Bayesian optimization enables balanced ambipolar transport in self‐doped random conjugated polymers

Abstract Random donor‐acceptor (D‐A) conjugated polymers offer exceptional mechanical compliance for flexible electronics, yet their vast compositional space makes rational optimization extremely challenging. Here, we introduce a combined side‐chain self‐doping and Bayesian optimization (BO) strategy that enables rapid, low‐sample‐count tuning of polarity in random terpolymers. Using a diketopyrrolopyrrole‐based polymer (PDPPBT) as a model system, we introduced dimethylamino (NMe 2 ) self‐doping groups to modulate carrier polarity. The optimization process, navigating a non‐linear structure–property landscape, converged to the optimal composition (PDPPBT‐NM15.218) after a single BO iteration with only six experimental data points. The resulting polymer exhibits nearly perfectly balanced ambipolar transport ( μ ave h/ μ ave e = 1) in both rigid and flexible OFETs. Comprehensive characterizations reveal that BO successfully located the critical trade‐off point where self‐doping effectively modulates frontier orbital energies, aligns Fermi levels, and subtly reorganizes thin‐film microstructures to enable balanced charge transport. This work demonstrates the power of BO in decoding complex composition‐function relationships, offering a generalizable route to accelerate the discovery of high‐performance functional polymers. image