Electric-Field-Induced Phase Change in Copper Oxide Nanostructures.

Transition-metal oxides such as cupric and cuprous oxides are strongly correlated materials made of earth-abundant chemical elements displaying energy band gaps of around 1.2 and 2.1 eV. The ability to design nanostructures of cupric and cuprous oxide semiconductors with in situ phase change and morphological transition will benefit several applications including photovoltaic energy conversion and photoelectrochemical water splitting. Here, we have developed a physicochemical route to synthesize copper oxide nanostructures, enabling the phase change of cupric oxide into cuprous oxide using an electric field of 105 V/m in deionized water via a new synthetic design protocol called electric-field-assisted pulsed laser ablation in liquids (EFA-PLAL). The morphology of the nanostructures can also be tuned from a sphere of ∼20 nm to an elongated leaf of ∼3 μm by controlling the intensity of the applied electric field. Futuristically, the materials chemistry occurring during the EFA-PLAL synthesis protocol developed here can be leveraged to design various strongly correlated nanomaterials and heterostructures of other 3d transition-metal oxides.

Introduction

Binary oxides of copper such as cupric (CuO) and cuprous (Cu2O) are p-type semiconducting materials with energy band gaps of ∼1.2 and ∼2.1 eV, respectively.[1] The crystal structure of CuO is monoclinic, which is described by the centrosymmetric space group C2/c, whereas Cu2O has a cubic crystal structure, which belongs to the pn3 space group. Within the copper ions, d electrons experience Coulombic repulsion, which tends to localize individual electrons at atomic lattice sites, while hybridization with the oxygen’s p electron states tends to delocalize the electrons.[2] These copper oxides (CuO; x = 1 or 2) are consequently materials having the appropriate energy band gap to enable the absorption of light and generate charge carriers via photovoltaic phenomena; and to split the water molecule into hydrogen and oxygen using solar energy through photoelectrochemical phenomena. Indeed, the energy required for water splitting is between 1.6 and 2.4 eV,[3] and CuO has a suitable energy band gap for oxidizing and reducing water. Furthermore, the CuO and Cu2O nanostructures (NSs) are also used for other applications such as batteries,[4,5] gas sensors,[6] catalysts,[7,8] and pigments.[9]

To further develop several technological applications such as photovoltaic energy conversion and photoelectrochemical water splitting, our ability to design strongly correlated structures of CuO semiconductors with in situ phase change and morphological transition is crucial. There exist several thermodynamic pathways to promote the phase change of CuO into Cu2O.[10] Recently, some studies have reported the synthesis of CuO NSs by the hydrothermal method[11] and CuO nanowires (NWs) by an electric-field-assisted thermal oxidation method.[12] Electric field plays a critical role in controlling the growth and morphology of the synthesized CuO NWs as ion diffusion and velocity can be tuned via an external electric field.[12] Earlier studies had also investigated the role of electrical discharges with various conductivities (chemically controlled) on the growth of Cu-based NSs.[13]

We present here a new pathway to induce a phase transition in CuO NSs by transforming CuO into Cu2O using an electric field, which is mechanistically understood via the Oswald ripening process. Indeed, by increasing the value of the electric field from 0 to 105 V/m, there is a structural phase change from CuO to Cu2O. Moreover, there is also a morphological transition, accompanying the structural change, evolving from a sphere of ∼20 nm to an elongated leaf of ∼3 μm. This paper reports for the very first time a synthesis of CuO NSs by electric-field-assisted-pulsed laser ablation in liquids (EFA-PLAL).

EFA-PLAL Synthesis, Mechanism, and Characterization

Background

CuO nanostructures (NSs) can be obtained by various methods including the hydrothermal synthetic method, chemical precipitation methods, thermal conversion of precursors, electrochemical method, thermal oxidation method, and pulsed laser ablation in liquids (PLAL).[1] PLAL is an interesting technique as it produces NSs with a naked surface. A complete list of CuO NSs performed by PLAL is given in Table .

Table 1

CuO NSs Synthesized by Laser Ablation in Liquids (LAL)a

ref	type of laser	repetition rate (Hz)	irradiation time (min)	solvent	energy/pulse or fluence	result
Femtosecond – PLAL
(14)	Ti:Sapphire at 800 nm	1000	9	deionized (DI) water, acetone	500 μJ/pulse and 50 μJ/pulse	Cu at Cu2O NPs and Cu2O at Cu NPs
Picosecond – PLAL
(15)	Nd:YAG at 532 nm	10	60	DI water, ethanol	31 mJ/pulse	Cu2O NPs
Nanosecond – PLAL
(16)	Nd:YAG at 355 nm	10	60	DI water	150 mJ/pulse	Cu/CuO NPs
(17)	Nd:YAG at 532 nm	10	15, 30	spinach water extract	200 mJ/pulse
(18)	Nd:YAG at 532 nm			DI water	20 mJ/pulse	CuO NPs
(19)	Nd:YAG at 532 nm	6	15, 30, 45, 60	DI water	30 mJ/pulse	CuO NPs
(20)	Nd:YAG at 532 nm	10	15	DI water with hydrogen peroxide (0–5%)	60 mJ/pulse	CuO and Cu2O NPs
(21)	Nd:YAG at 532 nm	10		DI water	25 mJ/pulse	Cu a tCu2O NPs
(22)	Nd:YAG at 1064 nm	1	5–20	DI water	40–200 mJ/pulse	CuO NPs
(23)	Nd:YAG at 1064 nm		10	DI water, acetonitrile, methanol, ethanol, hexane, 1-propanol, butanol, ethylene glycol	80 mJ/pulse
(24)	Nd:YAG at 1064 nm	10		poly(vinyl alcohol)	75 mJ/pulse
(25)	Nd:YAG at 1064 nm	10	60	DI water, ethylene glycol	27 J/cm2 and 80 J/cm2	hollow CuO NPs and Cu NPs
(26)	Nd:YAG at 1064 nm	10	4	ethanol	1.5 J/pulse	Cu2O NPs
(27)	Nd:YAG at 1064 nm			DI water	40 mJ/pulse	CuO and Cu2O NPs
(28)	Nd:YAG at 1064 nm	20		DI water, sodium hydroxide, hydrogen peroxide, ethyl alcohol
(29)	Nd:YAG at 532 nm	10	15, 30, 60	methanol, 2-propanol	30 mJ/pulse	Cu at Cu2O NPs
(30)	Nd:YAG at 1064 nm	10	60	PVP aqueous solution	80 mJ/pulse	Cu2O at CuO NPs
(31)	Nd:YAG at 1064 nm	10	5	DI water, acetone	130 mJ/pulse	CuO NPs, Cu NPs
(32)	Nd:YAG at 1064 nm	10	1.66	DI water	10 mJ/pulse	Cu/CuO NPs
(33)	Nd:YAG at 1064 nm	20	60	DI water	180 mJ/pulse	Cu2O NPs
(34)	Ce:Nd:YAG at 1064 nm	10	30	hydrogen peroxide	40, 70, 100 mJ/pulse	Cu/CuO flakes
(35)	Yb fiber at 1064 nm	21 000		DI water	1 mJ/pulse	CuO NPs
Continuous – LAL
(32)	diode laser at 530 nm	CW	120	DI water	/	Cu/CuO NPs
Nanosecond – EFA-PLAL
this work	Nd:YAG at 1064 nm	1000–15 000	30 + 30	DI water	350 J/cm2	CuO/Cu2O NSs

LAL, laser ablation in liquids; PLAL, pulsed laser ablation in liquids; EFA-PLAL, electric-field-assisted-PLAL.

Synthesis

The NSs were created using a two-step process. The first step was a regular PLAL setup (Figure a). A Q-switched Nd:YAG laser from Electro Scientific Industries operating at 1064 nm was used to irradiate a Cu target. Spherical Cu beads (99.99% from Sigma Aldrich) were used as targets in this experiment. The diameter of the beads was ∼2 mm. A 25 mL rounded flask was used as a container, and 3 mL of DI water was poured into it. Consequently, the height of the liquid above the surface of the target was measured to be 10 mm. The pulse repetition rate of the laser was varied from 1 to 15 kHz. Consequently, the pulse duration time varied slightly from 70 to 200 ns depending on the repetition rate. The laser shined a pulsed beam with an energy per pulse of around 5.5 mJ per pulse at 1 kHz. The beam was deflected by a flat mirror oriented at a 45° angle (with respect to the laser rail) to irradiate the target from the top and was then focused using an 83 mm focal length lens. The beam’s spot size on the target was measured by scanning electron microscopy (SEM) to be around ∼45 μm. Therefore, the intensity of the laser was determined to be around ∼3.5 × 105 W/cm2. At 1 kHz, the fluence was calculated to be ∼3.5 × 102 J/cm2. The Cu target was finally irradiated for 30 min.

Figure 1

(a, b) Sketch of the synthesis protocol. Step 1 uses a PLAL setup, while step 2 uses an EFA-PLAL setup.

The second step is an EFA-PLAL setup, as depicted in Figure b. EFA-PLAL differs from the PLAL setup in using a direct-current (DC) electric field with adjustable voltage from two parallel electrodes being applied on both sides of the Cu target.[36] The electrodes were two rectangular plates made of Cu. The laser beam was unfocused during the second step, and the target was removed from the container. The colloid obtained after the first step was irradiated for another 30 min. The container was a square cuvette with the electrodes placed face to face on two opposite walls of the cuvette. The potential difference applied between the electrodes was kept constant during the entire duration of the second step at 0, 100, 250, 500, 750, or 1000 V. All of the synthesis details are summarized in Table .

Table 2

Parameters of the EFA-PLAL Synthesis

parameter	step 1	step 2
mode	PLAL	EFA-PLAL
laser	Nd:YAG	Nd:YAG
wavelength (nm)	1064	1064
repetition rate (kHz)	5.1	5.1
duration (min)	30	30
target	Cu beads	_
beam	focused	unfocused
volume (mL)	3	3
container	25 mL single-neck rounded flask	3 mL square cuvette

Mechanism

The solvent used in this synthesis protocol, DI water in our case, confines the plasma plume and also provides a reactive medium to generate a compound based on the target’s chemical element, in this case, Cu. The electron configuration of Cu is [Ar]3d104s1, meaning that Cu has 11 valence electrons, 1 belonging to the 4s orbital and 10 belonging to the 3d orbital.[37] When the laser beam hits the Cu target, it starts ionizing Cu into Cu+ and 1e–, by removing the electron from the 4s orbital. Another electron can easily be removed from the 3d orbital to form Cu++. The laser beam also break downs the water molecule of the solvent into hydrogen ions (H+) and oxygen ions (O2–) or into hydrogen ions and hydroxide groups (OH–). When the plasma cools down (laser beam is off), the Cu+ ions, Cu++ ions, H+ ions, O2– ions, and electrons (e–) start reacting all together to form CuO or Cu2O according to the chemical reactions shown in Scheme .

Scheme 1

Possible Chemical Reactions Occurring during the EFA-PLAL Synthesis Protocol

Based on the ionic species present during the irradiation, Cu+ ions and Cu++ ions will react with O2– because the enthalpy of formation of CuO is more negative than the enthalpy of formation of copper hydrides (CuH or CuH2). Moreover, the chemical reaction among Cu2+, OH–, and e– will not occur under our synthesis conditions despite a favorable formation enthalpy. Indeed, the presence of the electric field influences the way Cu will be ionized. Without an electric field being present, Cu can be ionized two times into Cu++ as the Cu+ ions do not move (no electric field present) after being ionized the first time by the laser beam. Therefore, CuO is expected without an electric field. When an electric field is present, the Cu+ ions created by the interaction of the laser beam and the Cu target move along the electric field lines and consequently leave the ablation area; therefore, Cu+ ions cannot be reirradiated a second time by the laser beam to form Cu++. Thus, Cu2O is expected when an electric field is present.

Characterization

The colloids were then characterized by atomic emission spectroscopy (4210 MP-AES from Agilent), scanning electron microscopy (JEOL JSM-7000F, equipped with a field emission gun, and operating at 15 kV), and transmission electron microscopy (Thermo Fisher Scientific Talos F200X operating at 200 kV). To perform Raman, FT-IR, and SEM analyses, a droplet of colloids was deposited onto a silicon wafer and dried in an environmentally controlled glovebox. The samples for the TEM study were each prepared by the drop-casting of one droplet of the colloid onto a TEM grid followed by air-drying. Raman spectra were obtained with a RENISHAW inVia Qontor confocal Raman microscope with thermoelectric cooling for ultralow noise levels. The spectra were recorded in a backscattering configuration with ultrafast data collection (over 1800 spectra per second). The excitation line was at 532 nm, and the laser power was kept at 50% of the source power to avoid heating the samples. The IR measurements were performed using a Varian 670 FT-IR spectrometer.

Results and Discussion

The lifetime of the cavitation bubble can be determined by analyzing the concentration of the colloid versus the repetition rate of the laser (Figure a). At low repetition rates, the concentration of NSs increases with the repetition rate. However, by shining the laser beam too fast on the target, the laser will encounter the cavitation bubble, which will correspond to a decrease in concentration of NSs within the colloid. From Figure a, it is seen that the maximal production of NSs is reached at 5.1 ± 0.1 kHz; consequently, the cavitation bubble lifetime is estimated to be ∼0.196 ± 0.004 ms. At 10.2 kHz, there is a second peak that still produces some significant amount of NSs, but the concentration is less than the one at 5.1 kHz because approximately only half of the laser shoots reach the target.

Figure 2

(a) Concentration of CuO colloids synthesized at 0 V for various repetition rates from 1 kHz up to 15 kHz. The optimal repetition rate was determined at ∼5.1 kHz. (b) Concentration of CuO colloids synthesized at 5.1 kHz for various potential differences from 0 V up to 1000 V. Inset: Photo of the colloids synthesized at 5.1 kHz under various potential differences. (c) pH of the CuO colloids synthesized at 5.1 kHz versus the potential difference used during the synthesis. All of the pH values have been measured at room temperature (T ∼ 23 °C). (d) Temperature of the colloids synthesized at 5.1 kHz versus the potential difference used during the synthesis.

After determining the best repetition (5.1 kHz) to irradiate the Cu target in DI water, the electric field was activated during the synthesis to generate five colloids at various voltages of 100, 250, 500, 750, and 1000 V. Another sample (at 0 V) was synthesized without the electric field to serve as a reference. A picture of the sample’s series can be seen in the inset in Figure b. Figure b shows the concentration of Cu-based NSs within each of the colloids synthesized at various voltages. As seen in Figure b, the concentration is multiplied by ∼5 in comparison to the colloid synthesized without any applied voltage. This can be understood by having the electric field ON between the electrodes; the cavitation bubble is shifted spatially, allowing the laser to reach the target without encountering any shielding effect due to the cavitation bubble and, consequently, enhancing the ablation of Cu. Consequently, the colloids (0, 100, 250, 500, 750, and 1000 V) are acidic by exhibiting a pH around ∼6.7 (Figure c). Indeed, the presence of CuO nanoparticles (NPs) in the solvent decreases the pH of the solvent compared to a pure DI water solvent, whatever the synthesis protocol is, PLAL or EFA-PLAL. In the Supporting Information, the pH of the colloid is also plotted as a function of the concentration of CuO NSs within the colloid (Figure S1). From Figure S1, it is clear that the EFA-PLAL was very efficient to increase the concentration of NSs with respect to a regular PLAL protocol. In Figure d, the temperature of the colloid was monitored, and its value increased with the value of the electric field applied during the synthesis.

The presence of two phases of CuO was detected by Raman spectroscopy (Figure a). Indeed, in Figure a, each colloid displays some peaks belonging to both oxides CuO and Cu2O. However, it seems that by increasing the value of the electric field during the synthesis, Cu2O is promoted compared to CuO. CuO seems to dominate the colloid at a low electric field. At 0 V, the oxidation state of Cu is II, consequently forming CuO NSs. By increasing the potential difference during the synthesis, the oxidation state of Cu decreases to I and forms Cu2O NSs. The Raman peaks of CuO and Cu2O are listed in Table . In the Ag and Bg Raman modes, only the oxygen atoms move, with displacements in the b-direction for Ag and perpendicular to the b-axis for Bg modes. The presence of CuO and Cu2O was also confirmed by FT-IR (Figure b). The infrared active modes involve the motion of both O and the Cu atoms. The vibration modes of CuO and Cu2O are illustrated in Figure c. The presence of IR bands at 900, 1100, and 2950 cm–1 confirm the Cu2O phase, while the presence of bands at 600, 750, and 2350 cm–1 affirm the CuO phase.

Figure 3

(a) Raman spectra of the colloids synthesized at 5.1 kHz under various potential differences. (b) FT-IR spectra of the colloids synthesized at 5.1 kHz under various potential differences. (c) Illustration showing the vibrational modes for CuO and Cu2O.

Table 3

Theoretical and Experimental Raman Frequencies of CuO and Cu2O

material	vibrational mode	theoretical (cm–1)	experimentala (cm–1)
CuO	Ag	294,[38] 296,[39] 298,[40] 319[41]	276
	Bg	348,[38] 346,[39] 345,[40] 382[41]	327
	Au	421[38]	455
	Bg	624,[38] 631,[39] 632,[40] 639[41]	671
Cu2O	T2u/Γ25-	88[42]	not observed
	Eu/Γ12-	110[42]	145
	T1u/Γ15-	149,[42] 153[42]	216
	A2u/Γ2-	348[42]	not observed
	T2g/Γ25+	515[42]	not observed
	T1u/Γ15-	640,[42] 660[42]	672

This work.

The nature of the oxide present at the surface of the nanostructure has been investigated by X-ray photoelectron spectroscopy (XPS). The PLAL sample synthesized at 0 V looks very different from all of the other samples synthesized with a nonzero potential difference (EFA-PLAL), as can be seen in Figure . It is possible to distinguish the Cu oxidation states using the satellite peak features of Cu 2p.[43,44] In Figure a, there is a weak satellite peak around ∼944 eV, indicating the presence of Cu2O at the surface. In Figure b–f, there are two strong satellite peaks around ∼945 and ∼965 eV, indicating the presence of CuO at the surface of the NSs.

Figure 4

XPS of the CuO NSs synthesized at 5.1 kHz under various potential differences: (a) 0 V, (b) 100 V, (c) 250 V, (d) 500 V, (e) 750 V, and (f) 1000 V.

Finally, the morphology of CuO NSs was analyzed by SEM and is shown in Figure . At 0 V, the CuO NSs are spherical; however, at nonzero voltages, the NSs are elongated into leaves. The larger size of NSs obtained at higher potential differences can be explained by the larger temperatures reached by the colloid during the synthesis when the electric field is applied (Figure d). Indeed, the more intense the electric field (i.e., potential difference across the cuvette), the higher the temperature. Consequently, the NSs become bigger because of the Oswald ripening process.[45] This process is a heat-induced size change of NSs in solution. By increasing the potential difference from 0 to 1000 V, the surface-to-volume ratio of the NSs increases by having the morphology evolving from a sphere to a nanoleaf. Furthermore, the composition of CuO seems also to evolve from CuO to Cu2O when the potential difference increases from 0 to 1000 V (Figure a). This effect on the chemical composition may also be explained by the Oswald ripening process.[46] However, the Oswald ripening process explains why we observe larger structures at higher potential differences but does not explain the elongation. Indeed, the elongation is due to a growth-oriented attachment process occurring by aligning the seeds along the direction of the electric field, consequently creating nanoleaves when an electric field is applied. More SEM images are shown in Figure S2.

Figure 5

SEM images of the CuO NSs synthesized at 5.1 kHz under various potential differences: (a) 0 V, (b) 100 V, (c) 250 V, (d) 500 V, (e) 750 V, and (f) 1000 V.

In Figure , let us focus on the synthesized NSs synthesized under the two most extreme potential differences, i.e., 0 V and 1000. The sample at 0 V shows CuO NPs with sizes smaller than 20 nm, while the sample at 1000 V shows Cu2O NSs with the length reaching several hundreds of nanometers (inset). It is still possible to encounter spherical NPs in the sample at 1000 V as the spherical NPs produced during the first step process may not have crossed the laser beam path during the second step process. The electron diffraction of both samples confirms the crystallinity of the NSs produced.

Figure 6

(a, b) Tauc plots of the CuO NSs synthesized at 5.1 kHz under potential differences of 0 and 1000 V, respectively. Inset: SEM image of the CuO NSs.

The Tauc plot calculated from the absorbance curve obtained by UV–visible spectroscopy on these two extreme samples shows energy band gaps around 2.6 and 3.2 eV at 0 and 1000 V, respectively. Based on the previous discussion, it is evident that the CuO NSs displayed a band gap around 2.6 eV, while Cu2O NSs exhibited a band gap around 3.2 eV. This is in good agreement with what is observed in the literature.[47,48] These values are much larger than the bulk energy band gaps of CuO and Cu2O (1.2 and 2.1 eV, respectively). This is due to the small size of the CuO NSs, as shown in Figure . Indeed, as the size decreases, the energy band gap increases.[49,50]

Figure 7

(a, b) TEM images of the CuO NSs synthesized at 5.1 kHz under a potential difference of 0 and 1000 V, respectively. (c, d) Electron diffraction patterns corresponding to the TEM images shown in (a) and (b), respectively.

To further confirm the electric-field-induced phase transition in the synthesized CuO NSs, we performed TEM and selected area electron diffraction (SAED) studies. The typical diameter of CuO nanoparticles prepared at 0 V is observed to be ∼20 nm, which is shown in Figure a. The corresponding SAED pattern (Figure c) confirms the presence of the CuO phase with (022), (−202), and (002) crystal planes with concentric rings of distinct bright spots confirming the crystalline structure of CuO NSs. In contrast, the TEM topography image (Figure b) for the samples prepared at 1000 V shows elongated leaflike NSs of over 250 nm. The corresponding SAED pattern in Figure d demonstrates the presence of several crystal planes such as (331), (311), (022), (111), (−202), and (002), confirming the formation of both CuO and Cu2O crystalline phases.

Conclusions

CuO NSs have been successfully synthesized by the EFA-PLAL technique. The first step of the synthesis protocol serves to synthesize the CuO seeds, while the second step serves to elongate the seeds into their final shape. The intensity of the electric field has a huge influence on the morphology of the NSs and the crystalline phase of the oxide formed. When the electric field is OFF, the NSs are spherical, while when the electric field is ON, elongated nanoleaves are formed. The electric field used during the synthesis also helps control the oxidation state of Cu by promoting a phase transition from CuO to Cu2O. At low electric fields, Cu II (CuO) is favored, while at higher electric fields, Cu I (Cu2O) is preferred. Indeed, from the Raman and XPS analyses, it is clear that at 0 V the core and the surface of the spherical nanoparticles are made of CuO, while at 1000 V, Cu2O becomes predominant in the core and the surface. Also, the two structures displayed significantly different energy band gaps around 2.6 eV for the CuO NSs and around 3.2 eV for the Cu2O NSs. Moreover, the electric field also helped enhance the concentration of the colloid by a factor of ∼5. Finally, this new synthesis protocol could be extended to synthesize other 3d transition-metal oxides, which will pave the path to new oxide heterostructure designs.