Expand this Topic clickable element to expand a topic
Skip to content
Optica Publishing Group

4H-SiC wafer slicing by using femtosecond laser double-pulses

Open Access Open Access

Abstract

Silicon carbide (SiC) is promising as a key material for power electronics devices owing to its wide bandgap property. Meanwhile, by the convention wire-saw technique, it is difficult to slice off a thin wafer from bulk SiC crystal without reserving space for cutting. In this study, we have achieved exfoliation of 4H-SiC single crystal by the femtosecond laser induced slicing method. By using this technique, the exfoliated surface with the root-mean-square roughness of 5 μm and the cutting-loss thickness smaller than 24 μm was successfully achieved. We have also observed the nanostructure on the exfoliated surface in SiC crystal.

© 2017 Optical Society of America

1. Introduction

The silicon carbide (SiC) is well-known prospective material as its robustness properties of chemical stability, and heat resistance. Therefore, SiC has been widely used in applications ranging from polishing material, refractory material, to heating element. Among the over 200 of SiC polytypes, semiconductor 4H-SiC is transparent at wavelength from visible to near infrared. In addition, since 4H-SiC indicate the excellent characteristics of thermal conductivity, wide bandgap, and isotropic electron mobility compared to silicon, it is expected that 4H-SiC acts as a key material for power electronic devices. However, the problem is the cost for the production of SiC wafers. The dicing saw technique is conventionally employed to fabricate semiconductor wafer from ingot [1–4]. Although such wire-saw technique has advantages for mass production from its high throughput [5], it is generally performed under wet environment and requires the cleaning process of debris and cooling of frictional heat during wafer slicing [6]. Moreover, there are also various issues such as the reduction of the strength of wafers by chipping and cracking on the cut surface. Besides, the kerf-loss of material during wire sawing is typically more than 180 μm per a wire [5], and mean size of slurry are under 30 μm, finally the total loss eventually becomes around 250 μm. In the case of SiC with high wear resistance, which has a hardness comparable to diamond and boron nitride, the wire-saw technique is unfavorable from the viewpoint of the production costs. The advanced technique of wafer slicing method so-called “Smart Cut” technology for silicon-on-insulator (SOI) material has been applied by using proton-implantation and sacrificial oxide thin film. Recently, this technique has also attractive to cost-down for SiC applications [6–8]. The smart cut have advantages for high quality slicing the SiC thin film, it critically depends on the blister, defect and unexpected properties by ion-channeling. Another approach of novel dicing technology by pulse laser called as “stealth dicing” (SD) is attracting more attention [9]. Based on the local heating inside silicon by focused nanosecond pulse laser irradiation, the locally high-dense dislocations are generated in the focal depth [10]. Finally the silicon wafer can be easily divided under dry environment by this SD technique without debris. Since the coupling time between electron and phonon is generally shorter than picosecond order, this SD technique provided by nanosecond pulse laser is classified as a thermal process. On the other hand, the wire-slicing method is still widely used since the plural planes can be sliced off at one procedure. As far as we know, there is little knowledge to slice off the smooth planes from the bulk SiC crystal by using laser. While, the structural changes induced by femtosecond laser pulses have been studied about various bulk materials ranging from glass [11, 12], silicon [13], diamond [14] to SiC [15]. In the case of SiC, the formation of the small voids due to micro-explosion has been reported [15]. Owing to the nonlinear optical phenomenon and the spherical aberration, the modified region by the femtosecond laser pulses tends to elongate in the direction of laser propagation. It is expected that the slicing of bulk SiC crystal by femtosecond laser pulses with correction of the spherical aberration could reduce the kerf-loss compared to the conventional wire-sawing method. Here we report on the laser slicing of bulk SiC crystal by using femtosecond laser pulses. In particular, we compared the size of the modified regions after the laser writing by the single- or double-pulse configuration. Unexpectedly, we have observed that the elongation of the modified region in laser propagation direction can be suppressed by the femtosecond double-pulse configuration and controlling the pulse durations. To reveal the possibility of the laser slicing and the kerf-loss, we have also performed the exfoliation test after the femtosecond laser direct writing. Interestingly, the exfoliation force for the sample written by the femtosecond double-pulse trains were at least one-tenth smaller than that for the case of the single-pulse trains. Furthermore, the roughness on the exfoliated surface was improved by using the femtosecond double-pulse trains. We believe that this technique can be applied to slicing from the bulk ingot of SiC.

2. Experiments

In the experiments, we used a sample of commercially available N-doped 4H-SiC (0001) wafers with a thickness of 420 μm (Tankeblue semiconductor), which was cut in a 5 × 5 mm square. Figure 1(a) shows the schematic of the experimental procedure. Optical microscope images of the laser writing regions and the exfoliated surfaces viewed from the cross-sectional direction were also shown in Fig. 1(b) and 1(c). We used a femtosecond laser oscillator equipped with a regenerative amplifier (Cyberlaser; IFRIT). The center wavelength, pulse duration (τp) and repetition rate of this laser system were 780 nm, 220 fs and 1 kHz, respectively. To reveal the pulse duration dependence of the damage threshold, the τp was changed from 220 fs to 6 ps. The laser pulses were focused inside SiC sample in the [0001] direction. To enhance the light-matter interaction, we configured the Mach-Zehnder type of double pulse optical setup [16, 17]. The time delay (τdelay) between femtosecond double pulses was controlled by using an optical delay line. The total pulse energy (Ep) of the equally-divided double pulses was tuned by neutral density filter. To compensate the spherical aberration due to the high refractive index of SiC (n0 = 2.6 at λ = 780 nm), we used a spatial light modulator (LCOS-SLM, Hamamatsu Photonics; X10468). The pre-calculated CGH for the correction of spherical aberration at the focal depth of 260 μm was displayed on LCOS-SLM. The laser beam was focused inside a SiC wafer with a 50 × objective (Nikon, LU Plan Fluor; NA 0.80, Transmittance ≈80% at 780 nm). The transmittance of the sample was about 14% at the wavelength of 780 nm, which had been measured by UV-VIS-NIR spectrometer (JASCO; V-770ST). In the experiments, since we have compensated the spherical aberration, the beam diameter was obtained to be approximately 1.2 μm by the resolution equation. Since a typical laser energy before objective lens was to be 10 μJ, a typical laser fluence was calculated to be 5.7 × 102 J/cm2. In the calculation, the absorption coefficient of 4H-SiC at 780 nm was set to be 26 cm−1 [18]. We have also considered the surface loss due to the high refractive index of 4H-SiC. The peak power densities at τp of 220 fs and 6 ps were estimated to be 2.6 × 1015 W/cm2 and 9.5 × 1013 W/cm2, respectively. The laser-written tracks at a spacing of 25 μm in the SiC wafer were typically formed by scanning the wafer relative to the focus at 100 μm/s. After laser writing, the modified structures were inspected by optical microscope. The modified regions were located about 260 μm depth from surface. To characterize the modified structures, the samples were fractured in the plane perpendicular to the laser-written tracks and the Raman scattering spectra from the exposed modified-structures were observed by a confocal micro-Raman spectrometer with wavelength of 532 nm (Tokyo Instruments; Nanofinder). To reveal the detailed structural changes by the laser writing, the observation by using a high resolution transmission electron microscope (HR-TEM, JEOL; JEM-2200FS) was performed. Furthermore, to test the exfoliation of the laser-processed SiC wafer, after applying the instant glue on the sample surfaces, the samples were clamped by the SUS locking jig. By using a universal testing machine (Instron; 1122), a tensile stress was loaded at the normal direction relative to the sample surface. After the exfoliation test, the roughness of the separated surfaces were analyzed by surface profiler (KLA-Tencor; Alpha Step IQ).

 figure: Fig. 1

Fig. 1 (a) Schematic of the experimental procedure. (b) Optical microscope image of the laser writing region (yellow dotted area) viewed from the top surface. Optical microscope images of the exfoliated (c) upper and (d) lower surfaces viewed from the cross-sectional direction. The blue arrows show the exfoliated surfaces. The red Symbol of kw shows the laser propagation direction.

Download Full Size | PDF

3. Result and discussion

3.1 Threshold for SiC wafer internal laser-processing

To optimize the laser-processing condition, we investigated the damage threshold by using single-pulse trains with different pulse energy (Ep) and scanning speed. The pulse duration (τp) was fixed to be 220 fs. For the application to the wafer slicing, it would be preferable that the scanning speed is faster without the restriction in the depth direction. While Okada et al. reported that the structural modification inside 4H-SiC with femtosecond pulses (130 fs, 800 nm, 1 kHz) can be performed by using high NA oil immersion objective lens (100 × , NA 1.35) [15]. We determined the laser writing conditions for directly processing only the interior of SiC wafer by using lower NA 0.80 objective without damaging its surface. In the case of τp = 220 fs, the process window for the interior structural modification was very narrow and simultaneously accompanied with damage on the surface (pulse energy (Ep): 10 ~13 μJ, scanning speed: ≤ 100 μm/s). To enhance the light-matter interaction in longer time, we extended pulse duration as shown in Fig. 2(a) and 2(b). In our case, the black-colored structural modifications were observed in the case of the longer τp than 750 fs for the scanning speeds up to 2000 μm/s.

 figure: Fig. 2

Fig. 2 (a) Cross-sectional observations of the black-colored structural modifications inside 4H-SiC induced by the ultrashort pulse laser with the different pulse durations. Symbols of kw, and E show the laser propagation direction and polarization direction, respectively. (b) The size of the photoinduced structures in 4H-SiC as a function of the pulse duration. The Ep was fixed to be 10 μJ.

Download Full Size | PDF

Tomita et al. have obtained the time constant of 0.63 ps in 4H-SiC based on the transient absorption experiments [19]. They also ascribed this decay time to the inter-conduction band electron-phonon scattering time. While, it is known that the Shockley-Read-Hall (SRH) recombination is dominant for an indirect bandgap semiconductor [20, 21]. In the case of 4H-SiC, the bulk SRH lifetime was obtained to be ~800 ns [22]. Considering the relaxation of the photoexcited electrons, the thermalization by electron-electron scattering and the relaxation of the intra-band electron-phonon scattering become dominant process in the case of the pulse duration larger than the decay time of the inter-band phonon assisted transition. In the case of τp = 1 ps, the laser-energy range for the interior structural modification was from 8 μJ to 13 μJ. Figure 2(a) shows the optical microscope images of the cross-sections of the laser written lines inside SiC wafer with a fixed pulse energy of 10 μJ, and Fig. 2(b) shows the height and width of the cross-sections plotted against the pulse duration. Although the heights of modified regions were increased with increasing pulse duration, their widths were saturated to about 25 μm at 2 ps. Furthermore, in the case of longer pulse duration than 3 ps, the modified regions were fragmented in spite of the correction of spherical aberration. The critical power (Pcr) for the self-focusing can be estimated the following equation [23]:

Pcr=3.77λ28πn0n2
where λ is the laser wavelength, n0 and n2 ( = 2.7 × 1018 m2/W [24]) are the linear and nonlinear Kerr refractive index, respectively. The typical critical power density in the experiments was estimated to be 1.2 × 1012 W/cm2. The calculated critical power density for self-focusing was about one-tenth of the peak power density (~1.9 × 1014 W/cm2) at τp = 3 ps. Since the pulse duration is larger than the phonon assisted transition, the major part of the laser pulse can sufficiently interact with the trapped electrons during the indirect recombination via the recombination center. Therefore it is supposed that such fragmented modifications extended along with the laser propagation direction could be formed. Furthermore it is known that for longer pulse duration, self-focusing can occur even for powers lower than its critical power [25]. Although other possible mechanisms of the filamentation and the thermal lens effect should be also considered, detailed investigations should be required. To improve the narrow process window for the interior structural modification in SiC wafer, we employed the femtosecond double-pulse configuration with the pulse duration (τp) of 220 fs, because it is possible to elongate the interaction time between the excited electrons and the photon [16]. First, we have observed the shape and the laser-induced damage probabilities of the structurally modified regions inside SiC wafer by changing the total pulse energy as shown in Fig. 3. Although the probability of laser writing is stable, the heights were fluctuated. This is probably due to the defects such as “micro-pipe” in SiC crystal.

 figure: Fig. 3

Fig. 3 The size of widths, heights and the laser-induced damage probabilities of the photoinduced structures inside 4H-SiC as a function of the total pulse energy. The τdelay between femtosecond double pulse (τp = 220 fs) was 2 ps.

Download Full Size | PDF

The total pulse energy of the equally divided double-pulses with a delay time (τdelay) of 2 ps was changed from 1 to 50 μJ (i.e., each pulse energy was tuned from 0.5 to 25 μJ). The scanning speed was fixed to be 100 μm/s to prevent the discontinuous structural formation. Interestingly, although no apparent photoinduced structure inside 4H-SiC was observed by using single-pulse train with τp < 750 fs as shown in Fig. 2(a), the interior structural modifications were successfully induced by the femtosecond double-pulse train with τp = 220 fs at certain conditions. We have also confirmed that the damage threshold inside 4H-SiC was about 6 μJ (i.e., each pulse energy was 3 μJ) for double-pulse train with τdelay = 2 ps. In the case of femtosecond double pulse with τp = 220 fs, the total laser-energy range for the interior structural modification was from 6 μJ to 26 μJ (i.e., each pulse energy was from 3 μJ to 13 μJ), which is about four times wider than that by single pulse irradiation. Unexpectedly, the height of the photoinduced structures did not increase with increasing total pulse energy and saturated to be approximately 15 μm. We assumed that the photoinduced structures without elongation in the laser propagation direction was caused by the blacking structure, which linking to the reduction of nonlinearities, especially plasma defocusing. The slight increase of the widths and the probabilities of modifications with increasing total pulse energy can be explained by linear absorption process by the blackened structure.

We have investigated the shape change of the modified regions by tuning the τdelay between femtosecond double pulses as shown in Fig. 4(a). It should be noted that no apparent structural change can be observed at zero time delay. We assume that this phenomenon is probably derived from the interference of the pulses each other. Interestingly, the modified structures did not extend along with the laser propagation direction. Both of height and width of the modified regions were saturated to be 25 μm according to the longer τdelay as shown in Fig. 4(b). The modified regions were splitted in the case of longer pulse duration than 20 ps. This is probably due to the self-focusing. Interestingly, the height of the modified regions induced by the femtosecond double-pulse trains (τp = 220 fs) with τdelay of 2 ps was about one-third compared to the case of the single pulse trains (τp = 2 ps). We speculated that the structure induced by the first pulse can significantly absorb the second pulse energy, in the case of the femtosecond double-pulse trains, leading to prevent the elongation along with the laser propagation.

 figure: Fig. 4

Fig. 4 (a) Cross-sectional observations of the black-colored structural modifications inside 4H-SiC induced by the femtosecond double-pulses (τp = 220 fs). Yellow arrows point cracks. (b) The size of the photoinduced structures inside 4H-SiC as a function of the τdelay of double-pulse trains with the total pulse energy of 20 μJ.

Download Full Size | PDF

3.2 Characterization of modified structures

To characterize the photoinduced structure embedded in 4H-SiC, we observed Raman spectra before and after the irradiation of femtosecond double pulses with τdelay of 2 ps as shown in Fig. 5(a). The measurements were performed on the cross-sections (10-10) including the laser-written tracks. The Raman major peaks of the initial 4H-SiC wafer [the blue line in Fig. 5(a)] are consistent to those observed in previous studies [26, 27]. Based on the symmetry point group C6v for 4H polytype, the peaks at 206, 610, 778, 799, and 995 cm−1 can be assigned to E2 planar or transverse acoustic, A1 axial or longitudinal acoustic, E2 planar optical, E1 planar optical, and A1 longitudinal optical modes, respectively. The peaks at the wavenumber higher than 1450 cm−1 can be also assigned to the optical branch of the second-order Raman spectrum [28]. Since it is known that the frequency of longitudinal optical (LO) mode at around 995 cm−1 is influenced by the impurities (free electron) and the stress in SiC crystal, we evaluated the carrier density (n) in the initial SiC wafer by using simple equation [29]:

n[cm-3]=1.25×1017Δω
where Δω is the differential frequency of the LO mode between the sample and the pure SiC (964.1 cm−1). The carrier density in our sample was estimated to be 3.98 × 1018 cm−3. After the laser irradiation, the main peak of the folded transverse optical (FTO) band at 778 cm−1 was slightly broadened (the red line in Fig. 5(a)). This is probably due to stacking faults and/or residual strain [30, 31]. Noticeable spectral changes at two wavenumber regions at around 480 cm−1 and ranging from 1400 to 1600 cm−1 were observed in the photoinduced regions. The broad peak Raman band centered at 480 cm−1 is attributed to the amorphous silicon. In addition, the Raman signals from amorphous carbon were also observed as broad bands in the frequency region of 1400−1600 cm−1. In this frequency region, two Raman peaks are normally observed in disordered carbon structures; G and D peaks at 1580 and 1350 cm−1, respectively. The G peak is assigned to the optical zone center phonons of E2g symmetry involving the in-plane bond-stretching motion of all pairs of sp2 sites. The D peak is assigned to the K-point phonons of A1g breathing mode of C sp2 atoms in rings [32]. Furthermore, the band around 1490 cm−1 is interpreted as an amorphous carbon contributions of C-H vibrations [33]. Therefore, these changes in the Raman spectra in 4H-SiC after laser irradiation indicate the transformation from 4H-SiC to other phases and deformation during laser irradiation. Similar Raman spectra changes have been observed in the scratched surface of the 4H-SiC [32].

 figure: Fig. 5

Fig. 5 (a) Typical micro-Raman spectra before and after laser irradiation. Raman maps of the FTO peak on the cross-sectional surface including the laser-written tracks by the single- (upper row) and the double- (lower row) pulse trains. (b, e) intensity, (c, f) FWHM, and (d, g) peak position. The laser writing conditions by the single-pulse trains (b-d): Ep = 10 μJ, τp = 1 ps. The laser writing conditions by the double-pulse trains: Ep = 10 + 10 μJ, τp = 220 fs, τdelay = 2 ps.

Download Full Size | PDF

We have also carried out Raman mapping of the FTO peak on the cross-sectional surface including the laser-written tracks by the single- in Fig. 5(b-d) and the double-pulse trains in Fig. 5(e-g). It should be noted that there were no apparent differences between Raman spectra of modified region by single-pulse trains of τp = 1ps and that by double-pulse trains of τdelay = 1ps. For both of pulse trains, the FTO peaks in the center of the focal spot location became smaller and broader corresponding to Fig. 5(b,c) and 5(e,f). The change of the FTO peak indicates the deterioration in crystallinity of SiC. From the shift of the FTO peak, the residual stress can be evaluated according to the study by Liu et al., which revealed that the ratio of Raman shift to residual stress was 3.1 cm−1/GPa for hexagonal SiC [34]. The higher shift along the laser propagation axis as shown in Fig. 5(d,g) means that compressive stress was generated along the laser propagation direction with the maximum residual stress wave of 0.25 GPa exclude the irradiated center. The stress distributions along the laser propagation axis are similar to the previous study [15]. On the other hand, a tensile stress was also observed in the both sides of the photoinduced structures with the maximum residual stress of 0.25 GPa. We assumed that such stress distribution is due to the flat-shaped modification along c plane, which is preferable for the wafer slicing. To investigate the detailed structure of the black-colored structural modifications, we carried out the HR-TEM observations as shown in Fig. 6. FFT images at typical regions are also shown. The photoinduced structures by the single-pulse trains were consist of the brighter regions of the SiC crystal [the region “A” in Fig. 6(a) and the corresponding FFT image was shown in Fig. 6(c)] and the darker regions of the amorphous phase [the region “B” in Fig. 6(a) and the corresponding FFT image was shown in Fig. 6(d)], which are similar to the previous report [15]. On the other hand, in the case of double-pulse trains, although the regions of the initial SiC crystal arising from the discontinuous modification were observed [the region “D” in Fig. 6(b) and the corresponding FFT image was shown in Fig. 6(f)], the typical photoinduced structures were polycrystalline [the region “C” in Fig. 6(b) and the corresponding FFT image was shown in Fig. 6(e)].

 figure: Fig. 6

Fig. 6 HR-TEM images of the laser-written tracks induced by the single- (a) and the double- (b) pulse trains. FFT images at the regions of A (c), B (d), C (e), and D (f) are also shown. The laser writing conditions by the single-pulse trains: Ep = 10 μJ, τp = 1 ps. The laser writing conditions by the double-pulse trains: Ep = 5 + 5 μJ, τp = 220 fs, τdelay = 2 ps.

Download Full Size | PDF

3.3 Tensile test after the laser writing

After writing of the laser tracks with a pitch of 25 μm over the entire sample, the sample was subjected to tensile stress normal to the surface by universal tensile testing machine. The total pulse energies for the single- and double-pulses train were set to the same of 10 μJ. To perform the tensile test, the both sides of sample surfaces were attached strongly to two stainless plates with cyanoacrylate adhesive.

Figure 7(a) shows the tensile stresses and strains at which samples were separated completely [The SiC sample after separation is shown in Fig. 7(b, c)]. This graph clearly shows that the tensile stress required for separation of the samples is much smaller for the double-pulses than those for the single-pulses. It should be noted that there are the two points of the separation stress around 1 MPa in double-pulse experiments were smaller than the lower limit value of the machine (< 1 MPa). This easy separation is probably due to the flat-shaped modification along c plane and the crack formation along with <11-20> directions.

 figure: Fig. 7

Fig. 7 (a) Plots of tensile stress as a function of strain. (b) A photograph of the exfoliated sample surfaces after laser slicing by double-pulse. (c) Optical microscope image of the exfoliated surface after the laser writing by double pulse configuration. The laser writing conditions by the single-pulse trains: Ep = 10 μJ, τp = 1 ps. The laser writing conditions by the double-pulse trains: Ep = 5 + 5 μJ, τp = 220 fs, and τdelay = 1 ps.

Download Full Size | PDF

The separated sample surfaces were also observed by using optical microscope and surface profiler as shown in Fig. 8. The surface profiles were measured on the regions surrounded by the dashed yellow lines. Optical microscope images of the exfoliated surfaces viewed from the cross-sectional direction after the laser writing by single- and double-pulse trains are also shown in Fig. 8(b), (c). The profile of the separated surface by the single-pulse trains was irregular and the roughness in a wider range was up to 24 μm as shown in Fig. 8(a) and the blue line in profile. This roughness is probably due to the elongation of the structural changes along the laser propagation direction and/or the discontinuous structural modification as shown in Fig. 2. On the other hand, the separated surfaces induced by the double-pulse trains show a relatively smooth and regular morphology, which became smaller than 5 μm. The roughness of 5 μm is obviously smaller than the kerf-loss for the conventional wire-saw method. This laser slicing technique by the femtosecond double pulse can be applied not only SiC wafer but also other precious and hard materials such as diamond and AlN.

 figure: Fig. 8

Fig. 8 (a) The surface profiles on the exfoliated surfaces after the laser writing with the single- and double-pulse trains. Optical microscope images of the exfoliated surfaces viewed from the cross-sectional direction after the laser writing by (b) single- and (c) double-pulse trains are shown. The blue arrows show the exfoliated surfaces. The blue Symbol of kw shows the laser propagation direction. Optical microscope images of the exfoliated surfaces for (d) single- and (e) double-pulse trains are also shown. The laser writing conditions by the single-pulse trains: Ep = 10 μJ, τp = 1 ps. The laser writing conditions by the double-pulse trains: Ep = 5 + 5 μJ, τp = 220 fs, and τdelay = 1 ps.

Download Full Size | PDF

4. Conclusion

We have demonstrated that the thin SiC could be successfully sliced off from a wafer by focusing an ultrafast pulse laser with double pulse configuration. The structural changes are caused by the transformation to amorphous SiC and the decomposition to amorphous Si and C. The flat-shaped modification along c plane was photoinduced by the femtosecond double-pulses train. The separation force for double-pulse experiments was significantly lower than that for single-pulse experiments. It was shown that the kerf-loss for the femtosecond laser slicing method was much smaller than that for the conventional wire-saw method. It is expected that this technique cannot only open the door to new applications of the femtosecond laser processing but also provide the integrated processes between slicing and dicing by using the femtosecond light source.

Funding

JSPS KAKENHI (No. 16K13929), Cross-Ministerial Strategic Innovation Promotion (SIP) Program.

References and links

1. J. Ikeno, Y. Tani, A. Fukutani, and H. Sato, “Development of chipping-free dicing technology applying electrophoretic deposition of ultrafine abrasives,” Ann. CIRP 40(1), 351–354 (1991). [CrossRef]  

2. I. H. Cho, S. C. Jeong, J. M. Park, and H. D. Jeong, “The application of micro-groove machining for the model of PDP barrier ribs,” J. Mater. Process. Technol. 113(1), 355–359 (2001). [CrossRef]  

3. W. Peng, X. F. Xu, and L. F. Zhang, “Improvement of a dicing blade using a whisker direction-controlled by an electric field,” J. Mater. Process. Technol. 129(1), 377–379 (2002). [CrossRef]  

4. S. B. Lee, Y. Tani, T. Enomoto, and H. Sato, “Development of a dicing blade with photopolymerizable resins for improving machinability,” Ann. CIRP 54(1), 293–296 (2005). [CrossRef]  

5. W. Koch, A. L. Endrös, D. Franke, C. Häßler, J. P. Kalejs, and H. J. Möller, Handbook of Photovoltaic Science and Engineering, A. Luque and S. Hegedus, eds. (Wiley, 2002), Chap. 6.4.

6. L. Di Cioccio, F. Letertre, Y. Le Tiec, A. M. Papon, C. Jaussaud, and M. Bruel, “Silicon carbide on insulator formation by the Smart-Cut®process,” Mater. Sci. Eng. B 46(1), 349–356 (1997). [CrossRef]  

7. J. H. Lee, I. Bargatin, J. Park, K. M. Milaninia, L. S. Theogarajan, R. Sinclair, and R. T. Howe, “Smart-cut layer transfer of single-crystal SiC using spin-on-glass,” J. Vac. Sci. Technol. B 30(4), 042001 (2012). [CrossRef]  

8. V. P. Amarasinghe, L. Wielunski, A. Barcz, L. C. Feldman, and G. K. Celler, “Properties of H+ Implanted 4H-SiC as Related to Exfoliation of Thin Crystalline Films,” ECS J. Solid State Sci. Technol. 3(3), 37–42 (2014). [CrossRef]  

9. F. Fukuyo, K. Fukumitsu, and N. Uchiyama, “The stealth dicing technologies and their application” in Proceeding of 6th International Symposium on Laser Precision Microfabrication, #131, Williamsburg, USA, 4–8 Apr. 2005.

10. E. Ohmura, F. Fukuyo, K. Fukumitsu, and H. Morita, “Internal modified-layer formation mechanism into silicon with nanosecond laser,” JAMME 17(1), 381–384 (2006).

11. K. M. Davis, K. Miura, N. Sugimoto, and K. Hirao, “Writing waveguides in glass with a femtosecond laser,” Opt. Lett. 21(21), 1729–1731 (1996). [CrossRef]   [PubMed]  

12. Y. Shimotsuma, P. G. Kazansky, J. Qiu, and K. Hirao, “Self-organized nanogratings in glass irradiated by ultrashort light pulses,” Phys. Rev. Lett. 91(24), 247405 (2003). [CrossRef]   [PubMed]  

13. M. Mori, Y. Shimotsuma, T. Sei, M. Sakakura, K. Miura, and H. Udono, “Tailoring thermoelectric properties of nanostructured crystal silicon fabricated by infrared femtosecond laser direct writing,” Phys. Status Solidi., A Appl. Mater. Sci. 212(4), 715–721 (2015). [CrossRef]  

14. M. Shimizu, Y. Shimotsuma, M. Sakakura, T. Yuasa, H. Homma, Y. Minowa, K. Tanaka, K. Miura, and K. Hirao, “Periodic metallo-dielectric structure in diamond,” Opt. Express 17(1), 46–54 (2009). [CrossRef]   [PubMed]  

15. T. Okada, T. Tomita, S. Matsuo, S. Hashimoto, Y. Ishida, S. Kiyama, and T. Takahashi, “Formation of periodic strained layers associated with nanovoids inside a silicon carbide single crystal induced by femtosecond laser irradiation,” J. Appl. Phys. 106(5), 054307 (2009). [CrossRef]  

16. K. Sugioka, M. Iida, H. Takai, and K. Micorikawa, “Efficient microwelding of glass substrates by ultrafast laser irradiation using a double-pulse train,” Opt. Lett. 36(14), 2734–2736 (2011). [CrossRef]   [PubMed]  

17. E. Kim, Y. Shimotsuma, M. Sakakura, and K. Miura, “Ultrashort pulse laser slicing of semiconductor crystal,” Proc. SPIE 9983, 99831B (2016). [CrossRef]  

18. E. Biedermann, “The optical absorption bands and their anisotropy in the various modifications of SiC,” Solid State Commun. 3(10), 343–346 (1965). [CrossRef]  

19. T. Tomita, S. Saito, T. Suemoto, H. Harima, and S. Nakashima, “Structure dependent ultrafast relaxation time of photo-excited carriers in SiC,” J. Phys. Soc. Jpn. 73(9), 2554–2561 (2004). [CrossRef]  

20. R. N. Hall, “Electron-hole recombination in germanium,” Phys. Rev. 87(2), 387 (1952). [CrossRef]  

21. W. Shockley and W. T. Read Jr., “Statistics of the recombinations of holes and electrons,” Phys. Rev. 87(5), 835–842 (1952). [CrossRef]  

22. P. Ščajev, V. Gudelis, K. Jarašiūnas, and P. B. Klein, “Fast and slow carrier recombination transients in highly excited 4H– and 3C–SiC crystals at room temperature,” J. Appl. Phys. 108(2), 023705 (2010). [CrossRef]  

23. J. H. Marburger, “Self-focusing: Theory,” Prog. Quantum Electron. 4(1), 35–110 (1975). [CrossRef]  

24. F. De Leonardis, R. A. Soref, and V. M. N. Passaro, “Dispersion of nonresonant third-order nonlinearities in Silicon Carbide,” Sci. Rep. 7, 40924 (2017). [CrossRef]   [PubMed]  

25. J. Faure, V. Malka, J. R. Marqués, P. G. David, F. Amiranoff, K. Ta Phuoc, and A. Rousse, “Effects of pulse duration on self-focusing of ultra-short lasers in underdense plasmas,” Phys. Plasmas 9(3), 756–759 (2002). [CrossRef]  

26. D. W. Feldman, J. H. Parker Jr, W. J. Choyke, and L. Patrick, “Phonon dispersion curves by Raman scattering in SiC, polytypes 3C, 4H, 6H, 15R, and 21R,” Phys. Rev. 173(3), 787–793 (1968). [CrossRef]  

27. J. C. Burton, L. Sun, M. Pophristic, S. J. Lukacs, F. H. Long, Z. C. Feng, and I. T. Ferguson, “Spatial characterization of doped SiC wafers by Raman spectroscopy,” J. Appl. Phys. 84(11), 6268–6273 (1998). [CrossRef]  

28. J. C. Burton, L. Sun, F. H. Long, Z. C. Feng, and I. T. Ferguson, “First- and second-order Raman scattering from semi-insulating 4H−SiC,” Phys. Rev. B 59(11), 7282–7284 (1999). [CrossRef]  

29. S. Nakashima, T. Kitamura, T. Mitani, H. Okumura, M. Katsuno, and N. Ohtani, “Raman scattering study of carrier-transport and phonon properties of 4H−SiC crystals with graded doping,” Phys. Rev. B 76(24), 245208 (2007). [CrossRef]  

30. S. Nakashima, T. Mitani, M. Tomobe, T. Kato, and H. Okumura, “Raman characterization of damaged layers of 4H-SiC induced by scratching,” AIP Adv. 6(1), 015207 (2016). [CrossRef]  

31. S. Nakashima, Y. Nakatake, H. Harima, M. Katsuo, and N. Ohtani, “Detection of stacking faults in 6H-SiC by Raman scattering,” Appl. Phys. Lett. 77(22), 3612–3614 (2000). [CrossRef]  

32. A. C. Ferrari and J. Robertson, “Interpretation of Raman spectra of disordered and amorphous carbon,” Phys. Rev. B 61(20), 14095–14107 (2000). [CrossRef]  

33. J. Schwan, S. Ulrich, V. Batori, H. Ehrhardt, and S. R. P. Silva, “Raman spectroscopy on amorphous carbon films,” J. Appl. Phys. 80(1), 440–447 (1996). [CrossRef]  

34. J. Liu and Y. K. Vohra, “Raman modes of 6H polytype of silicon carbide to ultrahigh pressures: A comparison with silicon and diamond,” Phys. Rev. Lett. 72(26), 4105–4108 (1994). [CrossRef]   [PubMed]  

Cited By

Optica participates in Crossref's Cited-By Linking service. Citing articles from Optica Publishing Group journals and other participating publishers are listed here.

Alert me when this article is cited.


Figures (8)

Fig. 1
Fig. 1 (a) Schematic of the experimental procedure. (b) Optical microscope image of the laser writing region (yellow dotted area) viewed from the top surface. Optical microscope images of the exfoliated (c) upper and (d) lower surfaces viewed from the cross-sectional direction. The blue arrows show the exfoliated surfaces. The red Symbol of kw shows the laser propagation direction.
Fig. 2
Fig. 2 (a) Cross-sectional observations of the black-colored structural modifications inside 4H-SiC induced by the ultrashort pulse laser with the different pulse durations. Symbols of kw, and E show the laser propagation direction and polarization direction, respectively. (b) The size of the photoinduced structures in 4H-SiC as a function of the pulse duration. The Ep was fixed to be 10 μJ.
Fig. 3
Fig. 3 The size of widths, heights and the laser-induced damage probabilities of the photoinduced structures inside 4H-SiC as a function of the total pulse energy. The τdelay between femtosecond double pulse (τp = 220 fs) was 2 ps.
Fig. 4
Fig. 4 (a) Cross-sectional observations of the black-colored structural modifications inside 4H-SiC induced by the femtosecond double-pulses (τp = 220 fs). Yellow arrows point cracks. (b) The size of the photoinduced structures inside 4H-SiC as a function of the τdelay of double-pulse trains with the total pulse energy of 20 μJ.
Fig. 5
Fig. 5 (a) Typical micro-Raman spectra before and after laser irradiation. Raman maps of the FTO peak on the cross-sectional surface including the laser-written tracks by the single- (upper row) and the double- (lower row) pulse trains. (b, e) intensity, (c, f) FWHM, and (d, g) peak position. The laser writing conditions by the single-pulse trains (b-d): Ep = 10 μJ, τp = 1 ps. The laser writing conditions by the double-pulse trains: Ep = 10 + 10 μJ, τp = 220 fs, τdelay = 2 ps.
Fig. 6
Fig. 6 HR-TEM images of the laser-written tracks induced by the single- (a) and the double- (b) pulse trains. FFT images at the regions of A (c), B (d), C (e), and D (f) are also shown. The laser writing conditions by the single-pulse trains: Ep = 10 μJ, τp = 1 ps. The laser writing conditions by the double-pulse trains: Ep = 5 + 5 μJ, τp = 220 fs, τdelay = 2 ps.
Fig. 7
Fig. 7 (a) Plots of tensile stress as a function of strain. (b) A photograph of the exfoliated sample surfaces after laser slicing by double-pulse. (c) Optical microscope image of the exfoliated surface after the laser writing by double pulse configuration. The laser writing conditions by the single-pulse trains: Ep = 10 μJ, τp = 1 ps. The laser writing conditions by the double-pulse trains: Ep = 5 + 5 μJ, τp = 220 fs, and τdelay = 1 ps.
Fig. 8
Fig. 8 (a) The surface profiles on the exfoliated surfaces after the laser writing with the single- and double-pulse trains. Optical microscope images of the exfoliated surfaces viewed from the cross-sectional direction after the laser writing by (b) single- and (c) double-pulse trains are shown. The blue arrows show the exfoliated surfaces. The blue Symbol of kw shows the laser propagation direction. Optical microscope images of the exfoliated surfaces for (d) single- and (e) double-pulse trains are also shown. The laser writing conditions by the single-pulse trains: Ep = 10 μJ, τp = 1 ps. The laser writing conditions by the double-pulse trains: Ep = 5 + 5 μJ, τp = 220 fs, and τdelay = 1 ps.

Equations (2)

Equations on this page are rendered with MathJax. Learn more.

P cr = 3.77 λ 2 8π n 0 n 2
n [cm -3 ]=1.25× 10 17 Δω
Select as filters


Select Topics Cancel
© Copyright 2024 | Optica Publishing Group. All rights reserved, including rights for text and data mining and training of artificial technologies or similar technologies.