Solar panels don't grow on trees. It takes energy to make a solar panel. Manufacturing solar panels has it's toll on the environment just like anything else. I often wonder if it takes more energy to make a solar panel than it will give out in its lifetime? Think of all the making, eating, pissing, shitting and rooting that's required to just get a panel on the roof, before it's even switched on. We live in a system of diminishing returns. No one ever considers the Total Energy Expenditure.
Now where did I put my scrumpy?
Lots of people take the total lifecycle energy usage of these products very seriously. The snag is their figures seldom agree very well. Its pretty clear that the energy produced by solar panels exceeds the energy consumed for manufacture, deployment and support, as long as the system is kept deployed for a long time. However, by some measures the ratio might only be a few times as much energy produced. The casual observer, noting that these panels are designed to operate for at least 25 years, might expect a much higher multiple.
Everyone lies in the energy business. There's too much money involved for people to stay honest. Take all figures not presented with lots of solid supporting material with a pinch of salt. If there is supporting material, be highly suspicious of it, and probe its origins.
That said, I think the original post is actually referring to dirty power as in distorted power waveforms. Most approvable grid tie inverters are required to produce a fairly clean sine wave. It doesn't need to be perfect. The mains power is rarely that clean. Even a simple incandescent lamp causes 10% to 20% THD in its current waveform, due to the rapid heating and cooling of the filament through each half-cycle of the mains.