# Does entropy increase with heat flow?

In an exam, I had a scenario where 2 bodies with different temperatures were put together and over time their temperatures mixed and eventually became uniform. My intuition tells me that as a result, the entropy of this system increased, however the exam asked me to explain why entropy remained the same. Could someone tell me whether the exam is correct, and if so, why?

Edit: The problem was posted on our website (with no official answer unfortunately), so here it is in full:

Consider an isolated system composed of two bodies at slightly different temperatures T1 and T2 (T1 = T2 + dT) thas have been put in contact. a) What does the second law of thermodynamics say about the direction of heat flow between them? (1pt) b) Explain how the entropy of both bodies change. Show that the total entropy of that system is constant. (6pts)

(there's a picture showing the 2 bodies and a barrier around them labeled "perfect insulation")

• Was the system a closed one or an open one ? – Vincent Fraticelli Jan 27 at 11:40
• Unless you are leaving something out, it seems wrong to me. Can you think of any scenario where the temperatures of the two original bodies would spontaneously (without any outside influence, return to their original levels? Perhaps you can provide more details on the original question. – Bob D Jan 27 at 13:21
• If the combined system you described is adiabatic, the entropy of the system increased. – Chet Miller Jan 27 at 14:21

However, the exam explicitly states that the temperature difference is infinitesimal. This isn't uncommon in thermodynamics thought experiments; for example, if we consider the cooling of a hot object surrounded by a large thermal reservoir (e.g., a cup of hot coffee in your kitchen), then it's routine to assume that the environment is isothermal even though we know that the thermal energy lost by the coffee must heat up the reservoir to some degree. It's just that the amount can be assumed negligible. In this way, we can obtain that that environment gains $$Q/T$$ entropy, for example, where $$Q$$ is the thermal energy lost by the relatively small object and $$T$$ is the approximately constant temperature of the large reservoir.