Photo-Induced Active Lewis Acid–Base Pairs in a Metal–Organic Framework for H2 Activation

The establishment of active sites as the frustrated Lewis pair (FLP) has recently attracted much attention ranging from homogeneous to heterogeneous systems in the field of catalysis. Their unquenched reactivity of Lewis acid and base pairs in close proximity that are unable to form stable adducts has been shown to activate small molecules such as dihydrogen heterolytically. Herein, we show that grafted Ru metal–organic framework-based catalysts prepared via N-containing linkers are rather catalytically inactive for H2 activation despite the application of elevated temperatures. However, upon light illumination, charge polarization of the anchored Ru bipyridine complex can form a transient Lewis acid–base pair, Ru+–N– via metal-to-ligand charge transfer, as confirmed by time-dependent density functional theory (TDDFT) calculations to carry out effective H2–D2 exchange. FTIR and 2-D NMR endorse the formation of such reactive intermediate(s) upon light irradiation.


■ INTRODUCTION
The activation of H−H bond is challenging in various catalytic process for the production of fuels. 1 Traditional oxidative addition steps over an electron−rich transition metal center can lead to H−H cleavage into two M−H bonds.Recently, frustrated Lewis pairs (FLPs) involving synergetic sites can be created from stereoisolated Lewis acid and base sites in close proximity or derived from a weak intramolecular adduct instead of using single transition metal center to provide an alternative pathway for the heterolytic H−H bond cleavage by the charged, polarized bodies.−6 However, the majority of such FLP catalysts have the Lewis basic and acidic sites separately introduced, which prohibits precise control of the location and quantity of the active sites.Further advancement in synthesis methods would be required for the improved catalytic application of FLP for small-molecule activation by enhanced control of the active sites.
Molecular activation via photochemical means is generally achieved under comparatively milder conditions.Photocatalytic activation of the H−H bond would therefore be beneficial toward energy efficiency.Our approach for the design of such photoactive transient but polarized Lewis acid− base sites in metal−organic framework (MOF) materials are inspired by the characteristic superior surface area and porosity of the MOFs and their anchored transition metal(s) with designated linkers of tunable properties to capture light energy with facilitated polarization, which could create the required active Lewis acid and base pair for synergetic activation of substrate. 7In our previous work, we have utilized various MOFs as supports for the design of efficient catalysts thanks to their vast chemical tunability. 8MOFs are extending frameworks formed by the coordination of metal clusters with polyfunctional organic ligands, specifically the Universitet-i-Oslo (UiO) series. 9−22 Thanks to such tunability of the UiO series, the ruthenium bipyridine motives can be integrated into the MOF systems.Ruthenium has been deposited onto MOF UiO-67-bpydc with 2,2′-bipyridine-5,5′-dicarboxylic acid analogous to ruthenium bipyridine complexes 23,24 (denoted Ru/bpy).For comparison, ruthenium−amine complexes were also synthesized with rutheniumdecorated UiO-66-NH 2 (denoted Ru/NH 2 ), where 2-aminoterephthaclic acid was used as the organic linker.
In this work, we demonstrate the high activity for H 2 −D 2 exchange of our photoinduced FLP-like catalyst with ruthenium deposited UiO-67-bpydc (Ru/bpy) under illumination.Our synthesis method gives precise control of the FLPlike active site location and quantity in the solid catalyst.Synchrotron X-ray absorption and X-ray diffraction measurements were initially performed to determine the structures of the anchored ruthenium complexes on the frameworks with precision before theoretical calculations.The optical properties of different Ru-MOFs were also characterized with UV−vis, PL, and TRPL.Excited-state calculations based on timedependent density functional theory (TDDFT) were particularly performed to rationalize the charge transfer processes and the formation of photoinduced active sites in excited states.IR and 2D NMR indicated the synergetic formation of hydride and proton from dihydrogen via the formation of a transient Lewis acid−base pair upon illumination.

■ RESULTS AND DISCUSSION
Ruthenium-deposited UiO-67-bpydc (bipyridine linker) and UiO-66-NH 2 (benzylamine linker) were synthesized as prototypes for photoinduced FLP.Inductively coupled plasma-mass spectrometry (ICP-MS) results on both Ru/bpy and Ru/NH 2 showed similar Ru/Zr ratios of 0.352 and 0.344 (Table S1).Advanced techniques were then utilized to characterize the structures of their Ru loaded samples, namely, Ru/bpy (UiO-67-bpydc with Ru) and Ru/NH 2 (UiO-66-NH 2 with Ru).First, extended X-ray absorption fine structure (EXAFS) was performed on the two ruthenium-deposited MOFs to probe their local coordination environment.The scattering paths generated from bulk RuN, RuO 2 , and RuCl 3 were then used to fit the R-space EXAFS data (Figure 1a,b), with the k-space data and fit available in Figure S1 and fitting parameters available in Table S2.The goodness of fit in the EXAFS data is shown by the low R-factor of 0.59 and 0.04% for Ru/bpy and Ru/NH 2 , respectively.In both Ru/NH 2 and Ru/ bpy, it is unambiguous that there is no aggregation of the ruthenium species in the system by comparing the data to the scattering paths generated from the bulk references RuN, RuO 2 , RuCl 3 , and Ru foil (Figure S2).In total, an average of ca.6 coordination 6.4 (7) for Ru/bpy and 5.8 (5) for Ru/NH 2 can be fitted.The degeneracy obtained from fitting for Ru/bpy shows a value of 2.0(2) Ru−N from the bipyridine linker, 2.2(3) Ru−O from adsorbed H 2 O and 2.2(2) Ru−Cl.For Ru/ NH 2 , 0.9(1) Ru−N from linker, 1.9(1) Ru−O, and 3.0(3) Ru−Cl.To further determine their structures, synchrotron Xray diffraction (SXRD) was performed on all the samples and is shown in Figure S3.Notice that the positions of the Bragg peaks remain unchanged (space group: Fm3̅ m) for both samples, implying that the crystalline framework of the host MOF remains mostly unaltered by the Ru incorporation.Further Rietveld refinement of the model was built based on the bond lengths from EXAFS fitting resulted to generate an SXRD pattern and a satisfactory fit (R wp values of 15.46 and 10.58 for Ru/bpy and Ru/NH 2 , respectively) with acceptable parameters was obtained through the Rietveld method. 25The refined structures are shown in Figure 1c,d, with the Rietveld refined parameters and fit in Table S3.

■ STRUCTURE−ACTIVITY RELATIONSHIP
The hydrogen activation of Ru/bpy and Ru/NH 2 was evaluated by hydrogen−deuterium exchange.Reactions with and without visible light filtered Xe lamp irradiation were performed at 25 to 100 °C, and the results are presented in Figure 2a.The conversion factor of the signal of the quadruple mass spectrometer was determined by a calibration curve of pure HD gas (Figure S4).There is no noticeable exchange rate with both the Ru/bpy and Ru/NH 2 samples at room temperature.At 100 °C, small thermal activation of about 10 μmol HD per hour is formed and can be detected over Ru/bpy due to a thermal partition of the electrons for metal-to-ligand charge transfer to a small degree.Conversely, Ru/NH 2 still does not seem to show any measurable activity in the dark.However, under illumination, there is a dramatic light promotion to the exchange rate with at least ca.5.5-fold increase in the amount of HD formed in Ru/bpy, while Ru/ NH 2 still remains catalytically inert.The turnover frequency, with the Ru content based on ICP-MS results, is plotted against inverse of temperature to obtain an activation energy of 0.368 eV under light illumination (Figure S5), attributed to the diffusion of H 2 /D 2 molecules into the pores of Ru/bpy.This is much lower than the literature value of 0.672 eV for thermal chemical exchange between H 2 and D 2 , and this reduction is associated with the Ru−N catalyzing the reaction in an FLPlike manner. 26It is apparent that the two Ru-MOF samples do not seem to be active for H 2 activation via dissociative means over their metal center at room temperature.In fact, the more extensive conjugated aromatic π* in the bipyridine in Ru/bpy should in principle be able to withdraw electron from Ru by the back-donation than the electron richer Ru from the sigma Ru−N bond in Ru/NH 2 , hence attenuating the propensity for classical H 2 activation.However, this catalyst is clearly shown to be more active under light at elevated temperature than the Ru/NH 2 .The key question is why H 2 activation can be significantly promoted by light in the Ru/bpy but it does not apply to the related Ru/NH 2 .
To understand the mechanistic pathway of the H 2 −D 2 exchange reaction, Fourier transformed infrared (FTIR) spectroscopy at variable temperature was performed on the reaction intermediate (Figures 2b and S6).A simultaneous detection of Ru−H and N−H stretching at 2045 cm −1 of Ru− H and at 2973 cm −1 of N−H in the quenched sample at 100 °C supporting the unconventional heterolytic cleavage of H− H upon illumination is identified. 27,28Despite the difficulty in quantification by infrared, the Ru−H appears to be far smaller in size than that of N−H.It is well-known that charge transfer and proton migration can take place simultaneously from transition metal hydride with a proton acceptor in close proximity.For example, [FeFe]-hydrogenases can catalyze H 2 oxidation to protons exclusively via initiative FLP Fe−N sites and with the further conversion of H − on Fe by charge transfer to NH + .Similar chemistry for Ru−H charge transfer and proton migration to O(H + ) on MgO(111) is demonstrated. 29hus, Ru 3+ (2N) can generate Ru + and 2NH + with a substantial reduction/elimination of the Ru−H peak.Notice that one-dimensional (1D) 1 H solid-state nuclear magnetic resonance (SSNMR) spectra for sample before and after illumination are shown in Figure S7.Deconvolution of the spectrum of Ru/bpy in dark shows a 1:0.87:0.91 ratio of the peaks at 9.18, 8.30, and 7.14 ppm, respectively, of the three H environments on the linker, as depicted in Figure 2c and Table S4. 30,31The resonances at around 1.20 and 0.18 ppm are attributed to the bridging Zr(μ 3 -OH) groups 32 and the linker defects motifs, 33 respectively.Interestingly, a peak at 4.1 ppm, characteristic of trapped H 2 in the porous sample, is detected before light activation.For well-flushed and cold-trapped sample after illumination, its NMR spectrum is also deconvoluted and shown in Figure S8 and Table S5.Although no similar trace peak of Ru−H as IR is identified, presumably embedded in strong background, a distinctive peak at 2.4 ppm attributable to proton on amine can clearly be seen upon the illumination. 34The relative proton peak size of bipyridine (7− 9 ppm) to NH + (2.4 ppm) peak of 7.19:1 suggests the near conversion, with the stoichiometric quantity of Ru−N pairs relative to bipyridinic protons determined by ICP and thermogravimetric analysis (TGA, Figure S9 and Tables S6  and S7) as 6.90:1.To correlate their spatial relationship, a twodimensional (2D) 1 H− 1 H magic angle spinning (MAS) SSNMR experiment was performed (Figure 2c).Besides, the strong cross-diagonal correlation peaks which belong to the equivalent proton sites, both peaks at 4.

■ OPTICAL CHARACTERIZATION
To confirm the importance of MLCT over Ru/bpy over Ru/ NH 2 to create the active Lewis acid−base pair, the various charge transfer processes of different energies of the ruthenium-decorated MOFs were investigated by UV−vis spectroscopy (Figure 3a,b).Observed in both UiO-67-bpydc and Ru/bpy, the absorption peaks at 249 and 283 nm can be attributed the typical π to π* transition of a bipyridinic and the isolated pyridinic rings present in these MOF samples, respectively. 35For the absorption characteristics at visible regime, UiO-67-bpydc without Ru, gives a characteristic peak at 563 nm which can be ascribed to linker−linker charge transfer (LLCT), as it is also detectable in its free linker molecules (Figure S10).−38 Such strong adsorption peaks cover almost the entire visible region, giving the intense dark color of the sample.On the other hand, in the case of benzyl amine linker (UiO-66-NH 2 ) with the linker of lesser conjugation, the π to π* absorption peaks are found to be located at 219 and 257 nm, respectively. 35,39ditional peaks at 335 and 371 nm can be generally ascribed to the charge transfer processes from amine lone pair to organic linker π* transitions, 40 which remain unaltered upon Ru incorporation.With rather isolated conjugated benzyl ring from the sigma M−NH 2 there is no similar broad intense LLCT or MLCT absorption regions at the visible region in both samples in contrast to those observed in bipyridine systems.Only a small peak around 653 nm of Ru/NH 2 is detected, which may be attributed to the weak Ru d−d transitions. 41,42he photophysical processes in the samples were also characterized by photoluminescence (PL) at 375 nm excitation to understand the dynamic photoexcitation processes.It is interesting to note that the UiO-66-NH 2 with and without Ru only shows a main broad PL peak but with no vibrational or progression feature, whereas UiO-67-bpydc gives a clear shoulder peak at about 550 nm (triplet state from LLCT) matching with its visible absorption regime.The PL of Ru/bpy and UiO-67-bpydc can be deconvoluted into two Gaussian peaks to quantify the peak shoulder (Figure 3c), with the fitting parameters available in Table S8.Similar to the UV−vis spectrum, there is a further broadening of the PL shoulder in Ru/bpy to capture the most visible regime, presumably to the availability of the additional lower energy triplet intermediate state due to MLCT in the metal−bipyridinic system. 43ime-resolved photoluminescence (TRPL) was also employed at the energy of PL peak to examine the kinetics of such charge transfer processes (Figure S11), and the data are fitted against a biexponential decay function (Table S9).For the extensive conjugation in bipyridine without Ru, the dominant charge transfer process is shown to be the higher-energy LLCT, as stated.The introduction of Ru offers another lower energy pathway that would quench the LLCT transition.Notice that the average lifetime of Ru/bpy was calculated to be 0.110 ns, which is significantly shorter compared to 0.450 ns of UiO-67-bpydc. 44Such sharp attenuation in exciton lifetime suggests that an relaxation pathway is via the MLCT process, which is an alternative to the radiative recombination of LLCT.Alternatively, for the isolated benzyl amine systems, it is acceptable that the high energetic process involves transfer of lone pair electrons in N to the linker molecule (n → π*, see TDDFT calculations below), the Ru could thus facilitate the polarization by further coupling with its orbitals, hence prolonging the average lifetime of UiO-66-NH 2 and from 0.668 to 1.696 ns in the case of Ru/NH 2 .Thus, Ru is thus shown to suppress the rate of charge recombination for increased lifetime of charge carriers in this case. 45To further understand and evaluate the charge carrier dynamics, transient absorption spectroscopy (TAS) was performed on the Ru/bpy sample (Figure S12).Intriguingly, a positive peak at 423 nm can be observed, which is close to the PL peak determined at 426 nm.This is indicative of excited state absorption (ESA) on top of the radiative process probed by PL, echoing our proposed model for Ru/bpy.Both the broadening of the PL peaks and the ESA suggest there are various available triplet energy intermediate states due to MLCT in the Ru/bpy system, whose lifetime peaks at 2 ps.

■ EXCITED-STATE CALCULATIONS
From the optical characterization, the MLCT processes of the Ru/bpy catalyst have been seen by UV−vis and PL spectroscopy, while the charge transfer kinetics have indicated a decreased lifetime of Ru/bpy compared with its parent UiO-67-bpydc.To confirm the above charge transfer processes and the nature of polarization in the Ru-MOFs, excited-state calculations were performed based on TDDFT.The Ru/bpy and Ru/NH 2 structures were constructed by anchoring the corresponding Ru species on the linkers of the UiO-67-bpydc and UiO-66-NH 2 frameworks, respectively (Figure S13).To reduce the computational cost, the UiO-67-bpydc and UiO-66-NH 2 frameworks were presented by their truncated clusters (Figure S14), which has been proved to be effective in previous studies. 46,47The calculated most stable spin states of the cluster models of Ru/bpy and Ru/NH 2 are closed-shell singlet and doublet, respectively (Tables S10 and S11).The electronic transitions in these two clusters (with the most stable spin state) were further explored with the aid of TDDFT.The first 50 excited states were calculated for each cluster model, and those with an oscillator strength (f) larger than 0.01 are presented in Figure 4 and Table S12.The oscillator strength quantifies the probability of absorption or emission of electromagnetic radiation in transitions between energy levels.In the cluster of Ru/bpy, the S 0 → S 3 (f = 0.030), S 0 → S 20 (f = 0.019), S 0 → S 22 (f = 0.052), and S 0 → S 23 (f = 0.129) transitions are relatively strong.Interestingly, the S 0 → S 3 transition of MLCT in origin identified at 732 nm corresponds well to the experimental 690 nm peak in UV−vis.Similarly, the calculated S 0 → S 20 transition (415 nm) is correlated to the MLCT peak at 487 nm of UV−vis, while the energetically close S 0 → S 22 and S 0 → S 23 transitions (398 and 371 nm, respectively) are correlated to the absorption region at 339 nm of the UV−vis.Such transitions indeed feature the excited electron and hole being localized on the linker and Ru, respectively, confirming the generation of the photoinduced but transient Ru−N Lewis acid−base pair as the dominant transition under illumination.−51 In the framework of TDDFT, the mean absolute error of electronic transition energies calculated using the decent B3LYP functional 52 (which is also the functional used in the present study) is approximately 0.5 eV, as reported in previous benchmark studies. 53,54Therefore, in state-of-the-art TDDFT, the obtained transition energies agree well with the experimental results.
On the other hand, the absence of low energy available π overlap between Ru−N in Ru/NH 2 , with isolated benzyl amine and poor mixing with Ru orbitals, is evident; there are only two relatively strong electronic transitions, i.e., D 0 → D 29 (f = 0.053) and D 0 → D 32 (f = 0.023).The D 0 → D 29 transition at 413 nm is dominated by the electron from lone pair n in the N → aromatic π* configuration, with minimal involvement with Ru metal, which agrees well with the experimentally observed absorption peak at 371 nm in the spectrum of the Ru/NH 2 .The peak may also have the contribution from the D 0 → D 32 transition, which is generally localized within the Ru complex, reinforcing the absence of charge transfer between Ru and the linker (no polarization).

■ CONCLUSIONS
In conclusion, a photoinduced frustrated Lewis pair like active site is, for the first time, reported in the Ru/bpy catalyst as a proof of concept.Its design combines the high chemical tunability of the MOF and the strong metal-to-ligand charge transfer ability of ruthenium complexes.The successful anchoring of Ru onto MOFs is shown with EXAFS and Rietveld refinement on SXRD data, while the optical processes are characterized by UV−vis, PL, and TRPL.Upon light irradiation, the Ru−N Lewis acid−base pair in Ru/bpy is polarized as indicated by the TDDFT calculations, beneficial to the photo catalysis of the hydrogen−deuterium gas exchange reaction.We believe our approach for catalyst design can become transferable to the wider application of light-induced FLP-like chemistries in the activation of small molecules.

■ EXPERIMENTAL SECTION
Full details of the experiments can be found in the Supporting Information.
Sample Preparation.In a 20 mL scintillation vial, 0.150 mmol ZrCl 4 and 0.138 mmol 2-aminoterephthalic acid or 0.138 mmol 2,2bipyridine-4,4′-dicarboxylic acid for UiO-66-NH 2 and UiO-67-bpy, respectively, and 0.7 mL of ethanoic acid were added into 10 mL of DMF.The reaction mixture was sonicated for 10 min and then heated in an oven at 120 °C for 24 h.The formed UiO-66-NH 2 or UiO-67bpydc powder was collected by centrifugation at 5000 rpm for 10 min, washed by DMF three times and subsequently soaked in ethanol for 24 h three times.The products were dried overnight in a vacuum oven at 80 °C.For the synthesis of Ru/bpy and Ru/NH 2 , 200 mg of dried MOF was put into activation at 120 °C overnight and then added into 20 mL of ethanol.107.6 mg of RuCl 3 was then added to the UiO-67-bpydc (123.9 mg of RuCl 3 for UiO-66-NH 2 ) solution under stirring at room temperature for 24 h.The product was collected by centrifugation at 5000 rpm for 10 min, washed with ethanol 3 times.The products were dried overnight in a vacuum oven at 80 °C.
FT-IR Measurements.FTIR experiments were performed in the range of 650−4000 cm −1 on a Nicolet iS 50 FT-IR spectrometer using an MCT/A detector at a resolution of 4 cm −1 .Approximately 50 mg of Ru/bpy samples was placed in the reaction cell in the glovebox.
SS-NMR Measurements. 1 H MAS MAS NMR experiments were conducted on a Bruker Avance 400 spectrometer operating at 9.05 T with a 4 mm double-resonance MAS probe, with Larmor frequencies of 495.43 MHz for 1 H. Spinning speeds of 12 kHz, 32 scans, and 3.9 μs π/2 excitation pulse were used for 1 H single-pulse acquisitions.The standard two-dimensional (2D) three-pulse exchange (a.k.a.NOESY) sequence was used to monitor the chemical exchange between proton sites in the catalyst, with a dwell time in the indirect dimension set to 20 μs.Typically, 32 scans were acquired for each t1 increment, with final data sets consisting of 512 t1 × 2048 t2.Mixing times is 10 ms.The recycle delays used in the 2D exchange experiments is 5 s with MAS speeds equal to 12 kHz. 1 H chemical shift is referenced to adamantane at 1.78 ppm.
Theoretical Calculations.−57 The PBEsol 58 exchange−correlation functional with the D3 dispersion correction of Grimme 59 was applied for geometry optimizations.The core−valence electron interactions were treated by using the projector augmented wave (PAW) 60 method.A kinetic energy cutoff of 500 eV was used for all periodic DFT calculations.A 1 × 1 × 1 k-point mesh was used for the sampling of the first Brillouin zone.Both atomic positions and the shape of the cell were allowed to relax during optimizations, for which we used a Hellman−Feynman force criterion of 0.05 eV/Å for each ion.
The cluster models of the Ru/NH 2 and Ru/bpy structures were constructed by capping the [RuCl 3 (H 2 O) 2 (BDC-NH 2 ) 3 ] 6− and [RuCl 2 (H 2 O) 2 (BPYDC) 3 ] 6− anions (from the optimized periodic models shown in Figure S11) with protons, respectively (Figure S12).The ground-state structures were obtained by a two-step optimization procedure.In the first step, only the proton capping ions were optimized.In the second step, the COOH groups of the carboxylates were fixed to retain periodic constraints, while the other atoms were allowed to fully relax.Time-dependent density functional theory (TDDFT) with the adiabatic linear-response approximation 61 was used for excited-state calculations.The cluster calculations were performed using the Gaussian 16 program. 62The B3LYP 52 exchange−correlation functional and the def2-TZVP 63−65 basis set were used for the calculations; for Ru, effective core potential (ECP) was used.−69 The electron−hole distributions were calculated by using Multiwfn 70,71 and visualized by using VESTA. 72ASSOCIATED CONTENT * sı Supporting Information The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/jacs.3c05244.
Additional experimental details, materials, and methods (PDF) ■

Figure 1 .
Figure 1.Fourier-transformed magnitude of the experimental Ru K-edge k 3 -weighted R-space EXAFS data and fit of (a) Ru/bpy and (b) Ru/NH 2 , where |χ(R)| and Re[χ(R)] denote the magnitude and real part of the Fourier transformed k-space data.Crystal structures built from the EXAFS fitting parameters of (c) Ru/bpy and (d) Ru/NH 2 .Color scheme: Zr = light blue polyhedra, O = red, C = black, N = blue, Ru = pink, and Cl = light green.Hydrogen and adsorbed water O atoms have been omitted for clarity.

Figure 2 .
Figure 2. (a) Amount of HD formed with both catalysts, with and without light.(b) FTIR spectra of Ru/bpy in the dark and illuminated in reaction conditions.(c) Deconvoluted 1 D 1 H SSNMR spectra of Ru/bpy in the dark.(d) 1 H− 1 H COSY NMR spectrum of Ru/bpy in the dark (blue) and illuminated (brown).(e) Proposed mechanism of heterolytic activation of the H−H bond by transient frustrated Lewis acid−base like pair.
1 and 2.4 ppm show off-diagonal correlation peak with the bipyridinic protons, indicating their spatial close proximity.Evidently, transient N(H + ) can be identified in H 2 with no reduction to the bipyridine linker during illumination (proton remained in the same positions) after activation and charge transfer characterized by IR and ssNMR.It is thought that light activation is likely to create transient intramolecular Ru + −N − via metal-toligand charge transfer (MLCT) for H 2 activation.We therefore propose a photoinduced Ru−N polarization mechanism as depicted in Figure 2e.H 2 and D 2 molecules can be exchanged by light activation.

Figure 4 .
Figure 4. Electron−hole distributions (isosurface: 0.001 e/Å 3 ) of the electronic transitions with oscillator strength being larger than 0.01 in the cluster models of Ru/bpy and Ru/NH 2 structures.S and D denote singlets and doublet, respectively.Structure color scheme: C = black, N = blue, O = red, H = white, Ru = gray, and Cl = green.Isosurface color scheme: electron = yellow, hole = blue.