Hyperspectral Raman IMAging (RIMA) is used to study spatially inhomogeneous polycrystalline monolayer graphene films grown by chemical vapor deposition. Based on principal component analysis clustering, distinct regions are differentiated and probed after subsequent exposures to the late afterglow of a microwave nitrogen plasma at a reduced pressure of 6 Torr (800 Pa). The 90 × 90 µm² RIMA mapping shows differentiation between graphene domains (GDs), grain boundaries (GBs), as well as contaminants adsorbed over and under the graphene layer. Through an analysis of a few relevant band parameters, the mapping further provides a statistical assessment of damage, strain, and doping levels in plasma-treated graphene. It is found that GBs exhibit lower levels of damage and N-incorporation than GDs. The selectivity at GBs is ascribed to (i) a low migration barrier of C adatoms compared to N-adatoms and vacancies and (ii) an anisotropic transport of C adatoms along GBs, which enhances adatom-vacancy recombination at GBs. This preferential self-healing at GBs of plasma-induced damage ensures selective incorporation of N-dopants at plasma-generated defect sites within GDs. This surprising selectivity vanishes, however, as the graphene approaches an amorphous state.

高光谱拉曼IMAging（RIMA）用于研究通过化学气相沉积法生长的空间不均匀多晶单层石墨烯薄膜。基于主成分分析聚类，在随后暴露于6 Torr（800 Pa）的微波氮等离子体的余辉之后，区分并探测了不同的区域。90×90 µm ²RIMA映射显示了石墨烯域（GDs），晶界（GBs）以及在石墨烯层上方和下方吸附的污染物之间的区别。通过分析一些相关的能带参数，该映射进一步提供了等离子体处理的石墨烯中损伤，应变和掺杂水平的统计评估。已经发现，GBs比GDs显示出更低的损伤和氮结合水平。GBs的选择性归因于（i）C原子与N原子和空位相比迁移障碍低，以及（ii）C原子沿GBs的各向异性迁移，这增强了GBs的原子-空位重组。血浆诱导的损伤在GBs处的这种优先的自我修复可确保在GDs中血浆产生的缺陷部位选择性掺入N-掺杂剂。但是，这种令人惊讶的选择性消失了，