where we ignore the fact that 22021 is composite (22021 = 192 ⋅ 61).
A Descartes number is defined as an odd number n = m ⋅ p where m and p are coprime and 2n = σ(m) ⋅ (p + 1), whence p is taken as a 'spoof' prime. The example given is the only one currently known.
If m is an odd almost perfect number,[1] that is, σ(m) = 2m − 1 and 2m − 1 is taken as a 'spoof' prime, then n = m ⋅ (2m − 1) is a Descartes number, since σ(n) = σ(m ⋅ (2m − 1)) = σ(m) ⋅ 2m = (2m − 1) ⋅ 2m = 2n. If 2m − 1 were prime, n would be an odd perfect number.
If n is a cube-free Descartes number not divisible by 3, then n has over one million distinct prime divisors.[2] If is a Descartes number other than Descartes' example, with spoof-prime factor , then .[3]
John Voight generalized Descartes numbers to allow negative bases. He found the example .[4] Subsequent work by a group at Brigham Young University found more examples similar to Voight's example,[4] and also allowed a new class of spoofs where one is allowed to also not notice that a prime is the same as another prime in the factorization.[5] A generalization of Descartes numbers to multiperfect numbers has also been constructed. (Tóth (2025)).
^Currently, the only known almost perfect numbers are the non-negative powers of 2, whence the only known odd almost perfect number is 20 = 1.
^Banks, William D.; Güloğlu, Ahmet M.; Nevans, C. Wesley; Saidak, Filip (2008), "Descartes numbers", Anatomy of integers. Based on the CRM workshop, Montreal, Canada, March 13–17, 2006, Providence, RI: American Mathematical Society (AMS), pp. 167–173, ISBN978-0-8218-4406-9, Zbl1186.11004, retrieved 2024-05-13