It seems you want to "store" energy as light by keeping it bouncing around inside a chamber.
No, that's not going to work, at least not for more than a few 10s of nanoseconds for a chamber the size of a "large tank".
Typical mirrors reflect maybe 90% of the light. Let's say you have really great mirrors that reflect 99% of the light. Each time light bounces off such a mirror it is attenuated to 0.99 of its incident energy.
That may sound good. Anything in nature being 99% efficient is usually very good. However, what you are missing is how often this factor of 0.99 will be applied. Light propagates thru vacuum (air is close enough) at 300 Mm/s. Let's say your chamber is 3 m across. That means light will hit a mirror 100 M times per second, or every 10 ns.
The above means that the light intensity is multiplied by 0.99 every 10 ns. In 100 ns the intensity will be 0.9910 = 90%. After 1 µs 0.99100 = 37%. After 10 µs 0.991000 = 0.004%, or less 1/23,000 of the original energy.
It should be obvious that the above is essentially an exponential decay. We can therefore find the effective time constant or half life to make it easy to visualize. Since ½ is about 0.9969, the half life is only 690 ns. Similarly, we can compute that the light is attenuated by a factor of 1000 every 6.9 µs. That means you're down to 1 millionth in twice that time, or about 14 µs.
And no, impossibly good mirrors isn't going to fix this, only lengthen the decay some. Let's say your mirrors (and the intervening light path) only attenuate by 0.9999 each bounce. Congratulations! You've extended the half life to 69 µs, and the 1/10 decay time to 230 µs. You're down to 1/1000 the original energy in 690 µs, and 1 millionth in 1.4 ms. That's still much less than the blink of an eye.