GENESIS SOLAR WIND ALUMINUM ABUNDANCE CREATING AN ALUMINUM STANDARD FOR SIMS ANALYSESGENESIS SOLAR WIND ALUMINUM ABUNDANCE: CREATING AN ALUMINUM STANDARD FOR SIMS ANALYSESGENESIS SOLAR WIND ALUMINUM ABUNDANCE: CREATING AN ALUMINUM STANDARD FOR SIMS ANALYSES GENESIS SOLAR WIND ALUMINUM ABUNDANCE: CREATING AN ALUMINUM STANDARD FOR SIMS ANALYSES 1 2 1 1 1 3 1 1 1A E H f 1,2 J M P 1 D S B tt1 Y G 1 A J G J i 3 C M 1 d G R R 1 1C lif i I tit t f T h lA E Hofmann1,2 J M Paque1 D S Burnett1 Y Guan1 A J G Jurewicz3 C Ma1 and G R Rossman1 1California Institute of TechnologyA. E. Hofmann , J. M. Paque , D. S. Burnett , Y. Guan , A. J. G. Jurewicz , C. Ma and G. R. Rossman . California Institute of Technology, , q , , , , gy,

2Pasadena CA 91125 2Jet Propulsion Laboratory California Institute of Technology 4800 Oak Grove Dr M/S 183 401 Pasadena CA 91109Pasadena, CA 91125, Jet Propulsion Laboratory, California Institute of Technology, 4800 Oak Grove Dr M/S 183-401, Pasadena, CA 91109,Pasadena, CA 91125, Jet Propulsion Laboratory, California Institute of Technology, 4800 Oak Grove Dr M/S 183 401, Pasadena, CA 91109, 3

, , p y, gy, , , ,h f @j l 3SESE A i St t U i it T AZ 85287amy e hofmann@jpl nasa gov 3SESE Arizona State University Tempe AZ 85287amy.e.hofmann@jpl.nasa.gov. SESE, Arizona State University, Tempe, AZ 85287amy.e.hofmann@jpl.nasa.gov. SESE, Arizona State University, Tempe, AZ 85287

THE ALUMINUM PROBLEMTHE ALUMINUM PROBLEMTHE ALUMINUM PROBLEMTHE ALUMINUM PROBLEM The Genesis mission was designed to accurately measure the composi depth profile for San Carlos olivine (SCJ) using the Caltech CamecaThe Genesis mission was designed to accurately measure the composi- depth profile for San Carlos olivine (SCJ) using the Caltech Cameca g y pti f th l i d (SW) Th l t l h d i th

p p ( ) g7f G (O- i i 25 t i 200 fi ld t 75tion of the solar wind (SW) The sample return capsule crashed in the 7f Geo (O primary ion, 25 µm spot size, 200 µm field aperture, 75tion of the solar wind (SW). The sample return capsule crashed in the 7f Geo (O primary ion, 25 µm spot size, 200 µm field aperture, 75

d 3000 l i l 27Al+ fdesert fragmenting and contaminating samples (Figure 1) µm raster and 3000 mass resolving power to resolve 27Al+ fromdesert, fragmenting and contaminating samples (Figure 1). µm raster and 3000 mass resolving power to resolve Al from 26 +

, g g g p ( g )26MgH+) If the olivine Al content (deep part of profile) is known the

S i i l f th t diffi lt t k i i SW f thMgH ). If the olivine Al content (deep part of profile) is known, the

Surprisingly, one of the most difficult tasks in measuring SW from theg ) ( p p p ) ,

i l t fl b l l t d i [2]Surprisingly, one of the most difficult tasks in measuring SW from the implant fluence can be calculated, or vice versa [2].Genesis samples is calibrating standards for SIMS (secondary ion mass

implant fluence can be calculated, or vice versa [2]. Genesis samples is calibrating standards for SIMS (secondary ion mass p g ( yspectrometry) measurements Control pieces of Si mounted beside the olivine during implantationspectrometry) measurements. Control pieces of Si mounted beside the olivine during implantation p y) p g p

i th fl b i i t d d f l i fP i d h fil f l i d (SW) Al il bl f b k

receive the same fluence, becoming primary standards for analysis ofPrecise depth profiles of solar wind (SW) Al are available from back-

receive the same fluence, becoming primary standards for analysis of G i Si lPrecise depth profiles of solar wind (SW) Al are available from back- Genesis Si samples

side depth profiling [1] of Genesis silicon collectors Calculation of theGenesis Si samples.

side depth profiling [1] of Genesis silicon collectors. Calculation of the p p g [ ]l i d Al fl ( t / 2) f th i b d l b t Accurate electron microprobe (EMP) analyses of olivine Al at thesolar wind Al fluence (atoms/cm2) from these is based on a laboratory Accurate electron microprobe (EMP) analyses of olivine Al at the solar wind Al fluence (atoms/cm ) from these is based on a laboratory

i l d d b i l i l fl lp ( ) y

100 l l h ld b ibl I ti i ifi timplant standard but nominal implant fluences are only accurate to ≈100 ppm level should be possible In practice many significantimplant standard, but nominal implant fluences are only accurate to 100 ppm level should be possible. In practice, many significant p p yabout ±20% so independent calibration is required This can be done problems were encountered whose mitigations are described hereabout ±20%, so independent calibration is required. This can be done problems were encountered, whose mitigations are described here. , p q

d ib d i [2] d ill t t d i Fi 2 hi h h SIMSp , g

as described in [2] and illustrated in Figure 2 which shows a SIMSas described in [2] and illustrated in Figure 2, which shows a SIMS

RESULTSSan Carlos olivine RESULTSSan Carlos olivine Fi 5 RESULTSFigure 5 RESULTSFigure 5 RESULTS

S d I M SSecondar Ion Mass SpectrometrSecondary Ion Mass SpectrometrySecondary Ion Mass SpectrometryFig re 1 Secondary Ion Mass Spectrometry Figure 1 y p yg

Al fl f SCJ i l t R j ti th l filS C l li i SCJ i l t d ith 5 1014/ 2 Al fluence of SCJ implant Rejecting the anomalous profileSan Carlos olivine SCJ was implanted with 5x1014/cm2 Al fluence of SCJ implant. Rejecting the anomalous profile, 14

San Carlos olivine SCJ was implanted with 5x10 /cm the remaining 3 SCJ profiles give fluences of (units of 1014/Al at 80 keV (Figure 2) equivalent to about 15 20 ppm the remaining 3 SCJ profiles give fluences of (units of 10 /Al at 80 keV (Figure 2), equivalent to about 15-20 ppm the remaining 3 SCJ profiles give fluences of (units of 10 /

2) d h i d i i f i la 80 eV ( gu e ), equ va e o abou 5 0 pp

i EMP l i hi h i li ibl h SCJ cm2) 4 78 4 36 and 4 88 The estimated precision of a singlein an EMP analysis which is not negligible thus SCJ cm ) 4.78, 4.36, and 4.88. The estimated precision of a single in an EMP analysis, which is not negligible, thus SCJ ) , , p gfl l i i 4 8% Th b d fl t d d d i

y , g g ,t b d i t l t d d fluence analysis is 4 8% The observed fluence standard devi-cannot be used as an internal standard fluence analysis is 4.8%. The observed fluence standard devicannot be used as an internal standard .

ation is higher 6 0% which we adopt as the fluence preciation is higher, 6.0%, which we adopt as the fluence preci-Aluminum rich inclusions are also present in 2 out of 5 ation is higher, 6.0%, which we adopt as the fluence precii i i 14/ 2 h lib d fl f

Aluminum rich inclusions are also present in 2 out of 5 sion giving 4 67±0 30x1014/cm2 as the calibrated fluence of

pSCJ SIMS fil H h h d sion, giving 4.67±0.30x10 /cm as the calibrated fluence of SCJ SIMS profiles However the homogeneous and , g g

th SCJ i l t h th t d l i l d 2 2%SCJ SIMS profiles. However, the homogeneous and

the SCJ implant where the quoted error also includes a 2 2%p g

well calibrated EMP analyses (Figure 7) enable SC3 the SCJ implant where the quoted error also includes a 2.2% well-calibrated EMP analyses (Figure 7) enable SC3contribution from the error in the RSF The measured fluence

well calibrated EMP analyses (Figure 7) enable SC3 contribution from the error in the RSF. The measured fluenceto serve as an external standard to calibrate the SCJ contribution from the error in the RSF. The measured fluence i i hi f h i l 14 / 2to serve as an external standard to calibrate the SCJ is within 1 σ of the nominal 5x1014 atoms/cm2

Al i l t fl [1] T d th fil f SC3 is within 1 σ of the nominal 5x10 atoms/cm . Al implant fluence [1] Two depth profiles of SC3Al implant fluence [1]. Two depth profiles of SC3 Electron microprobe analyses: were bracketed between pairs of SCJ implant profiles The SC3Electron microprobe analyses: El t Mi b were bracketed between pairs of SCJ implant profiles. The SC3Electron microprobe analyses:

S i i i i Electron Microprobe were bracketed between pairs of SCJ implant profiles. The SC3 ÅSystematic instrument errors mitigated Electron Microprobe Al/Mg ratios deeper than 2000Å agree to better than 0 3% andSystematic instrument errors mitigated p Al/Mg ratios deeper than 2000Å agree to better than 0.3% and y g g p g

bi d ith th EMP Al t ti d fi l ti icombined with the EMP Al concentration define a relative sensi-With ti i d l ti l Fi 9Accurate electron microprobe (EMP) analyses of olivine combined with the EMP Al concentration define a relative sensiWith our optimized analytical con- Figure 9 Accurate electron microprobe (EMP) analyses of olivine tivity factor (RSF) [2] with a 1 σ error of 2 2%

With our optimized analytical con- gu e 9Al at the ≈100 ppm level should be possible In practice tivity factor (RSF) [2] with a 1 σ error of 2.2%.

p yditi 16/17 l thAl at the ≈100 ppm level should be possible. In practice, tivity factor (RSF) [2] with a 1 σ error of 2.2%. ditions, 16/17 analyses gave the

many significant problems were encountered whose Q i i SIMS l b d d dditions, 16/17 analyses gave the

many significant problems were encountered, whose Quantitative SIMS analyses must be done under steady state con-same counting rate within 1 σmitigations were described in [3] including (Figure 3): Quantitative SIMS analyses must be done under steady state con-same counting rate within 1 σ mitigations were described in [3] including (Figure 3): Q y yditi O ill t t i t t t f th f

gi i i i h1 Carbon deposition effect on count rates ditions Oscillatory transient structure was seen from the surfacecounting statistics errors with an1.Carbon deposition effect on count rates ditions. Oscillatory transient structure was seen from the surface

Å 24 30 56counting statistics errors with an

2 Burning of carbon coat with high current focused down to 1200 1500Å of all olivine 24Mg 30Si and 56Fe profilesg

f 69 3 Al (Fig re 7)2.Burning of carbon coat with high current focused down to 1200-1500Å of all olivine Mg, Si, and Fe profiles average of 69.3 ppm Al (Figure 7).beam

g, S , p(Fi 8) b hi d h l i ifi f

average of 69.3 ppm Al (Figure 7). beam (Figure 8) but this depth range over laps a significant par t ofEarlier analyses prior to optimiza3 Order of peak and background measurements is im- (Figure 8), but this depth range over laps a significant par t of Earlier analyses prior to optimiza-3.Order of peak and background measurements is im- ( g ), p g p g p

th SCJ Al i l t fil (Fi 2) At d th ll thy p p

i h d hportant the SCJ Al implant profile (Figure 2) At depths smaller thantion showed much more scatterportant the SCJ Al implant profile (Figure 2). At depths smaller than Å

tion showed much more scatter, 4 Selection of background points (Figure 4; from [4]) 600Å a large oscillatory transient is observed but in the 600hi h lth h t l t l4.Selection of background points (Figure 4; from [4]) 600Å a large oscillatory transient is observed, but in the 600-which, although not completely

Effects 1 3 eliminated by use of 50x100 micron rastered600 ge osc o y s e s obse ved, bu e 6001400Å h i f h M i ill i i ll

which, although not completely Effects 1-3 eliminated by use of 50x100 micron rastered 1400Å range the size of the Mg transient oscillations is smallunderstood may be due to surfacey

beam (Figure 3)1400Å range, the size of the Mg transient oscillations is small understood, may be due to surface beam (Figure 3).

g , g(<1%) d th M ti t b t l t d b

, yi i ll

( g )A d d i h l l Al “ i 25 25 ” (<1%) and the Mg counting rate can be accurately represented bycontamination as well as to sys-A standard with percent level Al, “enstatite 25-25,” was (<1%) and the Mg counting rate can be accurately represented by

Åcontamination as well as to sys-A standard with percent level Al, enstatite 25 25, was

d f l i i i i d d i i I a linear extrapolation from 1400Åy

t ti [3] Th SC3 lused for our analyses minimizing deadtime issues It a linear extrapolation from 1400Å. tematic errors [3]. The SC3 anal-used for our analyses, minimizing deadtime issues. It f ll h i d i i d d i

a ea e t apo at o o 00 .tematic errors [3]. The SC3 analwas carefully characterized using primary standards pri- D th 1400Å ti t d t i ll byses shows no evidence for effectswas carefully characterized using primary standards pri Deeper than 1400Å counting rates decrease monotonically be-yses shows no evidence for effects or to use Deeper than 1400Å, counting rates decrease monotonically be-y

f Al f i ior to use. p g ye f h i The SCJ fl e e i the d t f the RSF dof Al surface contaminationFor EMP analyses of San Carlos crystal SC3 (Figure 5) cause of charging. The SCJ fluence is the product of the RSF andof Al surface contamination. For EMP analyses of San Carlos crystal SC3 (Figure 5) cause of charging. The SCJ fluence is the product of the RSF and

we cleaned eleven 100 x 200 µm areas using the SIMS the (implant Al)/Mg depth integral [2]; this assumes that thewe cleaned eleven 100 x 200 µm areas using the SIMS the (implant Al)/Mg depth integral [2]; this assumes that the µ gto remove Al surface contamination revealed by XPS

( p ) g p g [ ];h i i i i h Al/M i i l Wi hFigure 7to remove Al surface contamination revealed by XPS charging variations in the Al/Mg counting rate ratios cancel WithFigure 7 y

analyses of SC3 [3; Figure 6]charging variations in the Al/Mg counting rate ratios cancel. With g

analyses of SC3 [3; Figure 6]. g g g g

ti t t Al/M ti b d ft tt Fi 9 y g

SC3 d th t tit t d d b t d t th one exception, constant Al/Mg ratios are observed after sputter- Figure 9SC3 and the enstatite standard were carbon coated at the one exception, constant Al/Mg ratios are observed after sputter Figure 9 SC3 a d t e e stat te sta da d we e ca bo coated at t eti t th t th thi k th b ing through the SCJ implants or beyond transients in the unimsame time to ensure that the thickness was the same be- ing through the SCJ implants or beyond transients in the unim- SCJ@2 Scaledsame time to ensure that the thickness was the same be

t th l d t d dg g p y

l d SC3 filSCJ@2 Scaled

tween the sample and standard. planted SC3 profilestween the sample and standard. planted SC3 profiles. SC3 was analyzed in two separate microprobe runs us-

p pSC3 was analyzed in two separate microprobe runs, us-

A “surface correction” for the Al implant fluence lost in the traning a 50 x 100 µm rastered beam 100 nA current 10 A surface correction for the Al implant fluence lost in the tran-ing a 50 x 100 µm rastered beam, 100 nA current, 10 A surface correction for the Al implant fluence lost in the trani i b l 600Å i b d l d h i l Sminute count time and 4 background points plus the Al sient region below 600Å is based on a scaled theoretical SRIMminute count time, and 4 background points plus the Al sient region below 600Å is based on a scaled theoretical SRIM

peak measured sequentiallyg

fil (Fi 9) Th di t d SRIM k d th i b t 33%peak, measured sequentially. profile (Figure 9) The predicted SRIM peak depth is about 33%profile (Figure 9). The predicted SRIM peak depth is about 33%

higher but when normalized to the peak of the measured profilehigher, but when normalized to the peak of the measured profile,higher, but when normalized to the peak of the measured profile, h d b 600Å d h k ll d ib d l h hthe data between 600Å and the peak are well described althoughFi 3 the data between 600Å and the peak are well described, although Figure 3 p , g

th d k i i ifi tl b d ll Th t ibFigure 3

the measured peak is significantly broader overall The contribu-the measured peak is significantly broader overall. The contribuÅS C l li i SC3 tion of the implant fluence from the surface to 600Å is calculatedSan Carlos olivine SC3 tion of the implant fluence from the surface to 600Å is calculatedSan Carlos olivine SC3 tion of the implant fluence from the surface to 600Å is calculated

f h l d fil d dd d h i l f hfrom the scaled SRIM profile and added to the integral of theFigure 6 from the scaled SRIM profile and added to the integral of the Figure 6 p gd (i l t Al)/M d t d th 600Å t i t t l

gmeasured (implant Al)/Mg data deeper than 600Å to give a totalmeasured (implant Al)/Mg data deeper than 600Å to give a total fluence For the adopted SCJ profiles the surface correctionsSIMS cleaned fluence. For the adopted SCJ profiles the surface correctionsSIMS cleaned fluence. For the adopted SCJ profiles the surface corrections

fareas

range from 21 to 29%areas

range from 21 to 29%. g

DISCUSSION AND CONCLUSIONSDISCUSSION AND CONCLUSIONSDISCUSSION AND CONCLUSIONSDISCUSSION AND CONCLUSIONS This work documents that an accurate calibration of an implant standard for the Genesis Al fluence is possibleThis work documents that an accurate calibration of an implant standard for the Genesis Al fluence is possible.This work documents that an accurate calibration of an implant standard for the Genesis Al fluence is possible. The large surface corrections for the SCJ profiles can be greatly reduced by using a higher implant energyThe large surface corrections for the SCJ profiles can be greatly reduced by using a higher implant energyThe large surface corrections for the SCJ profiles can be greatly reduced by using a higher implant energy (f t f 2 i ibl ) hifti l t ll f th i l t f th M t i t i A f 3 5% (1 ) i(factor of 2 is possible) shifting almost all of the implant from the Mg transient region An error of 3 5% (1 σ) in(factor of 2 is possible) shifting almost all of the implant from the Mg transient region. An error of 3.5% (1 σ) in ( p ) g p g g ( )

th f th t d d i l t fl d t t l f 5% i th l i d Al fl d 6% i th l i d Al/M fFi 4 the accuracy of the standard implant fluence and a total error of 5% in the solar wind Al fluence and 6% in the solar wind Al/Mg appears fea-Figure 4 the accuracy of the standard implant fluence and a total error of 5% in the solar wind Al fluence and 6% in the solar wind Al/Mg appears feaFigure 4 y p g ppibl F i th h t h i Al/M th t i h f 12% O k i t ti ll l bl f EMP tF B sible For comparison the photospheric Al/Mg among the most precise has an error of 12% Our work is potentially valuable for EMP traceFrom Batanova sible. For comparison, the photospheric Al/Mg, among the most precise, has an error of 12%. Our work is potentially valuable for EMP trace et al. (2015) p , p p g, g p , p yl t l i l O EMP i t th t li i Al t ti h th lit t d t i l

( )

element analyses in general Our EMP experience suggests that olivine Al concentrations are more homogenous than literature data implyelement analyses in general. Our EMP experience suggests that olivine Al concentrations are more homogenous than literature data imply. y g p gg g p yReferences: [1] Heber V S et al (2014) 45th LPSC Abstract #1203 [2] Burnett D S et al (2015) Geostandards and Geoanalytical Res 39 265 276 [3] Hofmann A E et al (2018) 49th LPSC AbstractReferences: [1] Heber V. S. et al. (2014) 45th LPSC, Abstract #1203. [2] Burnett, D. S. et al. (2015) Geostandards and Geoanalytical Res. 39, 265-276. [3] Hofmann A. E. et al. (2018) 49th LPSC, Abstract [ ] ( ) , [ ] , ( ) y , [ ] ( ) ,#1526 [4] Batanova et al (2015) Chemical Geology 419 149 157#1526. [4] Batanova et al. (2015) Chemical Geology 419, 149-157. [ ] ( ) gy ,