Optimal architecture for diamond-based wide-field thermal imaging

Nitrogen-Vacancy centers in diamond possess an electronic spin resonance that strongly depends on temperature, which makes them efficient temperature sensor with a sensitivity down to a few mK/$\sqrt{\rm Hz}$. However, the high thermal conductivity of the host diamond may strongly damp any temperature variations, leading to invasive measurements when probing local temperature distributions. In view of determining possible and optimal configurations for diamond-based wide-field thermal imaging, we here investigate, both experimentally and numerically, the effect of the presence of diamond on microscale temperature distributions. Three geometrical configurations are studied: a bulk diamond substrate, a thin diamond layer bonded on quartz and diamond nanoparticles dispersed on quartz. We show that the use of bulk diamond substrates for thermal imaging is highly invasive, in the sense that it prevents any substantial temperature increase. Conversely, thin diamond layers partly solve this issue and could provide a possible alternative for microscale thermal imaging. Dispersions of diamond nanoparticles throughout the sample appear as the most relevant approach as they do not affect the temperature distribution, although NV centers in nanodiamonds yield lower temperature sensitivities compared to bulk diamond.