N.34 NuDat 2 data se­lec­tion

The gamma de­cay data of fig­ures 14.63 and 14.64 were re­trieved from NuDat 2, [[12]], Oc­to­ber-No­vem­ber 2011.

In the data se­lec­tion, tran­si­tions were ig­nored if any am­bi­gu­ity at all was in­di­cated for any of the pri­mary data. The pri­mary data were the ini­tial en­ergy level, the re­leased en­ergy in the tran­si­tion of in­ter­est, the half life, the ini­tial and fi­nal spins and par­i­ties, the mul­ti­pole type and or­der, the rel­a­tive in­ten­sity of the gamma tran­si­tion of in­ter­est, (see be­low for more), the mix­ing ra­tio for tran­si­tions of mixed mul­ti­pole type, the con­ver­sion co­ef­fi­cient, (see be­low for more), and the de­cay rate in Weis­skopf units. If there was a de­cay process other than gamma de­cay in­di­cated, the gamma de­cay per­cent­age had to be given with­out am­bi­gu­ity. In­di­ca­tions of am­bi­gu­ity in­cluded paren­the­ses, square brack­ets, in­equal­i­ties, tildes, ques­tion marks, mul­ti­ple val­ues, and more.

Note that the given data un­cer­tain­ties were ig­nored. That is a weak­ness of the data, but pre­sum­ably not re­ally im­por­tant in view of the very large de­vi­a­tions from the­ory. In­clud­ing un­cer­tain­ties would make pro­cess­ing much more com­pli­cated.

As an over­all check on the data, the com­puted tran­si­tion rate was com­pared to the one pro­vided in terms of Weis­skopf units by NuDat 2 it­self. If the dif­fer­ence was less than 5% and there were no other con­cerns, as dis­cussed be­low, the tran­si­tion was ac­cepted au­to­mat­i­cally. Tests were also per­formed on whether the ini­tial and fi­nal en­ergy lev­els matched the en­ergy re­lease, and on spin and par­ity con­ser­va­tion. These tests were mainly to guard against ty­pos in the data base and no vi­o­la­tions were ob­served.

If there were any con­cerns, the data were printed out. A de­ci­sion was then made man­u­ally on whether to ac­cept the tran­si­tion as a po­ten­tial can­di­date for plot­ting. If the com­puted tran­si­tion rate was sub­stan­tially, (more than roughly 15%), above the NuDat 2 value, the tran­si­tion was re­jected out of hand. If the com­puted tran­si­tion rate was be­low the NuDat 2 one, it was ex­am­ined whether the NuDat 2 value was self-ev­i­dently miss­ing its cor­rec­tion for other de­cay types, for the other gamma in­ten­si­ties, the mix­ing ra­tio, or the con­ver­sion co­ef­fi­cient. That was ob­served in a rel­a­tively small num­ber of cases, usu­ally for a miss­ing con­ver­sion co­ef­fi­cient. In all other cases, for a sub­stan­tial dif­fer­ence in de­cay rates, over about 15%, the tran­si­tion was re­jected.

Even if the com­puted de­cay rate matched the Weis­skopf one, var­i­ous de­cays were man­u­ally re­jected. In do­ing so, if the gamma in­ten­sity was not given, it was as­sumed to be 100% only if there was only one gamma de­cay out of the en­ergy level. If there were other gamma de­cays out of the same en­ergy level, their in­ten­si­ties were, based on man­ual ex­am­i­na­tion, al­lowed to be omit­ted (as­sumed to be zero), spec­i­fied by an up­per limit if small (as­sumed to be half the up­per limit), or spec­i­fied as ap­prox­i­mate if small. If the con­ver­sion co­ef­fi­cient was not given, it was man­u­ally al­lowed to be zero if the in­com­press­ible ball­park value was be­low about 10$\POW9,{-4}$. Some ini­tial en­ergy lev­els with mul­ti­ple gamma de­cays were man­u­ally re­jected if the tran­si­tion of in­ter­est had very low in­ten­sity and only one or two dig­its were given. Mixed tran­si­tions were man­u­ally ex­am­ined, but it was not con­sid­ered cause for re­jec­tion as long as a valid mix­ing ra­tio was given. If the mul­ti­pole level was higher than needed, that was also an­nounced, but it too was not taken to be a rea­son for re­jec­tion.

No, in the man­ual se­lec­tions, the au­thor did not se­lect the worst nu­clei to make physi­cists look bad.

The plot range from 30 to 3 000 keV (in the plots re­duced to 2 500 keV) en­ergy re­lease was di­vided into 70 seg­ments for which one sym­bol to plot each. Tran­si­tions to plot were se­lected by com­par­ing them to the se­lected tran­si­tions in the other seg­ments. The se­lec­tion was de­signed to achieve a broad cov­er­age of tran­si­tions. For plot seg­ments for which there was only one avail­able tran­si­tion, that tran­si­tion was im­me­di­ately se­lected. Then the pro­gram it­er­ated over the seg­ments with more than one po­ten­tial can­di­date for plot­ting. In each seg­ment the best can­di­date to plot was se­lected ac­cord­ing to the fol­low­ing cri­te­ria:

Can­di­dates that were more dis­tant in terms of $Z$ from the cur­rently se­lected can­di­di­dates in the other seg­ments re­ceived pri­or­ity. Dis­tance was here de­fined as the dis­tance from the clos­est se­lected nu­cleus of the other seg­ments. The in­ten­tion was to cover the en­tire range of atomic num­bers as well as pos­si­ble.
In case of a tie, nu­clei that were dif­fer­ent from the most other se­lected nu­clei in terms of ei­ther pro­ton or neu­tron odd/even­ness re­ceived pri­or­ity. The in­ten­tion was to in­clude all vari­a­tions of even/odd­ness.
In case of a tie, nu­clei that were sta­ble re­ceived pri­or­ity. That was in the hopes that data on sta­ble nu­clei might be bet­ter qual­ity.
In case of a tie, nu­clei that were more dif­fer­ent in $A$ from the al­ready se­lected nu­clei re­ceived pri­or­ity.
In case of a tie, a ran­dom choice was made be­tween the nu­clei in the tie.
Be­cause these cri­te­ria de­pend on the se­lec­tions in the other seg­ments, it­er­a­tion was needed. The it­er­a­tions were ter­mi­nated if there were no more changes in the se­lected can­di­dates.

The data on the se­lected nu­clei, in­clud­ing log files, are avail­able in the web ver­sion of this doc­u­ment. If you have sug­ges­tions on how the data could be im­proved, let me know.