The peak intensity ratio of I Lα /I Lβ of various copper oxide compounds.
The size and shape of nanoscale materials provide excellent control over many of the physical and chemical properties, including electrical and thermal conductivity, magnetic properties, luminescence, and catalytic activity (Lieber, 1998). In particular, the synthesis and morphological control of nanosized particles, which exhibit surprising and novel phenomena based on the unique property called the quantum size effect, are attractive to chemists and physicists (Alivisatos, 1996). In recent years, nanoparticles are widely used in many applications ranging from biosensing (Anker et al., 2008; Elghanian et al., 1997; Lin et al., 2006), plasmonic devices (Ferry et al., 2008; Maier et al., 2003), and multifunctional catalysts (Hu et al., 1999; Lu et al., 2004). There are a wide variety of techniques that are capable of creating nanoparticles with various morphology and production yield. These nanoparticle formation approaches are typically grouped into two categories: ‘top-down’ and ‘bottom-up’. The first involves the breaking down of large pieces of bulk material into the required nanostructures. The second involves the building of nanostructures, atom-by-atom or molecule-by-molecule in a gas phase or solution. These two approaches have evolved separately and reached the limits in terms of feature size and quality, in recent years, leading to exploring novel hybrid approaches in combining the top-down and bottom-up methods. Colloidal chemists have gained excellent controlled nanosized particles for several spherical metal and semiconductor compositions, which has led to the discovery of quantum size effect in colloidal nanocrystals (Alivisatos, 1996). However, various bottom-up approaches for making morphologically controlled nanoparticles have been found; most of these solution methods are based on thermal process. On the other hand, top-down approaches have been developed for producing metal and semiconductor nanowires, nanobelts, and nanoprisms (Hu et al., 1999; Pan et al., 2001; Jin et al., 2001). In particular, the laser-induced ablation method has become an increasingly popular approach for making nanoparticles due to the applicability to various target materials in an ambient atmosphere (Jia et al., 2006a; Kawasaki & Masuda, 2005; Link et al., 1999; Sylvestre et al., 2004; Tamaki et al., 2002; Tull et al., 2006). Recently, various shape-controlled nanoparticles, such as nanowires (Morales & Lieber, 2008), nanotubes (Rao et al., 1997), and composite nanostructures (Zhang et al., 1998), have been fabricated by this technique. More recently, pulsed laser ablation in liquid has become profoundly intrigued for preparing nanoparticles from the viewpoint of the concise procedure and the ease of handling (Kawasaki & Masuda, 2005; Tamaki et al., 2002; Shimotsuma et al., 2007).
Besides, in recent decades, increasing interest has focused on the femtosecond laser field due to its unique characteristics based on the nonlinear optical effects deriving from ultrashort pulse width and high peak power density (Cavanagh et al., 1993; Stuart et al., 1995). Especially, by connecting the femtosecond laser technology with nanotechnology, an effective approach was provided to prepare and modify nanostructures in micro devices (Kabashin et al., 2003; Kanehira et al., 2005; ; Link et al., 2000; Muskens et al., 2006; Qiu et al., 2004; Shimotsuma et al., 2003, 2007). For example, the ripple nanostructures, which are aligned along the femtosecond laser polarization direction, have been observed on the surface of the metal and semiconductor caused by the interferences between the scattering incident femtosecond laser field and the surface plasmon-polariton waves (Jia et al., 2006a). We have recently reported the first experimental evidence of the periodic nanostructures embedded in silica glass after irradiation by a single focused beam of a femtosecond laser (Shimotsuma et al., 2003). This phenomenon can be interpreted in terms of interference between the incident laser light field and the generated bulk electron plasma waves, resulting in the periodic modulation of electron plasma concentration and the structural changes in transparent material. A great success has been achieved in forming zero dimensional nanoparticles by using intense ultrashort light pulses radiation. We have successfully fabricated a colourful micro-pattern composed of Au or Ag nanoparticles inside a glass matrix (Qiu et al., 2004; Shimotsuma et al., 2007). Podlipensky and co-workers reported the formation of anisotropic Ag nanoparticles in glass which were elongated in the direction parallel to the femtosecond laser polarization (Podlipensky et al., 2003). Link and co-workers reported the size and shape change of Au nanorods under both femtosecond and nanosecond laser irradiation in solution (Link et al., 1999). The limited prior exploration for preparing one-dimensional nanoparticles by a femtosecond laser was concentrated on their fabrication on the materials’ surface. Jia et al. reported ZnSe nanowires’ growth induced by the femtosecond laser. Yet the nanowires grew only on the ablation crater located on the surface of ZnSe wafer (Jia et al., 2006b). Mazur et al. reported the production of submicrometer-sized spikes that consisted of the silicon on the silicon wafer after irradiating a silicon surface with a femtosecond laser (Baldacchini et al., 2006; Shen et al., 2004). Thus it should be interesting to investigate the preparation and evolution of nanowires under intense ultrashort light fields, which might reveal the interaction between the photons and the excited plasmons (materials) in the nanoscale.
In this Chapter, we present formation of one-dimensional metal Cu (Nakao et al., 2008; Shimotsuma et al., 2007) and ZnO (Lee et al., 2008, 2009) nanoparticles under femtosecond laser irradiation in liquid. It should be interesting to investigate the preparation and evolution of nanowires under intense ultrashort light fields, which reveals that the growth mechanism of Cu and ZnO nanowires could be a nucleation growth process (Chang et al., 2008; Wu et al., 2010).
2. Photofragmentation and evolution of metal Cu nanowire
In this section, the photofragmentation from Cu micro-flakes to nanowires and nanospheres via femtosecond laser radiation in alcohol solution was described. This phenomenon has provided the two distinct surface plasmon resonances based on the characteristic shape. The observed Cu nanowires of 50 nm diameter could be fragmented from the initial flakes as a result of the interference between the light field and the surface plasmon wave. By observing the morphology transformation of nanowires with time, the mechanism of Cu nanowires growth with femtosecond laser irradiation in the solution was also presented. Interestingly, the growth of the Cu nanowires was influenced by the incident light polarization. Other starting materials, e.g. Cu microspheres, were also tested to fabricate nanowires under the same technique to verify the mechanism and procedure of nanowires’ growth by another process. Furthermore, we demonstrate that the morphological control of Cu nanoparticles, dispersed in various alcohol solutions, under femtosecond laser irradiation. Beyond the basic understanding, such Cu nanowires have possible applications in the areas of plasmonic devices (Liu et al., 2008), surface-enhanced Raman scattering (SERS) (Sauer et al., 2006), and medicinal imaging (Desvaux et al., 2005).
2.1. Photo-conversion of Cu micro-flakes to nanowires
We used commercially available Cu micro-flakes produced by a chemical reduction method, which were 5 μm in size and 100 nm thick. A small amount of the Cu flakes, 0.36 mg, was mixed with 4.5 mL of 99.5% alcohol solution in a rectangular quartz vessel of 1 × 1 × 5 cm3 with a vessel thickness of 1.25 mm. The laser radiation in Gaussian mode produced by a regenerative amplified mode-locked Ti:sapphire laser (Coherent Inc., 100 fs pulse duration, 1 kHz repetition rate) operating at a wavelength of 800 nm was focused via a microscope objective (Nikon; LU Plan Fluor, 20× 0.40 N.A.) into the alcohol-suspended Cu micro-flakes placed on a magnetic stirrer. The typical beam waist diameter and laser energy fluence were estimated at ~ 4 μm and 2.4 × 103 J/cm2, respectively. The polarization of the laser light was set linear or circular by a half-wave or a quarter-wave plate placed in the incident beam before the focusing optics. A schematic of the experimental setup is shown in Fig. 1.
Before laser irradiation, the suspension was deaerated by bubbling nitrogen gas into it for 15 minutes. To keep as many Cu micro-flakes as possible suspended in the solution, the solution was continuously stirred with a magnetic stirrer during laser irradiation. The generated gas during laser irradiation was trapped in a gas-tight syringe at atmosphericpressure, and then analyzed by a gas chromatography (Shimadzu Corp., GC-2014). After laser irradiation, absorption spectra of the Cu suspension were measured by a spectrophotometer (JASCO, V-570). In order to analyze the nucleation site just after femtosecond laser irradiation, the suspension was immediately dropped onto a silicon substrate and the solvent was evaporated at room temperature. To reveal the growth mechanism of Cu particles, after laser irradiation, the suspensions were left at rest in a temperature-controlled bath having a constant temperature of 40 or 60 °C. The Cu particles after femtosecond laser irradiation were characterized by a field-emission scanning electron microscopy (JEOL, JSM-6700F), a transmission electron microscope (JEOL, JEM-2010F), and a field-emission electron probe micro analyzer (JEOL, JXA-8500F). All of the experiments were carried out at ambient temperature and pressure.
We observed the focal spot during femtosecond laser irradiation into alcohol solution along and perpendicular to the laser axis by using CCD camera (Fig. 2). These photos clearly indicate the generation of white-light filaments. It is well known that when a high intense femtosecond laser pulse is launched in gases (Braun et al., 1995; Corkum et al., 1986), liquids (Liu et al., 2002), and solids (Tzortzakis et al., 2001), it self-focuses and self-transforms into a white-light emission during femtosecond laser filamentation. Filamentation is commonly explained in terms of a balance between two nonlinear effects: Kerr self-focusing inducing a change of the refractive index of the medium proportional to the laser intensity; plasma defocusing through multiphoton and/or tunnel ionization of the medium in the high-intensity zone.
The pulse width at the focus after propagating in the dispersion medium can be estimated by the following equation:
where Δτ in , Δτ out , λ, L, c, d 2 n/dλ 2 are the input pulse width, the output pulse width, the wavelength, the propagation length in dispersion medium, the speed of light, and the second derivative of refractive index with respect to wavelength, respectively. From this equation, the input laser pulse width of 100 fs was stretched to 110 fs after propagating through 10 mm in methanol (Fig. 3). Such elongation of pulse width in alcohol solution does not make any difference in nonlinear light-matter interaction.
Fig.4 (a) shows the sequence of absorption spectra of the suspension, taken as a function of laser irradiation time. Minor absorption was observed for the dilute suspension of Cu micro-flakes in the wavelength region from 330 to 800 nm before the femtosecond laser irradiation while there were two apparent surface plasmon peaks in the absorbance (λ abs = 380 nm and 600 nm) during laser irradiation. The fluctuation of the surface plasmon peaks are considerably weaker as compared to the background interband transition that predominates
the high-energy wing (λ< 320 nm) of the spectra (Kawasaki & Masuda, 2005). We also observed an increase in intensity of the surface plasmon absorption for the spherical particles at λ abs = 600 nm with a concomitant growth of a new peak at λ abs = 380 nm (Fig.4 (b)). Indeed, the surface plasmon absorption of 10 nm spherical Cu nanoparticles is located at 560 nm, which can be predicted by Mie theory with no free electron density correction (Lipinska-Kalita et al., 2005). After 10 min, the peak at 380 nm completely disappeared. In the case of long pulse laser irradiation, these unexplained peaks have not been observed (Kawasaki & Masuda, 2005).
Scanning electron images (SEIs), when correlated with the time-dependent absorption spectroscopic observations, show that the initial Cu micro-flakes (size: ~ 5 μm, thick: ~ 0.1 μm) were converted to nanowires after the femtosecond laser irradiation between 1 and 5 minutes (Fig.5 (a) ~ (d)). During initial stages of the nanowires formation, both micro-flakes and nanowires can be seen (Fig. 5 (b)). The diameter of the nanowire increases with laser-irradiation time. The growth rate of the diameter of nanowires increases with the pulse repetition rate from 10.2 to 54.2 nm/min (Fig. 6 (a)). After 10 minutes irradiation, nearly all of the nanowires are converted to the nanospheres of 10 ~ 70 nm diameter (Fig. 5 (e), (f)). Detailed FE-SEM observations revealed that the diameters of Cu nanowire were about 62, 159, 270 nm for the number of light pulses of 6 × 104, 180 × 104, 300 × 104, respectively, and for the pulse energy of 0.4 mJ, corresponding to intensity of 1.6 × 1016 W/cm2. This indicated a linear dependence of the nanowire diameter on the number of light pulses (Fig. 6 (c)). During the nanowire growth stage, the lengths of nanowires slightly increased from 3 μm and eventually were saturated to 6 μm (Fig. 6 (b)). The extremely small Cu clusters could be also generated by the fragmentation of a portion of nanowire and consumed in this nanowire growth. Finally, nanospheres of 10 ~ 70 nm were formed by the fragmentation of nanowires along with the termination in growth of the nanowires. The Cu clusters were also formed from either fragmentation or dissolution of the nanospheres. The nanowires act as the sources of nanospheres and clusters. These data clearly show that the Cu nanospheres are dissociated from the initial flakes via nanowires formation.
This photo-conversion is also indicated in the absorption spectral changes. Namely, the absorption peaking at λ abs = 380 nm could be assigned to the transverse surface plasmon resonance of Cu nanowires.
To understand the mechanism of one-dimensional Cu nanoparticles under the intense ultrashort light pulses and to compare the results from different starting materials, other starting materials, e.g. Cu microspheres and Cu sheet, were also tested to fabricate nanowires and nanorods under the same experimental procedure (Fig. 7). A similar structural evolution process was found for the microspheres (Fig. 7 (a)), which was also firstly formation of the short nanorods on the surface of microspheres (Fig. 7 (b)), then the.
amount and length of the nanorods increased at the cost of the microspheres and finally the microspheres were reduced in size to the nanospheres with increasing laser irradiation time, as shown in Fig. 7 (c). This means that the nucleation growth mechanism of nanoparticles is the same with another materials evolving in the condition of intense ultrashort light pulses. This process could find wide applications for the preparation of one-dimensional nanoparticles. Indeed, we have also observed the growth of nanowire on the surface of Cu sheet after femtosecond laser irradiation (Fig. 7 (d) ~ (f)). Compared with the typical formation of Cu nanowires of 85 nm diameter from micro-flakes, it was nanorods with the average length and diameter of 1 μm and 280 nm, respectively that formed in the methanol solution. In addition, it seemed that a slightly longer irradiation time (20 min) was needed for the transformation to the nanorods, indicating that higher laser energy was involved in the change to the nanorods from microspheres than that in the conversion from the micro-flakes to nanowires
2.2. Characterization of Cu nanowires
Fig. 8 (b), (c) show the typical SEIs of Cu nanowires, which were synthesized by femtosecond laser with an irradiation time of 5 min. It was clearly demonstrated that nanowires, with an average length of 1.0 μm and an average diameter of 85 nm were formed via simple femtosecond laser irradiation. Comparing with the original Cu micro-flakes, shown in Fig. 8 (a), it is clear that almost all of micro-flakes transferred to the nanowire structure in the methanol after 5 min irradiation, indicating that femtosecond laser irradiation provided a highly effective way for the nanowires formation. Some small nanospheres attached to the edge of nanowires (Fig. 8 (c)) were thought to contribute to the growth of nanowires. Interestingly, many clumps consisting of needle-like nanowires along all directions were found in the images, which implied that one nucleation site was located in the centre and grew up to form the nanowires.
A detailed TEM analysis of Cu nanowires and nanospheres prepared in ethanol has been performed. Fig. 9 shows TEM observations of Cu nanowires after femtosecond laser irradiation for 3 minutes. Schematic diagrams of the analysis methods are also shown. The conventional observations indicate that the nanowires’ surfaces are composed of polycrystalline Cu2O (Fig. 9 (a), (b), (e)). Furthermore, the cross-sectional observations clearly demonstrate that nanowires are partially oxidized from the surface to a depth of about 5 nm (Fig. 9 (d)). On the other hand, the inner part of the nanowires was composed of polycrystalline metallic Cu (Fig. 9 (c), (d), (f)). Indeed, the electron diffraction patterns of the inner and surface parts indicate that the observed areas were composed of metallic Cu (Fig. 9 (e)) and Cu2O (Fig. 9 (f)), respectively.
In order to reveal the chemical state inside Cu nanowires, we carried out chemical state mapping of Cu on the cross-sectional surface by using a field-emission electron probe micro-analyzer (JEOL, JXA-8500F). The chemical state of copper by using EPMA can be determined from the difference in the peak intensity ratio of I Lα /I Lβ X-ray fluorescence spectra of Cu (Kawai et al., 1993). Typical X-ray fluorescence spectra of various copper oxide compounds and the peak intensity ratio of I Lα /I Lβ were shown in Fig. 10 and Table 1, respectively. Based on these relationships, we converted the spectral mapping into the chemical state mapping (Fig. 11). These observations revealed that the main chemical composition inside the nanowires was metallic Cu. The red, orange, and yellow regions in the chemical state mapping indicate the presence of Cu metal, Cu2O, and CuO, respectively (Fig. 11 (b), (c)). Although the copper on the surface parts of nanowire was oxidized to Cu2O and/or CuO, the chemical state inside nanowire was maintained as metal Cu, which corresponds to the foregoing TEM analyses. Based on these observations, we speculate that these Cu nanowires have electrical conducting properties.
|Chemical state of Cu||Peak intensity ratio (ILα /ILβ )|
We have also confirmed that the nanospheres after long time laser irradiation in ethanol are composed of metallic Cu. Fig. 12 shows TEM observations of Cu nanospheres after femtosecond laser irradiation for 20 minutes in ethanol. A1 and A2 arrows in Fig. 12 (a) show the analysis points of the electron energy-loss spectroscopy (EELS). The EELS spectra near the C-K and Cu-L edges indicate that the Cu nanospheres are covered by a carbon layer (Fig. 12 (b), (c)). Results of the existence of the many metallic Cu nanospheres in the case of long time laser irradiation (Fig. 4 (e), (f)) indicate that these carbon layers, which could be produced by the photo-dissociation of the surrounding solvent molecules, act as a passivation layer preventing not only the aggregation and growth but also their oxidation. Through our experiments, it was found that the dispersed solvent was sensitive to the evolution of morphology of micro-flakes under laser irradiation. Methanol and ethanol were found to be suitable for producing the one-dimensional nanoparticles, such as nanowires and nanorods, while only spherical-like nanoparticles with diameters of about 40 nm were fabricated in the acetone and propanol solution. Moreover, there was no distinct change of the Cu nanospheres after storage in the methanol and ethanol solution for one month.
2.3. Morphology control factors of Cu nanoparticle
Fig. 13 shows the absorption spectra of the Cu micro-flakes dispersed in ethanol just after linearly or circularly polarized femtosecond laser irradiation for 5 minutes. The absorption spectra after a subsequent aging treatment at room temperature for 5 days are also shown. In each polarization case, absorptions peaking at 600 nm due to the surface plasmon resonance of Cu nanospheres are observed just after laser irradiation. An apparent difference in absorption spectra was observed after a subsequent aging treatment for 5 days. In the case of the circularly polarization, the maximum absorption was shifted to the short wavelength region (Fig. 13 (b)). This is attributed to the fact that the partial oxidation of Cu nanospheres causes the blue shift (Tilaki et al., 2007). On the other hand, a particular absorption peaking at 380 nm was observed after linearly polarized laser irradiation and subsequent aging treatment (Fig. 13 (a)).
Fig. 14 shows the SEIs of Cu nanoparticles corresponding to the absorption spectra in Fig. 13. In each polarization case, the nanospheres and unreacted starting Cu micro-flakes are observed just after laser irradiation (Fig. 14 (a), (b)), while nanowires can be observed only after linearly polarized laser irradiation and a subsequent aging treatment (Fig. 14 (c)). On the other hand, in the case of circularly polarized laser irradiation, there is no observation of nanowires just after laser irradiation and the subsequent aging treatment (Fig. 14 (b), (d)). These results indicate that the absorption peaking at 380 nm may be due to the surface plasmon resonance of partially-oxidized Cu nanowires. Indeed, it is well-known that the surface plasmon resonance frequencies depend not only on the size but also the shape of particles (Yim et al., 2007).
Why are the nanowires not formed in the case of the circular polarization? In order to understand detailed mechanism of the Cu nanowire formation, we compared experimental results to the finite difference time domain (FDTD) calculation of surface plasmon polariton (SPP) wave propagating along the surface of the Cu nanosphere in an ethanol. Furthermore, the TEM observations of the fragmented Cu nanoparticles just after the linear polarized and the circular polarized femtosecond laser pulse irradiation for 5 minutes in ethanol were performed (Fig. 15 (a), (b)). The ethanol suspended Cu nanoparticle was immediately evaporated at room temperature in order to quench particle growth; therefore, these images may indicate the pre-grown Cu nanoparticle. The differences in the shape of these pre-grown Cu nanoparticles evidently depend on the laser polarization (Fig. 15 (a), (b)). Fig. 15 (c) and (d) show FDTD simulation results of electric field amplitude of the SPP wave excited by the linear polarized and the circular polarized light. In this simulation, we used an input
electric field amplitude of E 0 = 2.4 × 1011 V/m corresponding to the laser energy fluence of 2.4 × 103 J/cm2, a wavelength of 800 nm, and a diameter of Cu nanosphere of 20 nm. In the case of linearly polarized radiation, the localized enhancement of the electric field along the light polarization can be expected (Fig. 15 (c)), while the region of the electric field enhancement rotates according to the circular polarization (Fig. 15 (d)). Such localized enhancement around the Cu nanosphere in the linear polarization induces an anisotropic distribution of the nonlinear photoionization of the Cu nanosphere, resulting in the anisotropic shape modifications.
Stalmashonak and co-workers have also reported that the laser induced shape modification of spherical Ag nanoparticles embedded in soda-lime glass was evoked by the surface plasmon assisted photoelectron emission of the electrons from the metal surface during femtosecond laser irradiation (Stalmashonak et al., 2009). They have also found that such transformation of the metal nanoparticles in glass, which takes place within a timescale of 1 ns, depends on the applied laser pulse intensity, and suggested that the directional memory is defined by the directed emission of hot electrons interacting with the laser field (Unal et al., 2009). Comparing experimental results with simulation results, the shape of the pre-grown Cu nanoparticle produced by the single laser pulse was defined by the laser polarization. Especially, Cu nanoparticles, which act as a nucleation site for Cu nanowire growth, were elongated in the direction parallel to the linear laser polarization. On the other hand, spherical Cu nanoparticles formed by irradiation with the circularly polarized pulses cannot grow into Cu nanowires. We have also confirmed that the one-dimensional growth of Cu nanoparticles occurs during the subsequent aging process.
Fig. 16 shows SEIs of the Cu nanoparticles after 5 minutes of the linearly polarized laser irradiation and subsequent aging treatment at 40 and 60 °C for several hours. These SEM observations reveal the diameter growth rate of Cu nanowires increases with the aging temperature (Fig. 16). In the initial stage during aging process at 40 °C, namely one-dimensional growth of Cu nanoparticles, nanoscale web-like aggregates of nanoparticles were formed (Fig. 16 (a)). This phenomenon is similar to the formation of unusual aggregated structures composed of both crystalline and amorphous silicon nanoparticles by the femtosecond laser ablation in the presence of a background gas (Tull et al., 2006). These nanoscale web-like aggregates are expected to be evolved into Cu nanowires (Fig. 16 (a) ~ (c)). Indeed, the inner part of the Cu nanowires was composed of polycrystalline copper (Fig. 9). Detailed SEM observations indicated the diameter of Cu nanowires was variable as a function of the aging time. After the subsequent aging treatment for 120 hours, the diameters of Cu nanowires were eventually about 68 and 185 nm, and the lengths were about 7.5 and 3.5 μm at the aging temperatures of 40 and 60 °C, respectively. This indicates that the aspect ratio of the Cu nanowires can be controlled by the change of the subsequent aging conditions. As described earlier, the shape of Cu nanoparticles was influenced by the surrounding solution. We have also investigated the effect on the growth of Cu nanoparticles by the surrounding solvent. Although the nanoparticles prepared by 3 minutes of laser irradiation in ethanol were almost wire-like, the fraction of nanospheres increased with the laser irradiation time (Fig. 17 (a), (d)). In contrast, the nanoparticles prepared by the same conditions of laser irradiation in methanol were observed to be cubic nanostructures (Fig. 17 (b)), while nanorods were formed in the case of the long time laser irradiation (20 minutes) (Fig. 17 (e)). Besides, in the case of long time laser irradiation in ethanol and methanol, many nanospheres still exist after a subsequent aging treatment for 5 days (Fig. 17 (d), (e)). However, the one-dimensional growth of the nanoparticles after laser irradiation and a subsequent aging treatment occurred in both the cases of ethanol and methanol but no morphological change was observed in the case of using propanol (Fig. 17 (c), (f)).
We speculated that redox state associated with the hydrogen gas generation is also responsible for determining the shape of nanoparticles. To evaluate this assumption, the generated hydrogen gas was detected during laser irradiation. Fig. 18 indicates the hydrogen gas generation rate during femtosecond laser irradiation as a function of the standard enthalpy change of formation, ΔfH liquid, of various alcohol solutions. Symbols of
2.4. Growth mechanism of Cu nanowires
The following explanation of the Cu nanowires’ growth is proposed: when an intense femtosecond laser pulse is impinging onto the surface of a metal Cu, multiphoton ionization rapidly occurs without significant ablation. The SPP waves, with a large electric field parallel to the surface, are resonantly excited by light and propagate along the surface. The SPP waves could absorb the light wave via inverse Bremsstrahlung heating (Kupersztych et al., 2004) and couple with the incident light wave only if it propagates in the plane of light polarization. This light-plasma coupling occurs at the flake surface over a narrow region with a depth of the order of the skin depth (d p << λ). Previous investigations suggested that either preimposed or self-generated deformations on the solid surface strongly affect laser energy absorption (Feurer et al., 1997). Evidence for small-deformations comes from the wide spreading of the reflected radiation observed in experiments (Zepf et al., 1998). The difference in the light-plasma coupling between linearly and circularly polarized laser pulses should be responsible for different morphology of nanoparticles. Numerical simulations suggested that electron oscillations may grow for a step density profile much faster than the typical time scale of ion motion, leading to an oscillatory “rippling” of the critical surface (n e = n c ) (Macchi et al., 2001). Such rippling is generated as a result of interference between the light field and the SPP wave launched by initial random surface inhomogeneities. Positive feedback leads to exponential growth of the periodic surface structures oriented perpendicular to the light polarization, which become frozen within the material. The nanowires were also generated by the fragmentation of the initial flakes via interference between the light field and the electric field of the surface plasmon-polariton wave. Indeed, the nanowire formation could not be observed using a circularly polarized beam just after the laser irradiation (Fig. 14, 15). We propose the two successive steps of the Cu nanowire growth mechanism:
photofragmentation via interference between the incident light field and the electric field of the SPP wave and
nanoparticle growth in a certain direction.
Small species of copper such as nanoclusters (and/or atoms) are fragmented from the initial Cu micro-flakes. This process corresponding to the nucleation is affected by the interaction between the electric field of the light wave and the electric field of the excited surface plasmon-polariton wave on the nanoscale. In fact, it is well-known that laser-ablation process of metal target under a liquid environment disperses many types of ablated material, which is called the plume, including small clusters, nanoparticles, free atoms, and ions. Such nanoclusters act as a nucleation site or a source for nanoparticle growth and aggregate into nanoparticles with a larger size compared to the nanoclusters. Self-aggregation of the nanoparticles suspended in the solvent should be prevented by hindering direct contact of the nanoparticles due to the interaction of solvent environment molecules and the surface of nanoparticles. Assuming femtosecond-laser irradiation into a continuous agitation of the suspension, the fragmented Cu nanoparticle with a radius of about 20 nm (Fig. 15) can move at least 86 nm within an interpulse time (τ int ) of 1 ms, which was estimated by the Brownian motion of suspended nanoparticles described by the Stokes-Einstein equation:
where κ B , T, η, d are the Boltzmann constant, the temperature, the viscosity of solvent (methanol), and the diameter of the fragmented Cu nanoparticle, respectively. From this estimation, the Cu nanoparticles generated from the initial particles were produced by the single femtosecond laser pulse. Such nucleation site starts growing after the nanoclusters saturate in the solution. Increasing the irradiation time, with more and more nanoclusters being provided at the cost of the micro-flake, the short nanorods on the micro-flake surface keep growing to nanowires. After the micro-flakes are exhausted, no resource can seed the nanoclusters, and then the growth of the nanowires is stopped after consumption of the nanoclusters. At this stage, the laser energy will be absorbed by the nanowires themselves and the nanowires start to fragment into small nanorods and nanospheres due to the drive of the energy of minimization to a more stable morphology (Xiang et al., 2006). After a long irradiation time, nanospheres would become the main product in the methanol solution (Fig. 5 (e), (f)). A similar mechanism of growth was also observed for Se nanowires by using a sonication method (Gates et al., 2002). Although a part of the aggregates are oxidized to Cu2O, most are kept the metallic state due to reducing environment provided by hydrogen gas generation and covering with the amorphous carbon layer which were photo-dissociated from the surrounding solvent molecules. By the competition between the oxidation to Cu2O and the aggregation of metallic Cu nanoclusters, Cu nanoparticles grow one-dimensionally into nanowires which have core-shell structure. In fact, it is well known that in the case of Cu2O crystals, as shown in Fig.19, the O2- ions in the (001) facet are more apt to hydrolyze, compared with those in the (111) facet, and the stacking along (001) directions therefore becomes energetically favorable (Chen et al., 2003).
3. Photo-initiated growth of ZnO nanowire
Zinc oxide (ZnO) nanostructures have attracted immense attention as they offer a wide bandgap and a large exciton binding energy of 3.37 eV and 60 meV, respectively, at room temperature (Ohta & Hosono, 2004) and ultraviolet emission (Huang et al., 2001a; Kong et al., 2001; Lin et al., 2007). Various nanostructures such as nanowires (Huang et al., 2001b; Zhang et al., 2005), nanotubes (Zhang et al., 2002a), nanobelts (X. Y. Zhang et al., 2004), nanohelixes, nanosprings and nanorings (Kong & Wang, 2003), nanonails (Kar et al., 2006) and nanodisks (Lin et al., 2007) have been successfully synthesized for various applications such as nanolaser (Huang et al., 2001a), gas sensor (Sberveglieri et al., 2007), biosensor (F. Zhang et al., 2004), field-effect transistor (Arnold et al., 2003), solar cell (Hosono et al., 2005) and field emission (Lee, et al., 2002). A number of physical and chemical synthesis processes have been studied for the growth of ZnO nanostructures. Some of the physical methods include the thermal evaporation and vapor transport approaches (Huang et al., 2001a; Pan et al., 2001), metal organic vapor-phase epitaxial growth (Park et al., 2002), molecular beam epitaxy (MBE) (Heo et al., 2002) and pulsed laser deposition (PLD) (Zhang et al., 2005), which are generally based on catalysed vapor-liquid-solid growth mechanism (Wagner & Ellis, 1964). In addition, the simple and low-cost chemical aqueous solution processing with hydrothermal treatments have also been thoroughly studied (Le et al., 2005; Yang et al., 2006; Zhang et al., 2002a; Zhang et al., 2002b; H. Zhang et al., 2004; X. Y. Zhang et al., 2004). However, all these processes require either high temperature, low pressure, complex procedures, extended growth period or the use of catalysts that could be embedded on the nanostructure tip, which are unfavourable conditions. In this section, a catalyst-free and surfactant-free synthesis process without the above-mentioned adverse conditions, comprising femtosecond laser irradiation to initiate heterogeneous nucleation in aqueous solutions at room temperature and pressure, followed by hydrothermal treatments at low temperatures for crystal growth, is demonstrated. This work exploits pulsed laser to induce nucleation for ZnO nanostructure growth, compared to previously reported laser processing as ablation tools (Zhang et al., 2005).
3.1. ZnO nanorods and nanotubes formation
0.016 M zinc acetate dihydrate (Zn(CH3COO)2∙2H2O) solution was first prepared at room temperature. 0.095 M ammonium hydroxide (NH4OH) was then added until pH 8 to create an alkaline environment. By using the above mentioned experimental setup (Fig. 1), the mixture solution was then immediately irradiated with femtosecond pulses with a regeneratively amplified mode-locked Er-doped fiber laser (Cyber Laser Inc.), operating at 780 nm wavelength at 1 kHz repetition rate. The laser beam was focused via a microscope objective (Nikon; LU Plan Fluor, 20× 0.40 N.A.) with typical pulse width and pulse energy of 215 fs and 200 μJ/pulse, respectively, into a rectangular quartz vessel of 1 × 1 × 3.5 cm3 filled with the pH 8 solution. The vessel was placed on a magnetic stirrer and the solution continuously stirred to maintain homogeneity. Irradiation was performed for 60 minutes and the solution subsequently transferred into furnaces for heat treatments at 60 °C and 80 °C for 120 minutes before being cooled down to room temperature. The grown ZnO nanorods were analyzed by field emission scanning electron microscopy (JEOL, JSM-6705F) to study their morphologies. Samples were prepared by drop-casting the solutions onto silicon substrates and allowed to evaporate at room temperature. X-ray diffraction (XRD) pattern was collected using Rigaku Rint2500HF to study the crystal structure. Room temperature photoluminescence spectrum was obtained via a Horiba Jobin Yvon FluoroMax-P spectrometer with a Xenon lamp excitation source at 300 nm.
Fig. 20 (a), (b) and (c), (d) show the SEIs for the pH 8 solutions without and with the laser irradiation process, after hydrothermal treatments at 60 °C and 80 °C, respectively, for 120 minutes. The results show that nanorods were grown only for the solution that underwent irradiation and subsequent treatment at 80 °C. This suggests that the femtosecond pulse irradiation plays a major role in the overall nanorod growth and also, the level of hydrothermal temperature is essential, affecting the solubility of the dissolved zinc species contributing to the ZnO crystal growth. This process shows significantly that no catalyst is required to initiate growth and no surfactant is necessary for preferential nanorod growth direction. Nanorods were grown after hydrothermal treatments for 120 minutes, compared to previously reported aqueous-based methods requiring more than 10 hours (Yang et al., 2006; Zhang et al., 2002b; H. Zhang et al., 2004). The precipitate observed in Fig. 20 (c) was filtered from the solution, washed, dried at room temperature and analyzed to study the nanostructures in details.
The result, as seen in Fig. 21 (a), shows that without laser irradiation, random homogeneous nucleation could spontaneously occur, leading to the growth of poor quality rod-like nanostructures with rough and porous surfaces. In comparison, Fig. 21 (b) ~ (d) show that large hexagonal nanorods with flat tops and smooth surfaces, with diameters up to 200 nm and lengths up to 1 μm were grown with laser irradiation, which appears to affect the morphology and quality of the nanostructures. We postulate that the femtosecond pulse irradiation initiates heterogeneous nucleation sites, due to induced solution inhomogeneity, acting as seeds for crystal growth into well-structured nanorods with smooth planes during the subsequent hydrothermal treatments where dehydration of the zinc complexation species to ZnO occurs. In addition, the zinc complexation species ions could interact with the laser light polarization and electric field, leading to specific growth directions. The level of hydrothermal temperature is critical and found to be 80 °C for optimal growth. Lower temperature is insufficient to dehydrate the zinc complexation species and no crystal growth is observed (Fig. 20 (b)) while higher temperature could result in the dissolution of the nanorods, leading to poor crystal quality. In Fig. 21 (a), (b), as indicated by arrows, there were instances where smaller secondary nanostructures grew on top of larger primary structures. This is due to additional ZnO nuclei attaching to the top flat surfaces of the primary nanostructures, thereby enabling secondary growth. Eventually, nanorods with lengths up to 1 μm were produced. Other features could also be observed, where the dashed arrows in Fig. 21 (c) indicate nanotubes while the double-headed arrows in Fig. 21 (d) indicate lateral splits along the nanorod side planes. The top plane of the nanorod is the (0001) polar and metastable crystal plane, which could attract hydroxide ions (OH-). These OH- ions attack the plane and erode the central parts of the nanorods, creating nanotubes. Prolonged exposure to OH- ions could also cause the non-polar and stable crystal side planes to split (Yu et al., 2007).
Fig. 22 (a) shows the XRD pattern for the ZnO nanorods, dispersed on a glass substrate, where the diffraction peaks can be indexed to the ZnO hexagonal structure with a = 3.24 Å and c = 5.19 Å. Fig. 22 (b) shows the idealized (Li et al., 1999) (left) and the actual (right) growth habit observed in this work for the ZnO crystal. For ideal growth, the growth rates of different crystal directions are found to be , with a tip at one end. The appearance of flat tops on the nanorods grown (Fig. 21 (b)) suggests that the growth rate of the (0001) crystal facet is relatively slower than ideal. At pH 8, when NH4OH is added to Zn(CH3COO)2∙2H2O, the Zn(NH3)4 2+ ammine complexation species is mainly formed as it prevented hydroxide formation and precipitation (Fratesi & Roventi, 1992). As the final growth unit for ZnO is Zn(OH)4 2- (Li et al., 1999), Zn(NH3)4 2+ has to first decompose to Zn(OH)4 2- before dehydration to ZnO at elevated temperature. This intermediate decomposition requires heat, thereby suppressing crystal growth. Hence, Zn(NH3)4 2+ acts as a buffer, slowly releasing the Zn(OH)4 2- growth unit prior to ZnO crystal growth. Furthermore, at pH 8, the nanorods tend to disperse and can be individually identified (Fig. 21) due to the weak basicity of Zn(NH3)4 2+ and the presence of NH3 which prevent aggregation.
Fig. 23 shows the room temperature photoluminescence of ZnO nanorods, with a broad UV emission peaking at ~380 nm that is ascribed to the ZnO band-edge emission due to the recombination of free excitons (Huang et al., 2001b). The broad green emission observed beyond ~530 nm is generally accepted as deep-level or trap-state emission due to radiative recombination of a photogenerated hole with an electron occupying the oxygen vacancy (Vanheusden et al., 1996). The broad spectrum and low intensity observed could be due to poor ZnO crystal quality caused by OH- attack, the low intensity Xenon lamp source used or the random alignment of the nanorods, resulting in poor luminescence. In addition, coatings of organic solutions could be present on the nanorod surfaces, affecting the results and possibly giving rise to the peaks seen at 450 ~ 470 nm.
3.2. Role of femtosecond laser irradiation
In order to reveal the role and feasibility of femtosecond laser irradiation in the overall ZnO nanostructure growth process, we performed the following experiments by changing the femtosecond laser irradiation condition (pulse width, pulse energy, and pulse repetition rate). Aqueous mixture solutions of 0.016 M Zn(CH3COO)2∙2H2O (zinc acetate dihydrate) and 0.20 M NH4OH (ammonium hydroxide) at a pH value of 11 alkaline environment were initially prepared and then immediately subjected to laser irradiation. The laser radiation in Gaussian mode produced by a regenerative amplified mode-locked Ti:sapphire laser (Coherent Inc., 100 fs pulse duration, 1 kHz or 250 kHz repetition rate) operating at a wavelength of 800 nm was focused by a microscope objective (Nikon; LU Plan Fluor, 20× 0.40 N.A.) to a spot size of ~ 2.4 μm diameter inside a rectangular quartz vessel of 1 × 1 × 3.5 cm3 filled with the pH 11 precursor solution. The quartz vessel was placed on a magnetic stirrer and continuously stirred to maintain homogeneity. Tables 2 shows the laser irradiation conditions for the femtosecond pulse irradiation performed at the repetition rates of 1 kHz and 250 kHz. Immediately after the laser irradiation, the solutions were placed in ovens for hydrothermal treatments at 80 °C and 100 °C for 120 minutes for the ZnO crystal growth, before being cooled down to room temperature.
|Sample||Pulse repetition rate|
Fig. 24 shows the SEIs of the nanostructures grown without and with laser irradiation (pulse energy, E pulse 200 μJ/pulse, irradiation duration 120 minutes, total irradiation energy E T 1.44 kJ and repetition rate 1 kHz), followed by hydrothermal treatments at 80 °C for 120 minutes. Without the laser irradiation, large flower-like structures comprising aggregated cone-shaped nanostructures can be observed, where their surfaces appear to be rough and porous. When the irradiation process was performed as one part of the synthesis process, ZnO nanowire growth can be observed. This shows that the ultrashort pulse laser irradiation plays a significant role in initiating heterogeneous nucleation sites for nanowire growth, compared to the random homogeneous nucleation process associated with hydrothermal growth without the laser irradiation. In a separate investigation, longer pulses of 860 fs were used for the laser irradiation process, instead of the 100 fs pulses used in this work in order to study the significance of ultrashort pulses for nanowire growth. For the study involving 860 fs (results not shown), short cone-tip nanostructures similar to Fig. 24 (b) were grown instead of thin elongated nanowires that were observed in this work.
The generally accepted reaction routes in aqueous ammonia solution at a high alkaline level are defined as below :
Upon mixing the starting reagents, the soluble complexation species of Zn(OH)4 2- were formed in a high alkaline environment (pH 11 in this work). Nucleation sites were then produced in the solutions via either random homogeneous nucleation or heterogeneous nucleation by the laser irradiation, which subsequently became aggregated. Acting as the growth units, Zn(OH)4 2- were then dehydrated to ZnO at elevated temperatures to construct the ZnO nanostructures. Fig. 25 summarizes the illustrated scheme for ZnO crystal growth. We postulate that the ultrashort pulse laser irradiation initiates heterogeneous nucleation seeds for subsequent crystal growth during the hydrothermal treatments. During the laser irradiation process, slight increased in solution turbidity was progressively observed. As similar but increased solution turbidity was observed after the solution was subjected to hydrothermal treatment which caused the dehydration of the zinc complexation species to ZnO, it can be inferred that the ultrashort pulses acted in a similar way to form ZnO heterogeneous nucleation seeds during the laser irradiation process. As seen in Fig. 24, these ZnO heterogeneous seeds lead to nanowire growth compared to homogeneous nucleation seeds. Hence, the laser-induced nucleation affects the morphology of the ZnO nanostructure grown. In addition, it might be possible that the complexation species interact with the light polarization and the electric field of the laser, which then direct the nanowire growth in a certain direction.
Fig. 26 shows the ZnO nanowires grown from the hydrothermal treatments at 80 °C and 100 °C for 120 minutes, after the laser irradiation at E T = 0.36 kJ and at 1 kHz repetition rate for different pulse energies (E pulse ) and irradiation durations while Fig. 27 shows the results for irradiation at E T = 0.72 kJ. Fig. 28 shows the ZnO nanowires after the laser irradiation at E T = 1.44 kJ at E pulse = 200 μJ/pulse for 120 minutes (Fig. 28 (a), (b)) and E pulse = 400 μJ/pulse for 60 minutes (Fig. 28 (c), (d)) and at E T = 2.16 kJ and E pulse = 300 μJ/pulse for 120 minutes (Fig. 28 (e), (f)), followed by hydrothermal treatments at 80 °C and 100 °C for 120 minutes, respectively. From Fig. 26, we can observe that the pulse energy levels and the hydrothermal treatment temperatures have significant effects on the growth of ZnO nanowires. By comparing the results after hydrothermal treatments at 80 °C and 100 °C for 120 minutes, there are higher density of nanowires grown at 100 °C, which are also more well-defined and longer. This is due to the higher solubility and mobility of the zinc growth units at a higher temperature, which encourages crystal growth. Analysing the results based on pulse energy, well-defined individual nanowires can be observed in the samples after the laser irradiation at E pulse = 200 μJ and 400 μJ, where they are arranged in a flower-like configuration aggregated at one end. This shows that a minimum of E pulse = 200 μJ pulse energy is required to initiate heterogeneous nucleation for nanowire growth. Typical nanowire structure can be observed after the laser irradiation at E pulse = 50 μJ but individual nanowires were not clearly separated from each other. As indicated by arrows in Fig. 26 and the subsequent figures, there are instances where individual nanowires fused together, possibly via oriented attachment (Pacholski et al., 2002), into larger nanowires. In Fig. 27, the results show that irradiation with E pulse = 100 μJ produced ZnO nanowires where individual nanowires could not be well-defined, as in the case for E pulse = 50 μJ in Fig. 26 (b). Increasing E pulse , while maintaining E T = 0.72 kJ by reducing the irradiation duration, produced well-defined and individually separated nanowires. This again shows the critical role of E pulse in initiating conducive nucleation sites for crystal growth. In Fig. 28, individual nanowires can be clearly observed for all samples as the threshold E pulse = 200 μJ required to initiate heterogeneous nucleation for favourable growth into ZnO nanowires was exceeded.
From the results in Fig. 26 ~ 28, it can be inferred that the pulse energy level, E pulse , is essential in initiating heterogeneous nucleation while the irradiation duration (and indirectly the total irradiation energy, E T ) determines the density of nanowire growth due to the abundant nucleation sites initiated. In addition, these two factors together with the hydrothermal treatment temperature affect the morphology of the nanowires as the rates of dehydration and consumption of the zinc growth units for numerous crystal growth processes would be affected. Generally, based on the amount of zinc growth units available in the quartz vessel and depending on the laser processing parameters and hydrothermal treatments conditions, ZnO nanowires with diameter 10 ~ 40 nm and length up to 1.5 μm can be grown.
Fig. 29 summarises the results on ZnO nanowire synthesis after the laser irradiation at E T = 1.44 kJ at 250 kHz repetition rate for different E pulse values and irradiation durations, followed by hydrothermal treatments at 80 °C and 100 °C for 120 minutes. For irradiation at E pulse = 0.8 μJ/pulse for 120 minutes, the ZnO nanostructures appear to have grown from random homogeneous nucleation similar to that of Fig. 24 (a), (b), suggesting that the low pulse energy had not been effective in initiating heterogeneous nucleation. For the laser irradiation at E pulse = 1.6 μJ/pulse for 60 minutes, the general shape of nanowires can be observed while the results for E pulse = 2.4 μJ/pulse for 40 minutes clearly show aggregations of ZnO nanowires. The difference in morphology and density of nanowires is most likely caused by the abundant nucleation seeds initiated by the higher pulse energy. At E pulse = 1.6 μJ/pulse, the soluble Zn(OH)4 2- species were consumed by less nucleation seeds, leading
to large nanowire dimensions. However, at E pulse = 2.4 μJ/pulse, the Zn(OH)4 2- growth units were consumed and distributed over a larger amount of nucleation sites for growth and thus reducing the dimensions of the nanowires grown. In addition, at E pulse = 1.6 μJ/pulse, the pulse energy might be too low to initiate stable heterogeneous nucleation so that two or more nucleation seeds might have merged to produce a larger seed and consequently the larger dimensions for the nanowires. This is highly likely due to the high frequency of pulses which initiate high amount of nucleation seeds in close proximity, thus enabling neighbouring seeds to merge to improve stability. This dependence of nanowire diameter on the size of the nucleation seeds is similar to the correlation between nanowire diameter and gold nanodot size, which acted as catalysts, observed by Fan et. al. (Fan et al., 2006). The high number of nucleation seeds near each other also enables them to produce large aggregations of nanowires, as seen in Fig. 29 (f), compared to those grown from 1 kHz repetition rate laser irradiation. Finally, homogeneous nucleation might have also played a part in the growth process for Fig. 29 (c), (d), which has been shown to produce larger nanostructures compared to that grown purely from heterogeneous nucleation. These large differences in the results in Fig. 29 suggest that there is a certain threshold level for E pulse to initiate nucleation sites in the solution. The results also agree with the results obtained via laser irradiation at the repetition rate of 1 kHz, which showed the significant role of E pulse . In addition, comparing the results of Fig. 26 ~ 29 show that well-defined nanowires could be grown beyond E pulse = 2.4 μJ/pulse at 250 kHz repetition rate. However, at 1 kHz repetition rate, E pulse beyond 200 μJ/pulse was required for well-defined ZnO nanowire growth. This difference in E pulse level could be attributed to the thermal effect induced by irradiation at 250 kHz repetition rate.
Fig. 30 shows the illustrated temperature profiles expected in the sample solutions irradiated by 1 kHz and 250 kHz ultrashort pulses. The thermal diffusion time was estimated to be approximately 3 μs using the thermal diffusion equation (Chartier & Bialkowski 1997), based on a laser spot radius of 1.22 μm estimated by the Airy disc approximation (Vogel et al., 2005) and the heat diffusivity of water = 1.38 × 10-7 m2 s-1 (Vogel et al., 2005). As seen in Fig. 30, the diffusion time is of the same order as the pulse interval for the 250 kHz pulses and hence thermal effect would be expected to build up within the localised laser spot volume and possibly contribute to the heterogeneous nucleation process. In contrast for the train of 1 kHz pulses, thermal effect was negligible as the heat induced by a single pulse would have completely dissipated by the time the next pulse arrived 1 ms later.
3.3. Formation mechanisms of ZnO nanorods
In order to reveal the formation mechanisms of ZnO nanorods with the additional femtosecond laser irradiation process, we performed the following experiments. Aqueous mixture solutions of 0.02 M ZnCl2 and 0.032 ~ 0.200 M NH4OH at pH value ranging from 8.5 to 10.5 alkaline environments were initially prepared. We used femtosecond laser pulses to focus in a liquid cell and efficiently transfer energy into the precursor solution. The initial solution exhibits a slightly white turbidity decreasing with a pH increase because the formation of zinc ammonia complex, which was subjected to femtosecond pulse irradiation at room temperature. The laser radiation in Gaussian mode produced by a regenerative amplified mode-locked Er-doped fiber laser (Cyber laser Inc., 230 fs pulse duration, 1kHz repetition rate, pulse energy 0.5 mJ/pulse) operating at a wavelength of 780 nm was focused by an objective (Nikon; LU Plan Fluor, 20× 0.40 N.A.) into a rectangular quartz vessel of 1 × 1 × 3.5 cm3 filled with the precursor solution, which was placed on a magnetic stirrer and continuously stirred to maintain homogeneity. Irradiation was performed for 60 min and the solution subsequently transferred into furnace for heat treatments at 80 °C or 100 °C for 120 minutes before being cooled down to room temperature. Samples were prepared by drop-casting the solutions onto silicon substrates and allowed to evaporate at room temperature. The grown ZnO particles were analyzed by field emission scanning electron microscopy (JEOL JSM-6705F) to study their morphologies. X-ray diffraction (XRD) pattern was collected using Rigaku Rint2500HF to study the crystal structure.
It is well known that the solid phase stability of Zn(OH)2 in the precursor solution has been determined by the pH value and the concentration of Zn(II) soluble species (Yamabi & Imai, 2002). Fig. 31 shows the phase stability diagrams for the Zn(OH)2−NH3 systems at 25 °C. The dashed lines indicate the thermodynamic equilibrium between the various Zn(II) soluble species, which are calculated by the following equilibrium equations (6) ~ (8).
Values of standard thermodynamic data and stability constant are taken from the literature (Goux et al., 2005; Hubert et al., 2007; Peulon & Lincot, 1998; Yamabi & Imai, 2002). The red solid line represents the boundary of the solubility of the solid Zn(OH)2. This diagram reveals that the solid Zn(OH)2 is thermodynamically stable at a pH value ranging from 7 to 12 in the precursor solutions ([Zn2+] = 0.02 M). Typical three precursor solutions with different pH values of 8.5, 9.5, and 10.5 were prepared in the present study.
Fig. 32 indicates XRD patterns of precipitates from mixed precursor solutions of ZnCl2 and NH4OH at pH 8.5, 9.5, and 10.5 without (a) and with (b) the femtosecond laser irradiation for 60 minutes and the successive thermal treatment at 80 °C for 120 minutes. The JCPDS standards of Zn5(OH)8Cl2 H2O, Zn(OH)2, and ZnO are also shown in Fig. 32.
The corresponding SEIs are shown in Fig. 33. No apparent diffraction peaks of ZnO were observed in the case of the thermal precipitates from precursor solutions at every pH condition without the laser irradiation (Fig. 32 (a)). These patterns were assigned to Zn(OH)2 and Zn5(OH)8Cl2 H2O, suggesting that the precursor solutions could not become supersaturated at 80 °C with respect to the homogeneous ZnO nucleation. On the other hand, the apparent diffraction peaks attributed to ZnO were observed in the samples which the laser irradiation process was applied before thermal treatment at the same temperature (Fig. 32 (b)). This indicates that the photo-initiated heterogeneous nucleation could be induced by the femtosecond laser irradiation in the precursor solutions at room temperature. The SEIs in Fig. 33 evidently indicate that in contrast to the formation of the scale-like or amorphous precipitate after the thermal treatment at 80 °C for 120 minutes, the ZnO hexagonal nanorods with a diameter of 40 ~ 80nm, which slightly decreases with a pH increase, were obtained by applying the laser irradiation (Fig. 33 (d) ~ (f)). Based on these results, we speculated that the nucleations of ZnO nanorods were initiated by the local supersaturation via femtosecond laser pulse irradiation even at the room temperature.
In order to reveal the nucleation and growth mechanisms of ZnO nanorods with the additional femtosecond laser irradiation process, the heating temperature during the successive thermal treatment was changed to 100 °C. ZnO nanoparticles were precipitated with or without laser irradiation (Fig. 34). The shape of ZnO nanoparticles was changed from nanorods to flower-like with increasing the pH in the thermal treatment (Fig. 34 (a), (c)). In addition, the smaller ZnO nanoparticles resulting from secondary nucleation were observed at the pH of 9.5 (Fig. 34 (b)). It is well known that the flower-like ZnO nanostructures are formed via twinned ZnO nuclei along the planes in the system with higher supersaturation of Zn(OH)4 2- (Zhang & Mu, 2007). On the other hand, much smaller ZnO nanorods with a diameter of 20 ~ 70 nm were formed by the additional laser irradiation compared to that of the thermal process regardless of pH, about 4 times thinner nanorods were especially formed at the pH of 8.5 (Fig. 34 (a), (d)). The size of ZnO nanorods obtained by the laser irradiation increased with an increase in pH (Fig. 34 (d) ~ (f)). Comparison of the effect of temperature on the size of ZnO nanorods between 80 °C and 100 °C with the laser irradiation indicates that the size decreases with increasing temperature at pH of 8.5 and 9.5, although the size increases with increasing temperature at pH of 10.5 (Fig. 33 (d) ~ (f), Fig. 34 (d) ~ (f)).
Based on the difference in the shape and size of ZnO nanostructures with and without the femtosecond laser irradiation before the subsequent thermal treatment, we deduce the formation mechanism of ZnO nanorods below. The possible reactions in our experiments can be summarized in the following equations (9) ~ (12).
In the experimental pH region, we could consider the soluble species of the uncomplexed Zn2+ ions and the zinc-ammonia complex ions of Zn(NH3)4 2+ at a much higher concentration of NH3. In addition, the insoluble compounds of Zn(OH)2 can be formed in this system. The calculated concentration of Zn(NH3)4 2+ and solid Zn(OH)2 in mixed precursor solutions are shown in Table 3.
|Diameter [nm]||Length [nm]|
|8.5||8.1 × 10-7||9.1 × 10-3||38 (7) a||164 (5) a|
|9.5||7.8 × 10-5||9.1 × 10-3||44 (5) a||233 (3) a|
|10.5||2.4 × 10-3||6.6 × 10-3||68 (24) a||515 (46) a|
|a The numbers in the parenthesis show the standard deviation for 20 samples.|
The size of obtained ZnO nanorods produced by the femtosecond laser irradiation at 0.5 mJ for 60 minutes and the successive thermal treatment at 100 °C for 120 minutes are also shown. The amount of such precipitation depends on the pH and the concentration of NH3 in the solution based on the solubility of Zn(OH)2 and the dissociation constants of Zn(NH3)4 2+. At the pH of 8.5 and 9.5, the reaction of Eq. (9) is dominant and the equilibrium moves to right, namely the nuclei of ZnO are predominantly formed from Zn(OH)2 by the laser irradiation. In contrast, a large amount of the soluble complexes ions of Zn(NH3)4 2+ in addition to the precipitation could be consumed by the formation of ZnO nuclei during laser irradiation, because the reaction of Eq. (11) is dominant at the pH of 10.5. Indeed, the energy absorption by the focusing of femtosecond laser pulses was almost same of 66 % regardless of the pH, although the scattered light intensities at the pH of 8.5 and 9.5 was about 2.5 times higher than that at the pH of 10.5. During the subsequent thermal treatment after the laser irradiation, ZnO nuclei formed by the different reaction path grow into ZnO nanorods along the c-axis direction (Baruah & Dutta, 2009). In lower pH solution (pH 8.5), the smaller ZnO nanorods are formed by the secondary nucleation and growth during the hydrothermal process (Fig. 34 (d)). On the other hand, the larger nanorods could be obtained because ZnO nuclei formed by the laser irradiation grow dominantly during the thermal treatment. It is noted that the standard Gibbs free energy changes of Eq. (10) and (12) are -3.94 and -47.2 kJ/mol, respectively.
To discuss the dynamics of ZnO nuclei formation during the femtosecond laser irradiation, we measured the evolution of spectral extinction of the precursor solutions during the femtosecond laser irradiation (Fig. 35). The transmitted visible light was detected by a photonic multi- channel analyzer (Hamamatsu Photonics, PMA-11). The components of the extinction in Fig. 35 include the sum of light scattering and absorption by ZnO nuclei formed by the laser irradiation. Assuming that the visible absorption of ZnO is negligible, we could estimate the dynamics of the photo-initiated nucleation process based on the Rayleigh scattering theory. In the Rayleigh scattering regime, the scattered light intensity is inversely proportional to the fourth power of wavelength, indicating the shorter wavelength will scatter more than the longer wavelength. While the scattered light intensity in the lower pH solution was substantially constant (Fig. 35 (a)), in the higher pH solution, the scattered light intensity in the shorter wavelength region increases with an increasing in laser irradiation time (Fig. 35 (c)). The results clearly indicate that ZnO nuclei are produced from the liquid phase, i.e. Eq. (11) and (12), in the higher pH solution. On the other hand, the scattering light intensity does not change dramatically because solid Zn(OH)2 already exists in the lower pH solution (Eq. (9) and (10)). Finally, ZnO nuclei produced through different reaction pathways grow into ZnO nanorods during the successive thermal treatment even in the higher pH solution.
In order to understand the origin of the observed phenomenon, the following explanation of the heating mechanism is proposed. Since the light intensity in the focus of the beam is of 1016 W/cm2, the plasma is produced by multiphoton ionization in the focal volume. Once a high free electron density is produced by multiphoton ionization, the material has the properties of plasma and will absorb the laser energy via absorption mechanism of inverse Bremstrahlung heating. Assuming that the electron temperature is proportional to the pulse energy, the electron temperature can be roughly estimated by a simple formula: Q = C e VT e ,where Q (= ηE) is the absorbed energy, η is the absorption coefficient, E is the pulse energy, V is the volume interaction, T e is the electron temperature, and C e is the electron heat capacity. Within the free electron gas model, The electron heat capacity can be approximately calculated by C e (T e ) = γT e , where γ = π2 neκ B 2/2I E (Lin et al., 2008), where n e is the electron number density, κ B is the Boltzmann constant, I E is the ionization potential. The electron temperature is estimated to be 1.3 keV (~ 1.5 × 107 K) by using the parameters in Table 4, which corresponds to the experimental results (Hatanaka et al., 2004).
|Density, ρ||1.0 × 103||[kg/m3]|
|Molar weight, M||1.8 × 10-2||[kg/mol]|
|Ionization potential, IE||6.5||[eV]|
|Laser wavelength, λ||7.8 × 10-7||[m]|
|Pulse width, τ p||2.3 × 10-13||[s]|
|Pulse energy, E||5.0 × 10-4||[J]|
|Absorption coefficient, η||0.2||-|
|Electron density, ne||1.0 × 1020||[cm-3]|
|Numerical aperture, NA||0.45||-|
Based on this calculation, not only the optical breakdown, bubble formation, but also the dissociation of the precursor solution could occur within the focal volume during the femtosecond laser irradiation. Such very high electron temperature decreases with an increase of the lattice temperature, then it reaches to the same temperature as lattice temperature with a time scale of several picoseconds. Assuming that the initial temperature of focal volume reaches ΔT 0 = 3000 K after the electron-phonon coupling, the thermal diffusivity can be calculated by the following equation. For simplicity, we used the thermal diffusivity coefficient of distill water (D th = 1.43 × 10-7 m2/s).
where ΔT 0 is the initial temperature just after the femtosecond single pulse irradiation, w 0 is the laser beam waist, t is the time after the irradiation, and r is the distance from the focus. Fig. 36 indicates the calculated temperature distributions just after the femtosecond single
pulse irradiation as a function of the distance from focus. In this calculation, the time after the laser irradiation were changed from 0 s to 100 μs. Since the repetition rate of 1 kHz, i.e. the interpulse time of 1 ms in the experiments, these calculations apparently indicate that the heat induced by the first pulse can diffuse away from the focal region before the arrival of the successive pulse. Indeed, no apparent temperature change occurred after the femtosecond laser irradiation for 60 minutes. The ZnO nucleation induced by the femtosecond laser irradiation could occur at the instantaneous high-temperature region surrounding the focal volume in precursor solution at room temperature.
In conclusion, we have successfully demonstrated the morphological control of Cu nanoparticles by femtosecond laser irradiation in an alcohol solution. Cu nanowires with a core-shell structure are formed depending on the surrounding solvent and laser irradiation time. The aspect ratio of the Cu nanowires can be controlled by a change of the subsequent aging conditions. The evolution of Cu nanowires from micro-flakes in the intense ultrashort light pulses, it was observed that first nanorods appeared on the micro-flake surface, then nanowires at the cost of initial microparticles and finally these were converted to nanospheres. The mechanism of nanowires’ growth was suggested to be a nuclear growth mechanism, that is Cu atoms and/or nanoclusters, produced by the femtosecond laser irradiation, then grew into nanorods and finally nanowires. The uniaxial growth mechanisms of Cu nanowires are interpreted in terms of a competition between the oxidation to Cu2O and the aggregation of metallic Cu nanospheres. Another interesting phenomenon is the polarization-dependence of the formation of Cu nanowires, resulting from interference between the incident light field and the SPP wave. Apart from the fundamental importance of the observed phenomenon, as the first evidence of photofragmentation of Cu nanowires fabricated by the “top-down” approach, could be useful for an optical polarization control medium, an electro-conductive nanomaterial, and a probe for SPM.
Furthermore, ZnO hexagonal nanorods and nanowires were successfully synthesised from heterogenous nucleation initiated by femtosecond laser irradiation in aqueous solutions with subsequent hydrothermal treatments, without catalyst and surfactant. The laser irradiation process improves the morphology and quality of the nanorod crystal compared to porous rod-like structures when no irradiation was performed. Due to the localized high supersaturation of precursor solution, the size of the obtained hexagonal ZnO nanorods with femtosecond laser irradiation and the subsequent thermal treatment is about 4 times thinner than that obtained by the thermal treatment. The nanorods have diameters up to 200 nm and lengths up to 1 μm, with flat tops due to the slow decomposition of Zn(NH3)4 2+ to Zn(OH)4 2- stunting the crystal growth rates, before dehydration to ZnO. Exposure to hydroxide ions causes erosion of the nanorods, resulting in nanotubes and splits on the side planes. Studies involving various pulse energy levels and total irradiation energy at 1 kHz and 250 kHz repetition rates show that there exists the critical threshold pulse energy to induce heterogeneous nucleation favourable for growth into well-defined and individually separated nanowires. Furthermore, studies involving pH variation indicate that ZnO nucleus produced through different reaction pathways according to the pH value of the precursor solution. The size of the obtained hexagonal ZnO nanorods is variable according to the pH of the precursor solution. The ZnO nanorods and nanowires exhibit a broad UV emission peaking at 385 nm, with green emission due to defect states.
The authors would like to thank Jianrong Qiu from Zhejiang University and Peter G. Kazansky from University of Southampton for helpful discussions. This research is partially supported by Grant-in-Aid for Scientific Research (B), Nippon Sheet Glass Foundation for Materials Science and Engineering (NSG Foundation), and MURATA Science Foundation and Engineering.