The photoconversion of CO2 to hydrocarbons
is a sustainable
route for its transformation into value-added compounds, which is
crucial to mitigating energy and climate crises. CuPt nanoparticles
on TiO2 surfaces have been reported to show promising photoconversion
efficiencies. For further progress, a mechanistic understanding of
the catalytic properties of these CuPt/TiO2 systems is
vital. Here, we employ ab initio calculations, machine
learning, and photocatalysis experiments to understand the photocatalytic
reduction of CO2 on CuPt/TiO2. We explore the
configurational space of the CO2@CuPt/TiO2 systems
and examine their structures and energetics. We find that the CuPt/TiO2 interface plays a key role in determining CO2 activation
and, thus, the conversion to hydrocarbons. The interface stabilizes
*CO and other intermediates containing CH groups, thus facilitating
a higher activity and selectivity for methane. A bias-corrected machine-learning
interatomic potential trained on density functional theory data enables
the efficient exploration of the potential energy surfaces of numerous
CO2@CuPt/TiO2 configurations using basin-hopping
Monte Carlo simulations, greatly accelerating the study of these photocatalyst
systems. Our simulations show that CO2 preferentially adsorbs
at the interface, with a C atom bonded to a Pt site and one O atom
occupying an O-vacancy site. The interface also promotes the formation
of *CH and *CH2 intermediates. For confirmation, we synthesize
CuPt/TiO2 samples with various compositions, analyze their
morphologies and compositions using scanning electron microscopy and
energy-dispersive X-ray spectroscopy, and measure their photocatalytic
activity. Our computational and experimental findings qualitatively
agree and highlight the importance of the interface design for the
selective conversion of CO2 to hydrocarbons.