Comment on J. Fernandez et al, "Requirements and sensitivity analysis for temporally- and spatially-resolved thermometry using neutron resonance spectroscopy," Rev. Sci. Instrum. 90, 094901 (2019)
aa r X i v : . [ c ond - m a t . m t r l - s c i ] J a n Comment on J. Fern´andez et al, “Requirements and sensitivity analysis fortemporally- and spatially-resolved thermometry using neutron resonancespectroscopy,” Rev. Sci. Instrum. 90, 094901 (2019).
Damian C. Swift ∗ and James M. McNaney Lawrence Livermore National Laboratory, 7000 East Avenue, Livermore, California 94551, USA (Dated: October 7, 2019 – LLNL-JRNL-795286)
Neutron resonance spectrometry (NRS) offers a uniqueway to measure the bulk temperature inside an optically-opaque sample such as a metal, with temporal resolutionsuitable to probe inertially-confined states of elevatedpressure produced by dynamic loading, but requires anintense neutron pulse and has previously been demon-strated only with a nuclear spallation source of o (10 )neutrons and shocked states induced by a centimeter-scale projectile (launched by high explosives) persistingfor the order of a microsecond [1, 2]. We have previ-ously undertaken research to investigate laser-generatedneutron pulses [3], assessing whether it would be feasi-ble to use such pulses for single- or few-pulse NRS mea-surements. We also considered whether such measure-ments could be optimized for use with laser-induced dy-namic loading, which would make the synchronization ofthe load and probe much more convenient and poten-tially less expensive than with the projectiles and parti-cle accelerator used previously. Our conclusion, which wedid not publish but did discuss with interested parties,was that such measurements were not feasible with the o (10 ) neutrons demonstrated in our experiments, evenif the measurement could be applied to a resonance ofsignificantly higher energy than the ∼
20 eV resonancesused previously [1].The article by Fern´andez et al argues that, becauseof advances made in laser-produced neutron pulses, itshould now be possible to obtain a temperature mea-surement by neutron resonance spectrometry from a sin-gle sub-picosecond laser pulse in the 300 J range pro-ducing o (10 ) neutrons. As a point design, the articletakes the moderator and detector configuration used inthe previous experiments at the LANSCE accelerator,and argues that, because a laser-driven neutron sourcewould not produce the γ -ray flash that accompanies theaccelerator-driven spallation event and which can blindthe neutron time-of-flight (NTOF) detector, the detec-tor can be moved much closer to the moderator: 3.1 minstead of 23 m. From standpoint of designing new ex-perimental facilities, this would make it possible to makeNRS measurements of dynamically-loaded states with alaser source of neutrons costing ∼ $10M rather than a par-ticle accelerator source costing ∼ $1B. The article impliesthat a future facility would induce transient high-pressurestates using a second laser, with a pulse ∼
200 ns.The article gives an estimate of the number of neu-trons N needed to be detected in the region of the res- onance for a temperature measurement of a given accu-racy, 2 /π / √ N . In our own analysis, we deduced that N should strictly be the number of neutrons that wouldhave been detected had it not been for the effect of theresonance, i.e. the number scattered by the resonancemultiplied by the efficiency of the detector, ǫ d . Apartfrom this subtlety, we arrived at much the same resultthat scientifically useful dynamic temperature measure-ments at the percent level require N ∼ /ǫ d neutrons.Again, we did not publish this result (though we did com-municate it to the authors). The challenge in NRS mea-surements is that, although pulsed neutron sources aredemonstrably capable of producing several orders of mag-nitude more neutrons, it is difficult to moderate enoughof them from their source energy to the energy range ofthe resonance and pass them through the sample while itis maintained at an inertially-confined state of elevatedpressure, to achieve this number.We note that short-pulse laser interactions used toaccelerate charged particles, and thereby neutrons, alsoproduce significant levels of electrical noise and energeticphotons. It is not clear whether NTOFs could be movedas close as 3 m from the moderator, though they couldlikely be closer than 23 m. In addition, there seems to beno particular reason why the dynamically-loaded sam-ple needs to be at ∼ A s must lie between A m and A d , the facing areas of the moderator and detector re-spectively. If these have the same shape, characterizedby a scale length l ∝ √ A , then the scale length of theloaded sample l s = l m + ( l d − l m ) LL + L d , (1)where L and L d are the moderator to sample and sam-ple to detector separations respectively. For the exam-ple configuration, with A m =56.25 cm , A d =1393 cm , L =1 m, and L d =3.1 m, A s ≃
212 cm . In practice, asshown in the figure in the article, the sample would likelybe oriented with its normal at an angle θ say to the neu-tron beam, in which case the actual loaded area wouldbe at least a factor 1 / cos θ larger. Putting it mildly, thiswould be an impractical area over which to induce dy-namic loading with a laser. Even the limiting case wherethe sample is butted up flat against the moderator wouldrequire at least 56.25 cm to be uniformly driven, whichis far in excess of any laser experiment we are aware of.The cost of a laser capable of such an experiment de-pends on other aspects, such as the peak pressure ofinterest, and whether it could be achieved by tampedablation [4–6] or whether ablation into vacuum wouldhave to be employed, but it could certainly exceed thecost of the National Ignition Facility (NIF) by a factor ofseveral, and could thus easily exceed $10B. Tamped ab-lation on timescales o (0 . µ s has not, to our knowledge,been demonstrated for pressures above 20 GPa [7]; freeablation at 100 GPa requires ∼
20 PW/m (or J/ns.mm ),i.e. ∼ , compared with ∼ O (100 keV-MeV) neutrons in polyethylene is ∼ ∼ ∼
20 eVrequires ∼
15 scatters, or a random walk of ∼ o (0 .
1) m in size, which willemit neutrons over an area o (0 .
01) m as at LANSCE.The design strategy used at LANSCE was to reducethe effective size of the neutron source by collimation,sacrificing neutrons. Another strategy is to use a moder-ator thinner than the mean free path to the first scatter- ing event, for instance selecting neutrons scattered nearlystraight back toward the source, which would have thelowest energy. This strategy again throws away most ofthe neutrons, but gives both a relatively small size and ashorter spread in time over the width of any given reso-nance. It could possibly be practical with neutron yields o (10 ) or more as have been demonstrated at NIF [8],but may still need to be combined with a drive pulse oflonger duration than available there.An avenue for further refinement is to consider higherenergy resonances, requiring less down-scattering of theneutrons and thus a thinner moderator and shorter min-imum duration of the load. However, these advantagesmust be offset against potentially less efficient neutrondetectors at higher energies, and larger distances to thedetector to achieve a given NTOF spectral precision.In conclusion, the increase in neutron yields fromshort-pulse laser sources, demonstrated since our previ-ous study, are not yet adequate for NRS temperaturemeasurements of dynamically-loaded samples. Simplybringing the detector closer to the moderator would nothelp as dynamic loading then becomes impractical. Or-ders of magnitude advances in neutron production areneeded before short-pulse laser facilities could usefullyperform single-shot NRS thermometry. Nuclear spal-lation at the LANSCE accelerator, and possibly iner-tial confinement fusion at NIF, are as yet the only neu-tron sources bright enough to contemplate employing forsingle-shot thermometry. LANSCE remains the only fa-cility in which such experiments are currently feasible,using ∼ microsecond scale dynamic loading. It is surpris-ing that greater advantage has not been taken of thisunique capability.This work was performed under the auspices of theU.S. Department of Energy under contract DE-AC52-07NA27344. ∗ Electronic address: [email protected][1] V. W. Yuan, J. D. Bowman, D. J. Funk, G. L. Morgan, R.L. Rabie, C. E. Ragan, J. P. Quintana, and H. L. Stacy,Phys. Rev. Lett. , 125504 (2005).[2] D. C. Swift, A. Seifter, D. B. Holtkamp, V. W. Yuan,D. Bowman, and D. A. Clark, Phys. Rev. B , 092102(2008).[3] D. P. Higginson, J. M. McNaney, D. C. Swift, T. Bartal,D. S. Hey, R. Kodama, S. Le Pape, A. Mackinnon, D.Mariscal, H. Nakamura, N. Nakanii, K. A. Tanaka, and F.N. Beg, Phys. Plasmas , 100701 (2010) and additionalunpublished work.[4] R. Fabbro et al, J. Appl. Phys. , 775 (1990).[5] J.D. Colvin et al, Phys. Plasmas , 2940 (2003).[6] D.L. Paisley, S.-N. Luo, S.R. Greenfield, and A.C. Koskelo,Rev. Sci. Instrum. , 023902 (2008).[7] M. Akin and D.C. Swift, unpublished.[8] S. Le Pape et al, Phys. Rev. Lett.120