Polar surfaces, which have a permanent dipole moment perpendicular to the surface, are of great importance in both physical and chemical applications1,2,3,4. Due to the very large energies of uncompensated surfaces, polarity compensation is required, generating different and complex surface configurations for these facets. Therefore, it is extremely difficult to understand the atomic-scale structure of polar surfaces, which is essential in order to design related nanomaterials for a targeted property5,6,7,8. By applying electron microscopy and computational modelling, a variety of polarity compensation mechanisms have been proposed, including ordered surface reconstructions9, surface disorder10, adsorption of environmental gas molecules11, surface metal layers deposition12, and subsurface oxygen vacancies13. Despite the many advantages of microscopy techniques, they are limited to the visualization of a small fraction of the sample which may not yield reliable quantitative information about the whole sample, and are typically performed at high vacuum conditions that may alter the surface environment14. Furthermore, light elements, such as hydrogen and oxygen which are of key importance for many materials, are difficult to probe with such techniques15. Although significant developments have been made in environmental electron microscopy, which allows materials to be investigated under adjustable pressure conditions and in variable gaseous environments16,17,18, the other disadvantages remain and complementary methods are required.

Solid-state NMR spectroscopy is a powerful method which can provide rich local structural information for solids19,20,21,22,23,24,25,26, complementary to the information obtainable from diffraction27 and microscopy techniques. Recently, 17O solid-state NMR spectroscopy has been developed as a new approach for determining the surface structure of oxide nanomaterials, with help from surface-selective labeling and DFT calculations. Oxygen ions in different layers of ceria nanostructures28 and at different facets of anatase titania nanocrystals29 can be distinguished according to the NMR shifts. However, only non-polar facets were studied previously and no attempt was made to investigate the more challenging oxide nanostructures with polar facets. Furthermore, quantification of different surface species was not possible using exclusively 17O NMR, due to the quadrupolar nature of 17O and the potentially non-uniform isotopic labeling procedure.

Ceria nanocubes expose (100) facets, which show exceptional properties as both the catalytically active plane and the support facet; this is a relatively simple polar surface, making ceria nanocubes an ideal model30,31,32,33. Using the example of ceria nanocubes, we introduce a strategy of qualitative 17O and quantitative 1H solid-state NMR spectroscopy combined with DFT calculations to characterize oxide nanocrystals with polar facets. We thereby quantitatively determine detailed polar surface structural information, specifically the presence and concentration of reconstructed Ce terminated structures (CeO4-t) and hydroxyl groups.


Morphology of the ceria nanocubes

Ceria nanocubes were hydrothermally synthesized with Ce(NO3)3·6H2O and NaOH (see methods). The X-ray diffraction (XRD) data (Supplementary Fig. 1) confirms the formation of ceria with a fluorite structure (JCPDS No. 34-0394). High-resolution transmission electron microscopy (HRTEM) images show that the samples adopt a cubic morphology with sizes of 18 to 40 nm, dominated by (100) polar surfaces before and after 17O enrichment (Supplementary Fig. 2). Inductively coupled plasma mass spectrometry (ICP-MS), elemental analysis and X-ray photoelectron spectroscopy (XPS) data show that there are no detectable Na+ or NO3 impurities (Supplementary Fig. 3 and Supplementary Table 1). In addition, the concentration of oxygen vacancies at the surface is only 1.0% according to analysis of Raman spectroscopy data and, therefore, their influences are not considered further (Supplementary Fig. 4 and Supplementary Note 1).

NMR spectra of the ceria nanocrystals

The 17O solid-state NMR spectra of ceria nanocubes enriched with 17O2 at 523 K (NCs-17O2) and H217O at 373 K (NCs-H217O) are shown in Fig. 1. The 17O NMR spectrum of NCs-17O2 is dominated by the peak at 877 ppm (see the untruncated spectrum in Supplementary Fig. 5) due to the OCe4 environment in the bulk of the ceria nanocubes, but other signals can also be observed at 970, 935, and 825 ppm. 17O NMR signals for ceria samples with (111) facets have previously been observed at 1040, 920, and 825 ppm due to oxygen ions in the first, second, and third (sub-)surface layers respectively28; the shoulder resonance at 825 ppm in the NCs-17O2 spectrum is therefore most likely due to a deeper sub-surface layer while the signals at 970 and 935 ppm, which have not previously been observed, may be tentatively assigned to the oxygen ions at the (100) surface.

Fig. 1
figure 1

17O Solid-State NMR spectra of ceria nanocubes. The 17O NMR measurements were performed at a spinning speed of 20 kHz for NCs-17O2 (a) and 16 kHz for NCs-H217O (b). A rotor synchronized Hahn-echo pulse sequence (π/6-τ-π/3-τ-acquisition) with 1H decoupling and a 0.5 s recycle delay was used. Stars denote spinning sidebands. In order to show the peaks due to OH better, a line broadening of 600 Hz is applied for the enlarged spectrum shown in light blue region.

For the spectrum of NCs-H217O, in addition to the bulk signal with a maximum at 877 ppm, two broad peaks centered at approximately 250 and 1012 ppm can be observed. The former is most likely to be related to surface hydroxyl groups (Ce-OH)29,34, and on closer inspection can also just be distinguished for NCs-17O2. The latter can again, based on its high frequency compared to the bulk resonance, be attributed to under-coordinated surface oxygen species. Surface-selective isotopic labeling is achieved by exposing the samples to 17O2 gas or to H217O vapor at relatively low temperatures, although there are differences between the spectra that will be discussed later.

A possible explanation for the higher frequency signal in the spectrum of NCs-H217O is the formation of some degree of the thermodynamic (111) surface, given that the first surface layer in this case has been observed at 1040 ppm. However, the HRTEM images show little evidence for (111) facets (Supplementary Fig. 2), and after re-enriching NCs-H217O with 17O2 gas, the surface signals are identical to those of NCs-17O2 and do not exhibit signals associated with (111) surfaces (Supplementary Fig. 6). This suggests that the high frequency signal is related to the (100) surface itself and the H217O enrichment.

To investigate the presence of hydroxyl groups or molecularly adsorbed H2O on the ceria surface, quantitative 1H NMR was performed. 1H chemical shifts are sensitive to hydrogen bonding35,36 and the lineshapes can be affected by the distribution of hydrogen bond distances; in particular, stronger hydrogen bonding results in higher frequency 1H NMR signals37. The spectrum of NCs-17O2 in Fig. 2a shows a broad resonance between 2–16 ppm (the sum of the green deconvoluted Lorentzian functions), along with a shoulder at 5.4 ppm (blue signal), and a relatively sharp peak at 2.5 ppm (brown signal); a summary of these deconvoluted signals is presented in Supplementary Table 2. The signal at 5.4 ppm is most likely due to molecularly adsorbed water molecules and the 2.5 ppm signal can be assigned to terminal hydroxyl species (-OHT)38; the remaining intensity is then assigned to bridging hydroxyl groups (-OHB) with a distribution of environments and degrees of hydrogen bonding. A similar 1H NMR spectrum was observed for NCs-H217O (Fig. 2b), but with a broader signal centered around 8 ppm, corresponding to bridging hydroxyl groups with an even greater distribution of environments and hydrogen bonding. Quantitative analysis of the integrated 1H intensities, combined with the specific surface areas as measured from the BET isotherms, yields adsorbed water contents for NCs-17O2 and NCs-H217O of 2.6 and 3.7 H2O molecules per surface unit (59.3 Å2), respectively. Of these, 0.2 H2O molecules are molecularly adsorbed for both NCs-17O2 and NCs-H217O, while the rest are dissociatively adsorbed to form hydroxyl groups (see Supplementary Fig. 7 and Supplementary Table 3). The quantitative 1H NMR spectroscopy indicates that there are similar concentrations of hydroxyl groups on both NCs-17O2 and NCs-H217O, however the intensity of signals from hydroxyl groups in the 17O NMR spectrum of the former is much lower. This is ascribed to unenrichment of the oxygen atoms of the hydroxyl groups for NCs-17O2 by natural abundance water, which is not completely dehydrated at the relatively low temperature of 523 K. In contrast, the surface water of NCs-H217O is replaced by H217O, so the same unenrichment does not occur.

Fig. 2
figure 2

1H Solid-State NMR spectra and 1H → 17O HETCOR NMR of ceria nanocubes. The 1H NMR measurements for NCs-17O2 (a) and NCs-H217O (b) were performed at a spinning speed of 12 kHz using a rotor synchronized Hahn-echo pulse sequence (π/2-τ-π-τ-acquisition) with a recycle delay of 2.0 s. Deconvolution of 1H spectra and the summarized 1H chemical shifts (obtained from DFT calculations, see below) are also shown. Background signals from the cap, rotor and probe etc. are removed by subtracting the spectra of an empty rotor obtained at the same condition from the initial spectrum. c The 1H → 17O HETCOR spectrum of NCs-H217O, recorded at a MAS rate of 14 kHz with a recycle delay of 2 s and a contact time of 0.5 ms.

The 1H → 17O HETCOR NMR spectrum of NCs-H217O (Fig. 2c) shows correlations between the 17O signals at around 250 ppm with the 1H signals due to hydroxyl groups, confirming the assignment of the 17O NMR spectra. Furthermore, hydroxyl sites with higher 17O shifts are associated with higher 1H shifts, and thus a stronger acidity (although stronger hydrogen bonding can also result in larger shifts)39. The conclusion that bridging hydroxyl groups are more acidic than terminal hydroxyls is in agreement with previous reports in zeolites39. Since the 17O NMR shift range is much wider than for 1H NMR shifts, 17O NMR spectroscopy may provide an alternative and more sensitive probe of the acidity and acid-catalysis reactivity for oxide nanomaterials. However, 17O NMR experiments often require spectra acquired at multiple magnetic fields, preferably higher fields, or with high resolution techniques (e.g., MQMAS40), in order to decrease the linewidths arising from quadrupolar interactions.

Spectral Assignments from DFT Calculations

DFT calculations have previously proved successful in aiding spectral assignment for surface oxygen ions in oxide nanostructures28,29. The differences between the calculated and experimental results are generally around or less than 10 ppm28, allowing reliable spectral assignment. An oxygen terminated (O-t) model of ceria (100) surface was previously investigated for DFT calculations28 (Fig. 3a), however, the calculated 17O NMR parameters (Supplementary Figs. 810 and Supplementary Table 4) are not in agreement with our NMR observations, i.e., no surface species in the calculations are associated with resonant frequencies at 970, 935 or ~1012 ppm and the calculated signal for the 1st layer two-coordinate oxygen ion (O2C), with a high frequency chemical shift (1117 ppm), is not observed in the experimental spectrum.

Fig. 3
figure 3

Structure of O-t and CeO4-t surface. Surface oxygen ions are represented as red balls and cerium ions as off-white, in the top and side views of ceria (100) facets with an O-t surface (a, b) and a CeO4-t surface (c, d). The top views (a, c) each show four surface units (each surface unit is 7.7 Å × 7.7 Å = 59.3 Å2), delimited by white boxes. Each O-t surface unit includes four O2C ions, while each CeO4-t surface unit contains two O2C ions and one CeO4 reconstruction (blue dashed circle), with four three-coordinated oxygen ions (O3C).

As seen from the 1H and 17O NMR spectra, a significant number of hydroxyl groups are present at the surface, which must be considered. A previous computational study reported that dissociative adsorption of water is much more favorable than the molecular adsorption of water on the (100) O-t surface41; accordingly, DFT calculations were performed on O-t models with 1, 2, 3 and 4 dissociatively adsorbed H2O molecules per O-t surface unit (Supplementary Figs. 1123 and Supplementary Tables 58). On reaction with water, the under-coordinated O2C sites with high chemical shifts are protonated to form hydroxyl groups with calculated shifts of 191–360 ppm, as observed in the experimental 17O NMR spectra; four dissociated H2O molecules per O-t surface unit are required to protonate all the O2C sites (monolayer hydroxylation, see Supplementary Fig. 20). However, there are no calculated resonances at 970, 935 or ~1012 ppm; this suggests that the structure of polar (100) facets in ceria nanocubes is more complicated than the simple O-t model.

Two recent studies suggested that a fraction of ceria (100) facets may form CeO4 terminated (CeO4-t) reconstructions, which yield a lower surface energy than cerium terminated (Ce-t) or O-t surfaces10,42. Therefore, DFT calculations were performed on a model comprising CeO4-t reconstructions linked by O2C sites (CeO4-t surface, Fig. 3b)—a pure CeO4-t reconstruction has previously been shown to have a high surface energy and is thus unstable10.

First, the relative energies of dissociative and molecular adsorption were calculated for a single H2O molecule per CeO4-t surface unit (Supplementary Figs. 2426). A comparison of the adsorption energies shows that H2O molecules also prefer to adsorb dissociatively on clean CeO4-t ceria (100) surfaces (Supplementary Table 9). The models M0, M1, M2 and M3 then correspond to 0, 1, 2 and 3 dissociatively adsorbed H2O molecules per CeO4-t surface unit. The most energetically favorable H2O adsorption sites in M1, M2, and M3 were then determined by calculating and comparing the adsorption energies of several possible structural models (the lowest energy structure for each model is shown in Supplementary Figs. 2628). The corresponding 17O NMR calculation results for M0, M1, M2, and M3 are shown in Figs. 4 and 5, Supplementary Figs. 2938 and Supplementary Tables 1013 (also see more discussion in Supplementary Note 2 and Supplementary Fig. 39).

Fig. 4
figure 4

17O NMR spectrum, calculated 17O NMR shifts and the structure model of ceria NCs-17O2. a 17O Solid-State NMR spectra of ceria NCs-17O2 (shown in Fig. 1). b The summarized 17O NMR shifts (δCGs) predicted for the CeO4-t model with two H2O molecules dissociatively adsorbed on each surface unit (M2). c The hydrated CeO4-t model used in the DFT calculations with NMR shifts (δCGs) for oxygen ions and the layer number shown on the right. Red, off-white, and white balls represent bulk oxygen, cerium, and hydrogen ions, respectively. Surface oxygen ions with different shifts are shown in different colors. The calculated 17O NMR parameters for each oxygen ion are shown in Supplementary Table 12. A line broadening of 600 was applied to the enlarged spectrum in the blue region, no line broadening was applied to the enlarged spectrum in the red region.

Fig. 5
figure 5

17O NMR spectrum, calculated 17O NMR shifts and the structure model of ceria NCs-H217O. a 17O Solid-State NMR spectra of ceria NCs-H217O (shown in Fig. 1). b The summarized 17O NMR shifts (δCGs) predicted for the CeO4-t model with three H2O molecules dissociatively adsorbed on each surface unit (M3). c The hydrated CeO4-t model used in the DFT calculations with calculated NMR shifts (δCGs) for oxygen ions and the layer number shown on the right. Red, off-white, and white balls represent bulk oxygen, cerium, and hydrogen ions, respectively. Surface oxygen ions with different shifts are shown in different colors. The calculated 17O NMR parameters for each oxygen ion are shown in Supplementary Table 13.

The O2C ions in the first layer of models M0 and M1 are associated with very high chemical shifts of 1162 and 1168 ppm respectively; such high frequency signals are not present in the experimental 17O NMR spectra. This again indicates that these high energy, under-coordinated species have reacted with water to form more stable hydroxyl species, as observed in the 1H and 17O NMR spectra.

The calculated 17O NMR shifts (center of gravity, δCG) based on the CeO4-t model with two or three H2O molecules dissociatively adsorbed on each surface unit (M2 and M3, Figs. 4 and 5, Supplementary Figs. 3538, Supplementary Tables 12 and 13) are in good agreement with the experimental spectra. All the O2C ions on the surface are consumed by reacting with H2O molecules to form hydroxyl groups, therefore there is no calculated signal at a shift higher than 1100 ppm, as observed experimentally. For the O3C surface species in the CeO4 reconstruction, the calculated shifts in M2 are 943 and 968 ppm, which match the observed resonances at 935 and 970 ppm in the experimental spectrum of NCs-17O2, while the calculated O3C shifts in M3 are 910 and 996 ppm, in accordance with the experimental signals at 907 and ~1012 ppm for NCs-H217O. The resonance at ~1012 ppm is quite broad, which is presumably due to a distribution of local environments caused by different arrangements of nearby dissociative H2O molecules. The observation that the resonant frequencies are higher for lower-coordinated oxygen ions on the surface is similar to the previous studies on ceria nanoparticles exposing (111) facets and titania nanostructures29,34.

The DFT calculations predict that the O4C ions in the 2nd or 3rd layer for both NCs-17O2 and NCs-H217O have resonant frequencies closer to the bulk shift of 877 ppm, which is consistent with the relatively broad component observed experimentally for the peak centered at 877 ppm. The predicted peak in the 2nd layer at 756 ppm for M2 is not experimentally observed; this environment, although fully coordinated, is highly distorted (see Supplementary Fig. 40). This distortion may be lost on addition of a bridging hydroxyl between the CeO4 reconstructions (e.g. H4, see Fig. 6), which is above the distorted environment, or the distortion may be averaged at non-zero temperatures due to rapid interconversion of different local environments with similar energies.

Fig. 6
figure 6

Structure of M2 and M3. White and off-white spheres represent hydrogen and cerium ions, respectively, in the top view of M2 (a) and M3 (b). Surface oxygen groups with different chemical shifts are in different colors. For each CeO4-t surface unit of M2, four hydrogen ions labeled as H1, H2, H3, and H4 form from the two dissociative H2O molecules while for each CeO4-t surface unit of M3, six hydrogen ions labeled as H1, H2, H3, H4, H5, and H6 form from the three dissociative H2O molecules.

Terminal hydroxyl sites (-OHT) are calculated to have 17O shifts (δCGs) of 69 ppm for M2 and 71 ppm for M3, while bridging hydroxyl sites (–OHB) for both samples have calculated 17O shifts (δCGs) of 206–291 ppm. These sites are predicted to have relatively large CQs (5.5–7.6 MHz), thus the corresponding resonances are expected to exhibit significant second order quadrupolar broadening. This agrees well with the broad resonance observed at lower shifts in the experimental spectrum of NCs-H217O; the hydroxyl groups have been largely unenriched for NCs-17O2, see above.

The above results indicate that the concentration of surface hydroxyl groups has a great impact on the NMR shifts of surface oxygen ions. Furthermore, in order to reproduce the experimental 17O NMR spectra, both dissociated water and CeO4-t reconstructions must be included in the calculations. Thus, the ceria (100) surface can be regarded as a combination of CeO4-t and O-t surface units, where the under-coordinated O2C ions have been converted to bridging hydroxyl groups and terminal hydroxyl groups have formed on some under-coordinated cerium ions.

The 1H NMR chemical shifts were also calculated using DFT, confirming the assignment of the 1H NMR signals at 2–16 ppm and 2.5 ppm to bridging (–OHB) and terminal (–OHT) hydroxyl groups on the surface (Fig. 6 and Table 1)38,39. The calculated 1H shift for -OHT is the most negative, and since the lowest frequency signal in the experimental 1H NMR spectrum is the relatively sharp peak at 2.5 ppm, this resonance is assigned to -OHT. The broad signal is attributed to –OHB. Different –OHB environments are associated with a range of chemical shifts and the distribution is wider for the model with three dissociatively adsorbed H2O molecules (M3) than that for two (M2); in particular, the very large shift of H4 in M2 is due to hydrogen bonding to the oxygen ions of H2 and H3. This at least partially explains why the experimental -OHB resonance is broad and why the spectral line width for NCs-H217O is broader than for NCs-17O2, given the higher hydroxyl content of the former. An inhomogeneous distribution of dissociated water and variable hydrogen bonding may also contribute to the broadness of the signals.

Table 1 Calculated 1H NMR chemical shiftsa for M2 and M3. The corresponding structures are presented in Fig. 6.

Based on the quantitative 1H NMR data and the above assignments from the DFT calculations, the fractions of CeO4-t and O-t surface units comprising the (100) facets of ceria nanocubes can be determined (Table 2). This is based on the fact that each CeO4-t surface unit contains one characteristic terminal hydroxyl group (–OHT) and either three (M2, Fig. 2a) or five (M3, Fig. 2b) bridging hydroxyl groups (–OHB), while each hydroxylated O-t surface unit contains eight bridging hydroxyls (-OHB) (Supplementary Fig. 21). The fractions of CeO4-t and O-t surface units calculated for NCs-17O2 (57.1% for O-t and 42.9% for CeO4-t) are very close to those for NCs-H217O (57.5% for O-t and 42.5% for CeO4-t), supporting the assignment of these models.

Table 2 The contents of terminal hydroxyls (-OHT and -OHB) and different surface units.

17O DNP NMR spectroscopy

Recent developments in dynamic nuclear polarization (DNP) provide new opportunities to characterize the surface structure of solid materials34,43,44. Direct DNP involves transferring polarization from unpaired electrons directly to the nucleus of interest, with the unpaired electrons typically being added in the form of organic biradicals; because the biradicals are external to the particles, and the hyperpolarization mechanism has a 1/r6 distance dependence, the surface can be selectively hyperpolarized and hence observed in the NMR spectrum (surface enhanced NMR spectroscopy, SENS). Indirect DNP, where 1H nuclei are first hyperpolarized before cross-polarization to the nucleus of interest, can yield greater enhancements in some cases, but for 17O NMR only oxygen atoms directly bonded to H can be observed, so indirect DNP cannot be used to detect the unhydroxylated (sub-)surface oxygen environments34. Therefore, direct DNP SENS was applied to study ceria nanocubes and aid spectral assignments. The intensities for the bulk 17O peak at 875 ppm are 5.2 times stronger with microwave irradiation than without, for both NCs-17O2 and NCs-H217O (Fig. 7a, b). For NCs-17O2, the enhancement factor for the peak at 970 ppm is ~8, while the surface signals for NCs-H217O can only be observed in the “on” spectrum, indicating that hyperpolarization is more efficient for these species than for the bulk, and therefore that these resonances are indeed surface oxygen species.

Fig. 7
figure 7

17O DNP NMR spectra and saturation recovery data of NCs-17O2 and NCs-H217O. Deconvoluted 17O direct DNP NMR spectra, with and without microwave irradiation of a NCs-17O2 and b NCs-H217O mixed with the TEKPol radical in dried TCE. The spectra were measured at 100 K at a spinning speed of 13.9 kHz. A pre-saturated rotor synchronized Hahn-echo pulse sequence (π/2-τ-π-τ-acquisition) and a recycle delay of 60 s were utilized. Stars denote spinning sidebands. c, d The corresponding 17O direct DNP saturation recovery build-up curves of the deconvoluted isotropic resonances in a and b. The build-up time (TDNP) was determined by deconvoluting the spectrum and fitting the peak areas to a stretched exponential function of the form \(\frac{{I(t)}}{{I_0}} = 1 - e^{ - \left( {\frac{t}{{T_{DNP}}}} \right)^\beta }\), where I(t) and I0 are the signal intensities at delay t and at equilibrium, respectively, and β is the stretching exponent (0 < β < 1) (Table 3).

The DNP build-up time, TDNP, can also be fitted to distinguish external and internal 17O nuclei34, since nuclei close to the surface hyperpolarize faster and hence have a shorter TDNP. For NCs-17O2, the TDNP values for the peaks at 970, 880, and 870 ppm are much smaller than for the bulk signal at 875 ppm, implying that the former arise from surface species in NCs-17O2 (Fig. 7c and Table 3). Broader peaks are observed for the 17O DNP SENS spectra than for the room temperature 17O spectra, which can be attributed to the freezing out of motional averaging of the dynamic surface sites at the low temperatures required for DNP, as previously observed for ceria (111) facets34.

Table 3 The build-up time (TDNP) and stretching exponent (β) of different peaks extracted from the 17O direct DNP saturation recovery build-up curves shown in Fig. 7.

The 17O DNP spectrum of NCs-H217O has spinning sidebands which overlap with other resonances due to the lower spinning speeds achievable at 100 K and the higher field at which the DNP experiments were performed (which results in a smaller separation of sidebands in ppm, for the same spinning frequency). Therefore, 17O DNP projection magic angle turning and phase adjusted sideband separation (MATPASS) NMR experiments were performed to resolve the isotropic resonances (Supplementary Fig. 43)45,46. Four non-bulk resonances at 1012, 970, 895, and 853 ppm can be deconvoluted, and were used to fit the saturation recovery data and obtain the TDNPs (Fig. 7d and Table 3). Again, these peaks are associated with shorter TDNPs than the peak at 875 ppm arising from the bulk part of the sample, in agreement with our assignments from conventional 17O NMR spectroscopy.


17O and 1H solid-state NMR spectroscopy, combined with DFT calculations, were employed to determine the surface structure of ceria (100) polar surfaces. The results obtained in this work provide compelling evidence that CeO4-t reconstructions and hydroxyl groups are present on the surface of ceria nanocubes, both of which are expected to reduce the surface energy and afford polarity compensation. The amount of hydroxyl groups alters the 17O NMR shifts of oxygen ions at the surface of the nanocubes, making 17O a very sensitive probe for the surface structure and, therefore, its properties. Furthermore, the fractions of CeO4-t and O-t surface units can be determined with quantitative 1H NMR measurements. DNP SENS results confirm the 17O spectral assignments, however, although DNP SENS spectroscopy generally provides a stronger signal-to-noise ratio compared to conventional NMR spectroscopy, due to the restricted spinning rates and broader surface signals at low temperature, certain species can be resolved better with the latter. The strategy introduced here can be applied to gain insight into the surface structures of oxide nanocrystals and materials with polar surfaces.


Preparation of ceria nanocubes

In a typical synthesis procedure47, 1.96 g Ce(NO3)3·6H2O was added into 40 mL distilled water. After stirring for 5 min, 30 mL NaOH solution (pH = 14) was slowly added into the mixture before it was vigorously stirred for another 30 min at room temperature. The mixture was then transferred into a 100 mL Teflon-lined hydrothermal reactor and heated at 453 K for 24 h before it was allowed to cool to room temperature. The resulting white sediment was centrifuged, washed with distilled water and dried at 353 K overnight. Finally, the solid was annealed in a tube furnace at 573 K for 5 h in flowing oxygen gas to obtain calcined ceria nanocubes.


Powder X-Ray Diffraction (XRD) characterization was performed with a Philips X’Pro X-ray diffractometer with Cu Kα irradiation (λ = 1.54184 Å) operating at 40 kV and 40 mA. High-Resolution Transmission Electron Microscope (HRTEM) images were recorded on a JEOL JEM-2010 instrument at an acceleration voltage of 200 kV. X-ray Photoelectron Spectra (XPS) were measured on a Thermo ESCALAB 250 X with Al Kα ( = 1486.6 eV) as the excitation source. The binding energies in XPS spectra were referenced to C 1 s = 284.8 eV. Brunauer–Emmett–Teller (BET) specific surface area information was obtained from nitrogen adsorption at 77 K on a Micromeritics ASAP 2020 system. Raman spectra were acquired with a Bruker Multi RAM FT-Raman spectrometer using 514 nm light from a He–Ne laser source. The content of Na ions was analyzed by an Optima 5300DV inductively coupled plasma mass spectrometer (ICP-MS) while the N content was determined with a Heraeus CHN-0-Rapid elemental analyzer.

17O Isotopic labeling procedure

Ceria nanocubes were 17O-labeled on a vacuum line with commercial 90% 17O-enriched O2 gas and H2O, respectively (Isotec Inc.). In a typical 17O isotopic labeling procedure, 200 mg sample was placed in a glass tube and calcined at 523 K under 1 × 10−3 Torr for 3 h, in order to remove most of the physically adsorbed water and surface hydroxyl groups. The nanocubes were allowed to cool down to room temperature before 17O2 gas or H217O vapor was introduced into the glass tube. The glass tube was sealed and then heated to the target temperature (523 K for 17O2 and 373 K for H217O labeling) for 10 h. After the enrichment with H217O, the samples were exposed to vacuum to remove physisorbed water.

Solid-State NMR Spectroscopy

17O and 1H Magic Angle Spinning Nuclear Magnetic Resonance (MAS NMR) experiments were recorded on a Bruker Avance III 400 spectrometer equipped with an 89 mm wide-bore 9.4 T superconducting magnet yielding Larmor Frequencies of 54.2 and 400 MHz, respectively. All of the samples were packed into 3.2 mm rotors inside a N2-filled glove box. 17O and 1H chemical shifts are referenced to H2O at 0.0 ppm and to adamantane at 1.92 ppm, respectively.

17O direct dynamic nuclear polarization (DNP) NMR experiments were performed at a Larmor frequency of 81.3 MHz on a 14.1 T Bruker Avance III HD 600 spectrometer equipped with a 395 GHz gyrotron microwave source and a 3.2 mm MAS probe. The microwave source power applied for 17O direct DNP measurements was 7.0 W. 17O labeled ceria nanocubes were mixed with radical solution (16 mM TEKPol44 in dried tetrachloroethane, TCE) in an Ar-filled glove box. 17O chemical shifts of the DNP NMR spectra were referenced to bulk ceria at 875 ppm at 100 K.

Details of DFT calculations

All spin-polarized DFT calculations were carried out using the Vienna Ab initio Simulation Package (VASP)48. The Perdew-Burke-Ernzerhof (PBE) functional49 with the Hubbard U correction (DFT + U)50 were used for all calculations. The effective U value of 5.0 eV was only applied to the localized Ce 4f orbitals51,52; our previous study shows that the calculated chemical shifts from PBE + U (5.0 eV) are in quantitative agreement with the experimental values28. The projector augmented wave method53 was used to describe the interaction between core and valence electrons. A plane-wave kinetic energy cutoff of 500 eV was used for all calculations. For geometry optimization, all of the atoms were allowed to relax until the Hellman–Feynman forces were lower than 0.02 eV Å−1. For electronic minimization, we used an energy convergence criterion of 10−5 eV for optimizing geometries and a higher criterion of 10−8 eV for chemical shift and electric field gradients (EFGs) calculations28. The optimized lattice parameter of ceria using PBE + U (5.0 eV) is 5.448 Å, which is in reasonable agreement with the experimental value (5.411 Å)54.

We used a 2 × 2 surface cell to model the ceria (100) surface. The ceria (100) surface slab model with 12 oxygen layers was found to be sufficiently thick, i.e., the middle layers of this model mimic the bulk environment in terms of chemical shift (Supplementary Figs. 9, 12, 15, 18, 22, 30, 33, 36 and 38). All the slabs contain a large vacuum gap (>10 Å) to remove the slab-slab interactions. The k-point mesh was sampled by using a 2 × 2 × 1 Monkhorst–Pack grid.

We used the same method as our previous work28 to calculate chemical shifts, quadrupole coupling constants (CQ) and asymmetry parameters (η). For the electric EFG calculations to obtain CQ and η of oxygen species, we used the experimental quadrupole moment (Q) of −0.02558 barns55 for 17O. For calculating the isotropic chemical shift (δiso), we used the following equation:

$$\delta _{{\mathrm{iso}}} = \delta _{{\mathrm{cal}}} + \delta _{{\mathrm{ref}}},$$

where δcal is the unaligned DFT chemical shift, δref is the reference chemical shift. The averaged value of the unaligned DFT chemical shifts of oxygen species in the middle layers (layers 4–9) of every prototype slab models is 835 ppm. By aligning 835 ppm to the corresponding experimental value of 877 ppm, we obtained the δref of 42 ppm.

The average adsorption energies of each water molecule (Εads) on the (100) surface with the O-t or CeO4-t model were calculated as the following:

$$E_{{\mathrm{ads}}} = 1/n\cdot \{ E[n{\mathrm{H}}_2{\mathrm{O}}/{\mathrm{CeO}}_2] - E[{\mathrm{CeO}}_2] - nE[{\mathrm{H}}_2{\mathrm{O}}]\} ,$$

where n is the number of adsorbed water molecules, E[nH2O/CeO2], E[CeO2] and E[H2O] are the DFT calculated total energies of the adsorption complex, the ceria substrate and the gas phase H2O molecule, respectively.

Thermodynamic stabilities of different water adsorption structures on O-t and CeO4-t surface units at given water partial pressure and temperature were determined by calculating the surface free energy per unit area (γ(p,T))41,56,57:

$$\gamma \left( {p,T} \right) = 1/A\!\cdot \!\left\{ {\!G_{{\mathrm{slab}}}\left[ {n{\mathrm{H}}_2{\mathrm{O}}/{\mathrm{CeO}}_2} \right]\left( {p,T} \right)-mG_{{\mathrm{bulk}}}\left[ {{\mathrm{CeO}}_2} \right]\left( {p,T} \right)- n\mu \left[ {{\mathrm{H}}_2{\mathrm{O}}} \right]\left( {p,T} \right)} \!\right\},$$

where A is the surface area of the slab, n is the number of adsorbed water molecules, m is the number of CeO2 bulk (i.e., Ce4O8) units in the slab model, μ is the chemical potential, and G is the Gibbs free energy.

We assumed that the surfaces are in thermodynamic equilibrium with gas phase H2O. So, μ[H2O] (p,T) can be calculated as follows:

$$\it \begin{array}{l}\mu \left[ {{\mathrm{H}}_2{\mathrm{O}}} \right]\left( {p,T} \right) = \!\!\!\! E[{\mathrm{H}}_2{\mathrm{O}}] + \Delta \mu \left[ {{\mathrm{H}}_2{\mathrm{O}}} \right]\left( {p,T} \right) = E[{\mathrm{H}}_2{\mathrm{O}}] + H\left[ {{\mathrm{H}}_2{\mathrm{O}}} \right]\left( {p^0,T} \right)\\ \!\!\!\!\!\!\!\!\!\!-\; H\left[ {{\mathrm{H}}_2{\mathrm{O}}} \right](p^0,0K) - TS\left[ {{\mathrm{H}}_2{\mathrm{O}}} \right]\left( {p^0,T} \right) + K_BTln\left( {p/p^0} \right),\end{array}$$

where p0 is the standard state pressure (0.1 MPa); enthalpy (H) and entropy(S) terms were taken from the website of NIST58. As the DFT total energies of the solid components can be regarded as good approximations of corresponding Gibbs free energies57, we then obtained:

$$\it \begin{array}{l}\gamma \left( {p,T} \right) = \!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!1/A\cdot \left\{ {G_{{\mathrm{slab}}}\left[ {{\mathrm{nH}}_2{\mathrm{O}}/{\mathrm{CeO}}_2} \right]\left( {p,T} \right)-mG_{bulk}\left[ {{\mathrm{CeO}}_2} \right]\left( {p,T} \right)-nE[{\mathrm{H}}_2{\mathrm{O}}]} \right.\\ \!\!\!\!\!\!\left. {-\;nH\left[ {{\mathrm{H}}_2{\mathrm{O}}} \right]\left( {p^0,T} \right) + nH\left[ {{\mathrm{H}}_2{\mathrm{O}}} \right]\left( {p^0,0K} \right) + nTS\left[ {{\mathrm{H}}_2{\mathrm{O}}} \right]\left( {p^0,T} \right) - nK_BTln\left( {p/p^0} \right)} \right\},\end{array}$$

Note that the vibration contributions and the pV (V denotes volume) term of solid components were not considered.