In refrigerator, heat energy got extracted from cold sink and dumped into hot source. This is not spontaneous so work must be done. I came across the following while reading a textbook:
For refrigeration, if no work is done, ΔS= qh/Th+qc/Tc;
-qh=qc, implying ΔS= qh/Th-qh/Tc =qh(1/Th-1/Tc)
qh>0, 1/Th-1/Tc<0, so ΔS<0 so not spontaneous. Therefore, we need to add heat energy to the stream of heat energy flowing from the cold sink to hot source, so that the entropy increase at hot source 'covers up' the decrease in entropy at cold sink, and the total entropy increases.
It does indeed make sense if we look at the value of entropy, but how can this be feasible physically? You want to move heat energy from cold sink to hot source. In terms of molecular motions, the molecules at hot source always vibrate more vigorously and give energy to molecules at cold sink spontaneously. To achieve your purpose, you literally make the molecules at hot source vibrate faster?
If you still do not get what I am asking, imagine you have a flask of water and you want it to flow from A to B. If you punch a hole at B to remove some water at B, the water will flow to B, but adding water to B definitely will not cause more water to flow to B. Now replace water with heat energy, A with cold sink and B with hot source, I think adding heat energy to hot source should not cause cooling of cold sink.
What is my misconception here?