Ga 3+ Focused Ion Beam for Piezo Electric Nano Structuration Fabrication

rejuvenate at least partially, the piezoelectric activity using a post annealing treatment (as it is currently proposed in the literature, e.g. for PZT films etched by RIBE 22 ), but irrespective of the conditions (temperature, atmosphere) we were unable to restore the initial properties of the films. These results suggest that a damaged layer exists on the surface of the film which affects strongly the piezoelectric performance of PZT Ion implantation presents a continuously evolving technology. While the benefits of ion implantation are well recognized for many commercial endeavors, there have been recent developments in this field. Improvements in equipment, understanding of beam-solid interactions, applications to new materials, improved characterization techniques, and more recent developments to use implantation for nanostructure formation point to new directions for ion implantation and are in book.


Introduction
Piezoelectric nanostructures are currently of the potential interest for the development of MEMS / NEMS (Micro / Nano-Electro-Mechanical Systems). Their main applications in the biomedical domains (micro / nano bio actuators and sensors), in automobile industry (actuators and sensors) and in aeronautic industry (health control monitoring) are based on PbZr 0.54 Ti 0.46 O 3 (PZT) recognized as leading material for these piezoelectric micro/nano devices. 1, 2, 3, 4 The fabrication of piezoelectric nanostructures on pure PZT (or "doped-PZT") films is frequently performed by means of Reactive Ion Etching (RIE or ICP (Inductively Coupled Plasma) -RIE with fluorine and/or chlorine gases), Reactive Ion Beam Etching (RIBE), Focused Ion Beam (FIB) Ga 3+ etching and electron beam direct writing. 5,6,7,8 In the present study we focus our results on Ga 3+ FIB etching of PZT nano structures. Typically, the obtained results show that it is possible to fabricate nanostructures with very small lateral size (50 nm on PZT films) when the films are etching in amorphous state followed by a post annealing treatment at the temperature which correspond to the perovskite formation temperature; the piezoelectric properties are similar to those obtained on un etched PZT films. Any degradation is observed. For crystallized films, the situation is completely different: no ferroelectric properties are observed after etching and the piezoelectricresponse is strongly degraded. A post annealing treatment in oxygen results in a partial recovery of the ferroelectric properties. The main goal of the present study is to evaluate the damage induced by FIB etching of PZT films: implantation of Ga 3+ and film amorphisation. The damaged layer that appears on the films' surface was thoroughly characterized in terms of the composition and charge implantation but a more detailed characterization seems to be necessary. To address the quality of the nanostructures, the measurement of local electromechanical activity via piezoelectric hysteresis loops acquisition using Piezoelectric response Force Microscopy (PFM) is certainly one of the best options. The conclusion of this work is the development of a suitable process of nano-structures manufacture without introducing piezoelectric property degradation.

Structure
In this work, we have studied two types of Si/SiO 2 /TiO x /Pt/PZT/Pt structures, namely PZT films being either in amorphous or crystallized state before etching 9 . The films in both www.intechopen.com Ion Implantation 316 states are deposited by radio frequency magnetron sputtering without substrate heating and therefore the films were amorphous. The film composition is fixed at the morphotropic phase boundary (Zr/Ti = 54/46). A post annealing was used to crystallize the PZT film in the Perovskite phase; the annealing temperature was 625°C (conventional annealing) in air during 1/2 hour 10 . The film thickness varied between 100 and 300 nm depending on the island diameter, i.e., for a small island diameter the film was thinner to reduce the etching time. Figure 1 presents a typical scheme of the PZT nanostructure (nano island) that we fabricate with the FIB system.

FIB system
A focused beam of Ga 3+ ions (STRATA DB 235 -FEI) was used to pattern the PZT film, Figure  2 shows our FIB system. The system was equipped with both the electron and ion beams. The advantage of this system is the possibility to monitor the etching process in real time. The acceleration voltage was 30 kV and ion current was in the range 100 -10pA depending on the island diameter. We have studied the relation between the current density of the ion beam and its diameter. The ion beam was at the 90° angle to the etched surface.

Piezoelectric characterization at nano-scale
At nano-scale, the Atomic Force Microscopy (AFM) is currently used [11][12][13][14][15][16][17] . Several AFM techniques, namely, Piezoelectric-response Force Microscopy (PFM) (and its spectroscopic tool which enables the measurement of local piezoelectric hysteresis loops) and Kelvin Force Microscopy (KFM) have been developed and these techniques were used in this work to characterize the piezoelectric nanostructures. For such investigations, a commercial AFM (Multimode, Nanoscope IIIA, Veeco) operating under environmental conditions was used; a scheme of the system is given Figure 3. The local piezoelectric hysteresis loops were obtained by plotting the amplitude of piezoelectric vibration A as a function of the DC voltage in so-called "in-field" mode in order to provide information about electromechanical activity. The frequency and the amplitude of the driving AC voltage were f = 2 kHz and Vac = 1.5 V, respectively. The applied DC voltage was gradually swept between -10 V and +10 V within the period of 200 s. To minimize the electrostatic force effect, Pt/Ir coated Si tips/cantilevers with relatively high spring constant of about 45 N.m-1 were used 16 . PFM domain imaging was also performed using the same cantilevers with the spring constant of 3.8 N.m -1 (the frequency and the amplitude of the driving AC voltage were f = 8 kHz and Vac = 1.5 V, respectively). KFM is a non contact AFM-based technique based on the electrostatic interaction between the tip and the surface. Surface potential maps are easily detected due to work function variation as a result of surface charges, electric dipoles or absorption layer

Results and discussion
Before the fabrication of Pt/PZT/Pt islands, we have evaluated the degradation induce by the Ga 3+ ion beam on the crystallized PZT films deposited on the entire substrate. A large zone 7 x 7 μm 2 was etched at different doses ( Figure 4); the Ga 3+ dose is in relation with number of path of the beam. More precisely, the number of ion beam paths was varied between 1 and 20. According to 18 , the relation between the number of paths and the Ga 3+ ion dose is the following: ion dose = 4 13 for 1 path and it become 3 15 for 20 paths  For the hardest conventional etching case, the Raman spectra recorded in the region exposed to the FIB shows an important modification in comparison of single PZT crystals 19 corresponds to the (E(LO3) mode). This result is certainly connected with the distortion of the tetragonal symmetry of the PZT crystalline lattice due to the ion implantation and the modification of the stress state 21,22 . The broadening of Raman scattering in 550-900 cm -1 region could be attributed to the existence of the amorphous material which is usually expected during the conventional FIB etching process. 23,24,25 In conclusion, whatever the FIB etching process, the peak position of Raman shift does not show any change, so the crystalline structure still exhibits perovskite phase.
We have compared the piezoelectric response of three samples ( Figure 6): un-etched crystallized PZT film (denoted as sample (a)), amorphous PZT film etched before crystallization (sample (b)) and PZT film etched after crystallization (sample (c)). In crystallized PZT films (sample (c)) we observed strong piezo-response degradation whatever is the ion dose during etching. In fact, no piezoelectric response was obtained even after 1st path. At the same time no ferroelectric domains could be imaged. The degradations induce by the FIB are not limited to the etched region; we have observed that even outside the exposed zone, some defects appears. Figure 7 shows this phenomenon, in this example we use the conventional FIB process. The PFM signal is restore at 100 µm outside the etched zone, the explanation can be in relation with the spatial extension of the ions beam. It is well known that the ions beam has a Gaussian form and as a consequence the zone outside the etched region is also in contact with the Ga 3+ ions beam.
We tried to rejuvenate at least partially, the piezoelectric activity using a post annealing treatment (as it is currently proposed in the literature, e.g. for PZT films etched by RIBE 22 ), but irrespective of the conditions (temperature, atmosphere) we were unable to restore the initial properties of the films. These results suggest that a damaged layer exists on the surface of the film which affects strongly the piezoelectric performance of PZT films. These damages could be: amorphisation of film surface, modification of the film composition, etc. In order to understand more the phenomena we analyse the film surface by KFM. The results are presented Figure 8.  The KFM image of Figure 8 shows the evolution of the tip grounded CPD (Contact Potential Difference) response as a function of the number of paths on the conventional etching sample. The KFM and the CAFM (see later) measurements were made at the same time that means with CAFM configuration the tip is virtually connected to the ground and the bias is applied to the sample. The change of contrast is, of course, directly related to the variation of the surface properties such as the composition and / or amorphisation of the material and induced structural defects. We cannot make more definitive conclusion for the moment but it is clear that after 20 paths the etched region is completely modified. A CPD profile (here the sample is grounded) over the three etched area indicates almost no modification of CPD for 1 path and apparent decrease by 350 mV and 500 mV for 10 and 20 paths, respectively. In our conditions (room and instrument calibration), the CPD of the surface was -50mV relative to a sample grounded. To compare the KFM signals obtained on the samples (b) and (c), we perform a second experiment under the same condition (tip, instrumentation configuration and sample grounding).
The results are presented in Figure 9.  It confirms that the potential increase between the un-etched surface and the surface of PZT etched after crystallization is very large (about 300 mV). The implanted charges are the same in both samples but for amorphous sample we can suppose that the annealing treatment induces these charges neutralization and a slight decrease (10 mV) of the CPD is only detected. The properties of the etched area are certainly different depending of the process. The difference, in terms of material is not only due the presence of charge on the sample surface but some crystal defects (dislocations), substitution between Pb and Ga 17 , implantation of metallic Ga 3+ probably were maintained on sample (c) but not in sample (b) due to the annealing treatment. Figure 4 shows the piezo-response amplitude A measured on etched zone after 1 path and 20 paths in the sample (c) (etched after crystallization). As we can observe, only weak piezoelectric activity is obtained even after one path. The poor ferroelectric properties are confirmed by the absence of the hysteresis on both samples. Therefore, the FIB etching is detrimental to functional properties of PZT. Annealing of these etched zones (in air or in O 2 atmosphere at different temperatures and times) had no effect on piezo-response. On the contrary, for PZT etched before crystallization (sample b), the piezoelectric signal (measured after crystallization) is the same as the one obtained for the sample (a) irrespective of the ion dose. Even if degradation is induced on the sample (b), the annealing treatment (crystallisation of the film) permits to eliminate the damages. These results confirm that the strong degradation occurs on the initially crystallized samples and no degradation happens if the FIB is conducted on amorphous samples.
To conclude this work it is evident that the surface potential is changed after etching, especially for the region exposed to conventional FIB, i.e. on crystallized PZT film. To see it more clearly, we shows the absolute surface potential value across the etching region ( Figure 10). It implies that the potential difference between the un-etched surface and the surface after amorphous etching is very small (only about 15mv), while the difference between the un-etched surface and the surface after conventional FIB etching is very large (about 300mv). Fig. 10. Absolute film surface potential measured by KFM So, it is evident that for nanostructures fabrication, the FIB process is very well adapted when the film is in amorphous state (probably due to the lake of grain). So, no degradation Ga 3+ Focused Ion Beam for Piezo Electric Nano Structuration Fabrication 323 is observed and maybe the post annealing treatment (at temperature equal to the temperature formation of the perovskite phase formation) following the etching process could suppressed the eventual degradations induced is the amorphous one.
In this second part of this work, we focused the results and discussions about the fabrication of PZT nano structures, named nano islands. The typical structures obtained by SEM are shown in Figure11. The size of the island shown in Figure 11 is 50 nm. Fig. 11. PZT nano island of 50 nm in diameter A Pt top electrode was used as a mask to protect the top of the island and, therefore, we take into account only the defects introduced into the islands walls. The etching anisotropy is probably due to the selectivity between Pt and PZT; this effect could be a limitation of the FIB process and it seems more difficult to decrease the island diameter with this technique. The comparative results of the PFM measurements on the PZT islands of 50 nm of size are shown in Figure 12. These results confirm that with conventional etching (on crystallized PZT) the piezoelectric response disappears completely and even after a post annealing treatment the signal could not be restored. On amorphous PZT etching followed by a post annealing treatment at perovskite formation temperature, the piezoelectric response is excellent. These results contradict to those obtained by Stanishevsky et al. 17 where the acceleration voltage of the Ga 3+ ions is larger (50 kV) and so it had to induce stronger degradation. In order to explain why the piezoelectric signal disappeared we have performed a local I (V) measurements with CAFM on the similar PZT island obtained with the conventional process (etching on crystallized PZT) with 250 nm diameter. The result is shown in Figure 13 for these experiments, a V DC bias was applied to the substrate holder and the top surface Pt was grounded. The current increased rapidly with the applied voltage indicating strong local conductivity due to the presence of an electrical conducting channel. This proves a very strong damage due to Ga 3+ ion implantation in the PZT nano-islands and associated loss of the piezoelectric activity. The leakage current becomes more important when the island diameter decreases, it is thus can be responsible for the results on 50 nm islands. This observed conductivity increase is thus responsible for the weak electromechanical activity observed with the conventional process.
To conclude, we present a new device for BioNEMs applications. It consist of a networks of nano islands ( Figure 14)

Conclusion
In conclusion, in this work we have evaluated the damages induced by the Ga 3° FIB process for the fabrication of PZT nano structures. Two processes have been developed: --the conventional process where the films are crystallized before etching followed by a post annealing to restore the initial piezoelectric electric properties, --the amorphous process where the etching process is made on amorphous films followed by a post annealing treatment at the Perovskite formation temperature.
In order to estimate the induced degradations by these two processes we have firstly etched large PZT film zones. The main conclusion is that when the PZT is crystallized in the Perovskite phase a strong degradation is observed and the piezoelectric of the film disappears even after a post annealing treatment. In amorphous state no degradation is observed. So, we have choice this process to fabricate nano piezoelectric structures.
By an optimization of the FIB Ga 3+ beam and by using amorphous PZT films we have fabricated piezoelectric nano structures: the minimum diameter size of the nano structure is 50 nm. Even at this dimension the nano structure gives always a piezoelectric response. To measure this piezoelectric activity on very small material we have developed and optimized many experimental set up such as : PFM, KFM, CAFM; these methods are relatively news and very well adapted to measurements at nano scale but these methods are very sensitive to external parameters: electrostatic charges, contact between the tip and the material surface,… Many precautions must be taken if you want that these measurements are a true reality of the nano structures performances.
For the nano piezoelectric performances, the conclusions are clear and similar to those obtained on film: with the conventional process, the piezoelectric response disappears (even for large islands (diameter = 800 nm) and even with a post annealing treatment the properties can't be restored. The piezoelectric activity of the nano structure is maintained with the amorphous process. So, the amorphous process is used to fabricate piezoelectric nanostructures. PZT islands of 50 nm in diameter were fabricated and the piezoelectric response was unchanged in comparison with un-etched PZT films. Identical results have been obtained on PMN-PT films. We try in future work to decrease the island diameter and maybe found a diameter where the piezoelectric activity disappears as predicted by some authors (when the dimension of the island = dimension of the domains). A limit exist with the FIB process (due to the anisotropy) and we will try to find this limit and maybe the electron beam lithography will be more adapted for very narrow PZT island (< 50 nm in diameter).