HAL Id: jpa-00224023
https://hal.archives-ouvertes.fr/jpa-00224023
Submitted on 1 Jan 1984
HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.
ABSOLUTE, PROMPT GAMMA-RAY
SPECTROSCOPY AND THE DETERMINATION OF FUNDAMENTAL CONSTANTS
R. Deslattes, G. Greene, E. Kessler, Jr.
To cite this version:
R. Deslattes, G. Greene, E. Kessler, Jr.. ABSOLUTE, PROMPT GAMMA-RAY SPECTROSCOPY
AND THE DETERMINATION OF FUNDAMENTAL CONSTANTS. Journal de Physique Colloques,
1984, 45 (C3), pp.C3-41-C3-46. �10.1051/jphyscol:1984309�. �jpa-00224023�
ABSOLUTE, PROMPT GAMMA-RAY SPECTROSCOPY AND THE DETERMINATION OF FUNDAMENTAL CONSTANTS
R.D. Deslattes, G.L. Greene and E.G. Kessler, Jr.
Quantum Metrology Group, National Bureau of Standards, Washington, D.C. 20234, U.S.A.
Résumé - Il existe une échelle de longueur d'onde absolue précise pour les radiations électromagnétiques qui s'étendent des microondes aux rayons Y ayant des énergies inférieures à 1 MeV. Cette échelle commence avec l'horloge à jet atomique de caesium (et ainsi le mètre SI) et continue jusqu'au laser HeNe stabilisé à l'iode. Un tel laser est alors utilisé pour déterminer l'espacement du réseau d'un cristal de Si utilisant l'interférométrie optique à rayons X. Des cristaux calibrés précisément sont alors utilisés dans un diffractomètre à cristal plat pour déterminer des longueurs d'onde absolues de rayon Y-
Nous proposons d'étendre cette échelle à la région 1-10 MeV. Ceci requiert l'emploi de sources intérieures à la pile pour pouvoir examiner les rayons Y rapides provenant de réactions n-y.
Abstract
There currently exists a highly accurate absolute wavelength scale for electromagnetic radiation which extends from microwaves to gamma-rays having energies less than ^ 1 MeV. This scale begins with the cesium atomic beam clock (and thus the SI meter) and continues through the iodine stabilized HeNe laser. Such a laser is then used to determine the lattice spacing of a single crystal of Si using x-ray/optical inter- ferometry. Accurately calibrated crystals are then used in a flat crystal diffTactometer to determine absolute gamma-ray wavelengths .
We propose to extend this scale to the region of "v 1-10 MeV. This requires the use of in-pile sources for the examination of prompt gamma-rays from n-y reactions.
Introduction
Powerful research reactors produce not only neutrons and neutrinos in abundance but also photons especially in the y_ r ay region. Such photons are likewise applicable, in certain cases, to questionsof fundamental phy- sical significance especially through accurate, high-resolution spec- troscopy. The main purpose of this report is to outline a program of research newly undertaken here at ILL. To make clear what significance such work can have and its limitations, we must, however, attend at least briefly to describing the system of spectroscopy which we use.
Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphyscol:1984309
C3-42 JOURNAL DE PHYSIQUE
The starting point for our measurements in the y-ray region is an iodine stabilized HeNe laser operating in the visible near 633 nm (0.5 eV). Such a laser is "locked" to a particular hyperfine component of the (Doppler- free) saturated absorption spectrum of molecular iodine. Although several
versions ofsuch lasers are in use, these are well related to one-another by means of heterodyne techniques and may be, for our purpose, regarded as equivalent. This choice of basis for y-ray measurement has a two-fold significance. On the one hand certain I2 stabilized HeNe lasers have been used in recent determinationsof the Rydberg constant, Rm, by interferome- tric comparisons with resolved components of Balmer a (3d + 2p) in H (and D).
Such a reference is an appropriate one for tests of Q E D in electronic and muonic atoms and for determining masses of (negatively) charged elementary particles (eg., pions and kaons) as has been previously discussed (1). On the other hand, I2 stabilized HeNe lasers are also well-connected to the conventional scales of wavelength and frequency which are nowadays equivalent owing to progress in respect of visible frequency synthesis (2). In the absence of a fundamental theory, such connection to the conventional scale(s) is appropriate for our planned work at ILL which aims, inter alia, at determining the neutron's mass and the composite physical constant, NAh/c.
As discussed below and in Ref. 1 , our y-ray spectrometer uses two flat crystals insymmetric Laue (transmission) geometry. This arrangement offers the possibility of accurate wavelength determinations if crystal spacings have been otherwise established and if diffraction angles are measured on an absolute basis. Unfortunately, such instruments have a very small effective solid angle of acceptance, R,and a relatively low efficiency,
E ; the product eQ is typically in the range 1
o-'
to 10-I for0.1 < E < 1.0 MeV. This feature of such instrumentation entails the
need to use rather intense sources with typical activities in the range of kilo Curies. In our earlier work at lower energies, 0.04 < Ey
<
1.0 MeV, it was possible to use relatively long-lived sources produced in the NBS reactor (10 MW) and carried to a measuring room located just outside of the reactor's containment vessel. For the newly undertaken measurements where Ey > 1 MeV, such long-lived, intense sources are not available so that we must work with sources internal to a reactor. This is uniquely convenient at ILL owing tosource handling capability developed in the GAMS program. By this means we have the possibility not only to obtain intense activities at high energies
but also to study certain important prompt y-transitions directly rather than accept increased errors arising from use of auxiliary transfer standards.
Our goal is to obtain accurate (sub-ppm) values for y-ray wavelengths in terms of the visible reference described above. The needed ratio is rather large ( % 10 6 ) and thus requires (in our opinion) at least one interme- diate step. Characteristic X-ray lines lie in an appropriate region to serve this purpose and have been used as "stepping stones" in the past.
Continuation in this way seems undesireable owing to the large width (+ 500 ppm) of such lines, their appreciable asymmetries and the presence of largely unresolved fine detail arising from spedator hole satellites.
~ a t t i c e repeat distances in well-chosen specimens of certain (mostly synthetic) crystals are more sharply defineable but are attributable only to particular individual samples. This is not a limitation in principle since, in a hypothetical chain from a visible wavelength to one in the y-ray region the (intermediate) crystal spacing has no influence on the final ratio provided that it is sharply defined and changes only slowly.
Both these conditions are well satisfied by currently available germanium and silicon.
For certain tecl~nical reasons, it is inconvenient to use the same crystal specimen for both comparison with the visible reference wavelength (by X-ray/optical interferometry) and for diffraction of y-rays. We therefore
introduced an extra step in the chain from visible to y-rays. In this extra step, crystals with different morphologies but having at least one near degeneracy in interplanar spacing are compared with one another exploiting the near degeneracy to reduce sensitivity to the unfortunate features of X-ray line shapes noted above. One can (properly) think of this step as an "almost non-dispersive" transfer where the datum obtained is the difference in lattice parameters which, if small enough, does not degrade the process.
Overall then, our procedures involve the following steps :
1. Determination of a Si lattice spacing by optical and X-ray interferometry.
2. Transfer of the results of step 1 to other specimens of Si and Ge.
3. Measurement of diffraction angles on an absolute basis.
The angle, 8 , measured in step 3 combined with lattice repeat distance, d, obtained from step 2 yields X via the usual Bragg-Laue relation :X = 2d sin 0.
This procedure has been described in detail in Ref. 1 to which the interes-
C3-44 JOURNAL DE PHYSIQUE
ted reader is referred. Here we only very briefly describe the techniques involved and the current status of each of the above steps.
The NBS measurement of the(220)repeat distance of a silicon crystal used Fabry-Perot interferometry on the optical channel and X-ray Moire inter- ferometry on the "lattice" side of the comparator. The result claimed an accuracy of 0.15 pprn (estimate of lo). This result was transferred to other specimens and these were used in y-ray measurements. Subsequently an X-rayloptical interferometer exercise was also completed at the PTB in Braunschweig ( 3 ) . This claims an accuracy of 0.06 pprn for the specimen used. The difference between NBS and PTB results, namely about 1.5 pprn is larger than likely sample-to-sample variability.
Efforts to resolve the problem through sample interchange and re-examination of systematic problems are still underway. At NBS this has involved re-esta- blishing the entire X-ray/optical interferometer experiment with numerous improvements including on-line monitoring of trajectory and diffraction phase-shift corrections. We anticipate improved precision and accuracy at the level of 0.1 pprn or less.
The lattice parameter-transfer measurement as practiced in the past used instrumentation not quite adequate to resolve a type of diffraction fine strucutre which attends high resolution, low dispersion comparators.
Also involved were a set of assumption about symmetry properties of actual samples and a "sibling" hypothesis. All of these are potentially troubling. We therefore have built a new comparator capable of resolving the diffraction fine-structure and exploiting it to gain higher precision
(better than 0.01 ppm) in the comparison exercise. We propose to eliminate the sibling hypothesis by direct use of the measured crystal in comparison with part of the y-ray diffraction crystal itself. Overall this step seems capable of refinement until its contribution to an overall error budget is at or below 0.01 ppm.
The y-ray diffraction angle measurements seem most robust. We have, however, in the meanwhile built a new instrument and refined our procedure in many ways. The new instrument was first of all needed to use a fixed
source as we must at ILL. Secondly, the instrument was made both more compact and versatile expecially getting a wider angular range and more rapid drives. Self calibration procedures have been made more efficient through the wider range and use of a compact autocollimator of novel design. Overall, our experience has been that self-calibration exercises indicate uncertainties near 0.03 ppm. With additional automation of
is being installed with operation anticipated during 1984.
An early aim in our program is to establish an improved value for the neutron's mass. This entails an accurate measurement of the 2.2 MeV y-transition corresponding to n + p + D. We propose to improve the y-energy presently best available from a Ge(Li) detector study with an uncertainty n e a r 3 p p m ( 4 ) t o a p p r e c i a b l y better than 1 ppm. The mass value is then obtained by combining this result with that available from high resolution mass spectroscopy of the mass doublet splitting between H2 and D :
M(n) = M(H) + E (2.2 MeV)
-
M(H2 - D) Ywhere E is understood to have been expressed on the nuclidic mass scale
Y -1
by multiplying its wavelength representation, A
,
by NAh/c.In the case that the wavelength measurement is improved to better than 1 ppm, the mass doublet uncertainty will dominate but this can also be improved.In a similar way, by extending both mass spectroscopy and wavelength determination to larger energy intervals, one has an opportunity,though a difficult exercise,to improve on present knowledge ofNAvc. There are several possible experimental approaches among which the 14~+n+15~+y(10.8MeV) appears promising. The mass doublet separation between 1 4 ~ ~ ~ 2 and 15
NH3 is already known to about 0.9 ppm and can likely be improved. The decay scheme for I5N* provides several groups of transitiors with Ey 'L 5 MeV which can be summed in various ways to give an over-determined estimate
- 1
of Ey (10.8 MeV). If one associates an effective A with the difference in binding energy between a neutron in D and a neutron in 1 5 ~ , then
N h/c = AMA A
where AM is the mass doublet separation.
SinceN v c is already known to better than 0.3 ppm indirectly, significant A improvement is a challenge to both spectroscopies. This challenge is not without interest since such a result would provide a robust, model-indepen- dent estimate of the fine-structure constant.
References
1. R.D. Deslattes, E.G. Kessler, W.C. Sauder and A. Henins,Annals of Physics
129,
378 (1980)2. This recommendation of the CCDM was accepted in October 1983
C3-46 JOURNAL DE PHYSIQUE
3 . P. Becker, P . S e y f r i e d and H. S i e g e r t , Z . Phys. B.
