From Passive to Programmable: Reading Notes on Weilai Yu’s REI Viewpoint
How one researcher’s accumulated experimental intuition became a complete cognitive map for interface design — and what that process reveals about how breakthroughs actually form.
Cite
(April 9, 2026). From Passive to Programmable: Reading Notes on Weilai Yu’s REI Viewpoint. AtomPub AP-2026-reading-notes-rei. <>Today I attended Professor Weilai Yu’s lecture in person, then came home and reread his recently published Viewpoint in Accounts of Materials Research1
The paper, in one sentence
Yu argues that the solid-electrolyte interphase (SEI) is not a passive protective layer but an active material — one that can and should be designed for programmable reactivity. He calls this reframing the Reactive Electrochemical Interphase (REI).
This is not a small semantic shift. Calling the SEI “passive” implicitly frames interface engineering as a containment problem: how do we prevent the interphase from doing damage? Calling it “active” reframes the problem entirely: how do we program the interphase to do work?
Why the lecture landed differently than the paper
Reading a Viewpoint article is a one-pass experience. The author has already done the work of linearizing a nonlinear body of knowledge into a sequence of claims. You follow the argument.
A lecture is something else. What Yu showed us — almost involuntarily, in the way he paused on certain figures, in the questions he turned back on the audience — was the shape of the underlying epistemic structure before it was linearized. You could see where the certainty lived and where it didn’t.
The Viewpoint is a milestone summary. The lecture was the process of arriving at the milestone, made briefly visible.
A pattern worth noting. Major cognitive leaps in materials science often accumulate quietly: a series of seemingly modest results, each individually publishable but not obviously connected, until a critical mass is reached and someone draws the map. Yu’s Viewpoint is that map. The experimental scaffolding behind it was assembled over years; the synthesis happened fast.
Three questions from the Q&A
The Q&A surfaced the three threads I find most generative for my own work:
1. The dissolution transition layer and macroscopic SEI structure. How does the composition of the semi-soluble transition layer near the REI influence the formation and spatial evolution of classical SEI structural models — the mosaic model, the multilayer model? The REI framing implies these models are not fixed archetypes but outputs of a dynamic process. The question is: what are the inputs?
2. Electrochemical mapping of 2D structural nucleation. Can we read the 2D structural evolution of a Cu-SEI interface back out of nucleation patterns in potentiostatic chronoamperometry? The I–t curve as a spatially encoded signal, not just a kinetic summary. This connects directly to my own work on deposition behavior and seems underexplored.
3. Closed-loop theory–experiment verification. Yu uses DFT to reveal electrolyte decomposition pathways. The question is how to close the loop: what other characterization techniques would form a rigorous cross-validation system? The answer probably depends on which decomposition pathway you’re targeting, but the framework for answering it is worth building explicitly.
Two adjustments to my own practice
This exchange sharpened two habits I want to hold myself to:
Show up. High-level academic exchange is the fastest way to compress years of experimental intuition into a few hours of conversation. The opportunity cost of missing a lecture like this is not recoverable.
Ask better questions. The Q&A is not a formality. A well-constructed question advances the topic; a careless one wastes everyone’s time. I want to hold myself to the standard of asking questions that would earn a place in the paper if the paper were being written after the lecture.