670 5 NOVEMBER 2021 • VOL 374 ISSUE 6568 science.org SCIENCEPHOTO: RUSH CAPPELLETTIsuch as a running engine or a steel girder.
Because neutrons act like little magnets,
they can reveal the atomic-scale patterns
of magnetism within materials.
Hore uses neutrons to probe the dynam-
ics of polymers, the chainlike molecules in
plastics and many biological materials. By
replacing hydrogen in a specific part of a
molecule with deuterium—which neutrons
are more likely to bounce off—researchers
can track that bit of the molecule in a
jumble of similar material, Hore says.
Neutrons also do less damage to delicate
samples than x-rays, says Tonya Kuhl, a
chemical engineer at UC Davis. “You can
just blast biological samples with neutrons
and it’s not an issue,” she says.
The NCNR reactor generates just
20 megawatts of heat—less than 1% as
much as a typical power reactor—and is
the smallest of the three main neutron
sources in the United States. The
other two are the 85-megawatt
High Flux Isotope Reactor (HFIR)
at the Department of Energy’s
(DOE’s) Oak Ridge National Labo-
ratory and the $1.4 billion Spall-
ation Neutron Source (SNS), also
at Oak Ridge, which fires protons
from an accelerator into a target
to blast out pulses of neutrons.
Nevertheless, NCNR’s 29 spectro-
meters, imagers, and other in-
struments nearly equal the total
at HFIR and SNS together, and
the lab serves more than 2600
researchers each year. Scientists
credit the lab’s staff of 194 with
its success. “In some odd way,
having a lower power reactor has
forced them to be more creative,”
Birgeneau says.
On 3 February, research at
NCNR came to an abrupt halt.
At 8 a.m. that Wednesday, operators be-
gan to restart the reactor after a refueling
stop, according to a report NIST submit-
ted to the Nuclear Regulatory Commission
(NRC) last month. At 9:07 a.m., the reac-
tor’s power plummeted from 15 megawatts
to 7 megawatts. Within 1 minute, monitors
sensed radiation in the reactor’s reinforced
concrete confinement building. At 9:
a.m., automated systems shut down the re-
actor and sealed the building.
Recognizing that the reactor’s uranium
fuel may have been damaged, the operator
immediately issued an alert, the most urgent
alarm a research facility can sound, says
Scott Burnell, an NRC spokesperson. Melt-
ing fuel is rare at research reactors, says Dale
Klein, a nuclear engineer at the University of
Texas, Austin, who chaired NRC from 2006
to 2009. “I used to be the director of our re-
search reactor,” he says, “and my worst night-
mare was having a fuel failure.”
Still, the safety systems worked. The
10 staff then in the confinement building
received a radiation dose roughly equal
to that from a computerized tomography
scan, NIST’s website says. Only trace
amounts of three radioactive isotopes es-
caped the building, and radiation at the
boundary of NIST’s 2.34-square-kilometer
campus never increased above background
levels, according to the website.
NIST traced the accident’s roots to a mis-
take in refueling the reactor a month earlier.
Typically, operators replace the oldest four
of 30 rodlike fuel elements and rearrange
the others. Working by feel, they use a spe-
cial tool to lock each element in place by
twisting it into a spring-loaded latching
mechanism, and an inexperienced crew
failed to secure one element. Circulatingcooling water then lifted it out of position,
impeding the flow around it. When the
reactor restarted, the element overheated
and part of its aluminum cladding melted.
Staff turnover and institutional cul-
ture played a role in the accident, accord-
ing to the NIST report. In 2011, nine of
NCNR’s 21 reac tor operators had more than
20 years of experience. Now, just three
of 22 do. NIST also relied too much on
hands-on training and too little on explicit
procedures to guide operators, the report
says. “We didn’t make that transition ef-
fectively from a skills-based workforce to a
knowledge-based workforce,” Dimeo says.
The reactor suffered no damage beyond
the overheated fuel element, Dimeo says.
Workers have removed all but three ele-
ments and now plan to clear debris from the
core and purify the deuterated cooling wa-ter, he says, which means the reactor cannot
restart before April. However, NCNR needs
NRC’s explicit permission to restart it, and
lab officials must convince the commission
they have eliminated the causes of the acci-
dent, which could take longer, Burnell says.
Still, he says, “The agency absolutely under-
stands the importance of the facility, and we
are going to do the most prompt and thor-
ough review we can.”
A radiation release led to the loss of an-
other neutron source 2 decades ago. In 1996,
researchers found tritium in the ground-
water near the reactor-based source at DOE’s
Brookhaven National Laboratory. Although
the leak was small and confined to the
Brookhaven campus, public outcry led DOE
to permanently shutter the reactor in 1999.
NIST hopes openness will forestall a simi-
lar outcry. It held a virtual meeting with resi-
dents of the suburbs surrounding the lab on
10 February, says Jennifer Huergo,
a NIST spokesperson. NIST is also
posting its communications with
NRC to its website and updating
residents by email, Huergo says.
“As soon as we get questions, we
respond as quickly as we can.”
The shutdown comes as the
United States has lost the lead
in neutron resources. DOE shut
down an accelerator-based neu-
tron source at Argonne National
Laboratory in 2008 and, in 2015,
stopped supporting basic re-
search at an accelerator-based
source still running at Los Alamos
National Laboratory. “Europe has
us beat by a factor of three, Asia
has us beat by a factor of two,”
Kuhl says. “It’d be catastrophic to
lose more capabilities.”
Even before the accident, re-
searchers fretted about the reactor-
based neutron sources. Both NCNR’s reactor
and 55-year-old HFIR run on “highly en-
riched” fuel, in which more than 90% of the
uranium is the fissile isotope uranium-235. In
principle, such fuel could be fashioned into
a nuclear bomb. As part of U.S. antiprolif-
eration efforts, both reactors are supposed
to be retrofitted to use a fuel containing
less than 20% uranium-235 when it be-
comes available, perhaps in the 2030s. In
recent years, expert panels have also ad-
vised NIST and DOE to plan to replace the
aging reactors.
Planning needs to start now because it
will take 10 to 20 years to build a new re-
actor, Birgeneau says. NCNR’s reactor is li-
censed through 2029, and Dimeo envisions
renewing its license even as NIST explores
replacing it. “From our perspective,” he
says, “neutrons are here to stay at NIST.” jA small research reactor in the Maryland suburbs has been a mainstay
of neutron beam studies for a half-century.NEWS | IN DEPTH