This is because an SSL certificate is issued for a particular domain name. If the certificate name doesn't match the visited domain, the browser will show an error.
One of the main functions of SSL is to prove to the user that they are really connecting to the site they requested, and not to an attacker masquerading as the end site. Without linking the domain name to the certificate this would not be possible.
It is conceivable that the browser certificate system could have been designed to include the IP address in the certificate, but this would make it difficult to use DNS load balancing or even to change hosting providers, as a new certificate would have to be issued each time this happened. If the certificate included just the IP address and not the domain, this would leave the user defenseless against DNS spoofing attacks. So the only way forward really was to use the domain alone.
As a matter of interest, it is possible to obtain an SSL certificate for an IP address - and as Google is their own certificate authority, they could issue themselves a certificate for 173.194.43.96 and thus make it possible to browse google securely by ip address, so long as they used SNI to serve up the correct certificate. It seems implausible that this would be worth the additional complexity however...
This is a nice introduction to SSL if you want to read more:
https://timnash.co.uk/guessing-ssl-questions/
与恶龙缠斗过久,自身亦成为恶龙；凝视深渊过久,深渊将回以凝视…
