Origin of Oxygen in Graphene Oxide Revealed by 17O and 18O Isotopic Labeling

Wet-chemical oxidation of graphite in a mixture of sulfuric acid with a strong oxidizer, such as potassium permanganate, leads to the formation of graphene oxide with hydroxyl and epoxide groups as the major functional groups. Nevertheless, the reaction mechanism remains unclear and the source of oxygen is a subject of debate. It could theoretically originate from the oxidizer, water, or sulfuric acid. In this study, we employed 18O and 17O labeled reagents to experimentally elucidate the reaction mechanism and, thus, determine the origin of oxo-functional groups. Our findings reveal the multifaceted roles of sulfuric acid, acting as a dispersion medium, a dehydrating agent for potassium permanganate, and an intercalant. Additionally, it significantly acts as a source of oxygen next to manganese oxides. Through 17O solid-state magic-angle spinning (MAS) NMR experiments, we exclude water as a direct reaction partner during oxygenation. With labeling experiments, we conclude on mechanistic insights, which may be exploited for the synthesis of novel graphene derivatives.


Isotopically labelled GO:
The general procedure for GO preparation, as outlined above, was downscaled using 50 mg graphite.In one reaction, 18 O4-labelled sulfuric was used and worked up with regular DI water (GO-18A-16W).In another reaction, regular sulfuric acid was used, but the reaction was worked up with 18 O-labelled water (GO-16A-18W).A total amount of 4 mL 18 O-labelled water was used to wash the sample for four times.The obtained material was freeze-dried and used for TGA-MS.
Reduction of GO: 300 nm SiO2/Si wafer cleaned with piranha solution for at least 4 h and subsequently rinsed with MilliQ.water.Then, a small amount of aqueous GO dispersion was diluted 1:1 with triple distilled methanol and either drop casted, or deposited by Langmuir-Blodgett technique on a freshly cleaned wafer 1 .The coating was dried under vacuum for at least 1 h.Then, the wafer was placed in a glass chamber, reduced for 10 min by vapor of hydrogen iodine and trifluoro acetic acid (equal in volumes, vapor generate on hotplate at 80 °C), and finally rinsed with MilliQ water.

Preparation of graphite intercalation compounds (GICs) for Raman analysis:
The preparation of GICs was conducted under inert gas condition in an Argon filled glove box (H2O and O2 < 0.01% ppm).Graphite (50 mg) and sulfuric acid (5 mL, 97.5%) were mixed in a small glass vial at room temperature with sodium persulfate, sodium nitrate or potassium permanganate and continuously stirred at room temperature.A maximum of one weight equivalent of oxidizer with respect to graphite was added each day at once, with a maximum of 3 weight equivalents.For analysis, some GIC PS and GIC N crystals were transferred together with a small amount of sulfuric acid on a glass slide.A cover slip was placed on top and common adhesive tape was used to seal the sample (Figure S1A).
Due to the intense color of the reaction mixture containing potassium permanganate, GIC PM particles were centrifuged for few for 10 times outside of the glove box.After each run, the supernatant was removed and replaced with fresh sulfuric acid until the mixture was colorless.To protect the GICs from ambient humidity during washing we added a certain volume n-hexane (Figure S1B).Obtained purified GIC PM crystals were transferred in the glove box on glass slides as aforementioned.For reference, one sample of graphite ("washed GIC PM ") was oxidized with only one weight equivalent of PM for two hours.It was washed with pure sulfuric acid by the above outlined procedure.
Exchange reactions with GO in H2O, D2O and 18 O-water: About 3 mg freeze dried GO-16A-16W was redispersed in approximately 3 mL DI water, D2O (99%), or 18 O-labelled water.After 24 h, the material was freeze dried and analyzed by TGA-MS and FTIR-analysis.
In another set of experiments, the oxygen-isotope exchange between the acid mixture and potassium permanganate was monitored by liquid NMR (Figure S7B) as follows: a) KMnO4 was dissolved in D2O at a concentration of 0.25 M and measured after 24 hours at room temperature; b and c) KMnO4 was dissolved in a mixture of D2O (800 µl) with 17 O-water (40%, 200 µl) at a final concentration of 0.25 M. the obtained solution as subsequently mixed with 1000 µl of 0.05 M HCl; d) KMnO4 (52.4 mg) was dissolved in 1 ml H2SO4 and stirred for at least 1 hour.The mixture reflects approximately the ratio of reagents used for the oxidation of graphite.After 30 minutes, 4 hours and 24 hours, a small amount of the mixture (100 µl) was diluted with D2O and 17 O-spectra were recorded within 5 minutes.All three spectra were almost identical. 17O ssNMR (67.8 MHz, Figure S7B): δ -3.0 -1.0 (H2O/H3O + ), δ 157.5 (H2SO4), δ 1210.9-1209.2(KMnO4).

Hydrolysis of organosulfates:
A freshly synthesized batch of GO was prepared according the general protocol of synthesis.Directly after synthesis, one share was of the material was freeze-dried overnight to determine the concentration of the material.The stock dispersion was adjusted to 0.6 mg/mL.After certain time intervals, a sample was centrifuged at 12.000 RCF for 30 minutes.The supernatant was analyzed by ion chromatography and the conductivity was determined.A set of 14mL deposable centrifugation tubes was filled with each 10 mL of the dispersion.One set of five was stored at 4°C in the refrigerator, another set of five stored at ambient conditions (25°C) in a locker.Two other sets of five were placed into a water bath with 40°C and 60°C, respectively.The same was done for the zero-sample at the start of the experiment directly after adjustment of the concentration (Figure 4D).For hydrolyzation at elevated temperatures over the course of several days, two almost identical GO dispersions were used (Figure 4E).One pair of 5 was incubated at 40°C and 60°C, the other pair was used at 60°C and 80°C.EA of GO-1: C = 42.50 %; H = 2.63 %; N < 0.5%; S = 3.91 %.GO-2: C = 38.90%; H = 2.52 %; N < 0.5%; S = 4.97.(2) Tarasov, V. P.; Kirakosyan, G. A. 16O/18O oxygen isotope exchange kinetics in MnO4 − as probed by 55Mn NMR.Russ.J. Phys.Chem. B 2016, 10 (4), 582-586. DOI: 10.1134/s1990793116040278.

Figure S1 :
Figure S1: (A) Setup for in situ oxidation and transfer of prepared graphite intercalation compounds (GICs) for Raman spectroscopy.(B) Washing procedure of GIC PM outside of the glove box with n-hexane at ambient conditions.

Figure S2 :
Figure S2: Raman spectrum of GO-16A-16W and GIC PM after 7 days in the reaction mixture with one weight equivalent of potassium permanganate and GO on 300 nm SiO2/Si wafer.The D and G Raman mode of functionalised graphene layers are indicated with letters at the corresponding positions, while G* indicates the Raman mode of stage-1 intercalated layers.The corresponding full-width-at-half-maximum of fitted Gauss-functions are given in italic numbers.

Figure S3 :
Figure S3: Photograph of oxidized graphite crystals with permanganate before (a) and after addition of water and hydrogen peroxide solution (b).

Figure S4 :
Figure S4: (A) Overview of the surface of a graphite particle, oxidized for 7 days in sulfuric acid with potassium permanganate.(B) Raman spectra of two different spots on the graphite surface.A higher intensity of G* stemming stage-1 intercalated domains can be observed in the blueish areas.(C, D) Heat maps of a selected area of the particle surface with respect to the G* and D mode intensity.

Figure S5 :
Figure S5: (A) Solid state NMR of GO-16A-16W.The corresponding functional groups are annotated together with their corresponding shift in bolt letters and the percentage of the integrated area in italic.(B) TGA-MS profile of GO-16A-16W.The corresponding m/z values of the cleaved fragments are shown in the inlet.(C) XPS survey spectra of GO-16A-16W and GO-18A-16W on gold substrate.The corresponding C1s high resolution spectra are shown in (D) for and (E).The functional groups are annotated together with their corresponding fraction in percent below.

Figure S6 :
Figure S6: TGA-MS spectra of the initial GO-16A-16W after incubation with (A) DI-water, (B) D2O, or (C) 18 O-labelled water for few hours.The samples were subsequently freeze driend and used as obtained for analysis.(D) Same TGA-MS spectrum as (C) but with a focus on m/z values for OH and OH2 (m/z 19, 20), labelled and unlabelled CO (m/z = 28, 30) and CO2 (m/z = 44, 46).(D) FTIR spectra of DI water, 18 O-labveleld water and D 2 O. (E) FTIR spectra of GO before and after incubation with 18 O-water or D2O for 24 hours.The inlet represents a magnifictation of the spectral range, where C=C double bonds (~1575cm -1) and vibration from adsiprbed water (~ 1620 cm -1 ) are visible.For D2O, no new vibrational bands are arising with respect to the spectrum of initial GO.Similarly, no shift in the band at 1620 cm -1 can be observed after incubation with 18 O-water.

Figure S7 :
Figure S7: 17 O-NMR reference spectra to monitor possible oxygen-echange bewen the water, sulfuric acid and potassium permanganate.(A) water cointaining 4% 17 O (black), pure sulfuric acid (~98%, red), and sulfuric acid slightly diluted with 17 Owater water after 5 minutes (blue) and 31 minutes (purple).The signal ratio reflects the equilibrium of ~94 % sulfuric acid.(B) 17 O-NMR spectrum of 0.25 M KMnO4 in D2O after 24 hours (black), 0.125 M KMnO4 containing 4% of 17 O-water and acidified with hydrochloric acid after 4 hours (red) and 24 hours of incubation (blue), and KMnO4 recovered from concentrated sulfuric by dilution of 100 µl of the acid mixture with 900 µl D2O (purple).As it can be seen in the inlet, after 24 hours of reaction time, a new signal at around 1210 ppm stemming from 17 O-KMnO4 appears in the acidified aqueous solution of KMnO4, which is absent for the other samples 2 .