In this study, we thoroughly explored the thermal transport and thermo-
electric properties of BN-co-doped phagraphene structures in the context of
first-principles computations combined with machine learning interatomic po-
tential (MLIP) techniques. Our results demonstrate that doping positions can
critically tune the thermal properties, offering potential advantages for both
thermoelectric and heat transfer applications. Notably, enhanced thermal con-
ductivity has been obtained for phagraphene with 10% co-doping. Additionally,
the rectangular structural symmetry of phagraphene plays an important role
for targeted thermal transport. Further, we observe negative Grüneisen pa-
rameter in these structures, suggesting negative thermal expansion. This will
serves as an unique mechanism for controlling the thermal conductivity at dif-
ferent frequencies. Interestingly, n-type behavior of the structures is indicative
of its negative Seebeck coefficient within the temperature range of 300 K to 900
K. Moreover, these structures display significantly larger electrical conductivity
compared to other two-dimensional (2D) materials. Apart from that, the cal-
culated figure of merit of the structures under a constant relaxation time at 300
K shows considerably better response for some specific doped structures. We
believe this study will play an important role in understanding the importance
of structural modifications in tailoring the thermal properties of carbon-based
2D systems.