# Why does heat flow from a region of higher temperature to one of lower temperature?

I have recently started thermodynamics and am unclear about reason behind the direction of heat flow, which is usually stated as a given fact.

From the zeroth law of thermodynamics, originally, i thought temperature was defined as a physical property that

                     1. at thermal equilibrium, temperatures are equal
2. Heat always flows from higher to lower temperature


and so the direction of heat flow was just a matter of definition.

However, recently, I came across an argument based on the 2nd law of thermodynamics:

Heat flows from a system of higher temperature to one of lower temperature as the decrease in entropy of the hotter system due to a decrease, du, in internal energy is less than the increase in entropy of the colder system due to an increase, du, in internal energy, so the entropy of the universe increases.

So, is the direction of heat flow because of definition or the 2nd law of thermodynamics?

• @AStudent4ever But that’s a microscopic analysis. Commented Mar 20, 2021 at 12:04
• I’m trying to understand heat and temperature from a macroscopic point of view. Assuming that the 0th law is used to define temperature, I don’t see any mention of particles or kinetic energy, so I don’t think this argument is useful Commented Mar 21, 2021 at 2:01

and so the direction of heat flow was just a matter of definition.

It is not just a matter of definition. It's a matter of observation. Heat is never observed to flow naturally or spontaneously from a low temperature substance to a high temperature substance. For that to happen, work must be done (as, for example, in the case of refrigerators and heat pumps.)

Before the second law was developed we had the first law. That law is a statement of conservation of energy. If heat heat were to flow from a cold to hot object, the heat lost by the cold object would equal the heat gained by the hot object and the first law would be satisfied. But since this never happens, a new law was needed that made such a flow of heat impossible .

Although the wording of the "argument" is rather awkward, at least to me, it is true that when heat transfers from the hotter system there is a decrease in both the entropy and internal energy (dU) of the hot system and there is an increase in both the entropy and internal energy of the cold system. The important distinction is internal energy is conserved but entropy is not, unless the heat transfer is "reversible".

The decrease in internal energy of the hot system exactly equals the increase in internal energy of the cold system, per the first law, for a total change in internal energy of zero. On the other hand, for any finite difference in temperature between the hot and cold system, the decrease in entropy of the hot system is less than the increase in entropy of the cold system, for a total entropy change greater than zero.

If the hot and cold systems are thermal reservoirs (constant temperature sources and receivers of heat), the entropy change of the hot system is

$$\Delta S_{H}=-\frac{Q}{T_H}$$

The entropy change of the cold system is

$$\Delta S_{C}=+\frac{Q}{T_L}$$

The total entropy change is

$$\Delta S_{tot}=+\frac{Q}{T_L}-\frac{Q}{T_H}$$

You will note that for all $$T_{H}>T_{L}$$, $$\Delta S_{tot}>0$$.

So, is the direction of heat flow because of definition or the 2nd law of thermodynamics?

No. The natural direction of heat flow from hot to cold is an observable fact of nature. The second does not dictate nature. The second law reflects nature. The second law says that

$$\Delta S_{tot}=\Delta S_{sys}+\Delta S_{sur}\ge0$$

Where the equal sign applies to a reversible transfer of heat, i.e., when the difference between $$T_H$$ and $$T_L$$ approaches zero.

If heat flowed from cold to hot, the signs for the entropy changes in the first two equations above would be reversed, resulting in $$\Delta S_{tot}<0$$ in violation of the second law.

Hope this helps.

• "The natural direction of heat flow from hot to cold is an observable fact of nature." Wouldnt this statement be circular if you're defining the "hotter" substance as the one which loses heat i.e if you re defining temperature in terms of heat flow. I think this boils down to whether temperature is defined in terms of entropy (which would make the first explanation wrong), or whether it is defined in terms of heat and thus entropy is defined in terms of temperature (making the second one wrong). Im not exactly sure which of this is true though Commented Mar 19, 2021 at 22:28
• @OVERWOOTCH I am not "re defining temperature in terms of heat flow". Temperature is a thermodynamic property, heat flow is not a property (it is mechanism for transferring energy, the other being work.). I am saying that heat is energy transfer due solely to temperature difference and that it is observed in nature that the transfer occurs naturally only from high to low temperature. Commented Mar 23, 2021 at 20:36

The zeroth law basically states that systems being at equilibrium with each other is a transitive property. if A is at equilibrium with B, and B is at equilibrium with C, then A is at equilibrium with C. From this we can map equivalence sets of systems at equilibrium onto the real numbers. If you do it in a particular way, you get "temperature."

However, this says nothing about how heat transfers between systems tagged with a larger temperature and systems tagged with a smaller temperature. That gets done with the second law.

And also, its worth remembering that when we see a "zeroth law," it usually means that the first few laws (like the 1st, 2nd, and 3rd) came first, and only later did we realize that we needed something more fundamental.

• Ok but when establishing scales using the 0th law, isnt deciding which object to assign a higher temperature simply a matter of choice? Wouldn't that work just as well? Sure, that would ruin the intuitive relationship with avg kinetic energy but that hasn't been worked out yet. (I'm trying to build up on thermodynamics from the 0th law, even though it was established much later) Commented Mar 19, 2021 at 22:38
• Correct. The mapping from equivalence sets to real numbers is a matter of choice. I don't believe you can derive the one-and-only concept of temperature from the 0th law. The 0th law is too specific.. But you can use it to say that if temperature measurement system measures A at 10 and B at 10, and then a second temperatur measurement system measures A at 25, then B must be at 25 as well Commented Mar 19, 2021 at 22:39

Suppose we mix two gases that initially have different temperatures. Higher temperature gas molecules move faster than lower temperature gas molecules. Collisions occur between the molecules of one gas and the other. Part of the energy from the faster molecules passes to the slower molecules. Gradually the temperature of both gases equalize. We have called Entropy to the natural tendency of the spread of energy , and we have given it the rank of law, it is the 2nd law of thermodynamics. Each time the energy goes from a smaller to a bigger space ( a bigger amount of matter or the same matter in a bigger space), the entropy increases. The quantification of this dispersion is the entropy value. Clausius expressed the 2nd law by saying that heat always goes from high to low temperature. However, Boltzman introduced a probabilistic characteristic to the 2nd law, he came to tell us that if the heat goes from high to low temperature it is not because there is a universal law that orders it so but because there is an overwhelmingly greater probability that it is so make it the opposite. Nature has a tendency towards equilibrium, when the balance is reached, the energy is no longer useful, that is why the Carnot machine needs both a hot and a cold focus, which is another expression of the 2nd law. The transfer of energy through heat flowing from hot to cold matter causes the useful energy to become energy from the particles in the environment from where nothing can extract it, useless energy. Every time this happens the entropy of the universe increases. When the universe reaches thermodynamic equilibrium it will have reached its maximum entropy and nothing will happen in it anymore. Obviously it will happen in a long long time.

Heat is, among others, kinetic energy of the disordered microscopic motion of atoms and molecules. You probably know that kinetic energy is exchanged during elastic collisions between two masses (think of pool billiards). It is the same for atoms and molecules: since there are so many of them, they are constantly and frequently bumping into each other, exchanging kinetic energy. A hot object is characterized by fast moving atoms/molecules, a cold object is characterized by slow moving atoms/molecules.

Now think about what happens if you put something hot close to something cold. The molecules inside the hot object will bump into each other just like before and nothing much will change there. The same inside the cold object. But at the interface between the hot object and the cold object, fast molecules from the hot object will collide with slow molecules from the cold object. To the effect that the formerly slow molecules become faster and the formerly fast molecules become slower. Of course, this also happens in reverse, which results in averaging out the differences between both objects.

Since the molecules near the interface stay in contact with molecules from the insides, this kinetic energy transfer progresses right towards the cores of the respective objects. Therefore, heat flows from the hot to the cold.

Since these motions are disordered, you cannot avoid that this is happening, just because you do not know what is happening in detail. You can't just turn one object by an angle of 10 degrees and hope that there will be less collisions. Nor can you put your finger on an individual molecule and keep it from bumping into another one. No matter what you do, disorder is increasing. Before, there was a tiny little remaining amount of order in that you knew that one object's molecules were faster than the molecules in the other object. Afterwards, both objects eventually have the same temperature, which means order has decreased (you know less about the molecules than before). This is the essence of the 2nd law of thermodynamics: disorder (aka entropy, aka negative information) will always increase without your consent.

• "Heat is very basically just kinetic energy of the disordered microscopic motion of atoms and molecules". That is not correct. Heat is energy transfer due to temperature difference. It is the transfer of kinetic energy, but it is not the kinetic energy itself. That is properly called the kinetic energy component of the total internal energy of a substance. Commented Mar 19, 2021 at 22:08
• I know, that is why I wrote "very basically". I could have restricted my statements to ideal gasses, then it would be correct. I just wanted something simple to tell to the OP, because he apparently has a lack of visual ideas about thermodynamics. I'll change it to something more neutral. Commented Mar 19, 2021 at 22:22
• Restricting it to ideal gases still doesn’t make it correct. Take your collision example. The collision is the process by which kinetic energy is transferred between the balls but the collision itself is not kinetic energy. It’s the same for heat. It’s a process for transferring energy not the energy itself Commented Mar 21, 2021 at 9:42
• @BobD: I am afraid, it does make it correct. The ideal gas has no way of storing potential energy, that is why it is called ideal. Hence, the only way it can store internal energy is by its kinetic energy. Nowhere did I say that a collision is kinetic energy, of course it is a process. So why do you think you need to correct me? Don't you think this discussion is already a little pedantic, given the beginner's level of the question? Commented Mar 21, 2021 at 12:25
• But to settle this, you are in fact correct, that the term "heat" is usually used in a sense of an internal energy difference that is not macroscopic work and that is exchanged in a thermodynamic process. In case there is potential energy in the system due to interactions, my original statement is manifestly wrong. Commented Mar 21, 2021 at 12:30