Secondary Ion Mass Spectrometer (SIMS and ToF-SIMS)

1. Theory

• Definition: A secondary ion mass spectrometer sputters a sample with a focused primary ion beam and mass-analyses the ejected secondary ions, giving dopant and impurity concentrations against depth with parts-per-billion sensitivity (EAG technique note; ASTM E1505).
• SIMS equation: The secondary current is the product of primary current, sputter yield, ionization probability, concentration and transmission; ionization probability depends strongly on the matrix (background knowledge, Wilson et al.).
• Quantification: Relative sensitivity factors from implanted standards in the same matrix convert intensity ratios to concentration (ASTM E1505; Drozdov et al.).
• Depth: Sputter time is converted to depth with the measured crater depth (ASTM E1505).

2. Working Principle

• Primary beams: O2+ enhances positive ions such as B and Fe; Cs+ enhances negative ions such as P, As, C, O and H (EAG notes).
• Dynamic SIMS: Continuous high-current sputtering with magnetic-sector or quadrupole analysers for depth profiles (EAG technique note).
• ToF-SIMS: A pulsed Bi cluster beam analyses while a separate Cs or O beam sputters, giving full mass spectra and surface imaging (Drozdov et al.; EAG technique note).
• Calibration: Implanted standards and crater profilometry fix concentration and depth scales (ASTM E1505).

3. Applications

• Dopant profiles: Junctions, halos and channel doping in silicon (ASTM E1505; EAG notes).
• Impurities: C, O, N, H and metals in wafers and epitaxy (EAG detection-limit note).
• Alloy composition: SiGe composition profiles in multilayers (Drozdov et al.).
• Surface chemistry: ToF-SIMS of contamination and organic residues (EAG technique note).

Signal and Quantification

Name	Formula	Related Formulae	Variables/Constants
Secondary ion signal	$I_s=I_{p}Y{\alpha }^{\pm }{\theta }_A\eta$	$Y=\frac{\text{atoms sputtered}}{\text{primary ion}}$	$I_s:secondaryioncurrentofspeciesA$ $I_p:primaryioncurrent$ $\begin{array}{r}Y:sputteryield,{\alpha }^{\pm }:ionization\\ probability\end{array}$ ${\theta }_A:fractionalconcentrationofA$ $\begin{array}{r}\eta :transmissionanddetectionefficiency;\\ SIMSequation(backgroundknowledge,\\ Wilsonetal.);EAGtechniquenote:O2+or\\ Cs+primarybeamsat100\mathrm{eV}to15\mathrm{keV}\end{array}$ $\begin{array}{r}(myreading:{\alpha }^{\pm }variesoverordersof\\ magnitudewithmatrix,sorawintensities\\ cannotbecomparedbetweenmaterials)\end{array}$
Quantification with relative sensitivity factors	$C_A=\frac{I_A{C}_R}{I_{R}RSFN}$	${\rho }_i=\frac{I_i}{I_m}RS{F}_i$	$C_A:concentrationofA$ $I_A:impurityintensity$ $I_R:referencematrixintensity$ $C_R:matrixreferenceconcentration$ $\begin{array}{r}RSF:relativesensitivityfactor(ASTM\\ form)\end{array}$ $N:isotopicabundance$ $\begin{array}{r}{I}_m:matrixintensity;ASTME1505:first\\ form,accuracywithin15percentRSD;\\ Drozdovetal.:secondformfordilute\\ impurities\end{array}$ $\begin{array}{r}(assumed:RSF=1\times 10^{22}\mathrm{c}{\mathrm{m}}^{-3}and{I}_B/I_{Si}=1\times 10^{-5};\\ computed:C_B=1\times 10^{17}\mathrm{c}{\mathrm{m}}^{-3})\end{array}$
RSF from an implanted standard	$RSF=\frac{\left(I_{A,integ}-I_{A,bkg}n\right)T{C}_{R}z}{F{I}_{R,avg}}$	$C_{A,max}=\frac{F}{2.5s}(\text{Gaussian})$	$I_{A,integ}:integratedcounts$ $I_{A,bkg}:backgroundperpoint$ $n:numberofpoints$ $T:mass-cycletime$ $z:sputterrate$ $F:implantfluence$ $\begin{array}{r}s:halfwidthat0.606ofmaximum;ASTM\\ E1505:integrationandGaussianmethods\\ (integrationmoreaccurate),peakbelow1\\ atomicpercent,peakatleast100times\\ background,fluences1\times 10^{13}to1\times 10^{16}\mathrm{c}{\mathrm{m}}^{-2},30\\ to400\mathrm{keV}\end{array}$ $\begin{array}{r}(assumed:F=1\times 10^{15}\mathrm{c}{\mathrm{m}}^{-2},s=30\mathrm{nm};\\ computed:C_{m}ax=1.3\times 10^{20}\mathrm{c}{\mathrm{m}}^{-3},0.27atomicper\\ cent,belowthe1percentlimit)\end{array}$

Depth Scale, Resolution and Limits

Name	Formula	Related Formulae	Variables/Constants
Depth scale	$z=\dot{z}t,\dot{z}=\frac{d_{crater}}{t_{total}}$	$\dot{z}=\frac{J_{p}Y}{q{N}_0}$	$\dot{z}:sputterrate$ $\begin{array}{r}{d}_{crater}:craterdepthbyprofilometryor\\ interferometry\end{array}$ $t_{total}:sputtertime$ $J_p:primarycurrentdensity$ $\begin{array}{r}{N}_0:atomicdensity;ASTME1505:\\ crater-depthmethod;Drozdovetal.:\\ constantaveragerateassumed,1\mathrm{keV}and\\ 80\mathrm{nA}Cs+at45degrees;fluxrelation\\ (backgroundknowledge,Wilsonetal.)\end{array}$ $\begin{array}{r}(assumed:300by300\mu \mathrm{m}raster,Y=2;\\ computed:5.6\times 10^{14}ionspersquare\\ centimetrepersecondand0.22nm/sinSi)\end{array}$
Depth resolution	$\sigma \left(z\right)={\sigma }_0+a\sqrt{z}$	$\text{MRI: mixing }w,\text{roughness }\sigma ,\text{information depth }\lambda$	$\begin{array}{r}\sigma (z):depth-dependentroughness\\ broadening;Drozdovetal.:relationwith\\ {\sigma }_0=2.3\mathrm{nm}anda=1.5forSNMS,ToF-SIMS\\ MRIparametersw=1\mathrm{nm},\sigma =2.3\mathrm{nm},\lambda =\\ 0.3\mathrm{nm},first-layerFWHM10.8\mathrm{nm}\\ (ToF-SIMS)against12.3\mathrm{nm}(HRXRD)and15\\ \mathrm{nm}(SNMS);EAGtechniquenote:monolayer\\ depthresolution(seetheQ/A);decay\\ lengthsof1to2nm/decadeatlowenergy\\ (backgroundknowledge,Wilsonetal.)\end{array}$ $\begin{array}{r}(computed:at100\mathrm{nm}depththeSNMS\\ broadeningis2.3+1.5\times 10=17.3\mathrm{nm})\end{array}$
Detection limit and dynamic range	$C_{DL}\approx \frac{RSF\cdot I_{bkg}}{I_m}$	$\text{ppba}=\frac{C}{N_0}\times 10^9,\text{ppbw}=\text{ppba}\frac{M_A}{M_{Si}}$	$\begin{array}{r}{C}_{DL}:detectionlimitsetbybackground\\ counts,N_0=5\times 10^{22}\mathrm{c}{\mathrm{m}}^{-3}forSi\end{array}$ $\begin{array}{r}M:atomicmass;EAGdetection-limitnote\\ (bulkSi):B1\times 10^{12}(0.008ppbw)andFe\\ 1\times 10^{13}(0.4)withO2+,P1\times 10^{13}(0.2),As\\ 1\times 10^{13}(0.5),N5\times 10^{13}(0.5),C1\times 10^{14}(3),O\\ 1\times 10^{15}(12),H5\times 10^{16}(36)\mathrm{c}{\mathrm{m}}^{-3}(ppbw)with\\ Cs+;EAGtechniquenote:ppmatoppba,up\\ tosevenordersofmagnitude,allelements\\ includingH\end{array}$ $\begin{array}{r}(computed:B1\times 10^{12}\mathrm{c}{\mathrm{m}}^{-3}is0.020ppbaand\\ 0.0077ppbw;H5\times 10^{16}\mathrm{c}{\mathrm{m}}^{-3}is1ppmaand36\\ ppbw,matchingthetable)\end{array}$

Alloys, ToF-SIMS and Practice

Name	Formula	Related Formulae	Variables/Constants
SiGe composition calibration	$\frac{I_{Ge}}{I_{Si}}=k\frac{x}{1-x},k=\frac{RS{F}_{Si}}{RS{F}_{Ge}}$	$x=\frac{I_{Ge}/I_{Si}}{k+I_{Ge}/I_{Si}}$	$x:Geatomicfraction$ $\begin{array}{r}k:calibrationconstant;Drozdovetal.:\\ relations,linearover0.092to0.78with\\ R^{2}of0.9999(GeC{s}_2^{+}/SiC{s}_2^{+})and0.9997\\ (negativeatomicions),intercepts0.0186,\\ 0.0974and-0.0091,SNMSk=0.2777,C{s}_2\\ clustersstrongerthansingle-Cs,matrix\\ effectscompensatedinSiGe\end{array}$ $\begin{array}{r}(computed:withk=0.2777,x=0.30gives\\ aratioof0.119)\end{array}$
Static against dynamic SIMS	${\Phi }_p<10^{12}\text{to}10^{13}\mathrm{c}{\mathrm{m}}^{-2}(\text{static limit})$	$\frac{{\Phi }_p}{N_s}\approx \frac{10^{12}}{1.5\times 10^{15}}$	${\Phi }_p:primaryiondose$ $\begin{array}{r}{N}_s:surfaceatomicdensity;staticlimit\\ (backgroundknowledge,Wilsonetal.);EAG\\ techniquenote:dynamicSIMSprofiles\\ whilesputtering,ToF-SIMSlimitedto\\ outermonolayersandgenerallynot\\ quantitative;Drozdovetal.:pulsed25\\ \mathrm{keV}1\mathrm{pA}B{i}^{+}andB{i}_2^{+}analysisbeamswith\\ Cssputteringindual-beamToF-SIMS,B{i}_3^{+}\\ giving3to5timesmoreyield\end{array}$ $\begin{array}{r}(computed:1\times 10^{12}ionspersquare\\ centimetredisturbsabout0.07percentof\\ amonolayer)\end{array}$
Analysis area and depth range	$A_o\gtrsim {\left(10\mu \mathrm{m}\right)}^2$	$z_{max}\sim \text{tens of }\mu \mathrm{m}$	$\begin{array}{r}{A}_o:analysedarea;EAGtechniquenote:\\ small-areaanalysisofatleast10\mu \mathrm{m},\\ depthfromafewnanometrestotensof\\ micrometres;ASTME1505:analysedarea{A}_o\\ intheRSFrelations\end{array}$ $\begin{array}{r}(computed:at0.22nm/s,10\mu \mathrm{m}ofSitakes\\ 12.6hours)\end{array}$

Important Questions

Ques: How does SIMS turn ion counts into concentration?

With relative sensitivity factors measured on standards in the same matrix. Raw signals follow $I_s=I_{p}Y\alpha \theta \eta$, and the ionization probability $\alpha$ varies by orders of magnitude between matrices (background knowledge, Wilson et al.), so ASTM E1505 normalizes to a matrix signal: $C_A=I_A{C}_R/(I_{R}RSFN)$. With an assumed RSF of $1\times 10^{22}$ $\mathrm{c}{\mathrm{m}}^{-3}$, a boron-to-silicon ratio of $1\times 10^{-5}$ gives $1\times 10^{17}$ $\mathrm{c}{\mathrm{m}}^{-3}$ (computed). The RSF comes from an implant of known fluence, by integration or a Gaussian fit; ASTM prefers integration and requires peak concentration below 1 atomic per cent and at least 100 times the background. A $1\times 10^{15}$ $\mathrm{c}{\mathrm{m}}^{-2}$ implant with 30 $\mathrm{nm}$ half width peaks at $1.3\times 10^{20}$ $\mathrm{c}{\mathrm{m}}^{-3}$, 0.27 atomic per cent (computed). Accuracy is about 15 per cent RSD (ASTM E1505). Reference: R. G. Wilson, F. A. Stevie and C. W. Magee, Secondary Ion Mass Spectrometry: A Practical Handbook for Depth Profiling and Bulk Impurity Analysis, Wiley 1989 (background knowledge; not opened); ASTM E1505-92 (reapproved 2001), ‘Standard Guide for Determining SIMS Relative Sensitivity Factors from Ion Implanted External Standards’ (fetch-tool summary).

Ques: Why are O2+ and Cs+ primary beams both used?

Because they enhance opposite ion polarities. Oxygen raises the yield of positive ions of electropositive elements, and caesium lowers the surface work function and raises negative-ion yields of electronegative elements (background knowledge, Wilson et al.). EAG’s bulk detection limits in silicon follow this split: B at $1\times 10^{12}$ and Fe at $1\times 10^{13}$ $\mathrm{c}{\mathrm{m}}^{-3}$ with O2+, and P, As and N at $1\times 10^{13}$ to $5\times 10^{13}$, C at $1\times 10^{14}$, O at $1\times 10^{15}$ and H at $5\times 10^{16}$ $\mathrm{c}{\mathrm{m}}^{-3}$ with Cs+. The parts-per-billion conversions check: B at $1\times 10^{12}$ $\mathrm{c}{\mathrm{m}}^{-3}$ is 0.020 ppba and 0.0077 ppbw, and H at $5\times 10^{16}$ is 1 ppma and 36 ppbw (computed), matching EAG’s 0.008 and 36. H, C and O limits are set by residual vacuum gases rather than ion yield (background knowledge). Reference: R. G. Wilson, F. A. Stevie and C. W. Magee, Secondary Ion Mass Spectrometry: A Practical Handbook for Depth Profiling and Bulk Impurity Analysis, Wiley 1989 (background knowledge; not opened); EAG Laboratories, ‘SIMS Detection Limits of Selected Elements in Si Under Bulk Analysis Conditions’, M-024517 (fetch-tool summary).

Ques: Is SIMS depth resolution a monolayer or about 10 nm?

The EAG technique note lists monolayer depth resolution, whereas Drozdov et al. measured a first-layer FWHM of 10.8 $\mathrm{nm}$ by ToF-SIMS against 12.3 $\mathrm{nm}$ by HRXRD, and 15 $\mathrm{nm}$ by SNMS. Their mixing-roughness-information model used 1 $\mathrm{nm}$ mixing, 2.3 $\mathrm{nm}$ roughness and a 0.3 $\mathrm{nm}$ information depth, and SNMS roughness grows as ${\sigma }_0+a\sqrt{z}$, about 17.3 $\mathrm{nm}$ at 100 $\mathrm{nm}$ depth (computed). The monolayer figure is more likely the information depth of the emitted ions, while measured interface widths of a few to ten nanometres reflect mixing and roughening (my reading). Low-energy beams near 1 $\mathrm{keV}$ give decay lengths of about 1 to 2 $nm/decade$ for steep dopant profiles (background knowledge, Wilson et al.). Reference: EAG Laboratories, ‘Secondary Ion Mass Spectrometry (SIMS) Services’ technique note M-006016 (fetch-tool summary); M. N. Drozdov, Y. N. Drozdov, A. Csik, A. V. Novikov, K. Vad, P. A. Yunin, D. V. Yurasov, S. F. Belykh, G. P. Gololobov, D. V. Suvorov and A. Tolstogouzov, ‘Quantitative depth profiling of Si1-xGex structures by time-of-flight secondary ion mass spectrometry and secondary neutral mass spectrometry’, Thin Solid Films 607, 25 (2016) (fetch-tool summary); R. G. Wilson, F. A. Stevie and C. W. Magee, Secondary Ion Mass Spectrometry: A Practical Handbook for Depth Profiling and Bulk Impurity Analysis, Wiley 1989 (background knowledge; not opened).

Ques: Did ToF-SIMS or SNMS give the right Ge content in the multilayer?

Drozdov et al. found about 40 atomic per cent Ge by ToF-SIMS and about 25 per cent by SNMS in a multilayer with a nominal 37 per cent, and their mixing-roughness model reconstructed 40 per cent in all layers. The ToF-SIMS value is more likely correct, because its calibration was linear from 9.2 to 78 per cent with R squared of 0.9999 and its depth resolution was better, 10.8 against 15 $\mathrm{nm}$ FWHM, so thin layers are less averaged with their Si neighbours (my reading). SNMS, with roughness growing as $2.3+1.5\sqrt{z}$, smears thin peaks and lowers their maxima. Their calibration $I_{Ge}/I_{Si}=kx/(1-x)$ with k = 0.2777 gives a ratio of 0.119 at x = 0.30 (computed), and inverting it returns x directly. Reference: M. N. Drozdov, Y. N. Drozdov, A. Csik, A. V. Novikov, K. Vad, P. A. Yunin, D. V. Yurasov, S. F. Belykh, G. P. Gololobov, D. V. Suvorov and A. Tolstogouzov, ‘Quantitative depth profiling of Si1-xGex structures by time-of-flight secondary ion mass spectrometry and secondary neutral mass spectrometry’, Thin Solid Films 607, 25 (2016) (fetch-tool summary).

Ques: How is the depth scale set, and how fast is sputtering?

From the final crater depth divided by total sputter time. ASTM E1505 measures the crater by profilometry or interferometry, and Drozdov et al. assume a constant average rate even across Si and SiGe layers, noting this does not greatly affect depth accuracy. The rate itself follows $\dot{z}=J_{p}Y/(q{N}_0)$ (background knowledge, Wilson et al.). Drozdov’s 80 $\mathrm{nA}$ Cs+ beam rastered over an assumed 300 by 300 $\mu \mathrm{m}$ area delivers $5.6\times 10^{14}$ ions per square centimetre per second; with an assumed yield of 2, silicon erodes at 0.22 $nm/s$ (computed). EAG gives a depth range from a few nanometres to tens of micrometres; 10 $\mu \mathrm{m}$ at that rate would take 12.6 hours (computed), so deep profiles use higher currents. Reference: ASTM E1505-92 (reapproved 2001), ‘Standard Guide for Determining SIMS Relative Sensitivity Factors from Ion Implanted External Standards’ (fetch-tool summary); M. N. Drozdov, Y. N. Drozdov, A. Csik, A. V. Novikov, K. Vad, P. A. Yunin, D. V. Yurasov, S. F. Belykh, G. P. Gololobov, D. V. Suvorov and A. Tolstogouzov, ‘Quantitative depth profiling of Si1-xGex structures by time-of-flight secondary ion mass spectrometry and secondary neutral mass spectrometry’, Thin Solid Films 607, 25 (2016) (fetch-tool summary); R. G. Wilson, F. A. Stevie and C. W. Magee, Secondary Ion Mass Spectrometry: A Practical Handbook for Depth Profiling and Bulk Impurity Analysis, Wiley 1989 (background knowledge; not opened); EAG Laboratories, ‘Secondary Ion Mass Spectrometry (SIMS) Services’ technique note M-006016 (fetch-tool summary).

Ques: What does ToF-SIMS add to dynamic SIMS?

Full mass spectra and surface imaging at very low dose. EAG notes that static SIMS, typically with a time-of-flight analyser, samples the outer monolayers and is generally not quantitative, while dynamic SIMS sputters continuously for depth profiles. Static conditions keep the primary dose below about $1\times 10^{12}$ to $1\times 10^{13}$ ions per square centimetre (background knowledge, Wilson et al.); $1\times 10^{12}$ disturbs about 0.07 per cent of a $1.5\times 10^{15}$ per square centimetre monolayer (computed). Drozdov et al. used dual-beam ToF-SIMS: 1 $\mathrm{keV}$ Cs+ for sputtering and pulsed 25 $\mathrm{keV}$ Bi+ and Bi2+ for analysis, with Bi3+ clusters giving 3 to 5 times more yield, and they quantified with GeCs2+ and SiCs2+ cluster ions, which were stronger than single-Cs ions. Reference: EAG Laboratories, ‘Secondary Ion Mass Spectrometry (SIMS) Services’ technique note M-006016 (fetch-tool summary); R. G. Wilson, F. A. Stevie and C. W. Magee, Secondary Ion Mass Spectrometry: A Practical Handbook for Depth Profiling and Bulk Impurity Analysis, Wiley 1989 (background knowledge; not opened); M. N. Drozdov, Y. N. Drozdov, A. Csik, A. V. Novikov, K. Vad, P. A. Yunin, D. V. Yurasov, S. F. Belykh, G. P. Gololobov, D. V. Suvorov and A. Tolstogouzov, ‘Quantitative depth profiling of Si1-xGex structures by time-of-flight secondary ion mass spectrometry and secondary neutral mass spectrometry’, Thin Solid Films 607, 25 (2016) (fetch-tool summary).

References

1. R. G. Wilson, F. A. Stevie and C. W. Magee, Secondary Ion Mass Spectrometry: A Practical Handbook for Depth Profiling and Bulk Impurity Analysis, Wiley, 1989 (cited from background knowledge).
2. ASTM E1505-92 (reapproved 2001), “Standard Guide for Determining SIMS Relative Sensitivity Factors from Ion Implanted External Standards,” ASTM International. Available: https://softbank.iust.ac.ir/Railway/Codes/ASTM-standard/ASTM_2004_Standards/ASTM%202004%20PART%20IA/ASTM%202004%20Volume%2003.06%20Molecular%20Spectroscopy_%20Surface%20Analysis/E%201505%20-%2092%20R01%20%20_RTE1MDU_.pdf
3. EAG Laboratories, “SIMS Detection Limits of Selected Elements in Si Under Bulk Analysis Conditions,” M-024517. Available: https://www.eag.com/wp-content/uploads/2017/09/M-024517-SIMS-Detection-Limits-of-Selected-Elements-in-Si-Under-Bulk-Analysis-Conditions.pdf
4. EAG Laboratories, “Secondary Ion Mass Spectrometry (SIMS) Services,” technique note M-006016. Available: https://www.eag.com/wp-content/uploads/2017/03/technique-note-secondary-ion-mass-spectrometry-SIMS-services-M-006016.pdf
5. M. N. Drozdov et al., “Quantitative depth profiling of Si1-xGex structures by time-of-flight secondary ion mass spectrometry and secondary neutral mass spectrometry,” Thin Solid Films, vol. 607, pp. 25-31, 2016. Available: https://arxiv.org/pdf/1701.01612

Related pages in this vault: X-Ray Photoelectron Spectrometer (XPS), X-Ray Diffractometer (XRD and HRXRD), Rutherford Backscattering Spectrometer (RBS), Atom Probe Tomography (APT), Spreading Resistance Profiler (SRP), Scanning Spreading Resistance Microscope (SSRM), Auger Electron Spectrometer (AES), Focused Ion Beam (FIB), FTIR Spectrometer, Residual Gas Analyzer (RGA), Beamline Ion Implanter, Plasma Immersion Ion Implanter, Diffusion Furnace, Ion Beam Milling System.