The SNAP-2 Development Reactor (S2DR) was the next stage in testing for the SNAP-2 program, Sometimes called the SNAP-2 Development System (S2DS or SDS), it was designed to refine the design of particular components of the SNAP-2 flight reactor design, test materials issues, and provide better calibrations and design optimization for the system that was meant to fly as part of the SNAPSHOT program.
By 1959, the test objectives for the S2DR had been submitted, and the design and construction schedule was set by September of that year. 1961 saw the completion of the facility needed for testing, with fuel elements being fabricated by April of 1961. Dry critical tests were completed over the summer, and the core was filled with coolant for thermal testing shortly after. The core underwent wet critical zero power testing in July 1961, and testing was begun at SNAP-10 design conditions shortly after, with over 1100 hours of testing completed by September. It would continue testing until
SNAP-2 Development Reactor Fuel And Core
This reactor had a roughly octagonal shape, rather than the hexagonal design of the SNAP-2 Experimental Reactor (S2ER), but the drum and safety assemblies for the reactor was very similar to that used in the S2ER. The fuel elements were modified several times during this process, although the overall dimensions didn’t change much, due to a number of issues seen in the S2ER fuel elements: welds had failed in the fuel elements, and the hydrogen diffusion barrier that was used (a boron-free enamel coating on the inside of the steel tube) had completely failed on occasion, leading to a loss of hydrogen (and therefore reactivity worth) in the fuel elements. While primarily designed to test the SNAP-2 reactor design, due to the similarities in core geometry (after the change from SNAP-10 to 10A), the reactor was also used at SNAP-10A power and temperature levels, and the first powered tests were carried out at this point.
The fissile fuel itself for the S2DR was 235UZrH, similar to that used in the S2ER. The uranium was enriched to 93%, with each fuel element being made up of 10% (about 128 grams) uranium metal. While the fuel portion of the element remained 10” long, it increased in diameter to 1.212”. The 1 1/2” internal reflectors were changed from Be metal to BeO. The clad was changed to Hastelloy N, coated internally with a hydrogen diffusion barrier made up of a glass enamel coating with samarium (Sm2O3) burnable poisons to maintain the reactivity of the reactor more effectively.
The core contained a total of 37 fuel elements, mounted on 1.26” centers, to form a similar triangular lattice hexagon, 8” across on flats and 9” across on corners. Rather than the largely hexagonal cross section of the outer reflector that was used in the S2ER, the S2DR had a more octagonal structure. Two control drums mounted opposite each other provided active reactor control during testing, forming two of the four wide sides of the external reflector. The rest of the S2DR’s core periphery is surrounded in two parts by a Be filler, with cutouts to match the control drums and one central, large flat portion and two smaller, slightly inset, flat portions used to place Be shims to control the reactivity in the core. The long sides, mounted at a 90 degree angle to the control drums, were made up of Be plates of variable thickness to tweak the reactivity of the core, which functioned as the scram safety mechanism in much the same way that the S2ER’s scram functionality was used. One side had an additional cutout immediately adjacent to the core for the coolant return pipe. The thermal equilibrium point of the core was at about 900F, with the maximum fuel element temperature being 1250F. There was an external reactivity source used on the reactor, probably mounted to one of the control drums, but I haven’t seen any direct evidence of this, and the SCA4 used a different mechanism for external reactivity than the S2ER, so it may have been configured differently.
S2DR Operational History
By the end of March 1961, fuel elements began to arrive at the now-completed facility, with the first elements being loaded in the core. Dry critical tests (first with 31, then 37 fuel elements), calibration and worth tests, flux distribution testing, and other experiments were done throughout April and May in the dry configuration until the latter half of May when the fuel elements were removed. June 1961 saw the reloading of the fuel elements, followed by the filling of the core for the first time with the NaK coolant and wet critical testing, as well as the thermal coefficient calibration testing. July’s testing schedule included hydrogen migration experimentation, as well as the first fission power operation at low power (7 watts), at temperatures from 600F to 882F. After a brief test at 140 watts, the reactor was shut down briefly in the beginning of August to prepare for the first operation at SNAP power and temperature conditions. This occurred from the middle of August to the end of September, when the xenon transient experiment was conducted under 1 kW of power and 500F temperature conditions. The temperature was increased for more hydrogen migration testing in the beginning of October until the middle of November, when control drum calibration testing was done and the reactor was shut down again. Further thermal testing was done, including the first high-temperature experimentation (at 1050F) for both primary system heat loss and hydrogen migration, until the end of the month. December was taken up with additional hydrogen migration experimentation at high temperature, low power (280 W, 1050F), until the reactor was shut down for maintenance until just before Christmas. No more power testing would occur until 1962.
January’s testing was largely consumed with SNAP-2 temperature but SNAP-10A power testing (30.5 kW, 1140F), with a source effect experiment and reactor shutdown at the end of the month. February saw more high-temp, low power (300W, 1050F) hydrogen redistribution testing, until the beginning of March, when the reactor was shut down for maintenance again. More hydrogen redistribution testing (30.5kW, 1050F) consumed the rest of the month. After a brief shutdown, the first half of April saw power coefficient testing, first at SNAP-10A conditions, then SNAP-2’s higher power and temperature. Sadly, the heat exchanger between the two coolant loops failed, requiring a reactor shutdown and repair cycle. After this was completed, the SNAP-2 power coefficient testing started back up, and continued until the beginning of May. The power and temperature were reduced for static power coefficient testing, followed by power and flow oscillation experiments, as well as fuel element-to-coolant heat transfer experiments. The reactor was shut down once again, and the external reactivity source was either modified or replaced, followed by additional control drum calibration experiments with the new source. These tests continued until the middle of June, when the reactor was restarted to SNAP-2 operating conditions. This round of S2-type testing would continue until the end of July, when the inter-loop heat exchanger failed yet again, causing another core shutdown. August saw experiments in ramped reactivity insertions, followed by the draining of the primary NaK coolant, and additional reactivity ramp insertion testing in a dry condition for the rest of the month and in to the beginning of September, when the core was refilled with NaK, the control drums were recalibrated, and the reactor returned to SNAP-2 conditions, until a suspected primary coolant leak caused yet another shutdown. No coolant leak was located after about two weeks of testing, so the reactor was restarted at various low-power, low-to-medium temperature conditions to calibrate the testing equipment, and in the middle of October the reactor was returned to SNAP-2 operating conditions until the end of November. Further high-temperature, SNAP-10A power level testing occurred from the end of November until the middle of december, when the reactor was shut down for the last time. A number of no-power coritical experiments were conducted after the final power-down, and the reactor stopped being used toward the end of the month.
By the end of the S2DR’s operational history in December 1962, it operated for 11,290 hours with a total energy release of 272,900 kWh. 2060 hours of this were at SNAP-2 operating conditions (55 kW, 1200F), 1544 hours at 30.5 kW and 1100F, and 1150 hours at SNAP-10 operating conditions (30.5 kW, 950F). The reactor was disassembled, and the fuel elements examined post-irradiation, in 1963.
Test Results of the S2DR Program
Much of the testing done as part of the S2DR program was focused on the fuel element behavior, since the basic reactor physics had been well-characterized by the S2ER. This made fuel element examination a big part of the test results (and also happens to be the part that I’ve been able to find). The U content burnup was estimated to be 0.027% over the course of the reactor’s lifetime, and the operational history was fed into a number of modeling programs to refine predictions on reactor behavior.
As far as overall fuel element behavior, both the radial power density for the fuel elements themselves and the core as a whole were well within expected values. The visual appearance of the fuel elements were also, in most cases, identical to when they were placed in the core, with the exception of two fuel elements which had chemical reactions between the Hastelloy clad and the Be end reflectors. This was likely caused by a failure of the chromium coating of the Hastelloy due to machining marks. Despite this, the testing continued with high hopes for the results. No significant fuel swelling was discovered, either radially or in length, however the density of the fuel itself dropped from 6.033 g/cc to 6.030 (+/- 0.005) g/cc.
One of the biggest challenges for the SNAP fuel program was hydrogen migration from the ZrH fuel element matrix. This type of fuel is wonderful in many ways, including moderating the neutron flux, allowing a homogenous fuel element composition that makes modeling reactor dynamics easy, and a number of other advantages (there’s a reason that the TRIGA reactor, used worldwide as a safe, effective research and training reactor to teach students and nuclear power operators, uses similar fuel), but it’s very temperature sensitive. The higher the temperature of the fuel element, the faster the hydrogen is going to migrate out of the fuel. In order to minimize this, a glass enamel is used to coat the inside of the clad tubes which is impervious to hydrogen. The behavior of these enamel coatings under thermal cycling, how well they contain the hydrogen, and any chemical reactions under the irradiation conditions seen in the core was a major focus of the program. Another area of concern when it comes to hydrogen is that it can migrate within the fuel element, creating reaction “hot” and “cold” spots within the fuel element and changing the temperature distribution within the element. This can cause thermal and chemical reaction problems (as it ended up doing), as well as problems for the power distribution within the core, which will only grow over time. Worse, if thermal distribution changes occur, they can cause unforseen chemical reactions in the fuel element casings, thermal failure in the enamel liner, or another problem with the clad, hydrogen will be released from the fuel elements and into the coolant, creating even greater reactivity changes!
Hydrogen loss was higher than expected in the fuel elements examined, but the overall migration was from 1.83% by weight to between 1.7 and 1.81% by weight post-irradiation. This means that, while the system could be (and was) better designed to capture the hydrogen, it was a verysmall change in the overall amount of hydrogen in the fuel element. Some changes in hydrogen permeation were unexpected, including migration from inner fuel elements to outer, but overall the FEs were not considered overly compromised by these migrations to prevent operation of the reactor.
All of the lining enamels are carefully made to eliminate any boron content, and in the case of the S2DR and later SNAP fuel elements a burnable fission poison (samarium, in the form of Sm2O3) was added. This technique allows for a more even amount of reactivity in the core, since the burnable poisons are destroyed as the fission products that act as fission poisons build. However, in order to do this effectively, the buildup of fission poisons needs to be modeled (in something like the CINDER MCNP code, for instance, but this wasn’t available at the time), or experimentally verified, and the burn-up rate of the Sm needs to be carefully matched. Failure of the lining enamel was reasonably common in the S2ER fuels, so it was important to study the new design, and its’ behavior, during the S2DR tests. This would continue through the Snap Critical Assembly 4 tests, all the way to the cancellation of the SNAP program, and inform clad designs for other hydride fueled reactors after the program was a footnote in history. While the lining was found to be chipped in some of the examined fuel elements, on the whole the new lining seemed to be a good step in the right direction.
Another concern in any fuel element is metallographic changes in the fuel element matrix. Changes in grain boundary size and orientation can change a homogenous structure, with similar neutron scattering throughout, into one that selectively reflects or refracts neutrons into a particular area causing uneven fuel burnup, variable thermal conductivity, materials failure due to neutron displacement, and other unwanted effects. This can also escalate, with defects growing if the initial defect can act as a see crystal for the growth of other metal crystals within the structure of the fuel element. This was observed in the ends of the fissile fuel, but the bulk of the fuel remained in good condition.
Overall, the test results from S2DR were promising for both the SNAP-2 and -10A designs. The criticality data gathered would provide key benchmarks for the SNAP-2/10 core going forward, and prepare the reactor for flight-ready status.
Further Reading and Sources
The quarterly status reports on the SNAP Reactors Program also carry more valuable details about the development, testing, and operation of these reactors, available on the SNAP main page.
SNAP 2 DEVELOPMENTAL REACTOR MOCKUP NO. 1 FINAL REPORT, Maki 1968 https://www.osti.gov/servlets/purl/4516729
THE FINAL REPORT ON THE SNAP 2 EXPERIMENTAL REACTOR (SER) OPERATION AND TEST PROGRAM 1962 https://www.osti.gov/servlets/purl/4591868
Fuel Element Development and Analysis
EVALUATION OF FUEL ELEMENTS FROM SNAP 2 DEVELOPMENT REACTOR CORE 1965 https://www.osti.gov/servlets/purl/4471462