Extended Data Fig. 1: Benchmarking FET performances of large area TMDs.

a. Benchmark data illustrating the I_on of MoS₂ FET relative to the integration scale. MoS₂ growth methods were categorized as thermal CVD (brown circle), Epitaxial growth on sapphire (navy square), and MOCVD (orange triangle). Here, CVD and epitaxial growth exhibit scaling limitations, as expected, but it appears that MoS₂ grown via epitaxial growth shows better performance. Although MOCVD clearly has advantages in terms of scaling, it shows limitations in performance. But this work overcomes these limitations of MOCVD and demonstrates performance as high as chip-scale research. b. Benchmark data for the I_on and channel length of large-area MoS₂ classified by patterning methods. Typically, photolithography is well-suited for large-scale integration; however, it leaves a significant amount of photoresist residue and involves relatively harsh wet processes, which can be detrimental to performance. On the other hand, e-beam lithography, whereas not well-suited for integration, utilizes a relatively clean e-beam resist (ER) and is suitable for fundamental studies through small-scale components, thanks to its high resolution. In this regard, the contact region, which is most influenced by photolithography, remains unexposed in the bottom contact geometry. Therefore, bottom contact devices exhibited performance equivalent or high to those fabricated using e-beam lithography and achieved a higher on-current compared to all other devices fabricated using photolithography. All data are based on V_ds = 1 V, and some values are approximately extracted, so they may not be highly precise. In cases I_on at V_ds = 1 V could not be obtained, values were unavoidably extracted at the specified V_ds values in each paper, indicated within parentheses.