Here are the results of the above benchmark running against three different Ruby VMs.
ruby 1.9.2p320 (2012-04-20 revision 35421) [x86_64-darwin11.4.0]
user system total real
single 0.170000 0.000000 0.170000 ( 0.171696)
double 0.170000 0.000000 0.170000 ( 0.171394)
ruby 1.9.3p194 (2012-04-20 revision 35410) [x86_64-darwin11.4.0]
user system total real
single 0.190000 0.000000 0.190000 ( 0.189292)
double 0.180000 0.000000 0.180000 ( 0.184996)
jruby 1.7.0.preview1 (ruby-1.9.3-p203) (2012-05-19 00c8c98) (Java HotSpot(TM) 64-Bit Server VM 1.6.0_35) [darwin-x86_64-java]
user system total real
single 0.420000 0.010000 0.430000 ( 0.221000)
double 0.240000 0.010000 0.250000 ( 0.118000)
As you can see, double quotes are actually faster in all three VMs. They are twice as fast in JRuby but only marginally faster in the two CRuby VMs. I don't have Rubinius on hand or I'd test that as well.
So, there you have it. Double quotes interpolate strings easily, support several nifty escape sequences like \n and \t, and for whatever reason they're faster than their counterpart. Maybe because there's 2X the quotes!
ruby 2.0.0p247