SIMS XVIII SIMS Course Depth Profiling

Transcription

1 SIMS XVIII SIMS Course Depth Profiling Fondazione Bruno Kessler Trento, Italy Fred A. Stevie Analytical Instrumentation Facility North Carolina State University Raleigh, NC USA

2 Outline Depth profiling Quantification Insulators / Residual gas species Instrumentation Ultra shallow analysis Applications

3 Depth Profiling Rationale for SIMS depth profiling Static versus dynamic sputtering Depth profile characteristics Parameter choices Selection of primary beam and secondary species Raster and gate Sputtering rate Sensitivity (detection limit) Count rate saturation Mass interferences (high mass resolution, voltage offset) Memory effects Small area analysis Depth resolution (sensitivity versus depth resolution) Ion beam induced roughening Analysis approach

4 Common Elemental Analysis Analytical Techniques AES XPS SIMS EDS Probe Species electron x-ray ion electron Detected Species electron electron ion x-ray Information Depth 2nm 2nm 0.3-1nm 0.1-1μm Lateral Resolution 20nm 10μm-1mm 7nm-10μm 5nm Elements Detected >He >He all >Na Detection Limit 0.1-1% 0.1-1% ppm - ppb 0.5-1% Chemical Info limited yes yes (ToF) no 4

5 Rationale for SIMS Depth Profiling Typical surface analysis techniques for elemental analysis AES, EDS, XPS, SIMS SIMS depth profiling used to provide: High sensitivity Good depth resolution Depth profiling can be achieved with magnetic sector, quadrupole, or Time-of-Flight instruments.

6 Static vs. Dynamic Conditions Parameter Static Dynamic Residual pressure Torr 10-7 Torr Primary current density µa/cm µa/cm 2 Analyzed area cm 2 5x10-4 cm 2 Atomic layer erosion rate /s /s Primary current 1 na 100 na Raster 1000 x 1000 µm x 250 µm 2 Primary current density <1 na/cm ma/cm 2 Sputtering rate 0.5 nm/hr 3.6 µm/hr Atoms in Si monolayer /cm 2 or 10 7 /µm 2 6

7 Static vs. Dynamic SIMS Static SIMS limit is ~10 12 ions/cm 2 (Si surface is atoms/cm 2 ) 1 na / (1000 µm x 1000 µm) = 6.25x10 11 ions/cm 2 -sec < 10 sec to reach static limit 10 pa / (1000 µm x 1000 µm) = 6.25x10 9 ions/cm 2 -sec > 800 sec to reach static limit Dynamic SIMS may have 150 na / (200 µm x 200 µm) = 2.3x10 15 ions/cm 2 -sec < sec to exceed static limit 7

8 STM Images Before and After Static SIMS Si surface Si surface exposed to 3x10 12 ions/ cm 2 H.J.W. Zandvliet, H. B. Elswijk, E. J. van Loenen, I. S. T. Tsong, SIMS VIII, A. Benninghoven, et al., eds. (1992) 3

9 Depth Profiling Need to remove material from region of interest Require a process that will work in vacuum, such as sputtering Electrons have very limited sputtering capability Laser ablation lacks uniformity of removal Sputtering typically done with ions Ions have charge Beam of ions can be rastered to create a crater Controlled removal possible Many different ion beams are used Raster and gate method employed 9

10 Depth Profile Characteristics J. Bennett, SEMATECH 10

11 Parameter Choices Primary beam species, energy, angle of incidence Sputtering rate (primary beam current and raster size) Raster and gate Secondary beam species Energy distribution of secondary species (voltage offset)

12 Absolute Positive Secondary Ion Yields Ar + 3 kev, 70º incidence Most elements show greater than x10 increase for oxygen covered surface compared with clean surface A. Benninghoven, Critical Rev. Solid State Sci. 6, 291 (1976)

13 Positive Secondary Ion Yields for O 2 + Bombardment of Si Relative Sensitivity Factors (RSFs) inversely proportional to ion yields More than 6 orders of magnitude variation across periodic table Highest yield for alkali elements SIMS, R. G. Wilson, F. A. Stevie, and C. W. Magee, Wiley, New York (1989)

14 Negative Secondary Ion Yields for Cs + Bombardment of Si Relative sensitivity factors (RSFs) inversely proportional to ion yields More than 5 orders of magnitude variation across periodic table Highest yield for halogens. Yields for O 2 + and Cs + are complementary SIMS, R. G. Wilson, F. A. Stevie. and C. W. Magee, Wiley, New York (1989)

15 Primary Beam / Secondary Polarity O 2 + and Cs + provide complementary secondary ion yields Evans Analytical Group 15

16 Choice of Profile Species for Detection Limit Best detection limit for In implant in Si obtained using O

17 Selection of Secondary Species SIMS Depth Profile of V Implanted in SiO 2 Profile of atomic and molecular ions indicates optimum choice for sensitivity VO 2 and VOSi have identical secondary ion yields (Mass spectrum on next slide taken at peak of implant) 17

18 Mass Spectrum of Vanadium Implant Mass spectrum taken at peak intensity of V ion implant depth profile showing relative intensities of species containing V. 28 Si + is saturated SIMS, R. G. Wilson, F. A. Stevie, and C. W. Magee, Wiley, New York (1989)

19 Choice of Species Improved detection limit for Ge in InP using molecular ion that contains an impurity atom and a matrix atom SIMS, R. G. Wilson, F. A. Stevie, and C. W. Magee, Wiley, New York (1989) 19

20 Choice of Species Improved detection limit for Dy in GaAs using molecular ion that contains an impurity atom and the primary beam species SIMS, R. G. Wilson, F. A. Stevie, and C. W. Magee, Wiley, New York (1989) 20

21 Secondary Species Choices Matrix species for normalization Major or minor isotope Atomic or molecular ion Mass interferences / mass resolution required Count time per point (data density vs. statistics) Voltage offset 21

22 Crater Issues for SIMS Depth Profile A) Surface layer with atomically sharp interface B) Rounded crater C) Flat bottomed crater - sidewall contributions D) Secondary ions due to neutral species E) Non-uniform sputtering F) Knock-on SIMS, C. G. Pantano, Metals Handbook Ninth Edition, Vol. 10, R. E. Whan, ed., American Society for Metals, Metals Park (1986)

23 Raster & Gate Relationship to Depth Profile Shape Reduce sidewall contributions by detecting ions from center of larger crater SIMS, R. G. Wilson, F. A. Stevie, and C. W. Magee, Wiley, New York (1989) 23

24 Depth Profiling: Raster and Gate Method Raster Circular gate (aperture) Square gate (electronic) 24

25 Beam Diameter Increases Crater Size Crater size will be raster chosen plus beam diameter For electronic gating, detected area is beam diameter added to gated length and width SIMS, R. G. Wilson, F. A. Stevie, and C. W. Magee, Wiley, New York (1989) 25

26 Parameters for Profilometer Trace of SIMS Crater Beam Crater width is sum of raster width and beam diameter. SIMS, R. G. Wilson, F. A. Stevie, and C. W. Magee, Wiley, New York (1989) 26

27 Detected Area versus Raster Size Arsenic implant in Si Profile A: 220 µm x 220 µm Profile B: 80 µm x 80 µm Profile B illustrates the distortion introduced by too small a ratio of rastered area to detected area. The raster is insufficient to reject sidewall contributions. The shaded area represents a 60 µm diameter detected area. Profile A is more accurate. SIMS, R. G. Wilson, F. A. Stevie, and C. W. Magee, Wiley, New York (1989) 27

28 Sputtering Rate Sputtering rate determines time for depth profile Sputtering rate can be varied with raster size or beam current Sensitivity typically improved with higher sputtering rate For sample with layers, sputtering rate can be different for each layer

29 Dynamic Range Depth profile of B implanted Si at near normal incidence O 2 + bombardment with 13% linear electronic gate. Background signal is 15 cts/s. Almost 6 orders of magnitude dynamic range. K. Wittmaack and J. B. Clegg, Appl. Phys. Lett. 37, 285 (1980) 29

30 Sensitvity (Detection Limit) K + 1 ppma 10 ppt detection limit obtained for K in Si using O ppba CAMECA IMS-6f 10 ppta F. A. Stevie, R. G. Wilson, J. M. McKinley, and C. J. Hitzman, SIMS XI Proceedings, G. Gillen, et al., eds., Wiley, Chichester (1998)

31 Count Rate Saturation Determine gated or detected area to rastered area ratio to convert measured constant count rate to instantaneous count rate. Instantaneous count rate is constant count rate times ratio of rastered area to detected area. SIMS, R. G. Wilson, F. A. Stevie, and C. W. Magee, Wiley, New York (1989) 31

32 Count Rate Saturation Depth profiles of fluorine in Si with different count rates. The profiles are adjusted to show the effect of different levels of count rate saturation on profile shape. SIMS, R. G. Wilson, F. A. Stevie, and C. W. Magee, Wiley, New York (1989) 32

33 Count Rate Saturation Compare ratio of 28 Si + / 30 Si + for Si + and Si + Mass spectra taken with detected area equal to 4 and 36% of the sputtered area, showing that 28 Si + is saturated in both cases and the ratio of 28 Si + to 30 Si + is incorrect. Analyzed using quadrupole. SIMS, R. G. Wilson, F. A. Stevie, and C. W. Magee, Wiley, New York (1989) 33

34 High Mass Resolution 31 P and 30 Si 1 H 31 P H Si P mass separation from 30 Si 1 H is amu or 7.8 mmu Mass resolution required is M/ΔM = 31/ = Fe and 28 Si 2 56 Fe Si Fe mass separation from 28 Si 2 is amu or 18.9 mmu Mass resolution required is 56/ = Ar and 40 Ca 40 Ar Ca Mass resolution required is? 34

35 Secondary Ion Energy Distribution Normalized energy distributions of sputtered Si and Si molecular ions. Analyzed using quadrupole instrument. K. Wittmaack, Phys. Lett. 69A, 322 (1979) 35

36 Voltage Offset Placement of energy window to achieve maximum separation between Si + atomic and Si 2 O + molecular ions. Atomic and molecular distributions 1) energy window open 2) energy window translated 3) energy window translated and sample voltage offset CAMECA operating manual 36

37 Voltage Offset Mass interference for 75 As is 29 Si 30 Si 16 O Improvement in detection limit for As implant in Si analyzed using O 2 + and Cs + primary beams at different energy window positions Energy window moved with sample offset voltage. SIMS, R. G. Wilson, F. A. Stevie, and C. W. Magee, Wiley, New York (1989) 37

38 Voltage Offset on Mass Spectrum No offset 75 V offset Reduction of molecular ion intensities in InSb mass spectrum with application of voltage offset. SIMS, R. G. Wilson, F. A. Stevie, and C. W. Magee, Wiley, New York (1989) 38

39 Memory Effects Magnetic Sector Close proximity of sample and immersion lens cover plate permits re-sputtering of material previously analyzed Example: prior analysis of InP substrate will affect detection of P in Si. The effect shows at about 5 orders of magnitude below the matrix density Quadrupole Open geometry reduces memory effect Time of Flight Low sputtering rate reduces this effect 39

40 Magnetic Sector Geometry Primary Beam Memory effect ~1 part in primary beam 2-sputter and deposit 3-resputter and deposit on sample 5 mm 1 Immersion lens cover plate 2 3 ground Sample 4.5 kv 40

41 Quadrupole Geometry Primary Beam Extraction 250 V 1-2 cm primary beam 2-sputter and deposit 3-resputter and deposit on sample Sample ground 41

42 Analysis Volume SIMS analysis volume is small Analysis 100 m Raster 100 m 30 m 1 cm Specimen 1 cm 42

43 Analysis Volume SIMS analysis volume 100 µm x 100 µm raster 30 µm diameter analyzed area and 1 µm depth analysis volume = 7.1x10-10 cm 3 Amount sampled Si density x analysis volume = 5x10 22 atoms/cm 3 x 7.1x10-10 cm 3 = 3.6x10 13 atoms x 28 gm/6.02x10 23 atoms = 1.7 ng Need more than one analysis point per sample 43

44 Relationship of Analysis Area to Detection Limit 44

45 Small Area Analysis on Silicon Patterned Wafers Dedicated SIMS Patterns Analysis patterns in grid areas between devices e.g.: n source and drain, p source and drain, poly/gate oxide/si 45

46 SIMS Patterns on Patterned Wafers Nthinox & Pthinox in Grid areas 90µm x320µm pattern in 100µm grid spacing 46

47 SIMS Patterns on Patterned Wafers 100 µm x 125 µm patterns. The crater in the pattern on the right is from a 75 µm x 75 µm raster. F. A. Stevie, et al., J. Vac. Sci. Technol. A10, 2280 (1992) 47

48 Counts (cts/sec) Counts (cts/sec) 1E+06 1E+05 C, O, N in Nb Cs kev Sensitivity vs. Depth Resolution 1E+04 1E+03 1E+02 1E+01 1E C Counts 16O Counts 93Nb Counts 93Nb14N Counts Typically better depth resolution but lower counts by reducing impact energy Time (s) 1E+07 1E+06 1E+05 1E+04 1E+03 1E+02 1E+01 C, O, N in Nb Cs + 6 kev 12C Counts 16O Counts 93Nb Counts 93Nb14N Counts 14.5 kev 20nA 120µm raster 30µm analyzed diameter, MR kev 15nA 130µm raster 30µm analyzed diameter, MR2000 1E Time (s)

49 Concentration (atom/cm3) Quantification of C, N, O in Nb Counts (cts/sec) 1E+21 93Nb 1E+04 Ion implantation Dose: atoms/cm 2 1E+20 1E+19 16O 93Nb14N 1E+03 1E+02 C: 1E15 N: 1E15 O: 2E15 1E+18 1E+17 12C Depth (um) 1E+01 1E+00 SIMS analysis CAMECA IMS-6F 6keV impact energy P. Maheshwari, et al., Surface and Interface Analysis (2010, 42) 49

50 -3 ) Depth Profiling Using TOF-SIMS Intensity (counts) Depth Profile of C, O, N in Nb C - O - Nb - NbN Depth (nm) Intensity (counts) Analysis Beam Sputtering Beam Ions Bi 3 + Cs + Energy 25 kev 10keV Raster area 50 x 50 µm x 120 µm 2 Current 0.2 pa 20 na ToF Energy 2 kv ION-TOF TOF-SIMS V C. Zhou, NC State University Depth (nm)

51 Interlaced Mode Sample is sputtered during the flight time of the secondary ions in the analyzer between the analysis gun pulses Analysis Gun <1 ns Extraction 10 µs Cycle time 100 µs Delay 5 µs Leadoff 5 µs Flood Gun 80 µs Sputter Gun Delay 10 µs Leadoff 10 µs 60 µs C. Zhou, NC State University

52 Non-Interlaced Mode The sputtering and analysis are sequentially organized. Analysis Gun Analysis 1.64 s 100 µs cycle time, cycle 1 scan of 128 x 128 pixel, 1 shot/pixel Flood Gun Flood gun 1.66 s Sputter Gun Sputter 1.46 s Pause 0.2 s C. Zhou, NC State University

53 Interlaced Mode vs. Non-Interlaced Mode Concentration (cm Concentration (cm -3-3 ) ) Intensity (counts) Intensity (counts) Analysis Beam Sputtering Beam Concentration (cm Concentration (cm Ions Bi 3 + Cs + Energy 25 kev 10keV Raster area 50 x 50 µm x 120 µm 2 Current 0.2 pa 20 na ToF Energy 2 kv H, C, O, N implanted in Si H and O profiles in non-interlaced mode show higher background signals due to the influence of the residual gas during the analysis phase Concentration (cm Concentration (cm -3-3 ) ) -3-3 ) ) Interlaced mode H - H - C - - O - - Si - - SiN SiN Depth (nm) Depth Depth (nm) (nm) H H SiN Non-Interlaced mode C C - C - O - - Si - - SiN SiN - - SiN Intensity (counts) Intensity (counts) Intensity (counts) Intensity (counts) C. Zhou, NC State University Depth (nm)

54 Depth Resolution Simulation Effect of 10% unevenness on Gaussian distribution D. S. McPhail, et al., Scanning Microscopy 2, 639 (1989) 54

55 Depth Resolution Simulation Effect of 1% and 10% unevenness on crater bottom for a sinusoidal dopant distribution, according to the uneven etching model D. S. McPhail, et al., Scanning Microscopy 2, 639 (1989) 55

56 Depth Resolution Factors that improve depth resolution Larger angle of incidence from normal for bombarding species Lower bombarding energy Increased mass of bombarding species Depth Resolution improves as ion penetration is reduced Depth resolution measurement Decay length depth for 1/e intensity change Leading and trailing edges decay length for increasing and decreasing ion intensity Interface width 84-16% of maximum Full width at half maximum 56

57 Penetration for Different Analysis Conditions TRIM Monte Carlo Simulations in GaN 0 R p = 4.8nm DR p = 2.9nm R p = 1.6nm DR p = 1.0nm R p = 1.4nm DR p = 0.7nm 20 nm O kev Θ = 41.3º O kev Θ = 48.5º C s+, 1.25 kev Θ = 48.5 º GaN density 6.15gm/cm

58 Depth Resolution Depth profile parameters for analysis of interface described in terms of sputter time or depth. Error function is derivative of interface curve and +/- sigma points correspond to 84 and 16% of maximum intensity. SIMS, R. G. Wilson, F. A. Stevie, and C. W. Magee, Wiley, New York (1989) 58

59 Depth Resolution Illustration of depth resolution for GaAs/Si interface in structure of GaAs/Si/Al 2 O 3. As and Si are quantified. 125 V offset on As. 50% of maximum point designates interface. SIMS, R. G. Wilson, F. A. Stevie, and C. W. Magee, Wiley, New York (1989) 59

60 Depth Resolution Profile of delta doped Be in GaAs showing less than 3 nm FWHM O kev 60 from normal Data from H. Luftman, E. F. Schubert, and R. F. Kopf F. A. Stevie, J. Vac. Sci. Technol. B10, 323 (1992) 60

61 Sensitivity versus Depth Resolution Cannot simultaneously optimize sensitivity and depth resolution Best sensitivity High sputtering rate Large detected area - more counts Best depth resolution Low impact energy, reduced ion penetration into sample Low sputtering rate Small detected area reduce effect of sample variations

62 Non-Uniform Sputtering Ion bombardment can cause non-uniform sputtering Polycrystalline materials (metals) expose surfaces at varied angles to ion beam Observed even for crystalline materials under certain conditions Check for roughness Optical crater bottom becomes dark if significant roughening Profilometer SEM AFM 62

63 Sputter Induced Roughness O kev profile of (100) Si. Arrows match depths for SEM micrographs SEM micrographs of SIMS crater bottoms at depths of a) 2.1, b) 2.8, and c) 4.3 µm Roughening of crater affects secondary ion yields F. A. Stevie, P. M. Kahora, D. S. Simons, and P. Chi, J. Vac. Sci. Technol. A6, 76 (1988) 63

64 Reduction of Roughening Choice of primary species Energy of species Incidence angle Sample rotation 64

65 RMS Roughness vs Incidence Angle Different O 2 + impact energies AFM roughness measurement Z. X. Jiang and P. F. A. Alkemade, Appl. Phys. Lett. 73, 315 (1998) 65

66 Crater Roughness at Oblique Incidence 1keV O w/ O 2 flood UHV 2E-5 Pa 1.3E-4 Pa Z. X. Jiang and P. F. A. Alkemade 66

67 Sample Rotation Al and Ga profiles from 3 kev O 2 + analysis of GaAs/AlGaAs superlattice a) without rotation b) with rotation E.-H. Cirlin, et al., J. Vac. Sci. Technol. A8, 4101 (1990) 67

68 Sample Rotation During Ion Bombardment Surface Sputter through layer Sputter through layer SEM micrographs of a) aluminum surface, b) bottom of crater sputtered through 1 µm aluminum layer into underlying silicon without rotation and c) with rotation F. A. Stevie and J. L. Moore, Surf. Interf. Anal. 18, 147 (1992) 68

69 Sample Rotation During Ion Bombardment without rotation with rotation SIMS profiles of 11 B ion implantation into 1 µm Al/Si. With sample rotation, B at interface is clearly defined and silicon from Al-Si-Cu layer shows movement to Al/Si interface F. A. Stevie and J. L. Moore, Surf. Interf. Anal. 18, 147 (1992) 69

70 Depth Profiling Analysis Approach Choose primary beam species / secondary polarity Choose primary beam parameters (voltage, raster, current, ) Sample preparation if necessary (remove over-layers) If insulator, prevent sample charging Decide on quantification, determine if standards can be made Depth profile: determine sputtering rates, identify layers Obtain mass spectra: check mass interferences and interfacial contaminants Depth profiles to obtain at least two matching analyses If analyses do not match, check for particles and anomalies 70