Patterned Carboxylation of Graphene Using Scanning

Apr 7, 2015 - A simple, direct, and versatile scanning electrochemical microscopy (SECM) approach for local carboxylation of multilayered graphene on ...
0 downloads 0 Views 9MB Size
Download PDF
Article pubs.acs.org/Langmuir

Patterned Carboxylation of Graphene Using Scanning Electrochemical Microscopy Kristian Torbensen,† Mikkel Kongsfelt,‡ Kyoko Shimizu,‡ Emil B. Pedersen,‡ Troels Skrydstrup,‡ Steen U. Pedersen,‡ and Kim Daasbjerg*,‡ †

Physicochimie des Electrolytes et Nanosystèmes Interfaciaux (PHENIX), Université Pierre et Marie Curie, 4 Place Jussieu, 75005 Paris, France ‡ Department of Chemistry and Interdisciplinary Nanoscience Center (iNANO), Aarhus University, Langelandsgade 140, DK-8000 Aarhus, Denmark S Supporting Information *

ABSTRACT: A simple, direct, and versatile scanning electrochemical microscopy (SECM) approach for local carboxylation of multilayered graphene on nickel is demonstrated, in which carbon dioxide serves as the carboxylation agent under reductive conditions in N,N-dimethylformamide. The use of SECM gives control over both the spatial dimensions and the degree of carboxylation. While the pattern size, in general, is governed by the dimension of the SECM tip, the degree of modification, expressed as the surface coverage of carboxylate groups introduced at the graphene substrate, is found to be controlled by the electrolysis time. This is supported by electrochemical measurements, two-dimensional X-ray photoelectron spectroscopy, Raman spectroscopy mapping, and He ion microscopy. Surprisingly, intercalation of the supporting electrolyte in the multilayered graphene on nickel occurs to a relatively small extent when compared to corresponding results obtained in previously described carboxylations of this kind of multilayered graphene.

■

INTRODUCTION Graphene is a single atom-thick layer of carbon atoms bonded in a hexagonal honeycomb-like arrangement. Because of the essentially planar structure and extended π system, graphene possesses remarkable properties. Among these are exceedingly high electrical and thermal conductivities with ballistic charge transport,1−3 high transparency in the visible spectrum,4 extremely high mechanical strength,5 and superior barrier properties.6,7 These properties have led to intensive research in not only manufacturing processes8−10 but also modification procedures,11,12 with the aim of controlling and adjusting the properties. One of the most versatile procedures for doing a chemical modification exploits the reduction of substituted benzenediazonium salts, both spontaneously13 and electrochemically,14,15 to graft covalently a large variety of functional groups through radical chemistry. The disadvantage is that the radical-based reactions are difficult to control, usually resulting in multilayer structures atop the graphene sheets. This multilayer grafting may introduce a significant distance between the substrate and the outermost functional groups, which is undesirable in a number of applications, such as sensor devices. Noteworthy, Hirsch and co-workers16 were able to exfoliate single-layered aryl-substituted graphene by mixing a sterically hindered aryldiazonium salt with graphite under strongly reducing conditions. Alternative routes to graphene modification are hydrogenation,17,18 fluorination,19 and plasma treatment,20,21 of © XXXX American Chemical Society

which, in particular, the latter involves the risk of introducing defects and holes in the graphene lattice by outright etching. Also, non-covalent modification of graphene through π−π and van der Waals interactions has been reported.22,23 This type of modification has the advantage that it does not introduce additional defects in the graphene lattice, and the electronic performance of the graphene is therefore less perturbed by the interactions. On the other hand, the stability of the modification can be questioned in some cases. Recently, we reported on an efficient electrochemical carboxylation of multilayered graphene on nickel (Ni−Gra) through either a two-step procedure consisting of charging Ni− Gra, followed by reaction with added CO2 or, alternatively, in a one-step electrolysis of CO2 using Ni−Gra as the working electrode.24 These methods allow for the extent of carboxylation to be controlled through either the number of repetitive cycles employed in the former case or the electrolysis time in the latter. Importantly, the carboxylate groups are directly attached to the substrate as opposed to the multilayer structures obtained using diazonium grafting. In addition, the carboxylate group is a versatile linker for further modification of carbon materials.25 This has mainly been exploited for graphene oxide,26 where small amounts of carboxylate groups are Received: November 18, 2014 Revised: February 21, 2015

A

DOI: 10.1021/la504500m Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

Figure 1. (A) Two-electrode cyclic voltammograms recorded at Ni−Gra in the direct-mode SECM setup at a sweep rate of 0.1 V s−1 in CO2-saturated 0.1 M Bu4NBF4/DMF. (B) Current transients recorded in the direct-mode SECM setup during electrolysis at a cell voltage of −2.6 V for Ni−Gra−5 (solid line), Ni−Gra−10 (dashed line), and Ni−Gra−30 (dotted line) with electrolysis times of 5, 10, and 30 s, respectively, and a tip−substrate distance of 13 μm in CO2-saturated 0.1 M Bu4NBF4/DMF. The inset shows a blow-up of the first 7 s of the electrolysis processes. μm; the average number of layers is 4. All samples were analyzed using XPS and Raman spectroscopy prior to use. Electrochemical Setup. The CH Instrument 900b SECM equipment was employed for the electrochemical experiments. The Ni−Gra substrate was mounted in a homemade Teflon cell and aligned to the SECM stage controller. A homemade SECM platinum tip electrode (radius a = 12.5 μm and RG = 3) was positioned 13 μm above the Ni−Gra substrate using approach curves under conditions as described for the mapping experiment below. The electrolyte solution used for the local modification, i.e., 0.1 M Bu4NBF4/DMF saturated with CO2, was introduced into the cell, and it was covered with a blanket of CO2. With SECM in the direct-mode configuration, a two-electrode cyclic voltammogram of the CO2-saturated electrolyte solution could be recorded, thereby allowing for the appropriate cell voltage for the subsequent electrolysis (−2.6 V) to be selected. The electrolysis was performed using Ni−Gra as the working electrode (cathode) and the SECM tip as the counter electrode (anode). The exact potentials of the two electrodes could not be monitored during the experiment, but considering the large active electrode area of the Ni−Gra electrode together with the high concentration of CO2, the potential of Ni−Gra can be assumed to settle at the reduction potential of CO2, ECO2. Thus, the potential of the SECM tip becomes 2.6 V relative to ECO2, which corresponds to or is slightly positive of the oxidation potential of oxalate (generated at Ni−Gra; vide inf ra). Without CO2 in the electrolyte solution, the exact potentials are unknown. For the SECM mapping experiment, a standard three-electrode configuration was employed using the same tip electrode as for the modification, a large silver wire as the quasi-reference electrode, and a platinum wire as the counter electrode in a solution of 1 mM ferrocene in 0.1 M Bu4NBF4/MeCN. The substrate was left unbiased, and the tip potential was fixed at 0.7 V versus Ag/Ag+. Post treatment of the samples consisted of thorough rinsing with DMF, followed by drying under argon flow. XPS. XPS analyses were conducted using a Kratos Axis Ultra-DLD instrument employing a hemispherical analyzer for spectroscopy and a spherical mirror analyzer for imaging. Spectra and images were recorded with a monochromatic Al Kα X-ray source at a power of 150 W. Survey scans were acquired by accumulating in the 0−1350 eV range at a pass energy of 160 eV, and high-resolution scans were acquired at a pass energy of 20 eV, with the analysis spot being 27 μm in diameter at the center of each modified area. The pressure in the main chamber during the analysis was in the 10−9 Torr range. Imaging data of the relevant species were recorded at a pass energy of 160 eV by scanning an area of 120 × 120 μm2 covering both the modified spots and surrounding unmodified areas. Background region images were recorded at a binding energy of 10 eV below and above each

indigenously present. Especially important is the diimide chemistry, used intensively for the attachment of biomolecules for sensor applications. As such, the controlled grafting of carboxylate groups is important for further modifications of graphene.27,28 Previously, the scanning electrochemical microscopy (SECM) technique has been used for local deposition of reduced graphene oxide (rGO),29 patterning of rGO,30 investigation of graphene reactivity and heterogeneous electron transfer, 31−34 and the study of adsorbate mobility on graphene.35 While SECM has been used extensively for patterning purposes, in general,36 no methods for patterned and controlled covalent functionalization of pristine graphene has yet been reported to our knowledge. In this work, it is shown that the specific location as well as the extent of carboxylation of graphene may be achieved and controlled using SECM, resulting in a high local surface coverage of carboxylate groups. This is supported by electrochemical measurements, two-dimensional X-ray photoelectron spectroscopy (XPS), Raman spectroscopy mapping, and He ion microscopy. This possibility of designing specific patterns may open for the future development of graphene-based arrays of, e.g., sensors or fabrication of electronic components with tunable band gaps determined by the carboxylation degree and/ or the exact geometrical patterns introduced.

■

EXPERIMENTAL SECTION

Materials. N,N-Dimethylformamide (DMF) was purchased from Sigma-Aldrich, and MeCN (anhydrous, 99.9%) was purchased from Lab-Scan. The supporting electrolyte tetrabutylammonium tetrafluoroborate, Bu4NBF4, was prepared using standard procedures. The electrolyte solution consisting of 0.1 M Bu4NBF4/DMF or 0.1 M Bu4NBF4/MeCN was dried prior to electrochemical experiments by passing it through a column of activated (i.e., heated to 350 °C under vacuum) Al2O3 (99.99%, Sigma-Aldrich). After this treatment, the water content was determined to be 0.055 wt % by Karl Fisher titration (TitraLab 980, Radiometer Analytical). Ferrocene was purchased from Sigma-Aldrich. The CO2 gas was supplied by AGA Denmark and used as received. CO2-saturated 0.1 M Bu4NBF4/DMF solutions were prepared by purging with CO2 for 15 min. Substrates. Chemically vapor-deposited graphene on Ni (itself deposited as a thin film on a silicon wafer) was purchased from Graphene Supermarket. According to the specifications, the Ni−Gra samples are composed of patches of 1−7 layers with a size of 3−10 B

DOI: 10.1021/la504500m Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir relevant peak. Background-corrected images were obtained by subtraction of the two background region images from that of interest. Charge compensation was achieved using an electron flood gun. A binding energy of 284.4 eV for CC components of the graphene C 1s peak was used as a reference for charge correction. The generated XPS data were processed using the computer software CasaXPS (version 2.3.15). Atomic surface concentrations were determined by fitting the core level spectra using Gaussian−Lorentzian line shapes and a Shirley background. The systematic error is estimated to be of the order of 5−10%. Raman Spectroscopy. Measurements were performed using a Renishaw InVia Raman microscope equipped with a 50 mW, 514 nm argon ion laser. In general, areas of approximately 200 × 200 μm2 were mapped with a step size of 1 μm using a 100× objective that provides a laser spot size of ∼0.6 μm in diameter. This means that ∼40 000 spectra were analyzed for each selected area by means of the Wire 4.0 software from Renishaw plc. All spectra were automatically fitted using as few Lorentzian peaks as possible on a polynomial background. He Ion Microscopy. Imaging was performed using a Zeiss Orion NanoFab microscope with a working distance of 7.5 mm, a blanker current of 7.0 pA, an acceleration voltage of 28.0 kV, and a dwell time of 1.0 μs. ImageJ software was used to determine spot sizes. Optical Microscopy and Condensation Imaging. A Nikon Eclipse Ci light microscope equipped with a 10× objective and an ImagingSource camera was used for optical microscopy. Imaging of the water condensation on the modified plates could be generated by recording a video sequence while breathing on the samples. Images with the highest contrast were then simply extracted from this sequence.

as shown recently by us for Ni−Gra, a small fraction of the reduction current may go to electrografting of CO2 onto graphene (path B),24 thereby making pattern formation based on carboxylation in SECM a viable pathway. Because the SECM tip anode is placed very close (13 μm) to the Ni−Gra cathode, a substantial part of the formed oxalate will diffuse to the anode and be oxidized back to CO2. The potential difference of the two electrode processes, i.e., reduction of CO2 (−2.2 V versus SCE in acetonitrile) and oxidation of C2O42− (0.3 V versus SCE in acetonitrile), is −2.5 V,40 which is close to the applied cell voltage herein. Hence, these two complementary reactions constitute the main electrode processes, with only a small fraction of the reduction current expected to lead to carboxylation of Ni−Gra through a reaction between graphene and CO2• −. On this basis, it was decided to perform the local modification of Ni−Gra using the direct-mode SECM setup36 and a cell voltage, ΔE, of −2.6 V, with Ni−Gra serving as the working electrode and a platinum tip positioned 13 μm above the Ni−Gra substrate as the counter electrode. Figure 1B shows the cell current recorded for 30 s for the two-electrode electrolysis carried out at ΔE = −2.6 V in a CO2-saturated electrolyte solution. Similar traces, only shorter, were recorded for 5 and 10 s electrolysis (see the inset of Figure 1B). In between the three experiments, the SECM tip was moved in lateral steps of 300 μm to generate three modified samples denoted Ni−Gra−X, where X = 5, 10, or 30. Interestingly, the current does not go to zero even for Ni−Gra−30 but reaches an almost steady-state level after 10−15 s. This is attributed to the coupling of the two reverse catodic and anodic electrode processes that “feed” each other and allows for the modification to continue essentially unimpeded throughout the electrolysis. Also, it implies that the modification of the Ni−Gra substrate does not occur to an extent, where the conductivity of the Ni− Gra electrode is lost. Finally, integration of the electrolysis current provides the total charge consumption of the cell, which can be used to evaluate the grafting efficiency (vide inf ra). A disadvantage of using the direct-mode SECM is the poor control on the exact electrode potentials applied. A number of experiments, introducing a reference electrode to the system, were conducted to determine the potentials applied during the SECM experiment. However, the interpretation of such data is not straightforward because of the strong and non-uniform electric field between the tip and substrate electrodes and the strong coupling of the two redox processes in the small gap between the electrodes. This being said, the direct-mode SECM approach possesses some distinct advantages, and it was chosen herein because of (1) the simplicity of the technique and, importantly, in this case, (2) the possibility of exploiting the CO2/oxalate redox system to exert control on the reduction and oxidation processes. The validity of exploiting this redox system was further substantiated by the fact that no substrate modification was observed, if −ΔE was lowered beyond 2.6 V. Figure 2A shows SECM mapping images obtained of Ni− Gra−5, Ni−Gra−10, and Ni−Gra−30 using ferrocene as a solution-based redox probe. Significant drops in the normalized tip current are seen at the three modified spots, which are caused by hindered electron transfer from the substrate to the redox probe. This strongly suggests that Ni−Gra indeed has been modified during the electrolysis of CO2 just below the SECM tip electrode. Defects introduced in the graphene lattice because of the covalent modification will decrease the lateral conductivity. In addition, the fact that the graphene layers in

■

RESULTS AND DISCUSSION Electrochemical Analysis. Figure 1A shows the cyclic voltammogram for reduction of CO2 using the direct-mode SECM setup36 with Ni−Gra serving as the working electrode and a platinum tip positioned 13 μm above the Ni−Gra substrate as the counter electrode with the purpose of establishing the right conditions for the SECM electrolysis experiments. In the presence of CO2, a reduction wave is observed for a cell voltage, ΔE, between −2.2 and −2.8 V, which is not seen in the absence of CO2 and agrees with similar features already obtained at various carbon substrates.37,38 As illustrated in Scheme 1 and in accordance with early literature work, the main reaction occurring upon reduction of CO2 to CO2• − in aprotic media is a fast dimerization of the radical anion to generate oxalate, C2O42− (path A).39 However, Scheme 1. Illustration of the Two-Electrode Processes, in Which Carbon Dioxide upon Reduction at the Ni−Gra Substrate Either Dimerizes to the Oxalate (Path A) or Reacts with Graphene (Path B)a

a

Oxalate is reoxidized to carbon dioxide at the tip electrode. C

DOI: 10.1021/la504500m Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

Figure 2. (A) SECM mapping of the Ni−Gra−5, Ni−Gra−10, and Ni−Gra−30 spots using 1 mM ferrocene as the redox probe in 0.1 M Bu4NBF4/ MeCN. The scale bar refers to the recorded tip current, iT, normalized with respect to the bulk current, iT,inf. (B) Line scans of Ni−Gra−5 (solid line), Ni−Gra−10 (dashed line), and Ni−Gra−30 (dotted line) using the SECM tip position during electrolysis as a reference point.

Figure 3. XPS mapping of the 1s core level electrons for O, C, and the O/C ratio (maximum and minimum values are shown at the intensity scale bars), displaying the trends in the modification of Ni−Gra−5, Ni−Gra−10, and Ni−Gra−30.

spots. This provides the full width at half maximum (fwhm), which is 57, 78, and 83 μm, respectively, in the three cases. The generation of such well-defined foot prints of sizes in accordance with diffusion-controlled processes is, furthermore, similar to those obtained elsewhere using similar tip geometries but on other chemical systems.36 The spatial extent of modification increases significantly as the electrolysis time is prolonged from 5 to 10 s, while a further increase in the electrolysis time to 30 s has little effect. This would seem to be in line with a self-limiting process that is governed by the spatial distribution of the electric field lines at the tip−substrate gap. XPS Analysis. To extract information on the chemical nature of these spots, XPS survey and high-resolution spectra of the samples were therefore recorded (see Figure S2 of the Supporting Information). Figure 3 shows mapping images of

the multilayered structure will experience expansion because of the introduction of the functional groups along with a concomitant intercalation of Bu4N+ (vide inf ra) may further decrease the conductivity perpendicular to the substrate. The dimension and location of these regions are a reflection of the size and lateral position of the tip electrode as well as the electrolysis time. Noteworthy, these regions, being much larger than the tip dimension (because of the expanding diffusion field), are still reasonably conducting because the normalized tip current at no sites is found to be less than unity. No sign of modification was observed at a sample prepared under the same conditions without CO2 (see Figure S1A of the Supporting Information). Figure 2B shows the pertinent SECM line scans across the central areas of the Ni−Gra−5, Ni−Gra−10, and Ni−Gra−30 D

DOI: 10.1021/la504500m Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

Figure 4. High-resolution XPS carbon 1s core level spectra recorded at (A) Ni−Gra, (B) Ni−Gra−5, (C) Ni−Gra−10, and (D) Ni−Gra−30.

Table 1. Atomic Concentrations of C 1s, N 1s, O 1s, and Ni 2p, Classification of Carbon-Bonding Types, and Surface Concentration of Carboxylate Groups Obtained by XPS for Ni−Gra and Ni−Gra−X (X = 5, 10, or 30) atom percenta sample

CC

C−C

C−O/C−N

Ni−Gra Ni−Gra−5 Ni−Gra−10 Ni−Gra−30

81.5 69.6 48.9 26.3