1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

2 downloads 0 Views 24MB Size Report
Joshua D. Landis, Mukul Sharma*, Devon Renock, and Danielle Niu ...... Ivanovich M, Harman RS (1992) Uranium-‐Series Disequilibrium: Applications to Earth, ...



Rapid  desorption  of  radium  isotopes  from  black  shale  during  hydraulic  fracturing.  1.  


Source  phases  that  control  the  release  of  Ra  from  Marcellus  shale.    




Joshua  D.  Landis,  Mukul  Sharma*,  Devon  Renock,  and  Danielle  Niu    


Department  of  Earth  Sciences,  Dartmouth  College,  Hanover  NH  03755  




*correspondence  to  [email protected]  



9   10   11   12   13   14   15   16   17   18   19   20   21   22   23   24   25   26   27   28   29   30   31   32   33   34   35   36  

Hydraulic   fracturing   of   the   Marcellus   Shale   produces   wastewaters   that   are   hypersaline   and   highly   enriched   in   isotopes   of   radium.   Radium   is   understood   to   derive   from   the   Marcellus   Shale   itself,   but   its   source   phases   and   their   contributions   to   wastewater   production   have   not   been   described.     Using  sequential  extractions  and   experimental   leachates,   we   characterize   two   distinct   end-­‐members   that   could   contribute  Ra  to  wastewaters,  (1)  a  labile   228Ra  mineral  phase  with   226Ra/228Ra  atom   ratios   ~250,   and   (2)   an   exchangeable   226Ra   organic   phase   with   226Ra/228Ra   ~10,000.     In  leaching  experiments  we  observed  rapid  extraction  of  Ra  from  these  phases,  with   high   ionic   strength   solutions   leaching   up   to   14%   of   Ra   from   the   shale   in   just   hours.     Radium   concentrations   and   226Ra/228Ra   ratios   increase   with   [Ca2+]   of   the   leaching   solution,   and   solutions   approaching   1M   Ca2+   produce   226Ra/228Ra   ratios   compatible   with  Marcellus  wastewaters.    In  contrast,  pure  water  removes  450   C.     Additional   descriptions   of   these   rocks   are   given   in   Renock  et  al.  (2016)  and  Niu  et  al.  (2016).         Hand   samples   were   split,   selected   for   clean   faces   and   powdered   in   a   ring   mill   to   a   median   particle   size   of   2-­‐3   µm   with   measured   BET   surface   area   of   20-­‐30   m2   g-­‐1   (Renock   et   al.   2016).   We   note   that   the   rock   powder  particles  are  considerably  larger  than  typical  shale   pores,  which   have  median   sizes   of   a   few   nanometers.     We   anticipate   that   the   exchange   properties   and   presence   of   pore   brine   are   maintained  through  crushing  (see  Balashov  et  al.  2015).     3.2.  Sequential  extractions  under  oxidizing  conditions   To  provide  insight  into  sources  of  Ra  within  the  Marcellus  shale  we  performed  sequential  extractions   that  target  specific,  operationally-­‐defined  phases  of  the  shale  (Tessier  et  al.  1979).    Major  phases  of  the   shale  including  calcite,  organic  matter  and  clay  minerals  may  each  be  expected  to  host  U  or  Th  and   their  decay  daughters  including   226Ra  and   228Ra.    Operationally,  these  phases  would  be  dissolved  by   acetic   acid,   acidic   peroxide   (H2O2)   and   hydrofluoric   (HF)   acid,   respectively   (see   also   Galindo   et   al.   2007,   Stewart   et   al.   2015,   Phan   et   al.   2015).     Major   elements   and   U,   Th   were   measured   for   all   fractions.     Ra   isotopes   were   measured   on   leachate   fractions   for   all   three   shale   samples,   and   on   refractory   phases   for   the   Chenango   Co.   sample   only.   These   extractions   were   performed   under   ambient  (oxic)  conditions  with  5-­‐gram  aliquots  of  shale  powder  and  30  mL  of  leachate.         The   following   extractions   were   adapted   from   Eagle   et   al.   (2003):     (f1)   deionized   water   to   dissolve   soluble   salts   and   labile   components;   (f2)   1M   CaCl2   to   remove   exchangeable   cations;   we   chose   Ca2+   as   an   exchanger   to   best   replicate   exchange   conditions   in   wastewaters;   (f3)   4N   acetic   acid   to   dissolve   carbonates;   (f4a)   repeated   extractions   using   30%   H2O2   at   pH   ~1.5   with   HNO3   to   target   oxidizable   organic   matter,   continued   until   the   shale   residual   was   a   uniform   tan-­‐gray   color,   visually   indicating   loss   of   organic   matter;   pyrite   will   also   be   dissolved;   (f4b)   1M   NH4-­‐acetate   rinse   to   desorb   species   released  during  oxidation  but  resorbed  to  residual  solids;  (f5)  0.2M  hydroxylamine  hydrochloride  to   reduce  oxides;  (f6a)  repeated  additions  of  HF  in  increasing  strengths  to  destroy  silicates  (primarily   clays  and  quartz)  while  minimizing  precipitation  of  alkaline  earth  fluorides;  (f6b)  saturated  AlCl3  to   dissolve   insoluble   fluorides   formed   during   HF   digestion;   (f7)   0.5   N   Na2CO3   carbonate   replacement   to   dissolve   residual   barite   (Curie   and   Debierne   1904),   either   a   primary   component   of   the   shale   or   produced   secondarily   during   preceding   chemical   treatments;   (f8)   strong   acid   digestion   to   attack   refractory   minerals.     Note   that   for   step   f4a   we   used   acidified   H2O2   as   oxidant   as   opposed   to   commonly   used   bleach   (NaOCl)   to   prevent   trace   metal   precipitation   as   oxides   at   high   pH   of   hypochlorite,   as   a   compromise   between   selective   decomposition   of   target   phases   and   successful   extraction  of  target  elements.  Scanning  electron  microscope  characterization  of  residue  (90%.  For   these   samples   we   added   1M   Na2SO4   to   provide   SO42-­‐   equimolar   to   Ba,   with   Ra   yields   of   >95%   as   measured   by   γ-­‐spectrometry   and   gravimetrically   by   weighing   the   precipitate   and   assuming   BaSO4   stoichiometry.     Prior   to   further   processing   the   precipitates   were   rinsed   repeatedly   in   deionized   water   to   a   pH   of   ~5   that   signaled   the   removal   of   excess   H2SO4   from   the   precipitates.     The   wastewaters  are  enriched  in  Ba  that  is  sufficient  to  produce  several  mg  of  Ba(Ra)SO4  precipitate.         Sulfate  precipitates  were  dissolved  by  a  carbonate  replacement  technique  developed  by  Marie  Curie   (Curie  and  Debierne  1904,  Cohen  and  O’Nions  1991):    approximately  10  mg  of  Sr(Ra)SO4  or  Ba(Ra)SO4   were  transferred  to  Teflon  vials  in  10  mL  0.5N  Na2CO3  and  were  oven-­‐heated  at  105°C  for  about  12   hours.    This  permits  a  complete  replacement  of  the  sulfate  precipitate  by  acid-­‐soluble  carbonate   precipitate.  After  cooling  the  samples  were  centrifuged  and  rinsed  in  deionized  water  until  rinses   showed  a  pH  of  ~  5.    The  resulting  Ba  (or  Sr)  carbonates  were  dissolved  in  2  mL  of  1N  HCl.    These   solutions  were  then  evaporated,  re-­‐dissolved  in  3N  HNO3,  and  passed  through  columns  packed  with   Eichrom  Sr-­‐spec  resin  to  separate  Ra  from  Sr  or  Ba  (Chabaux  et  al.  1994;  Fig.  3b).    Samples  were   passed  twice  on  2  mL  columns,  then  twice  on  0.1  mL  micro-­‐columns.  In  the  final  pass  the  samples   were  eluted  in  0.03N  HNO3  to  minimize  elution  of  hydrolyzed  resin  organic  material.    To  further   reduce  organic  interference,  the  final  Ra  fraction  was  combined  with  aqua-­‐regia  and  evaporated,   then  re-­‐dissolved  in  3N  HNO3  and  irradiated  with  ultra-­‐violet  light.    Samples  were  finally  evaporated   to  dryness  and  loaded  onto  filaments  for  thermal  ionization  mass  spectrometry.     3.4.5.  Thermal  ionization  mass  spectrometry  (TIMS)  




  332   333   334   335   336   337   338   339   340   341   342   343   344   345   346   347   348   349   350   351   352   353   354   355  

Radium   isotope   analysis   was   performed   on   the   Dartmouth   Triton   thermal   ionization   mass   spectrometer.     Purified   Ra   fractions   were   loaded   on   tungsten   (W)   filaments   in   2.5N   HCl   between   a   ‘sandwich’  load  of  tantalum  chloride  solution.  Ra+  ions  were  measured  using  a  Mascom  secondary  ion   multiplier  (SEM)  operated  in  ion  counting  mode.  To  preserve  our  detector  fromaccruing  dark  noise   due   to   radiation   damage,   we   adopted   a   static   analytical   protocol   (see   Yokoyama   and   Nakamura   2004).     Samples   were   sequentially   baked   at   1200°C   and   1240°C   to   remove   final   traces   of   organics.   The   filament   temperature   was   then   raised   to   1320°C   for   analysis.     We   measured   30-­‐100   cycles   as   necessary   to   achieve  an   internal   precision   of   ~3%.   Effective   ionization   efficiencies   were   in   the  range   of  1-­‐10%.    SEM  dark  noise  was  monitored  throughout  our  analysis  period  and  was  maintained  at  less   than  0.05  counts  per  second.    Ion  counts  were  corrected  for  the  dark  noise  contribution.       The  final  226Ra/228Ra  atom  ratio  of  our  228Ra  tracer  measured  using  TIMS  is  6.60±0.04  (n  =11),  within   error  of  certified  OKA2  ore  value  of  6.38±0.14.    Standards  with  different  sized  Ra  loads  and  running   temperatures   do   not   show   fractionation   beyond   analytical   uncertainty.     Final   concentration   of   the   228Ra   tracer   solution   was   determined   by   gamma   spectrometry  (Landis   et   al.   2012)   with   an   estimated   expanded  uncertainty  (2σ)  of  3.4%.    We  used  this  primary   228Ra  spike  solution  to  determine  sample   Ra  concentration  by  isotope  dilution  using  TIMS.    We  also  created  a  mixed  spike  from  our   226Ra  and   228Ra   primary   tracer   solutions   to   provide   cross-­‐calibration   of   the   two   solutions.     Based   on   γ-­‐ calibration   of   tracer   solutions   the   expected   226Ra/228Ra   ratio   of   the   mixed   spike   was   1065±72   (mean±2σ).     TIMS   analyses   yield   a   226Ra/228Ra   ratio   of   1043±14   (n=9).     All   tracer   228Ra   concentrations   and   226Ra/228Ra   ratios   are   decay-­‐corrected   to   a   common   reference   time   of   1/1/2015.     Total   procedural   Ra   blanks   were   too   low   to   be   distinguished   from   0.1   fg   loads   of   our   228Ra   spike   during  measurement  by  isotope  dilution,  and  are  less  than  0.01  fg  at  99%  confidence  interval.  


4.  Results  and  Discussion  


Quantitative   XRD   reveals   that   the   black   shale   samples   contain   clay   minerals   (illite   and  


chlorite),   quartz,   calcite,   organic   matter,   and   pyrite   ±   barite   (Table   1).   Evaluation   by  


scanning   electron   microscopy   reveals   that   calcite   is   present   as   intergranular   cement,   and  


organic   matter   and   clay   minerals   form   intercalated   organo-­‐clay   composites   at   scales   that  


vary   from   sub-­‐mm   to   nanometer   (Figure   4).     Ra   concentrations   (=   [Ra])   and   226Ra/228Ra  


ratios  of  three  black  shale  samples  are  given  in  Table  1,  as  are  activities  of  the  Ra  isotopes  


and  their  U  and  Th  parents.    All  rocks  appear  to  be  in   226Ra-­‐238U  and   228Ra-­‐232Th  radioactive  


equilibrium.    Due  to  the  hosting  of  U  in  organic  matter,  bulk  rock  [Ra]  and  226Ra/228Ra  ratios  


increase  with  total-­‐organic-­‐carbon  (TOC;  Figure  2a)  -­‐-­‐  this  drives  a  strong  Ra  isotope  mixing  


relationship  between  organic  and  mineral  solids  of  Marcellus  Shale  (Figure  2b).  




4.1  Sequential  extractions  under  oxidizing  conditions,  atmospheric  pressure  


Table   2   gives   the   sequential   extraction   yields   of   U,   Th,   Na,   Ca   as   well   as   the   ionic   strength   of  


the   solutions   produced   under   oxidizing   conditions   at   room   temperature   and   atmospheric  


pressure.     Table   3   gives   the   sequential   extraction   yields   of   Ra,   Ba   and   S   and   their  


corresponding   226Ra/228Ra  ratios.    Prior  to  describing  these  results,  we  note  that  elements  


can   and   do   fractionate   from   each   other   during   leaching   procedures,   which   are  






“operationally  defined”  according  to  both  the  general  properties  of  the  extracting  reagents  


and  the  nature  of  the  target  phase  (Tessier  et  al.  1975,  Bacon  and  Davidson  2008).    Different  


elements   in   decay   couples   such   as   U-­‐226Ra   or   Th-­‐228Ra   may   also   be   differentially   susceptible  


to   extraction   based   on   elemental   behavior   rather   than   properties   of   their   source.     In   fact,  


decay   couples   provide   the   best   means   for   observing   elemental   effects   in   extractions   since  


these  elements  are  known  to  originate  from  the  same  source;  when  released  in  equilibrium,  


as  evidenced  by  the  same  activity  and  mass  percent  of  bulk  rock,  no  elemental  fractionation  


is   implicated.     But   where   disequilibrium   is   apparent,   elemental   chemistry   has   dictated  


release   over   mineralogy,   due   to   either   selective   extraction   of   one   radionuclide,   or  


precipitation   or   resorption   of   another   (e.g.,   Lucey   et   al.   2007).     Repartitioning   of  


radionuclides  to  solids  can  be  expected  to  be  favored  in  the  order  Th  >>  Ra  >  U  (Ivanovich  


and  Harman  1992).  




4.2.1.  Reservoirs  of  238U  and  226Ra  in  Marcellus  Shale    


Among   shale   samples   the   amount   of   U   removed   in   the   oxidizable   (f4)   fraction   increases  


approximately   exponentially   with   TOC.     This   is   in   agreement   with   the   control   of   organic  


matter   abundance   on   shale   U   content   (Leventhal   1981;   Figure   2a).     However,   consistent  


with  the  findings  of  Phan  et  al.  (2015),  the  fraction  of  total  U  removed  in  the  oxidizable  (f4)  


fraction   constitutes   a   minority   (

Suggest Documents