Results and discussion.

In fig. 2 we show the matrix elements of the axial charge for the relativistic and non relativistic cases of eqs. (11) and (14) respectively as a function of q. We show the results for the #math42#1p1/2→2s1/2 and #math43#1p3/2→1d3/2 transitions on 16O. One observes that the strength of the latter transition is about a factor 2 larger than for the first one. In both cases the relativistic calculation of the matrix element yields larger values than the non-relativistic one derived from the equivalent non relativistic potential. We also observe that the matrix elements are weakly dependent on the momentum transfer q up to values of #math44#q #tex2html_wrap_inline1382# 100 MeV/c. In fig. 3 we show the ratio of the relativistic versus non relativistic matrix elements as a function of q for the two transitions in 16O. The values of the ratios at q = 0, relevant to β decay, are about 1.33 and 1.20 for the #math45#1p1/2→2s1/2 and #math46#1p3/2→1d3/2 transitions, respectively.

The strength of #math47#Σs(r ) from eq. (6) at r = 0 is in our case #math48#Σs #tex2html_wrap_inline1392# -384 MeV, close to the value typical for nuclear matter at saturation ( #math49#Σs #tex2html_wrap_inline1394# -400 MeV) which we have considered in our estimates of section 2. In the perturbative approach of section 2 we would have obtained a ratio of 1.41 whereas the non perturbative nuclear matter estimate would even yield a ratio of 1.69 for this value of #math50#Σs. We can see that the calculations performed directly for the finite nuclei yields results which are significantly smaller than those estimates from nuclear matter. The reason for this difference is the fact that the calculation of matrix elements for finite nuclei requires a radial integration which is dominated by the integrand at the surface. This is due to the fact that the integrand contains a product of wavefunctions for a particle- and a hole-state. The nuclear density at the surface, however, is smaller than for r = 0 or the saturation density of nuclear matter. Consequently also the relativistic effects due scalar potential Σs, leading to an enhacement of the small component of the Dirac spinor, will be smaller at these relevant densities than at the center of the nucleus or at the saturation density of nuclear matter. Similar, although a bit smaller, reductions with respect to the nuclear matter approach were also found in the finite nuclei perturbative approach of [#8##1###], though the results were found to be sensitive to the short range correlations assumed. Here short range correlations are incorporated in the problem in a selfconsistent way.

From these considerations we can also understand that the relativistic renormalization of the axial charge operator in the case of the #math51#1p3/2→1d3/2 transition is smaller than the one in the #math52#1p1/2→2s1/2 case. The smaller renormalization in the case of the #math53#1p3/2→1d3/2 transition can be interpreted in terms of the centrifugal barrier which pushes the d state more to the surface of the nucleus where the potential #math54#Σs is weaker. Furthermore we observe a slight increase of the renormalization as a function of q. At a larger momentum transfer one tends to probe more the higher densities in the interior of the nucleus.

These results are confirmed by our calculations for the nucleus 40Ca. In fig. 4 we show the relativistic and non relativistic matrix elements for the #math55#1d3/2→2p3/2 transition in 40Ca and in fig. 5 the ratio of the relativistic to non relativistic matrix elements. The ratio is of the order of 1.23, rather independent on the momentum transfer. This result for the renormalization of the axial charge is very similar but slightly larger than the ratio obtained for the #math56#1p3/2→1d3/2 in 16O. Once again the centrifugal barrier is responsible for a reduced renormalization compared to the expectations of nuclear matter approach.