Hyperstage Graphite: Electrochemical Synthesis and Spontaneous Reactive Exfoliation

Abstract

Covalent modification of the π-electron basal planes of graphene enables the formation of new materials with enhanced functionality. An electrochemical method is reported for the formation of what is referred to as a Hyperstage-1 graphite intercalation compound (GIC), which has a very large interlayer spacing d₀₀₁ > 15.3 Å and contains disordered interstitial molecules/ions. This material is highly activated and undergoes spontaneous exfoliation when reacted with diazonium ions to produce soluble graphenes with high functionalization densities of one pendant aromatic ring for every 12 graphene carbons. Critical to achieving high functionalization density is the Hyperstage-1 GIC state, a weakening of the van der Waals coupling between adjacent graphene layers, and the ability of reactants to diffuse into the disordered intercalate phase between the layers. Graphene functionalization with 3,5-dinitrophenyl groups provides for exceptional dispersibility (0.24 mg mL⁻¹) in N,N-dimethylformamide and for conjugation with amines.

Advances in graphene science and technology depend critically on manipulating new forms of graphene into functional materials. Chemical methods are attractive because they can tailor the physical, electronic, mechanical, thermal, and/or transport properties of graphenes for specific applications.^1–4 Many functionalization schemes take advantage of defect sites and oxygen groups on graphene oxide (GO) or reduced graphene oxide (rGO).^5–7 However, the highly degraded carbon framework of GO and rGO ultimately limits the performance of these materials. Therefore, efficient methods for the bulk production of high-quality functionalized graphenes are needed. Graphite can be activated by reducing the intersheet van der Waals attractions by inserting atomic or molecular layers of alkali metals, acids, or FeCl₃, between the graphene sheets. Despite the application of harsh chemical and thermal conditions, only partial graphite intercalation compounds (GICs) are typically produced in these methods.^8–10 Proper analysis of these processes requires rigorous studies of the crystal structure of GICs. Although the use of GICs as activated intermediates in chemical functionalization has enjoyed success, typical exfoliation schemes require prolonged high power ultrasonication, which degrades the graphene.^11–13 Additionally, the resulting graphene dispersions are often meta-stable and re-aggregation generally occurs.^{14, 15}

We demonstrate, herein, a gentle room temperature method for reducing the van der Waals coupling between graphene sheets (d₀₀₁ > 15.3 Å) and subsequent spontaneous exfoliation to give soluble functionalized graphene. This study is not the first demonstration of electrochemical activation and functionalization of graphene; however, previous investigations,¹⁶ including those from our own group, failed to quantitatively produce pure isolated sheets, giving instead inseparable mixtures that contain some multilayered graphenes. Our methods involve the intercalation of tetrabutylammonium (TBA⁺) into the graphene galleries that electrostatically balance the negatively charged π-electron system created electrochemically. Increased d₀₀₁ spacing in Stage-1 GIC reduces van der Waals interactions between graphenes. However, as we described, the crystalline organization of the TBA⁺ ions¹⁷ restricts reactant diffusion into the network and thereby reduces reactivity. At high negative potentials, it is possible to both increase the intercalation of large density of TBA⁺ ions and introduce amines by electrolytic decomposition of the TBA⁺. The result is a further expanded gallery with a disordered interstitial phase between the basal planes to give what we refer to as Hyperstage-1 GICs. The result has great utility because every graphene layer is accessible to react with diazonium salts in this process and the graphene spontaneously exfoliates without any deliberate mixing or sonication. The diazonium reaction of the Hyperstage-1 GIC results in exceptionally high densities of functional groups on the graphene. We further demonstrate that by functionalization with 3,5-dinitrophenyl (3,5-DiNP) groups enables the formation of Meisenheimer complexes between functionalized graphene and n-butylamine.

Our goal is to create extended functionalized graphenes with minimal defects in the hexagonal graphene lattice. As a result, highly oriented pyrolytic graphite (HOPG) was chosen as a high-purity graphite source. In contrast to randomly oriented graphite, HOPG conserves the monolithic structure and electrical connectivity of the graphite domains throughout the necessary expansion that accompanies the electrochemical generation of the different GICs. The electrochemically driven intercalation process is depicted in Figure 1. In this scheme, high reducing potentials of HOPG result in TBA⁺ intercalation between the graphene sheets. The solvent is key to this process, and we find that a mixture of acetonitrile (MeCN) and dimethylformamide (DMF) enhances this process over our previous procedures.¹⁶ A continuous electrochemical potential ramp maintains a driving force for full intercalation of TBA⁺, which is accompanied by a dramatic volumetric expansion of HOPG.

Figure 1.

Illustration of the generation of a Hyper-3-Stage-1 GIC by electrochemical intercalation of TBA⁺ ions and amine generation. This periodic solid reacts with diazonium ions and spontaneously exfoliates (disperses) to form solutions of functionalized graphenes.

In this process, we identified five voltage ranges (I, II, III, IV, and V in Figure 2) that correspond to distinct phases of GIC staging. These include the Stage-1, Hyper-1-Stage-1, Hyper-2-Stage-1, and Hyper-3-Stage-1, which, as shown in Figure 2, have varying degrees of ion intercalation and intersheet spacing. Since the electromigration of the larger molecule TBA⁺ in graphene galleries was predictably sluggish, we experimentally observed specific current peaks associated with changes in staging only with extremely slow voltage sweeps (−40 μV s⁻¹).

Figure 2.

Voltage sweep of HOPG working electrode with a graphite counter electrode and Ag/Ag⁺reference electrode, in 1.5 m tetrabutylammonium perchlorate (TBAP) in MeCN/DMF, recorded at a scan rate of −40 μV s⁻¹ at 25 °C. The structural models of the associated GICs illustrate the different levels of intercalation and interlayer spacing. Linear voltage ramping conditions: Stage-1 (−40 μV s⁻¹), Hyper-1-Stage-1 (−40 μV s⁻¹), Hyper-2-Stage-1 (−2 μV s⁻¹), and Hyper-3-Stage-1 GICs (−3 μV s⁻¹).

The different GIC stages were identified by X-ray diffraction (XRD), which reveals different crystallographic lattice arrangements of intercalants and graphite layers along the c-axis (Figure 3).¹⁸ As intercalation takes place, the graphite (002) peak vanishes and new peaks appear. The dominant diffractions in Stage-1 GIC appear in the (00ℓ) reflections are a direct measure of the interlayer distance (d₀₀₁ ≈ 8.17 Å). The introduction of the intercalated layer weakens the attractive potential between graphite layers; however, the material is still relatively inert as a result of the ordered intercalation ions. The interlayer distance is in agreement with the expected size of TBA⁺ in a flattened conformation (≈4.8 Å) within the graphene galleries.¹⁹ With increasing voltage, the graphene/TBA⁺ matrix exhibits an increasing d₀₀₁ spacing (3.35 Å → 8.17 Å → 12.70 Å → 15.30 Å → disordered) (Figures S1 and S2, Supporting Information). The structural disorder of the intercalation phase is highest for Hyper-3-Stage-1 GIC, and we attribute this to the reductive decomposition of TBA⁺ ions within graphene galleries.²⁰ Our hypothesis is that reductive fragmentation of TBA⁺ ions decomposes into tributylamine, butene, and alkanes (Figures S3 and S4, Supporting Information). These fragments disrupt much of the crystalline organization of the ions, which, as will be discussed, facilitate functionalization. However, we still observe minor diffraction peaks that are ascribed to the electrolyte.

Figure 3.

XRD diffraction data of HOPG at different levels of electrochemical activation. Stage-1, Hyper-1-Stage-1, Hyper-2-Stage-1, and Hyper-3-Stage-1 GICs and functionalized graphene. The additional peaks, not related to graphene stacking, appeared at around 8.4° and 8.7° are assigned to the organized TBA⁺ domains in Hyper-2-Stage-1 and Hyper-3-Stage-1 GICs.

The disorder in the Hyper-3-Stage-1 GIC produces greatly enhanced reactivity and transferring this material as an activated electrode to a solution containing 0.1 M 3,5-dinitrobenzenediazonium tetrafluoroborate (3,5-DiNBD) and 1 m tetrabutylammonium perchlorate (TBAP) in MeCN with a negative potential ramp from −1.2 to −2.0 V (vs Ag/Ag⁺, −10 μV s⁻¹) results in efficient reductive functionalization (Figure S5, Supporting Information). Under these conditions, we observe remarkable behavior with the graphite undergoing spontaneous exfoliation to give MeCN solutions of soluble functionalized graphene. XRD data of the purified functionalized graphene revealed a completely amorphous material that lacks typical intergraphene sheet diffractions (Figure 3). This is unusual for a functionalized graphene, and we attribute this to the high degrees of functionalization that tend to prevent the sheets from organizing into 2D structures.

Raman spectroscopy was used to characterize HOPG, Stage-1, Hyper-2-Stage-1, Hyper-3-Stage-1 GICs, exfoliated graphene, and functionalized graphene (Figure 4a).²¹ The D- and 2D-bands displayed similar shifts with TBA⁺ intercalation and, in transition from HOPG to Hyper-2-Stage-1 GICs, these peeks gradually shifted to lower frequencies (−41.1 cm⁻¹ for D-band and −74.1 cm⁻¹ for 2D-band). These shifts are indicative of n-doping resulting from electrochemical reduction of the graphene with concurrent TBA⁺ intercalation (Figure S6, Supporting Information).²² The D- and 2D-bands broaden with increasing TBA⁺ density, and the G-band splits and shifts upon formation of different staged compounds (Figures S6 and S7, Supporting Information).²³ The G-band depends strongly on the charge carrier densities, and the observed shift to higher frequencies is consistent with recent reports on electron doping by electrochemical gating.^{24, 25} The G-band of graphite appears at 1580 cm⁻¹ and shifts to 1603.0–1605.1 cm⁻¹ in Stage-1/Hyper-2-Stage-1-GICs, respectively (Figure 4a; Figure S6, Supporting Information). The G-band of Hyper-3-Stage-1 GIC also exhibits complexity with multiple overlapped peaks that can be deconvoluted into four distinct peaks (Figure 4a; Figure S6, Supporting Information). These peaks for Hyper-3-Stage-1 GIC can be assigned to n-doping with TBA⁺ intercalation as well as the presence of highly undulated graphene layers.^26–28

Figure 4.

Comparison of the Raman and XPS spectra of HOPG, GICs, exfoliated graphene, and functionalized graphene. a) Measured ex situ Raman data points (D-, G-, and 2D-bands) of HOPG, Stage-1, Hyper-2-Stage-1, Hyper-3-Stage-1 GICs, exfoliated graphene and functionalized graphene. b) High-resolution XPS spectra (black circles) of HOPG, Stage-1, Hyper-2-Stage-1, Hyper-3-Stage-1 GICs, exfoliated graphene, and functionalized graphene (magenta: Lorentzian/Gaussian peak fitting, red: convoluted line, and green: Shirley-based line) in the region of C 1s. High-resolution N 1s spectra of c) Hyper-3-Stage-1 GIC, d) 3,5-dinitrobenzenediazonium as a reference, and e) functionalized graphene with 3,5-dinitrophenyl (3,5-DiNP) groups. f) Photographs showing functionalized graphene with 3,5-DiNP groups on a polytetrafluoroethylene filter membrane (left) and in MeCN (right).

We confirmed that electrochemical intercalation does not result in covalent functionalization of the graphene sheets by analyzing the exfoliated (ultrasonication) and purified (washed) material. Specifically, we find that the intensities of the Raman I_D/I_G and I_D/I_D′ bands of exfoliated graphene from Hyper-3-Stage-1 GIC are 0.3 and 2.4, respectively (Figure 4a; Figure S8, Supporting Information). These exfoliated defect-free graphenes readily reassemble into stacked sheets when isolated, and the powders display the characteristic (002) reflection of graphite (Figure S9, Supporting Information). As a result, the electrochemical activation does not lead to irreversible functionalization of the graphene sheets. In contrast, after spontaneous exfoliation of functionalized graphene with 3,5-DiNP groups from Hyper-3-Stage-1 GIC, the Raman spectra indicate extensive covalent functionalization on the basal plane. The dramatically broadened lines of the D- and G-bands and the I_D/I_G ratio (≈0.71) confirm the introduction of a high percentage of sp³-hybridized carbon atoms into the sp²-hybridized graphene (Figure 4a; Figure S6, Supporting Information).

X-ray photoelectron spectroscopy (XPS) provides additional insight into the chemical nature of the GICs and functionalized graphene. The deconvolution of the C 1s peak reveals the presence of sp² C–C (≈284.5 eV), sp³ C–C/C–OH (≈285.5 eV), C–O/C–N⁺ (≈286.3 eV), C=O (287.6 eV), and a shake-up peak (≈291 eV) in Figure 4b.^{29, 30} The relative intensities of the different sp³ component peaks (≈285.5 and ≈286.3 eV) in Stage-1, Hyper-2-Stage-1, and Hyper-3-Stage-1 GICs are all consistent with the amount of intercalated TBA⁺.31 The analysis of the nitrogen peaks is also relevant, and Hyper-2-Stage-1 and Hyper-3-Stage-1 GICs provide a pure intercalated phase with a peak at 402.8 eV corresponding to the N 1s of the quaternary TBA⁺ (Figure S10, Supporting Information). The additional N 1s peaks at 400.3 and 398.4 eV in Hyper-3-Stage-1 GIC are attributed to TBA⁺ reduction products and solvent molecules (MeCN and DMF) in the graphene galleries (Figure 4c; Figure S10, Supporting Information).^{30, 32}

XPS of graphene functionalized with 3,5-DiNP after purification produces an exceptionally strong N 1s peak in the XPS spectra. The C/N ratio of ≈8.5 reveals an extraordinarily high functionalization density of one 3,5-DiNP group per 12 carbon atoms (Figure S6C, Supporting Information). A high-resolution N 1s analysis reveals three peaks centered at 398.5, 400.5, and 404.7 eV (Figure 4d,e).^33–35 The major peak at higher binding energy (404.7 eV) is assigned to the nitrogen of the nitro groups, thereby confirming the presence of 3,5-DiNP groups attached to the graphene (Figures S10 and S11, Supporting Information). The broad and lower binding energy N 1s peaks at 400.5 and 398.5 eV are assigned to nitrogens from reduction of the nitro groups produced under the electrochemical conditions. Specifically, we expect that reduction of the nitro groups will give rise to Ph–NHOH and/or Ph–NH₂units. These processes are also facilitated by the H atoms that can be extracted from the solvent and the products generated by the reductive decomposition of the TBA⁺ electrolyte. The XRD data reveal the reason for this lack of reactivity, and peaks associated with ordered TBA⁺ ions in the gallery likely provide for a rigid network that blocks diffusion of other reagents into the interior of the material. The introduction of amine reduction products in Hyper-3-Stage-1 GIC, confirmed by the XPS N 1s peaks, produces disorder in the graphene galleries thereby allowing diazonium reagents to diffuse into the galleries (Figure 2f).

The functionalized graphene sheets were investigated by transmission electron microscopy (TEM) and atomic force microscopy (AFM) in order to obtain additional microscopic evidence of the covalent functionalization. Exfoliated single-layer graphene produced by ultrasonic exfoliation of the Hyper-3-Stage-1 GIC exhibits a single set of sharp peaks associated with a graphitic hexagonal diffraction pattern (Figure 5a). The relative intensity ratio of I_{1100}/I_{2110}is ≈1.35, which is in agreement with previous reports.³⁶ In the case of our functionalized graphene, the surface of the graphene sheets appears to have some heterogeneity suggesting that sections are heavily covered with pendant phenyl groups (Figure 5b) and others may have lower functionalization. In spite of the functionalization, the sample still displays hexagonal crystalline domains; however, these domains are distorted with d_{1100} = 2.12–2.28 Å as compared to graphene with d_{1100} = 2.13 Å as determined by electron diffraction. An average thickness of ≈4.3 nm was observed for functionalized graphenes according to the AFM profile (Figure 5c). Figure 5d shows a histogram of thickness of functionalized graphene and reveals two distinct thicknesses of 2.4 and 4.4 nm. These heights correspond to two-sided functionalized single- and double-layer graphene structures, respectively, and suggest that our method specifically produces dominantly single-/double-layer exfoliations.³⁷ The double-layer structures may result from postfunctionalization aggregation. AFM analysis provides an average sheet area of 0.032 μm², which is smaller than the average domain size of the precursor graphite (intraplanar microcrystallic size: 1–10 μm) (Figure 5e; Figure S12, Supporting Information).³⁸

Figure 5.

TEM and AFM characterizations of graphenes. a) TEM image and the corresponding electron diffraction pattern of graphene produced by ultrasonic assisted dispersion of Hyper-3-stage-1 GIC. The crystal lattice retains its characteristic hexagonal diffraction peaks (1–210), (0–110), (–1010), and (–2110) diffractions. b) TEM image of functionalized graphene produced from diazonium ion reactions with Hyper-3-stage-1 GIC. This material shows apparent domains of functionalization but it still maintains a {1100} diffraction pattern. c) Topographic image on mica by AFM of functionalized graphene and a height profile measured along the white dashed line. d,e) The height distribution and particle size of the functionalized graphenes, respectively.

We seek to take advantage of the ability of the electron-deficient 3,5-DiNP groups in our functionalized graphene to form Meisenheimer complexes with n-butylamine. A schematic illustration of Meisenheimer complex formation from 3,5-DiNP-functionalized graphene and n-butylamine is shown in Figure 6a. When compared with the functionalized graphene with 3,5-DiNP groups, the Meisenheimer complex graphene provides improved dispersion stability in MeCN (Figure 6b). We noticed that graphene dispersed with n-butylamine does not form a stable dispersion and the attachment of this molecule via a Meisenheimer complex increases the graphene’s dispersibility in MeCN to give indefinitely stable solutions at a concentration of 0.9 mg mL⁻¹. The UV–vis absorption spectra of exfoliated graphene, functionalized graphene, and Meisenheimer complex graphene dispersions in MeCN (Figure 6c) support the proposed structures. The characteristic feature at 270 nm corresponds to a π–π* plasmon peak where van Hove singularities occur.³⁹ For the functionalized graphene, this peak blueshifts to ≈195.5 nm, suggesting that the electronic conjugation within the graphene is severely restricted by the sp³ defects.^{37, 40, 41} After functionalization, a new absorbance peak associated with the 3,5-DiNP group appears at 249.1 nm (Figure 6c; Figure S13, Supporting Information). The strongly electron-withdrawing nitro groups in 3,5-dinitrobenzene favor Meisenheimer complexes with primary amines.⁴² This chemical characteristic is preserved in 3,5-DiNP-functionalized graphene, and Figure 6a illustrates the formation of Meisenheimer complex with n-butylamine. The Meisenheimer complex graphene produces only small variations in the optical spectra (Figure 6c); however, the complex is confirmed by the weakening of the NO₂ infrared bands with a −7.9 cm⁻¹ shift on ʋ_asym(N–O) and −1.5 cm⁻¹ on ʋ_sym(N–O) (Figure 6d). The capability of functionalized graphene to form Meisenheimer complex provides a new method to control electronic structure and conjugate to biologically relevant species.

Figure 6.

Meisenheimer complex graphenes produced by reactions with amines. a) Proposed mechanism for the formation of Meisenheimer complex graphene, wherein pendant 3,5-DiNP groups react with n-butylamine (R = n-butyl). b) The photographs of MeCN solutions of functionalized graphene with 3,5-DiNP groups and the Meisenheimer complex graphene (dispersions after 1 d). c) Normalized absorption spectrum of exfoliated graphene, graphene with 3,5-DiNP groups, and the n-butylamine Meisenheimer complex graphene in MeCN. The π→π* plasmon peak of the graphene is blueshifted (yellow arrow: −74.5 nm) after the functionalization. d) Attenuated total reflectance–Fourier transform infrared spectroscopy (ATR-FTIR) spectra of functionalized graphene with 3,5-DiNP groups and formation of the Meisenheimer complex graphene results in changes in the NO₂ bands.

To conclude, we have developed a method for spontaneous exfoliation of highly functionalized graphene directly from Hyper-3-Stage-1 GIC by reaction with an aryl diazonium salt solution under electrochemical reducing conditions. The successful covalent functionalization of the sp² carbon network of graphene with one group per 12 graphene C was obtained by first weakening the van der Waals attractions between graphene sheets by cation–π interacted GICs. Key to this process is a reduction of the organization (crystallinity) of the intercalated ions, which is accomplished by partial reductive decomposition of the TBA⁺ cations. Highly expanded graphenes, characterized as Hyper-3-Stage-1 GIC (d₀₀₁spacing > 15.3 Å), were produced without any evidence of creating new covalent defects that disrupt the sp² lattice. The different intermediate Hyper-Stage-1 GICs were unambiguously characterized by XRD, Raman spectroscopy, TEM, and XPS. Raman analysis further confirmed the conversion of delocalized graphene sp² states to localized sp³ bonds with the functionalization with 3,5-DiNP groups. The formation of a Meisenheimer complex between 3,5-DiNP-functionalized graphenes with amines has utility for creating new forms of functional graphenes.