13
Methods in Water Science and Technology Scientific research Cross validation of two different COD test kits (Kit with Hg and kit without Hg) Written by: Eric Clayderman CAZOLI December 2014 University of Stavanger Department of Mathematics and Natural Sciences 4036 Stavanger

Module 3 in Methodology course

Embed Size (px)

Citation preview

Page 1: Module 3 in Methodology course

                   

Methods  in  Water  Science  and  Technology      

Scientific  research    

Cross  validation  of  two  different  COD  test  kits    (Kit  with  Hg  and  kit  without  Hg)  

   

 

 

 

 

Written by: Eric Clayderman CAZOLI

December 2014

University  of  Stavanger  Department  of  Mathematics  and  Natural  Sciences  

4036  Stavanger  

Page 2: Module 3 in Methodology course

Abstract:    COD  measurements  on  COD  standard  solutions  and  wastewater  samples  

were  done  by  using  two  different  COD  test  kits  (kit  with  Hg  and  kit  without  Hg).  Data  

obtained  from  both  kits  were  analyzed  by  looking  essentially  at  statistical  parameter  

mean   X   and   standard   deviation   σ.   The   objective   of   this  work  was   to   compare   the  

analyzed  data   in  order   to  check  the  cross  validation  of   the  two  kits.  Results  of   this  

work   showed   that   the   two   kits   are   the   same   when   measuring   COD   standard  

solutions,  while  they  do  not  really  overlap  when  measuring  wastewater  samples.    

Key  words:  COD,  wastewater,  test  kits,  mean,  cross  validation  

 1.  Introduction:    

Chemical   Oxygen   Demand   (COD)   is   a   term   used   in   both   water   and  

wastewater  treatment  to  measure  the  amount  of  a  specified  oxidant  reacting  with  a  

given  sample  under  controlled  conditions   (Al-­‐Momani,  2003).  The  diochromate   ion  

(Cr2O72-­‐)   is   the   specified   oxidant   in   colorimetric   method   and   its   amount   is  

expressed  in  terms  of  its  oxygen  equivalence.  

Under  the  presence  of  catalysts  (sulphuric  acid  H2SO4,  mercuric  sulphate  AgSO4  and  

sulfamic   acid   H3NSO3),   the   dichromate   (Cr2O72-­‐)   oxidizes   organic   material   in   a  

sample  after  incubation  of  2h  at  150°C.  This  oxidation  reduces  Cr2O72-­‐  (hexavalent)  

into  Cr3+  (trivalent).  Each  of  these  chromium  species  has  a  direct  relationship  with  

oxygen  consumed  (Association,  Association,  Federation,  &  Federation,  1915).    

Colorimetric  method  for  COD  analysis  is  a  time  consuming  since  you  have  to  

prepare  both  digestion  and  catalyst  solutions.  Nowadays,  COD  test  kits  are  available  

for  an  easy  and  quick  COD  analysis.  A  COD  test  kit  is  like  a  small  glass  tube  (vial)  on  

which   there   is   a   unique   barcode   label   that   is   automatically   read   by   a  

spectrophotometer   to   identify   the   appropriate   method   and   take   the   COD  

measurement.  COD  test  kits  also  contain  both  digestion  and  catalyst  solutions  (like  in  

the   colorimetric  method)   that   react  with   samples   to   be  measured.   A   COD   test   kit  

may   present   a   risk   for   the   environment  when   it   contains   harmful   chemicals.   A   kit  

Page 3: Module 3 in Methodology course

containing  Hg  is  harmful  for  the  environment  compared  to  the  one  that  has  Hg  free.  

Regarding  the  environmental  aspect  and  the  price  of  kits,  it  would  be  recommended  

to  use   the  one  with  Hg   free  and   the  one   that   is   less   costly.     For   this   reason,   I   did  

some  COD  analysis  on  COD  standard  solutions  and  wastewater  samples  by  using  two  

types  of  COD  test  kits  (kit  with  Hg  and  kit  without  Hg).  For  this  COD  analysis,  three  

known   COD   concentrations   of   standard   solution   were   prepared   from   Potassium  

hydrogen  phthalate  (C8H5KO4).    

The  objective  of  this  work  was  to  check  the  cross  validation  of  the  two  kits  by  

analyzing   the   statistical   parameter  mean  and   standard  deviation  of  obtained  data.  

Data  obtained  from  this  work  and  the  results  of  statistical  analysis  are  presented  and  

explained  in  the  result  section  of  this  paper.    

 

2.Theory  

A  COD  test  kit  contains  all  necessary  chemicals  that  digest  organic  matters  in  

a   given   sample   and   catalyse   reactions   that   happen   inside.   The   following  

stoichiometry  shows  these  reactions  and  the  relationship  as  well  as  the  theoretical  

ratios  between  chromium  species  and  O2.    

Oxidation:                  C6H12O6  +  6  H2O                  ===    >          6  CO2    +  24  e-­‐    +  24  H+  

Reduction:              24  e-­‐    +  24H+  +  32  H+  +  4  Cr2O72-­‐    ===    >          8  Cr3+    +  28  H2O  

   Redox  reaction:  C6H12O6    +  +  32  H+  +  4  Cr2O72-­‐                      ===    >                        8  Cr3+    +  6  CO2    +    22  H2O  

     In  reality,  O2  is  the  electron  acceptor:    C6H12O6  +  6  H2O    ===    >        6  CO2    +  24  e-­‐    +  24  H+                                      24  e-­‐    +  24  H+  +  6  O2      ===    >  12  H2O  

From  the  reactions  above,  we  can  see  that  1  mole  of    O2    takes  up  4  e-­‐      and  1  mole  of  

Cr2O72-­‐    takes  up  6  e-­‐  .  

=    >    4  e-­‐      /6  e-­‐    *  mole  Cr2O72-­‐/  mole  O2  =  4  mole  Cr2O72-­‐/  6  mole  O2  =    >  Δ  Cr2O72-­‐    =  

3/2*ΔO2  =  COD  

Page 4: Module 3 in Methodology course

=    >    8  Cr3+  /4  Cr2O72-­‐  *  4  Cr2O72-­‐/6  mole  O2    =    >    4  mole  Cr3+/3  mole  O2  =    >  Δ  Cr3+=  

¾*  ΔO2  =  COD  

Practically,   we   measure   Cr3+   for   the   high   range   COD   (100   and   900   mg/L).   The  

relationship   between   theoretical   COD   and   ΔCr3+   is   obtained   by   a   standard   curve  

calibration.    

 

Reagents:   According   to   (Association   et   al.,   (1915),   different   reagents   are   needed  

during  the  set  up  of  analysis  in  order  to  have  a  complete  oxidation  reaction  and  also  

to   remove   any   possible   interferences.     Specifically,   these   reagents   are   mercuric  

sulphate,   sulfuric   acid   and   sulfamic   acid.     Mercuric   sulphate   is   added   to   remove  

complex   chloride   ions   present   in   the   sample.    Without   the  mercuric   sulphate,   the  

chloride   ions   would   form   chlorine   compounds   in   strong   acid   media   used   in   the  

procedure.     These   chlorine   compounds   would   oxidize   the   organic   matter   in   the  

sample,  resulting  in  a  COD  value  lower  than  the  actual  value.  Sulfamic  acid  is  added  

to   remove   interferences  caused  by  nitrite   ions.    Without   sulfamic  acid,   the  COD  of  

the   sample  would  measure   higher   than   the   actual   value.  Potassium   dichromate   is  

used  as   the  oxygen   source  with   concentrated   sulfuric   acid  added   to   yield  a   strong  

acid  medium.      

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Page 5: Module 3 in Methodology course

 

 

 

3.  MATERIALS  AND  METHODS  

3.1.  Materials  

Different  materials  (devices,  chemical  and  instruments)  were  used  during  the  

work  in  a  laboratory.  These  materials  are  listed  in  the  following  table  (tab.  1).  

Table  1:  list  of  materials  that  were  used  during  the  experiments,  and  their  functions  

 Material   Function  

Potassium  hydrogen  phthalate  (C8H5KO4)   Chemical   used   to   prepare   COD   standard  solutions  

Distilled  water   Used  for  dilution    

Analytical  balance     Used   to   weight   the   amount   of   C8H5KO4   to   be  used  

2  types  of  COD  test  kits:  kit  with  Hg  and  Kit  without  Hg  

 

Contain  chemicals  needed  to  react  with  samples  

Pipettes   Used  to  take  a  precise  volume  of  sample    

Flask  Erlenmeyer   Used  for  mixing  chemicals  and  solutions  

Gloves   Hands  protection  

Glass   Yes  protection  

Incubator   Used  to  cook  COD  kits  containing  samples  

Spectrophotometer   Measures   COD   concentration   of   sample   as   a  function  of  the  color  intensity.  

Stop  watch   Record  time    

 

3.2.  Methods  

First  of  all,  I  prepared  three  known  concentrations  of  COD  standard  solution  

from  Potassium  hydrogen   phthalate   (KHP=   C8H5KO4).     This   preparation  was   done  

according  the  American  standard  for  COD  analysis  (Association  et  al.,  1915).  KHP  has  

a   COD   of   1.176   gO2/g   KHP.   This   value   is   obtained   by   the   following   reaction   and  

calculations:  

Page 6: Module 3 in Methodology course

 C8H5KO4        +  aO2  +  H+    =    >    8  CO2  +  K+      +  3  H20          ===    >        a=  15/2  *  mole  

O2/mole  KHP  

=    >        7.5  mole  O2/mole  KHP  =    >        γO2/KHP  =  7.5  mole  O2/mole  KHP  *  (32gO2/moleO2)  /  (204  gKHP/mole  

KHP)  =    >      1.176  gO2/gKHP          =  1.176  gCOD/gKHP  

 (1  gCOD/L)/  (1.176  gCOD/gKHP)  =  0,85  gKHP/L.  

   

Preparation  of  the  three  known  COD  concentrations:    

C1=  stock=  1500  mg/L      (add  1275mg  KHP  into  1000mL  distilled  water)  

C2=  525  mg/L                                      (add  35  mL  stock  into  100  mL  volumetric  flask  

containing  distilled  water)  

C3=  150  mg/L                                      (add  10  mL  stock  into  100  mL  volumetric  flask  

containing  distilled  water)  

 

The  second  part  of  this  work  was  to  measure  the  COD  of  these  three  known  

COD  concentrations  and  measure  the  COD  of  wastewater  samples.    The  procedures  

for  COD  measurement  are  explained  in  the  appendix  i.    

The  third  part  of   this  work  was   to  analyse   the  data  obtained   from  the  COD  

measurement.   Mean   and   standard   deviation   were   the   two   main   statistical  

parameters  analysed  to  compare  the  data  obtained  from  both  Kits.  This  analysis  was  

done  under  Excel  software.    

 

 

 

 

 

     

Page 7: Module 3 in Methodology course

         4.  Results  and  discussion  

4.1.  Raw  data  

The  raw  data  obtained  from  all  measurement  during  this  work  is  presented  in  table  

2.  

For  simplification,  let´s  use  the  terms  “Hg”  and  “Hg  free”  respectively  for  kit  with  Hg  

and  kit  without  hg.  

Table  2:  Raw  COD  data  obtained  from  the  two  kits.  

 Expected  concentrations  of  the  three  known  COD  standard  solutions  (mg  COD/L)  

C1   C2   C3  1500   525   150  

   

COD  concentrations  of  blanks  (CODmg/L)       Blank  1   Blank  2  

         

Hg   83   65        

   Hg  free   49   0  

         

   COD  concentrations  of  COD  standard  solution  after  the  measurement  (mgCOD/L)  

    C1   C1   C1      

   Hg   1556   1569   1551  

       

Hg  free   1528   1561   1515      

           

    C2   C2   C2   C2   C2   C2  Hg   569   571   573   571   577   570  Hg  free   530   533   537   531   534   530      

             

    C3   C3   C3      

   Hg   205   202   200  

       

Hg  free   153   158   157      

       

             

   COD  concentrations  of  wastewater  samples  (mgCOD/L)  

    sample  1   sample  2   sample  3   sample  4   sample  5   sample  6  Hg   1304   409   388   400   354   392  Hg  free   441   422   637   465   354      

   

Page 8: Module 3 in Methodology course

 

 

 

4.2.  Analysed  data  

4.2.1.  Mean  (X)  and  standard  deviation  (σx)  

During   the   statistical   analysis,   parameter   mean   X   and   standard   deviation   σ   were  

calculated.  Table  3  and  4  show  respectively  the  mean  and  the  standard  deviation  of  

COD  values  obtained  from  the  two  kits.  The  average  COD  concentrations  of  standard  solutions  and  wastewater  samples  were  subtracted  by  COD  content  of  blanks.    

Table  3:  Average  COD  concentrations  of  standard  solutions  (C1,  C2,  C3)  and  wastewater  samples  

    COD  concentrations  (  mgCOD/L)  

  Blank   C1   C2   C3   Wastewater  Hg   74   1485   498   128   467  

Hg  free   26   1510   508   131   439    Table  4:  Standard  deviation  of  COD  values  (σx)    

  C1   C2   C3   Wastewater  Hg     9   3   3   374  

Hg  free   24   3   3   105      

4.2.2.  Rejection  of  data  

By  looking  at  the  table  2,  we  can  see  that  some  of  the  COD  concentrations  of  

wastewater  samples  look  specious.  The  value  1303  mgCOD/L  (from  kit  with  Hg)  and  

the  value  637  mgCOD/L  (from  kit  without  Hg)  seem  anomalously  large.  By  applying  

the  Chauvenet´s  criterion,  we  can  decide  the  rejection  of  these  two  values.  

Assuming   provisionally   all   COD   measurement   of   wastewater   samples   is  

legitimate.    

a.   N=6   (1303,   409,   388,   400,   354,   392);   the  mean  X   is   here   467   and   the   standard  

deviation  σx    is  374.    The  difference  between  the  suspect  Xsus=1303  and  the  mean  X=  

467  is  836,  or  2.2  standard  deviations;  that  is,  

Tsus=  (xsus-­‐x)/  σx    =  (1303-­‐467)/374    =  2.2  

Page 9: Module 3 in Methodology course

Referring  to  the  table  in  Appendix  ii,  the  probability  that  a  measurement  will  differ  

from  X  by  2.2σx  or  more  is:  

Prob(outside  2.2σx)      =  1-­‐  Prob(inside  2.2σx)  

                     =  1-­‐  0.972                        =  0.028  

In  6  measurements,  I  would  expect  to  find  0.168  of  one  measurement  as  deviant  as  

the   suspect   result.   Since   0.168   is   less   than   the   0.5   set   by   Chauvenet´s   criterion,   I  

should   reject   the   suspect   Xsus=   1303   mgCOD/L.   So,   the   new   mean   and   standard  

deviation   for   COD   of   wastewater,   which   was   measured   by   kit   with   Hg,   would   be  

respectively  315  mgCOD/L  and  21.    

 

b.  N=5  (441,  422,  637,  456,  354);  the  mean  X  here  is  439  and  standard  deviation  σx  is  

105.  The  difference  between  the  suspect  Xsus=637  and  the  mean  X=  439  is  198,  or  

1.88  standard  deviations;  that  is,  

Tsus=  (xsus-­‐x)/  σx    =  (637-­‐439)/105    =  1.88  

Referring  to  the  table  in  Appendix  ii,  the  probability  that  a  measurement  will  differ  

from  X  by  1.88σx  or  more  is  

Prob(outside  1.88σx)      =  1-­‐  Prob(inside  1.88σx)  

                         =  1-­‐  0.939                            =  0.061  

In  5  measurements,  I  would  expect  to  find  0.305  of  one  measurement  as  deviant  as  

the   suspect   result.   Since   0.305   is   less   than   the   0.5   set   by   Chauvenet´s   criterion,   I  

should   reject   the   suspect   Xsus=   637   mgCOD/L.   So,   the   new   mean   and   standard  

deviation  for  COD  of  wastewater,  which  was  measured  by  kit  without  Hg,  would  be  

respectively  421  mgCOD/L  and  48.    

4.2.3.  Correct  estimator  for  all  measurements  

If  we  assume  95  %  confidence   for  our  measurements,   the  average  COD   for  

our  samples  would  be:    

COD=  X  ±  (σx/√n)*tα/2,n-­‐1  

where:  

 x:  mean  σx  :  standard  deviation    

 n:  number  of  measurements    α:  accepted  error=  5%=  0.05  

Page 10: Module 3 in Methodology course

 

For  95%  confidence,  the  real  mean  of  COD  (mg/L)  for  the  standard  solutions  and  the  

wastewater  sample  are  the  following:  

 

 

Conclusion:  By   looking  at  the  real  mean  of  COD  for  standard  solutions,  we  can  see  

that  the  two  test  kits  overlap.  Therefore,  we  can  say  that  the  two  tests  are  the  same.    

For   the   wastewater   samples,   we   can   see   that   the   two   kits   almost   overlap.   The  

reason,  why  they  do  not  exactly  overlap  for  wastewater  samples,  could  be  from  the  

subtraction  of  the  COD  concentrations  of  samples  by  the  mean  of  the  COD  of  blanks  

that  seem  incorrect.    

 

 

 

 

 

 

 

 

 

Reference:  

 Al-­‐Momani,  F.  (2003).  Combination  of  photo-­‐oxidation  processes  with  biological  

treatment:  Universitat  de  Barcelona.  Association,  A.  P.  H.,  Association,  A.  W.  W.,  Federation,  W.  P.  C.,  &  Federation,  W.  

E.  (1915).  Standard  methods  for  the  examination  of  water  and  wastewater  (Vol.  2):  American  Public  Health  Association.  

Taylor,  J.  (1997).  Introduction  to  error  analysis,  the  study  of  uncertainties  in  physical  measurements  (Vol.  1).  

   

  C1   C2   C3   wastewater  Hg   1485±  22   498±  3   128±  7   315±  26  

Hg  free   1510±  60   508±  3   131±  7   421±  77  

Page 11: Module 3 in Methodology course

 Appendix  i    a.  Procedures  for  COD  analysis  using  Kit  with  Hg  

 

Page 12: Module 3 in Methodology course

Source:  http://www.google.no/url?sa=t&rct=j&q=&esrc=s&source=web&cd=2&ved=0CB8QFjAB&url=http%3A%2F%2Fenvirotest.com.my%2Fproductsupport%2Fdownload%2F57&ei=LGKuVLKWLYLyaNyegig&usg=AFQjCNESMTZBXvJApC7q3eBZfUkcSuhTvg&sig2=88DGxtDdmga_9a9tpWl1EA        

 b.  Procedures  for  COD  analysis  using  Kit  without  Hg  (Hg  free)      

 

Page 13: Module 3 in Methodology course

Source:  http://www.google.no/url?sa=t&rct=j&q=&esrc=s&source=web&cd=2&ved=0CB8QFjAB&url=http%3A%2F%2Fenvirotest.com.my%2Fproductsupport%2Fdownload%2F57&ei=LGKuVLKWLYLyaNyegig&usg=AFQjCNESMTZBXvJApC7q3eBZfUkcSuhTvg&sig2=88DGxtDdmga_9a9tpWl1EA        

     

   Source:  (Taylor,  1997)  

Appendix  ii