Antonio  Simonetti   University  of  Notre  Dame   —  Precision and accuracy on individual isotopic measurements suffer due to matrix effects and isobaric interferences? —  Availability of suitable reference materials, in relation to matrix-matching in studies of mass-dependent isotopic fractionation, and laser-induced isotopic fractionation —  Factors that contribute to the accuracy and precision of in-situ measurements- —  Interplay related to sample, laser operating conditions and processes in the mass spectrometer Pearson et al. (2008) —  Fall into 2 groups: —  Measurement of radiogenic isotopes for trace elements in common rock-forming minerals; e.g., Sr in clinopyroxene, carbonate, feldspar; Hf in zircon; Pb in feldspar, clinopyroxene —  Study of mass-dependent isotopic fractionation of elements that are major constituents in minerals of interest; e.g., Cu in chalcopyrite, Fe in pyrite —  Size of mineral grain and sensitivity of mass spectrometer will dictate the sampling strategy —  Use ‘rastering’ so as to maximize volume of sample ablated and therefore improve precision (i.e. higher ion signal) and ‘homogenize’ the sample —  Use ‘stationary’ analysis, mimic a microprobe and hence attempt to decipher internal variations combined with trace element data, images, etc —  Probably is the most important factor affecting the accuracy and external precision —  Mass bias is isotopic fractionation (i.e. artificial change in isotope ratios) produced by variable transmission of the ion beam in mass spectrometer —  In the MC-ICP-MS instrument, this occurs primarily in the plasma and interface regions —  For MC-ICP-MS instruments, —  The ‘measured’ isotope ratio > the ‘true’ ratio —  E.g., measured 87Sr/86Srsample > ‘true’ 87Sr/86Srsample —  For TIMS (thermal ionization mass spectrometers) instruments, —  The ‘measured’ isotope ratio < the ‘true’ ratio —  E.g., measured 87Sr/86Srsample < true 87Sr/86Srsample —  In-­‐situ  analyses  are  also  affected  by  the  “matrix”  of  the   samples,  which  may  produce  isotopic  interferences   (isobaric  and  molecular  overlaps)  –  these  also  need   to  be  corrected  (if  possible)!   —  Isobaric  (equivalent  atomic  mass)  interferences  in   radiogenic  isotope  systems  include:   —  87Rb è 87Sr —  144Sm è 144Nd —  176Lu è 176Hf —  For example, matrix-based molecular interferences include: —  40Ca + 31P + 16O = mass 87 = 87Sr —  The  principal  inorganic  component  of  enamel  is   Hydroxyapatite    -­‐  3Ca3(PO4)2.CaX  [where  X  can   represent  a  mixture  of  F,  Cl,  CO2,  OH] —  In-situ LA-MC-ICP-MS analyses of fossilized teeth from human remains for their Sr isotope composition (Simonetti et al., 2008) —  Tracing migration patterns for ancient civilizations Simonetti  et  al.  (2008,  Archaeometry)   Simonetti  et  al.  (2008,  Archaeometry)   —  Magnitude (i.e., deviation of measured ratio compared to the ‘true’ value) of mass bias for MCICP-MS instrument is larger than that associated with TIMS (thermal ionization mass spectrometry) —  However, the same fractionation laws are applied to correct for instrumental mass bias —  Previous studies have shown that instrumental mass bias can be corrected using a “generalized power law” —  Rtrue = Rmeas Ÿ f (M 2 – M 1 ) —  Rtrue = true isotope ratio of the two isotopes of mass M1 and M2 (M2/M1) —  Rmeas = is the isotope ratio (M2/M1) measured by the mass spectrometer —  f = mass fractionation coefficient —  Equation can be rewritten into “Exponential Law” form: —  Rtrue = Rmeas Ÿ (M2 / M1) f —  Correction of instrumental mass bias is best achieved by “internal normalization” —  Mass fractionation coefficient (f) can be determined using a pair of stable isotopes with a known or “true” isotopic ratio —  E.g., —  86Sr/88Sr = 0.1194 —  146Nd/144Nd = 0.7219 —  179Hf/177Hf = 0.7325 —  205Tl/203Tl = 2.3871 Nu  Plasma  II  -­‐  Collector  Configuration   Rtrue = Rmeas Ÿ (M2 / M1) f Pearson et al. (2008) Pearson et al. (2008) —  Linear relationship on log-log plots also holds for when isotope ratios of different elements are plotted, which formed the basis for the ‘external’ or ‘doped’ correction procedure. —  This technique has been successfully applied to a number of applications as shown by previous investigations —  Cu-Zn (e.g., Maréchal et al., 1999) —  Pb-Tl (e.g., Belshaw et al., 1998; Woodhead, 2002) Pearson et al. (2008) Pearson et al. (2008) Pearson et al. (2008) —  This method measures standards (samples) of known isotopic composition interspersed with unknown samples to monitor change in instrumental mass bias over time —  This technique is used when there are no internal isotope pairs (of constant ratio) or the addition of an external dopant (i.e., element of similar mass that is a nonisobaric overlap) is not a viable option (e.g., Li, Mg) —  The mass bias value for the unknown sample is interpolated from the inferred mass bias values obtained from a pair of standard runs (one before, and one following the sample) —  Assumptions here are that: - the mass bias changes uniformly with time - matrices of sample and standard are identical; this may be of critical importance for laser ablation analyses —  For accurate measurements of in-situ radiogenic isotope ratios such as 87Sr/86Sr, 143Nd/144Nd, and 176Hf/177Hf, we need to address the issues of mass bias correction and evaluate isobaric interferences —  The mass bias corrections for Sr, Nd, and Hf are relatively straightforward as each element has a pair of ‘interference-free’ stable isotopes that can be used for normalization —  E.g., 86Sr/88Sr = 0.1194, 146Nd/144Nd = 0.7219, 179Hf/ 177Hf = 0.7325 —  However,  isobaric  interferences  in  radiogenic  isotope   systems  include:   —  87Rb è 87Sr (effects 87Sr/86Sr) —  144Sm è 144Nd (effects 143Nd/144Nd) —  176Lu è 176Hf (effects 176Hf/177Hf) —  Thus,  we  will  discuss  the  various  approaches  that  can   be  used  to  overcome  the  problem  of  mass  bias   correction  for  the  isobaric  interferences  using  the  Lu-­‐ Hf  isotope  system  as  an  example   Nu  Plasma  II  Collector  Configuration   Nu Plasma II -In-situ Hf, collector configuration University of Notre Dame —  METHOD 1: —  Assume ƒHf = ƒLu = ƒYb —  This can be evaluated by doping solutions of JMC 475 Hf standard with Yb and Lu (using desolvating nebulizing system – ‘dry plasma’) —  The isotope compositions of the Yb and Lu are varied until the 176Hf/177Hf are accurate —  Assume that mass bias for ‘dry plasma’ are identical as those for laser ablation analyses and instrument conditions are stable —  METHOD 2: —  ƒHf , ƒYb are measured independently —  Measure 173Yb/171Yb, 172Yb/171Yb ratios, then assume 176Yb/ 172Yb = 0.5865 (Segal et al., 2003) —  Natural abundances of 171Yb and 173Yb are 14.28% and 16.13%, respectively. This translates into low ion signals (≤ 20 millivolts) —  Hence, the uncertainty associated with the measurement in the ƒYb will affect and contribute to the uncertainty on the corrected 176Hf/177Hf ratio —  METHOD 2: —  Wu et al. (2006) adopted using the mean 173Yb/171Yb ratio to calcuate the mean ƒYb for each analysis, and they showed a two-fold improvement in the precision of the 176Hf/177Hf isotope measurements. —  METHOD 3: —  ƒLu = ƒYb —  Assume a constant relationship between the ƒLu and ƒYb over a long period of time and between different types of analyses (both ‘dry plasma’ vs. laser ablation modes) Pearson et al. (2008) —  Choice of method for the correction of an isobaric overlap requires careful assessment of the mass bias both in terms of within-run variations, analytical precision, and longer term instrument stability —  In order to do this, it is important to consider the factors that control instrumental mass bias —  Fundamental previous studies have shown that mass bias is generated within the plasma, within the interface region (between sample and skimmer cones), and immediately behind skimmer cone —  Physical properties of the plasma control the vaporization, atomization and ionization of the sample —  Parameters such as gas kinetic temperature and electron density parameters have a significant effect on diffusion rates in the ICP and kinetic energy of the ion transmitted through the interface —  It is commonly accepted that only ions from the central channel of the plasma can be effectively sampled into the mass spectrometer —  Processes that result in preferential vaporization of light isotopes from dry aerosol particles or mass-dependent diffusion contribute to mass bias and element fractionation —  Temperature in central channel and in front of the sample cone orifice is critical to the degree of ionization The  mutual  repulsion  of  ions  of  like  (similar)   charge  limits  the  total  number  of  ions  that  can   be  compressed  into  a  beam  of  given  size   —  Which instrument operating parameters exert an important control on the mass bias? —  vs. —  Those produced by the laser ablation process itself (plasma loading)? —  Instrumental parameters that effect instrumental mass bias include: —  Nebulizer gas flow —  Extraction lens voltage —  RF power —  Torch Depth Position —  Cone design and condition —  Nebulizer gas composition Pearson et al. (2008) Pearson et al. (2008) Pearson et al. (2008) Pearson et al. (2008) —  Laser Induced Isotopic Fractionation: —  Particle size distribution is important – incomplete ionization of larger particles in the plasma results in increased transmission of lighter isotopes (Jackson & Günther, 2003; Kühn et al., 2007, Günther & Koch, 2008) —  Although the laser is controlling the particle size distribution, the isotopic fractionation is occurring within the plasma —  Laser Induced Isotopic Fractionation: —  Intrinsic physics of the laser design and operating conditions are considered to be the most important factors that contribute to LIEF —  Laser pit width, wavelength, repetition rate, and energy density —  Experiments conducted using UP213 laser ablation system and Mud Tank zircon Pearson et al. (2008) Pearson et al. (2008) —  Laser Induced Isotopic Fractionation: —  “Mass (Plasma) Loading” – amount of material introduced into the ICP, has been shown to have a large effect on elemental fractionation —  The influence of plasma loading can be evaluated in a plot of mass bias coefficient ‘ƒ’ versus total ion signal (volts) —  Hopefully, you will not see a correlation between ƒ and the corrected isotope ratios of interest! —  Laser Induced Isotopic Fractionation: —  “Matrix Effects” – Several previous studies indicate there is an important effect of the sample matrix on the observed mass bias —  However, Pearson et al. (2008) clearly show that although different concentrations of matrix elements clearly effect the calculated mass bias coefficient “ƒ”, the varying matrix conditions did not have an impact on the calculated isotope ratio of interest! —  There  are  3  categories  of  factors  that  effect  the   accuracy  and  precision  of  in  situ  isotope  ratio   measurements  by  LA-­‐MC-­‐ICP-­‐MS:   —  SAMPLE   —  MASS  SPECTROMETER   —  LASER  ABLATION  SYSTEM   —  Accuracy  is  mainly  a  function  of:   —  Correction  procedures  for  mass  bias   —  Isobaric  interferences   —  Sample  matrix  effects   —  Precision  is  a  function  of:   —  Ion  signal  intensity  –  which  is  dependent  on   concentration  of  element  in  sample,  laser  energy,  spot   size,  mass  spectrometer  sensitivity,  and  time  of  laser   ablation  analysis   —  Correction  for  mass-­‐dependent  instrumental  mass   bias  ‘ƒ’  is  the  most  critical  factor  controlling  the   accuracy!!   —  Instrument  operation  parameters  that  affect  mass  bias   for  in-­‐situ  laser  ablation  analysis  include:   —  Nebulizer  gas  flow  rate   —  RF  power   —  Voltages,  torch  position,  and  cone  design   —  Matrix  (sample-­‐related)   —  Laser  operating  conditions   —  The  interplay  of  all  these  factors  that  influence  the   accuracy  and  precision  emphasizes  the  need  for   reference  materials  that  are  well  characterized,  in   terms  of  their  isotopic  composition  AND  their   major  and  trace  element  constituents   —  Trace  ancient  crustal  differentiation  events   —  Hydrothermal  activity/diagenetic  processes   —  Trace  magmatic/mantle  processes   —  Nd  isotope  analyses  by  ID-­‐TIMS  (isotope  dilution-­‐ thermal  ionization  mass  spectrometry)  using  whole   rocks  generates  high  precision  results;  however,   isotopic  heterogeneities  at  the  scale  of  individual   minerals  is  lost   —  However,  compared  to  a  ID-­‐TIMS  analysis,  1000  times   less  sample  is  consumed  for  a  typical  LA-­‐MC-­‐ICP-­‐MS   result,  therefore  application  of  this  method  is  dictated   by  the  absolute  concentrations  of  Nd  (ppm)  in  the   sample   —  Hence,  the  in-­‐situ  Nd  isotopic  method  is  applicable  to   LREE-­‐  enriched  accessory  minerals  –  most  notably   —  Monazite,  allanite,  titanite,  and  apatite   —  Naturally,  the  interesting  aspect  of  applying  the  in-­‐ situ  Nd  isotope  technique  to  minerals  such  as  apatite   and  monazite,  is  that  you  can  also  obtain  in-­‐situ  U-­‐Pb   ages  for  these  phases   —  Hence,  this  combination  of  U-­‐Pb  ages  and  Nd  (Sr,  Pb)   isotope  ratios  becomes  a  powerful  technique  for   tracing  a  variety  of  igneous  processes        143Nd/144Nd  =    143Nd/144Nd0  +  147Sm/144Nd  (eλt  –  1)   147Sm  decays  to  143Nd   λ  =  6.54  x  10-­‐12   —  1-­‐  Removal  of  the  isobaric  interference  of  144Sm  on   144Nd   —  Assessing  the  accuracy  of  corrected  ratios   —  Obtaining  an  accurate  147Sm/144Nd  ratio   McFarlane & McCulloch (2008) —  Sm/Nd  ratios  in  REE-­‐bearing  accessory  minerals  can   be  quite  variable  but  typically  vary  between  0.05   (monazite,  allanite)  to  1.0  (xenotime)   —  Always  have  isobaric  interference  of  144Sm  on  144Nd,   thus  no  choice  but  to  strip  this  overlap  in  order  to   obtain  accurate  143Nd/144Nd  ratios   —  However,  Sm  correction  is  complicated  since  the   146Nd/144Nd  ratio  is  used  for  the  internal   normalization,  which  is  also  affected  by  the   interference   —  This  interference  can  be  addressed  in  several  ways:   —  1-­‐  by  using  an  iterative  approach  as  outlined  by  Foster   &  Vance  (2006);   —  2-­‐  using  146Nd/145Nd  (normalized  to  146Nd/144Nd   measured  using  pure  Nd  solutions)   —  3-­‐  Monitoring  an  additional  interference  free   invariant  ratio  such  as  147Sm/149Sm  to  correct  144Sm/ 149Sm  –  then  applying  this  to  the  146Nd/144Nd  for  mass   bias  correction  purposes  (McFarlane  &  McCulloch,   2007)     Iterative  approach  -­‐  Foster  &  Vance  (2006);   —  Use  the  NIST  SRM  610  international  glass  standard  to   establish  144Sm/147,149Sm  working  values  in  order  to   obtain  the  appropriate  Nd  isotope  values   —  Isotopic  homogeneity  of  reference  standard  is  critical,   especially  for  samples  with  Sm/Nd  >0.1     McFarlane  &  McCulloch  (2007)   —  Invariant  ratios  for  Sm  ratios  were  measured  with  pure   Sm  solutions   —  Interference  of  144Sm  on  144Nd  is  then  corrected  by   monitoring  149Sm  and  using  a  value  for  144Sm/149Sm   obtained  by  doping  Nd  standard  solutions  with  variable   Sm  contents  and  iteratively  refining  the  144Sm/149Sm  to   give  the  true  Nd  isotope  values   McFarlane & McCulloch (2008) McFarlane & McCulloch (2008) McFarlane & McCulloch (2008) McFarlane & McCulloch (2008) McFarlane & McCulloch (2008) McFarlane & McCulloch (2008)