In Blundell's Concepts in Thermal Physics, page $117$, the author defines an adiabatic expansion as follows,

The word adiathermal means ‘without flow of heat’. A system bounded by adiathermal walls is said to be thermally isolated. Any work done on such a system produces an adiathermal change. We define a change to be adiabatic if it is both adiathermal and reversible.

On the other hand, this Wikipedia article clearly specifies that in order for an adiabatic process to happen, the expansion/compression must happen relatively fast. But Blundell's definition for an adiabatic process incorporates the condition of reversibility, which requires the process to happen infinitly slowly. The wiki article didn't add the reversibilty condition in its definition.

Is Blundell's definition incorrect?


3 Answers 3


It's important to distinguish between several things:

  1. A process which does not change the entropy
  2. A process which involves no transfer of heat
  3. A reversible process

Blundell is correct in that (2) and (3) together imply (1). However, conventions differ on whether (2) is called adiabatic, or (1) is called adiabatic. When you're doing more sophisticated thermodynamics, it's more useful to call (1) adiabatic, because the entropy is so important. However, in high school you'll often see the definition (2) used, mostly because irreversible processes are barely mentioned at all, so (1) and (2) end up being equivalent. It goes without saying that neither definition is "wrong", it's just a different way of using words.

Under definition (1), statements that adiabatic processes are "slow" and "fast" are simultaneously correct. They have to be much faster than isothermal processes, as otherwise heat transfer would become significant. But they also have to be much slower than processes like free expansion, which would be irreversible. It would probably be better to say that adiabatic processes are neither very slow nor very fast.


There are a few points to address here.

First, I would hesitate to say that a definition can be "incorrect" in the first place. Definitions are, in a sense, conventions, especially when it comes to terminology. A definition can certainly be unclear, or self-contradictory, or inconsistent with later usage, and in those cases you could certainly say that a definition is poorly-chosen, but the fact that what Author A means by a term is not the same thing as what other authors mean does not indicate that Author A's definition is incorrect. As long as Author A explicitly clarifies what is meant by the terms they use in the works they use, and as long as Author A uses their terms in a self-consistent manner, there is nothing incorrect about their work, and the only thing that can be said is: "This author uses a different definition for these terms than the majority of other authors."

You can see a similar situation in many electromagnetism textbooks: some authors define the magnetic field to be the vector field $\mathbf{B}$, while other authors (usually older textbooks, in this case) prefer to define the magnetic field as the vector field $\mathbf{H}$. Neither convention is incorrect, and as long as you pay attention to the definitions outlined in the work you're reading, you'll usually avoid confusion.

Second, adiabatic processes do not have to be "relatively fast", or at the very least, the meaning of "relatively fast" can encompass processes that are usually considered very slow. The defining condition for (the usual definition of) an adiabatic process (i.e. what Blundell calls an "adiathermal process") is that no heat enters or leaves the system. If you have a system that's totally isolated from its environment (like a container with ideal-insulator walls, or the entire universe), then an adiabatic process can occur as slowly as you like. In non-ideal situations, this translates to the following rule: A process is well-approximated as adiabatic if the time taken for the process to occur is small compared to the rate of heat transfer to the surroundings. If your container is well-insulated, then slow processes can be adiabatic too.

Blundell's definition of an "adiabatic" process is equivalent to the usual definition of an isentropic process, which is a reversible adiabatic process (and therefore one that preserves entropy).


The word adiathermal means ‘without flow of heat’. A system bounded by adiathermal walls is said to be thermally isolated. Any work done on such a system produces an adiathermal change. We define a change to be adiabatic if it is both adiathermal and reversible.

I disagree with the statement that a change is an adiabatic change if it is both adiathermal and reversible in that it implies the change has to be reversible to be adiabatic. An adiabatic process can be reversible or irreversible. If the process is both adiathermal and reversible, it is then called an isentropic process, meaning it is a constant entropy process.

What's more the use of adiathermal walls is a sufficient but not necessary condition for a process to be adiabatic. As Wikipedia points out a process that is carried out so fast that there is no time for heat to transfer is also an adiabatic process. However, such a process is inherently an irreversible adiabatic process because a reversible process has to be carried out extremely slowly (quasi-statically).

Bottom line: Blundell is misleading in implying that an adiabatic process has to be both adiathermic and reversible. It can also be irreversible. It is also misleading in that it suggests adiathermal walls are required for a process to be adiabatic. A process can be carried out without adiathermal walls so fast that it is in effect an adiabatic process.

Hope this helps.

  • $\begingroup$ Physics language has changed over the last 60 years in this area, and physics language does not always agree with other disciplines such as engineering. It's ok to define terms as Blundells do as long as you make clear what your definitions are (and it is helpful to note that other discussions may be using some terms differently). $\endgroup$ Oct 1, 2022 at 11:27
  • 1
    $\begingroup$ @AndrewSteane In my 30 years experience with standards writing committees one of the most difficult things was getting agreement on definitions! $\endgroup$
    – Bob D
    Oct 1, 2022 at 16:18
  • $\begingroup$ Yes! My own take on "adiabatic" is that although it was for many years said to be defined as "adiathermal" and strictly that, nevertheless people went ahead and used it in practice to mean "isentropic". For example, $\kappa_S$ was called by the name "adiabatic compressibility" and lines of constant entropy were called "adiabats". Then in quantum physics "adiabatic" means slow change in system parameters such that energy level populations don't change. In other words, isentropic again. In view of all this, I judge that the ship of "adiabatic means isentropic" has sailed, in physics at least. $\endgroup$ Oct 1, 2022 at 16:38
  • $\begingroup$ ... I think that way of defining terms is ok, as long as it is understood that we mean reversible as well (i.e. not a case where an irreversible process generated entropy which immediately left the system as heat). $\endgroup$ Oct 1, 2022 at 16:44
  • $\begingroup$ @AndrewSteane I agree. I also hope we can both agree that isentropic means there is no entropy transfer (which is satisfied as long as there is no heat transfer, whether because of thermal isolation or because of the speed of the process) and there is no entropy generation (which requires the process be reversible, which a speedy process rules out). $\endgroup$
    – Bob D
    Oct 1, 2022 at 17:28

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.