I know that lcm(n) or Carmichael's totient function $\lambda(n)$ is now used instead of $\varphi(n)$ or Euler's totient function to generate the public key in RSA encryption? Why is this so? Are there any mathematical, computer scientific reasons behind this related to key security or efficiency?
Defining the public and private exponents as inverses modulo φ(N), as originally done for RSA, provides a sufficient (but not necessary) condition for the decryption rule to recover the plaintext from any ciphertext. Since Carmichael’s lambda function is, by definition, the smallest number m such that $a^m \equiv 1 \pmod N$, for any integer a that is relatively prime to N. Thus, it is sufficient to define the public and private exponents as inverses modulo any multiple of λ(N).