H3PW12O40 is a Keggin heteropolyacid (HPA) that nowadays attracts more and more attention as catalyst for the gas phase dehydration of methanol to dimethylether. Here, we investigate its performance in the latter reaction at 250-300 degrees C, namely under coking conditions. Actually, we address the question whether the elsewhere elucidated procedure allowing to activate the bulk of the H3PW12O40 solid is beneficial or not under conditions leading to the rapid coking/deactivation of H3PW12O40'S surface (through which the bulk gets accessed). Unprecedented, our results show that pre-activating the bulk is actually crucial in the latter conditions, as it allows drastically increasing H3PW12O40's lifetime (up to 5 times at 300 degrees C). Indeed, once activated, the bulk stays accessible beyond the deactivation of the surface. However, with its 24 times more protons than provided by the surface, and with the conversion over the latter protons being diffusion-limited, it then contains continuously enough fresh protons being able to take over the job of the other bulk protons that are deactivated (in the sense that they are capped within coke molecules,) until having all been solicited/deactivated. As a consequence, the catalyst bed gets fully exploited until the last of its available bulk protons has contributed to the reaction, instead of getting only scarcely exploited as it is the case when the reaction is purely surface-type. So, the results of this work bring considerable insight on how to use a Keggin HPA-based catalyst in a more sustainable way. Furthermore, they allow understanding the performance obtained upon submitting a Keggin HPA to a temperature programmed reaction of methanol from 25 to 300 degrees C. Indeed, throughout such a reaction, the deactivation profile (e.g. drastic conversion drops vs slow ones) is governed by the nature of the protons (surface vs bulk-type) that are getting deactivated. (C) 2017 Elsevier B.V. All rights reserved.