I forgot to mention that I had added the diode and resistor. Your right that it was shorting to ground. The circuit "works" with a 22k resistor and a normal bjt but it alters the charging time too much on the capacitor which throws off the duty cycle. I built it and saw it happen.
I measured 5v at the collector of Q2 when it switches on. My regular bjt had hfe(min) = 100. The transistor was rated for 0.8A, so Rc=Vcc/Ic which comes to Rc=9/0.8=11.25ohm. And Ib=Ic/hfe(min) according to a textbook I have, so Ib(min) to saturate transistor is 0.8/100= 8mA. 8mA is small but it still throws off the cycle.
When I use a darlington with hfe(min)=30k, rated at 0.4A Ic, I get Rc=9v/0.4A= 225ohm. Ib comes to 0.4/30k=13uA. 13uA is way less than 8mA. So I used a base resistor of Rb=Vc(from Q2)/Ib = 5/13uA = 384kohm. I use a value less than that because as the battery drains it won't be 5v at the Q2 collector anymore.
I've built this and it seems to work. Thanks for the help guys.