and hot spot identification of COTS gamma imaging systems: (a) BeRP ball and (b) DU hemishells

Fig. 15. Example of small plutonium samples that were placed in the Thor center piece for neutron multiplicity measurements

Fig. 16. Setup of the Thor core measurements

Fig. 17. University consortia measurements

Fig. 18. (a) A subset of RF HEU hemishells. (b) Individual RF hemishells can be replaced with aluminum hemishells. (c) Rocky Flats hemishells 3–30 have a mass of 21.8 kg with dimensions shown

RTO operations can be further broken down into five types, which will be discussed in the following sections: benchmark experiments (Sec. IV.A), detector characterization (Sec. IV.B), university consortia measurements (Sec. IV.C), NDSEs (Sec. IV.D), and classes (Sec. IV.E).

IV.A. Benchmark Experiments

Subcritical benchmark measurements (sponsored by the Nuclear Criticality Safety Program) provide experimental data for nuclear data and computational method validation. These benchmarks are part of the International Criticality Safety Benchmark Evaluation Project⁵¹ (ICSBEP) and include evaluation of all statistical and systematic uncertainties. In order to minimize uncertainties and make the configurations easier to model, simple geometries with few well-characterized materials are preferred. This is why the reflector materials shown in Fig. 9 are spherical and composed of pure materials with few impurities. The Los Alamos RTO benchmarks are shown in Table II.

TABLE II Los Alamos RTO Benchmarks

Identifier	Location	SNM	Reflectors	Number of Configurations	Year
SUB-PU-MET-FAST-001 (Ref. 52)	LACEF	BeRP	HDPE	5	2003
FUND-NCERC-PU-HE3-MULT-001 (Ref. 54)	NCERC	BeRP	Nickel	7	2014
FUND-NCERC-PU-HE3-MULT-002 (Ref. 57)	NCERC	BeRP	Tungsten	8	2016
FUND-NCERC-PU-HE3-MULT-003 (Ref. 14)	NCERC	BeRP	Copper and HDPE	17	2019
To be determined (Ref. 62)	NCERC	${ }^{237}$Np sphere	Nickel	8	Upcoming

The first benchmark listed in Table II was performed at LACEF, prior to the move to NCERC. This was a subcritical, source-driven, noise-analysis experiment, led by Oak Ridge National Laboratory in 2002 (Ref. 52). The detection system used in this experiment consisted of four ³He tubes inside an HDPE moderator. This source-driven experiment utilized a ²⁵²Cf source contained in a hemispherical ionization chamber. The RTO configurations included the BeRP ball with varying HDPE thicknesses of 0.00 to 7.62 cm. This experiment utilized the ²⁵²Cf Source-Driven Noise Analysis method,⁵² a neutron noise technique that utilizes correlations between different detectors (cross-power spectral density) and the same detector (auto-power spectral density) to infer the system $k_{eff}$.

The BeRP/nickel (Refs. 53 and 54) and BeRP/tungsten (Refs. 55, 56, and 57) experiments were the first benchmark experiments performed at NCERC. For these experiments, seven and eight configurations were respectively performed with variable reflector thicknesses (both ranging from 0.00 to 7.62 cm). Both used the NPOD detection system,⁵⁴ which consists of 15 ${ }^3$He tubes moderated with HDPE. The measured results were analyzed using the Hage-Cifarelli formalism⁵⁸ of the Feynman variance-to-mean⁵⁹ method. This is a neutron noise method in which the deviation of the measured data from a Poisson distribution within time gates is used to infer multiplicity. For these benchmarks, the values and uncertainties of three parameters were reported: the singles count rate $R_1$, which is the rate at which one neutron from a fission chain is detected; doubles count rate $R_2$, which is the rate at which two neutrons from a fission chain are detected; and leakage multiplication $M_L$. The setup for these experiments is shown in Fig. 1c, where both the SNAP and high-purity germanium (HPGe) detectors were also used.

The Subcritical Copper-Reflected $\alpha$-phase Plutonium (SCR$\alpha$P) experiment, shown in Fig. 11, was performed at NCERC in 2016 and published in ICSBEP in 2019 (Refs. 14 and 60). This experiment included 17 configurations with the BeRP ball. The reflectors included various combinations of copper and HDPE ranging from 0.00 to 10.16 cm. These reflector combinations were chosen to obtain a large range of system multiplication values in addition to maximizing copper nuclear data sensitivities.⁶¹ This experiment utilized the NoMAD detection system, which, like the older NPOD system, is also a multiplicity detector with 15 ${ }^3$He tubes moderated with HDPE. The same benchmark parameters ($R_1$, $R_2$, and $M_L$) from the BeRP/nickel and tungsten experiments were evaluated for SCR$\alpha$P. The French technical safety organization Institut de Radioprotection et de Surete Nucleaire (IRSN) and Lawrence Livermore National Laboratory both performed independent, external reviews for the SCR$\alpha$P experiment.

The Neptunium Subcritical Observations (NeSO) experiment⁶² was performed at NCERC in 2019. This experiment utilized the ${ }^{237}$Np sphere, where eight configurations were measured including the bare sphere and varying nickel hemishells, with a maximum thickness of 9.14 cm. Nickel was chosen as the reflector since it led to the highest multiplication factor among several reflectors investigated including copper, tungsten, iron, beryllium, beryllium oxide (BeO), and molybdenum (Ref. 63). The only reflector material that outperformed nickel was tungsten carbide; however, the marginal improvements in performance were judged not to be worth the large cost difference. Note that ${ }^{237}$Np is a threshold fission nuclide; thus, the best reflectors for ${ }^{237}$Np are very different from those of fissile nuclides. The NeSO experiment has a much larger sensitivity to ${ }^{237}$Np fission than any experiments currently in the ICSBEP Handbook (including the Planet neptunium critical mass experiment utilizing the same ${ }^{237}$Np sphere). The NeSO experiments utilized the NoMAD detection system, similar to the SCR$\alpha$P experiment, and the Detective X device (a mechanically cooled HPGe COTS system). During preliminary measurements prior to NeSO, it was determined that the neutron count rate was higher than anticipated from simulations but lower than given from measurements in the early 2000s. These results seemed to indicate that perhaps a nuclide with a high spontaneous fission rate was present in addition to the already known ${ }^{240}$Pu. Neutron and gamma data indicated that ${ }^{244}$Cm was the most likely nuclide, but additional work is being performed to confirm if this is the case and to ideally quantify the amount of ${ }^{244}$Cm present. The NeSO experiment is shown in Fig. 12.

Subcritical benchmark measurements enabled R&D related to the advancement of methods. Measurement methods that utilized these subcritical benchmark experiments include research into the binning of Feynman histograms,⁶⁴ a figure of merit associated with these histograms,⁶⁵ and fitting of neutron noise parameters.⁶⁶ Additional research was performed in the area of neutron noise measurement uncertainty,^67–69 reproducibility,⁷⁰ the application of Bayesian sampling,^71,72 and the use of parameters that are independent of detector efficiency.^73–75

Benchmark experiments are also valuable for validation of simulation methods, largely focused on development of new capabilities within MCNP®.^a Examples of this research include list-mode simulations,^76–80 evaluation of multiplicity capabilities associated with the CGMF (Ref. 81) and FREYA (Ref. 82) physics event generators,^83,84 and use of a genetic algorithm to recommend nuclear data changes for prompt neutron multiplicity.⁸⁵

IV.B. Detector Testing and Characterization

While benchmark measurements focus on characterizing a system for nuclear data and computational method validation, RTO operations are also performed to test and characterize detector systems (both hardware and algorithms). This work is performed for a variety of sponsors and utilizes various detector systems. The majority of measurements use well-characterized nuclear material, such as the BeRP ball and RF hemishells, to ensure that there are known values for comparison. Often, however, RTO configurations are also chosen to challenge the detection systems in specific ways. Some detector testing involves blind tests in which the RTO contents are unknown to those performing data acquisition and analysis. Configurations may be constructed at NCERC and taken to other locations; Fig. 13 shows an example of a handheld COTS detector measuring a configuration in the field. Often, newer equipment will be tested against current state-of-the-art COTS systems. New detection systems often have goals that include reduction in size, weight, and operational cost of the equipment while maintaining, or improving, the scientific fidelity and precision needed for assessment.

Testing of gamma spectrometers involves several performance metrics. These include technical metrics, such as detector resolution and available energy range, in addition to operational metrics, such as battery life, weight, etc. Gamma spectra are often used to estimate uranium and plutonium isotopics, where the accuracy and uncertainties of these estimates are determined and compared between systems. Shielded detection and RIID performances are also assessed. RIID identification performance is based not only on hardware but also on software. For isotope identifications, true positives, false positives, and false negatives are evaluated using various RTOs. Since many of the detection systems are designed for nonproliferation applications, it is important that they perform well on actual SNM, and NCERC RTO operations provide one of the few locations to conduct such tests with greater than significant quantities of material.

Testing of gamma imaging systems utilizes RTOs designed to challenge hot spot location [i.e., the location(s) of SNM within a specific configuration]. These systems use photopeak and Compton scattering analysis coupled with optical information to identify the hot spot location(s). An example of this can be seen in Fig. 14 for two COTS imaging systems. Some imaging systems include spectroscopic and isotope identification capabilities. All types of gamma detection systems are utilized for detector testing and characterization. Examples include OS, IS [including sodium iodide (NaI), cesium iodide (CsI), lanthanum bromide (LaBr${ }_3$), bismuth germanium oxide (BGO)], and semiconductors [including HPGe, cadmium zinc tellurium (CZT), and others].

Neutron systems of all types have also been used for testing and characterization. These include proportional counters (such as ${ }^3$He, ${ }^4$He, and BF${ }_3$), OSs, and solid-state detectors (boron or lithium based). Objectives for testing neutron systems include accuracy and uncertainty of neutron emission, neutron multiplication, SNM mass, and neutron energy. The large quantity of SNM and nonnuclear components make NCERC an ideal location for testing of neutron systems. RTO configurations range from low neutron emission (tens of neutrons per second) to high neutron emission ($>$10${ }^8$ n/s) and from low multiplication (M$\sim$1) to high multiplication (M$>$10). Configurations can be built that bound those expected for most applications.

One example of neutron measurements for detector testing and characterization used the Thor core with the NPOD neutron multiplicity detectors in 2012 (Refs. 86 and 87). As mentioned in Sec. II, the Thor core consists of three parts (referred to as the upper, center, and lower pieces), with a combined mass of approximately 9.5 kg. The center piece has a 0.5-in.-diameter hole in which samples can be inserted. Various pieces of plutonium were inserted into the core, such as those shown in Fig. 15. Use of these samples allowed for the measurement of 14 configurations (all consisting of bare plutonium), ranging from 200 g (using the samples only) to 9.8 kg (consisting of all three Thor core pieces and plutonium samples). The setup for this experiment is shown in Fig. 16. These experiments provided characterization data to show the limitations of the detection system; the minimum mass difference between configurations that could be distinguished depends upon the system multiplication and was previously documented.⁸⁶ Data from these measurements were used to test new features within MCNP including list-mode⁸⁷ and fixed-source sensitivity⁸⁸ capabilities.

IV.C. University Consortia Measurements

From 2015–2019, over 40 experiments were performed at NCERC for two Defense Nuclear Nonproliferation (DNN) university consortia. This included the Consortium for Nonproliferation Enabling Capabilities,⁸⁹ led by North Carolina State University (NCSU) and the Consortium for Verification Technology⁹⁰ led by the University of Michigan. These measurements were performed on RTOs, primarily consisting of configurations containing the BeRP ball, RF hemishells, or the ${ }^{237}$Np sphere. Some experiments, shown in Fig. 7b, also included AI. Table III gives a summary of the consortia measurements. Over the 5-year duration, eight universities participated in these experiments: NCSU, the University of Michigan, the University of Illinois at Urbana-Champaign, Purdue University, Princeton University, North Carolina Agricultural and Technical State University, the University of Florida, and the Georgia Institute of Technology. In total, 15 faculty members and 41 students or postdocs participated in these experiments. Forty unique experiments were performed, and 110 configurations were measured. Figure 17 shows some of the measurements performed during these measurement campaigns.

TABLE III Summary of University Consortia Experiments

Year	Universities	Faculty	Students/Postdocs	Experiments	Configurations
2015	4	4	8	6	9
2016	4	6	12	8	13
2017	6	2	17	11	25
2018	4	5	10	8	17
2019	5	3	10	7	46

The experiments had varying goals and included different detection systems. NCSU investigated anisotropy in neutron-neutron coincidences⁹¹ (shown on the left of Fig. 17) and spatial characteristics of SNM (Ref. 92) by performing measurements on the BeRP ball and RF hemishells with OS. The University of Michigan studied uranium isotopics of the RF hemishells using gamma spectroscopy with CZT detectors^93,94 and delayed neutron information from AI with lithium glass and PVT detectors.⁹⁵ The University of Michigan used OSs (Ref. 96), ISs (Refs. 97 and 98) (shown in the top-right of Fig. 17), and silicon photomultipliers⁹⁹ to study neutron and gamma imaging. NCSU and the University of Florida also performed imaging experiments using optically stimulated luminescence dosimeters¹⁰⁰ and a system that utilized OS and Lidar.¹⁰¹ Princeton University performed measurements with NaI for template matching to provide information barriers within treaty verification.^102,103 The University of Michigan performed Rossi-$\alpha$ measurements with OS and ${ }^3$He detector systems (shown in the bottom-right of Fig. 17) (Refs. 104, 105, and 106).

IV.D. Neutron Diagnosed Subcritical Experiments

The NDSEs were designed to measure the fission chain decay of RTOs initiated by neutrons from an external source. NDSE measurements have been performed at NCERC since 2016. Three static objects have been constructed to date: Object 1 (shown Figs. 19a and 19b) consists of RF hemishells 3–30 (21.8 kg) (shown in Fig. 18c) reflected by 6.35-cm-thick HDPE. Object 2a (shown in Fig. 19c) uses RF hemishells 1–34 (29 kg) (shown in Fig. 18a) reflected by 2.00-cm-thick HDPE. Object 2b is identical to Object 2a except it does not contain any HDPE (Ref. 107). Objects 2a and 2b are above the water-reflected critical mass for HEU metal¹⁶ and therefore require unique criticality safety requirements to assemble. Some configurations include the replacement of RF hemishells with aluminum hemishells (shown in Fig. 18b).

Fig. 19. (a) NDSE Object 1 measurements. (b) NDSE Object 1 being assembled for DPF operations. (c) NDSE Object 2a during construction

Fig. 20. Example of an NCERC class (BeRP ball demonstration shown)

Measurements with these objects were performed with multiple neutron sources including ${ }^{252}$Cf, AmBe, and a D-T neutron generator. Multiple detection systems have been employed for NDSE including OS, BGO IS, and ${ }^3$He systems. The scintillator systems are primarily used for gamma-ray die-off analysis¹⁰⁷ while the neutron detectors utilize neutron noise methods.^108,109 As shown in Fig. 19c, some measurements were performed outside the DAF. Here, Object 1 was assembled for measurements using a dense plasma focus (DPF) neutron source.

IV.E. Classes

As mentioned in Sec. IV, NCERC has a large inventory of SNM; a large collection of detection systems; and experts in neutron detection, gamma detection, and radiography. This makes NCERC a great place to conduct hands-on training. Students/attendees of NCERC classes include criticality safety practitioners, glove box operators, managers of fissionable material handlers, and personnel working in nonproliferation. A demonstration with the BeRP ball and ${ }^{237}$Np sphere (shown in Fig. 20) has been a part of the Nuclear Criticality Safety Program training for criticality safety practitioners and managers since the establishment of NCERC in 2011. Starting in 2017, NCERC has partnered with the DNN Nuclear Science and Security Consortium¹¹⁰ (led by the University of California, Berkeley) to give workshops for the Keepin summer school.¹¹¹

Classes are tailored based upon the objectives and attendees. Most classes include a hands-on demonstration with the BeRP ball and ${ }^{237}$Np sphere complemented with discussion of criticality safety, thermal properties of plutonium, neutron emission and multiplication (for plutonium, neptunium, and uranium systems), reflection, and neutron absorption. Some classes also include overviews of neutron and/or gamma detection and may include radiography. Demonstrations on the Planet,⁸ Comet,⁹ Flattop,¹⁰ and/or Godiva IV (Ref. 11) critical assembly machines may also be performed.

V. CONCLUSIONS AND FUTURE WORK

Radiation test objects are uniquely useful because their geometry, material characteristics, and radioactive emissions provide surrogates for real-world items of concern or interest. This makes RTOs especially useful for detector testing and validation. The multiplication and neutron emission of RTOs also make them useful for benchmarks, which can be used to validate nuclear data and code capabilities. Highlights from the first 10 years of RTO operations at NCERC were presented. Over 900 RTO configurations were constructed in the first 10 years at NCERC, primarily for measurements involving neutron detection, gamma detection, and/or radiography. While RTO operations slowed in 2020 due to COVID-19, the future of these operations is bright and will include many measurements, experiments, and classes.

Acknowledgments

This work was supported by the DOE Nuclear Criticality Safety Program, funded and managed by the National Nuclear Security Administration for the DOE. Work detailed within this document was performed with the support of many sponsors including the Nuclear Criticality Safety Program, Nuclear Counter-Terrorism, Defense Threat Reduction Agency, DNN, and others.

This work was supported by the US Department of Energy through the Los Alamos National Laboratory. Los Alamos National Laboratory is operated by Triad National Security, LLC, for the National Nuclear Security Administration of the US Department of Energy under Contract No. 89233218CNA000001.

The authors would like to thank all involved in the success of these measurements (including all logistics to perform measurements in a nuclear facility). We would like to specifically thank the following individuals for their significant contributions to RTO operations in the past 10 years: Jennifer Alwin, Rian Bahran, Timothy Beller, Benoit Richard, Timothy Dugan, James Dyson, Jeffrey Favorite, Ryan Kamm, Keith Lash, Donnette Lewis, Alex Lynn, Justin Martin, John Mattingly, Mark Mitchell, Dane Spearing, Clell Solomon, Avneet Sood, Kenneth Valdez, Timothy White, and Mission Support and Test Services/National Security Technologies operations.

Notes

^a MCNP® and Monte Carlo N-Particle® are registered trademarks owned by Triad National Security, LLC, manager and operator of Los Alamos National Laboratory. Any third party use of such registered marks should be properly attributed to Triad National Security, LLC, including the use of the designation as appropriate. For the purposes of visual clarity, the registered trademark symbol is assumed for all references to MCNP within the remainder of this paper.