Highly Efficient and Reversible Covalent Patterning of Graphene: 2D‐Management of Chemical Information

Abstract Patterned graphene‐functionalization with a tunable degree of functionalization can tailor the properties of graphene. Here, we present a new reductive functionalization approach combined with lithography rendering patterned graphene‐functionalization easily accessible. Two types of covalent patterning of graphene were prepared and their structures were unambiguously characterized by statistical Raman spectroscopy together with scanning electron microscopy/energy‐dispersive X‐ray spectroscopy (SEM‐EDS). The reversible defunctionalization processes, as revealed by temperature‐dependent Raman spectroscopy, enable the possibility to accurately modulate the degree of functionalization by annealing. This allows for the management of chemical information through complete write/store/erase cycles. Based on our strategy, controllable and efficient patterning graphene‐functionalization is no longer a challenge and facilitates the development of graphene‐based devices.


S1. Patterned graphene fabrication.
The sample was spin-coated with a Polymethylmethacrylate (PMMA) double layer (PMMA 200 k 100 nm, PMMA 950 k 200 nm), followed by a bake step after each layer (Layer 1: 180°C, 60 s; Layer 2: 180°C, 90 s). Subsequently two different patterns (Fig. S1 and S2) in the double layer was created by performing e-beam lithography with a Zeiss Supra SEM (10kV). Such conditions avoid any radiationrelated defects in the graphene layer 1 . Irradiated PMMA areas were removed by isopropanol-methyl isobutyl ketone solution. Figure S1. Patterned graphene within G A . The dark blue cyclic dots are areas where PMMA is removed such that the graphene is exposed. Figure S2. Patterned graphene within G B . The dark blue FAU logos are areas where the PMMA is removed such that

SUPPORTING INFORMATION
S4 the graphene is exposed.

S2. Patterned graphene functionalization of G A and G B .
Inside the glovebox (˂ 0.1 ppm O, ˂ 0.1 ppm H2O, Ar), the above prepared patterned graphene (GA and GB) were initially activated by reducing with Na/K (molar ratio 1:3) alloy. Specifically, a drop of liquid Na/K alloy (molar ratio is 1:3) is dripped onto the surface of the sample and kept for 1.5 h (the reduction is saturated), resulting in the exclusive reduction of unprotected graphene. This new activation procedure by reducing graphene directly with the liquid Na/K alloy differs considerably from our previously developed method employing Na/K-DME (DME = dimehtoxyethane) solution. The use of neat Na/K alloy avoids the application of the PMMA-damaging DME solvent thus guaranteeing compatibility with the mask lithography process. Two different diazonium salts including 4-Bromobenzenediazoniumtetrafluoroborate and 4-Nitrobenzenediazonium-tetrafluoroborate were respectively dissolved in dried and degased ethanol (0.5mmol / mL). Then the liquid Na/K alloy can be simply blown away with inert Ar. As soon as removing of the Na/K alloy one drop of each diazonium salt solution was added immediately for 15 min. Afterwards, the reactions were terminated by removing reactants with ethanol.
Subsequently, the two samples were exported from the glove-box and washed with additional 20 mL ethanol and 20 mL water. Then acetone was used to remove PMMA layer to give rise to the final GA and GB.
Raman spectroscopy and in particular, scanning-Raman spectroscopy-and microscopy (SRS and SRM) represents a very powerful monitoring tool for the investigation of covalent addend binding to graphene.
This technique was therefore applied to characterize our 2D-patterned sheet architectures GA and GB. As shown in Fig. S3 and S4, with the increase of reduction time for GA and GB, the intensity of D-band increased progressively, suggesting that the degree of functionalization increases gradually. Finally, the average ID/IG ratio of GA/GB can be enhanced to ca. 2.6/2.8. Thus, it's interesting that our method can control the degree of functionalization by manipulating the reduction time of graphene, which is of great importance for graphene chemistry. Figure S3. Raman spectra of patterning functionalizaition of graphene (G A ) upon different reduction time. Figure S4. Raman spectra of patterning functionalization of graphene (G B ) upon different reduction time. Table S1: Raman data of patterning functionalization of G A and G B upon different reduction time, λ exc = 532 nm

S3. Quantified local degree of functionalization of G A and G B .
Previous studies have pointed out the correlation between ID/IG and mean defect distance LD, with a maximum ID/IG ratio at a certain LD-crit value. [2,3] In general, the LD-crit can be used as a boundary to distinguish the low density functionalization from the high density functionalization. To determine the LD value under consideration, the width of the Raman bands is the core. As a consequence, the degree of functionalization of graphene can be quantified based on ID/IG ratio along with full width at half maxium (FWHM) of D peak. The determined ID/IG ratio combined with observed FWHM value (˂ 30 cm -1 ) of D peak indicate the low density functionalization of GA and the mean distance between defects (LD) was calculated to be 6.2 nm. Following the quantified graphene functionalization method we introduced earlier, [4] the local degree of functionalization can be quantified to 0.060%. Based on the observed ID/IG ratio as well as FWHM of value (˂ 30 cm -1 ) of D peak, the LD was calculated to be 5.9 nm corresponding to the local degree of functionalization of 0.066% for GB. Furthermore, the local degree of functionalization for both GA and GB under different annealing temperature has been quantified as well (Table S2). Besides, on basis of this method, we also quantified the degree of functionalization of previously reported patterning graphene functionalization for comparison (see Table S3). Table S2: Quantified local degree of functionalization (θ) of G A and G B upon different annealing temperature. Table S3: Comparison of quantified degree of functionalization (θ) of previous reported cases (patterning graphene addition) and this work.