Bulk and Surface Chemistry of the Niobium MAX and MXene Phases from Multinuclear Solid-State NMR Spectroscopy

MXenes, derived from layered MAX phases, are a class of two-dimensional materials with emerging applications in energy storage, electronics, catalysis, and other fields due to their high surface areas, metallic conductivity, biocompatibility and attractive optoelectronic properties. MXene properties are heavily influenced by their surface chemistry, but a detailed understanding of the surface functionalization is still lacking. Solid-state nuclear magnetic resonance (NMR) spectroscopy is sensitive to the interfacial chemistry, the phase purity including the presence of amorphous/nanocrystalline phases, and the electronic properties of the MXene and MAX phases. In this work, we systematically study the chemistry of Nb MAX and MXene phases, Nb 2 CT x and Nb 4 C 3 T x , with their unique electronic and mechanical properties, using solid-state NMR spectroscopy and examine a variety of nuclei ( 1 H, 13 C, 19 F, 27 Al and 93 Nb) with a range of one-and two-dimensional correlation, wideline, high-sensitivity, high-resolution, and/or relaxation-filtered experiments. Hydroxide and fluoride terminations are identified, found to be intimately mixed, and their chemical shifts are compared with other MXenes. This multinuclear NMR study demonstrates that diffraction alone is insufficient to characterize the phase composition of MAX and MXene samples as numerous amorphous or


Introduction
[3][4] MXenes are a similarly diverse class of 2D compounds derived from MAX phases via etching of the A-site atoms. 5It is wellestablished that MXenes are surfaceterminated, and thus better represented by the formula Mn+1XnTx (T = termination), however the nature of the surface composition, coordination, and spatial arrangements are exceptionally challenging to ascertain (Figure 1).Both material families have extremely diverse applications: MAX phases are used for industrial components such as heating elements and gas burner nozzles with high temperature applications 1,3,6 and MXenes are being tested as functional materials for applications in energy storage, 7,8 hydrogen 9,10 and oxygen 11,12 evolution catalysis, cancer therapeutics, 13 and electromagnetic interference shielding 14,15 .The electronic and functional properties of MXenes depend strongly on the surface terminations, [16][17][18] so a deeper understanding of the surface chemistry is required to optimize the performance in these applications.As the first reported MXene, Ti3C2Tx has been the most thoroughly studied member of this family, while many other MXenes have been, and continue to be, synthesized and applied to diverse functions.In this work, on the other hand, we focus on the less studied 4d transition metal-based niobium MAX and MXene phases including Nb4AlC3, Nb4C3Tx, Nb2AlC, and Nb2CTx, and Nb2SnC.While Nb2AlC 19 and Nb2SnC 20 were reported in the 1960s, it was not until the 2000s that Nb4AlC3 emerged 21 via a synthesis strategy borrowed from the synthesis of the Ta MAX series.The syntheses of Nb2CTx 22 and Nb4C3Tx 23 MXenes were then reported in 2013 and 2014, respectively, by selective etching of Al from Nb2AlC and Nb4AlC3, following the same methodology as for Ti3C2Tx.Neither the intermediate compound Nb3AlC2 nor its MXene derivative have been reported; density functional theory (DFT) calculations suggest that this 312 MAX phase is dynamically unstable. 24The Nb MAX and MXene phases have various interesting properties and potential applications including in biomedical 13,25 and optical 26 applications.Nb2CTx 22 and Nb4C3Tx 27 have attracted recent interest as lithium-ion battery anodes.Nb4C3Tx is one of the most conductive and strongest MXene phases, 23 which makes it promising for field-effect transistors and mechanical applications. 28Nb2SnC 29 and Nb2AlC 30 are superconducting up to 7.8 K and 0.440 K, respectively.Superconductivity was recently observed in Nb2CTx at 4.5 to 7.1 K, the temperature depending on the surface termination. 18e structures of the Nb MAX and MXene phases are shown in Figure 1, with the firstshell coordination numbers and geometry given in Table 1.Nb4AlC3 and Nb4C3Tx contain two distinct crystallographic Nb sites and two C sites.Nb2 and C2 are the sites nearer to the Al/etched layer and will thus be referred to as 'external' while Nb1 and C1 are denoted 'internal' following the assignments in the Inorganic Crystal Structure Database crystal structure files 160755 21 and 606236 19 .Nb2AlC and Nb2CTx contain one Nb site and one C site; these are necessarily 'external'.Both MAX phases contain a single Al environment with trigonal prismatic coordination to six niobium atoms.Note that the MXene coordination in Table 1 is described in the absence of termination groups (T).
Solid-state NMR spectroscopy is a versatile toolkit that is sensitive to bonding and electronic properties with the ability to probe different elements individually.Additionally, there is no requirement for long-range periodic order in NMR, therefore, it can be readily applied to defect and surface chemistry as well as to nanoparticles.In particular, NMR can reveal the presence of amorphous impurities which are invisible to diffraction techniques.All the elements in the niobium aluminum carbide phases are NMR active and relatively accessible: 13 C is I = ½ with a natural abundance of 1.1%, while 93 Nb (I = 9/2) and 27 Al (I = 5/2) are 100% naturally abundant quadrupolar nuclei.The quadrupolar nature of the latter two nuclei means that they are extremely sensitive to the symmetry of their local environment.Despite the analytical utility, NMR investigations of MAX and MXene phases are sparse.Lue et al. measured static 27 Al NMR spectra of the 3d transition metal MAX series Ti2AlC, V2AlC, and Cr2AlC, 31 finding relatively long nuclear relaxation times and systematically decreasing nuclear quadrupole coupling constants and Knight shifts from Ti toward Cr. 31 The first NMR study of a MXene was by Harris et al. who identified -OH and -F surface terminations in V2C by 1 H and 19 F NMR spectroscopy and demonstrated the proximity of 1 H nuclei to the MXene surfaces by using 1 H→ 13 C cross polarization (CP). 32Hope et al. later quantified the concentrations of -OH and -F terminations on Ti3C2 MXene, which were found to be highly dependent on the synthesis method, and showed the terminations to be intimately mixed using 1 H- 19 F NMR correlation experiments. 33These observations have been extended to further Ti3C2Tx syntheses and preparation conditions in a pair of very recent reports. 34,35 this work, a comprehensive multinuclear NMR study ( 93 Nb, 27 Al, 13 C, 1 H, and 19 F) of the bulk structures and surface chemistry of the known Nbn+1AlCn MAX and Nbn+1CnTx MXene phases is performed with a range of one-and two-dimensional techniques, with additional comparisons to Nb2SnC and the 47/49 Ti NMR of Ti3AlC2 MAX phase.These experiments are supported by DFT-based electronic structure calculations of the electric field gradient (EFG) at the nucleus.The quadrupolar parameters determined by high-resolution magic-angle spinning (MAS) and static wideline 93 Nb and 27 Al spectra give insights on the local structures of the MAX phases, while 93 Nb and 13 C spectra reveal the nature of the etching to the corresponding MXene phases, with the 93 Nb NMR further confirming that the latter are surface terminated. 1 H, 19 F and 93 Nb two-dimensional NMR spectroscopy then identifies the chemistry and connectivity of the MXene surface terminations.Finally, the presence of diffraction-silent side-products is confirmed, including the observation of aluminum oxides via the 27 Al NMR spectra.Overall, these results have important implications for the synthesis, characterization, and functional properties of Nb MAX and MXene phases.

Experimental
Synthesis.Nb2AlC MAX phase was prepared as described per Naguib et al. 22 Elemental powders of niobium (99.9% metals basis, 325 mesh), aluminum (99.8% purity, 300 mesh), and carbon black (99% purity, 300 mesh) were mixed in a plastic jar with a ratio of 2:1.1:1Nb:Al:C and ball milled.The powder was placed into an alumina boat and heated under Ar flow to 1600 °C, with a ramp rate of 4 °C•min -1 , and sintered for 4 h.After natural cooling to room temperature, the MAX phase was milled into powder and sieved through a 400-mesh sieve.
To prepare Nb2CTx MXene, 1.0 g of Nb2AlC powder was transferred into 20 mL of HF (aqueous, 49%, Millipore-Sigma) over 60 s and stirred for 48 h at 50 °C.The mixture was washed several times by centrifugation at 3500 rpm (5 min/cycle), the supernatant was decanted, and deionized (DI) water was added until the supernatant reached a pH >6.
Multilayered MXene powder was collected by filtration through a cellulose acetate membrane (0.45 µm pore size).The delamination of Nb2CTx was carried out by transferring 1.0 g of MXene powder into 10 mL of a diluted tetramethylammonium hydroxide (TMAOH) solution (25% in H2O, Millipore-Sigma).The solution was stirred overnight (~18 h) at room temperature.To remove excess TMAOH, the mixture was washed by centrifugation with DI water (50 mL) at 3500 rpm for 20 min followed by decantation of the supernatant.The process was repeated until the supernatant reached a pH <8.Then, the sediment was dispersed in 50 mL of deionized water and sonicated for 1 h in an ice-bath with Ar bubbling.Finally, the mixture was centrifuged for 1 h at 3500 rpm, and the supernatant was carefully removed to leave the MXene sediment in the bottom of the tube.
To prepare Nb4AlC3, powders of niobium (99.9% metals basis, 325 mesh), aluminum (99.8% purity, 300 mesh), and carbon black (99% purity, 300 mesh) were mixed and the synthesis was performed as previously reported. 27To prepare Nb4C3Tx, 0.4 g Nb4AlC3 powder was added to 30 mL HF solution (aqueous, 49%, Millipore-Sigma) and stirred at room temperature (20-25 ℃) for 6 days.The produced acidic mixture was washed by DI water followed by centrifugation (3500 rpm, 2 min per cycle).After each centrifugation cycle, the supernatant was discarded, and the sediment was dispersed in DI water until neutral pH (~7) was reached.In order to delaminate the Nb4C3Tx, 1 mL of TMAOH (25% in H2O, Millipore-Sigma) was mixed with 9 mL DI H2O, added to Nb4C3Tx, and shaken for 15 minutes at room temperature.The excess TMAOH was separated from the product by repeated centrifugation at 3500 rpm.kHz with a 2.5 mm HX probe with recycle delays of 1 s or 3 s. 27Al and 93 Nb spectra were recorded at 16.4 T with a 1.3 mm or 4.0 mm HXY probe under static or MAS conditions with a single pulse or a Hahn echo pulse sequence and a recycle delay of 0.3-100 s for 27 Al and 0.1-0.5 s for 93 Nb, as described in each figure.Broad static 93 Nb spectra were collected with the assistance of an external automatic tuning/matching (eATM) robot 37 .These variable-offset cumulative spectra (VOCS) measurements were recorded in steps of 170 kHz (~1000 ppm). 47,49Ti spectra were recorded with a CPMG pulse sequence using VOCS acquisition in steps of 80 kHz (~2000 ppm) with 115 echoes of 0.8 ms each and recycle delays of 0.1 s; the subspectra were combined by taking the skyline projection: 38 for each point, the highest intensity value from the overlapping sub-spectra is taken, rather than adding their intensities together.
1 H spectra were referenced to adamantane at 1.81 ppm, 13 C spectra to the tertiary carbon of adamantane at 38.5 ppm, 19 F spectra to LiF at −203 ppm, 27 Al spectra to AlF3 at -15 ppm, 93 Nb spectra to LiNbO3 at -1004 ppm 39 , and 47/49 Ti spectra to the 49  The Rose convention (ZYZ) was used to define the Euler angles relating the orientations of the shift and quadrupolar tensors.
Calculations.Density functional theory (DFT) calculations were performed with the CASTEP planewave pseudopotential code. 41The calculations used the Perdew-Burke-Ernzerhof (PBE) exchange-correlation functional 42 and Vanderbilt ultrasoft pseudopotentials. 43MAX crystal structures were used as starting models. 19,21Prior to the EFG calculations, atomic positions were optimized (with lattice parameters held constant) until the force on any atom was smaller than 10 meVÅ -1 .All calculations used a planewave energy cut-off energy of 600 eV and a Monkhorst-Pack 44 grid with a spacing finer than 2  0.04 Å -1 to sample the Brillouin zone.
Computed EFG parameters were used as the starting point to fit the experimental 27 Al and 93 Nb spectra.

Results and Discussion
MAX bulk structure characterization.XRD analysis of the Nb2AlC sample (Figure S1) showed Nb4AlC3 as a minor (~5% from Rietveld refinement) secondary phase.No additional peaks were visible beyond those of the two MAX phases.The Nb4AlC3 diffraction pattern showed a NbC-like impurity (~8%) and a small, broad unindexed peak just above the background signal at 26.7; no Nb2AlC was detected in the Nb4AlC3 sample. 27Al NMR spectra of the MAX phases Nb2AlC and Nb4AlC3 are given in Figures 2 and   3, respectively.For each compound, a simultaneous fit was performed over the static and MAS spectra (Table 2).The DFT calculated quadrupolar parameters served as useful starting points to guide the fits (Table 2).Static spectra were sensitive to the value of the CSA while MAS removed CSA broadening and resulted in spectra that were more sensitive to the isotropic shift The static spectra were T1-filtered to facilitate efficient recording of the fast-relaxing MAX signal whereas the MAS spectrum was recorded under quantitative conditions to observe all aluminum environments.Spinning sidebands are denoted with asterisks.See Table 2 for fitting parameters.  2 for fitting parameters.The missing intensity in the fit of the static spectrum is ascribed to the AlO6 environment at ~16 ppm.
and quadrupolar tensor.The positions of discontinuities in the powder lineshape from satellite transitions, where observed, were also used to determine the quadrupolar parameters.The calculated quadrupolar coupling constants matched the experimental values to within <5%.
In addition to aluminum in the MAX phases, 27 Al NMR revealed the presence of secondary phases.At short recycle delays, the MAX phases, with fast spin-lattice, T1, nuclear relaxation from conduction electrons, appear relatively pure (Figure S2-3).However, at longer free induction decay (FID) intervals, the relative intensities of the impurities increase; an indication that the impurities are diamagnetic with longer T1 relaxation times.The Nb2AlC sample exhibits an unexpected resonance at 16 ppm while the Nb4AlC3 sample shows three additional 27  The oxide environments could arise from secondary oxide phases, or from the oxide layers previously observed by XPS to form on the surface of MAX particles. 47The resonance at 115 ppm appears at a higher frequency than for typical aluminum oxides or fluorides 48 ; it does coincide with the shift range of aluminum carbide Al4C3 (111-120 ppm) but the maximum CQ value for the observed resonance is ca.4.5 MHz while Al4C3 has two Al sites with CQ magnitudes of 14-17 MHz. 49,50yes et al. identified aluminum oxycarbide impurities with smaller CQs in their recent study of commercial Al4C3 so a related aluminum species is plausible here. 502][53][54] Note that the 27 Al signal from Nb2AlC covers a range of ~90-170 ppm, so the impurity at 115 ppm is less pronounced but it can be clearly seen as a shoulder that appears as the interval between scans increases (Figure S2).Quantitative spectra recorded with recycle delays of 90-100 s (Figure S2) revealed the relative ratios of the MAX and impurity (or surface) species in the samples from the perspective of aluminum content. 47For the Nb2AlC sample, a ratio of 89:9:2 was found for the MAX signal at 169 ppm and the impurity resonances at 16 and 115 ppm, respectively.For the Nb4AlC3 sample, a ratio of 56:34:7:3 was found for the MAX signal at 319 ppm and the impurity resonances at 17, 5, and 115 ppm, respectively.in NbC, 55 depending on the preparation conditions, with NbCx (0.7 ≤ x < 1) yielding quadrupolar 93 Nb signals. 56,57Overlapping lines preclude an accurate fitting of the quadrupolar parameters of the NbCx impurity in Nb4AlC3 here but a small CQ (~2 MHz) is consistent with the lineshape.For Nb2SnC (Figure S6), a sharp 93 Nb signal was observed with an isotropic shift of -189(3) ppm; this is 300 ppm higher frequency than the isostructural Nb2AlC.In the wideline static VOCS measurement, Nb metal was also observed as a secondary phase at ~7200 ppm.
The 13  V2AlC (208 ppm), 32 but less than for Ti3AlC2 (566 ppm) 33 .There is a significant shoulder for the 324 ppm peak which is ascribed to NbC, the single carbon site of which resonates at ~315 ppm (Figure S5b,c Nb4AlC3, although it is not observed in the X-ray diffraction pattern (Figure S1) because (i) it is present in nanoparticulate or amorphous form and/or (ii) its identification is obstructed by peak overlap.Note, that phases such as amorphous carbon and low levels of Nb2O5 would be difficult to observe via these NMR experiments due to the broad nature of those signals and/or overlap with other signals.The presence of alumina suggests that niobia could also be present, and Sarycheva and Gogotsi observed amorphous carbon and titania in a degraded Ti3C2Tx sample via Raman spectroscopy. 58

MXene bulk structure characterization
By etching out the aluminum, the MAX phases were converted to MXenes.However, the Nb2CTx also has a significant amount of unreacted Nb2AlC, while NbC remains present as an impurity in Nb4C3Tx, as previously seen for other MAX/MXene samples. 23,59e 93 Nb NMR spectra of Nb2CTx and Nb4C3Tx are shown in Figure 7 and Figure S7.
Since niobium atoms are located at the outer surface of the 2D MXene sheets, the number of distinct Nb local environments and NMR signals depends on the surface terminations.The 93 Nb lineshape of Nb2CTx was simulated with two quadrupolar lines, providing a reasonable fit to the data (Figure 7); the minor but sharper signal is consistent with Nb2AlC, the broader signal was attributed to Nb2CTx and was fit with iso = -790(30) ppm, CSA = -350(50) ppm, CSA = 0.5(3), CQ = 77(3) MHz, Q = 0.2(1), Euler angles ,, = 0, 50(10), 0.The CQ for Nb2CTx is more than double that of Nb2AlC, providing spectral evidence for the reduction of symmetry at the MXene surface.Furthermore, the non-zero ηQ indicates a breaking of the threefold symmetry of the Nb site at the surface, presumably due to the surface terminations.
Calculations for a bare Nb2C surface predict a 93 Nb CQ of 122 MHz, so the lower observed CQ is also evidence for the surface terminations which result in a more symmetric charge distribution than a bare surface.Basic ordered models for surface-terminated Nb2CTx (Tx = F2, (OH)2, (OH)F, O2, O) give calculated CQ values ranging from 36-109 MHz (Table S2), which are consistent with the experimental CQ of 77 MHz.The best agreement is for fluoride and oxide termination, which may suggest a greater proportion of fluoride and oxide terminations than hydroxyl terminations, as previously observed for Ti3C2Tx. 33Meanwhile, the 93 Nb lineshape of Nb4C3Tx is substantially broader.An exact fitting of the lineshape could not be obtained due to the difficulty of distinguishing a potential distribution of surface terminations, the two Nb sites in Nb4C3Tx, and the NbCx impurity, which all overlap.Nevertheless, based on the linewidth a maximum CQ of ~100 MHz can be determined, ruling out a significant proportion of bare surface (calculated CQ = 135 MHz for the outer Nb2 site, Table S3).Given that the calculated quadrupolar parameters for ordered models of surface terminations (Table S3) are in the range one side, but terminations on the other, or nanoscale clusters of unetched aluminum.After etching Nb4AlC3 (Figure 6b, top), a broad high frequency signal can be observed at ~400 ppm, corresponding to Nb4C3Tx MXene, although the inner and outer carbon sites cannot be resolved.After etching, the NbC impurity is still evident, as well as signals from other impurities in the sample at 31 and 111 ppm; the former is correlated with a 1 H signal at 1 ppm in the 1 H→ 13 C HETCOR spectrum (Figure S8c), suggesting that this impurity is predominantly alkyl, while the latter can be observed in a 19 F→ 13 C cross polarization (CP) experiment (Figure S8b), which, combined with the chemical shift, suggests a -CF2moiety in a fluorinated organic species.Notably, the 13 C shift increases on etching from the MAX to the MXene for both Nb2CTx and Nb4C3Tx; this is due to an increased Knight shift, corresponding to an increased density of states at the Fermi level for the carbon atoms in the MXene phases.The same increase in 13 C shift was observed for V2CTx (from 208 ppm to 260 ppm), 32 however, in contrast, the 13 C Knight shift actually decreases for Ti3C2Tx (from 566 ppm to 380-410 ppm) 33 .
A decrease in 13 C shift was also observed on etching Mo2Ga2C to Mo2CTx (from 178 ppm to 125 ppm), 61 although in this case the shift is well within the diamagnetic range, implying there is not a significant density of states at the Fermi level localized on the carbons.
As seen from the presence of the Nb2AlC MAX signal in the 93 Nb, 27 Al, and 13 C spectra, not all aluminum had been etched out during the HF treatment, consistent with prior work on other MXenes: V2AlC was observed in V2CTx samples and Ti3AlC2 in Ti3C2Tx samples in their respective 13 C NMR spectra. 32,33The 27 Al NMR spectra of the MXene samples (Figure S9) also show an AlO6-like aluminum oxide environment, while the Nb2CTx sample also shows a signal consistent with AlF3nH2O, a common impurity in MXenes 60 and one that can be readily identified with NMR spectroscopy.
Since diffraction did not reveal the Al-O, Al-F, or Al-(C,N,O) compounds identified here, these secondary phases must be amorphous, present as small nanoparticles, or hidden under the MAX/MXene diffraction peaks.These findings again highlight that the picture from diffraction alone can be misleading.Furthermore, NbCx was almost entirely hidden under the Nb4AlC3 XRD peaks.A method that is sensitive to amorphous phases such as solid-state NMR and/or quantitative diffraction with an internal standard should be routinely performed to characterize MAX and MXene phases.V2CTx at 85 and 27 ppm 32 and Ti3C2Tx between 12.5 and 20 ppm, the exact shift values depending on the synthesis procedure 33,34 .The free H2O signal can be reduced by drying the MXene at 200 C in vacuo (Figure 8b), after which an additional resonance can be distinguished at 12.2 ppm, as well as sharper signals between −0.5 and 1.6 ppm; the former could be due to a further -OH termination environment, or bound H2O partaking in strong hydrogen bonding, while the latter are assigned to intercalated volatile contaminants, as was also observed for Ti3C2Tx.There is also a loss of intensity for the 26.0 ppm signal in particular, which may be due to loss of -OH terminations as H2O during the drying process. 62 The assignment of the 1 H NMR resonances at high chemical shift to surface species on the MXene layers can be confirmed by 1 H/ 93 Nb TRAPDOR experiments: 63 for these spectra, a 1 H spin echo is recorded with continuous irradiation of 93 Nb during only the first half of the echo, thus any dephasing due to recoupled 1 H-93 Nb interactions is only partially refocused, resulting in a loss of signal intensity for 1 H environments in the vicinity of 93 Nb; when the difference is taken between spin echoes recorded with and without 93 Nb irradiation, these 1 H environments (near 93 Nb) can be selectively observed.The signals at 12.2, 19.5 and 26.0 ppm can all be clearly observed in the TRAPDOR difference spectra (Figure 8c and Figure S10), confirming their vicinity to Nb atoms at the MXene surface; the signals at less than 5 ppm, on the other hand, are suppressed in the TRAPDOR difference spectra.A weaker TRAPDOR effect is observed for the free H2O resonance at 6.3 ppm, which suggests that only some of the free H2O is proximal to Nb and/or that it is mobile; this may be evidence of chemisorbed water, as proposed by Persson et al. 64 The sideband-separated 19 F NMR spectrum of Nb2CTx, recorded with a MATPASS pulse sequence, 40 shows a number of different resonances (Figure 8d).The signal at −158 ppm is due to AlF3⋅nH2O, a by-product of the etching process also seen in the 27 Al spectrum (Figure S8); 33,65 this resonance is suppressed in the 19 F/ 93 Nb TRAPDOR difference spectrum (Figure 8f), leaving the signals at −20, −132 and −330 ppm, which must be near Nb. 1 H/ 19 F HETCOR experiments (Figure S11) are completely dominated by the H2O↔AlF3 correlation, so cannot be used to aid assignment of the 19 F spectrum.After drying at 200 C in vacuo the AlF3⋅nH2O resonance shifts to −164 ppm due to changes in hydration, so that the −132 ppm signal can be seen as a shoulder more clearly, and there is also a small, sharp peak at −121 ppm from a fluorinated organic impurity.The 19 F signal at −20 ppm is largely removed by drying, so may be tentatively ascribed to -F terminations with strong hydrogen bonding to H2O molecules, or adjacent -OH terminations (given that the intensity of the -OH signals in the 1 H NMR also decreases on drying).Conversely, the major 19 F signal at −132 ppm is assigned to -F terminations without strong H bonding.The minor signal at −330 ppm is strongly ionic; it may be a niobium fluoride species formed as a byproduct from excessive etching and could therefore be minimized by using milder etching conditions. 66e 1 H NMR spectrum of Nb4C3Tx (Figure 8g) is dominated by a peak at 1.0 ppm due to the organic impurity identified from the 1 H→ 13 C HETCOR experiment (Figure S8c).Further 1 H signals at 5.5 ppm and 15 ppm are assigned to H2O and -OH terminations, respectively.

Surface terminations of
Both the organic contaminant and the H2O signals can be reduced by drying in vacuo at 200 C overnight (Figure 8h), while the -OH terminations can be identified in the 1 H/ 93 Nb TRAPDOR difference spectrum (Figure 8i and Figure S10).In the TRAPDOR experiment, the 1.0 ppm peak is almost, but not entirely, suppressed; the residual signal and satellite peaks are due to the differential phase (Bloch-Siegert) shift 67,68 introduced by 93 Nb irradiation, which results in imperfect cancellation in the difference spectrum.
The 19 F MATPASS NMR spectrum for Nb4C3Tx is dominated by a resonance at −121 ppm (Figure 8j,k), which is consistent with the fluorinated impurity identified by 19 F→ 13 C CP.
Two other signals can also be distinguished, centered at ca. −190 and −80 ppm; these are assigned to -F terminations as they can clearly be observed in the 19 F/ 93 Nb TRAPDOR difference spectrum (Figure 8l), although the sidebands of the terminal -F signals are unresolved.Further evidence for the assignment of the -OH and -F terminations can be seen in the 1 H/ 19 F HETCOR spectra: in the 19 F→ 1 H spectrum (Figure 9b) a broad correlation is observed between both the H2O (6.0 ppm) and -OH (15 ppm) 1 H resonances and 19 F intensity between ca.50 and −300 ppm that corresponds to the signals from -F terminations at −190 and −80 ppm, with unresolved spinning sidebands.This supports the assignments of the 1 H and 19 F spectra, as well as showing that the terminations are intimately mixed, rather than being segregated into different regions, as also found for Ti3C2Tx. 33In contrast, the 1 H→ 19 F spectrum (Figure 9a) only shows correlation between the 1 H H2O (6.0 ppm) signal and the −80 ppm -F termination 19 F signal.The lack of signal from the -OH terminations is ascribed to a short 1 H T1 for this resonance, so that the transverse magnetization decays before developing significant cross polarization.The fact that the H2O resonance correlates only with the −80 ppm 19 F signal suggests that these -F termination environments are associated with interlayer water via hydrogen bonding, whereas those at −190 ppm are not.The higher frequency 19  Table 3 shows a comparison of the shifts of the observed surface species for the different MXenes studied by 1 H and 19 F NMR to date.For the -OH terminations, V2CTx is the exception with an extremely large shift of 85 ppm; this is due to a metallic Knight shift, as confirmed by the short T1 relaxation of this signal, 32 Interestingly, the -F terminations of V2CTx do not appear to be significantly Knight shifted and are observed at a similar frequency to those of Ti3C2Tx.The -OH 1 H shifts obtained for Ti3C2Tx have been shown by both Hope et al. 33 and Anayee et al. 34 to depend on the synthesis procedure, but in both cases fall in the range 12.5-20 ppm.In contrast, Kobayashi et al. 35 observed no 1 H signals at frequencies above 7 ppm, instead ascribing signals at 0.5-2.0 and 3.6 ppm to -OH terminations on the basis of DFT calculations and the assignment of a Ti2CTx 1 H spectrum in a separate report by Sugahara 69 .
These differences may be due to sample preparation, however, we note that the studies of Kobayashi et al. and Sugahara et al. do not provide direct evidence for their terminal -OH assignments (c.f. the 1 H- 13 C HETCOR employed for V2CTx and Ti3CTx, 32,33 and the 1 H{ 93 Nb} TRAPDOR used here).Furthermore, (i) Sugahara et al. found an estimated composition of Ti2C(OH)0.3O0.7F0.6Cl0.4 with a similar etching procedure to Kobayashi et al. (LiF, HCl) 69 and (ii) Ti3C2Tx is electronically conducting (so much so that the sample in the Kobayashi et al. study was diluted by a factor of 20 to perform magic angle spinning) but the DFT calculations for -OH shifts include neither -F termination nor Knight shift contributions, which are expected to substantially affect the 1 H shifts. 35 It is possible that due to the high conductivity of their samples, Kobayashi et al. and Sugahara et al. are not able to observe the -OH terminations due to rf penetration or bulk magnetic susceptibility broadening effects.
The -OH 1 H shifts for the Nb MXenes are similar to those observed for Ti3C2Tx by Hope et al. 33 and Anayee et al. 34 , although as the 26 ppm signal for Nb2CTx is outside the diamagnetic range for 1 H NMR, there must also be at least a small Knight shift, as for V2CTx.
The -F terminations, on the other hand, are observed at higher frequencies for the Nb MXenes than for V2CTx and Ti3C2Tx, which could be due to more covalent Nb-F bonding caused by the more diffuse 4d orbitals.In general, the multiple NMR signals observed for -OH and -F terminations, and the dependence on the synthesis, can be ascribed to distinct local environments due to different arrangements of the surface terminations (-OH, -O, -F and termination vacancy), as well as interlayer H2O, as shown, for instance, for the -F terminations in Nb4C3Tx.4.
and outer Ti environments in Ti3AlC2 could be assigned from the overlapped spectrum (Figure 10) by considering that the CQ values for the outer/inner metal sites are expected to differ by an order-of-magnitude (c.f. 93Nb in Nb4AlC3), and by using the fixed relationships between the nuclear properties of 47/49 Ti.The quadrupolar parameters of the two titanium sites for both nuclei are given in Table 4: since the frequency separation and ratio of the quadrupolar coupling constants for the 47 Ti and 49 Ti signals are fixed by their respective nuclear properties, there are fewer independent parameters than it might seem, and the simultaneous agreement of both signals with the measured spectrum affords a high confidence in the NMR parameters.Ti quadrupolar parameters for the two crystallographic Ti sites in Ti3AlC2 extracted from Figure 10.The estimated errors in the last digit are given in parentheses.Note that the simulated spectra were insensitive to the CSA parameters, which were therefore not determined.a The 47 Ti shifts are referenced to 49 Ti and as such are 266 ppm less than the corresponding 49 Ti shift, rather than equal.b This represents an upper bound, based on the linewidth of the 47 Ti signal.c The inherent linewidth is too broad to determine ηQ from the lineshape of the central transition, and the turning points of the satellite transitions cannot be identified, however ηQ = 0 is consistent with the crystallographic symmetry.

Conclusions
An extensive survey of the NMR properties of the Nb MAX and MXene phases has been presented.The multinuclear approach adopted in this study provides complementary measurements of the local structural and electronic properties of the MAX and MXene phases, identifies the nature and connectivity of MXene surface terminating species, and enables unambiguous detection of a number of impurity phases.Following the etching process, the conversion to form MXenes was confirmed by 13 C NMR, although the presence of unreacted Nb2AlC is apparent in the 93 Nb, 27 Al, and 13 C spectra of the Nb2CT sample. 93Nb environments become notably more asymmetric in the MXenes, demonstrating the reduction of symmetry at the surface.The quadrupolar coupling is still lower than calculated for a bare surface, however, pointing to the presence of surface terminations;

Figure 1 -
Figure 1 -Crystal structures of the niobium MAX and surface-terminated MXene compounds.In (a) Nb2AlC and (b) Nb4AlC3, the crystallographic sites are labeled and the unit cells are outlined.Table 1 describes the local coordination around each atom.The corresponding MXenes (c) Nb2CT and (d) Nb4C3T, showing -F, -O and -OH terminations; two layers and interlayer water are shown for Nb2CT.
Ti signal of SrTiO3 at −843 ppm.The isotropic shift iso is defined in the Haeberlen convention   =   +  +  3 with the chemical shift anisotropy CSA defined as   =   −   and the shift asymmetry CSA defined as   =   −    −  .With these definitions, the principal components of the shift tensor are ordered such that |ZZ-iso|  |XX-iso|  |YY-iso|.N.b.This definition of CSA is sometimes referred to as the reduced anisotropy, which is equal to 2/3 of the 'full' anisotropy ∆ =   −   +  2 used by some authors and programs.The quadrupolar coupling constant, CQ, is defined by the nuclear quadrupole moment Q (QNb-93 = 32(2) fm 2 ; QAl-27 = 14.66(10) fm 2 ) 40 and the largest principal component VZZ of the EFG at the nucleus according to   =   ℎ where e is the electric charge and h is Planck's constant.The quadrupolar asymmetry parameter Q is also defined by the EFG tensor components as   =   −    ordered such that |VZZ|  |VYY|  |VXX|.

Figure 2 -
Figure2-27 Al NMR spectra of Nb2AlC at 50 kHz MAS (top) and static (bottom).MAS rate and magnetic field are given for each spectrum.The recycle delays were 100 s and 0.35 s for the MAS and static spectra, respectively.The static spectra were T1-filtered to facilitate efficient recording of the fast-relaxing MAX signal whereas the MAS spectrum was recorded under quantitative conditions to observe all aluminum environments.Spinning sidebands are denoted with asterisks.See Table2for fitting parameters.

Figure 3 -
Figure3-27 Al NMR spectra of Nb4AlC3 at 33 kHz MAS (top, middle) and static (bottom).Note the different x-axis scales.MAS rate and magnetic field are given for each spectrum.The recycle delays were 90 s and 0.05 s for the MAS and static spectra, respectively.Spinning sidebands are denoted with asterisks.See Table2for fitting parameters.The missing intensity in the fit of the static spectrum is ascribed to the AlO6 environment at ~16 ppm.

Figure 4 -δ
Figure 4 -93 Nb NMR spectra of Nb2AlC.MAS rate and magnetic field are given for each spectrum.See Table2for fitting parameters.

c
The fits are only sensitive to the  angle because  is nearly zero.dThe isotropic shift is temperature-dependent and thus varies with MAS rate.The static 93 Nb spectrum of Nb2AlC was fit with -475(10) ppm while the 50 kHz spectrum was fit with -445(5) ppm.n/d = not determined, either due to insufficient signal-to-noise ( 13 C) or because the fit was insensitive to this parameter.Cells corresponding to the quadrupolar properties are left blank for13 C because it is a I = ½ nucleus.Quadrupolar 93 Nb NMR spectra of the compounds were recorded under MAS and static conditions.A fit of the 93 Nb NMR data of Nb2AlC reproduced the central and satellite transitions of the spectrum (Figure 4, Table 2), with calculated and measured nuclear quadrupolar coupling constants in close agreement.The Hahn echo and quadrupolar Carr-Purcell-Meiboom-Gill (QCPMG) 93 Nb MAS spectra of Nb4AlC3 (Figure 5) showed a number of overlapping peaks that could not be resolved by variable MAS rates due to their individual linewidths (full-width at halfmaximum (FWHM) > 40 kHz) and the static spectrum also showed overlapping lineshapes (FigureS4).To overcome this challenge, a MATPASS experiment was performed; the isotropic slice showed three broad, featureless resonances at -840, -1700, and -2340 ppm (Figure5).However, only two distinct Nb environments are expected in Nb4AlC3.The low frequency93  Nb resonance matches that of NbC (Figure5a,b, FigureS5a), as do the13 C NMR data (vide infra).Stoichiometric NbC should give a sharp93  Nb signal with CQ = 0 due to the cubic site symmetry of the 12-coordinate niobium atom, however nonstoichiometry is common

Figure 5 -
Figure 5 -93 Nb MAS NMR spectra of the Nb4AlC3 sample (45 kHz MAS, 16.4 T).Several methods were used to study this sample including (a) the isotropic slice of a pjMATPASS experiment; (b) a Hahn echo-deconvoluted in grey with the internal site Nb1 in green, the external site Nb2 in orange, and NbC in red; and (c) a QCPMG measurement.
C NMR spectrum of Nb2AlC (Figure 6a, bottom) shows a single resonance at 293 ppm corresponding to the single carbon environment in the 211 phase.The 13 C NMR spectrum of Nb4AlC3 (Figure 6b, bottom) shows two resonances at 324 ppm and 238 ppm corresponding to the outer (C2) and inner (C1) carbon sites of Nb4AlC3, respectively.For both Nb MAX phases, the large 13 C shift is evidence of metallic conduction resulting in a Knight shift; the shifts are larger than observed for

Figure 6 -
Figure 6 -13 C NMR spectra of (a) Nb2CTx MXene and Nb2AlC MAX phase and (b) Nb4C3Tx MXene and Nb4AlC3 MAX phase at 25 kHz MAS and 16.4 T. Resonances are labeled, dashed lines are guides to the eye.

15- 70 Figure 7 -
Figure 7 -Static 93 Nb NMR central transition spectra of the Nb2CTx and Nb4C3Tx samples at 16.4 T. For Nb2CTx, the deconvolution into signals from Nb2CTx and Nb2AlC is shown.For full VOCS spectra, see Figure S7.
Nb4C3Tx and Nb2CTx.1 H and19 F NMR spectra were recorded in order to investigate the surface functionalization of the MXene phases directly.The 1 H NMR spectrum of Nb2CTx as synthesized shows a large signal at 6.3 ppm due to free H2O as well as two resonances at large chemical shifts of 19.5 and 26.0 ppm (Figure8a); these latter signals are assigned to -OH surface terminations.-OH terminations were observed in the MXenes

Figure 8 - 1 H
Figure8-1 H and19 F NMR spectra of Nb2CTx (a-f) and Nb4C3Tx (g-l).The 1 H/ 93 Nb and19 F/ 93 Nb TRAPDOR difference spectra (c,f,i,l) show 1 H (c,i) or19 F (f,l) atoms near93  Nb atoms.The echo and MATPASS experiments were performed at 50 kHz MAS and the TRAPDOR experiments at 30 kHz, all at a magnetic field strength of 11.7 T. Sidebands are marked with asterisks.
F shift observed for H-bonded -F terminations in Nb4C3Tx is also consistent with the assignment of the −20 ppm 19 F signal in Nb2CTx to strongly H-bonded -F terminations.

Figure 9 :
Figure 9: 1 H/ 19 F HETCOR spectra of Nb4C3Tx recorded at 11.7 T and 20 kHz MAS with a 2 ms contact time.The recycle delays were 3 s and 1 s respectively for a) and b).The 19 F MATPASS spectrum is shown for comparison.Note that the axes for b) have been swapped relative to convention to allow comparison of the 19 F axis.

aFigure 10 -
Figure 10 -47/49 Ti static QCPMG spectrum of Ti3AlC2 acquired at 16.4 T by taking the skyline projection of five VOCS sub-spectra.The simulated quadrupolar 47 Ti and 49 Ti patterns for the inner and outer Ti sites are shown below the experimental spectrum.Simulation parameters are given in Table4.

27
Al and 93  Nb NMR spectra of the MAX phases Nb2AlC and Nb4AlC3 have been recorded under static and MAS conditions with high resolution and wideline techniques, allowing extraction of the full anisotropic shift and quadrupolar tensors.The insights gained from the Nb MAX phases also then enabled the interpretation of the47/49  Ti spectrum of Ti3AlC2.The quadrupolar parameters extracted from the spectra were in close agreement with DFT calculations, affording a detailed view of the local coordination environments.The27 Al,   93  Nb and13 C spectra also reveal impurity phases in both Nb2AlC and Nb4AlC3, including aluminium oxide, hydrated aluminium fluoride, aluminium carbide/oxycarbide/nitride, niobium metal, and (possibly nonstoichiometric) NbCx species.Many of these impurities are not apparent from diffraction alone and these results should serve as a guide for future synthesis efforts.
these surface terminations are key in determining the performance of MXenes in various applications.-OH and -F terminations have been identified for both Nb2CT and Nb4C3T in the 1 H and19 F NMR spectra, their proximity to the surface demonstrated with93  Nb TRAPDOR experiments, and a comparison made with previous studies on Ti3C2T and V2CT MXenes.The atomic-scale intermixing of -OH and -F terminations for Nb4C3T is seen in the 1 H/ 19 F HETCOR spectra, which would be challenging to identify by other techniques, and the different19 F termination environments observed for both MXenes are correlated with the degree of hydrogen bonding to interlayer H2O.number DMR-1740795.This work made use of the IMSERC X-ray and NMR facility at Northwestern University, which has received support from the National Science Foundation (NSF DMR-0521267), the Soft and Hybrid Nanotechnology Experimental (SHyNE) Resource (NSF ECCS-1542205), the State of Illinois, and the International Institute for Nanotechnology (IIN).Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation.The authors are grateful to M. Alhabeb and Dr. D. Pinto (Drexel University) for their help with the synthesis of samples for this study.

Table 1 -
Local atomic sites in Nb MAX and MXene phases.

Table 2 -
Experimental (shift and quadrupolar) and calculated (quadrupolar only) NMR parameters of Nb2AlC and Nb4AlC3.aSeeTableS1for chemical shift tensor component values in the standard convention.
a Static and magic angle spinning spectra (in some cases at multiple fields) were fit to a single model.Estimated errors in the last digit are given in parentheses, derived from lineshape fitting.bThesign of the quadrupolar coupling can be calculated but is not determined experimentally from the spectrum of a polycrystalline sample.

Table 3 -
19mmary of the chemical shifts of 1 H and19F surface species observed for different MXenes.