GaN

Taper cutting with maximum cutting depth of 1 μm is carried out on GaN surface. The results are shown as Fig. 29. GaN is hard and brittle, normally with a shallow brittle-ductile transition point of about 176 nm.
It is found that of GaN brittle-plastic transition depth has a close relationship with cutting speed. The transition depth would increase when the cutting speed reduces. Figure 30 shows the result of decreased cutting speed taper cutting and the corresponding AFM results. It can be seen that the brittle-ductile transition depth is 234 nm and the surface fracture is significantly reduced.

Nanoindentation is also carried out to measure the microhardness of GaN and to give references for the nano-cutting process. Measured average hardness is 20.248GPa. Figure 31 shows the loading and unloading curve in the indentation process. It can be seen that GaN have the mechanical properties similar to monocrystalline silicon, with brittle-plastic transition phenomenon in the loading process (the pop-in phenomenon).
The loading and unloading curve is compared with <100> crystal plane p-type high-resistance silicon nanoindentation results in Fig. 32, with average hardness of 12.992 GPa and modulus of 170.617 GPa. The hardness and modulus of Si are significantly less than that of GaN. This shows that the GaN would cause more severe tool wear than that Si but have better plasticity from the unloading curves.

As shown in Fig. 33, single-point diamond turning generates surfaces with obvious tool marks and brittle fracture. It has small possibility of achieving good surface quality while tool wear is serious, which makes the material not suitable for direct cutting. Ion implantation offers a way to reduce the surface hardness.
Experiments of F-ions implanted into the GaN surface with a dose of 5 x 1014 ion/cm2 and energy of 6 MeV are carried out. A nanoindenter is used to measure the surface properties, as shown in Fig. 34. Modified GaN has a hardness and elastic modulus for the average of 22.719 and 274.877 GPa, respectively. The primitive GaN in the same test conditions result in 20.248 GPa for hardness and 294.627 GPa for modulus. The reduction of hardness is not obvious, but the modulus is, indicating that the resistance of the material on the cutting process significantly reduced.
The pop-in phenomenon upon loading disappears. This shows that the brittle behavior of materials is suppressed, which would be very beneficial in reducing the occurrence of tool wear during cutting. Moreover, the slope of unloading curve of the materials under the same load to 8,000 μN elevated, that means, plasticity of the material increases, and the material is easier to be processed.

GaP

The mechanical property of GaP is first tested using taper cutting. A single-point diamond tool with a nose radius of 0.5 mm is used for the taper cutting experiments. AFM and SEM are used to observe the brittle-ductile transition zone. The ductilebrittle transition point is measured to be 159.2 nm.
Ion implantation of GaP has been carried out; considering the larger density of material, Cl ions of 1 x  1015 ion/cm2 and Si ions of 5 x 1015 ion/cm2 are implanted. After implantation, the brittle-ductile transition depth is observed, and it is found that the implanted material has better cutting performance. The AFM test results show that the two different kinds of particles enhance the brittle-ductile transition depth of more than 50 nm (211.346 nm for Cl ions and 203.902 nm for Si ions).
Micro structures are machined on GaP using flying cutting as shown in Fig. 35. It can be seen that good surface quality can be achieved and surface modification using ion implantation improves the cutting efficiency.

LiNbO3

LiNbO3  is an important optoelectronic material due to its suitable electronic and acousto-optic constants, high chemical and mechanical resistance, high Curie temperature, and large single domain crystals which are available at low costs.
LiNbO3 has a brittle nature, which prevents it from mechanical manufacturing. Traditionally it is fabricated by chemical etching, ion bombardment, etc., which are time-consuming and with low efficiency.
A bare possibility in LiNbO3  of ductile chip removal and mirror-finished surface can be obtained by micro machining using highly accurate machine tool under vibration-free operation. Ion implantation is a superb method for modifying the surface properties of materials since it offers accurate control of dopant composition and structural modification at any selected temperature.

Surface Annealing

Machining would inevitably cause subsurface damages such as phase transformations, dislocations, and fractures. It is the depth and nature of the subsurface damages that influence the mechanical, optical, and electronic performances of silicon products (Yan et al. 2005). Surface-modified silicon, for the same reason, needs to be annealed to recover the damages.
Many factors could influence the annealing results, such as the temperature, annealing time, heating rate, and even the environmental pressure. Simulation and experiment are both needed for studying the phenomenon. Several methods could be applied for the damage recovering process, such as thermal annealing and laser radiation annealing. Thermal annealing is widely used as it is efficient and easy to control.
Thermal annealing was carried out on (100) CZ silicon implanted with B (200 keV, 500 keV) and He (350 keV, 500 keV) ions. Ion fluences are 1 x  1015 ion/cm2 and 2 x  1015 ion/cm2, respectively. Experiment parameters and result are shown in Table 2 and Fig. 36. Structural recovery is measured using changes of bulk resistivity. Implanted ions are blocked at interstitial positions. Thermal annealing allow interstitial ions enter the substitutionally points which could reduce the resistivity.

The results show that temperature is more influential than time. The resistivity descending speed is quite slow when the temperature increases. It is because the stable structure which ion implantation induced requires high temperature for annealing. Longer time can reduce resistivity as well, but it is not as effective as temperature. As time increases, only the damages which can be decomposed at annealing temperature disappear; resistivity reduces more slowly with time. In addition, it will form more stable damage structures when temperature is not high enough (as 300o C annealing in Fig. 36). This phenomenon indicates that simple defect (e.g., Frenkel pairs) can move and accumulate. The new stable damages will survive the annealing and then absorb more defects, which increases the resistivity.
Another group of experiments was also carried out, aiming to study how temperature change rate (TCR) and multiple-stage annealing influence resistivity. The parameters and result are shown as Table 4 and Fig. 37.

For the same annealing condition (30 min at 400o C), small TCR (~2o C/min, furnace cooling) has better effect. While temperature is increasing slowly, damages have enough time to be dissolved and annihilated at each “temperature stair” than large TCR (~10o C/min). Actually, though both of them keep 30 min at 400o C, small TCR process has longer annealing time which reduces resistivity more.
Multiple-stage annealing (400o C, 120 min + 485o C, 60 min) has very little improvement compared with single 120 min, 485o C, annealing. Multiple-stage annealing spends more time at 400o C, so it can dissolve more damages before reaching 485o C. But the maximum temperature has not been changed. In other words, they have the same recovery capability. So the final resistivity doesn’t change obviously. Only time delay cannot compensate temperature difference. As time increasing to 3 h and 4 h at 400o C, resistivity reduces slowly. This is consistent with the analysis above.
Appropriate high temperature and long time are beneficial for better annealing, while the latter may have negative effect at low temperature. Temperature is more influential and cannot be compensated only through annealing time. Reduce the TCR (temperature change rate) if possible. Multiple-stage annealing has no obvious improvement to the result; however, it may be combined with laser annealing.