Nanoscale Material Removal Mechanism of Soft-Brittle ... - Springer Link

3 downloads 27070 Views 553KB Size Report
Feb 12, 2012 - У Springer Science+Business Media, LLC 2012. Abstract Damage-free subsurfaces of soft-brittle HgCdTe. (MCT) single crystals were directly ...
Tribol Lett (2012) 46:95–100 DOI 10.1007/s11249-012-9924-9

ORIGINAL PAPER

Nanoscale Material Removal Mechanism of Soft-Brittle HgCdTe Single Crystals Under Nanogrinding by Ultrafine Diamond Grits Zhenyu Zhang • Yaxing Song • Fengwei Huo Dongming Guo



Received: 14 January 2012 / Accepted: 27 January 2012 / Published online: 12 February 2012 Ó Springer Science+Business Media, LLC 2012

Abstract Damage-free subsurfaces of soft-brittle HgCdTe (MCT) single crystals were directly achieved after nanogrinding by a developed ultrafine diamond wheel. This is different from those of hard-brittle semiconductors, where there is usually a damaged layer found after mechanical machining. Two chips induced by nanogrinding with thicknesses varying from 23 to 27.1 nm attached on the ground MCT surface were observed, which is consistent well with a proposed model of chip thickness. Nanoscale material removal mechanism was investigated using high resolution transmission electron microscopy. Twins and nanocrystals were observed within the chips found. Keywords Non-ferrous alloys  Grinding  SEM  TEM  Wear mechanisms

1 Introduction An amorphous damaged layer on the topmost layer of hardbrittle semiconductors, such as silicon (Si), germanium (Ge), and gallium arsenide (GaAs), is usually found after mechanical machining [1–4]. Nevertheless, a crystalline damaged layer of soft-brittle semiconductors after nanogrinding was observed in our recent work [5]. As material removal mechanism governs the damaged layer left after mechanical machining, it aims to discover a distinct material removal mechanism of soft-brittle semiconductors from that of hard-brittle ones. We do find a tendency that

Z. Zhang (&)  Y. Song  F. Huo  D. Guo Key Laboratory for Precision and Non-Traditional Machining Technology, Ministry of Education, Dalian University of Technology, Dalian 116024, People’s Republic of China e-mail: [email protected]

the finer of grits involved in diamond wheel, the smoother of the ground surface of soft-brittle semiconductors [6]. It is therefore a damage-free subsurface is expected with the finer of grits employed in nanogrinding. Nanogrinding is used to remove the damaged layer of soft-brittle semiconductors by slicing from an ingot, such as mercury cadmium telluride (HgCdTe or MCT) [7]. MCT is a representative of soft-brittle semiconductors, and is widely used in the infrared detectors [7, 8]. However, little has been reported on the material removal mechanism involved in mechanical machining [7, 8]. A chip is essential to investigate the material removal mechanism, while it is difficult to obtain [9]. As a result, different models were built to predict the chip thickness involved in grinding [9]. For example, Snoeys and Peters proposed a model based on continuous equation [10]. Reichenbach and Mayer suggested a chip thickness model, accounting for the two dimensional shape of diamond wheel [11]. Malkin deduced a detailed chip thickness model according to the geometric shape of diamond wheel [12]. These models have significantly improved our understanding on the chip thickness, both technologically and scientifically. However, none of these models are verified, due to the absence of direct experimental evidence. For the ultrafine resin bond diamond wheel, it is easy to induce burn damage on ground specimens, due to the low thermal conductivity and holding force of resin materials [13]. Metal bond diamond wheel generated higher normal grinding force [14], easily resulting in chipping and cracking on the ground MCT specimens. To overcome the disadvantages of resin and metal bond ultrafine diamond wheels, a novel ultrafine ceramic bond diamond wheel is necessary to develop. In this study, we report our recent results on direct experimental evidence of chips, as well as material

123

96

Tribol Lett (2012) 46:95–100

removal mechanism of soft-brittle MCT semiconductors involved in nanogrinding, by a developed ultrafine ceramic bond diamond wheel.

2 Experimental Details As-received Hg0.22Cd0.78Te (111) single crystal wafers were grown by modified Bridgman method [9]. The wafer is 10-mm in diameter, and 1-mm in thickness. The grit mesh size is 15,000, equivalent to a grit size of 700 nm. The diamond concentration is 200, equivalent to a volume fraction of 50%. The ceramic bond consisted of silicon carbide (SiC), silica (SiO2), sodium chloride (NaCl), and sodium bicarbonate (NaHCO3). The diamond wheel consisted of 56 blocks. Each block was 18-mm long, 3-mm wide, and 5-mm thick. The coolant used was deionized water. Nanogrinding was conducted on an Okamoto ultraprecision grinder equipped with air spindles. Preliminary nanogrinding tests were performed to achieve a better surface finish. The optimal grinding conditions are listed in Table 1. Samples A and B were designated as two different kinds of nanogrinding conditions. For each nanogrinding condition, three tests were conducted. Surface roughness (Ra) and peak-to-valley (PV) were measured by a high resolution non-contact ZYGO surface profiler. The scanning area was 70 9 50 lm2. For each measurement, five random locations were recorded, and the average was used as Ra. The Ra of samples A and B is 1.7 and 1.6 nm, respectively, as shown in Table 1. Surface topography and

composition were characterized by a field emission environmental scanning electron microscopy (SEM, Quanta 200 FEG, FEI, Netherlands) equipped with energy dispersive spectroscopy (EDS). Cross-sectional transmission electron microscopy (TEM) samples were prepared using a tripod technique [15], followed by ion beam thinning in a Gatan precision ion polishing system with a low energy level of 2.5 eV and an ion beam tilt of 4° to avoid any damage on the soft-brittle MCT specimens. Tripod technique is advanced than conventional ion milling. Ion milling usually has a detrimental effect on TEM specimens, such as permanent damage resulting amorphously artificial surface on the thinned area, which can be effectively avoided by tripod technique [15]. TEM examinations were carried out on a FEI Tecnai G2 F30 S-twin high resolution TEM operated at 300 kV. To achieve the chip attached on the ground MCT surface, a distinct nanogrinding test was conducted on sample B. The distinct test was as follows: the diamond wheel was sharply stopped during feeding, and then receded immediately to achieve a chip attached on the ground surface at a more possibility.

3 Results Figure 1 shows the SEM image of developed ultrafine ceramic bond diamond wheel and its corresponding EDS spectrum. The surface has uniformly porous microstructure, as shown in Fig. 1a, indicating a relatively high porosity was achieved in this diamond wheel, regardless of

Table 1 Nanogrinding conditions, surface roughness, and maximum undeformed chip thickness Sample

Wheel speed (rpm)

Table speed (rpm)

Feed rate of wheel (lm/min)

Surface roughness

Maximum undeformed chip thickness (nm)

Ra (nm)

PV (nm)

A

2,000

120

1

1.7

13.6

18.1

B

1,800

80

8

1.6

12.4

26.2

Fig. 1 Surface a SEM image of ultrafine ceramic bond diamond wheel and corresponding. b EDS spectrum of the white square marked in (a)

123

Tribol Lett (2012) 46:95–100

97

Fig. 2 Surface a SEM image of sample A after nanogrinding and corresponding. b EDS spectrum of the white square marked in (a)

use of ultrafine diamond grits. EDS spectrum consists of C, O, Na, Si elements, derived from diamond grits, SiC, SiO2, NaCl, and NaHCO3. Figure 2 shows the surface SEM image of sample A after nanogrinding and its corresponding EDS spectrum. The surface of sample A is bright (not shown here), free of burn damage, crack, and imbedding. The EDS confirms that the O element is absent from the collected peaks, indicating free of burn damage. Figure 3 shows the cross-sectional TEM images of samples A and B at low and high magnifications. Damagefree subsurface of MCT is achieved, as well as the perfect selected area electron diffraction (SAED) patterns and monocrystalline lattice. This is different from those of hard-brittle semiconductors, such as Si, Ge, and GaAs, where there is usually a damaged layer found on the topmost after mechanical machining [1–4]. Figure 4 shows the cross-sectional TEM images at low and high magnifications of sample B under a distinct nanogrinding test. Two chips were observed attached on the ground surface, which is directly experimental evidence to verify the validity of a model of chip thickness. The thicknesses of two chips range from 23 to 27.1 nm, as shown in Fig. 4a. SAED shows an ordered double-dot pattern, indicating a sign of twins. Twins are observed in Fig. 4b, c. Nanocrystals are observed in Fig. 4c, d with random orientations. All the nanocrystals belong to MCT (111) plane, which is consistent with the SAED pattern, where there is no extra diffraction spots, except for twins of (111) plane.

4 Discussion SiC is a kind of ceramic material, which has high temperature strength and hardness, wear and corrosion resistance, thermal stability, and chemical inertness and

stiffness [16]. The more important is SiC as a bond material could take away the nanogrinding generating heat effectively. This is attributed to the high thermal conductivity of SiC (120 W/m–K) [17], which is 480 times as that of resin bond material (0.25 W/m–K) [13]. SiO2 and SiC could bond well [18]. The different ceramics had various volume shrinkage rates during sintering, which was beneficial for increasing porous rate within the ultrafine diamond wheel. NaCl was used to refine the grains during sintering [19]. NaHCO3 aimed to generate gas to increase porous rate [20] during sintering. As a consequence, the uniformly porous microstructure was achieved, as shown in Fig. 1a, which benefits to achieve damage-free subsurface and avoid burn damage occurring effectively. Two chips with thicknesses varying from 23 to 27.1 nm are observed in Fig. 4a, which is a direct evidence to verify and build a model to calculate the chip thickness. As the model of Malkin [12] is widely used to characterize the overall grinding conditions involved in an individual grinding event [9], it is expressed as [12]: "    #1=2 4 vw a 1=2 hm ¼ ð1Þ Cr vs de where C is the active surface grit cutting density, r is the ratio between the width and depth of an undeformed chip, vw is the workpiece velocity, vs is the wheel velocity, a is the depth of cut, de is equivalent wheel diameter. C is calculated as follows [21]: C¼

4f dg2 ð4p=3vÞ2=3

ð2Þ

where dg is the equivalent diameter of a grit, f is the surface active fraction of grits, v is the volume fraction of grits within the diamond wheel. dg is obtained from [12]:

123

98

Tribol Lett (2012) 46:95–100

Fig. 3 Cross-sectional TEM images at low magnifications of samples A (a), B (c) and corresponding high magnifications (b) and (d) taken from the black squares. The inset shows the corresponding SAED patterns taken from the black circles

dg ¼ 28 M 1:1

ð3Þ

Based on Eq. 1, the calculated maximum undeformed chip thickness is 5.1 nm, which is not consistent with the experimental result. The difference between the experimental and calculated results may be due to the absence of effects of elastic modulus for both diamond wheel and workpiece. The elastic modulus of ultrafine diamond wheel developed is 780 GPa, measured by a TriboIndenterÒ indenter. The elastic modulus of MCT is 50 GPa [8]. Taking elastic modulus into account, Eq. 1 is expressed as:  0:6 "   1=2 #1=2 E1 4 vw a hm ¼ ð4Þ Cr vs de E2

123

where E1 and E2 are the elastic modulus of the wheel and workpiece, respectively. The corrected chip thicknesses of samples A and B according to Eq. 4 are 18.1 and 26.2 nm, respectively, as shown in Table 1. The chip thickness of sample B is calculated as 26.2 nm, which is consistent well with chip thicknesses varying from 23 to 27.1 nm observed from Fig. 4a. A chip is also used to investigate the nanoscale material removal mechanism involved in nanogrinding of softbrittle MCT semiconductors. Twins in Fig. 4 are attributed to the low stacking fault energy of MCT. The stacking fault energy of MCT varies from 10 to 14 mJ m-2 [22], which is much lower than those of hard-brittle Si (69 ± 7 mJ/m2) [23], Ge (60 ± 8 mJ/m2) [24], and GaAs (48 ± 6 mJ/m2) [25] semiconductors. It has a significant effect on

Tribol Lett (2012) 46:95–100

99

Fig. 4 Cross-sectional TEM images at (a) low and (b), (c), (d) high magnifications of sample B taken from the black square marked in (a) under a distinct nanogrinding test. The inset shows the SAED pattern from the marked black circle area

deformation response and hardness of semiconductors. For example, the hardness of MCT is 0.5 GPa [23], which is also much lower than those of hard-brittle Si (12–14 GPa), Ge (10 GPa), and GaAs (6 GPa) semiconductors [26]. Slips were easy to occur for MCT under stress induced by nanogrinding, owing to the low stacking fault energy [7, 27]. Then twins generated after slips, as shown in Fig. 4b, c [7, 27]. The slips and twins could effectively relax the stress generated by nanogrinding, releasing the stress effectively. Therefore, the stress might not reach the critical value for amorphization. As the stress induced by nanogrinding was complicated, twins transformed into nanocrystals with random orientations, as shown in Fig. 4c, d. The boundaries between nanocrystals limited the further movement of slips, followed

by the broken of nanocrystals, due to the weak Hg–Te bonds [8]. They are further weakened by the addition of Cd atoms [8]. A damage-free subsurface was generated, after the nanocrystals were gone under the stress generated by nanogrinding, as shown in Fig. 4d. On the other hand, the boiling point of MCT is 250 °C [28], which is much lower than melting points of hard-brittle Si (1,410 °C), Ge (936 °C), and GaAs (1,238 °C) semiconductors [29]. MCT has not a fixed melting point, and it might melt at around 150 °C. When the chip thickness is below 1 lm, grinding generating heat is significantly increased [30], leading to the temperature increasing of ground MCT surface. This accelerated the broken of Hg–Te bonds between nanocrystals, followed by their departure, generating a damage-free subsurface.

123

100

5 Conclusions In summary, two chips with thicknesses varying from 23 to 27.1 nm attached on the ground MCT surface were observed, which is the direct experimental evidence. A model of chip thickness was proposed, which is consistent well with the experimental result. Damage-free subsurface was achieved after nanogrinding by a developed ultrafine ceramic bond diamond wheel. Burn damage free surface was achieved under nanogrinding. Twins and nanocrystals with random orientations were observed within the chips. Acknowledgments The authors are grateful for the financial supports from the National Natural Science Foundation of China (91123013), and the Science and Technology Project of Dalian City of China (2009A18GX014).

References 1. Yan, J.W., Takahashi, H., Tamaki, J., Gai, X., Kuriyagawa, T.: Transmission electron microscopic observation of nanoindentations made on ductile-machined silicon wafers. Appl. Phys. Lett. 87, 211901 (2005) 2. Zarudi, I., Nguyen, T., Zhang, L.C.: Effect of temperature and stress on plastic deformation in monocrystalline silicon induced by scratching. Appl. Phys. Lett. 86, 011922 (2005) 3. Bradby, J.E., Williams, J.S., Leung, J.W., Swain, M.V., Munroe, P.: Nanoindentation-induced deformation of Ge. Appl. Phys. Lett. 80, 2353–2651 (2002) 4. Wasmer, K., Wojtan, M.P., Gassilloud, R., Pouvreau, C., Tharian, J., Micher, J.: Plastic deformation modes of gallium arsenide in nanoindentation and nanoscratching. Appl. Phys. Lett. 90, 031902 (2007) 5. Zhang, Z.Y., Guo, D.M., Kang, R.K., Gao, H., Jin, Z.J., Meng, Y.W.: Subsurface crystal lattice deformation machined by ultraprecision grinding of soft-brittle CdZnTe crystals. Int. J. Adv. Manuf. Technol. 47, 1065–1081 (2010) 6. Zhang, Z.Y., Meng, Y.W., Guo, D.M., Wu, L.L., Tian, Y.J., Liu, R.P.: Material removal mechanism of precision grinding of softbrittle CdZnTe wafers. Int. J. Adv. Manuf. Technol. 46, 563–569 (2010) 7. Rogalski, A.: HgCdTe infrared detector material: history, status and outlook. Rep. Prog. Phys. 68, 2267–2336 (2005) 8. Martyniuk, M., Sewell, R.H., Musca, C.A., Dell, J.M., Faraone, L.: Nanoindentation of HgCdTe prepared by molecular beam epitaxy. Appl. Phys. Lett. 87, 251905 (2005) 9. Gopal, A.V., Rao, P.V.: A new chip-thickness model for performance assessment of silicon carbide grinding. Int. J. Adv. Manuf. Technol. 24, 816–820 (2004) 10. Snoeys, R., Peters, J.: Significance of chip thickness in grinding. Ann. CIRP 23, 227–237 (1974) 11. Reichenbach, G.S., Mayer, J.E., Kalpakcioglu, S., Shaw, M.C.: Role of chip thickness in grinding. Trans. ASME 78, 847–859 (1956)

123

Tribol Lett (2012) 46:95–100 12. Malkin, S.: Grinding technology, theory and applications of machining with abrasives. Ellis Horwood, Chichester (1989) 13. Tomlinson, W.J., Stapley, D.: Thermal conductivity of epoxy resin-aluminium (0–50%) composites. J. Mater. Sci. 12, 1689– 1690 (1977) 14. Inasaki, I.: Creep feed grinding with continuous dressing. Ann. CIRP 36, 227–230 (1987) 15. Eberg, E., Monsen, A.F., Tybell, T., van Helvoort, A.T.J., Holmestad, R.: Comparison of TEM specimen preparation of perovskite thin films by tripod polishing and conventional ion milling. J. Electron Microsc. 57, 175–179 (2008) 16. Agarwal, S., Rao, P.V.: Grinding characteristics, material removal and damage formation mechanisms in high removal rate grinding of silicon carbide. Int. J. Mach. Tools Manuf. 50, 1077–1087 (2010) 17. Xu, L.M., Shen, B., Shih, A.J.: Vitreous bond silicon carbide wheel for grinding of silicon nitride. Int. J. Mach. Tools Manuf. 46, 631–639 (2006) 18. Fan, S.W., Zhang, L.T., Cheng, L.F., Tian, G.L., Yang, S.J.: Effect of braking pressure and braking speed on the tribological properties of C/SiC aircraft brake materials. Compos. Sci. Technol. 70, 959–965 (2010) 19. Goswami, A.P., Das, G.C.: Role of fabrication route and sintering on wear and mechanical properties of liquid-phase-sintered alumina. Ceram. Int. 26, 807–819 (2000) 20. Avalos, J.C.R., Ramirez, A.M., Limon, J.M.Y., Garcia, M.E.C., Guzman, E.M.A., Hernandez, J.G.: Development and characterization of an inorganic foam obtained by using sodium bicarbonate as a gas generator. Constr. Build. Mater. 19, 543–549 (2005) 21. Xu, H.H.K., Jahanmir, S., Ives, L.K.: Effect of grinding on strength of tetragonal zirconia and zirconia-toughened alumina. Mach. Sci. Technol. 1, 49–66 (1997) 22. Renault, P.O., Barbot, J.F., Girault, P., Declemy, A., Rivaud, G., Blanchard, C.: Properties of dislocations in HgCdTe crystals. J. Phys. III. 5, 1383–1389 (1995) 23. Foll, H., Carter, C.B.: Direct TEM determination of intrinsic and extrinsic stacking fault energies of silicon. Philos. Mag. A 40, 497–510 (1979) 24. Ray, I.L.F., Cockayne, D.J.H.: Investigation of dislocation geometries in the diamond cubic structure. J. Microsc. 98, 170–173 (1973) 25. Gomez, A.M., Hirsch, P.B.: The dissociation of dislocations in GaAs. Philos. Mag. A 38, 733–737 (1978) 26. Gogotsi, Y.G., Domnich, V., Dub, S.N., Kailer, A., Nickel, K.G.: Cyclic nanoindentation and Raman microspectroscopy study of phase transformations in semiconductors. J. Mater. Res. 15, 871–879 (2000) 27. Sabinina, I.V., Gutakovsky, A.K., Sidorov, Y.G., Latyshev, A.: Nature of V-shaped defects in HgCdTe epilayers grown by molecular beam epitaxy. J. Cryst. Growth 274, 339–346 (2005) 28. Material Safety Dada Sheet for MCT. Spitfire Semiconductors Inc. http://www.spitfirenz.com/ (2004) 29. Material Safety Dada Sheet for Si, Ge and GaAs. Sciencelab.com, Inc. http://www.sciencelab.com/ (2005) 30. Sanjay, A., Venkateswara, R.P.: Experimental investigation of surface/subsurface damage formation and material removal mechanisms in SiC grinding. Int. J. Mach. Tools Manuf. 48, 689–710 (2008)