Value-added fabrication of NiO-doped CuO nanoflakes from waste flexible printed circuit board for advanced photocatalytic application

Crystallographic structure, morphological, and spectroscopic analysis of NiO-doped CuO nanoflakes

The recorded XRD pattern of CuO-based nanomaterial was processed using HighScore plus software (supplier: PANalytical BV) for a candidate search match in the ICSD database, followed by a Rietveld fit to identify and quantitatively establish the abundances of Tenorite (CuO) and NiO (as shown Fig. 2). An acceptable Rietveld fit was obtained with these two phases (Rwp: 9.91). Nanoflakes produced in the system were confirmed to be predominantly Tenorite, CuO (98.5% ± 4.5) with traces of nickel oxide, NiO (1.5% ± 0.1).

Figure 2

(a) XRD pattern of nanoflakes with all peaks explained by CuO as a major phase and NiO as a trace phase. (b) Rietveld fit analysis for composition with relative abundances of Tenorite (CuO) and NiO. Inset: The unit cell of CuO.

After refinement, the lattice parameters were calculated to be: a = 4.696(2) Å, b = 3.414(3) Å, c = 5.137(9) Å, and β = 99.2°, with cell volume 81.31 Å3. This can be compared to ICSD data card values of a = 4.683(0) Å, b = 3.459(0) Å, c = 5.130(0) Å, and β = 99.309° with cell volume 81.29 Å3 for CuO. A full pattern Rietveld fit with size31 and strain-specific parameters32 was applied to approximate the average size of the crystallites and developed microstrain. The instrument broadening was taken into account by measuring a Silicon standard sample (Si 640c). The CuO crystallite size was found to be 7.7 nm, and the associated microstrain was estimated as 0.207%. The following equation was used to calculate dislocation density (δ), which is defined as the length of a dislocation line per unit metre square of the crystal:

$$ \delta = \frac{1}{{D^{2} }} $$


The obtained value of dislocation density, dislocation length, and lines per unit volume of a crystal structure is: δ = 0.0169 and strain ɛstr = 0.207%. This signifies a superior crystallisation and good quality of the CuO nanoflakes, which may be appropriate for applications in photovoltaics. This study found that the measured strain resulting from the dislocation density and lattice dislocation was deficient and had no effect on the broadening peaks for CuO tetrapods.

The microstructure, crystal structure and elemental mapping of the NiO-doped CuO nanoflakes were investigated using HR-TEM imaging, SAED, and TEM-EDS analysis represented in Fig. 3. The nanoflakes demonstrate an irregular pattern and feather-like morphology with a varied flake size. The irregular morphologies of the flakes are formed by assembling many 1D nanorods visible in the HR-TEM images. As the flakes are not a single structure but the assembly of several 1D nanorods, the width varies from  10 to 50 nm and the length varies from  30 to 80 nm. From the HR-TEM images, it is clearly understood that many rod-like particles are accumulated together to develop the structure of the nanoflakes which could be ascribed to the well-known Ostwald ripening phenomena40. In this procedure, at the beginning of the synthesis of the nanoflakes, particles ranging from small to large are generated in the non-equilibrium solution. The smaller particles dissolve easily and create free atoms which are transferred to the surface of the bigger particles. This process continues because the bigger crystals are energetically favourable compared to the smaller crystals and promote more solubility for the smaller crystals. The reprecipitation of the smaller crystal on the surface of the larger crystals creates a compact structure which is favourable for the use of this CuO in solar cells for the transportation of photocurrent. The CuO nanoflakes show well-defined fringes, which are attributed to the single crystal of CuO. The measured lattice spacing was 2.75 Å, which is attributed to (110) interplanar spacing. The SAED pattern also confirms the absolute monoclinic structure, which corresponds to the XRD pattern. The TEM-EDS mapping shows Cu and O distribution in the nanoflakes.

Figure 3
figure 3

(a,b) Bright-field TEM image, (c) lattice structure of nano CuO flakes with electron diffraction pattern.

The porosity and the specific surface area of CuO nanoflakes were measured using N2 isotherms for adsorption–desorption at 77.4°K and pore size distribution (BJH) measurements. The isotherm exhibits type IV hysteresis, as represented in Fig. 4, and the relative pressure (P/Po) and loop are from 0.65 to 1.0, which further indicates the structure is mesoporous. The collected CuO nanoflakes have a specific surface area of 115.703 m2 g−1, and the pore size distribution (BJH) indicates that the CuO nanoflakes have a mesoporous structure with an average 6 nm pore diameter.

Figure 4
figure 4

Isotherms for N2 adsorption–desorption at 77 °K and pore size distribution (BJH) of the synthesised nano CuO.

The surface chemistry of the CuO nanoflakes was explored using UV–Vis spectroscopy, represented in Fig. 5a. The reflectance characteristics at different wavelengths were also analysed with UV–Vis spectroscopy. Due to the initial stability and lance changing for UV–Vis spectroscopy setup, there are a few smaller humps enlarged with scales. Those humps are very common in UV–Vis spectroscopy for nanomaterials in the range of 300–500 nm. The nanoflakes show around 70–75% reflectivity in the visible and infrared region, which is a good indication of the applicability of this material in solar energy harvesting41. The indirect bandgap can be estimated using a Kubelka–Munk plot or Tauc plot (shown in the inset of Fig. 5a). The relationship between the reflectance and estimated bandgap can be written as:

$$ F\left( R \right) = \frac{{\left( {1 – R} \right)^{2} }}{2R} $$


where R is the reflectance percentage of CuO measured by the UV–Vis spectroscopy at different wavelengths. The bandgap measured for this CuO by the Kubelka–Munk plot is ~ 1.57 eV, which is relatively higher than that of the bandgap of bulk CuO42. The increase in the energy of the bandgap can be ascribed to quantum confinement in the nanocrystalline arrangement. This quantum confinement can be caused by the presence of another oxide (for example, NiO) in the original composition of CuO36,37.

Figure 5
figure 5

(a) UV–Vis spectrum (inset: Kubelka–Munk plot) and (b) Photoluminescence analysis for CuO.

The photoluminescence of a material varies with factors including the composition, synthesis technique, and storing system. The spectra can be derived from the combination of the free transporters in the defect in the energy state. It is a sign to estimate the bandgap of the material. The photoluminescence spectra for CuO are shown in Fig. 5b. There are few emission peaks at varying wavelengths, and all four major peaks are in the visible range. The peak at 407 nm is for violet emission, at 451 nm for the blue region, and the significant one at 572 nm can be attributed to the green emission region43. Some other minor peaks are derived from the free excitation of electron–hole pairs and their recombination42. The photoresponse of NiO is also identified at 691 nm. This emission peak is considered to be due to the participation of NiO in the mixture of CuO nanoflakes42,44.

Electrochemical properties of CuO nanoflakes

In the next step, the nanoflakes were rapidly heat-treated at a low temperature (5 min at 400 °C). They were then used to make ink for the fabrication of a film electrode on the surface of a current collector (FTO). In the first step, the photocurrent onset potential (Eonset) and the capability of the film to generate photocurrent were examined. Eonset is the potential where the minority electron carriers in the photocathode trigger a Faradic reaction (which in this research is hydrogen evolution reaction (HER)) at the interface of solid/liquid45). Figure 6a illustrates the j–V plots of the film at a very low current density, where the small photocurrents can be identified under illumination. It can be seen that the potential 1.15–1.20 V versus RHE can be selected as Eonset, and the nanoflake film electrode behaves like a photocathode (p-type)46,47. It is well known that the negative shift of the onset voltage potential is suitable for cooperation with the anodes to construct a non-biased PEC water splitting cell48. The plot for j–V of the film is shown in Fig. 6b. This was obtained using linear sweep voltammetry. The ratio \(\frac{{j}_{light}}{{j}_{dark}}\) for the film was approximately 33, and the photocurrent value of 1.9 mA cm−2 is amongst the highest values reported in the literature for CuO47.

Figure 6
figure 6

Photoelectrochemical results of the film electrode made from heat-treated (at 400 °C for 5 min) NiO-doped CuO nanoflakes employed in 2 M KOH solution and under illumination intensity of AM 1.5 G; (a) photocurrent (j–V) plot of the film at low current density for identification of onset potential, (b) three-electrode photocurrent plot of the film, the scanning rate of 20 mV s−1 and chopped light illumination at ~ 0.5 Hz frequency, (c) Nyquist plot of the film at − 0.3 V versus SCE under illumination and in dark conditions, in the frequency range of 0.1 Hz to 0.1 MHz under 10 mV AC amplitude, (d) calculated functional bias photon to electron conversion efficiency (ABPE) of the film electrode extracted from the data obtained from continuous illumination and continuous dark conditions [see (b)], (e) photon incident to current conversion effectiveness (IPCE) for the film electrode at − 0.3 V versus SCE, and (f) two-electrode photocurrent plot of the film.

Figure 6c depicts the Nyquist plot derived from the EIS test to study the transfer charge rate across the interface of the electrolyte and electrode. The diameter of the semicircle indicates the resistance for the charge transfer (Rct)47,49; hence, the smaller semicircle indicates a remarkable increase in electron conductivity. In the EIS of photoelectrodes, the diameter of semicircles presents the charge transfer or/and mass transfer resistance across the electrode/electrolyte interface50. The presence of one semicircle means the lack of faradic reactions, which means only charge transfer occurred across the Stern layer51. In another word, the bigger the semicircle, the more insulator the electrode. This enhanced charge transfer at the interface is associated with the photoinduced surge of carrier density51, proving the photocatalytic activity of the electrode in harvesting solar powder. The initial resistance value (known as the electrode resistance, or Rp) for dark and illuminated conditions is identical at ~ 8 Ω. This value is related to the sum of the resistance of the working electrode and the contact resistance between the current collector and electrode50. The real resistance in the semicircle can be attributed to the mass transfer and charge transfer rates at the interface of the film and electrolyte44, where the value for the film under dark conditions (~ 11,000 Ω) is orders of magnitudes greater than that of under illumination (~ 90 Ω). This significant difference reflects the outstanding photocatalytic activity of the film, which gives it great potential for energy harvesting from sunlight.

According to Fig. 6a, the photoelectrode is not able to generate current beyond 1.15 V versus RHE, although 1.23 V is a thermodynamical requisite for splitting water. Hence, the performance of the photoelectrode film in the photocatalytic water-splitting process was measured by determining the employed bias photon to electron conversion effectiveness (ABPE) (where bias is varied between working and counter electrodes) using the following formula52,53:

$$ {\text{ABPE}}\;(\% ) = \left( {\frac{{\left[ {{\text{J}}_{{\text{p}}} \times \left( {1.23 – \left| {{\text{V}}_{{\text{b}}} } \right|} \right) \times\upeta _{{\text{F}}} } \right]}}{{{\text{P}}_{{{\text{total}}}} }}} \right) \times 100\% $$


where 1.23 V versus RHE indicates the minimum thermodynamic voltage for the splitting of water molecules54, Jp is the density of the photocurrent at the used bias voltage (mA cm−2), Vb is the applied bias voltage (V), Ptotal is the incident light’s intensity (mW cm−2), and ηF is Faradic efficiency (taken in this study to be 0.8, with a conservative approach). Figure 6e represents the ABPE of the CuO nanoflake film. The measured efficiency peaks at 0.52 V with ABPE effectiveness of 1%, which is quite promising for water splitting. It should be noted that the water-splitting potential for a given material should not exceed the thermodynamic potential. No sacrificial donors or chemical bias were used in this analysis, and the bias of the electrodes counter along with the reference electrode was reported.

As a function of excitation wavelength, the incident photon-to-current efficiency (IPCE) corresponds to the ratio of the photocurrent and the rate of incident photons from a light source. Monochromatic light sources were employed to irradiate the electrode, and the IPCE factor was calculated using the following equation45:

$$ IPCE\;(\lambda ) = \left( {\frac{{\left| {J_{p} } \right| \times 1240}}{{P_{mono} \times \lambda }}} \right) $$


where 1240 nm is a multiplication of plank’s constant (h), Pmono is the power intensity of the monochromated illumination (mW cm−2), and λ is the illuminated light’s wavelength (nm).

The plot of IPCE versus wavelength is illustrated in Fig. 6d. IPCE of the film shows a continuous increase from 800 to 400 nm, which is more promising than the results in the literature55,56 for water-splitting applications. This increase implies that the bandgap of the nanoflakes is somewhat less than 1.6 eV, as the nanoflakes can be stimulated with the light of 800 nm wavelength. This bandgap could be due to the presence of NiO in the CuO phase and is in good agreement with previously derived results. The plot peaks at 432 nm to ~ 18% under − 0.3 V versus SCE. These values are greater than those reported in the literature55,56.

As two-electrode systems are employed in practical applications of photocatalytic materials, the j–V plot of the film electrode was plotted using a two-electrode system, where the Pt wire and film electrode were applied as counter and working electrodes, respectively. The results are plotted in Fig. 6f. Photocurrent flowed from − 0.3 V and peaked at ~ 1.2 V with a current density of − 1.9 mA cm−2, which is equal to that in the two-electrode system. The overall shape of the j–V plot for the two-electrode system implies that when external bias is applied, this nanomaterial can be effectively employed beyond the potential of − 0.3 V for water-splitting applications.

In this research, similar to the most studies presented in the Table S2, we indirectly measured the performance of the photoelectrode in water splitting via analysing the onset potential, comparing the dark current and under illumination photocurrent, percentage of ABPE in a potential window of 0.6 V, and percentage of IPCE between the wavelength of 400 to 800 nm (which is the light wavelength), and measuring the photocurrent at a two-electrode system, all shown in Fig. 6. These indirect methods are well-described in Ref.54.

It has been proved that the conduction band of nano CuO is not able to provide enough negative potential to generate hydrogen57. To decompose water molecules, the electrode performance needs to be consistent with the high-level conduction band side, which is normally found to be − 0.2 to − 0.6 V versus RHE in recent literature57,58. The energy band of NiO is positioned between the energy bands of CuO and Cu2O Y58,59. Consequently, adding a small amount of NiO can enhance the electrochemical performance of CuO. The photo-generated electrons are transferred to the copper oxide materials and trapped for hydrogen evolution, which further is improved by the liberation of NiO once there is a homogeneous distribution in the base materials. If there are any cathodic transient peaks in the photoelectron activity, the doping materials, such as NiO and Pt, can assist elimination of those unexpected transitions and make the process relatively steady, which upturns the efficiency of hydrogen evolution in water-splitting. To increase the significance, a comparative study of the results of this work with existing data in literature works is tabulated and presented in the supplementary (Table S2).


Leave a Reply

Your email address will not be published. Required fields are marked *