The paper studies a special family of Carmichael numbers, which are composite numbers that mimic primes in a specific arithmetic test (Fermat's primality test). The numbers in question have the form two raised to some power, times a prime raised to an odd power, plus one. The paper is written as a tribute to mathematician Florian Luca, who proved shortly before his death that for any fixed odd exponent greater than four, only finitely many primes can appear in such Carmichael numbers. The authors build on his work by making his result more concrete and extending it in new directions.
The first major contribution is turning Luca's existence proof into explicit numerical bounds. Where Luca showed finiteness without saying how large things could get, the authors now give actual inequalities bounding how large the exponents and the numbers themselves can be. They also bound how many prime factors such a Carmichael number can have. One practical consequence is a count of how many such Carmichael numbers can exist below a given size. The second contribution is a generalization: Luca's key technical step involved powers of two, but the authors show the same logic works for any even base, broadening the scope of the theory considerably.
The third contribution is computational. The authors searched exhaustively and found that the only Carmichael number of this form up to one million trillion (ten to the eighteenth power) with an odd exponent between three and thirty-seven is 1729, the famous Hardy-Ramanujan taxicab number. No examples with odd exponent between five and thirty-seven were found even when the search was extended to ten to the twenty-first power. The rarity of these numbers in practice lines up well with the theoretical prediction that only finitely many should exist for each fixed odd exponent.