Evidence against correlations between nuclear decay rates and Earth–Sun distance

Abstract

Unexplained periodic fluctuations in the decay rates of 32 Si and 226 Ra have been reported by groups at Brookhaven National Laboratory (32 Si), and at the Physikalisch-Technische-Bundesandstalt in Germany (226 Ra). We show from an analysis of the raw data in these experiments that the observed fluctuations are strongly correlated in time, not only with each other, but also with the distance between the Earth and the Sun. Some implications of these results are also discussed, including the suggestion that discrepancies in published half-life determinations for these and other nuclides may be attributable in part to differences in solar activity during the course of the various experiments, or to seasonal variations in fundamental constants. Following the discovery of radioactivity by Becquerel in 1896 [1] an intense effort was mounted to ascertain whether the decay rates of nuclides could be affected by external influences including temperature, pressure , chemical composition, concentration, and magnetic fields. By 1930, Rutherford, Chadwick, and Ellis [2, p. 167] concluded that "The rate of transformation of an element has been found to be a constant under all conditions." (For decays resulting from K-capture, or for beta-decays in strong ambient electromagnetic fields, the situation is slightly more complicated, since these decays are influenced by the electron wave functions which can be affected by external pressure or fields [3, 4, 5].) For 32 Si and 226 Ra, which decay by beta-and alpha-emission, respectively, fluctuations in the counting rates (in the absence of strong external electromagnetic fields) should thus be uncorrelated with any external time-dependent signal, as well as with each other. In what follows we show that neither of these expectations is realized in data we have analyzed for 32 Si and 226 Ra, thus suggesting that these decays are in fact being modulated by an external influence. Between 1982 and 1986, Alburger, et al. [6] measured the half-life of 32 Si at Brookhaven National Laboratory (BNL) via a direct measurement of the counting rate as a function of time. If N (t) denotes the number of surviving atoms starting from an initial population N 0 at t = 0, then the familiar exponential decay law, N (t) = N 0 e −λt , leads to ˙ N ≡ dN/dt = −λN 0 e −λt where λ = ln(2)/T 1/2. A plot of ln ˙ N (t) as a function of time is then a straight line whose slope is λ, which then gives the half-life T 1/2. At the time this experiment was initiated, the 32 Si half-life was believed to be in the range of 60 T 1/2 700 yr, and hence a multi-year counting experiment was needed to obtain a mea-sureable slope. As in other counting experiments, the counting rate for 32 Si was continually monitored in the same detector against a long-lived comparison standard, which in the BNL experiment was 36 Cl (T 1/2 =301,000 yr). Since the fractional change in the 36 Cl counting rate over the four year duration of the experiment was only O(10 −5), which was considerably smaller than the overall uncertainty of the final...

Power spectrum analyses of nuclear decay rates

“…This observation naturally raises the possibility that either the detectors or the decays in question are being affected in some fashion by an external influence such as seasonal temperature variations or solar radiation. The reported periodic variations in decay rates motivated other experiments and analyses, which both supported and questioned the original results [9,10,11,12]. Data obtained during the solar flare of 2006 December 13 exhibited a significant dip in the counting rate of 54 Mn nearly coincident in time with a solar flare, thus supporting the suggestion of a connection between nuclear decay rates and solar radiation [9].…”

“…Data obtained during the solar flare of 2006 December 13 exhibited a significant dip in the counting rate of 54 Mn nearly coincident in time with a solar flare, thus supporting the suggestion of a connection between nuclear decay rates and solar radiation [9]. On the other hand, in an analysis of data from several decay measurements, Norman et al [10] reported no evidence for a deviation from Eq. (1).…”

“…For a non-integer number of cycles, the expression is more cumbersome but still depends on the relative phase and ranges from -1 to 1. For phases on the order of those observed in the data the application of r or χ 2 ν , as is done by Norman et al [10], would result in a poor fit statistic when compared to an otherwise good periodic fit. We show this in detail in Ref.…”

“…[25], where we re-examine the 22 Na/ 44 Ti raw data of Ref. [10] which have been generously provided to us. Using the power spectrum analysis methods provided above, the 22 Na/ 44 Ti data yield a frequency of 1.09 ± 0.33 year −1 with more than a 97% confidence the peak is statistically significant and a corresponding phase of −25 degrees.…”

“…In addition to verifying the annual periodicity by three different analyses, we show that these phases provide an important tool in discriminating among possible explanations of the observed annual periodicities. As we discuss below, the new phase information has two important implications: phase can have a significant impact on the Pearson correlation coefficient, r, and chi-squared, χ 2 ν , values used by Norman et al [10], and properly incorporating phases can influence the conclusions to be drawn from their data.…”