Dear Sudip Banerjee thanks for the interesting question. Please see the following closely related discussion which could provide you with some aditional answers to your question. The link is entitled
Can I cite an accepted, pre-publication paper without a DOI and the citations get counted correctly?
My personal feeling is: at any time during the process of the publication of a scientific manuscript there should be a DOI (nowadays!).
Some journals use a simple procedure. Each submission gets a specific DOI consisting of the Journal-code, the year and the submission number. This DOI is defined at the moment of submission.
For example:
DOI: https://doi.org/10.1515/zkri-2020-0072
DOIs of rejected manuscripts are lost using such a procedure, but numbers at the end of the DOI have never been a rare thing as far as I know.