Electrochemical performance of Ti₃C₂T_x MXene in aqueous media: towards ultrasensitive H2O2 sensing

Abstract

An extensive characterization of pristine and oxidized Ti₃C₂T_x (T: =O, -OH, -F) MXene showed that exposure of MXene to an anodic potential in the aqueous solution oxidizes the nanomaterial forming TiO₂ layer or TiO₂ domains with subsequent TiO₂ dissolution by F^- ions, making the resulting nanomaterial less electrochemically active compared to the pristine Ti₃C₂T_x. The Ti₃C₂T_x could be thus applied for electrochemical reactions in a cathodic potential window i.e. for ultrasensitive detection of H₂O₂ down to nM level with a response time of approx. 10 s. The manuscript also shows electrochemical behavior of Ti₃C₂T_x modified electrode towards oxidation of NADH and towards oxygen reduction reactions.

1. Introduction

2D nanomaterials with a high specific surface area have a variety of promising properties, making them useful as carriers, novel electronic materials and as a part of optical devices, sensors and energy storage devices [1–4]. Since the description of properties of the first 2D nanomaterial – graphene in 2004 [5], there is an explosion of papers describing other 2D nanomaterials [6–10].

In 2011, a new family of 2D MXene nanomaterials were introduced [11], exhibiting more complex (layered) structure compared to graphene, having many specific properties, like metallic conductivity and hydrophilicity due to presence of a negative charge on the surface [12–14]. MXenes belong to a family of exfoliated transition metal carbides and carbonitrides synthe-sized by hydrofluoric acid (HF) etching of the “A” group element from “MAX” phase powders [15] (where “M” is a transition metal, “A” is an element mostly from groups 13 and 14 of a periodic table, and “X” is a carbon or a nitrogen atom [16]) resulting in 2D layered structure similar to graphenes [17].

The as-obtained MXene sheets are terminated with oxygen- and/or fluorine-containing functional groups (=O, -OH, -F). However, alkalization and calcination post-treatments were shown to remove these surface groups, thus enhancing electrical conductivity of the nanomaterial [18]. Up to now, the most common applications of MXenes are high capacity electrode materials for batteries [19–25], as supercapacitors [26] and pseudocapacitive cathode materials [27,28] or as an electromag-netic interference shielding material [4]. Environmental removal of Pb(II) ions using this nanomaterial was also reported [29]. Ti₃C₂T_x in a colloidal solution exhibits an antimicrobial activity, which is higher compared to graphene oxide (GO) [30]. An adsorption and a photocatalytic decomposition of organic molecules in aqueous solutions were observed, as well [31].

From a sensing point of view, Ti₃C₂ either in a pristine form [32] or combined with TiO₂ nanoparticles [33] was shown to provide an excellent immobilization matrix for hemoglobin-based mediator- free biosensor for H₂O₂ detection with a limit of detection (LOD) down to 14 nM and an excellent biosensor stability [33]. The same platform was also used to investigate detection of NaNO₂ with LOD of 120 nM [34]. The most recent study suggests immobilization of glucose oxidase on Ti₃C₂T_x MXene modified by gold nanoparticles as a biointerface for sensitive detection of glucose [35]. Also, an adsorption of different gases (NH₃, H₂, CH₄, CO, CO₂, N₂, NO₂ and O₂) on Ti₂CO₂ monolayer was studied, resulting in an adsorption of only NH3 molecules, making this material applicable not only as a battery material, but also as a potential gas sensor and NH₃ capturer with a high selectivity [36]. The most recent study described application of MXene patterned field-effect transistor for probing neural activity by detection of dopamine [37].

In this study, we focused on investigation of electrochemical performance of Ti₃C₂T_x in an aqueous solution for potential sensing applications. Ti₃C₂T_x was investigated for its ability to detect oxygen and hydrogen peroxide and to oxidize NADH, for potential future construction of biosensors. Ti₃C₂T_x modified electrode proved to be extremely sensitive for detection of H₂O₂ with LOD of 0.7 nM. Detection of H₂O₂ as an analyte is of importance in chemical and food industry (applied as an oxidizing agent) and for detection in clinical, pharmaceutical and environmental samples (see an excellent paper reviewing construction of H₂O₂ sensors [38]). Furthermore, H₂O₂ is a byproduct of enzymatic action of various oxidases so efficient detection of H₂O₂ is very important for development of oxidase-based biosensors utilizable in numerous applications [39–42].

2. Materials and methods

2.1. Materials

All chemicals (i.e. K₃[Fe(CN)₆], K₄[Fe(CN)₆].3H₂O, H₂O₂, H₂SO₄, NADH, NaOH, DMSO) and phosphate buffer (PB) components (KH₂PO₄ and K₂HPO₄, pH 7.0), were of ≥99% purity or p.a. grade and were purchased from Sigma Aldrich (USA). 50 wt% HF was obtained from Fisher Scientific, USA. All solutions were freshly prepared in 0.055 μS ultrapure deionized water (DW) and filtered prior use using 0.2 μm sterile filters.

2.2. Ti₃C₂T_x MXene synthesis

Ti₃AlC₂ was synthesized as described previously [43]. Multi-layer Ti₃C₂ MXene was prepared by HF treatment protocol with a minor modification. Briefly, Ti₃AlC₂ was added slowly to an aqueous HF solution (50 wt%) for 18 h at room temperature followed by intercalation with DMSO. The reaction mixture was washed several times with DW until pH 6 was reached. The colloidal solution of delaminated Ti₃C₂T_x dispersion was obtained by sonication of Ti₃C₂T_x powders (1 mg) in 2 mL of DW water, which was purged with argon for 60 min prior sonication, followed by centrifugation of the dispersion at 3,000 rpm for 1 h with a final collection of the supernatant.

2.3. Electrochemical procedures

All electrochemical procedures were run on a laboratory potentiostat/galvanostat Autolab PGSTAT 302N with an impedi-metric module (Ecochemie, Utrecht, Netherlands) with a glassy carbon electrode (GCE, d= 3 mm, Bioanalytical systems, USA) used as a working electrode. Chronoamperometric detection of H2O₂ at 300 rpm was performed on a rotating disc electrode employed as a working electrode. An Ag/AgCl/3 M KCl reference electrode and a counter Pt electrode (Bioanalytical systems, USA) were applied in a three-electrode cell system. Measurements were run under Nova Software 1.10, and data acquired were evaluated using OriginPro 9.1.

Chronoamperometry was applied as a useful method for determination of real surface area of Ti₃C₂T_x modified GCE, which was calculated from a Cottrell equation:

$$i=\frac{nFA{c}_j^0\sqrt{D_j}}{\sqrt{\pi t}}$$ (eqn. 1)

where n is the number of electrons exchanged, F is a Faraday constant (96,485 C mol^-1), c⁰_j is concentration of the electrochemical mediator i.e. ferricyanide (mol cm^-3), D is the diffusion coefficient (7.6 10^-6 cm² s^-1 for ferricyanide solution used in this study), t is time (in s) and A is the real surface area (in cm²). The experiment was conducted by applying two potentials (0.2 V and -0.6 V, respectively) for the reduction of 1 mM ferricyanide solution in 0.2 M KCl. Under diffusion control, a plot of i vs. t^-1/2 is linear and from the slope, the value of A could be obtained. Real surface area of Ti₃C₂T_x modified GCE was 13.3 mm², while a geometric surface area of GCE was 7.1 mm².

Electrochemical impedance spectroscopy (EIS) can provide characteristics of an interfacial layer using a redox probe. The result of EIS analysis is presented in a Nyquist plot, from which such characteristics can be obtained. EIS was measured in an electrolyte containing 5 mM potassium hexacyanoferrate (III), 5 mM potassium hexacyanoferrate (II) and 0.1 M PB, pH 7.0. The analysis was run at 50 different frequencies (ranging from 0.1 Hz up to 100 kHz) under Nova Software 1.10 (Ecochemie, Netherlands). The results were presented in a form of a Nyquist plot, with an equivalent circuit R(Q[RW]) applied for data fitting.

2.4. Electrode modifications

First, the GCE was polished with a 1.0 μm alumina slurry and diamond polishing paste and sonicated in DW. The cleaned GCE was subsequently dried using a purified nitrogen stream. In order to obtain homogeneous Ti₃C₂T_x dispersion, Ti₃C₂T_x solution was sonicated for 1 min, if not specified otherwise, under Ar atmo-sphere to prevent potential oxidation of MXene. The Ti₃C₂T_x modified electrode was prepared by a simple drop-casting method. The final volume of 30 mL of a MXene dispersion was pipetted on the GCE in two steps (2 x 15 μL) and allowed to dry at room temperature in a laminar box.

2.5. Preparation of oxidized Ti₃C₂T_x (o Ti₃C₂T_x)

Oxidation of Ti₃C₂T_x was performed in 0.1 M PB pH 7.0 by a linear sweep voltammetry (LSV) running from 0 mV to 500 mV at a sweep rate of 100 mV s^-1.

2.6. Characterization of Ti₃C₂T_x and o Ti₃C₂T_x

Raman spectra were measured with a DXR Raman Microscope (Thermo Scientific, USA) with 532 nm laser in the region from 3,350 to 52 cm^-1 (laser power 0.5 mW, exposure time 20 s, number of exposures 10, slit: 50 μm).

Contact angle measurements were run on a portable instrument System E (Advex Instruments, Czech Republic) to reveal contact angle and free surface energy for Ti₃C₂T_x modified interfaces. The droplet volume was 2 μL and the testing liquid was distilled water. Free surface energy was determined using the two-liquid Owens-Wendt (OW) method, where the total surface energy γ consists of disperse γ^d and polar γ^p components. Water and diiodomethane were used as test liquids (surface tension values according to Strom). In order to minimize measurement error, 5 contact angles were measured, with the highest and the lowest value eliminated. For each sample, the water and diiodomethane contact angle was obtained as an average value of assays performed using 3 droplets.

A peak force tapping mode atomic force microscopy (AFM, Scan Asyst, Bruker, USA) in air was carried out on a Bioscope Catalyst instrument and Olympus IX71 microscope in conjunction with NanoScope 8.15 software at a scan rate of 0.5 line s^-1 with the tip set automatically for optimal gain. AFM mica substrates (grade V-1, d = 12 mm, SPI Supplies, USA) modified with Ti₃C₂T_x were scanned using a SCANASYST-AIR silicon tip on a nitride lever (Bruker, USA, with f₀ = 50–90 kHz and k = 0.4 N m^-1), sharpened to a tip radius of 2 nm.

The same modified substrates i.e. modified square shaped gold chips (Arrandee, Germany) as for AFM imaging were applied for obtaining scanning electron microscopy (SEM) images using Carl Zeiss EVO 40HV apparatus (Germany) after Au CVD treatment to observe the structure of Ti₃C₂T_x.

XPS signals for Ti₃C₂T_x and o Ti₃C₂T_x were recorded on modified square shaped Au chips (Arrandee, Germany) using a Thermo Scientific K-Alpha XPS system (Thermo Fisher Scientific, UK) equipped with a micro-focused, monochromatic Al K alpha X-ray source (1486.6 eV). An X-ray beam of 400 μm size was used at 6 mA x 12 kV. The spectra were acquired in the constant analyzer energy mode with pass energy of 200 eV for the survey. Narrow regions were collected with pass energy of 50 eV. Charge compensation was achieved with the system flood gun that provides low energy electrons (~0 eV) and low energy argon ions (20 eV) from a single source. The argon partial pressure was 2 x 10^-7 mbar in the analysis chamber. The Thermo Scientific Avantage software, version 4.84 (Thermo Fisher Scientific), was used for digital acquisition and data processing. Spectral calibra-tion was determined using the automated calibration routine and the internal Au, Ag and Cu standards supplied with the K-Alpha system. The surface compositions (in atomic %) were determined by considering the integrated peak areas of atoms and the respective sensitivity factors.

Ti₃C₂T_x was further characterized using X-Ray powder diffrac-tion (XRD) measurements to investigate crystal structure of the material. Typical samples were prepared by pipetting of 20 μL of Ti₃C₂T_x dispersion (1.5 mg mL^-1 in DW, sonication for 30 min unless stated otherwise) on a glass slide, dried under reduced pressure, gently rinsed with DW to wash out any particles not incorporated into the formed Ti3C2Tx film and dried in Ar stream. The samples were characterized using XRD equipment Empyrean with irradiation source Cu Kα1 (λ=0.15406 nm) at tension 45 kV and current 40 mA and detector PIXcel1D with stage platform with adjustable Z-height (all from PANalaytical). Value of Z-height was determined for each sample using micrometer. X-ray diffraction was observed in gonio mode in 2Θ range of 4°–20° with step size 0.0066° and scan speed 0.055° s^-1. Size of the main lattice distance (d) corresponds to the peak position, i.e. to the angle of the diffracted beam, according to the Bragg's law λ = 2d sin Θ, where λ is 0.15406 nm and sin Θ was calculated from the position 2Θ of the given XRD peak. The domain size can be calculated using Scherrer formula s = K λ/β cos Θ where s, K and β stands for a mean domain size, a shape factor (0.9) and a broadness of the peak in half its maximum intensity (FWHM) in radians, respectively.

Secondary Ion Mass Spectrometry (SIMS) is a technique for sensitive chemical surface analysis of samples [44]. The analysis is not limited by the origin or type of a sample, that can be substantially any, inorganic, organic and biological. SIMS employing Time-of-Flight (TOF) analyzer provides elemental, chemical state and molecular information from surface layers or thin film structures with high sensitivity on the level of ppm-ppb. Besides, TOF SIMS IV spectrometers could provide high mass resolution, lateral resolution of 100 nm and a depth resolution of 1 nm. With the primary ion beam scanning across the sample surface, even the 2D chemical imaging of elements or molecules can be obtained providing data on a spatial distribution of predefined species [44]. Mass spectrometry measurements were performed using a TOF-SIMS IV (ION-TOF, Muenster, Germany), a reflectron type of time-of-flight mass spectrometer equipped with a Bismuth ion source. Pulsed 25 keV Bi⁺ were used as primary ions with ion current of 1.1 pA. The TOF-SIMS spectra were measured by scanning over the 100 μm x 100 μm analysis area with a total primary ion dose density below the static limit of 10¹³ ions cm^-2. SIMS images were measured by scanning over the 200 μm x 200 μm analysis area, with a lateral resolution of 5 mm. All assays were performed in a positive and a negative polarity.

3. Results and discussion

3.1. Microscopic characterization of Ti₃C₂T_x modified GCE

SEM images revealed formation of aggregates on the surface differing in size i.e. having few mm in size (Fig. 1 left) or with size larger than 10 μm (Fig. 1 right).

Fig. 1.

Representative SEM images of Ti₃C₂T_x sonicated for 1 min. Magnification: 20,000x.

3.2. Electrochemical oxidation of Ti₃C₂T_x (preparation of oTi₃C₂T_x)

Initial cyclic voltammetry (CV) experiments confirmed that an anodic oxidation of Ti₃C₂T_x is an irreversible process with an anodic peak appearing only in the first CV scan at a potential of 430 mV and could not be observed in the subsequent CV scans (Fig. 2). Further experiments revealed that this anodic peak appeared only once and could not be seen if the Ti₃C₂T_x modified GCE was further reduced by running CV in the potential window from 0 mV to -500 mV (as shown in Fig. 3), from -500 mV to -1,000 mV (data not shown) or after the Ti₃C₂T_x modified GCE electrode was kept at an open circuit potential for couple of minutes. This really indicates irreversible oxidation of Ti₃C₂T_x upon exposure to an anodic potential, which could not be re-reduced.

Fig. 2.

CV of GCE/Ti₃C₂T_x showing several consecutive scans run in a potential window from 0 V to 1 V at a sweep rate of 100 mV s^-1 in 0.1 M PB pH 7.0. Ti₃C₂T_x dispersion was prepared by 1 min sonication. Further details are provided in the Experimental section.

Fig. 3.

CVs performed in 0.1 M PB pH 7.0 at bare GCE and at GCE/Ti₃C₂T_x run at a scan rate of 100 mV s^-1. Ti₃C₂T_x dispersions deposited on GCE were prepared by sonication of the mixture for 1, 10, 30 or 60 min.

Optimization of sonication time for preparation of Ti₃C₂T_x dispersions was performed also electrochemically either by running CV in the potential window from 0 mV to -500 mV (Fig. 3) or in the potential window from 0 mV to 1,000 mV (Fig. S1). Such CV experiments confirmed that optimal sonication time for preparation of Ti₃C₂T_x dispersion was 1 min. An increase of sonication time from 1 min to 10 min led to decrease of a Faradaic current (Fig. S1) or a capacitive current (Fig. 3). Sonication time longer than 10 min (up to 60 min) did not have a detrimental effect on the electrochemical behavior of Ti₃C₂T_x modified GCE (Fig. 3 and Fig. S1). When Ti₃C₂T_x modified GCE was exposed to potentials above 200 mV, it was possible to observe partial dissolution of the material from the modified GCE by a naked eye.

3.3. Raman spectra analysis

Results indicated that optimal power density for obtaining Raman spectra was 0.5 mW (Fig. S2) and an optimal sonication time was 1 min, in an agreement with results obtained from electrochemical assays. The Raman spectrum of Ti₃C₂T_x modified GCE showing peaks at 200, 380 and 610 cm^-1 (Fig. 4) is in an agreement with results obtained in a previous study [45].

Fig. 4.

Representative Raman spectra of Ti₃C₂T_x and oTi₃C₂T_x acquired with laser power set to 0.5 mW.

Moreover, intensity of D-band (1,391 cm^-1) compared to G-band (1,596 cm^-1) is very low, in agreement with a previous study, as well [45]. When oTi₃C₂T_x was inspected by Raman spectroscopy the spectra looked very similar to Raman spectra of Ti₃C₂T_x, with the only difference i.e. presence of a well-developed D-band at 1,391 cm^-1, indicating induction of disorder within oTi₃C₂T_x (Fig. 4) [46]. Interestingly, in the Raman spectrum of oMXene a major peak at 144 cm^-1 and other minor peaks observed at 394, 513 and 635 cm^-1 (attributed to anatase TiO₂) [22,45] are not visible, indicating low density of anatase TiO₂ on the surface of oTi₃C₂T_x.

3.4. Contact angle measurements

Contact angle measurements obtained using an Owens-Wendt- Rable-Kaeble model are summarized in Table S1. A decrease of contact angle in water (from 36° to 30°) and an increase of a polar component of a free surface energy (from 30 mJ m^-2 to 35 mJ m^-2) for oTi₃C₂T_x compared to Ti₃C₂T_x indicate that oxidation of the sample by LSV introduced polar functional groups into the sample of oTi₃C₂T_x or removed F^- groups making the surface of oTi₃C₂T_x more hydrophilic compared to Ti₃C₂T_x. The contact angle of 36° measured on Ti₃C₂T_x is in an excellent agreement with a previous study showing a value of 34° [17]. Images from such measurements are shown in Fig. S3.

3.5. AFM measurements

AFM measurements were performed to see surface morphology of Ti₃C₂T_x or oTi₃C₂T_x layers and results indicated a substantial decrease in the value of mean square roughness (R_q) for Ti₃C₂T_x with a value of 1.6 nm compared to oTi₃C₂T_x modified interface with a value of 0.2 nm. Presence of either higher or deeper features on Ti₃C₂T_x interface compared to oTi₃C₂T_x could be anticipated when looking on a parameter of image R_max (i.e. a value of 10.5 nm for Ti₃C₂T_x and a value of 1.4 nm for oTi₃C₂T_x) (Table S2) [22]. Typical AFM images of Ti₃C₂T_x and oTi₃C₂T_x modified gold chips are shown in Fig. 5. However, oTi₃C₂T_x modified gold surface is not completely flat as could be indicated from Fig. 5 right, but exhibits a moderate roughness (Fig. S4).

Fig. 5.

Typical AFM image of Ti₃C₂T_x (left) and oTi₃C₂T_x (right) modified Au chip, showing rougher surface of Ti₃C₂T_x compared to oTi₃C₂T_x. In both cases z-axis was set to 11 nm.

AFM height profile analysis of a more concentrated sample deposited compared to Fig. 5 revealed a decrease in Ti₃C₂T_x film thickness (correlating with a decrease observed for R_q value) from (7.8 ± 0.8) nm to (2.2 ± 0.9) nm for oTi₃C₂T_x, respectively. Moreover, layers of Ti₃C₂T_x were clearly visible at the edge of each flake (as seen on Fig. S5 left). Thickness of these layers ranges from (0.9 ± 0.1) nm for oTi₃C₂T_x to (1.1 ± 0.1) nm for Ti₃C₂T_x, respectively, what correlates well with the previously published value for a Ti₃C₂T_x monolayer (1.0 ± 0.2) [47–49]. Average size of the isolated flakes on the surface was (123 ± 7) nm (correlating with the value of 100–200 nm range published previously) [50,51], as shown on Fig. S5 right, and the average surface density for the MXene films prepared was G = (88 ± 10) flakes mm^-2 (2D projection of the 3D surface map).

3.6. XPS spectra

Analysis of both types of samples showed a significant decrease (from (18.8 ± 0.5) atomic % to (3.8 ± 1.7) atomic %) in F1 s content resulting in an increased hydrophilicity of oTi₃C₂T_x (Table S3, Fig. S6). The content of Ti decreased from (9.2 ± 0.1) atomic % for Ti₃C₂T_x to (4.4 ± 3.2) atomic % for oTi₃C₂T_x (as shown in Fig. S6 left) with an increase of carbon content from (42.9 ± 0.3) atomic % for Ti₃C₂T_x to (62.9 ± 5.7) atomic % for oTi₃C₂T_x. The content of oxygen within both samples is approximately the same i.e. (29.2 ± 0.9) atomic % for Ti₃C₂T_x or (29.0 ± 0.8) atomic % for oTi₃C₂T_x. It is quite interesting to point out to the fact that measuring an atomic composition of Ti₃C₂T_x by XPS was more reproducible with an average RSD of 1.9%, while an average RSD for measuring atomic composition of oTi₃C₂T_x was 32.3%, what can indicate more heterogeneous chemical and/or morphological composition of oTi₃C₂T_x surface compared to Ti₃C₂T_x.

3.7. XRD analysis

For all samples, two major XRD peaks were observed (see Fig. S7). One at 2Θ = 6.09 ± 0.03° with calculated d lattice of 14.5 Å and an average domain size of 14 nm and the other one at 9.519 ± 0.001° with calculated d lattice of 9.28 Å and an average domain size of 143.4 nm. The former peak indices disintegration of some larger particles into smaller ones with more distant individual sheets confirming that the performed sonication leads to an exfoliation of Ti₃C₂T_x (see for example [17]). The former peak, on the other hand, revealed a presence of large particles (domain size of 143.4 nm, in agreement with AFM data discussed above) with smaller distance between individual sheets (i.e. 9.28 Å, again with agreement with AFM data discussed above), assigned most probably to the non-exfoliated material. This hypothesis is supported by a value of (11 ± 3) nm reported for particle size in 0001 direction for different MAXs after the HF treatment [11].

Besides the abovementioned two major peaks, small features at 19.107 ± 0.004° were detected and a pattern with one peak at 38.780 ± 0.004° and a second one at 38.876 ± 0.003°. Except for the first one, positions of all the peaks exhibited very narrow dispersions, suggesting that they should be assigned to crystallini-ty patterns which are not influenced by inevitable random variances in Ti₃C₂T_x exfoliation and deposition steps. On the other side, the first peak is broad with a larger position values interval, suggesting that it originated from less regular structures, probably exfoliated Ti₃C₂T_x sheets loosely stacked to each other via van der Walls or hydrophobic interactions. It should be also noted, that the difference between d parameters of particles represented by the two major peaks, i.e. 5.22 nm, is very close to the value reported by Lukatskaya et al. [52] for a lattice size change caused by intercalation of K⁺ or NH₄⁺ ion between Ti₃C₂T_x sheets.

3.8. Secondary ion mass spectrometry (SIMS) analysis

SIMS is a technique for sensitive chemical surface analysis of samples [44]. In Fig. S8 and Fig. S9 there are shown representative mass spectra obtained for Ti₃C₂T_x and oTi₃C₂T_x by SIMS. LSV procedure to prepare oTi₃C₂T_x was performed on Ti₃C₂T_x deposited on bare Au chip (prepared by CVD method). In the mass spectrum of Ti₃C₂T_x (Fig. S8) it was possible to see peaks, which were attributed to Ti⁺ and TiO⁺, but such peaks were not observed, when oTi₃C₂T_x was analyzed indicating most likely removal of an outer Ti layer from Ti₃C₂T_x during exposure of Ti₃C₂T_x to an anodic potential. A closer look at data presented in Fig. S8 showed an interesting fact that from Ti₃C₂T_x, ions of carbon with lower amount of oxygen (i.e. C₂₂H₂₉O₂⁺, C₂₃H₃₁O₂⁺ and C₂₄H₃₃O₂⁺) were released and such ions could not be generated from oTi₃C₂T_x and that oTi₃C₂T_x contains highly oxidized carbon molecules indicated by presence of ions such as C₂₅H₂₅O₄⁺, C₂₆H₂₇O₄⁺ and C₂₇H₂₉O₄⁺, which were not present in Ti₃C₂T_x spectra. Furthermore, SIMS analysis revealed presence of F^- in the spectra acquired from Ti₃C₂T_x, while a peak attributed to presence of F^- ions in the spectra of oTi₃C₂T_x was less intensive in order of magnitude compared to Ti₃C₂T_x (Fig. S10 and Fig. S11).

In Fig. S12 and Fig. S13 there are representative SIMS 2D images of fragments released from Ti₃C₂T_x and oTi₃C₂T_x shown. Fragments detected in both samples (from upper left to middle right) are as follows: Li⁺, CH₃⁺, Na⁺, C₂H₅⁺, Ti⁺ and Au⁺. For bare Ti₃C₂T_x, isolated islets of Ti (10–20 mm) could be observed, in contrast to oTi₃C₂T_x, where the Ti coverage is more uniform, however, with lower Ti intensity compared to Ti₃C₂T_x (Fig. 6).

Fig. 6.

SIMS 2D images for Ti⁺ fragments of Ti₃C₂T_x (left) and oTi₃C₂T_x (right) in a positive polarity (upper row) and SIMS 2D images for F^- fragments of Ti₃C₂T_x (left) and oTi₃C₂T_x (right) in a negative polarity (lower row).

From all these characterization techniques we can make some conclusions: 1) from XPS and SIMS experiments it is clear that especially outer Ti containing layer of Ti₃C₂T_x is influenced by application of an anodic voltage since SIMS image showed lower intensity of the upper Ti layer in the oTi₃C₂T_x sample compared to Ti₃C₂T_x, while XPS spectra with a laser beam reaching deeper into the oTi₃C₂T_x layer provided evidence about presence of Ti species in the sample of oTi₃C₂T_x; 2) XPS measurements confirmed dramatic decrease of fluoride content from (18.8 ± 0.5) atomic % for Ti₃C₂T_x to (3.8 ± 1.7) atomic % for oTi₃C₂T_x (Fig. S6 right) and SIMS experiments confirmed that the intensity of F^- ions is an order of magnitude lower in oTi₃C₂T_x sample compared to Ti₃C₂T_x sample (Fig. S14 and Fig. S15), while F^- were present in Ti₃C₂T_x sample (F^- peak area of 3,243,291 for Ti₃C₂T_x and 284,426 for oTi₃C₂T_x with intensity of F^- peaks normalized to total ion intensity for both samples, as shown in Fig. 6); 3) dissolution of some material from GCE/Ti₃C₂T_x during exposure to an anodic potential seen by a naked eye.

Based on these observations and using data already published in the literature we could propose the following equation behind oxidation of an outer layer of Ti₃C₂T_x during exposure to an anodic potential:

$${\text{Ti}}_3{\text{C}}_2{\text{F}}_{\text{x}}+8{\text{H}}_2\text{O}\to 3{\text{TiO}}_2+2\text{CO}+16{\text{H}}^{+}+\left(16-\text{x}\right){\text{e}}^{-}+{x\text{F}}^{-}$$ (eqn. 2)

When Ti₃C₂F_x containing F^- ions is oxidized by the anodic potential it is oxidized to TiO₂ and most likely CO/CO₂ is emitted, an assumption based on oxidation of another type of MXene (i.e. Ti2CTx-based one) by H₂O₂ performed in DW [22]. After formation of TiO₂ on the outer layer of Ti3C2Fx by an anodic applied voltage in presence of F^- ions a further step would be a dissolution of TiO₂ by forming a complex with F^- ions according to this equation [53]:

$${\text{TiO}}_2+4{\text{H}}^{+}+6{\text{F}}^{-}\to {\left[{\text{TiF}}_6\right]}^{2-}+2{\text{H}}_2\text{O}$$ (eqn. 3)

This proposed mechanism explains both decrease of Ti and F^- content in the oTi₃C₂F_x sample as measured by XPS and SIMS. In a previous study formation of TiO₂ layer or TiO₂ islands on the surface of Ti₂CT_x MXene was observed, while exposing MXene to H₂O₂ in distilled water [22]. Since this reaction was performed in distilled water with concentration of H⁺ ions too low, chemical etching via F^- ions destroying TiO₂ layer could not proceed (i.e. eqn. (3)). In the second study, formation of either rutile or anatase TiO2 nanoparticles were formed on the surface of Ti₃C₂T_x MXene by exposure to flash oxidation conditions (950 °C; 1 min) or to slow heating (450 °C; 2 h) [54]. This MXene sample equally as our Ti₃C₂T_x sample contained F^- ions [54], but the main reasons why F^- did not dissolve forming TiO₂ nanoparticles in the previous study most likely was that high temperature applied effectively removed F^- ions, as suggested previously [18].

3.9. Electrochemical impedance spectroscopy (EIS) analysis

Using a R(Q[RW]) Randles equivalent circuit, the Rct values obtained for bare GCE, Ti₃C₂T_x modified GCE and oTi₃C₂T_x modified GCE were (164 ± 36) Ω, (7,130 ± 600) Ω and (8,500 ± 1,000) Ω, respectively (Fig. 7), suggesting an increase of resistivity of oTi₃C₂T_x compared to Ti₃C₂T_x indicating that oTi₃C₂T_x is less conductive than oTi₃C₂T_x.

Fig. 7.

Representative Nyquist plots at bare GCE, GCE/Ti₃C₂T_x and GCE/oTi₃C₂T_x obtained in 5 mM ferricyanide/ferrocyanide solution in 0.1 M PB pH 7.0. For the assay 50 different frequencies in the range from 1.0 Hz to 10 kHz were applied.

3.10. Electrochemical oxidation of NADH

An interesting analyte applicable for proper function of enzyme-based biosensors is NADH, which is a byproduct of enzymatic action of dehydrogenases. There is a large anodic peak of NADH oxidation on Ti₃C₂T_x modified GCE indicating a beneficial redox behavior of Ti₃C₂T_x towards oxidation of NADH (Fig. 8). The onset potential for NADH oxidation is close to 200 mV, what is an attractive feature for potential construction of dehydrogenase- based biosensors operating with NADH as a cofactor and a value similar to the value obtained with the electrode modified by chemically reduced GO [55]. Oxidized Ti₃C₂T_x exhibited much lower oxidation current of 23 mA (325 mA cm^-2, expressed per geometric surface area as usually described in the literature) compared to Ti₃C₂T_x with a value of 144 mA (i.e., 2.04 mA cm^-2, both read at an applied potential of +780 mV) in presence of 2 mM NADH. A control experiment performed by oxidation of 2 mM NADH on GCE revealed a current of 32 mA, while a value of 19 mA (both read at a potential of 718 mV) was observed at GCE/oTi₃C₂T_x surface.

Fig. 8.

CVs performed in 0.1 M PB pH 7.0 and in 2 mM NADH solution at bare GCE electrode and Ti₃C₂T_x modified GCE run at a scan rate of 100 mV s^-1.

When CV experiment was performed with several scans, the experiment revealed that the beneficial redox behavior of Ti₃C₂T_x substantially dropped in the 2^nd scan and a further decrease of an anodic current with an increased number of scans was observed (Fig. S16). Furthermore, chronoamperometric experiment (Fig. S17) confirmed that exposure of Ti₃C₂T_x modified GCE to an anodic potential of 700 mV resulted in a decreased ability of Ti₃C₂T_x modified GCE to oxidize NADH with the same current observed after 1 h on bare GCE and on Ti₃C₂T_x modified GCE.

3.11. Oxygen reduction reaction (ORR)

From previous electrochemical investigations it was clear that Ti₃C₂T_x modified GCE was possible to apply only for redox reactions, which occur in a cathodic potential window. This is why we tested performance of Ti₃C₂T_x modified GCE for ORR and for reduction of H₂O₂, reactions, which can be applied for construction of sensors or biosensors. ORR is of high importance for many applications, e.g. hydrogen-oxygen fuel cells, metal-air batteries and biosensors, as well [56].

The results indicate that especially acidic environment i.e. 0.1 M H₂SO₄ is suitable for ORR to occur with a possibility to achieve a moderate current density in presence of oxygen and again it was confirmed that oTi₃C₂T_x has only a limited ability to reduce oxygen compared to Ti₃C₂T_x modified GCE (Fig. 9). The current density of 330 mA cm^-2 (at -590 mV) for oTi₃C₂T_x and current density of 500 mA cm^-2 (at -590 mV) for Ti₃C₂T_x were observed when CV assayed in presence of air were subtracted from CV obtained under N₂ atmosphere.

Fig. 9.

CVs of ORR run at bare GCE, GCE/Ti₃C₂T_x and GCE/oTi₃C₂T_x in 0.1 M NaOH (left) or 0.1 M H₂SO₄ (right) under N₂ and air atmosphere. The experiments were run at a sweep rate of 100 mV s^-1.

In an alkaline solution, subtracted CV for ORR measured in 0.1 M NaOH showed an onset potentials at -500 mV for oTi₃C₂T_x and at -550 mV for Ti₃C₂T_x with a maximal current density of 80 mA cm^-2 for oTi₃C₂T_x and 63 mA cm^-2 for Ti₃C₂T_x. In a previous study an onset potential for ORR in 0.1 M NaOH of -450 mV and -700 mV was reported for graphene with a current density of 180 mA cm^-2 using a rotating disc electrode (1,000 rpm) [57]. The same study showed that N-doped graphene showed an onset potential for ORR in 0.1 M NaOH of -200 mV with a limiting current density of ~800 mA cm^-2 using a rotating disc electrode (1,000 rpm), while Pt/C electrode showed a limiting current density of ~220 mA cm^-2 [57]. Furthermore, effectivity of ORR reaction can be enhanced by attachment of nanoparticles (i.e. Co nanoparticles) to N-doped graphene with an onset potential at -200 mV and a limiting current density of 4–5 mA cm^-2 [58]. Thus, even though oTi₃C₂T_x and Ti₃C₂T_x without any further modifications are not as good catalysts for ORR in acidic and alkaline solutions as the best catalysts described in the literature (such are for example noble metal-free, nitrogen and sulphur co-doped graphene/carbon-nanotube material decorated with Co nanoparticles offering current density of ~7 mA cm^-2 in alkaline and acidic media [59]), Ti₃C₂T_x or oTi₃C₂T_x could be a good substrate for accommodation of various types of nanoparticles for subsequent effective ORR reactions in both media. Additional possible application of Ti₃C₂T_x to perform ORR is in construction of enzymatic biosensors using oxidases, which upon enzymatic action consume oxygen as a co-substrate [42] for analysis of a wide range of analytes.

3.12. Electrochemical reduction of H₂O₂

Reduction of H₂O₂ on Ti₃C₂T_x modified electrode started with an onset potential of -160 mV, comparable to the results obtained on chemically reduced GO [55] or carbon nanotube modified electrode [60]. Thus, it can be concluded that Ti₃C₂T_x modified electrodes could be applied in oxidase-based biosensing as effectively as graphene-based devices. The results indicated that H₂O2 can be effectively reduced by Ti₃C₂T_x modified GCE and less effective by oTi₃C₂T_x modified GCE (Fig. S18). Moreover, intercalation of DMSO into Ti₃C₂T_x during Ti₃C₂T_x sonication resulted in a less effective redox behavior towards H₂O₂ reduction both for Ti₃C₂T_x and oTi₃C₂T_x modified GCE compared to GCE modified only by Ti₃C₂T_x or oTi₃C₂T_x without being intercalated with DMSO (Fig. S18). This is why we tested Ti₃C₂T_x modified GCE as a sensor for detection of H₂O₂ at an applied potential of -500 mV (Fig. 10). The noise of the sensor prior H₂O2 addition was approx. 30 nA and if we apply S/N=3 for calculation of a limit of detection (LOD) then we could get LOD of 0.7 nM and if we take into account noise level of approx. 150 nA, after addition of H₂O₂, then we can calculate LOD as 3.5 nM (Fig. 10). The sensor towards H₂O₂ exhibited sensitivity of detection of 596 mA cm^-2 mM^-1 (Fig. S19). The response time for detection of H₂O₂ was approx. 10 s (Fig. 10). The H₂O₂ sensor based on Ti₃C₂T_x is much more sensitive compared to previously published H₂O₂ sensors with sensitivity up to 1.08 mA mM^-1 cm^-2 and LOD down to 20 nM [61–64]. There are however some papers reporting similar sensitivity or lower detection limits. For example Prussian blue based nanoelectrode array could reductively detect H₂O₂ down to 10 nM with a sensitivity of 60 mA μM^-1 cm^-2 [65], Prussian blue at Pt nanoparticles and carbon felt could detect H₂O₂ down to 1.2 nM with a sensitivity of 41 mA mM^-1 cm^-2 [66], Au-Pt nanoparticle-modified ionic liquid composite electrode could reductively detect H₂O₂ down to 0.3 nM with a sensitivity of 3.98 mA mM^-1 cm^-2 [67] and 3D porous Prussian blue layer deposited on graphene nanocomposite could detect H₂O₂ down to 5 nM [68].

Fig. 10.

Chronoamperogram recorded for Ti₃C₂T_x modified rotating disc electrode (RDE) in 0.1 M PB pH 7.0 with H₂O₂ additions at a working potential of -0.5 V. Arrow in the inset picture shows the first addition of stock H2O2 solution. Limit of detection was calculated as S/N = 3 from the 1^st H₂O₂ injection.

4. Conclusions

The study showed that exposure of Ti₃C₂T_x to an anodic potential induces formation of TiO₂, which is subsequently most likely etched from the Ti₃C₂F_x surface by present F^- ions. However, pristine Ti₃C₂T_x could be effectively applied in a cathodic potential window for sensing purposes. Results suggested that Ti₃C₂T_x exhibits low catalytic activity for ORR run either in acidic or alkaline media, but Ti₃C₂T_x was proved as an excellent catalyst for reduction of H₂O₂, and the H₂O₂ sensor based on Ti₃C₂T_x is the most sensitive device described so far with a detection limit of 0.7 nM comparable to the best device described so far (i.e. 0.3 nM) [67]. It is possible that further modification of Ti₃C₂T_x by metallic nanoparticles could further enhance performance of modified Ti₃C₂T_x to detect H₂O₂.

Supplementary Material

Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.electacta.2017.03.073.