GB2605261A - Method of polishing a polycrystalline diamond body - Google Patents

Abstract

Method of polishing a polycrystalline diamond body by providing a femtosecond laser with maximum average power or 1.1kW and minimum average power 900W, providing a polycrystalline diamond to be polished and operating the laser to generate femtosecond laser beam comprising a laser pulse with a pulse duration less than 900fs and a wavelength between 900-1100nm, delivering a fluence or fluence per pulse at the surface of the diamond body of less 50J/cm2, ablating the surface of the diamond with the laser beam to reduce its surface roughness.The laser may operate in burst mode of a single burst, 3-pulse, 5-pulse or 7-pulse mode. Also claimed is use of a UV femtosecond laser to polish a polycrystalline diamond body, the laser having maximum average power of 1.1kW and minimum average power 900W, fluence per pulse at the surface of the diamond body of less 20J/cm2. Also claimed is use of a UV femtosecond laser to polish a polycrystalline diamond body, the laser having maximum average power of 1.1kW and minimum average power 900W, and operating in burst mode.

Description

METHOD OF POLISHING A POLYCRYSTALLINE DIAMOND BODY

FIELD OF THE INVENTION

The invention relates to the laser ablation of diamond products. In particular, it relates to a system for polishing a polycrystalline diamond (PCD) body using a femtosecond laser and a polishing method using the same.

BACKGROUND ART

Previous studies on the laser polishing of CVD diamond disclose mostly nanosecond lasers, utilized because of their high removal rate reaching from a few minutes up to several hours per square centimeter depending on the laser source, the number of processing steps, and the initial roughness, Ra. The roughness of thin CVD diamond films (< 100 pm) with an initial Ra varying between 0.1 -1 pm can be reduced by a factor between 2 & 4, and higher reductions are achievable for thick films with an Ra varying between 20 -30 pm.

UV femtosecond pulsed lasers have since been introduced to overcome the modification of the surface composition observed with nanosecond pulsed lasers. In a previous study, after irradiation with a KrF (A = 248 nm and r = 500 fs) or a Titanium: sapphire (X, = 825 nm and r = 120 fs) based laser focused with a 100-150 pm spot size, no amorphous carbon or graphite was present on both CVD diamond wafers and a natural single diamond crystals type IIA, leading to the conclusion that no phase transformation occurred during ablation with UV femtosecond lasers. However, UV femtosecond lasers do not produce a surface entirely free of imperfections, as rippled surfaces or streaks can be visible. Even though UV nanosecond lasers have a higher material removal per unit of time, UV femtosecond lasers reach a higher material removal per pulse at low fluence (To < 20 J/cm2).

A higher removal rate per time is theoretically achievable if the power of femtosecond lasers could technically match the one delivered by nanosecond lasers.

After the recent development of high-power femtosecond lasers, these results on CVD diamond have encouraged the inventors to develop a laser polishing process on PCD materials using a high-power femtosecond laser. As part of their work, the inventors reviewed previous studies in the field, which are outlined below.

Study at lower average power (P., < 80W) Previous work by the inventors on PCD laser polishing started with a 5W average power femtoseconds laser before upscaling up to 80W where various processing parameters were demonstrated to influence the surface quality of the PCD surface.

First, the existence of heat accumulation during ultrashort pulse laser ablation of PCD was demonstrated when the scanning speed was too low and secondly, when the pulse overlap po% reached over a threshold po%,th. When po% > po%,th, the surface quality was highly deteriorated due to thermal effects, for which po%,th = 95% at Pay = 5W, and po%,th = 90% at Pa, = 80W.

It was also found that the number of scanning passes must be limited to a certain number (20 passes as assessed on fine PCD grade), otherwise the surface roughness rises with more passes.

For both fine and medium PCD grades, the surface roughness (Sa and Sz) reduced with increasing fluence until rising at higher fluences (To > 25 J/cm2). The cobalt binder material present in PCD is ablated prior to diamond due to its lower fluence threshold compared to diamond. When the ablation of diamond initiates, the surface roughness begins to reduce, and reaches an average areal surface roughness, Sa that is 36% below the initial roughness on a fine PCD grade, before heat starts accumulating which induces thermal damage.

The introduction of pulse burst leads to a further reduction of the surface roughness at low fluences per pulse (To < 9 J/cm2) reaching up to 27%, compared to single pulse on fine PCD grades. However, the burst mode was found to exacerbate the heat accumulation effects in the coarser grades which have a higher thermal diffusivity than fine grades, and the surface quality deteriorated compared to single pulse mode.

In another study, the thickness of a graphite layer was measured after processing CVD diamond with ultra-short pulse lasers with a pulse duration down to r = 100fs. In a different study, no graphitization was observed after laser ablating PCD tools (4 pm diamond grain size) with a picosecond laser (X. = 1030nm, r = 10ps). However, graphitization of diamond was later demonstrated by the inventors to occur in PCD material with ultra-short pulse lasers (X = 1030nm, r = 400fs) like that observed in the aforementioned CVD diamond study. It was found that this phase transformation is determined by the fluence and is dependent on the thermal conductivity of the PCD grade. Indeed, an amorphous carbon phase only forms on the surface of a fine PCD grade when (po > 19% (see Figure 1), whereas no graphitization can be detected on medium and coarse PCD grades, as the laser induced heat conducts faster into the bulk of the PCD and the local temperature at the ablated spot remains lower than the graphitization temperature.

Figure 1 shows the Raman spectra of fine PCD grade surfaces after laser processing at increasing peak fluences (po (X = 1030nm, r = 400fs, Pay = 5W) from a previous study. The Raman spectrum of the fine PCD grade surface before laser processing is also displayed (lapped surface).

The spectra of the surfaces processed for coo < 11% are similar to the spectrum of the lapped surface with a sharp diamond peak at 1332cm-1 and a weak wide peak in the G band at 1583cm-1 highlighting the presence of traces of graphite between diamond grain boundaries.

When cpo = 19% (red curve), the intensity of the peak in the G band at 1583cm-1 becomes sharper and more intense. The diamond peak is no longer centered at 1332cm-1 and widens over 1350cm-1, consequences of the appearance of a D peak and its addition to the diamond signal. Both diamond and disordered graphite are present. When coo > 28% (green curve), the diamond peak disappear and, instead, a wide peak D centered at 1346cm-1 appears while a G peak is centered at 1580cm-1, characteristic of the presence of amorphous carbon.

With ultra-short pulse lasers, the maximum ablation rate on PCD linearly increases with the average power -see Figure 2a. Furthermore, as displayed in Figure 2b with a 2-pulse burst (red squares), the burst mode appears to increase the ablation rate in comparison to single pulse (black curve) at fluences (pa > 3.5 J/cm2, while no increase is recorded at lower fluence. This increase of ablation rate with the burst mode at high fluence is considered to be related to heat accumulation which reduces the energy required to reach the vaporization temperature of POD (diamond/cobalt). The occurrence of heat accumulation during burst ablation also helps to explain the rapid increase in surface roughness with high fluence measured previously.

Finally, in another study, a Sa reduction of the initial surface by 50% was achieved with a twice higher removal rate than mechanical polishing on fine POD grade with a femtosecond laser at Pay = 80W (r = 400 fs, 3-pulse burst).

A higher average power ultra-short pulse laser is required in order to further increase the volume removal rate and develop a high throughput process. It is an object of the invention to provide such a process.

SUMMARY OF THE INVENTION

In a first aspect of the invention, there is provided a method of polishing a polycrystalline diamond body, the method comprising: providing a femtosecond laser system with a maximum average power of 1.1 kW and a minimum average power of 900W; -providing a polycrystalline diamond body to be polished; - operating the laser system to generate a femtosecond laser beam comprising a laser pulse with a pulse duration less than 900 femtoseconds and a wavelength between 900 and 1100 nm, the laser system configured to deliver a fluence ya or fluence per pulse at the surface of the polycrystalline diamond body of less than 50 J/cm2, and - ablating the surface of the polycrystalline diamond body with the femtosecond laser beam to thereby reduce its surface roughness. Preferable and/or optional features of the first aspect of the invention are provided in claims 2 to 15.

In a second aspect of the invention, there is provided a system for polishing a surface of a polycrystalline diamond body, the system comprising: - a fixture to support a polycrystalline diamond body, a laser system with a maximum average power of 1.1 kW and a minimum average power of 900W that generates a femtosecond laser beam, the femtosecond laser beam comprising a laser pulse with a pulse duration less than 900 femtoseconds and a wavelength between 900 and 1100 nm, the laser system further configured to deliver a fluence To or fluence per pulse at the surface of the polycrystalline diamond body of less than 50 J/cm2; and a controller configured to control relative positioning of a surface of the polycrystalline body and the femtosecond laser beam.

Preferable and/or optional features of the second aspect of the invention are provided in claims 17 to 22.

In a third aspect of the invention, there is provided use of a UV femtosecond laser system to polish a polycrystalline diamond body, the UV femtosecond laser system having a maximum average power of 1.1 kW and a minimum average power of 900W, and configured to deliver a fluence cpo per pulse of less than 20 J/cm2to the surface of a polycrystalline diamond body.

In a fourth aspect of the invention, there is provided use of a UV femtosecond laser system to polish a polycrystalline diamond body, the UV femtosecond laser system having a maximum average power of 1.1 kW and a minimum average power of 900 W, and operating in burst mode.

BRIEF DESCIPTION OF THE DRAWINGS

Non-limiting examples of the invention will be described with reference to the accompanying drawings, in which: Figure 1 shows Raman spectra of fine PCD grade surfaces after laser processing at increasing peak fluences rpo (X = 1030nm, r = 400fs, Pay = 5'A.

Figure 2a shows the normalized maximum ablation rate as a function of the average power on fine and medium PCD grades (X = 1030nm and r = 400fs). A linear fit is applied on the experimental results for both PCD grades.

Figure 2b shows the normalized ablation rate as a function of the peak fluence coo per pulse in a burst for a fine POD grade. 1 (single pulse) to 6 pulses in a burst were tested (X = 1030nm, r = 400fs, Pav = 80W, and AtB = 23ns).

Figure 3 shows a schematic of the laser system used for laser ablation trials.

Figure 4a shows normalized surface roughness, Sa, of the fine POD grade surface as a function of the peak fluence per pulse yo for single pulse, 3-pulse burst and 5-pulse burst (A = 1030nm, r = 500fs, and 5 scanning passes), and SEM images of the surfaces processed at: (b) single pulse To = 1.98J/cm2, (c) single pulse pa = 7.64J/cm2, (d) 5-pulse burst To = 0.54J/cm2, (e) 5-pulse burst To = 2.03J/cm2 as highlighted in Figure 4a.

Figure 5a shows roughness S. of the medium POD grade surface as a function of the peak fluence per pulse To for single pulse, 3-pulse burst and 5-pulse burst (X = 1030nm, T = 500 fs, and 5 scanning passes), and SEM images of the surfaces processed at: (b) single pulse To = 1.98 J/cm2, (c) single pulse (P13 = 6.85 J/cm2, (d) single pulse pa = 16.27 J/cm2, (d) 3-pulse burst To = 4.07 J/cm2 as highlighted in Figure 5a.

Figure 6a shows roughness Sa of the coarse POD grade surface as a function of the peak fluence per pulse To for single pulse, 3-pulse burst and 5-pulse burst = 1030 nm, r = 500 fs, and 5 scanning passes), and SEM images of the surfaces processed at: (b) single pulse pa = 4.27 J/cm2, (c) single pulse (pa = 9.28 J/cm2, (d) 7-pulse burst (pa = 1.09 J/cm2, (e) 7-pulse burst (pa = 1.89 J/cm2 as highlighted in Figure 6a.

Figure 7 shows SEM images of fine POD grade surfaces after laser processing with single pulse at both: (a) P., = 260W, To = 2.34 J/cm2 and (b) P." = 990W, To = 2 30.65J/cm2, and with 3-pulse burst at both: (c) Pay = 260W, pa = 0.78 J/cm2 and (d) Pay= 990W, pa = 0.88 J/cm2 (X = 1030 nm, r = 500 fs, and 25 scanning passes), along with (e) the Raman spectra of these surfaces and the Raman spectrum of the surface processed at 5-pulse To = 2.03 J/cm2. The S. and 57 values are relative to the values characterizing the initial surfaces.

Figure 8 shows normalized ablation rate as a function of the peak fluence per pulse TO for different average powers Pay. Results are displayed at single pulse and 3-pulse burst when Pay > 260W.

DETAILED DESCRIPTION

Experimental set-up A femtosecond laser delivering an average power of 1110W was integrated into a test laser system. The main characteristics of the laser are summarised in

Table 1.

Parameters Specifications

Wavelength 1030 nm ± 5 nm Pulse duration 500 fs -800 fs Max. average power* 1110W Max. frequency 1 MHz Max. pulse energy 2.0 mJ at 500 kHz Max. pulses per burst 27 Infra burst distance 22.7 ns Beam quality TEMoo, M2<1.6 Polarization Linear

Table 1

A Lasea station was used to direct the laser beam, as illustrated in Figure 3. A galvoscanner was used to deflect the laser beam onto the workpiece, with a maximum average power of 990 W after absorption from the optics. Three POD grades with different compositions and diamond grain sizes (fine, medium and coarse) were tested.

The samples were placed on conventional CNC stages (X, Y, Z) to allow linear movements for successive ablation trials and for positioning of the POD top surface at the required distance from the focusing optics depending on the targeted spot size. The laser beam was linearly polarized when interacting with the POD material as no quarter wave plates were set on the optical path, due to their optical absorption being too elevated for the prescribed average power.

Surface areas measuring 5 mm x 5 mm were laser-ablated on each PCD body of material. The area roughness parameter S. was measured in the bottom surface of the squares by focus-variation microscopy with an Alicona G4 (mag x50) over an area of 285 pm x 216 pm. Five measurements were taken per sample (four in the corners and one in the middle of the squares) and averaged to minimize the standard deviation up a maximum value of 10% across all trials. Secondary electron images of the ablated surfaces were recorded with a Hitachi 5-4800 FEG Scanning Electronic Microscope (SEM). The ablation rate was measured as described previously. The pulse energy could only be modified by changing the average power of the laser (diode efficiency), therefore the trials were carried out with different pulse energies/average powers, and the final upscaling was done at 2 mJ / 1 kW.

Results Study of the surface roughness after processing at Pay = lkW As a result of the high average power and the large spot diameter, a wider range of fluence is accessible compared to previously reported studies. The peak fluence goes from po = 0.88 J/cm2 to 50.42 J/cm2 with single pulse (Figure 4a, Figure 5a and Figure 6a). This approach allows more regimes to be distinguished where the S. reduces for both single pulse and pulse bursts instead of one regime according to previous work discussed earlier.

Using fine PCD grade, two fluence ranges of S. reduction can be distinguished: a first regime where cobalt is ablated/melted only before a second regime where the diamond grains are ablated at higher fluences (Figure 4a). Indeed, during this first regime, clusters of diamond grains are not ablated due to the presence of flat valleys of ablated cobalt laying between diamond peaks as visible in Figure 4b and Figure 4d. Then, in the second regime, the diamond peaks are flattened as shown in Figure 4c and Figure 4e. In previous work, the trend of the experimental data shows a 5-pulse burst could potentially improve the surface roughness more than a 3-pulse burst. This is confirmed by the results in Figure 4a where the minimum S. achieved in the second regime is S. = 59% with a 5-pulse burst and S. = 70% with a 3-pulse burst. The reduction of S. values in the second regime with a 5-pulse burst (S. = 59%) compared to single pulse (So = 66%) is confirmed by SEM: the polishing of diamond peaks appears enhanced with a 5-pulse burst in Figure 4e in comparison to single pulse in Figure 4c. The optimal laser ablation of diamond, where the Sa is the lowest in the second regime, is performed at a lower fluence per pulse with 5-pulse burst ((p0 = 2.03 J/cm2) than single pulse (To = 7.64 J/cm2) or 3-pulse burst (To = 3.38 J/cm2) according to Figure 4a. This can be related to incubation effects which causes a reduction of the fluence threshold of diamond with an increased number of pulses in a burst. Due to the energy distribution of the pulses (Gaussian here considering the M2 of the beam), the pulse energy is partially deposited in the material under the form of residual heat where the fluence of the pulse is lower than the fluence threshold of the material. This amount of residual heat is lower in the case of a pulse with less energy and therefore less thermally induced damage is generated. In conclusion, less residual heat is generated when diamond is ablated with a 5-pulse burst explaining the smoother surface roughness achieved.

On medium PCD grade, three regimes with a Sa reduction can be differentiated at single pulse, 3-pulse, and to a lesser extent 5-pulse (Figure 5a). At single pulse, in comparison previous work where the fluence is limited, two additional regimes can be observed at fluences To < 9 J/cm2. Similar to the fine grade results in Figure 4, these two regimes correspond to the ablation of cobalt followed by the ablation of diamond. Clusters of diamond are visible in Figure 5b after ablation at po = 1.98 J/cm2 whereas they are flattened at a higher fluence of To = 6.85 J/cm2 as shown in Figure Sc. Figure Sc also reveals the formation of cavities on the surface with a size < 5pm. At fluences To > 9 J/cm2, a third regime is highlighted at single pulse where the surface roughness Sa decreases with increasing fluences until a constant value of yo > 25 J/cm2 is reached. The access to higher fluences in this work allows to confirm this observation from 25 J/cm2 to 50 J/cm2. These cavities are more numerous and larger in size > 5pm in this third regime as shown in Figure 5d. These cavities can represent the holes left after diamond grains are expelled off the surface by spallation, mechanism of material removal with ultrashort pulse laser and larger diamond grains are expelled at higher fluences. During ultra-short pulse laser ablation, the ultrafast heating (> 1014 K/s) and thermalization of the electron/lattice systems (-100 ps in the case of aluminum) generate tensile waves within the material. When the tensile strength of the material is exceeded, subsurface voids are formed and grow until fractures are induced parallel to the surface leading to the ejection of layers of material. This spallation mechanism is enhanced in the case of a composite materials like PCD where cobalt and diamond have different thermal expansion coefficients: the variation of volume expansion between the two materials is favorable to the formation and propagation of cracks at the interface diamond/cobalt causing diamond grains to be ejected. An identical behaviour is observed with bursts of pulse with the formation of cavities in a second regime and their enlargement and greater number in a third regime (Figure 5e). The burst mode does not bring any improvement in terms of S. compared to single pulse (Figure 5a) as proven by the consistency between the smoothest surface topographies achieved in the third regime at single pulse (Figure 5d) and 3-pulse burst (Figure 5e). As observed with the fine grades, the incubation effect with the burst mode are responsible for lowering fluences where the minimum S. is achieved per regime.

However, due to the higher conduction of medium grade over fine grade, the heat is spread at a higher rate in the bulk of the medium grade material and less localized around the targeted area which causes less thermal stress. Therefore, the reduction of residual heat with the use of burst mode does not reduce the level of thermal damage in comparison to single pulse, as previously measured when (pa < 4.07J/cm2 at 3-pulse. However, the higher conductivity causes stronger heat accumulation and surface damage occurring at pa > 4.07J/cm2at 3-pulse and responsible for the higher S. values at 5-pulse burst.

For the coarse POD grade, two regimes of fluences with a reduction of S. occurs at single pulse. The number of regimes for 3-pulse and 5-pulse bursts is debatable, nonetheless three regimes are clearly apparent at 7-pulse burst (Figure 6a). At single pulse, as previously shown for the fine POD grade, ablation of cobalt takes place in the first regime before the ablation of diamond follows in a second. The SEM of the surface processed at pa = 4.27 J/cm2 at single pulse shows that diamond grains -25pm wide, corresponding to the diamond grain size in coarse POD grade, are not ablated while the area surrounding the grains, corresponding to the location of the binding material, is ablated (Figure 6b); then, at a higher fluence pa = 9.28 J/cm2, ablation of the diamond grains takes place as illustrated by the change of the diamond grains morphology in Figure 6c. At 7-pulse burst, diamond grains are pulled out leaving cavities with a width < 5pm on the surface in the second regime as highlighted in Figure 6d at (pa = 1.09 J/cm2. The formation of these cavities on coarse POD surfaces after ultrashort laser ablation was already detected by focus variation microscopy in a previous study. In a third regime, the number of cavities may appear more numerous in Figure 6e with the increase of fluence at (pa = 1.89 J/cm2, and their size remains in the same range < 5 pm. Considering the grain size distribution of coarse POD grades, some diamond grains with a size < 5pm compose the surface and can therefore be expelled during laser ablation as observed earlier with medium PCD grades, while larger grains remain on the surface. As partially determined in a previous study, Figure 6a confirms that the addition of pulses per burst reduces the roughness Sa up to 55% during the ablation of coarse grade surfaces. The SEM analyses corroborates the Sa measurements: a smoother surface is achieved with 7-pulse bursts (Figure 6d) compared to single pulse burst (Figure 6c). Coarser grades have a higher fluence threshold as determined in previous studies, and a higher number of pulses per burst allows the reduction of the fluence threshold via incubation effects and effectively ablate large diamond grains at lower fluence with minimal residual heat. A significant reduction of the Sa is achieved in the diamond ablation regime ultimately: from Sa = 42% at single pulse (90 = 9.28 J/cm2) to Sa = 19% at 7-pulse burst (To = 1.09 J/cm2).

Final results at Pay = lkW The process was scaled up from Pay = 260W and eventually to Pay = 990W. This was achieved by defocusing the beam to reach the targeted fluences according to the above surface roughness study.

After optimization of the process parameters, the initial surface roughness Sa of a fine PCD grade was polished by 14% after laser processing with a 3-pulse burst and To = 0.88 J/cm2 at Pay = 990W (25 scanning passes). A 5-pulse burst and processing with a peak fluence per pulse To in the second regime could potentially improve the surface roughness Sa further (Figure 4a); unfortunately, the Sa measurements were ulterior to the upscaling trials which were carried out with a 3-pulse burst following the results from previous studies.

The surface topography is represented by SEM in Figure 7d along with the topography of the surfaces processed at lower power, where Pay = 260W (Figure 7c) and at single pulse for both Pay = 260W (Figure 7a) and Pay = 990W (Figure 7b). No significant differences in topography between the surfaces processed at Pay = 260W and Pay = 990W are noticeable which means the process can successfully be scaled up in terms of average power ramp up Furthermore, the surfaces appear more homogenous after 3-pulse burst rather than single pulse processing and confirms the possible reduction of roughness Sa with the burst mode compared to single pulse processing with fine POD grade surfaces as demonstrated in Figure 4a.

Therefore, Raman analyses were carried out on these four aforementioned fine grade surfaces to inspect if the diamond structure is still present on the top surface after laser processing at high power (Figure 7e). The spectra are similar to the spectrum taken on a lapped surface pre-laser ablation (Figure 1) with a diamond peak at 1332cm-1 as well as a wide G band centered at 1600cm-1, and no peak in the D band. Therefore, the occurrence of graphitization can be excluded. This is expected for single pulse processing as the surfaces are processed with a peak fluence 90 < 2.65 J/cm2 which is lower than the fluence threshold of graphitization as determined in Figure 1. No additional thermal effects capable of graphitizing the surface arise from the use of a 3-pulse burst (like heat accumulation) at low fluence per pulse, which is consistent with the constant ablation efficiency between single pulse and burst mode processing at (p0 < 3.5 J/cm2 (Figure 2b). The Raman spectrum taken off a surface processed with a 5-pulse burst is also displayed to confirm the presence of diamond and the absence of thermal effects with the burst mode.

This Sa reduction of 14% at Pay = 990W (T = 500 fs, 3-pulse burst) is performed with an ablation rate over 70 times higher than at Pa, = 80W. The increase of ablation rate with the average power from BOW to 260W and further 990W is linear as predicted in Figure 2a. Converting from single pulse to 3-pulse burst reduces the ablation rate as plotted in Figure 7 at Pay = 260W and Pay = 990W. As seen in Figure 2b, the ablation efficiency is constant between single pulse and burst mode when wo < 3.5J/cm2 without the generation of additional thermal effect as confirmed by Raman spectroscopy (Figure 8). Therefore, this reduction of ablation rate is only the consequence of processing at a lower fluence per pulse.

Concerning the other grades, no improvement of the initial surface roughness is achieved after the optimization trials of the laser ablation process at Pay = 260W and Pay =1kW. Laser polishing coarser POD grades is more challenging than fine grades. The cavities formed by the spallafion of the diamond grains do not allow to perform a reduction of the initial roughness of medium and coarse grade surfaces. The additional results did not reveal any potential improvements regarding the medium grades; however, the results with 7-pulse burst on coarse grade surfaces showed a promising trend of surface smothering but were not reproducible at high power yet.

Summary

A 1kW femtosecond laser system (X. =1030 nm and r = 500 fs) was first tested to investigate the variation of surface roughness Sa as a function of the fluence per pulse To for a wide range of values from cpo = 0.88 J/cm2 to 50.42 J/cm2 at single pulse. A burst mode was also experimented to test the effects of the number of pulses in a burst on the surface roughness compared to single pulse processing.

This investigation demonstrates the existence of several ablation regimes for each PCD grade and the specific evolution of the surface roughness Sa upon laser ablation at single pulse and burst mode processing per PCD grade.

For fine PCD grade, two regimes take place starting with the ablation of cobalt followed by the ablation of diamond at higher fluences. The surface roughness reduces with the addition of pulses per burst: by 10% from single to 5-pulse. The repetition of pulses in a burst on the PCD surface causes incubation effects, which allows processing of the diamond at a lower fluence, and therefore reduces the residual heat and the related thermal damage.

For medium PCD grade, the ablation of diamond results in the spallation of diamond grains after the ablation of cobalt at lower fluence. The diamond ablation proceeds in two distinct regimes where the size and number of the cavities (< 5pm) left from the expulsion of the diamond grains increase from one regime to another of higher fluence. Even though incubation effects occurs with the burst mode, the reduction of residual heat does not improve surface roughness of medium grade compared to single pulse.

The high thermal conductivity of medium grade (higher than fine grade) allows the evacuation of the residual heat into the bulk of the material and the minimization of localized thermal effects at single pulse already.

For coarse PCD grade, the ablation process at single pulse follows the one described for fine grade with two regimes; however, the addition of pulses in a burst leads to an enhancement of the spallation of diamond grains and clearly results in three regimes at 7-pulse burst like for medium grades. Considering the high fluence threshold (higher than fine grade), the burst mode allows a more efficient ablation of the large PCD grains (-25um) thanks to incubation effects and results in a Sa lower by 55% than achieved at single pulse.

The laser ablation process was then scaled up to an average power Pa, = 1 kW.

Finally, a polishing laser operation was developed for fine grade PCD surfaces which is capable of reducing by 14% the initial surface Sa with a 140 times increase of volume removal rate compared to a traditional mechanical process (70 times increase compared to previous study at Pay = 801/10. The process is carried with a 3-pulse burst at a fluence per pulse below the fluence threshold of graphitization, which is estimated at single pulse. No phase transformation of the diamond structure is detected which confirms the absence of additional thermal effect with burst mode at qm < 3.5 J/cm2.

The cavities formed by the spallafion of the diamond grains on coarser PCD grade surfaces prevent a reduction of their initial roughness during the laser polishing trials at Pay =1kW to date.

The laser polishing of fine grades can be optimized further on fine grade surfaces with a 5-purst and potentially a higher number of pulses per burst in order to achieve a higher Sa reduction than 14%.

Regarding coarser grades, the formation of cavities must be prevented to allow an efficient polishing of the initial surface.

Claims (24)

1. CLAIMS 2. 3. 4. 5. 6. 7. 8.
2. A method of polishing a polycrystalline diamond body, the method comprising: providing a femtosecond laser system with a maximum average power of 1.1 kW and a minimum average power of 900W; providing a polycrystalline diamond body to be polished; operating the laser system to generate a femtosecond laser beam comprising a laser pulse with a pulse duration less than 900 femtoseconds and a wavelength between 900 and 1100 nm, the laser system configured to deliver a fluence wo or fluence per pulse at the surface of the polycrystalline diamond body of less than 50 J/cm2, and ablating the surface of the polycrystalline diamond body with the femtosecond laser beam to thereby reduce its surface roughness.
3. The method as claimed in claim 1, wherein the polycrystalline diamond body is high pressure high temperature (HPHT) diamond.
4. The method as claimed in claim 1 or 2, comprising operating the laser system in burst mode.
5. The method as claimed in claim 3, comprising operating the laser system in a single pulse burst mode.
6. The method as claimed in claim 3, comprising operating the laser system in 3pulse burst mode.
7. The method as claimed in claim 3, comprising operating the laser system in 5-pulse burst mode.
8. The method as claimed in claim 3, comprising operating the laser system in 7-pulse burst mode The method as claimed in any preceding claim, comprising configuring the laser system to deliver a fluence Po or fluence per pulse of less than 20 J/cm2.
9. 9. The method as claimed in claim 8, comprising configuring the laser system to deliver a fluence poor fluence per pulse of less than 9 J/cm2
10. 10. The method as claimed in claim 9, comprising configuring the laser system to deliver a fluence To or fluence per pulse of between 2 and 8 J/cm2.
11. 11. The method as claimed in claims 8 or 9, comprising configuring the laser system to deliver a fluence poor fluence per pulse of less than 1 J/cm2.
12. 12. The method as claimed in any preceding claim, wherein the polycrystalline diamond body is a fine grade having average diamond grain size of less than 10 pm.
13. 13. The method as claimed in any of claims 1 to 11, wherein the polycrystalline diamond body is a medium grade having average diamond grain size in the range of 10 to 20 pm.
14. 14. The method as claimed in any of claims 1 to 11, wherein the polycrystalline diamond body is a coarse grade having average diamond grain size greater than 20 pm.
15. 15. The method as claimed in any preceding claim, wherein the polycrystalline diamond body is a PCD wafer.
16. 16. A system for polishing a surface of a polycrystalline diamond body, the system comprising: - a fixture to support a polycrystalline diamond body, - a laser system with a maximum average power of 1.1 kW and a minimum average power of 900 W that generates a femtosecond laser beam, the femtosecond laser beam comprising a laser pulse with a pulse duration less than 900 femtoseconds and a wavelength between 900 and 1100 nm, the laser system further configured to deliver a fluence To or fluence per pulse at the surface of the polycrystalline diamond body of less than 50 J/cm2; and -a controller configured to control relative positioning of a surface of the polycrystalline body and the femtosecond laser beam.
17. 17. The system as claimed in claim 16, wherein the pulse duration is between 500 and 800 femtoseconds.
18. 18. The system as claimed in claim 16 or 17, wherein the laser system is configured to deliver a fluence To or fluence per pulse at the surface of the polycrystalline diamond body of less than 20 J/cm2.
19. 19. The system as claimed in claim 18, wherein the laser system is configured to deliver a fluence TO or fluence per pulse at the surface of the polycrystalline diamond body of less than 9 J/cm2.
20. 20. The system as claimed in claim 19, wherein the laser system is configured to deliver a fluence To or fluence per pulse at the surface of the polycrystalline diamond body of between 2 and 8 J/cm2.
21. 21. The system as claimed in any of claims 16 to 19, wherein the laser system is configured to deliver a fluence 90 or fluence per pulse at the surface of the polycrystalline diamond body of less than 1 J/cm2.
22. 22. The system as claimed in any of claims 16 to 21, further comprising a galvanometer scanner for deflecting the femtosecond laser beam onto the polycrystalline diamond body.
23. 23. Use of a UV femtosecond laser system to polish a polycrystalline diamond body, the UV femtosecond laser system having a maximum average power of 1.1 kW and a minimum average power of 900W, and configured to deliver a fluence To per pulse of less than 20 J/cm2 to the surface of a polycrystalline diamond body.
24. 24. Use of a UV femtosecond laser system to polish a polycrystalline diamond body, the UV femtosecond laser system having a maximum average power of 1.1 kW and a minimum average power of 900W, and operating in burst mode.