This week's Nature (subscription! possible workaround using Bugmenot ) has a brief note about why liquid water is considered so essential for life. I have been wondering about this for some time, and have only read a few reviews about alternative life chemistries.
The article starts with Steve Benner, who actually wrote the review I had read last December. Benner's position is that liquid water is literally a two-edged sword. It's an excellent solvent and is chemically both an electron donor and acceptor-- but these flexibilities also mean it will chemically attack just about any information storage molecule you care to think of. (My RNA preps can testify to that!)
The interesting part comes with a third property of water- it is highly polar (the hydrogens are weakly positive, and the oxygen weakly negative) which allows a much weaker class of bond called hydrogen bonding to exist. They get made and get broken, over time and energy scales we can live with. These hyrogen bonds, for example, zip DNA molecules together, and can be readily undone when it's time to copy them. When you get out of the temperature domain of liquid water, you'd have to rely on a new class of bonds-- or operate on a timescale wholly different than we think of as alive.
But we only have the single example of life on earth, and maybe our thinking is limited by only studying life which has clearly evolved with water in mind. Benner's favorite alternatives are cold domains, with liquid ammonia or nitrogen (which have subsets of water's properties). Cold means **slow**, but probably information storage is greatly simplified. Would we recognize something with a lifespan of millenia?
The Benner Homepage
Astrobiology encyclopedia for liquid ammonia life
An earlier Nature pdf file on extreme life.
The physical chemistry of liquid water