﻿JUNE 
  4, 
  1922 
  PROCEEDINGS 
  : 
  PHILOSOPHICAL 
  SOCIETY 
  267 
  

  

  Knowledge 
  of 
  the 
  thermoelectric 
  effect 
  of 
  small 
  traces 
  of 
  impurities 
  is 
  

   limited. 
  A 
  thermoelectric 
  difference 
  is 
  irrefutable 
  evidence 
  of 
  difference 
  of 
  

   the 
  two 
  metals. 
  (These 
  should 
  be 
  carefully 
  annealed 
  so 
  that 
  the 
  physical 
  

   condition 
  is 
  constant.) 
  No 
  difference, 
  that 
  is 
  a 
  negative 
  result, 
  raises 
  doubt. 
  

   Perhaps 
  more 
  than 
  one 
  impurity 
  gives 
  opposite 
  effects. 
  

  

  Special 
  study 
  has 
  been 
  made 
  of 
  pure 
  platinum, 
  pure 
  gold 
  and 
  pure 
  palladium, 
  

   the 
  purity 
  being 
  of 
  the 
  highest 
  order. 
  In 
  the 
  three 
  cases 
  the 
  purest 
  samples 
  

   as 
  selected 
  by 
  other 
  tests 
  have 
  been 
  the 
  most 
  negative 
  thermoelectrically. 
  

   As 
  an 
  illustration 
  of 
  the 
  magnitude 
  of 
  the 
  e. 
  m. 
  f 
  's. 
  afforded 
  by 
  traces 
  of 
  impurity, 
  

   two 
  pieces 
  of 
  platinum 
  one 
  spectroscopically 
  free 
  of 
  any 
  detectable 
  impurity 
  

   and 
  one 
  indicating 
  the 
  faintest 
  trace 
  of 
  calcium 
  (or 
  lime) 
  as 
  an 
  impurity, 
  

   gave 
  a 
  thermal 
  e. 
  m. 
  f. 
  of 
  about 
  twenty 
  micro-volts 
  at 
  1200° 
  C. 
  Probably 
  

   there 
  is 
  a 
  very 
  wide 
  variation 
  in 
  the 
  effect 
  of 
  constant 
  amount 
  of 
  different 
  

   impurities. 
  

  

  As 
  a 
  beginning, 
  a 
  series 
  of 
  platinum 
  rhodium 
  alloys, 
  containing 
  respec- 
  

   tively 
  0.001%, 
  0.01%, 
  0.1%, 
  0.5%, 
  and 
  1.0% 
  of 
  rhodium 
  were 
  made. 
  Spec- 
  

   troscopic 
  examination 
  gave 
  no 
  evidence 
  of 
  impurity. 
  The 
  e. 
  m. 
  f 
  . 
  of 
  each 
  alloy 
  

   against 
  platinum 
  at 
  1083° 
  C. 
  was 
  measured. 
  It 
  was 
  found 
  that 
  the 
  e. 
  m. 
  f. 
  was 
  

   proportional 
  to 
  the 
  rhodium 
  content 
  up 
  to 
  about 
  1.0%. 
  The 
  accuracy 
  of 
  

   the 
  whole 
  procedure, 
  including 
  the 
  synthesis 
  and 
  electrical 
  measurement, 
  

   was 
  =i= 
  4 
  microvolts 
  from 
  strict 
  proportionality. 
  All 
  these 
  alloys 
  are 
  pos- 
  

   itive 
  to 
  platinum. 
  This 
  is 
  not 
  the 
  case 
  with 
  gold-palladium 
  alloys 
  which 
  

   will 
  be 
  tried 
  next 
  to 
  show 
  whether 
  or 
  not 
  the 
  thermoelectric 
  effect 
  of 
  a 
  trace 
  

   of 
  an 
  impurity 
  is 
  the 
  same 
  in 
  sign 
  as 
  the 
  two 
  metals 
  of 
  the 
  alloy. 
  Is 
  99.999% 
  

   Pd-.001% 
  Au 
  positive 
  to 
  palladium? 
  

  

  W. 
  F. 
  Meggers 
  : 
  Spectrographic 
  tests 
  for 
  the 
  purity 
  of 
  some 
  metals 
  (illus- 
  

   trated) 
  . 
  

  

  (Author's 
  abstract.) 
  In 
  connection 
  with 
  the 
  purification 
  of 
  certain 
  

   elements 
  in 
  the 
  platinum 
  group 
  carried 
  out 
  by 
  Dr. 
  Wichers 
  at 
  the 
  Bureau 
  of 
  

   Standards, 
  spectrographic 
  analysis 
  was 
  employed 
  to 
  indicate 
  the 
  progress 
  of 
  

   purification 
  and 
  to 
  detect 
  the 
  traces 
  of 
  impurities 
  present 
  in 
  the 
  final 
  product. 
  

   The 
  metals 
  were 
  vaporized 
  and 
  ionized 
  in 
  a 
  high 
  potential 
  electrical 
  spark 
  

   with 
  capacity 
  and 
  self 
  -inductance 
  in 
  the 
  circuit. 
  Alloys 
  or 
  mixtures 
  of 
  ele- 
  

   ments 
  made 
  luminous 
  in 
  this 
  source 
  give, 
  simultaneously, 
  the 
  spectra 
  of 
  all 
  

   the 
  elements 
  present 
  and 
  these 
  spectra 
  are 
  recorded 
  photographically 
  with 
  

   the 
  aid 
  of 
  a 
  quartz 
  or 
  a 
  concave 
  grating 
  spectrograph. 
  If 
  one 
  observes 
  the 
  

   spectra 
  of 
  a 
  series 
  of 
  mixtures 
  in 
  which 
  one 
  element 
  is 
  progressively 
  diluted 
  

   it 
  is 
  seen 
  that 
  the 
  spectrum 
  of 
  this 
  element 
  becomes 
  simplified, 
  more 
  and 
  more 
  

   lines 
  disappear, 
  as 
  the 
  dilution 
  increases, 
  until 
  a 
  single 
  line 
  remains 
  faintly 
  

   visible 
  when 
  a 
  mere 
  trace 
  of 
  the 
  element 
  is 
  present. 
  These 
  most 
  sensitive 
  

   lines 
  were 
  called 
  "raies 
  ultimes" 
  by 
  de 
  Gramont^ 
  to 
  whom 
  much 
  credit 
  is 
  

   due 
  for 
  developing 
  the 
  principles 
  of 
  this 
  method 
  of 
  analysis. 
  

  

  The 
  raies 
  ultimes 
  are 
  exceptionally 
  sensitive 
  for 
  all 
  metals 
  and 
  remain 
  visi- 
  

   ble 
  when 
  the 
  concentration 
  of 
  an 
  element 
  in 
  an 
  alloy 
  or 
  mixture 
  is 
  even 
  less 
  

   than 
  0.0001 
  per 
  cent. 
  A 
  trained 
  observer 
  after 
  studying 
  the 
  partial 
  spectra 
  

   of 
  carefully 
  prepared 
  standard 
  samples 
  can 
  apply 
  this 
  information 
  to 
  making 
  

   rapid 
  and 
  fairly 
  accurate 
  quantitative 
  analyses 
  of 
  similar 
  alloys 
  of 
  unknown 
  

   percentage 
  composition. 
  The 
  empirical 
  basis 
  for 
  such 
  spectrographic 
  tests 
  has 
  

   been 
  developed 
  at 
  the 
  Bureau 
  of 
  Standards, 
  especially 
  for 
  the 
  analysis 
  of 
  

   mint 
  gold 
  and 
  of 
  platinum 
  metals. 
  Some 
  of 
  the 
  physical 
  constants 
  of 
  metals, 
  

  

  3 
  DE 
  Gramont. 
  Ann. 
  Chim. 
  et 
  Phys. 
  VIII, 
  17: 
  437. 
  1909. 
  

  

  