Assessment of the precision and accuracy of laser ablation ... · PDF fileanalyses in the Fluids Research Laboratory within the Department of Geosciences, Virginia Tech, Blacksburg,

1

Assessment of the precision and accuracy of laser ablation-ICPMS

analyses in the Fluids Research Laboratory within the Department of

Geosciences, Virginia Tech, Blacksburg, VA U.S.A.

The following document describes results of laser ablation ICP-MS analyses of several

standards of known composition conducted using the LA-ICPMS system installed in the

Department of Geosciences at Virginia Tech to evaluate the precision and accuracy of

analyses. The values reported here apply in general to analyses of solid materials in

which ablation starts at the sample surface, and do not necessarily apply to analyses of

fluid inclusions where the accuracy and precision are related to inclusion size and depth

beneath the surface, among other factors.

To asses the accuracy and precision of analyses, NIST standard 610 was used as the

reference, and then each of several other standards were analyzed as “unknowns”.

Accuracy was assessed by comparing the elemental abundances predicted by our

analyses with the “known” values. Each standard (USGS standards BCR-2G, BHVO-2G,

BIR-1G, NKT-1G and NIST standards NIST612 and NIST614) was analyzed five times

according to the specifications indicated in Table 1.

USGS standards BCR-2G, BHVO-2G and BIR-1G are natural basalt glasses

distributed by the USGS. NKT-1G is a natural nephelinite glass distributed by the USGS

(http://crustal.usgs.gov/geochemical_reference_standards/basaltbcr2.html). NIST 612 and

614 are commercial standard reference materials obtained from the National Institute of

Standards and Technology.

The data were reduced using the AMS software (Mutchler et al., 2008), using NIST610

glass as external standard and assuming that the elements analyzed represent 100 % of

the sample (as oxides). The average concentration in ppm, average limit of detection in

ppm, standard deviation in ppm and number of analyses above the detection limit (n) of

the 5 analyses for each element are reported in Table 2.

The precision of the measurements is reported as percent relative standard deviation

(%RSD) for each element, which is calculated as the standard deviation divided by the

average concentration. The %RSD provides an estimate of the consistency of the

analyses of the same standard, or in other words, the reproducibility of the measurement.

2

The relative standard deviation is mostly within 5-10 percent (Fig. 1) indicating that

repeated analyses of the same material give results that are consistent to within 5-10 %.

The accuracy of the measurement is reported as the average percent error for each

element, and is calculated as the average of the differences between each individual

analysis and the known concentration for the standard. The known concentrations used to

evaluate the accuracy of the LA-ICPMS analyses are the preferred concentrations in the

standards reported by the Max-Plank-Institut für Chemie at http://georem.mpch-

mainz.gwdg.de. The average percent error provides an estimate of the difference between

the calculated and known concentration. The average percent error is mostly within 5-

15% (Fig. 2) indicating that the measured concentrations agree with the known

concentrations within 5-15 %.

Mutchler SR, Fedele L & Bodnar RJ (2008) Analysis Management System (AMS) for

reduction of laser ablation ICPMS data. In Laser-Ablation-ICPMS in the Earth Sciences: Current Practices and Outstanding Issues (P. Sylvester, ed.) Mineralogical Association of Canada Short Course Series, Vol. 40, 318-327.

The following data are the result of work conducted in the LA-ICPMS laboratory in

the Department of Geosciences, Virginia Tech, Blacksburg, VA. These data are

provided as a mean to evaluate the basic analytical performance of our laboratory.

Please reference this report in your publications using this link:

http://www.geochem.geos.vt.edu/fluids/laicpms/VT_LAICPMS_Accuracy_Precision

.pdf.

If you use this site as a reference for the accuracy and precision of your data, please

email the information regarding your publication to Dr. Robert Bodnar

([email protected]).

Raw data are available upon request. Please email Luca Fedele ([email protected])

3

LA−ICPMS instrument and data acquisition parameters Excimer 193−nm ArF laser GeoLasPro Output Energy 150 mJ Energy Density on sample ~ 7 − 10 J/cm2 Repetition Rate 5 Hz Pit Size Between x and x µm Ablation Cell Volume ~ 1.5 cm3 Cell Gas Flow (He) ~ 1 L/min

Agilent 7500ce quadrupole ICP−MS Auxiliary gas flow 1.03 l/min Ar RF power 1500 V Detector Mode Dual 8 orders of magnitude linear dynamic range Quadrupole Settling Time 2 ms

Data acquisition parameters Sweeps per reading 1 Reading per replicate 200 − 300 Replicates 1 Dwell time per isotope 10 ms Points per peak 1 per measurement External Standard NIST610 glass Spot size 60 µm Isotopes Analyzed Li (7), Be (9), B (11), Na (23), Mg (25), Al (27), Si (28),

K (39), Ca (40), Sc (45), Ti (49), V (51), Cr (52), Mn (55), Fe (56), Co (59), Ni (62), Cu (63), Zn (66), Ga (69), As (75), Rb (85), Sr (88), Y (89), Zr (90), Nb (93), Ag (107), Sn (120), Sb (121), Cs (133), Ba (138), La (139), Ce (140), Pr (141), Nd (143), Sm (147), Eu (153), Gd (157), Dy (163), Er (166), Yb (172), Lu (175), Hf (178), Ta (181), Au (197), Pb (208), Th (232), U (238)

4

Fig. 1 - Precision of the analyses reported as relative standard deviations (RSD %) for each of the elements

that had concentrations above LOD in standards BCR-2G, BHVO-2G, BIR-1G, NKT-1G, NIST612 and NIST614. The blue area corresponds to an RSD of <5 %, and the red box corresponds to RSD ≤10 %.

Precision

0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0 45.0 50.0

Li*(7)Be*(9)B*(11)

Na*(23)Mg*(25)

Al*(27)Si*(28)K*(39)

Ca*(40)Sc*(45)Ti*(49)V*(51)

Cr*(52)Mn*(55)Fe*(56)Co*(59)Ni*(62)Cu*(63)Zn*(66)Ga*(69)As*(75)Rb*(85)Sr*(88)Y*(89)

Zr*(90)Nb*(93)

Ag*(107)Sn*(120)Sb*(121)Cs*(133)Ba*(138)La*(139)Ce*(140)Pr*(141)

Nd*(143)Sm

Eu*(153)Gd*(157)Dy*(163)Er*(166)Yb*(172)Lu*(175)Hf*(178)Ta*(181)Au*(197)Pb*(208)Th*(232)U*(238)

RSD %

BCRT2GBH

VOT2G

BIRT1GN

ISTT612N

ISTT614N

KTT1G

5

Fig. 2 - Accuracy of the analyses reported as the average percent error for each of the elements that had concentrations above LOD in standards BCR-2G, BHVO-2G, BIR-1G, NKT-1G, NIST612 and NIST614. The blue area corresponds to an average percent error of ±5 %, and the red box corresponds to an average percent error of ±15 %.

Accuracy

!40.0 !30.0 !20.0 !10.0 0.0 10.0 20.0 30.0 40.0

Li*(7)Be*(9)B*(11)

Na*(23)Mg*(25)

Al*(27)Si*(28)K*(39)

Ca*(40)Sc*(45)Ti*(49)V*(51)

Cr*(52)Mn*(55)Fe*(56)Co*(59)Ni*(62)Cu*(63)Zn*(66)Ga*(69)As*(75)Rb*(85)Sr*(88)Y*(89)

Zr*(90)Nb*(93)

Ag*(107)Sn*(120)Sb*(121)Cs*(133)Ba*(138)La*(139)Ce*(140)Pr*(141)

Nd*(143)Sm

Eu*(153)Gd*(157)Dy*(163)Er*(166)Yb*(172)Lu*(175)Hf*(178)Ta*(181)Au*(197)Pb*(208)Th*(232)U*(238)

Average % Error

BCR!2GBH

VO!2G

BIR!1GN

IST!612N

IST!614N

KT!1G

6

Fig. 3 – Ratios of Measured over reference values for USGS Basaltic glass BCR-2G. Shaded area represents ±10% variation. Red box represents ±20% variation. Reference values are from GeoReM (http://georem.mpch-mainz.gwdg.de/). Some ratios were not plotted either because the element was not detected or a reference value was not available (cfr. Table 1).

Measured / STD

7

Fig. 4 – Ratios of Measured over reference values for USGS Basaltic glass BHVO-2G. Shaded area represents ±10% variation. Red box represents ±20% variation. Reference values are from GeoReM (http://georem.mpch-mainz.gwdg.de/). Some ratios were not plotted either because the element was not detected or a reference value was not available (cfr. Table 1).

Measured / STD

8

Fig. 5 – Ratios of Measured over reference values for USGS Basaltic glass BIR-1G. Shaded area represents ±10% variation. Red box represents ±20% variation. Reference values are from GeoReM (http://georem.mpch-mainz.gwdg.de/). Some ratios were not plotted either because the element was not detected or a reference value was not available (cfr. Table 1).

Measured / STD

9

Fig. 6 – Ratios of Measured over reference values for USGS Nephelinite glass NKT-1G. Shaded area represents ±10% variation. Red box represents ±20% variation. Reference values are from GeoReM (http://georem.mpch-mainz.gwdg.de/). Some ratios were not plotted either because the element was not detected or a reference value was not available (cfr. Table 1).

Measured / STD

10

Fig. 7 – Ratios of Measured over reference values for NIST glass NIST612. Shaded area represents ±10% variation. Red box represents ±20% variation. Reference values are from GeoReM (http://georem.mpch-mainz.gwdg.de/). (cfr. Table 1).

Measured / STD

11

Fig. 8 – Ratios of Measured over reference values for NIST glass NIST614. Shaded area represents ±10% variation. Red box represents ±20% variation. Reference values are from GeoReM (http://georem.mpch-mainz.gwdg.de/). (cfr. Table 1).

Measured / STD

B (6.6)

Sc (55.6)

12

Fig. 9 – Typical example of LAICPMS signal collected on the USGS basaltic glass BHVO-2G (analyzed as unknown). Shaded areas represent integration regions. Typically we select 50s for background signal and 40s or more for sample signal. (for more details: http://www.geochem.geos.vt.edu/fluids/laicpms/images/PosterWeb.pdf)

1 10

100

1000

1000

0

1000

00

1000

000

0 20

40

60

80

10

0 12

0

Counts

Tim

e (s

)

Li7$

Be9$

B11$

Na23$

Mg25$

Al27

$

Si28$

K39$

Ca40

$

Sc45$

Ti49$

V51$

Cr52

$

Mn5

5$

Fe56$

Co59

$

Ni62$

Cu63

$

Zn66$

Ga69

$

BH

VO

-2G

13

Table 1 – Data for Fig. 3 to 8

14

Table 1 – Cont.

All elemental concentrations are in (ppm), except in RED -> %oxide [Na2O, MgO, Al2O3, SiO2, K2O, CaO, FeO]. When the numbers for these elements are in blue then concentrations are expressed as elemental (ppm) due to low amount. ND = Not detected; NR = Concentration not reported in the literature. All reference concentration data are from GeoReM (http://georem.mpch-mainz.gwdg.de/). Avrg (5) = average of 5 measurements. STD = Reference value

15

Table 2 - Basic Statistic: STDEV = Standard Deviation - %RSD = Relative Standard Deviation LOD = Limits of Detection

16

Table 2 – Cont.

17

Table 2 – Cont.

18

Table 3 – Color coded representation of the analytical variations presented in Table 1 (Measured/STD) and in graphs 3 to 8.

19

Table 4 – Data for Fig. 1 and 2