Black ZnO nanoparticles synthesized by a green chemistry process

In the present work, the standardization of the methodology to obtain black ZnO nanoparticles from Arabica coffee extract as a reducer agent and stabilizer of the reaction is presented for the first time through a scalable combustion green chemistry process without obtaining dangerous byproducts. The size distribution of the nanoparticles was found between 15 and 30 nm. High-resolution transmission electron microscopy shows distorted regions from the atomic column. Whereas, the estimated energy band gap measured by UV–vis spectroscopy is 2.22 eV, which is 30% value below the typical band gap for bulk ZnO. XPS measurements show a change in the binding energy of black ZnO compared to commercial ZnO. From experimental evidence, it is proposed that the black color of zinc oxide resulted from vacancies in the ZnO structure. The vacancies in the structure were theoretically modeled considering a variation in the Coulomb interaction between Zn—O atoms by applying the Hubbard + U DFT approximation. The theoretical electronic distribution of the influence of vacancies ZnO was compared with the experimental results obtained by Raman, FTIR and the experimental profile of the valence band region. These results open the exploration of green synthesized black zinc oxide nanoparticles to possible technological applications related to catalysis.


Introduction
Nanoscience and nanotechnology are under constant research because all the possible applications with positive impact such as water remediation, wastewater treatment, nanocomposite production to enhance thermophysical properties, enhancement of properties related to use efficiently the solar energy, alternative methods for time an energy saving [1][2][3][4][5][6].
Among the several metabolites that have been documented to be contributors of reducing and stabilizing zinc salts into Zn ions to the formation of zinc oxide nanostructures; the most reliable are flavonoids, steroids, terpenoids, tannins and polyphenols [28,35,39].Coffee is known to contain all these phytochemicals in substantial proportions [40], and thus, successful green synthesis of ZnO has been reported previously by using coffee leaf extract, spent coffee grounds and coffee powder extract assisted with microwave irradiation; nevertheless, the green chemistry pathways proposed are low scalable procedures [36][37][38]41].
Black ZnO was reported earlier by Ting Xia et al [42] in 2014, obtained by hydrogenation treatment, reducing NaBH 4 and pulsed laser irradiation, where the change in color from white to black ZnO was attributed to oxygen vacancies.N. Zhang and co-workers [43], in 2016, synthesized black colored ZnO nanowires by metal-organic chemical vapor deposition system under oxygen-deficient conditions and studied in photocatalytic hydrogen evolution.H Zhang et al [44], in 2018, reported the attaining of gold nanoparticles with hybrid black ZnO nanorods synthesized by sol-gel method with plausible applications as humidity sensor.In 2018, Zuñiga-Ibarra [45] synthesized black TiO 2 through pulsed laser irradiation in liquid while S Chen [46] obtained black TiO 2 by thermal treatment of amorphous Ti(OH) x in vacuum and reported the enhancement in the photocatalytic activity degrading Rhodamine B and Methylene blue dyes.L Han [47] performs several annealing processes of 3 h long per each step in the thermal treatment from w-TiO 2 to obtain black TiO 2 and successfully tested in a photocatalytic activity degrading phenol.All these previous studies explore that black color in ZnO and TiO 2 is due to the presence of vacancies.Recently, Badreldin et al [48] reported bulk black ZnO synthesized from several thermal treatments from sealed glass container to calcination in argon atmosphere followed by a physiochemical reduction procedure while S Sharma et al [49] obtained nanostructured black ZnO by nanosecond pulsed laser irradiation in liquid.
The present work reports, for the first time, the synthesis methodology of black ZnO nanoparticles within the green chemistry scheme by using Coffee Arabica L extract, its characterization and its influence in the band gap, offering an antecedent to possible exploration for potential applications in catalysis.

Materials and methods
Ground beans of Coffea Arabica L type (CaL) were purchased from a Mexican commercial coffee market.The coffee extract was prepared by mixing 12 g of CaL with 100 ml of distilled water with a magnetic stirrer hot plate for 15 min at 90 °C, then, the extract was filtered with a commercial coffee filter.Then, 2 g of zinc nitrate (Zn(NO 3 ) 2 •6H 2 O, Sigma Aldrich, 99.0%) were dissolved in 10 ml of distilled water and stirred without heat for 10 min.The zinc nitrate solution was then combined with 30 ml of CaL extract in a flask to hold a 1:3 volumetric ratio between zinc solution and coffee extract.
The blend obtained was heated without stirring at 85 °C until the complete vaporization of water; the material produced after evaporation of water was placed in a crucible in a conventional electric furnace for a rapid combustion heat treatment, for 5 min at 600 °C, and then rapidly cooled in air to room temperature.The powder achieved, was washed with 30 ml of distilled water and centrifugated for 10 min at 4500 RPM to eliminate any remaining organic compound from the coffee extract (figure 1).
X-ray diffraction patterns were acquired in a Bragg-Brentano geometry with a Siemens D5000 diffractometer with Cu-Kα radiation and a Ni filter in an interval from 10°to 90°with a scan step of 0.02°in 2θ.The ZnO crystal structure was compared with the ICSD -PDF card # 890511 through a Rietveld refinement analysis done employing the MAUD software [50].Transmission electron microscopy in different configuration modes (TEM, HRTEM and SAED) was acquired using a JEOL TEM-2010 FEG electron microscope with 200 kV of accelerating voltage with a point resolution of 0.19 nm.The micrographs were acquired with an ORIUS CCD camera and processed with the Digital Micrograph software by GATAN.ThermoFisher GENESYS TM 150 UVvis spectrophotometer was used in a wavelength interval from 280-750 nm.Raman measurements were done with a dispersive Raman Trivista 557 spectrometer in a range from 80-3000 cm −1 .X-ray photoelectron spectroscopy (XPS) measurements were performed in an ultra-high vacuum (UHV) system XPS PHI 5000 Versa Probe II, with an Al K α x-ray source (hν = 1.4866 keV).
The XPS spectra were obtained at 45°to the normal surface in the constant pass energy mode, E 0 = 100 eV (surface survey) and 10 eV (high-resolution narrow scan).A least square process was made in deconvolution adjustment for the XPS spectra with the Spectral Data Processor software.

Computational details
Computational calculations were performed within the density functional theory (DFT) framework [51,52] as implemented in the Cambridge Serial Total Energy Package code [53].Geometry relaxation of the crystal structure and vibrational property was performed under the local density approximation (LDA) with the Ceperley-Alder and Perdew -Zunger (CA-PZ) exchange-correlation functional together with the linear response density functional perturbation theory (DFPT) [54][55][56], considering for zinc and oxygen atoms the 3d 10 4s 2 and 4s 2 2p 6 orbitals as valence electrons, respectively.The tightly bound core electrons were represented by a norm-conserving pseudopotential of the Vanderbilt type [57].Electronic properties were computed within the Hubbard corrected semi-empirical local density approximation (LDA + U) to correctly reproduce the experimental ZnO band gap.The wave functions were expanded in a plane-wave basis set with cut-off energy of 830 eV with a k-point sampling inside the first Brillouin zone constructed using the Monkhorst-Pack scheme with 5 × 5 × 4 grids [58].The equilibrium properties were obtained via the geometry optimization process in the Broyden-Fletcher-Goldfarb-Shanno (BFGS) minimization scheme.The tolerance parameters to achieve the optimization were 5 × 10 -6 eV/atom for the total energy convergence, 0.01 eV/Å for the forces between atom pairs, 5 × 10 -4 Å for the ionic displacement and 0.02 GPa for stress.The ZnO were modelled with a hexagonal symmetry, P6 3 mc space group (No. 186), with 4 atoms distributed in the 2b Wyckoff sites, where the initial lattice parameters were set to a = 3.22 Å and c = 5.20 Å.

Results and discussion
The occurrence of the ZnO phase is evidenced by the characteristic x ray diffraction pattern shown at figure 2, in which broadened peaks observed are commonly related to nanoparticle distributions, with a crystallite size    the (002) plane in ZnO is related with a strong interaction between Zn-O atoms, evidencing the failure in the peak fitted intensity by influence of vacancies.
The figures 3(a) and (b) shows the formation of ZnO NPs with regular morphology and size distribution between 15 to 20 nm.It could be assumed that concentration of the coffee arabica L extract, associated with the phytochemicals present in coffee, promotes the ZnO NPs formation and stabilization, in concordance with previous results of similar biosynthesis process when the concentration of the coffee extract is modified [59].Selected area electron diffraction pattern (SAED), figure 3(c), evidence the crystalline nature of the NPs.Moreover, the occurrence of the (002), (101), ( 102) and (103) planes are linked to ZnO phase, according to PDF card 89-0511.
The HRTEM micrograph, figures 3(d) and (e), shows an interplanar distance matching with the (101) planes in ZnO, principally composed by Zn atoms.As it is shown by the simulated ZnO supercell oriented along the (101) planes, it is observed non-uniform zones which is suggestive of the presence of vacancies in the ZnO structure, since oxygen-deficient regions modifies interatomic distances and thus deforms the atomic columns distribution seen by HRTEM, as it is displayed in the micrographs [60,61].
Figure 4(a) shows the UV-vis spectrum of the synthesized black ZnO nanoparticles with a main absorbance below 400 nm with the typical contribution for ZnO nanoparticles, at 366 nm, originated by the surface plasmon resonance due to strong quantum confinement effects [62], with a slightly 10 nm blue shift respect to the 376 nm bulk ZnO surface plasmon [63].The symmetrical shape of the plasmon band suggest a uniform scattering coming from regular symmetric nanoparticle distribution [64], and in concordance with TEM.
It is known that the optical bandgap is related to the minimum energy required to excite a photon; while the electronic bandgap is associated with the exciton binding energy; thus, for semiconductor materials with a small exciton binding energy, the optical and electronic bandgap could be considered quite similar among them.Therefore, by means of the Tauc approximation ( ) it is plausible to estimate the energy bad gap (E g ); where a is the absorption coefficient, hn is the incident photon energy and A is a constant of proportionality frequently linked with the absorbance through the Beer-Lambert law.
Through a linear fit, an E g = 2.22 eV for the black ZnO nanoparticles was experimentally estimated, inset figure 4(a).In order to theoretically describe the electronic structure of the ZnO NPs, the electronic band gap was calculated within the LDA+U approximation in order to reproduce the experimental energy band gap.Previous ZnO theoretical works have reported values for the semi-empirical Hubbard U parameter between 6 eV to 8 eV for zinc (U Znd ) and oxygen (U Op ) atoms in the ZnO compound to reproduce the electronic band gap value close to the experimental 3.3 eV, assessed for bulk ZnO [66,67].In the present work, different combinations of both semi-empirical U Znd and U Op Hubbard parameters were tested in order to correctly reproduce the experimental electronic band gap value of 2.22 eV.The best configuration was achieved when both Hubbard parameters was set as Uzn d = UO p = 5 eV, reaching an electronic band gap equal to 2.24 eV, figure 4(b), which is 1% near to the experimental E g data assessed from Tauc approximation.The values for the semi-empirical Hubbard parameters (U Znd and U Op ) below 6 eV can be associated with a redistribution of the electronic density around Zn and O atoms, from bulk ZnO to black ZnO NPs, due to the presence of oxygen vacancies, keeping in mind that the Hubbard parameter models the coulomb interaction of valence electrons [68].Therefore, the main influence on the conduction band (figures 4 and 7(c)) comes from the oxygen atoms; thus, it could be inferred that reduction from 3.3 eV to 2.24 eV in the theoretical electronic band gap is due to oxygen vacancies.
Survey XPS spectra from commercial ZnO and black ZnO nanoparticles shows similar contribution from the core level region (figure 5).The XPS high-resolution region from 288 to 283 eV binding energy shows the contribution of adventitious carbon at 284.84 eV (figure 6   respectively.It is also observed a contribution at 286.08 eV for the black ZnO NPs that could be related with organic materials capping the ZnO nanoparticles as it was reported before in green synthesis process [36][37][38]41].Figures 6(c) and (d), show deconvolution of the 2p 1/2 and 2P 3/2 orbital contributions for the commercial and black ZnO NPs with a doublet difference of 23 eV related to the spin-orbit coupling, table 2. The commercial ZnO NPs binding energy region is in good concordance with those values reported in NIST [69] 1044.7 eV (2p 1/2 ) and 1021.5 eV (2p 3/2 ).However, for the black ZnO NPs in the 2p core-level region it is observed the presence of a small shift into higher energies; compared against the commercial zinc oxide nanoparticles, and being indicative of an electronic density redistribution around the Zn atoms [70][71][72] supporting the assumption of vacancies in the black ZnO, as HRTEM results suggested and as it was previously reported in similar works for black ZnO [42,48,49].
Deconvolution of the O1s core-level XPS region is shown at figures 6(e) and (f) for commercial and black ZnO NPs, respectively.The orbital binding energy for the commercial ZnO located at 530.07 eV is due to the Zn -O density charge distribution (NIST database [69]), while the core-level energy contribution at 531.66 eV could be related with adventitious carbon interacting with ZnO [49,73].For the black ZnO NPs the contribution to the Zn -O interaction in the O 1 s core level region is located at 531.12.eV; where it is observed a shift of almost 1 eV to higher energies, being indicative of redistribution of charge density between the Zn-O atoms; and thus, correlating with the presence of vacancies.The considerable contribution at 532.59 eV figure 6(b), could be related with the capping organic material on the surface of the black ZnO nanoparticles, figure 6(f) [49].
Finally, figures 6(g) and (h), show the valence band region, letting to estimate the valence band maxima (VBM) of 2.82 eV and 3.64 eV for the commercial and black ZnO NPs, respectively.The calculated electronic structure for black ZnO was compared with three different experimental results: FTIR, Raman and XPS in the valence band region.
As can be seen from figure 7, there is a good agreement between experimental and theoretical results, elucidating that the electronic structure of the black ZnO is well described by incorporating Hubbard correction UZnd = UOp = 5 eV.Even when a general trend coincidence exists between experimental and theoretical results, there are many possible reasons for a slight disagreement between both schemes.One of the main reasons is related to the fact that DFT calculations consider that the systems under study are at 0 K, avoiding different thermal influences, for example, Debye vibrational lattice contributions or with structural defects.There are different possibilities to improve the DFT model to obtain more accurate results compared to the experimental ones, for example, a model could be constructed for the ZnO nanoparticles as a cluster with specific sites assigned as vacancies to distribute the vacancies along the cluster.However, this scheme is more expensive in computational resources, which would be another independent investigation.Since one of the main purposes of the present work was to evidence that vacancies are responsible for the narrowed band gap of 2.22 eV in black ZnO nanoparticles when it was synthesized within the green chemistry by using coffee extract as a reducing and stabilizing reagent, it was considered that the DFT + U semi-empirical model was enough.
The line broadening in FTIR spectra from experimental data, figure 7(a), could be a consequence of different chemical environments around Zn atoms owing to the presence of the oxygen vacancies.Considering that ZnO has an irreducible representation where A 1 and E 1 are both Raman and IR active, E 2 is only Raman active and where B 1 have two inactive branches.Theoretical results show that IR main contribution comes from the 437 cm −1 and 415 cm −1 wavenumber energies, related with the E 1 and A 1 IR modes, figures 8(a) and (b).There are two intense Raman peaks (figure 7(b)), the first at 100 cm −1 , with a 17% difference with the 83 cm −1 Raman band calculated; assigned to the E 2L mode and which is related to the vibrations of the Zn sublattice as it is shown at figure 8(c).The second contribution at 435 cm −1 , with a 6% difference with the 463 cm −1 Raman contribution calculated by DFT is related to the E 2H modes, figure 8(d).
Figure 7(c), shows a qualitative comparison between the calculated partial density of states (PDOS) in the valence band region and the experimental XPS valence band.From the PDOS, it could be seen that the main influence near the Fermi energy comes from the p and d states while the contribution to the conduction band comes from the s and p states.Moreover, the change from the 3.2 eV (commercial ZnO) to 2.22 eV (black ZnO) is due to the modification of PDOS in the conduction band as a consequence of s and p states.
Through the estimation of the valence band maximum (VBM) and the energy band gap (E g ) from XPS and UV-vis spectroscopy; it was possible to build a schematic band edge representation of the two ZnO compounds, figure 9.It is observed that vacancies in the black ZnO structure causes a modification of the band edges moving into a region within a more oxidative character, proposing to black ZnO as a reasonable promotor of OH radicals [14,18]; and thus, opening their exploration as a feasible photocatalyst in the UV visible region; in good concordance with preceding reports where vacancies in the structure of the compounds promotes their catalytic performance [74][75][76][77][78][79].

Conclusions
It is proposed for the first time a new scalable green synthesis methodology to produce black ZnO nanoparticles mediated by coffee arabica L extract.The crystallographic and microstructural analysis shows particle sized distribution between 15 and 30 nm.The presence of vacancies in the ZnO structure was supported experimentally by Rietveld analysis, HRTEM, XPS and theoretical calculus by DFT scheme.UV-vis spectroscopy estimates a narrowed band gap of 2.22 eV while the theoretical DFT+U Hubbard semi-empirical model accurately reproduces the electronic structure of black ZnO by comparison with Raman, FTIR and XPS in the valence band region, when it is considered UZnd = UOp = 5 eV.The presented results are foregoing to exploring green synthesized black ZnO nanoparticles in plausible applications promoting catalytic activity by influence of vacancies.

Figure 1 .
Figure1.Schematic green synthesis process followed to achieve black ZnO NPs.

Figure 2 .
Figure 2. Rietveld refinement of the black ZnO NPs.Inset 1: Photograph of visual inspection showing the color of the ZnO powder.Inset 2: Schematic representation of the (002) plane in ZnO crystal structure.

Figure 3 .
Figure 3. (a)) and (b) TEM micrograph of the ZnO NPs and their particle size distribution.Figure 3(c) SAED image with planes associated to the ZnO crystal structure.Figurex 3(d) and (e) HRTEM micrograph with a super imposed supercell model of ZnO oriented along the (101) plane.

Figure 5 .
Figure 5. Survey XPS spectra from the commercial 5. (a) and (b) black ZnO synthesized from the coffee extract.It is observed a similar core level region contribution related with Zn -O interaction.

Figure 6 .
Figure 6.High resolution XPS spectra of commercial and black ZnO NPs.Figures 6(a) and (b) shows the core level region associated to the adventitious carbon C1s.Figures 6(c) and (d) shows the 2p core level region of Zn atoms while figures 6(e) and (f) displays the 1s region linked with oxygen atoms in ZnO.Finally, it is shown the valence band region of both commercial and black ZnO nanoparticles, figures 6(g) and (h), respectively.

Figure 7 .
Figure 7. Experimental FTIR, Raman and XPS spectrum in the valence band region compared against theoretical DFT calculations for the black ZnO NPs.

Figure 8 .
Figure 8. IR and Raman modes from the optimized ZnO computed by DFT with the LDA+U approximation with U Znd = U Op = 5 eV.

Figure 9 .
Figure 9. Schematic band edge representation of the two ZnO compounds.

Table 1 .
Structure parameters from the Rietveld refinement analysis for black ZnO.

Table 2 .
XPS core lever region measured for the commercial and black ZnO nanoparticles.