# Fileset

[MassProductionGraphene_Tanks.pdf](https://mdr.nims.go.jp/filesets/85ab6f9b-bff7-41b6-b973-df6431bc7d3e/download)

## Creator

Yoshihiko Arao, Riichi Kuwahara, Kaoru Ohno, [Jonathon Tanks](https://orcid.org/0000-0002-0232-8240), Kojiro Aida, Masatoshi Kubouchi, Shin-ichi Takeda

## Rights

[Creative Commons BY Attribution 4.0 International](https://creativecommons.org/licenses/by/4.0/)

## Other metadata

[Mass production of low-boiling point solvent- and water-soluble graphene by simple salt-assisted ball milling](https://mdr.nims.go.jp/datasets/d454c039-6cbd-464d-a03a-f6ed2d5a956f)

## Fulltext

Mass production of low-boiling point solvent- and water-soluble graphene by simple salt-assisted ball millingNanoscaleAdvancesPAPEROpen Access Article. Published on 21 November 2019. Downloaded on 12/11/2019 1:26:43 AM.  This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.View Article OnlineView Journal  | View IssueMass productionaTokyo Institute of Technology, School of MaO-okayama, Meguro-ku, Tokyo, Japan. E-mabDassault Systèmes, ThinkPark Tower 2-1-1cDepartment of Physics, Yokohama NationalYokohama, JapandNational Institute for Materials Science, SeeStructures and Advanced Composite ReseAgency (JAXA), 6-13-1 Osawa, Mitaka-shi, T† Electronic supplementary informa10.1039/c9na00463gCite this:Nanoscale Adv., 2019, 1, 4955Received 30th July 2019Accepted 29th October 2019DOI: 10.1039/c9na00463grsc.li/nanoscale-advancesThis journal is © The Royal Society of Cof low-boiling point solvent- andwater-soluble graphene by simple salt-assisted ballmilling†Yoshihiko Arao, *a Riichi Kuwahara,b Kaoru Ohno, c Jonathon Tanks, dKojiro Aida,a Masatoshi Kubouchia and Shin-ichi TakedaeDeveloping a mass production method for graphene is essential for practical usage of this remarkablematerial. Direct exfoliation of graphite in a liquid is a promising approach for production of high qualitygraphene. However, this technique has three huge obstacles to be solved; limitation of solvent, low yieldand low quality (i.e., multilayer graphene with a small size). Here, we found that soluble graphiteproduced by mechanochemical reaction with salts overcomes the above three drawbacks. Solublegraphite was exfoliated into monolayer graphene with more than 10% yield in five minutes of sonication.The modified graphite was easily exfoliated in a low-boiling point solvent such as acetone, alcohol andwater without the aid of a surfactant. Molecular simulation revealed that the salt is adsorbed to the activecarbon at the graphite edge. In the case of weak acid salts, the original bonding nature between thealkali ion and the base molecule is retained after the reaction. Thus, alkali metals are easily dissociated ina polar solvent, leading to negative charge of graphene, enabling the exfoliation of graphite in lowboiling point solvents. The approach proposed here opens up a new door to practical usage of theattractive 2D material.IntroductionGraphene has received enormous attention in the eld ofmicroelectronics and composite materials. A wide range ofapplications such as high-sensitivity sensors, thin lm transis-tors, transparent conductive lms, and anti-corrosion coatingshave been proposed up to now.1–6 Commercialization of gra-phene for these attractive applications highly depends on theprogress of graphene production technology. Bottom-upapproaches like chemical vapor deposition can fabricate large-area high-quality graphene, but the productivity is usually at themilligram-scale and it is not likely to become a mainstreammass production technique.7 Thus, a top-down approach—i.e.,exfoliation of graphite—is the only feasible method to producegraphene at the ton scale.Graphene was rst produced by mechanical cleavage usingscotch tape to literally peel off layers from natural graphite.8 Ofterials and Chemical Technology, 2-12-1il: yoshihiko.arao@gmail.comOsaki, Shinagawa-ku, Tokyo, JapanUniversity, 79-5 Tokiwadai, Hodogaya-ku,ngen 1-2-1, Tsukuba, Ibaraki, Japanarch Unit, Japan Aerospace Explorationokyo, Japantion (ESI) available. See DOI:hemistry 2019course, this method requires considerable labor and is notappropriate for mass production. It is well known that severalatomic or molecular species can be chemically inserted betweengraphene layers of the host graphite, a process known asintercalation. Graphite intercalation compounds (GICs) can beproduced by means of anodic or chemical oxidation; in general,0.11 wt. equiv. of potassium permanganate (KMnO4) is addedinto concentrated sulfuric acid (98%) to covert 1 wt. equiv. ofgraphite to stage 1 GICs.9 Expanded graphite is produced byintercalation of a strong acid following gasication of theintercalant.10–12 Aer that, the expanded graphite is exfoliated toproduce multi-layer graphene (>10 layers). This method is thesimplest and its feasibility for mass production has alreadybeen established. However, there are no repulsive forcesbetween the platelets in dry powder form (i.e., no solvent), onlyattractive van der Waals forces. This results in agglomeration ofthe multi-layer graphene platelets that are difficult to dispersedue to their large face-to-face contact area. When 6 wt. equiv. ofKMnO4 is added during the intercalation process, the GIC isconverted to graphene oxide, from which graphene can be ob-tained by reduction.13,14 This route tends to generate highmaterial and processing costs. In addition, the defects that areinduced in the basal plane of graphene during oxidation are notable to be fully recovered by the reduction process. Thus, in thepresent state, it is very challenging to develop a low-costproduction route to graphene with high quality via grapheneoxide.Nanoscale Adv., 2019, 1, 4955–4964 | 4955http://crossmark.crossref.org/dialog/?doi=10.1039/c9na00463g&domain=pdf&date_stamp=2019-11-28http://orcid.org/0000-0003-1534-3420http://orcid.org/0000-0002-1980-5971http://orcid.org/0000-0002-0232-8240http://creativecommons.org/licenses/by/3.0/http://creativecommons.org/licenses/by/3.0/https://doi.org/10.1039/c9na00463ghttps://pubs.rsc.org/en/journals/journal/NAhttps://pubs.rsc.org/en/journals/journal/NA?issueid=NA001012Nanoscale Advances PaperOpen Access Article. Published on 21 November 2019. Downloaded on 12/11/2019 1:26:43 AM.  This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.View Article OnlineWith respect to electrochemical exfoliation, either cathodicor anodic potentials drive guest ions into graphene layers of thehost graphite, which leads to expansion of graphite akes.Gasication of the intercalated species then facilitates theexfoliation of the expanded graphite.15 Although high qualitygraphene can be obtained by this method, electrical connec-tivity to the graphite feed materials must be maintained duringthe entire process, and if contact is broken the graphene yielddramatically decreases. Furthermore, the high quality stockgraphite materials that are required—such as rods, foil orhighly oriented pyrolytic slabs—lead to high material cost. Itwas proposed that there is a fundamental need to re-engineerthis method so that the electrochemical driving force can beapplied to graphite materials more efficiently and effectively.16Graphene cannot be stored in dry powder form, asmentioned above; in a liquid, however, repulsive forces such assteric and electrical repulsion are available to stabilize thegraphene dispersion. A method known as liquid-phase exfolia-tion (LPE) was introduced in 2008, in which graphite is sub-jected to sonication or shear mixing in a liquid to produce few-layer graphene.17 Shear mixing is the simplest LPE method interms of equipment and it is scalable,18 so it has strong poten-tial for encouraging the commercialization of non-oxide gra-phene. To achieve good exfoliation and dispersion performancewith LPE, it is imperative to choose a compatible solvent—meaning its surface tension is close to that of graphene (thesurface energy is�68 mJ m�2), such that the enthalpy of mixingis minimized.17,19,20 Although the LPE method holds promise forthe mass production of non-oxide graphene, it faces threeissues that must be overcome before widespread practicalapplication can be realized.The most serious one is the limited selection of compatiblesolvents: N-methyl-2-pyrrolidone (NMP) and dimethylforma-mide (DMF) are commonly acknowledged as the most appro-priate solvents, because the surface tension of these solventsmatches that of most nanocarbons.17,18 However, drawbackssuch as a high boiling point make it difficult to prepare thesedispersions for subsequent processing such as ink jet printingor coating, and toxicity is also a concern for human health. Forexample, a typical solvent exchange process involves switchingthe exfoliated graphene from NMP to chloroform or alcohol;1,2however, the ltration of well-dispersed graphene is time-consuming, creating a bottleneck that impedes so-called “onepot” synthesis strategies and ultimately slows down the indus-trial usage of graphite exfoliated via LPE.The second problem is polydispersity of the thickness anda low aspect ratio. To make effective use of graphene'sphenomenal properties in applications such as high-perfor-mance polymeric nanocomposites, a high aspect ratio (i.e., largelateral size with a small thickness) is ideal.21,22 The average layernumber (N) and aspect ratio of exfoliated graphene are usuallybetween 3 and 7 and 50 and 100, respectively.18 If we separatemonolayer graphene with a higher aspect ratio from few-layergraphene (FLG) with a lower aspect ratio in terms of theirmaterial properties used in engineering design, it is clear thata variable mixture of monolayer/FLG (polydispersity) would adddifficulty to the design and manufacture of the nal product.4956 | Nanoscale Adv., 2019, 1, 4955–4964Thus, efficient production of graphene with uniform dimen-sions is ideal.The third issue is the low yield of graphene, particularlymonolayer graphene. A yield in the range of 0.1–10% aer 1 h ofsonication has commonly been reported, depending on theinitial graphite structure and centrifugation conditions.23,24Recently, microuidization was proposed by several authorsas an alternative to shear mixing or sonication.25–27 Micro-uidization is a homogenization process in which a dispersionis forced through a narrow gap by the application of highpressure. The uid–particle interaction inside the channel canbe controlled by changing the uid dynamics (e.g. turbulentow, laminar ow, cavitation, and collision). Graphene yieldusing microuidization was reportedly 10 times higher thanthat using high-power probe sonication.25 Karagiannidis et al.produced conductive printable graphene inks with a 100%exfoliation yield (i.e., graphene rather than graphite akes).27Although it appears that the low yield of graphene by LPE mightbe drastically improved by microuidization, the exfoliationdegree (i.e., layer number) of graphene by this method is usuallyinsufficient; 96% of akes fall in the 4–70 nm thickness range.27Moreover, as mentioned above, the limitation of compatiblesolvents such as NMP and DMF systems is still the critical issue,and these facts restrict the potential of the microuidizationmethod and thus the application of exfoliated graphene.Until now, there was no single method that could address allthree of the above drawbacks. In this paper, we present a newmodied graphite that undergoes exfoliation in low boilingpoint solvents to monolayer graphene, with a total yield of morethan 10%. Carbon radicals induced by fragmentation ofgraphite in a ball milling process react with some selected saltmolecules (Fig. 1a). This mechanochemical reaction producesedge-functionalized graphite that has the ability to inducenegative charge in polar solvents and facilitate exfoliation inwhat are typically incompatible solvents, such as water, acetone,alcohol, and so on (Fig. 1b). The edge-functionalization isconrmed by detailed chemical analysis. In addition, molecularsimulation reveals that the salt is adsorbed on the active carbonat the graphite edge. The soluble graphite developed in thiswork can denitely expand the effective usage of these advanced2D materials.Results and discussionMechanochemical reaction of graphitePlanetary ball milling is conducted to induce the mechano-chemical reaction of graphite with salts. The approach toproduce graphene via ball milling has been widely investi-gated.28–35 Baek's group found that edge-functionalized gra-phene nanoplatelets could be prepared by dry ball milling in thepresence of hydrogen, carbon dioxide, or sulfur trioxide whichacts as an electrocatalyst for the oxygen reduction reaction.28,29Salt such as sodium chloride or sodium sulfate is added tofacilitate the delamination of graphite.30,31 However, the crystalsize of graphene gradually decreases because of the fragmen-tation caused by the impact of the balls, making it difficult tomaintain a layered structure aer a long milling time (>48 h).This journal is © The Royal Society of Chemistry 2019http://creativecommons.org/licenses/by/3.0/http://creativecommons.org/licenses/by/3.0/https://doi.org/10.1039/c9na00463gFig. 1 Schematic diagram of the mechanochemical reaction and exfoliation of anionic graphene. (a) Carbon radicals are generated by frag-mentation of graphite by ball milling. Some salts react with the carbon radicals, forming a functionalized edge. (b) Alkali metals at the edgedissociate in a polar solvent, leading to negative charge. The enhanced electrical repulsion enables exfoliation of graphite in incompatiblesolvents such as water and alcohol.Paper Nanoscale AdvancesOpen Access Article. Published on 21 November 2019. Downloaded on 12/11/2019 1:26:43 AM.  This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.View Article OnlineBecause so many defects are introduced if ball millingcontinues for long durations, the milled graphite tends tobecome very ne with an amorphous structure, rather thancrystalline graphene.36,37 To avoid this problem, a protectiveagent such as naphthalene, pyrene, melamine, or ammonia gasis introduced during ball milling.32–35 Although these additiveshelp produce high-quality FLG, the thickness is usually notuniform and the milled powder can only be dispersed in NMP,DMF, or a water/surfactant solution.In our process, we added weak acid salts such as potassiumcarbonate or sodium acetate to modify the graphite structure.The milling time, quantity of salt, and type of salt all havedominant effects on the solubility and quality of the resultantgraphite; hence we carefully investigated these parameters.The functionalization process is extremely simple. Themixture of graphite and salt was milled for an arbitraryamount of time, and then the mixture was washed severaltimes until the pH of water became neutral. The ltered cakewas dried at 60 �C overnight, and then ground in a mortar toobtain soluble graphite powder. To assess the quality of thepowder, Raman and XRD analyses, as well as SEM observation,were conducted. In order to investigate the solubility ofgraphite, 0.3 g of the produced graphite was added into 100 mlof isopropanol and sonicated for 5 min, aer which thedispersion was centrifuged at 1500 rpm for 30 min to removeunexfoliated akes. An optical absorbance A of the dispersionat 660 nm was measured with a spectrophotometer. Theconcentration of graphene C in the liquid was calculated usingthe Lambert–Beer law A ¼ acl where l is the light path length, cis the graphene concentration, and a is the absorption coef-cient. The absorption coefficient was 3300 L g�1 m�1.Theyield of graphene (C/Ci) was calculated by dividing the outputgraphene concentration by the initial graphite concentration(Ci ¼ 3 g L�1).This journal is © The Royal Society of Chemistry 2019Typical Raman spectra of milled graphite powder are shownin Fig. 2a. It should be noted that the laser spot size was 1.0 mm,smaller than the diameter of milled graphite powder (5–30 mm,depending on the milling time), and larger than the size ofindividual graphene aer LPE (0.4 mm on average). For char-acterization of graphene, we used a graphene thin lm depos-ited on lter paper. ID/IG values were scattered, depending onthe laser position in this characterization. Therefore, wemeasured Raman spectra of at least 5 spots (usually 10 spots),and the average ID/IG was obtained. This method is oen used tocharacterize the defect or size of graphene, which has a smallerdiameter than the laser spot.18 The spectra are normalized bythe intensity of the G band (�1590 cm�1). The D peak (�1350cm�1) and D0 peak (�1620 cm�1), which indicate graphitedefects, increased with the increasing time. The type of defectsthat appear in graphene—sp3 defects, vacancy-like defects, andedge defects—can be distinguished by the intensity ratio of theD and D0 peaks.38 A low ID/ID0 value of approximately 2.1 wascalculated for all the milled samples, indicating that only edgedefects are introduced during the ball milling process ratherthan basal plane defects (Fig. S1†). The intensity ratio of the Dand G peaks reects the defect density of graphite, so it is oenused to quantify the quality of graphene. The ID/IG value grad-ually increased with the increasing milling time (Fig. 2b),exceeding 0.5 aer 2 h of milling time. The ID/IG value ofoxidized graphite is in the range of 0.8–1.5, meaning thatpowder milled for more than 2 h is either graphite oxide oramorphous carbon.36,37Fig. 2b shows the yield of graphene by the LPE process, inwhich the graphene yield without salt-assisted ball milling wasonly 0.05% in isopropyl alcohol (IPA). This is due to the lowsurface tension of IPA (21.8 mN m; around 40 mN m�1 ispreferable for exfoliation) and short sonication time. Althoughexfoliation was conducted in an incompatible solvent, the yieldNanoscale Adv., 2019, 1, 4955–4964 | 4957http://creativecommons.org/licenses/by/3.0/http://creativecommons.org/licenses/by/3.0/https://doi.org/10.1039/c9na00463gFig. 2 Characterization of graphite powder after salt-assisted ball milling. (a) Typical Raman spectra of graphite powders after salt-assisted ballmilling. (b) Effect of milling time on the ID/IG ratio and graphene yield after liquid-phase exfoliation in IPA. (c) XRD profiles normalized by theintensity of the silicon (110) peak. (d) Content of potassium and oxygen in graphite powder after ball milling with potassium carbonate measuredby EPMA. (e) FT-IR survey of salt-assisted milled graphite. (f) XPS survey of salt-assisted milled powder. K 2p spectrum of milled powder (inset).Nanoscale Advances PaperOpen Access Article. Published on 21 November 2019. Downloaded on 12/11/2019 1:26:43 AM.  This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.View Article Onlineof graphene reached 25% when we used the salt-modiedgraphite aer milling for 30 min. The yield increased to 60%when 2 h of salt-assisted ball milling was applied, but beyond 2h the graphene yield decreased with increasing milling time—presumably due to the structural changes described above.We varied the salt content during ball milling to study theeffect of salt quantity on the resulting solubility. Milling withoutsalt yielded 0.09%, only slightly improved compared to that ofpristine graphite, while the yield increased linearly with saltcontent until reaching a saturated value of 30% at a salt contentof 2 g (2 wt. equiv.) (Fig. S2†). It should be noted that theaddition of salt did not facilitate the fragmentation of graphite.The ID/IG value of milled powder gradually decreased with theincreasing salt content (Fig. S3†) indicating that some of themilling energy was consumed by grinding the salt into nerpowder, thereby reducing the energy contribution towardgraphite fragmentation. The type of salt also has a dominanteffect on dispersibility when using the LPE process. Inorganic(strong acid) salts such as sodium sulfate and sodium chlorideare oen used as a gliding assistant to improve the yield ofgraphene by LPE to some degree, usually between 2 and 100times higher than that without any salt.30 In contrast, organic(weak acid) salts like sodium acetate, potassium sodiumtartrate, and potassium carbonate improve the yield of gra-phene 700 timesmore thanmilling without salt (Table S1†). Thereason for this difference between strong acid and weak acidsalts will be discussed later in this paper. We also found thatgraphite powder milled with acetic acid or tartaric acid (i.e., not4958 | Nanoscale Adv., 2019, 1, 4955–4964in salt form) showed low yield (<0.03%), which points to theimportant role of the cation in the mechanochemical reactionand dispersion mechanisms. Salts including alkali metals (e.g.,potassium, sodium) improve graphene yield, whereas coppercarbonate and copper phosphate have no effect on solubility ofexfoliated graphene (Table S1†). Cations with a greater ioniza-tion tendency are required to improve the solubility of graphenein a solvent.Aer the milled graphite strongly aggregates into sphericalagglomerations aer being washed and dried due to the relativestrength of van der Waals forces, the size decreases with theincreasing milling time (Fig. S4†). Powder XRD analysis indi-cates that no intercalation occurs with salt-assisted ball milling(Fig. 2c and Table S2†). The d-spacing of milled graphitecalculated from the peak position of the (002) reection was0.3358–0.3364 nm, which is almost the same d-spacing Fig. 2bof natural graphite (0.3355). The d-spacing slightly increasedaer ball milling because the graphene structure changed fromregular stacked layers to turbostratic. This value is lower thanthe d-spacing of reduced graphene oxide (0.356) or electro-chemically exfoliated graphene (0.348).39,40 This implies thatthere is no intercalant and no functional group on the basalplanes of graphene, which is consistent with the results ofRaman analysis (defects are introduced only at the edge). Theintensity of the 002 peak gradually decreased as ball millingprogressed. The crystallite thickness Lc can be obtained fromthe full-with at half-maximum (FWHM) of the 002 prole, andthe crystallite size La is determined from a 110 prole of carbonThis journal is © The Royal Society of Chemistry 2019http://creativecommons.org/licenses/by/3.0/http://creativecommons.org/licenses/by/3.0/https://doi.org/10.1039/c9na00463gPaper Nanoscale AdvancesOpen Access Article. Published on 21 November 2019. Downloaded on 12/11/2019 1:26:43 AM.  This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.View Article Online(Fig. S5†).41 The crystallite thickness decreased from 910 nm to21.1 nm aer 2 h of ball milling, meaning that the ball millingexfoliates graphite to multilayer graphene (layer number N >10), but not to few-layer graphene (N < 10). In the case of powdertreated by 12 h of milling, the Lc value could not be determinedcorrectly because the 002 peak was asymmetric and small. Thegraphite structure became amorphous aer 12 h of milling,with La ¼ 23 nm; this might make it difficult to maintain thelayered structure, leading to a spherical agglomeration (Fig.-S4d†). These experimental results indicate that the powdersaer salt-assisted ball milling are not few-layer graphene, butsimply functionalized graphite (i.e., essentially no exfoliation).To check whether potassium carbonate is chemically bondedwith graphite during ball milling, electron probe micro analysis(EPMA), Fourier transform infrared spectroscopy (FT-IR), and X-ray photoelectron spectroscopy (XPS) were performed. Based onthe elemental analysis of EPMA, the potassium content andoxygen content almost linearly increased with milling time.These elements gradually decreased with increasing washingtime but could not be removed completely aer more than vetimes. There are no functional groups on the graphene basalplane; thus it is reasonable to consider that the potassiumcarbonate adsorbs only along the edge of graphite during salt-assisted ball milling. In the FT-IR results, the characteristicpeak for carboxylate (COO�) appears around 1720 cm�1(Fig. 2e), while 1180 cm�1 corresponds to C–O stretching. Thisindicates that the carbonate ion has reacted with the ball-milledgraphene to form a carboxylate-functionalized edge. Addition-ally, the symmetric stretching and asymmetric stretching of–CH2 are located at 2850 cm�1 and 2930 cm�1, respectively; thesignicant increase in these peaks for the ball-milled grapheneis a result of passivation of the active carbons around the edgesof the fractured graphene. The broad peak around 3300–3600cm�1 is typical of O–H stretching, and is likely caused by theformation of alcohol or carboxylic acid groups. The C/O ratio ofmilled powder was approximately 25 up to 2 h of milling(Fig. 2f), but decreased to 15 aer 12 h of milling. Conversely,the C/O ratio of reduced graphene is usually less than 10,indicating that the milled powder retains its original highquality for short milling times. Based on a detailed survey of theC 1s spectrum, the C]O groups increased from 2% to 9%(Fig. S6 and Table S3†) aer 12 h of milling. Potassium was alsodetected by a detailed survey of the K 2p spectrum. These resultsindicate the interaction of graphite with potassium carbonateduring ball milling. However, the degree of functionalization ofgraphite produced by salt-assisted ball milling is far less thanthat of traditional graphene oxide; thus it is reasonable toconsider that the adsorption occurs only at the edge of thenanoplatelets.Liquid phase exfoliation of soluble graphiteTo produce high quality graphene, the ID/IG should be main-tained at less than 0.3—preferably less than 0.2, because the ID/IG value reects the defects and lateral size of graphene.42 Basedon our comprehensive work with salt-assisted ball milling, wesuccessfully produced soluble modied graphite with an ID/IG ofThis journal is © The Royal Society of Chemistry 20190.21 using ourmethod (ball milling for 20min using 3 wt. equiv.K2CO3). To check the solubility of graphite, the modiedgraphite with salt was added to various solvents and then mixedwith 5 min of sonication followed by 30 min of centrifugation.As shown in Fig. 3a, concentrated graphene dispersions wereobtained in not only NMP but also tetrahydrofuran (THF),water, methyl ethyl ketone (MEK), IPA, and acetone. Forexample, the graphene concentration was 0.6 mg ml�1 in THFwhen the initial concentration of milled powder was 3 mg ml�1.This means the yield of graphene reaches 20% aer only 5 minof sonication. The graphene concentration in acetone was 0.32mg ml�1, indicating a 10% yield. Preparing concentrated gra-phene dispersions with these solvents is typically regarded asimpossible because of the signicant difference in surfacetensions between the solvents and graphite. In fact, the gra-phene concentration was 0.0005 mg ml�1 in acetone when weuse milled graphite without salt (Table S1†). In spite of this, theyield of graphene reached 40% in NMP and 10–20% in low-boiling point solvents. Note that the yield for milled graphitewithout salt in a low-boiling point solvent is less than 0.1%,which strongly reinforces the assertion that the exfoliationefficiency of graphite is signicantly improved by salt-assistedball milling. However, the modied graphite could not beexfoliated in toluene and hexane, which are non-polar solvents.The surface tension of the solvent can be varied by control-ling the water/alcohol content. The cosolvent approach is oenapplied to nd the best surface tension of the solvent for exfo-liation of layeredmaterials. For milled graphite without salt, thebest surface tension was conrmed to be around 40 mN m�1(Fig. 3b), which is the same as that in previous reports.19,20 Thesoluble graphite (salt-functionalized) showed the same generaltrend as the milled graphite without salt, with a peak around 40mN m�1 followed by a gradual decrease. However, the solublegraphite showed higher yield values compared to the non-functionalized graphite for the entire range of surface tensionvalues. This indicates that the surface characteristics of gra-phene did not vary aer salt-assisted ball milling, which isconsistent with the Raman and XRD analysis (i.e., no functionalgroup on the basal plane).Following from the results of Fig. 3a and b, a more dominantparameter related to a stable dispersion of soluble graphiteshould be discussed, in addition to the surface tension of thesolvent. There are three factors to obtain a stable dispersion:compatibility of the solvent and solid (related to surfacetension), steric repulsion, and electrical repulsion. The rst twofactors are not expected in a low boiling point solvent. There-fore, it is presumed that the electrical repulsion in a liquid isenhanced by salt-assisted ball milling. The zeta potential ofgraphene in a wide range of pH values was measured (Fig. 3c).The milled graphite without salt exhibited �30 mV at pH 7,which is the limiting value for a stable dispersion. While thegeneral trend was again similar between graphite milled withand without salt, the absolute value of the zeta potential wasnoticeably increased by salt-assisted ball milling. The organicsalt—which is chemically bonded to the graphene edges—canbe dissociated in a polar solvent. When a cation such as sodiumor potassium is dissociated from the edge, negative charge ofNanoscale Adv., 2019, 1, 4955–4964 | 4959http://creativecommons.org/licenses/by/3.0/http://creativecommons.org/licenses/by/3.0/https://doi.org/10.1039/c9na00463gFig. 3 Liquid-phase exfoliation of soluble graphite. (a) Graphene yield of soluble graphite in various solvents. Yield of graphene exceeds 10% inpolar solvents (dielectric constant 3 > 5), though the concentrated dispersion cannot be obtained in nonpolar solvents (3 < 5). (b) Graphene yieldof salt-assisted milled powder and milled powder in an IPA/water cosolvent. The surface tension of the cosolvent is controlled by changing thewater content. (c) Zeta-potential of graphene as a function of pH. (d) Effect of sonication time on the graphene yield of anionic graphite. Thefitting curve is obtained using the power law equation.Nanoscale Advances PaperOpen Access Article. Published on 21 November 2019. Downloaded on 12/11/2019 1:26:43 AM.  This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.View Article Onlinethe graphite is enhanced. Dissociation of salts can be deter-mined by the conductivity of deionized water. The conductivityof deionized water increases when soluble graphite is added,whereas no conductivity change was observed for unmodiedgraphite (Fig. S7†). The mechanism of enhanced electricalrepulsion is that cations such as sodium and potassium diffuseinto the solvent, but most of them are attracted to the negativelycharged particle surface. In the vicinity of the particle surface,the electrical double layer is formed. The neutralization ofsurface charge by the attracted cations is incomplete due to thethermal motion of said cations. Therefore, electric eld aroundparticles is negatively charged, and this generates the electro-static repulsion between layers.In the case of electrical repulsion, the dielectric constant 3 ofthe liquid has a dominant effect on the stability of the disper-sion.43 If the dielectric constant 3 is sufficiently large (3 > 5),some electrolyte dissociation does occur which generates anelectrostatic repulsion force strong enough to stabilize thedispersion. However, if 3 < 5 then negligible dissociation occursand the electrical double layers are extended, making it difficultto obtain an electrostatically stable dispersion system. Thesephenomena explain why the soluble graphite could not beexfoliated in hexane (3¼ 1.9) or toluene (3¼ 2.3). These solventsare categorized as non-polar solvents, so an electrical repulsionforce is not expected.4960 | Nanoscale Adv., 2019, 1, 4955–4964The effect of sonication time on the yield of graphene wasalso investigated (Fig. 3d). The yield of graphene in the IPA/water cosolvent reached 80% (2.4 mg ml�1) aer 2 h of soni-cation, which is the highest value obtained for the LPE process.In the case of pure IPA, a yield of 20% was obtained aer 10 minof sonication. It was reported that it took 300 h to obtaina dispersion of few-layer graphene with a yield of 15% whennatural graphite was used.35 The yield of graphene in IPAgradually decreased aer more than 10 min of sonication; thereason for this phenomenon is not clear, but possibly aging ofIPA might occur during sonication.44The quality of dispersed graphene was determined by AFM,TEM, and Raman analyses. Surprisingly, the average thicknessof graphene in IPA and the IPA/water cosolvent was 0.63 and0.59 nm, respectively, which are the smallest values obtained forliquid exfoliated graphene (Fig. 4a, S8 and S9†). The thicknessof monolayer graphene measured by AFM depends on thesubstrate and measuring conditions, and the value is usually0.6–1 nm.18 We measured the thickness of over 200 graphenenanoplatelets, and none with a thickness above 0.8 nm could befound. Thus, we assumed that nanosheets dispersed in thesolvent were monolayer graphene. The proportion of mono-layers was 100%. Few-layer graphene was not observed in thisstudy, which is unusual because the centrifugation conditionswere mild (1500 rpm, 30 min) and few-layer graphene is usuallyThis journal is © The Royal Society of Chemistry 2019http://creativecommons.org/licenses/by/3.0/http://creativecommons.org/licenses/by/3.0/https://doi.org/10.1039/c9na00463gFig. 4 Quality of graphene after LPE using anionic graphite. (a) AFM image of graphene with its thickness profile. (b) Histogram of the graphenelength in the IPA/water cosolvent measured by AFM. (c) Typical Raman spectrum of graphene deposited on the filter paper. (d) TEM image ofrestacked graphene sheets. (e) The magnified image of the edge marked by the circle in (d).Paper Nanoscale AdvancesOpen Access Article. Published on 21 November 2019. Downloaded on 12/11/2019 1:26:43 AM.  This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.View Article Onlineobtained under these conditions. We presume that the repul-sive force generated by negative charge further contributes toexfoliation during sonication. The average length of graphenewas 366 nm for the IPA/water cosolvent and 227 nm for the IPAsolution (Fig. 4b and S10†). The IPA/water cosolvent is morecompatible with graphene than IPA, so clean exfoliationfrequently occurs rather than fragmentation during LPE. In anoptimal solvent such as that described above, the aspect ratio ofgraphene was 580. In the case of LPE using natural graphite, theaspect ratio of few-layer graphene is typically 50–100.18 Thishigh aspect ratio of soluble graphene will contribute to theimprovement of mechanical, thermal, and barrier properties ofpolymer nanocomposites.21,22 The quality of graphene exfoliatedand dispersed by the LPE process was also checked by Ramanspectroscopy, and the ID/IG value was around 0.22 by Ramananalysis (Fig. 4c), which indicates high quality graphene. Thevalue did not change even with 2 h of sonication (Fig. S11†).Thus, the quality of graphene is more inuenced by the ballmilling conditions than by the sonication time.It should be noted that the agglomeration of grapheneimmediately occurs as the solvent evaporates because there isno repulsive force without the solvent. The isolated graphenewas obtained on mica, but we could not directly observe exfo-liated graphene, only its agglomeration on the Si and SiO2substrate, presumably due to the weak interaction betweengraphene and substrates. With respect to TEM observation,graphene agglomeration was observed as shown in Fig. 4d. Onlya single dark line was observed at the edge of graphene (Fig. 4e),indicating that the agglomeration is composed of randomlystacked monolayer graphene. Energy dispersive X-ray spectros-copy was performed for graphene (Fig. 4e) to check the presenceThis journal is © The Royal Society of Chemistry 2019of potassium at the edge of graphite; approximately 0.7 weight%potassium was detected (Fig. S12†), while no potassium wasdetected at the basal plane of graphite (Fig. S13†). These resultsindicate that the graphite edges are indeed functionalized bysalt-assisted ball milling.The productivity of graphene obtained by a combination ofsalt-assisted ball milling and subsequent LPE is compared withthat obtained by other processes in the literature. It was re-ported that pyrene or melamine facilitated the exfoliation ofgraphite by p–p interaction with graphite.33–35 We have con-ducted the same milling process using pyrene or melamine asadditives during ball milling, and the graphene yield aer theLPE process was determined. The yield of graphene in acetoneaer centrifugation (1500 rpm, 30 min) was 0.01% for pyrene-assisted milled graphite and 1.03% for melamine-assistedmilled graphite, whereas the yield of salt-assisted milledgraphite exceeds 3.0% (Table S1†). These results indicate thatweak acid salts are the best choice for additives in ball milling toproduce soluble graphene.Molecular simulationIt should be claried how the active carbon at the graphiteedge reacts with the salts and why weak acid salts are moreefficient in improving solubility than strong acid salts. Toanswer the above question, we have performed a rst-princi-ples molecular simulation of a chemical reaction betweena salt molecule (K2CO3, K2SO4, CH3COOK, and KNO3; K may bechanged to Na) and a rectangular graphene fragment withthree sides terminated by hydrogen by using DMol3,45,46 inMaterials Studio with the DNP+ basis set and the meta-GGAfunctional SCAN47 in density functional theory. Aer structuralNanoscale Adv., 2019, 1, 4955–4964 | 4961http://creativecommons.org/licenses/by/3.0/http://creativecommons.org/licenses/by/3.0/https://doi.org/10.1039/c9na00463gFig. 5 Molecular simulation of the mechanochemical reaction at the graphene edge. (a and b) Front and side views of the graphene fragmentreacting with K2CO3. (c and d) Front and side views of the graphene fragment reacting with K2SO4. (e) Electrostatic potential map with the valueof the Hirshfeld charge for graphene reacting with K2CO3. (f) Electrostatic potential map with the value of the Hirshfeld charge for graphenereacting with K2SO4.Nanoscale Advances PaperOpen Access Article. Published on 21 November 2019. Downloaded on 12/11/2019 1:26:43 AM.  This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.View Article Onlineoptimization, we found that a –CO2, –SO2, or –NO2 base isadsorbed on top of one edge carbon atom in a “Y”-shapeperpendicular to the graphene plane (see Fig. 5a–d for K2CO3and K2SO4 and Fig. S14a–d† for CH3COOK and KNO3; see alsothe green circle in Fig. S15a–d†) with an adsorption energymore than 5 eV (see Table S4†) indicating that the nalstructure is energetically very stable. The electrostatic poten-tial felt by each electron is plotted in Fig. 5e and f for K2CO3and K2SO4 and Fig. S14e and f† for CH3COOK and KNO3together with the value of the Hirshfeld charge, which isdened as the difference between the molecular and unrelaxedatomic charge. Obviously, the electrostatic potential of weakacid salts (blue region) is much lower than that of strong acidsalts (yellow region). Additionally, the Hirshfeld charge of theY-shaped base is much more negative for weak acid salts(typically ��0.5) than for strong acid salts (typically ��0.1).This difference is clearly due to the difference between thebase atoms (C, S, or N) of the Y-shaped adsorbent; a –CO2 baseis more negatively charged than a –SO2 or –NO2 base. Thismeans that the original bonding nature between the alkali ionand the base molecule is retained in the case of weak acidsalts, but the alkali atom ion is more strongly bonded to thegraphene edge in the case of strong acid salts (see Table S4† forthe Hirshfeld charge of the Y-shaped base and Table S5† forthe bond length). Therefore, the alkali ions are more easilydissociated in aqueous solution and form the electrical doublelayer. This fact corresponds to the negatively large zetapotential of the ball milled graphene sample in aqueoussolution in Fig. 3c.4962 | Nanoscale Adv., 2019, 1, 4955–4964ConclusionsWe have found a new mechanochemical reaction for theproduction of high-quality soluble graphene, whereby the acti-vated carbon radicals induced by fragmentation react withorganic salts in a pathway for edge-functionalization. Ingeneral, salt is used to precipitate a stable colloid by weakeningthe electrical repulsion of the colloid via ionic dissociation.However, if the salt is chemically bonded to the colloid, it actsinstead as a dispersing agent. Dissociation of salts from theparticle increases the negative charge of particles in the liquid.Thus, the salt-modied graphite can be exfoliated in low-boilingpoint solvents, which is commonly believed to be impossible forachieving stable dispersions. By using soluble graphite with theLPE method, essentially monolayer graphene can be obtainedwith a high yield (more than 10%) with just a few minutes ofsonication. The process requires only common chemicals suchas carbonate and acetate, which are cheap and environmentallyfriendly. Moreover, the process is simple (milling and washing)and scalable; thus the new mechanochemical route proposedhere will be a key technology for mass production and effectiveusage of attractive nanomaterials.MethodsNatural graphite (Sigma-Aldrich, d ¼ 500 mm) was used formodication by planetary ball milling (P-6, Fritsch). In mostcases, seven steel balls with a diameter of 20 mm were placed inan 80 ml container. The rotation speed was controlled at 500This journal is © The Royal Society of Chemistry 2019http://creativecommons.org/licenses/by/3.0/http://creativecommons.org/licenses/by/3.0/https://doi.org/10.1039/c9na00463gPaper Nanoscale AdvancesOpen Access Article. Published on 21 November 2019. Downloaded on 12/11/2019 1:26:43 AM.  This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.View Article Onlinerpm. The milled powder was washed several times until thewater pH became neutral. The washed samples were driedovernight and were ground in a mortar to obtain ne powder.Typically, 0.3 g of the powder was added into 100 ml of a solventand probe-sonication (UH-600, SMT) was conducted to exfoliatethe powder. The obtained dispersion was poured into a 50 mlcentrifugation tube. The dispersion was centrifuged (model2420, Kubota) at 1500 rpm for 30 min. The top half of thedispersion was carefully extracted using a pipette and stored foruse. The characterization methods are all described in detail inthe ESI.†Conflicts of interestThere are no conicts to declare.AcknowledgementsThis work was supported by JSPS KAKENHI Grant Number15H05504 and the Fujikura Foundation. We acknowledge theCenter for Advanced Materials Analysis at the Tokyo Institute ofTechnology for TEM and Raman analysis.Notes and references1 C. S. Boland, U. Khan, G. Ryan, S. Barwich, R. Charifou,A. Harvey, C. Backes, Z. Li, M. S. Ferreira, M. E. Mobius,R. J. Young and J. N. Coleman, Science, 2016, 354, 1257–1260.2 A. G. Kelly, T. Hallam, C. Backes, A. Harvey, A. S. Esmaeily,I. Godwin, J. Coelho, V. Nicolosi, J. Lauth, A. Kulkarni,S. Kinge, L. D. Siebbeles, G. S. Duesberg andJ. N. Coleman, Science, 2017, 356, 69–73.3 Z. Liu, K. Parvez, R. Li, R. Dong, X. Feng and K. Mullen, Adv.Mater., 2015, 27, 669–675.4 S. Pourhashem, M. R. Vaezi, A. Rashidi andM. R. Bagherzadeh, Corros. Sci., 2017, 115, 78–92.5 C. Chen, S. Qiu, M. Cui, S. Qin, G. Yan, H. Zhao, L. Wang andQ. Xue, Carbon, 2017, 114, 356–366.6 L. Tan, C.Wang, M. Zeng and L. Fu, Small, 2017, 13, 1603337.7 S. Bae, H. Kim, Y. Lee, X. Xu, J. S. Park, Y. Zheng,J. Balakrishnan, T. Lei, H. R. Kim, Y. I. Song, Y. J. Kim,K. S. Kim, B. Ozyilmaz, J. H. Ahn, B. H. Hong and S. Iijima,Nat. Nanotechnol., 2010, 5, 574–578.8 K. S. Novoselov, A. K. Geim, S. V. Morozov, D. Jiang, Y. Zhang,S. V. Dubonos, I. V. Grigorieva and A. A. Firsov, Science, 2004,306, 666–669.9 N. E. Sorokina, M. A. Khaskov, V. V. Avdeev andI. V. Nikol'skaya, Russ. J. Gen. Chem., 2005, 75, 162–168.10 K. Kalaitzidou, H. Fukushima and L. T. Drzal, Carbon, 2007,45, 1446–1452.11 L. Paliotta, G. De Bellis, A. Tamburrano, F. Marra, A. Rinaldi,S. K. Balijepalli, S. Kaciulis and M. S. Sarto, Carbon, 2015, 89,260–271.12 D. D. L. Chung, J. Mater. Sci., 2015, 51, 554–568.13 D. C. Marcano, D. V. Kosynkin, J. M. Berlin, A. Sinitskii,Z. Sun, A. Slesarev, L. B. Alemany, W. Lu and J. M. Tour,ACS Nano, 2010, 4, 4806–4814.This journal is © The Royal Society of Chemistry 201914 A. M. Dimiev and J. M. Tour, ACS Nano, 2014, 8, 3060–3068.15 C. T. J. Low, F. C. Walsh, M. H. Chakrabarti, M. A. Hashimand M. A. Hussain, Carbon, 2013, 54, 1–21.16 Y. L. Zhong, Z. Tian, G. P. Simon and D. Li, Mater. Today,2015, 18, 73–78.17 Y. Hernandez, V. Nicolosi, M. Lotya, F. M. Blighe, Z. Sun,S. De, I. T. McGovern, B. Holland, M. Byrne, Y. K. Gun'Ko,J. J. Boland, P. Niraj, G. Duesberg, S. Krishnamurthy,R. Goodhue, J. Hutchison, V. Scardaci, A. C. Ferrari andJ. N. Coleman, Nat. Nanotechnol., 2008, 3, 563–568.18 K. R. Paton, E. Varrla, C. Backes, R. J. Smith, U. Khan,A. O'Neill, C. Boland, M. Lotya, O. M. Istrate, P. King,T. Higgins, S. Barwich, P. May, P. Puczkarski, I. Ahmed,M. Moebius, H. Pettersson, E. Long, J. Coelho,S. E. O'Brien, E. K. McGuire, B. M. Sanchez,G. S. Duesberg, N. McEvoy, T. J. Pennycook, C. Downing,A. Crossley, V. Nicolosi and J. N. Coleman, Nat. Mater.,2014, 13, 624–630.19 J. Shen, J. Wu, M. Wang, P. Dong, J. Xu, X. Li, X. Zhang,J. Yuan, X. Wang, M. Ye, R. Vajtai, J. Lou and P. M. Ajayan,Small, 2016, 12, 2741–2749.20 M. Wang, X. Xu, Y. Ge, P. Dong, R. Baines, P. M. Ajayan,M. Ye and J. Shen, ACS Appl. Mater. Interfaces, 2017, 9,9168–9175.21 R. J. Young, I. A. Kinloch, L. Gong and K. S. Novoselov,Compos. Sci. Technol., 2012, 72, 1459–1476.22 D. Pierleoni, Z. Y. Xia, M. Christian, S. Ligi, M. Minelli,V. Morandi, F. Doghieri and V. Palermo, Carbon, 2016, 96,503–512.23 H. Tao, Y. Zhang, Y. Gao, Z. Sun, C. Yan and J. Texter, Phys.Chem. Chem. Phys., 2017, 19, 921–960.24 A. Ciesielski and P. Samori, Chem. Soc. Rev., 2014, 43, 381–398.25 Y. Arao, Y. Mizuno, K. Araki and M. Kubouchi, Carbon, 2016,102, 330–338.26 A. E. Del Rio Castillo, V. Pellegrini, A. Ansaldo,F. Ricciardella, H. Sun, L. Marasco, J. Buha, Z. Dang,L. Galiani, E. Lago, N. Curreli, S. Gentiluomo, F. Palazon,M. Prato, R. Oropesa-nuňez, P. S. Toth, E. Mantero,M. Crugliano, A. Gamucci, A. Tomadin, M. Polini andF. Bonaccorso, Mater. Horiz., 2018, 5, 890–907.27 P. C. Karagiannidis, S. A. Hodge, L. Lombardi, F. Tomarchio,N. Decorde, S. Milana, I. Goykhman, Y. Su, S. V. Mesite,D. N. Jojmstone, R. K. Leary, P. A. Midgley, N. M. Pugno,F. Torrisi and A. C. Ferrari, ACS Nano, 2017, 11, 2742–2755.28 I. Y. Jeon, H. J. Choi, S. M. Jung, J. M. Seo, M. J. Kim, L. Daiand J. B. Baek, J. Am. Chem. Soc., 2013, 135, 1386–1393.29 I.-Y. Jeon, S.-Y. Bae, J.-M. Seo and J.-B. Baek, Adv. Funct.Mater., 2015, 25, 6961–6975.30 O. Y. Posudievsky, O. A. Khazieieva, V. V. Cherepanov,V. G. Koshechko and V. D. Pokhodenko, J. Nanopart. Res.,2013, 15, 2046.31 O. Y. Posudievsky, O. A. Khazieieva, V. G. Koshechko andV. D. Pokhodenko, Theor. Exp. Chem., 2014, 50, 103–109.32 T. Xing, S. Mateti, L. H. Li, F. Ma, A. Du, Y. Gogotsi andY. Chen, Sci. Rep., 2016, 6, 35532.Nanoscale Adv., 2019, 1, 4955–4964 | 4963http://creativecommons.org/licenses/by/3.0/http://creativecommons.org/licenses/by/3.0/https://doi.org/10.1039/c9na00463gNanoscale Advances PaperOpen Access Article. Published on 21 November 2019. Downloaded on 12/11/2019 1:26:43 AM.  This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.View Article Online33 V. Leon, J. M. Gonzalez-Dominguez, J. L. Fierro, M. Prato andE. Vazquez, Nanoscale, 2016, 8, 14548–14555.34 V. Leon, A. M. Rodriguez, P. Prieto, M. Prato and E. Vazquez,ACS Nano, 2014, 8, 563–571.35 M. Buzaglo, I. P. Bar, M. Varenik, L. Shunak, S. Pevzner andO. Regev, Adv. Mater., 2017, 29, 1603528.36 P. Dash, T. Dash, T. K. Rout, A. K. Sahu, S. K. Biswal andB. K. Mishra, RSC Adv., 2016, 6, 12657–12668.37 A. E. D. Mahmoud, A. Stolle and M. Stelter, ACS SustainableChem. Eng., 2018, 6, 6358–6369.38 A. Eckmann, A. Felten, A. Mishchenko, L. Britnell,R. Krupke, K. S. Novoselov and C. Casiraghi, Nano Lett.,2012, 12, 3925–3930.39 K. Parvez, Z. S. Wu, R. Li, X. Liu, R. Graf, X. Feng andK. Mullen, J. Am. Chem. Soc., 2014, 136, 6083–6091.4964 | Nanoscale Adv., 2019, 1, 4955–496440 H. Feng, R. Cheng, X. Zhao, X. Duan and J. Li, Nat. Commun.,2013, 4, 1539.41 N. Iwashita, C. R. Park, H. Fujimoto, M. Shiraishi andM. Inagaki, Carbon, 2004, 42, 701–714.42 U. Khan, A. O'Neill, M. Lotya, S. De and J. N. Coleman, Small,2010, 6, 864–871.43 P. C. Van Der Hoeven and J. Lyklema, Adv. Colloid InterfaceSci., 1992, 42, 205–277.44 A. O'Neill, U. Khan, P. N. Nirmalraj, J. Boland andJ. N. Coleman, J. Phys. Chem. C, 2011, 115, 5422–5428.45 B. Delley, J. Chem. Phys., 1990, 92, 508–517.46 B. Delley, J. Chem. Phys., 2000, 113, 7756–7764.47 J. Sun, A. Ruzsinszky and J. P. Perdew, Phys. Rev. Lett., 2015,115, 036402.This journal is © The Royal Society of Chemistry 2019http://creativecommons.org/licenses/by/3.0/http://creativecommons.org/licenses/by/3.0/https://doi.org/10.1039/c9na00463g Mass production of low-boiling point solvent- and water-soluble graphene by simple salt-assisted ball millingElectronic supplementary information (ESI) available. See DOI: 10.1039/c9na00463g Mass production of low-boiling point solvent- and water-soluble graphene by simple salt-assisted ball millingElectronic supplementary information (ESI) available. See DOI: 10.1039/c9na00463g Mass production of low-boiling point solvent- and water-soluble graphene by simple salt-assisted ball millingElectronic supplementary information (ESI) available. See DOI: 10.1039/c9na00463g Mass production of low-boiling point solvent- and water-soluble graphene by simple salt-assisted ball millingElectronic supplementary information (ESI) available. See DOI: 10.1039/c9na00463g Mass production of low-boiling point solvent- and water-soluble graphene by simple salt-assisted ball millingElectronic supplementary information (ESI) available. See DOI: 10.1039/c9na00463g Mass production of low-boiling point solvent- and water-soluble graphene by simple salt-assisted ball millingElectronic supplementary information (ESI) available. See DOI: 10.1039/c9na00463g Mass production of low-boiling point solvent- and water-soluble graphene by simple salt-assisted ball millingElectronic supplementary information (ESI) available. See DOI: 10.1039/c9na00463g Mass production of low-boiling point solvent- and water-soluble graphene by simple salt-assisted ball millingElectronic supplementary information (ESI) available. See DOI: 10.1039/c9na00463g Mass production of low-boiling point solvent- and water-soluble graphene by simple salt-assisted ball millingElectronic supplementary information (ESI) available. See DOI: 10.1039/c9na00463g Mass production of low-boiling point solvent- and water-soluble graphene by simple salt-assisted ball millingElectronic supplementary information (ESI) available. See DOI: 10.1039/c9na00463g