A graphite cavity chamber for absolute exposure measurements of 6oCo gamma rays is presented. The chamber is designed to allow the determination of the exposure to rely as far as possible on a theoretical analysis previously given by Allisg. As an extension of this analysis, the problem of the influence of the cavity thickness is treated. The values to be used for the physical constants involved and the correction factors are discussed as well as the related uncertainties.The chamber dealt with in this paper is currently used as the cobalt exposure standard of the Bureau International des Poids et Mesures (BIPMA). In its design we have been guided by the desire to determine exposure from experimental results by a calculation relying on a theoretical basis and to reduce thereby the amount of purely empirical corrections. * To Andre Allissy on the occasion of his fiftieth birthday. ** This work was performed in the gamma-ray laboratory of the Ionizing Radiation Section of the Bureau International des Poids et Mesures, Pavillon de Breteuil, 92310 S h e s , France (present address of the authors). 12 = zm-21 for the intermediate region, and 13= (za-z1)/2, which is half of the total chamber e 2 thickness, for the side wall region.

An experimental study of the saturation conditions of parallel-plate free-air ionization chamhers is presented. The results may be explained by assuming that two different types of ion recombination (initial and volume recombination) take place simultaneously. Their relative values can be calculated from the results, and are found to agree with theories proposed hy Boag, by Scott and Greening, and by Kara-JIichailova and Lea. It was found that for the saturation current to be determined by extrapolation, the collecting field strength should he high enough for volume recombination t o be negligible compared with initial recombination. The ext,rapcllation is of the reciprocal of ionization current against the reciprocal of collecting field strength.

Two experimental determinations of W, the average energy required to produce an ion pair in dry air, for electrons arising from 6oCo radiation, are presented, One is based on calorimetric and ionometric measurements of absorbed dose in graphite, the other on activity and exposure measurements. The values obtained by using the most recent data for the electron stopping powers are W / e = 33.96 f 0.08 and 33.8 I f 0.14 J C", respectively. The good agreement with the W values from other authors whose methods do not involve stopping powers gives support to the new stopping-power values.

Comparisons forgamma rays between the exposure standards of the Bureau International des Poids et Mesures (BIPM), the National Bureau of Standards (NBS) and the Physikalisch-Technische Bundesanstalt (PTB) are reported. For measuremente made at about one meter from the source, and if the same physical constants are used, the differences are 0.26% between the BIPM and NBS standards, and a t most 0.4% between the BIPM standard and the PTB standard chambers. The differences are consistent with the estimated systematic uncertainties (= 0.5 %).The difference between the NBS standard and the PTB chamber measurements, as determined through the ratios of each to the BIPM chamber, ranges from 0.17 to 0.66%. The difference in exposure standards for NBS and PTB actually ranges from 0.48 to 0.97%, since different stopping power corrections are used by these two laboratories.An indirect comparison of the BIPM and NBS standards indicates that the difference of 0.26% will increase to 0.55% a t a distance of two meters. RBsumBOn relate des comparaisons entre Btalons de mesure de l'exposition dans un faisceau de rayons gamma du 6OCo: ceux du Bureau International des Poids et Mesures (BIPM), du National Bureau of Standards (NBS) et de la Physikalisch-Technische Bundesanstalt (PTB). Pour les mesures effectuBes B un metre environ de la source et si Yon utilise les m6mes constantes physiques, les Bcarta sont 0,26y0 entre les &talons du BIPM et du NBS, et au plus 0,4 % entre 1'Btalon du BIPM et les chambres Btalons de la PTB. Ces diffhrences sont compatibles avec les incertitudes systBmatiques estimbes (= 0,5 %).Les diffbrences entre les mesures de 1'6talon du NBS et des chambres de la PTB, telles qu'on peut les dBduire de leurs rapports respectifs la chambre du BIPM, sont comprises entre 0,17 et 0,66%. Les d8Brences entre lea Btalons #exposition du NBS et de la PTB sont en rBalit6 comprises entre 0,48 et 0,97 % puisque ces deux laboratoires utilisent des corrections de pouvoir de ralentissement d8Brentes.Une comparaison indirecte des Btalons du BIPM et du NBS montre que la diffkrence 0,26% deviendrait 0,55% B unc distance de deux mhtres.

No abstract