High-efficiency exfoliation of large-area mono-layer graphene oxide with controlled dimension

In this work, we introduce a novel and facile method of exfoliating large-area, single-layer graphene oxide using a shearing stress. The shearing stress reactor consists of two concentric cylinders, where the inner cylinder rotates at controlled speed while the outer cylinder is kept stationary. We found that the formation of Taylor vortex flow with shearing stress can effectively exfoliate the graphite oxide, resulting in large-area single- or few-layer graphene oxide (GO) platelets with high yields (>90%) within 60 min of reaction time. Moreover, the lateral size of exfoliated GO sheets was readily tunable by simply controlling the rotational speed of the reactor and reaction time. Our approach for high-efficiency exfoliation of GO with controlled dimension may find its utility in numerous industrial applications including energy storage, conducting composite, electronic device, and supporting frameworks of catalyst.


Results and Discussion
GtO is typically fabricated by the inter-layer oxidization of natural graphite with oxidizing agents such as KMnO 4 41-43 . As synthesized GtO contains carbonyl groups on the edge sites and hydroxyl and epoxy groups in the basal plane, which expand the inter-layer spacing of the natural graphite 17,21 . GtO, with its inter-layer space being expanded, can then be readily exfoliated into GO with single-or multi-layers due to the weakened inter-layer van der Waals forces 18,23 .
GtO produced via the oxidation reaction (60 min) in shearing stress reactor was spectroscopically analyzed as shown in Fig. 1 44 . Fig. 1a shows the x-ray diffraction (XRD) spectra of the produced GtO powders where the main peak of graphite at 26.5° (lower) is reduced to 10.3° after the oxidation reaction (upper). The inter-planar spacing of GtO (0.85 nm), measured by Bragg's Law ( λ θ = n dsin 2 ), was larger than that of the natural graphite (0.34 nm) due to the oxygen functional groups present on the carbon sheet 17,21 . Figure 1b shows the x-ray photoelectron spectroscopy (XPS) C1s spectra of GtO. Three main peaks from different functional groups of carbon atoms are clearly observed: carbon bond (C-C) at 284.5 eV, epoxy/hydroxyls group (C-O) at 286.7 eV, and carbonyl group (C=O) at 288.2 eV. Typical Raman spectra of graphene have two distinctive peaks at 1590 cm −1 and 1350 cm −1 , G mode representing the sp 2 C element and D mode representing the degree of defects in the sp 2 domain, respectively. As expected and shown in Fig. 1c, Raman spectra of GtO exhibit G band at 1589 cm −1 and pronounced D band at 1349 cm −1 originating from the reduced plane size of sp 2 domains owing to the intensive oxidation.  illustrates three different approaches (sonication-upper, homogenization-mid, and shearing stress-lower) for the GtO exfoliation process studied in this work. The GtO was prepared under the same experimental conditions and the exfoliation time was kept constant for all cases. Sonication is the most widely used approach for the exfoliation process and the single-layered GO flakes can be readily synthesized by the ultrasonic force (directionless force) 18,33,34,45 . In the homogenization, exfoliation takes place from the lateral side of GtO by the shearing stress and mechanical force, which is known to produce single-layered GO with relatively large lateral size 45 . The shearing stress reactor comprises two coaxial cylinders. While the outer cylinder remains standstill, the inner one rotates at controlled speed. When the rotational speed of the inner cylinder reaches a threshold value, doughnut-shaped vortexes are generated which rotates in opposite directions with constant arrays along the cylinder axis [38][39][40] . This shearing stress flow induces highly effective radial mixing and uniform fluidic motion within each vortex cell, enabling enhanced mass transfer of the reactants. The toroidal motion also generates high wall shear stress, which can facilitate the GtO exfoliation (Fig. 3a).
In shearing stress reactor, the hydrodynamic condition of the fluids depends on the rotating speed of the inner cylinder. The shearing stress flow is formed when the Taylor number (Ta) proportional to the angular velocity of the inner cylinder exceeds a critical value, which is determined by the following relation [38][39][40] : where R 1 is the radius of inner cylinder, ω 1 is the angular velocity of the inner cylinder, d is the width of the annular gap, and ν is the kinematic viscosity. In this work, the viscosity was set at 200 cP (water-dispersed GtO used for forming the shearing stress flow) and the exfoliation was carried out from 500 to 1500 rpm for 1, 3, and 5 hr. Figure 3b,c show the shearing stress reactor and its capability of large-volume GO production. The generated shearing stress, dependent on the speed of rotating inner cylinder, is plotted in Fig. 3d: 355.0 Pa at 500 rpm, Scientific RepoRts | 7: 16414 | DOI:10.1038/s41598-017-16649-y 744.2 Pa at 1000 rpm, and 1250.8 Pa at 1500 rpm, respectively. As expected, shearing stress increases with rotational speed, indicating more efficient GtO exfoliation at higher shear force.
The efficacy of various exfoliation methods was evaluated by analyzing the aggregated non-exfoliated GtO particulates. GtO was exfoliated in water, and the exfoliated GO and non-exfoliated GtO were isolated by centrifugation (500 rpm) after the exfoliation process. During centrifugation, non-exfoliated GtO particulates were precipitated out, while exfoliated GO remained in the supernatant. We note that the exfoliated GO in the supernatant was mostly single-or few-layer, and its measured weight relative to that of the initial GtO was used to determine the recovery rate.
Field emission scanning electron microscopy (FE-SEM) images and recovery rates of exfoliated GO obtained from the shearing stress are shown in Fig. 4. The exfoliation time was varied from 1 to 5 hr, while the rotating speed of inner cylinder was set as 500, 1000, and 1500 rpm. At 500 rpm (Fig. 4a), the lateral size of exfoliated GO was ~70, ~50, and ~30 µm and the recovery rate was 83.2, 88.1 and 91.8%, at each time duration. At 1000 rpm, the recovery rate increased over 93% even at 1 hr of exfoliation time, and the overall flake size was ~50 µm (Fig. 4b).
At 1500 rpm, the GO flake size ranged 10 ~ 30 µm, and the recovery rate was over 98% at all exfoliation times, indicating that most of GtO was successfully exfoliated and well-dispersed in water (Fig. 4c). Figure S1 shows the UV-vis spectrum according to the recovery rates of GO dispersed in water. GO only absorbs in the violet and UV range of light and has a maximum absorption at 231 nm. As the yield of single-or few-layer GO increased, the concentration was increased and the absorption rate increased accordingly. Table 1 shows the lateral dimension and recovery rate of exfoliated GO via sonication and homogenization method under various experimental conditions (Supplementary Figures S2, 3). For the sonication method at 80 W of sonication power, the lateral size of GO was large in general: ~50, ~40, and ~20 µm with sonication time of 1, 3, and 5 hr, respectively. However, the recovery rate was only 16.2% even at 5 hr of sonication time, indicating that most of the GtO precipitated without further exfoliation. Under this mild-sonication condition, large-size GO can be obtained, but achieving single-layer GO with high-yield is limited since the exfoliation cannot fully occur. At higher sonication power of 150 W, the exfoliation process was almost completed (recovery rates were 92.1, 98.3, and 98.9% at 1, 3, and 5 hr, respectively) but with decreased flake size (~0.5 to 5 µm range) and nonuniform flake distributions. With the sonication power further increased at 200 W, the exfoliation of GtO almost fully occurred: the lateral size and recovery rates were ~0.8, ~0.5, and ~0.3 µm, and 98.8, 99.0 and 99.1% at 1, 3, and 5 hr, respectively. As shown here, sonication method is intrinsically limited in producing single-layered GO flakes with large lateral size since the sonication generates acoustic wave agitation which can easily damage the starting GtO.
For the homogenization method at 3000 rpm rotational speed, the lateral size of exfoliated GO were ~40, ~30, and ~15 µm with homogenization time of 1, 3, and 5 hr, respectively, and the corresponding recovery rates were only 6.1, 13.5, and 26.7%, leaving most of the GtO unexfoliated. At 6000 rpm, the flake size was ~25 µm or less at all exfoliation time but the recovery rate was still 32.2% even at 5 hr of homogenization time. The recovery rate was somewhat improved at 9000 rpm (71.2% at 5 hr), but the corresponding flake size was less than 10 µm. Overall, the homogenization process results in low recovery rates with nonuniform size distributions, and the frictional force generated between the inner and outer blades during the rotation causes significant damages to the exfoliated GO flakes. Consequently, the yield of large-area and single-layer GO is also quite low in this method.
Based on these results, we conclude that sonication method typically produces GO flakes with small sizes due to the destructive acoustical wave agitation in the solution, and homogenization method results in GO with various flake size distributions because the GO sheets are readily destroyed by the mechanical forces of the blade. On the other hand, the shearing stress process promotes the non-destructive exfoliation of GtO owing to the slippage of the GO induced by the in-plane directed shear stress, which is described as the rheologically derived GO.
Atomic force microscopy (AFM) was performed to analyze the thickness of GO flakes prepared by three exfoliation methods -sonication, homogenization, and shearing stress -for exfoliation time of 1 hr and after centrifuging at 6000 rpm for 30 min (Fig. 5). The height profiles reveal uniform flake thickness of ~0.75 nm, typical of the exfoliated individual GO sheet, whereas the lateral size exhibited distinctive size variations: ~5, ~20, and ~50 µm range for sonication (150 W), homogenization (6000 rpm), and shearing stress (1000 rpm), respectively.  This difference in size distribution positioned the dispersed exfoliated flakes at different locations of the conical tube (upper, mid, and lower region) due to varying flake weights as shown in the digital image of Fig. 5. Compared to the conventional sonication and homogenization method, the recovery rate of shearing stress improved more than 90% with associated high-yield of single-or few-layered GO with large lateral size. The strong shear stress generated from the vortex in the reactor also enabled non-destructive exfoliation. These results suggest that shearing stress approach can afford effective exfoliation of GtO flakes with desired size tunability and quality, realizing facile production of single-or few-layer GO in high yields.

Conclusions
In summary, we have demonstrated a facile method to promote the non-destructive exfoliation process of GtO into GO using shearing stress reactor. We found that the shear stress generated inside the reactor enabled efficient exfoliation of GtO with tunable size and high yield of single-or few-layer GO sheets. In this work, large-area single-layer GO flakes over 80% yield was obtained, surpassing the conventional sonication or homogenization method. Our approach for facile and large-scale production of GO thus has great potentials in various industrial applications including energy storage and harvesting, conducting composites, or electronic devices.

Materials and Methods
Preparation of GtO. Graphite flakes (150 µm, Alfa Aesar) were oxidized using the shearing stress reactor (Supplementary Figure S4, Lamina co. ltd). 7 g of graphite flakes was added to 250 mL of sulfuric acid (H 2 SO 4 , 95%, Sigma Aldrich). Then, 32 g of potassium permanganate (KMnO 4 , Sigma Aldrich) was slowly added to the mixture at ~10 °C, and stirred for 30 min. Shearing stress reactor (length: 500 mm) consists of two coaxial cylinders with the fixed outer cylinder (radius: 68 mm) and the rotating inner cylinder (radius: 60 mm). After the mixture was introduced into the gap between the two cylinders, the inner cylinder was rotated. Oxidation of graphite inside the shearing stress reactor led to brown-colored slurry. Then, 250 mL of purified water and 15 mL of hydrogen peroxide (H 2 O 2 , 30%, Sigma Aldrich) were added to the mixture, and stirred for 30 minutes. For purification, centrifugation was used to separate the GtO from the impurities. Finally, the dried GtO powders were obtained by freeze-drying.
Exfoliation of GtO for GO. In this study, three different exfoliation methods were carried out to prepare GO from GtO. All exfoliation methods were conducted for 1, 3, and 5 hr duration. Variable parameters were power (80, 150, and 200 W) for the sonication method, rotating speed of axis (3000, 6000, and 9000 rpm) for the homogenization method, and rotating speed of inner cylinder (500, 1000, 1500 rpm) for the shearing stress reactor, respectively. In the case of sonication and homogenization method, GtO was dispersed in water with concentration of 1 mg/mL, whereas the viscosity of GtO dispersed in water was set at 200 cP for the shearing stress reactor. Non-exfoliated GtO particles were precipitated out by centrifugation at 500 rpm for 10 min. The supernatant containing single-or few-layer GO was obtained and coated on Si substrate for further characterization.
Characterization. The microstructure and lateral size of the samples was investigated using field emission scanning electron microscopy (JSM-7600F, JEOL). The thickness and lateral size of GO sheets were obtained using atomic force microscopy (SPA-300HV, SII). X-ray diffraction patterns of oxidized graphite were determined by D8 ADVANCE (Bruker Corporation) with Cu-Ka X-ray source. Raman spectra were obtained by micro-Raman system (Bruker FRA 160/S, Bruker) with excitation energy of 2.41 eV. X-ray photoelectron spectroscopy spectra of GO samples were obtained by VG Microtech ESCA 2000 (JEOL) with a monochromatic Al-Ka X-ray source at 250 W. The UV-Vis absorption spectrum was obtained by using a TIDAS 100 spectrophotometer (J&M, Germany). The spectrum has been recorded by measuring a 0.02 wt% solution of GO dissolved in water.