sample analysis designasimonet/engv60500/fall 2013/lectures... · sample analysis design –...

Sample Analysis Design – Isotope Dilution

•  Most accurate and precise calibration method available

•  Requires analyte with two stable isotopes

•  Monoisotopic elements cannot be determined via isotope dilution

•  Spike natural sample with enriched isotope spike of analyte


•  The amount of spike is selected so that the resulting ratio between spiked isotope and unspiked isotope is near unity – maximizes precision

•  Typically use the most abundant isotope as the reference -- maximizes sensitivity


•  Check isotope ratio in unspiked sample to determine if the “natural ratio” in the sample matches with the predicted ratio

•  If not -- interference in acting on one or both of the isotopes

•  Always attempt to use interference free isotopes


•  Prepare the spike to desired concentration

•  Add spike as early as possible – after equilibration of spike and sample you don’t have to have complete sample recovery

•  During any stage of the process complete equilibration is absolutely necessary


•  Analyze the solution on the ICP using many repetitive scans (to maximize precision)

•  Need to measure isotopic ratios on standards of a known ratio in order to correct for machine mass discrimination

•  Use previous equation to calculate concentrations!


•  Advantages:

–  Most accurate and precise method for quantitative elemental concentrations

–  Partial loss of analyte during preparation is compensated for since physical and chemical interferences are not an issue -- will cancel out as they will affect each isotope identically

–  Ideal form of internal standardization since another isotope of the same element is used in this capacity


•  Disadvantages:

– Generally only applicable to multiple-isotopic elements

– Need an enriched isotope spike for the analyte of interest - not always available or sometimes at very high cost

– Need two interference free isotopes – VERY time consuming

Sample Analysis Design

STEP 3 – INTERNAL STANDARDIZATION & INSTRUMENT

DRIFT CORRECTION

Sample Analysis Design – Internal Standard

•  Every sample should be analyzed with an internal standard (IS)

•  What is an internal standard (IS)?

–  element that is added to EVERY sample/ blank/calibration standard/QA sample/etc., that is not expected to be in the sample in appreciable quantities and is not an element of interest

–  use IS to monitor machine drift (both short and long term) and matrix effects


•  Choice of IS depends upon which elements you are quantifying

•  The IS should have similar properties in the plasma as element(s) of interest

•  ICP-MS: similar in mass/ionization potential


•  Example:

–  attempting to quantify U - use Th

–  attempting to quantify most transition metals - use As

–  attempting to quantify REEs - use Re

–  115In and 103Rh are common IS for general use

–  alternatively, you can add several IS to each sample


•  From previous slide, we assume that samples have little or no Th, As, or Re

•  It’s important to have an idea of what’s in your sample prior to quantitative analysis

•  Solid samples can use a naturally occurring element as IS, provided that you know the concentration in each sample


•  Procedure for IS use:

•  Calculate the concentration of the IS in each centrifuge tube – the latter will contain an aliquot of your sample and an aliquot of the IS

•  Divide the measured ion signal (CPS) by the concentration of your IS to derive the factor = CPS/ppb

•  Divide CPS/ppb of each tube by the CPS/ppb for those measured for the blanks since these are not influenced by possible effects due to sample matrices

•  The latter yields a dimensionless correction factor (I refer to it as a normalization factor)

•  Use correction factor to adjust analyte counts for drift or matrix effects


•  Advantages:

•  Fluctuations are monitored in each sample/ calibration / blank

•  Disadvantages:

•  Assume that behavior of IS is the same as the analyte

Sample Analysis Design – Instrumental Drift

•  Correct for instrument drift with:

•  Internal standardization is a common procedure

•  Use of drift corrector solutions (DCS)


•  Drift Corrector Solutions (DCS):

•  Measure the same solution intermittently throughout the course of the analytical session

•  Change in ion signal is assumed to be linear between each DCS measurement


•  The DCS should contain all elements of interest and can be matrix matched to samples

– Example: use standard reference materials (SRMs) for DCS


•  Apply a linear correction to samples between DCS solutions

•  DCS1 + ((DCS2 - DCS1)*F)

•  F = position dependent fraction


•  Advantages of DCS correction:

–  all analytes are monitored for drift

–  nothing added to sample solutions

•  Disadvantages of DCS correction:

–  assume change is linear

–  cannot easily monitor matrix effects

Sample Analysis Design – Background & blanks

•  Standard blank - blank used to monitor polyatomic ion interferences, gas peaks, and contamination from reagents; used for background subtraction

•  Procedural blank - blank used to monitor contamination acquired during all stages of sample preparation; grinding, digestion, acidification, powdering, etc


•  Use of blanks during an analytical session:

•  ALWAYS begin an analytical session with at least one standard blank

•  Analyze standard blanks periodically throughout the course of the session in particular to monitor memory effects

•  Process and analyze at least one procedural blank at some point during your research study; for its analysis, it’s preferable to measure it early in order to avoid any potential memory effects


•  The more standard blanks that are run during an analytical session, the more information you will have with regards to monitoring change(s) in background levels throughout the entire session


•  How to determine “the background”:

•  1. just use the first standard blank

•  2. average all standard blanks

•  3. take median of all standard blanks

•  4. apply statistical analysis to standard blanks and select some of them


•  Outlier tests:

•  1. I know the truth

•  2. Looks different

•  3. Statistical “proof”


•  Option 1 should be avoided - unscientific and invalid

•  Option 2 is better but only if the measurement is repeated

•  Option 3 is the best approach, but needs to be carried out carefully in order to avoid false negatives and positives


•  Huber Outlier Test

•  take median of all values

•  calculate absolute deviation |xi - xm|

•  take mean of absolute deviations (MAD)

•  multiply MAD by coefficient (k = 3-5)

•  anything higher than k*MAD is rejected as outlier


•  Calculation of Limit of Detection (DL) and Limit of Quantification (QL)

•  Easy way: •  LOD = 3*STDEVblank; •  LOQ = 10*STDEVblank

Sample Analysis Design – SUMMARY

•  A ‘good’ analytical method will:

•  1. provide the means to calculate an accurate background level

•  2. allow for correction of instrument drift

•  3. use Internal standardization to monitor matrix effects

•  4. provide some method for monitoring/ correcting interferences

•  5. Use a proper calibration strategy


PART II

Sample Analysis Design •  Generating high quality, validated results is the

primary goal of elemental abundance determinations

•  It is absolutely critical to plan an ICP-MS analysis carefully – from sample gathering to final analysis on the ICPMS

•  Always create and follow an analysis design that shall permit you to follow the procedure for future samples, AND know what you did when looking at old data

Sample Analysis Design •  Goal: high quality, validated, quantitative determination

of elemental concentrations/isotope ratios

•  This goal can be achieved by:

•  Specificity - only detecting isotope(s) of interest, not interferences, OR making sure that matrix effects do not play a role

•  Sensitivity - can we differentiate the isotope of interest from the background signal?

•  Accuracy - does the analysis represent the real value?

•  Precision - how repeatable is the value?


•  Sensitivity and Precision: – generally defined by how well the instrument

is tuned (assuming dilution factor is appropriate)

•  Specificity and Accuracy: – generally governed by how well you prepare

your sample and set up your analysis method/parameters


•  Spectroscopic interferences

•  Matrix effects

Sample Analysis Design •  Interference – analytical artifact which causes

the ion signal in a sample to vary from an idealized (true) signal –  Interferences cause inaccurate analyses

•  Two main types:

– 1. Spectroscopic

– 2. Nonspectroscopic

Sample Analysis Design •  Spectroscopic Interferences in ICP-MS

– 1. Isobaric overlap

– 2. Polyatomic ions

– 3. Refractory oxides

– 4. Doubly charged ions

Sample Analysis Design •  Isobaric Overlap

•  Two different elements having the same nominal mass

•  E.g., 40Ca (96.9%) and 40Ar (99.6%)

•  Monoisotopic elements:

•  B, Na, Al, Sc, Mn, As, Nb, Y, Rh, I, Cs, Pr, Tb, Ho, Tm, Au, Th

•  None have isobaric interferences

Sample Analysis Design – Isobaric Overlap

•  A few multi-isotope elements have no interference free isotopes:

–  E.g., In - 113In → 113Cd, and 115In → 115Sn

•  Usually possible to choose an interference free isotope or one that is unlikely to be interfered upon strongly (either due to very low natural abundance of interfering element (isotope) or the interfering element (isotope) is not present in sample)

Sample Analysis Design – Isobaric Overlap

•  IF an isobaric interference is unavoidable, you can correct for it by measuring the counts on a non-interfering isotope of the interfering element

•  CPSanalyte = CPStotal -CPSinterferent*(Xa/Xb)

•  Xa = abundance of interfering isotope •  Xb = abundance of non-interfering isotope

Sample Analysis Design – Polyatomic Interferences

•  More serious than isobaric interferences

•  Result from possible, short-lived combination of atomic species in the plasma or during ion transfer

•  Common recombinants are Ar, H, and O

•  Dominant elements in reagents also form polyatomic interferences - N, S, and Cl


•  Analyzing deionized water with ICP-MS instrument – we have solely H and O present (within the ‘matrix’)

•  Thus, peaks would be visible at masses:

–  41 (40Ar + 1H) → 41K

–  56 (40Ar + 16O) → 56Fe

–  80 (40Ar + 40Ar) → 80Se

–  As well as other minor peaks from the minor isotopes of Ar and O


•  If you acidify the deionized water with HNO3 or H2O2 , these behave in the same manner as deionized water

–  these acids are considered “ideal matrices” as they don’t add unnecessary polyatomic interferences

•  Acidification with HCl or H2SO4 changes the situation since Cl and S can be used to form polyatomic ions


•  HCl matrix:

– 35Cl + 16O = 51V

– 37Cl + 16O = 53Cr

– 40Ar + 35Cl = 75As

– 40Ar + 37Cl = 77Se


•  H2SO4 matrix:

– 32S + 16O = 48Ti

– 32S + 16O + 16O, or 32S + 32S = 64Zn, 64Ni

– 34S + 16O = 50Ti, 50V, 50Cr

– 33S + 16O + 16O = 65Cu

Sample Analysis Design •  These interferences will be present just from the

gas and solvent

•  Therefore, if possible, solutions should be prepared with a weak (1-5% v/v) HNO3 acid matrix

•  Why not deionized water? –  acid helps keep elements from sticking to sides of test

tubes and tubing during transport –  ionization efficiency is significantly lower relative to

HNO3


•  The most serious polyatomic interferences are formed from the most abundant isotopes of H, C, N, O, Cl, and Ar

•  So, if polyatomic interferences are present in plain deionized water with no other ions present, what happens in natural (complex) samples/matrices?


•  Need to know the most abundant elements present in sample

–  Is (are) there any one (or more) elements present that are in very high abundances?

•  If so - then it’s likely interferences could be formed from such abundant element(s)

•  E.g., rock samples - high Si


•  Typically, most polyatomic interferences are only present below mass 82 - since Ar, H, and O are by far the most abundant isotopes in the plasma, they form most interferences

•  Counteractive measures:

–  Proper machine settings (especially nebulizer flow rate and RF power) can minimize formation


•  If a polyatomic interference is unavoidable, then a correction equation like the one for isobaric interferences can be used (assuming you are certain of the ‘species’ of the interference)

Sample Analysis Design – Refractory Oxides

•  Refractory Oxides

•  Occur because of incomplete sample dissociation or from recombination

•  Always occur as an interference as an integral value of 16 mass units above the interfering element

•  MO, MO2, or MO3 where M is the interfering element


•  Elements with high oxide bond strength are most likely to form refractory oxide interferences

•  Severity is expressed as MO/M as a percentage

•  MO/M should be as low as possible, typically <5%

•  MO/M minimized by adjusting nebulizer flow rate, z-axis position, and RF power


•  Si, Ce, Zr, Ti, Sm, Mo, and P all form strong oxide bonds and may have severe oxide interferences

•  Usually monitor the CeO/Ce ratio to make sure it’s below 5%


•  Examples:

•  LREE on the HREE

•  BaO on Eu

•  MoO on Cd


•  Again, correction equations can be applied but these lead to increased error, especially if the interference is large

•  Oxide formation tends to not be stable and the corrections accumulate error quickly

•  If there is a severe oxide interference, it’s usually better to try to separate the interference from the analyte


•  Typically,

– MO > MO2 > MO3 > MO4

Sample Analysis Design – Doubly Charged Ions

•  Form when the 2nd ionization potential is less than the first ionization potential of Ar

•  Typical of alkaline earths, a few transition metals, and some REE

•  Low nebulizer flow rates increase doubly charged ion formation

•  Interference signal is always M/2 where M is the element becoming doubly charged

How to avoid spectral MS interferences?

•  Attempt to ‘dry’ solutions prior to introduction to plasma - remove O and H gets rid of many polyatomic species (i.e. use a desolvating introduction system – e.g. DSN 100)

•  Optimize instruments settings so that formation is minimized and corrections are not severe

•  Use simple matrices whenever possible

•  Process samples prior to analysis to isolate elements of interest or remove potential interfering elements; e.g. ion exchange chromatography

How to avoid spectral MS interferences?

•  Alternatively - use higher mass resolutions, MR – medium resolution or HR – high resolution

•  Most overlaps are not exact using medium and high resolution modes

•  Can completely separate the isotope of interest from interfering species

•  However, consequence is loss of sensitivity

sample analysis designasimonet/engv60500/fall 2013/lectures... · sample analysis design –...

Documents