Ultrashort Pulsed Laser Ablation of Magnesium Diboride: Plasma Characterization and Thin Films Deposition

AMgB 2 target has been ablated byNd:glass laser with a pulse duration of 250 fs.Theplasma produced by the laser-target interaction, showing two temporal separated emissions, has been characterized by time and space resolved optical emission spectroscopy and ICCD fast imaging. The films, deposited on silicon substrates and formed by the coalescence of particles with nanometric size, have been analyzed by scanning electron microscopy, atomic force microscopy, X-ray photoelectron spectroscopy, micro-Raman spectroscopy, and X-ray diffraction.The first steps of the films growth have been studied by Transmission ElectronMicroscopy.The films deposition has been studied by varying the substrate temperature from 25 to 500C and the best results have been obtained at room temperature.


Introduction
Metal borides films, which show peculiar characteristics at nanoscale level, are currently used in a large number of applications, in particular as corrosion resistant coatings and thermal and diffusion barriers [1].Among metal diborides, magnesium diboride has attracted great interest due to its superconductivity at about 40 K [2].In fact, since MgB 2 is the nonoxide material with the highest transition temperature, the deposits of this material in the form of thin films can be very important for many electronic applications.In particular, magnesium diboride thin films have been investigated for applications in high-field magnets, radio frequency cavities used in accelerators, Josephson junctions, and digital circuits [3][4][5] even if all the most probable applications remain at the moment niche applications and only in the future could this material have good chances to be used in a wider market [6] and many different techniques have been used to produce thin films of this material [3].In general, one of the most useful techniques to deposit thin films of materials with technological interest, such as metal borides, is Pulsed Laser Deposition (PLD) [7,8].PLD has been already used to deposit magnesium diboride thin films but, even if some works concern films produced directly by laser ablation [9][10][11], in many cases PLD was only the first step of a procedure including ex situ or in situ annealing of the produced deposits, often in Mg rich atmosphere [12][13][14][15][16][17][18][19][20][21][22].The necessity of a postannealing is not surprising since all the cited ablation-deposition works were performed by using lasers with pulse durations of the order of nanoseconds.In ns PLD the film formation depends strictly on the condensation of atomic species, in this case Mg and B, on the substrate, followed by their surface diffusion and nucleation.The sticking coefficient of magnesium is in general very low, especially if the substrate temperature is higher than 200 ∘ C, implying that there is only a small probability that it could react with boron to form MgB 2 on the surface [11].The best way to enhance this probability is to increase the flux of Mg on the surface.
When the PLD is carried out by a laser with pulse duration lower than 1 ps (ultrashort pulse) the mechanism of films formation is very different.In fact, depending on the ablation mechanism, the main constituents of the laser produced plasma, often called plume, are a large number of particles with nanometric size and these nanoparticles (NPs) give the larger contribution to the formation of the deposited films [8,23,24].The origin of NPs is still a matter of debate but there is a growing consensus about their origin directly from target fragmentation and ejection of molten material [8,[25][26][27][28][29] even if other hypotheses, such as gas phase reaction, have been also proposed [30].So, if we consider that in ultrashort PLD condensation from gas phase atomic species is probably only a secondary process in the film formation, the application of this technique to the deposition of magnesium diboride thin films could be useful to minimize the problems related to the low sticking coefficient of Mg.
In this work, with the aim of both verifying ablationdeposition mechanisms already proposed [28,31,32] and studying the effect of the substrate temperature on the thin films composition, a target of commercial MgB 2 has been ablated in vacuum by Nd:glass laser with a pulse duration of 250 fs.Considering that no data were present in literature about either the plasma produced by fs ablation of magnesium diboride or the films deposited in that way, we have used the techniques already utilized for other systems [28,31] to characterize the ablation and deposition of MgB 2 .The plasma produced by the laser-target interaction has been characterized by time and space resolved optical emission spectroscopy and ICCD fast imaging.The films, deposited on silicon substrates, have been analyzed by scanning electron microscopy, atomic force microscopy, X-ray photoelectron spectroscopy, micro-Raman spectroscopy, and X-ray diffraction.The size distribution of the particles forming the films has been studied by Transmission Electron Microscopy.The influence of the deposition parameters on films characteristics has been studied by varying the substrate temperature from 25 to 600 ∘ C. The results have been explained considering an ablation-deposition mechanism involving melted NPs directly ejected from the target.

Experimental
Ablation and deposition processes were performed by means of an experimental apparatus, already described [32], consisting of a stainless steel vacuum chamber evacuated by a scrollturbomolecular pumping system to a pressure of 1.5×10 −4 Pa.The chamber was equipped with a support for the target, rotating with a speed of 1 rpm to minimize the craterization effects, quartz windows for the laser beam inlet and the plasma emission detection, and a resistively heated substrate holder.Cold pressed pellets from commercial MgB 2 powder (Aldrich) were used as ablation targets.The substrates ((100) oriented silicon wafers from Aldrich) were kept in front of the target at a distance of 2 cm.The deposition substrates were Formvar carbon film coated copper grids only in the case of samples for transmission microscopy analyses.The ablation laser source was a frequency doubled Nd:glass laser with the following characteristics:  = 527 nm,  = 3.0 mJ, repetition rate = 10 Hz, and pulse duration = 250 fs.The laser beam, focused by a suitable optical system, impinged on the target surface with an angle of 45 ∘ and the spot area was 0.1 mm 2 .The deposition time was fixed at 2 hours and the substrate temperature was varied between 25 and 500 ∘ C. The deposition time was 3 min only in the case of films deposited for Transmission Electron Microscopy.
Optical emission spectroscopy (OES) was carried out by a Princeton ICCD device (1024 × 1024 pixels) coupled with a 50 cm spectrograph (ARC SP-500i).The width of the entrance spectrograph slit was 80 m and the gratings employed were 2400, 1200, and 150 grooves/mm.The gated system had a time resolution of 2 ns and each acquisition was integrated over 50 laser shots in order to increase the signalto-noise ratio.Varying the position of the optical elements by a micrometric translation stage, it has been possible to obtain space resolved emission spectra at different distances from the ablated target surface.The same ICCD system, equipped with 105/35 mm quartz Nikkor lens, was used for acquiring gated images of the lateral view of the overall plasma plume emission in the range of 200-800 nm.The acquisition was performed by accumulating up to 20 shots and the spatial resolution achieved by this system was 37 m.Both ICCD fast imaging and time resolved spectra were accomplished by delaying the data acquisition of the plasma emission in the range of 10-100,000 ns with respect to the laser pulse.In all cases every data acquisition was carried out on a not previously irradiated surface.
The structure of the films was evaluated by X-ray diffraction (XRD) and by micro-Raman spectroscopy.The XRD measurements were performed by a Philips X'Pert Pro diffractometer, using Cu K radiation.The Raman spectra were recorded in a backscattered configuration using a HORIBA LABRAM HR 800 micro-Raman spectrometer, with 632.8 nm radiation from He-Ne laser source.The deposits' morphology was studied by atomic force microscopy (AFM-Park XE 120) and by a scanning electron microscopy (SEM) apparatus (Philips-FEI ESEM XL30).High resolution Transmission Electron Microscopy (HR-TEM, Fei-TECNAI G2 20 TWIN), operating at 200 kV was used to characterize size and distributions of the NPs forming the deposited films.X-ray photoelectron spectra (XPS), acquired by LH-Leybold 100 spectrometer using unmonochromatized Mg K radiation, were used to analyze the films composition and stoichiometry.Wide and detailed spectra were collected in Fixed Analyzer Transmission (FAT) mode with pass energy of 50 eV and channel widths of 1.0 and 0.1 eV, respectively.The resolution of the detailed spectra is 1 eV.

Results and Discussion
3.1.Plasma Characterization.The ICCD fast imaging evidences the presence of a plasma plume starting at the end of laser pulse and lasting few hundreds of nanoseconds.This plume is characterized by a front velocity of 4.5 ± 0.5 × 10 6 cms −1 and an anisotropic angular distribution parameter  = 1.9. is the cosine exponent in the expression () =  0 cos  , where () is the flux intensity along a direction forming an angle  with the normal to the target surface and  0 is the intensity corresponding to  = 0.
OES of the same plume shows many spectral lines assigned to neutral and singly ionized magnesium and boron (Mg I, Mg II, B I, and B II). From the optical emission spectra the time of flight curves for the different species, at Due to the blazing of our gratings, the intensity of B I peaks was very low and so it was impossible to obtain time of flight curves for this species.Comparing these results with those of the plasmas obtained by ns PLD in Ar atmosphere [10,33], it is evident that the atomic species present in both plasmas are the same except for the absence, in the case of fs ablation, of Mg III, found only by D' Amico et al. in the plasma produced by ns PLD performed by a doubled Nd:YAG laser [10].This absence is probably related to the lack of interaction between the expanding plume and the laser radiation and to the fact that our experiments have been carried out in vacuum instead of a buffer gas.A comparison of the velocities of the species in ns and fs regimes shows results of the same order of magnitude when we consider a low buffer gas pressure (10 −2 and 0.5 Pa) for the ns experiments.OES data were used for estimating the Mg I excitation temperature (  ), assuming that the local thermodynamic equilibrium (LTE) condition was fulfilled.Different spectral series, obtained by a fixed detection gate of 25 ns, and starting at delays of 75 ns from the laser pulse, respectively, were accomplished.  was determined by using the Boltzmann plot method using the following equation: where  is the wavelength of the transition line,  is its intensity,  is the transition probability,  2 is the energy of the upper state [34], and  B is the Boltzmann constant.The results reported in Figure 2(a) show that in the ultrashort laser induced plasma the excitation temperature reaches the value of about 6500 K during the first nanoseconds and then drops down reaching a plateau temperature of about 4500 K in about 150 ns.
The temporal behavior of the electron densities induced by the ultrashort laser ablation of the MgB 2 target was evaluated by the Full Width at Half Maximum (FWHM) Stark broadening of the Mg I line centered at 383.8 nm according to the relation where  and  ∘ are the Stark and ion broadening parameters, respectively [35], and Λ is the number of particles estimated in the Debye sphere.For nonhydrogenic ions the broadening due to electron contribution is the main one, whereas the perturbation caused by ions is negligible compared to electrons [36].Intensity and FWHM were estimated by fitting experimental curves with symmetrical Voigt profiles.Different spectral series, obtained by a fixed detection gate of 25 ns, and starting at a delay of 5 ns from the laser pulse, were accomplished.The results reported in Figure 2(b) show that in the ultrashort induced plasma the electron density starts at a value of about 3.5 × 10 17 cm −3 in the first nanoseconds and then suffers an exponential decay reaching a plateau value of about 1 × 10 16 cm −3 after few hundreds of nanoseconds.After the quenching of the primary plume, one microsecond after the laser pulse, a second emission of material has been observed by ICCD fast imaging.This secondary plume, with a cosine exponent  = 1.1, has a front velocity of 9.2 × 10 4 cms −1 and it lasts about 10 s.OES achieved in this temporal range shows blackbody-like spectra, suggesting the presence of hot particulate.In order to determine the temperature of the front of the secondary plume at different time delays, its emission in the whole UV-Vis detected range has been fitted by using Wien's equation in the form   ∝ (1/ 5 ) −ℎ/ B  with, in the Mie approximation, the emissivity  considered to be proportional to 1/ [37].The results show an initial temperature of 3490 ± 10 K which decreases to 3200 ± 20 K after 5 s.For longer times the low intensity of the overall emission did not allow temperature calculation.

Thin Films Characterization.
Even if the commercial target is nominally MgB 2 , XPS analysis shows that the real B/Mg ratio is 1.2, evidencing an excess of magnesium, probably due to the synthesis process.On the other hand, XRD analysis of the same target shows the presence of both crystalline MgB 2 and MgB 4 , with the first as the main component and the second presenting only in a small amount.These results are not completely unexpected, since it is known that commercial powders are often nonstoichiometric [38], but, of course, they add further uncertainty to the interpretation of the results obtained in the thin films deposition.
From XRD analyses the films appear to be amorphous in structure, since the spectra show only the presence of the substrate peaks, independently from the substrate temperature.These data are confirmed by micro-Raman spectra that do not evidence the presence of peaks assignable to magnesium borides.
SEM images registered at different deposition temperatures show that the films' morphology does not change with the substrate temperature.In all cases the deposits are uniform and formed by the coalescence of a large number of NPs (Figure 3), as already observed for other metal boride films deposited by ultrashort PLD technique [31].These NPs are visible in a clearer way in the AFM images reported in Figure 4.
The formation of the films by coalescence of NPs is also confirmed by HR-TEM micrographs, showing the first steps of the film growth (Figure 5).From these images it has been possible to calculate the distribution of the NPs diameters.It is clearly log-normal distribution, the mean value of which is 14 nm (Figure 6).
XPS analyses performed on films deposited at different temperatures (25 ∘ C, 300 ∘ C, and 500 ∘ C) reveal that the films composition varies with the substrate temperature.Since, under the resolution of our XPS apparatus, it is not possible to separate the XPS contribution of magnesium boride from oxide in Mg 2p region (both at about 51 eV), only the peak-fitting of the B 1s region is reported.The analysis of the B 1s region (Figure 7) evidences, in the case of films deposited at 300 and 500 ∘ C, that three contributions at 187.4, 189.7, and 192.0 eV are present.These contributions are assigned, following the literature indications [39,40], to magnesium-bonded boron, atomic boron, and boron oxide (B 2 O 3 ), respectively.In the case of the target and of films deposited at room temperature, only two peaks are present, assigned to magnesium bonded boron (187.0 eV) and to boron oxide (192.3 eV).Considering the difficulty of separating the contributions of magnesium boride from oxide, the correct stoichiometry of the magnesium boride films is not so easy to establish.By comparing the molar ratios obtained from XPS data for the films with those obtained for the target (Table 1), some considerations can be done.First of all, a large loss of Mg in the films, in particular in the films deposited at 300 and   500 ∘ C, is evident.If we consider the total B boride /Mg total ratio in the films we see that the larger quantity of Mg is present in the film deposited at room temperature (B boride /Mg total = 1.7).Of course, not all the magnesium present in the film will be bonded to boron (e.g., in the MgB 2 target, the ratio between boron-bonded and non-boron-bonded magnesium is approximately 1/1.5, supposing that the boride forming the target is stoichiometric) so the stoichiometry of the film deposited at 25 ∘ C will probably be MgB  with  comprised between 2 and 4.This should mean a mixture of MgB 2 and MgB 4 .For higher substrate temperatures, further depletion of magnesium is evident in the films, with the occurrence of atomic boron.The large presence of boron oxide in both target and films could be due to surface oxidation taking place in the presence of atmospheric oxygen and is clearly a surface phenomenon, since no boron oxide peaks were found in the XRD spectrum of the target.

Ablation-Deposition Mechanism.
From the data obtained by the analyses of the two plumes produced by fs ablation of MgB 2 , it is evident that the characteristics of the primary plume are similar to those of the plasmas produced by the ablation of the same target performed by ns laser sources [10,33].Indeed, as already seen, the atomic species are the same, while their velocities are comparable.Due to the lack of literature data about the excitation temperature and the electron density of the plasmas produced by ns laser ablation, it is impossible to verify whether these parameters match those reported in this paper for fs ablation but, considering the results from a paper where such comparisons were made, even if on other systems [41], we can hypothesize also in this case a similarity.On the other hand, the characteristics of the deposited films are very different in the two regimes.Since the fs primary plume is very similar to the single plume formed by ns ablation, the different films structure and composition should be related to the presence, in fs ablation, of a secondary plume.As already seen, this plume is formed by NPs.These NPs, probably ejected directly from the target, represent the main component of the ablated material and are responsible for the formation of a nanostructured film [8,24].In the case of MgB 2 PLD, the reported experimental results can be interpreted considering an ablation-deposition mechanism already proposed for other systems deposited by ultrashort PLD [28,31,32].In those previous works we have proposed a mechanism considering the fact that the NPs, forming the secondary plume, were ejected directly from the target and presented initially the target stoichiometry and all the composition changes took place during their flight to the substrate.Indeed, the composition changes depend on the different equilibrium vapor pressures of the elements forming the NPs.From this point of view magnesium diboride, with the large difference of vapor pressure between magnesium and boron, is a good system to test the hypothesis.As previously reported, the initial temperature of the particles was about 3500 K and it decreased to 3200 after 2 s.In these conditions, considering the Mg-B phase diagram [42], the system is formed by a liquid phase containing mainly boron in equilibrium with a gaseous phase containing mainly magnesium.During the flight to the substrate, which lasts about 20 s, the NPs will experience loss of magnesium, leading to a final composition consisting, according again to the Mg-B phase diagram, of a mixture of MgB 2 and MgB 4 .It should be pointed out that probably the presence in the target of a large excess of magnesium could partly compensate the loss of the same element during the flight of the NPs.Of course, due to the high vapor pressure of Mg, the heating of the substrate will lead to a further decrease of the concentration of this element in the deposited films.In fact, the films with the higher Mg concentration were those deposited at room temperature.On the other hand, the increase of the substrate temperature did not improve the films' crystallinity and eventually resulted to be counterproductive.

Conclusions
In conclusion, the laser ablation of the magnesium diboride target performed by an ultrashort pulse laser has confirmed the importance of the nanoparticles forming the secondary plume in the deposition of thin films and the low importance of the species forming the primary plume.Indeed, the films are formed by the coalescence of a large number of NPs and their composition can be related to the dynamics of melted NPs ejected directly from the target.The films composition depends on the NPs composition and on the substrate temperature.From this point of view, the best films have been obtained at room temperature, since heating of the substrate increases the loss of magnesium and does not improve the film's crystallinity.

Figure 1 :
Figure 1: Normalized time resolved emission plots for (a) B II (412.0 nm) and (b) Mg II (448.0 nm) recorded at different distances from the target.The lines connecting the experimental data were added to facilitate the understanding of the graphics.

Figure 2 :
Figure 2: Temporal evolution of excitation temperature (a) and electron density (b) for magnesium diboride primary plume plasma at a distance of 200 m from the target surface.

Figure 3 :
Figure 3: SEM micrographs of films deposited at different substrate temperatures: (a) 25 ∘ C, (b) 300 ∘ C, and (c) 500 ∘ C. The deposition time was 2 hours for all samples.

Figure 4 :
Figure 4: AFM images of films deposited at a substrate temperature of 25 ∘ C (a) and 500 ∘ C (b).In both cases the deposition time was 2 hours.

Figure 5 :Figure 6 :
Figure 5: TEM micrograph showing the initial steps of the films formation.

Table 1 :
Molar ratios obtained from XPS data for the target and the deposited films.