Niels
Kubitza
a,
Isabel
Huck
a,
Hanna
Pazniak
b,
Curran
Kalha
c,
David
Koch
d,
Bo
Zhao
d,
Pardeep K.
Thakur
e,
Tien-Lin
Lee
e,
Aysha A.
Riaz
c,
Wolfgang
Donner
d,
Hongbin
Zhang
d,
Benjamin
Moss
f,
Ulf
Wiedwald
b,
Anna
Regoutz
c and
Christina S.
Birkel
*ag
aDepartment of Chemistry, Technische Universität Darmstadt, Germany
bFaculty of Physics and Center for Nanointegration Duisburg-Essen, University of Duisburg-Essen, Germany
cDepartment of Chemistry, University College London, London, WC1H 0AJ, UK
dInstitute of Materials Science, Technische Universität Darmstadt, 64287 Darmstadt, Germany
eDiamond House, Harwell Science and Innovation Campus, Fermi Ave, Didcot, OX11 0DE, UK
fDepartment of Chemistry, Molecular Science Research Hub, White City Campus, Imperial College London, London W12 0BZ, UK
gSchool of Molecular Sciences, Arizona State University, Tempe AZ-85282, USA. E-mail: Christina.Birkel@asu.edu
First published on 18th April 2024
MAX phases are almost exclusively known as carbides, while nitrides and carbonitrides form a significantly underrepresented subgroup even though they have been shown to possess enhanced properties in comparison to their carbide counterparts. One example is the nitride phase Cr2GaN which exhibits a spin density wave magnetic state below T = 170 K, while the metallic carbide phase Cr2GaC follows the MAX phase-typical Pauli-paramagnetic behavior. To investigate the influence on the materials/functional properties of mixing carbon and nitrogen on the X-site, this study aims to synthesize and comprehensively characterize the hitherto unknown carbonitride phase Cr2GaC1−xNx and compare it to the parent phases. Due to the challenging synthesis of (carbo)nitrides in general, a sol–gel-assisted approach is applied which was recently developed by our group. This process was further improved by using time-efficient microwave heating, leading to a highly phase pure product. STEM-EDX analyses reveal a C/N ratio of roughly 2:
1. Temperature-dependent XRD measurements confirm the literature-known magnetic phase transition of the parent nitride phase Cr2GaN, while the incorporation of carbon suppresses the latter. Nonetheless, magnetic characterization of the phases reveals that the magnetic behavior can be specifically influenced by changing the composition of the X-site, resulting in an increase of the susceptibility by increasing the nitrogen amount. Overall, these findings further substantiate the big potential in nitrogen-containing MAX phases, which will also serve as starting materials for future doping studies, i.e. on the M- and A-site, and as precursors for novel 2D MXenes.
Liu et al. comprehensively characterized Cr2GaN using temperature-dependent magnetic, as well as resistivity measurements and found a spin-density-wave transition (SDW) state below TN = 170 K possibly due to Fermi surface nesting.30 Subsequent temperature dependent X-ray diffraction studies by Tong et al. showed an anomalous c/a increase of Cr2GaN at around 170 K that further substantiated the origin of the occurring SDW state.33 On the contrary, the isostructural carbide pendant Cr2GaC did not show any anomalous temperature-dependent structural changes and thus no anomalies in the physical properties.33 The magnetism of the pure parent phase without any doping elements is restricted to Pauli paramagnetism.30,33 In 2015, further investigations within the Cr2GaN system were made by Li et al. by doping Cr2GaN with germanium to induce superconductivity.34 Even though suppression of the SDW state in Cr2GaN was realized by germanium-doping, no superconducting behavior was observed.34 Instead of doping on the A-site of the MAX phase, producing a solid solution between the carbide and nitride (hence, doping on the X-site) is also expected to slightly change the density of states at the Fermi level (EF) and thus possibly leading to interesting cooperative phenomena like (anti)ferromagnetism or superconductivity. Therefore, we here report the synthesis and comprehensive characterization of the carbonitride phase Cr2GaC1−xNx. The hitherto unknown carbonitride phase was synthesized by a sol–gel assisted solid-state method that has recently been developed by our group in order to circumvent the above stated challenges accompanied with the synthesis of (carbo)nitrides.35 Additionally, microwave heating was used for heating the reaction mixtures and to drastically reduce the reaction time for the synthesis of the new carbonitride and the carbide parent MAX phase. The products were structurally characterized by means of (temperature-dependent) X-ray powder diffraction (XRD), electron microscopy (SEM/TEM), and soft and hard X-ray photoelectron spectroscopy (SXPS/HAXPES). Additionally, functional properties were evaluated by magnetometry, as well as electronic transport and heat capacity measurements. Density functional theory (DFT) calculations support the structural physical characterization of the materials.
![]() | ||
Fig. 1 Synthesis scheme for the preparation of the target materials Cr2GaC, Cr2GaN, and the new carbonitride Cr2GaC1−xNx. |
All precursor amounts were based on 0.5 g of the desired MAX phase. A detailed summary can be found in the ESI† (Table S2). First, gallium flakes (Alfa Aesar, > 99%) were cut under atmospheric conditions and subsequently transferred into an argon-filled glovebox. Inside the glovebox, the gallium flakes were loosely mixed with the remaining reactants (chromium, 99%, Sigma-Aldrich; chromium nitride, Alfa Aesar; graphite, >99%, Alfa Aesar) according to the reaction equations above and pressed into a dense pellet (ϕ = 10 mm, 3 t, 5 s). Afterwards, the pellets were transferred into fused silica ampoules and heat treated in the microwave oven (CEM Microwave Technology Ltd) under flowing argon or vertical tube furnace (Carbolite) under vacuum (Table S3, ESI†). Prior to characterization of the samples, the pellets were finely ground using an agate mortar and stored under atmospheric conditions.
Ab initio calculations based on density functional theory (DFT) were performed using the Vienna Ab initio Simulation Package (VASP),38,39 which the projector augmented wave implemented. The generalized gradient approximation (GGA) of Perdew–Burke–Ernzerhof was used as the exchange correlation functional.40 The wave functions were expanded with a 24 × 24 × 6 Monkhorst–Pack k-sampling grid and cutoff energy of 500 eV. Calculations for different ratios of N and C were performed within the virtual crystal approximation (VCA) model.41 The precision of total energy convergence for the self-consistent field (SCF) calculations was 10−6 eV. All structures were fully optimized until the maximal Hellmann–Feynman force was less than 10−3 eV Å−1.
SEM images were taken at the XL30 FEG (Philips) using an acceleration voltage of 20 kV adapted with an APOLLO X-SDD detector (EDAX) for collecting EDX data. The EDX data were evaluated using the software EDAX GENESIS.
Bright-field (BF) high-resolution TEM (HRTEM) and scanning TEM (STEM) images were acquired with a JEOL 2200FS transmission electron microscope at an acceleration voltage of 200 kV using a 2k × 2k GATAN UltraScan1000XP CCD camera. The local chemical composition was determined using EDX in STEM mode with an Oxford windowless 80 mm2 SDD X-MaxN 80 TLE detector with 0.21 sr solid angle. HRTEM and EDX data were analyzed using Gatan Micrograph Suite and Oxford's Aztec software, respectively.
SXPS measurements were initially conducted using a laboratory-based Thermo Scientific K-Alpha XPS instrument (hv = 1486.7 eV), however, at this photon energy the family of Ga LMM Auger peaks (located between kinetic energies of 790–1100 eV) overlaps with the key N 1s and Cr 2p core levels (Fig. S3/S2, ESI†). For this reason, synchrotron-based XPS measurements were conducted, where the tunability of the soft X-ray energy enables moving the Auger lines relative to the core levels. Soft and hard X-ray photoelectron spectroscopy (SXPS and HAXPES) measurements were conducted at beamline I09 (surface and interface structural analysis) at the Diamond Light Source, UK. SXPS measurements were conducted at a photon energy of 1794.3 eV (1.8 keV) to avoid the overlap of crucial core states with the Ga LMM Auger peaks. This photon energy was achieved by using a 400 lines per mm plane grating monochromator, which provided a total energy resolution of approximately 400 meV (determined by extracting the 16%/84% width of a clean polycrystalline gold foil Fermi edge). HAXPES measurements were conducted at a photon energy of 5926.7 eV (5.9 keV), achieved using a Si(111) double crystal monochromator and a Si(004) post-channel-cut crystal, providing a total energy resolution of approximately 300 meV. The end-station at I09 operates at a base pressure of 3 × 10−10 mbar and is equipped with a VG Scienta Omicron EW4000 high-voltage electron analyzer with a ±28° wide acceptance angle. Samples were mounted on adhesive conductive carbon tape with the X-ray spot for both soft and hard X-rays converging on the same sample position. Survey, key core level (Ga 2p3/2, Cr 2p, O 1s, N 1s, C 1s), and valence band (VB) spectra were acquired at both photon energies.
Electronic transport, vibrating sample magnetometry (VSM) and specific heat measurements were studied in a PPMS DynaCool system (Quantum Design). For VSM, dried powders (20–30 mg) were weighed and put into polymer capsules. Measurements were taken within the field range of ±9 T at variable temperatures ranging from 3 to 400 K. Resistivity measurements were carried out in four-point in-line geometry at a constant current of 10 mA using a custom-made sample holder and the electrical transport option (ETO). Dense pellets were prepared by placing powder samples into a 12 mm diameter cylindrical mold and applying 7.5–9 tons using a manual hydraulic press. Pressed samples with an average thickness of 0.3 mm (Cr2GaN) and 0.44 mm (Cr2GaC1−xNx) were carefully cut into a rectangle (5 mm wide and 10 mm long). Specific heat was measured using the heat capacity option. The Cr2GaC1−xNx sample powder (92 mg) was pressed in cylindrical shape with 3 mm diameter by a hydraulic press at 0.5 tons.
![]() | ||
Fig. 2 Results of the Rietveld refinements (orange lines) and residuum curves (grey) of the X-ray powder diffraction data (black dots) of Cr2GaN, Cr2GaC and Cr2GaC1−xNx. The refinements were conducted based on structural models of Cr2GaN42/Cr2GaC43 (black), Cr3GaN (red),44 CrGa445 (blue), Cr3Ga46 (purple). |
Phase name | Cr2GaC | Cr2GaC1−xNx | Cr2GaN |
---|---|---|---|
Space group | P63/mmc | P63/mmc | P63/mmc |
Lattice parameters/Å | a = 2.884(2) | a = 2.891(3) | a = 2.882(3) |
c = 12.605(2) | c = 12.618(2) | c = 12.722(2) | |
Cell volume/Å3 | 90.82(2) | 91.36(2) | 91.52(2) |
Background order | 15 | 15 | 15 |
R P | 3.74 | 3.15 | 3.19 |
R wp | 5.73 | 5.09 | 4.55 |
R exp | 2.09 | 1.82 | 2.04 |
GOF | 2.74 | 2.80 | 2.23 |
Besides room-temperature X-ray powder diffraction data, temperature dependent (20–300 K) diffraction data were obtained (Fig. 2) in order to investigate possible structural anomalous behavior that was initially observed for the nitride phase by Tong et al.33 The latter was attributed to a spin density wave (SDW) transition, shown with an abrupt increase of the c/a lattice parameter ratio at 170 K. Thus, temperature-dependent diffraction data can be used as an indicator for an SDW state in the carbonitride phase. Analogous to the room temperature diffraction data, Rietveld refinements were conducted to extract the temperature-dependent lattice parameters of the samples. As shown in Fig. 3(a), the a-lattice parameter decreases monotonically with decreasing temperature for all three compounds. The c-lattice parameter, however, decreases for both, the Cr2GaN and Cr2GaC phase, whereas the c-lattice parameter of the nitride phase anomalously increases below 140 K (Fig. 3(b)). In general, these results are in good agreement with the reported data of Tong et al.,33 however, the anomalous c-lattice parameter change appears at ∼30 K lower temperatures than what was reported previously. For the carbonitride phase Cr2GaC1−xNx, no anomalous structural change was observed, rather the behavior was similar to the carbide phase Cr2GaC. Since no anomalous change in the structure is obtained, the SDW state is suppressed in the carbonitride but can be reproduced for the nitride.
For further characterization of the synthesized MAX phases, electron microscopy studies were conducted. However, due to the literature known parent phases, the following data are restricted to the new carbonitride Cr2GaC1−xNx. SEM micrographs reveal its morphology, which can be described as a mixture of typical MAX phase layered structures (Fig. 4(c)), as well as particles, whose surfaces are covered with smaller drop-like substructures (Fig. 4(a)), typical for sol–gel based syntheses.47,48 On the other hand, TEM micrographs show a mixture of elongated and spherical particles (Fig. 4(b)), whereas the corresponding HRTEM micrograph (Fig. 4(d)) reveals the high crystallinity of the sample which is supported by the fast Fourier transformation (FFT) shown in the inset of Fig. 4(d).
Additionally, the FFT of the HRTEM image was used to calculate the averaged d-spacings from the distances of all spots to the origin. This allows to determine the zone axis to [001] and to assign Miller indices hkl to the obtained spacings of d100 = 2.45 Å and d110 = 1.37 Å. These results are in reasonable agreement with the d-spacings derived by the XRD data (d100 = 2.504 Å and d110 = 1.446 Å) by considering a larger error bar of the FFT evaluation.
Due to the calculated and measured non-linear evolution of the lattice parameters of the carbonitride phase based on the amount of the X-element, XRD cannot deliver the stoichiometry. Instead, STEM-EDX mappings were conducted to estimate the C/N ratio in the MAX phase. In Fig. 5, a representative elemental map of a particle of the investigated phase is shown. All expected atomic signals of the MAX phase are detected, however, the carbon and nitrogen signals are slightly inhomogeneously distributed. This can most likely be explained by amorphous carbon residues as a result of the sol–gel-based synthesis approach as well as by a varying thickness or self-shadowing effects in the direction of the EDX detector. Nonetheless, both elemental signals are present over the whole particle. Averaged over the whole area, the elemental ratio of the phase can be quantified as Cr51Ga24C17N8, leading to an estimated overall C/N ratio of ∼ 2:
1. The detected oxygen signal can be explained by surface oxidation of the particles originating from the sol–gel based synthesis procedure and treatment under ambient conditions.
![]() | ||
Fig. 5 STEM-EDX mappings showing the distribution of the elements in the carbonitride MAX phase, including a representative EDX spectrum for determining the C/N ratio. |
To complement the XRD data and infer the chemical state of the material, SXPS and HAXPES data on the three samples were collected at beamline I09 at photon energies of 1.8 and 5.9 keV, respectively. Changing the photon energy (hv) between the soft and hard X-ray regimes during the measurements allows for control over the probing depth, a strategy exploited in the authors' previous work.35 The maximum relativistic inelastic mean free path (IMFP) of photoelectrons exiting the Cr2GaC sample surface can be calculated using the TPP-2M predictive formula implemented in the QUASES software package.49 Cr2GaC does not exist in the QUASES database. Therefore, a new material was created (assuming a bulk density of 7.09 g cm−3,50 atomic mass of 185.73 u, 19 valence electrons, and metallic behaviour (no band gap)). The maximum relativistic IMFP of photoelectrons from Cr2GaC at a photon energy of 1.8 and 5.9 keV is calculated to be 2.7 and 7.2 nm, respectively, equating to an estimated probing depth of 8.1 and 21.6 nm (assuming 3 × IMFP = probing depth). The survey spectra collected at both photon energies (Fig. S4, ESI†) display strong signals from the expected elements (Ga, Cr, O, C and N). Minor signals from chlorine (≈200 and 270 eV) and sodium (≈1071 eV) are also observed in the spectra of the Cr2GaN and Cr2GaC samples, respectively. SXPS shows that using a photon energy of 1.8 keV ensured that the Ga LMM Auger peaks did not overlap with the key core levels. Fig. 5 displays the Ga 2p3/2, Cr 2p, N 1s, and C 1s core level spectra as a function of photon energy. The O 1s and valence band spectra collected with SXPS and HAXPES can be found in Fig. S5 (ESI†). The collected valence band spectra (see Fig. S5(b/d), ESI†) show that for all samples a distinct Fermi edge is observed, with it being more distinct when measured with HAXPES, indicating metallic character. The presence of a Fermi edge enables the binding energy (BE) scale of all core level spectra to be referenced to the intrinsic Fermi energy (EF) of the respective samples. Due to the nature of the samples and the lack of charge compensation mechanisms at the synchrotron, partial charging occurred during the measurements, leading to different levels of distortion on the higher BE side of the spectra. This was also experienced in the authors' previous work on the V-Ga-C/N MAX phases.35 Nevertheless, the spectra still provide valuable insight into the chemical states of the samples, but quantification of the chemical states cannot be performed. The Ga 2p3/2 spectra displayed in Fig. 6 (a) show that when measured with SXPS, a main intensity peak at 1118.6 eV is observed, commensurate with an oxygen-terminated gallium (Ga–O) chemical environment.35 This peak is the most intense for the mixed carbonitride phase. A subtle asymmetry is also observed on the lower BE side of this peak when measured with soft X-rays, which is attributed to the Ga-C/N environment.35Fig. 6 (e) shows that with increasing photon energy (and therefore increasing probing depth), the intensity of this Ga-C/N environment (BE = 1116.5 eV) significantly increases relative to the Ga–O peak intensity. This suggests that the surface of these samples is oxidized, whereas the bulk is carbonitride-rich. An additional peak and/or broadening is observed on the higher BE side of the Ga–O peak with both soft and hard X-rays (labelled with an asterisk, *). The assignment of this spectral feature is difficult to confirm owing to the partial charging of the spectra and, therefore, cannot be discussed further. Similar charging effects are found in the other core level spectra. Additionally, the XRD data showed side phases (albeit at low concentration), and the resulting chemical environments will contribute to the core level spectra. The Cr 2p spectra in Fig. 6(b) and (f) display similar attributes to the Ga 2p3/2 core level in that a low BE peak at 574.3 eV (Cr 2p3/2) is observed, and its signal intensity is enhanced with HAXPES. This peak is attributed to the primary Cr-C/N environment,51,52 and given the depth sensitivity of this peak when measured with HAXPES, it again suggests that the carbonitride is situated toward the bulk of the sample. The higher BE features observed in the Cr 2p spectra are associated with an oxygen-terminated Cr environment (Cr–O).51,52 As for the Ga 2p3/2 core level, the side phases observed in XRD will also contribute to the Cr 2p spectra. The N 1s and C 1s spectra displayed in Fig. 6(c) and (g) and (d) and (h), respectively, provide direct confirmation that the desired carbonitride phases were obtained for the three samples. The lower BE peak at approx. 283.0 eV in the C 1s spectra is commensurate with a metal–carbide environment (C–Cr/Ga),35,53 and is only observed in the Cr2GaC and Cr2GaC1−xNx samples. In the N 1s spectra, the lower BE peak at 397.8–398.0 eV is commensurate with a metal–nitride environment (N–Cr/Ga),35,54 and is only observed in the expected sample. In both the C/N 1s spectra, the carbonitride environments are more easily observed with HAXPES, which indicates that the bulk of these samples is carbonitride-rich, whereas the surface is oxidized. Furthermore, in both the N and C 1s spectra, the intensity of metal carbide/nitride peaks increase with respect to the total Ga 2p3/2 in the expected trend. Higher BE features are also observed in both C/N 1s spectra, with the C 1s spectra dominated by graphitic sp2 carbon, adventitious carbon (C0) and carbon–oxygen species, whereas in the N 1s spectra a broader peak is observed at +2.8 eV from the N–Ga/Cr and is most likely attributed to organic nitrogen species (e.g. C–NHx).55 Lastly, it is noted that with respect to the total Ga 2p3/2 peak intensity, the mixed carbonitride sample has the most intense C 1s signal, in both soft and hard XP spectra. Fig. 6(d and h) show that the total normalized C 1s signal intensity of the mixed carbonitride sample is roughly three times that of the signal measured on the nitride and carbide samples. The reason for the dominant C signal in the carbonitride sample can be related to the synthesis procedure. A sol–gel synthesis approach was employed, using urea as a gelling agent. During the heat treatment step, urea will be converted into amorphous carbon that is also retained in the product phase. The SEM images (Fig. 4) verify that free amorphous carbon particles are indeed present at the surface and throughout the sample. This also explains why the dominate peak within the C 1s spectra for the mixed carbonitride sample displays an asymmetric profile, commensurate with a graphitic sp2-like environment.52 To conclude, the variance in probing depth achieved with the combination of SXPS and HAXPES confirms that all samples are carbo/nitride-rich in the bulk with an oxide overlayer on the surface. Furthermore, we obtain the expected carbo/nitride chemical environments for each sample as well as showing the influence of the synthesis procedure on the final products.
![]() | ||
Fig. 7 (a) Temperature-dependent magnetization curves of Cr2GaC, Cr2GaC1−xNx, and Cr2GaN in a field of B = 10 mT and a temperature range of 3–400 K. (b) Field-dependent magnetization of Cr2GaC1−xNx at various temperatures between 3–300 K. (c) Magnetic susceptibility from the field-dependent magnetization data of Cr2GaC, Cr2GaC1−xNx, and Cr2GaN at 300 K. The susceptibility is extracted from linear fitting of the signal in the interval |B| = 4–9 T. The nitrogen amount of the carbonitride is based on the EDX measurements presented in Fig. 4. |
Footnote |
† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d4tc00067f |
This journal is © The Royal Society of Chemistry 2024 |