Neutron anisotropic evaporation and scission emission in fission

Experimental neutron distributions have been investigated in the spontaneous fission of Cf at IPHC in Strasbourg. The CORA experiment associating the CODIS twin ionisation chamber and the neutron multi-detector DEMON aimed to solve an long-standing problem in fission: the possible emission of scission neutrons and/or the presence of a dynamical anisotropy in the neutron evaporation by the moving fission fragments. A new method allowing to establish the dynamical anisotropy in an independent way is presented. The results obtained from a comparison with simulations based on GEANT4 are shown.


Introduction
It is commonly considered that during the fission process the major part of the neutrons are emitted isotropically by the fully accelerated fragments. But many works, experimental and theoretical, showed up discrepancies with such a purely isotropic emission. The question arises to find out the origin of this deviation, either due to some dynamical anisotropy in the centres of mass (CM) of the fission fragments (FF) or to neutron emission at scission. Already in 1962, H. Bowman et al. [1] made a detailed measurement of the neutrons emitted in spontaneous fission of 252 Cf. In the evaluation the authors had to introduce 10% of scission emission to explain their results. The possibility of some anisotropy in the CMs of the fragments was also suggested. B. Franklyn et al. [2], when they analysed K. Skarsvåg and K. Bergheim's data on thermal neutron induced fission of 235 U, had to assume up to 20% scission emission to reproduce the measured distributions. Even with

CORA experiment
The setup for the CORA experiment was composed of the angle-sensitive twin ionization chamber CODIS [7] for the detection of fission fragments and sixty modules of the DEMON [8] neutron multi-detector as shown in figure 3. CODIS provides the energies, mass ratios and emission angles of the FFs from the central 252 Cf source while DEMON gives the energies and angles of the correlated neutrons. Sixty DEMON modules have been used in the actual CORA setup.
In a five-month experiment, about 10 9 triple or higher coincidences between any FF and two or more associated neutrons have been collected.
The geometrical configuration used for DEMON, almost spherical around the CODIS chamber, allowed to study for the first time both processes, dynamical anisotropy and scission neutron emission, simultaneously in the same experiment.

CORA analysis
For the many parameters measured in the CORA experiment several experimental biases concerning particularly the neutron detection had to be considered carefully. The geometrical acceptance of DEMON was about 20% of 4π. A common neutron energy threshold was set at 0.9 MeV. The intrinsic efficiency of the DEMON modules and the fact that only the centres of the DEMON cells can be used to define the detection geometry had to be taken into account. Another important effect is the cross talk, when one emitted neutron interacts successively with two detectors, mainly in neighbouring detectors. In the actual geometry, the cross talk is quite low but, as the effect of interest is also very weak, it may impact the result.
Detailed simulations based on GEANT4 have been performed in order to consider the experimental biases but also to get a feeling about the extracted quantities and to check the relevance of the analysis.
Three correlations were addressed: ϕ nnF as defined above, the neutron-neutron relative angles θ nn and the angles between a light fragment (LF) and the neutrons, θ nLF .
The implementation of the simulations are described in detail in references [9,10] and preliminary results were presented in [11,12]. The characteristics of the FFs (mean velocities, mean temperatures) and the neutrons (multiplicities, variances, covariance) were taken from literature [13,14,15]. As to the neutrons, the angles have been generated randomly for an anisotropic emission in the CMs of the FFs. The dynamical anisotropy has  Figure 6 presents the resulting correlations ϕ nnF (left), θ nn (right-top) and θ nLF (rightbottom).

Results
The ϕ nnF distribution has been fitted following eq. 1. The fit result is shown in figure 6 (red curve). The deduced anisotropy ratio a 2 = 0.0028 ± 0.0015 is consistent with I. Guseva's prediction, a 2 = 0.0025 as indicated in figure 4, corresponding to an anisotropy parameter A nJ = 0.16. However the large reduced χ 2 value of the fit shows that beyond the statistical error there is a systematic error which has been estimated. This leads to a total error estimated to be of the order of 4 x 10 -3 .
This value was then fixed in the simulation to extract the scission emission component by comparing the experimental θ nn and θ nLF distributions (blue points) to the simulated ones (pink curves) by a χ 2 minimisation. The best fit is obtained for a scission emission probability ω sci = 8% which is also in agreement with the result of [15].
Only the angles θ nn > 53° are presented in figure 6. This restriction, applied also to the ϕ nnF distribution, was necessary to remove the impact of cross talk events to the analysis. Fig. 6. Summary of the experimental results after the three distributions ϕ nnF , θ nn and θ nLF have been corrected for the detector response by normalisation to uncorrelated events: -left: The resulting distribution has been fitted by eq. 1. The fit parameters are presented in the insert.
right: The normalised θ nn (top) and θ nLF (bottom) experimental distributions (blue points) have been compared to the corresponding simulated ones (pink curves) by a χ 2 minimisation leading to a scission emission probability ω sci = 8% for both distributions. The agreement between experiment and simulation is very good.
The slightly high threshold at 0.9 MeV of the measured neutron energies shouldn't influence the present results as its effect has been taken into account in the simulations. Concerning the anisotropy, according to reference [3], its effect increases drastically with neutron energy. Thus the major part of those neutrons have been detected in the CORA experiment. As for the scission neutrons, their distribution is not really known up to now. In our simulation a temperature of Tsci = 1.2 MeV adopted from [15] has been used in the Weisskopf formula. If this approach is realistic, also a reduced amount of those scission neutrons only have been missed in the experiment.

Conclusion
The CORA experiment constitutes a new approach to address the problem of ambiguities in the interpretation of the fission neutron angular distributions. CORA which associated CODIS and DEMON is the first experiment giving access simultaneously to both the dynamical anisotropy and the scission neutron emission.
It allowed to accede to the contributions of these two processes quantitatively. The values obtained up to now, A nJ = 0.16 and ω sci = 8%, are consistent with theoretical calculations performed by I. Guseva. Complementary work is needed to confirm the value obtained for the A nJ anisotropy parameter as its contribution is very weak and the error attached to the a 2 determination is still quite high. However, the present work demonstrates the relevance of the new approach.
This work was partially supported by the JINR-IN2P3 agreement for which the authors are very grateful.