PB86-216751
    Oxygen-18 Study of S02 Oxidation in
    Rainwater by Peroxides
    Argonne National Lab., IL
    Prepared for

    Environmental Protection Agency
    Research Triangle Park, NC
    Jul 86
U.S. Department of
      Technical I

-------
                                   TECHNICAL REPORT DATA
                            (Please read lauruciions on the reverse before completing)
1. REPORT NO.

 EPA/600/3-86/035
             3. RECIPIENT'S ACCESSI Of* NO.
                       2 1 6 y 5 1
4. TI-..E AND SUBTITLE
     OXYGEN-13 STUDY OF S02 OXIDATION IN
     RAINWATER BY PEROXIDES
             b. REPORT DATE
               July  1936
             6. PERFORMING ORGANIZATION CODE
7. AUTHORIS)

     B.  D.  Holt and R. Kumar
                                                           8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
     Chemical Technology Division
     Argonne National  Laboratory
     Argonne, Illinois   60439
                                                            10. PROGRAM ELEMENT NO.
                 CCVN1A/Q2-3058 (FY-86)
             11. CONTRACT/GRANT NO.
               IAG DW 89930060
  . SPONSORING AGENCY NAME AND ADDRESS
     Atmospheric Sciences Research Laboratory-RTP, NC
     Office of Research and  Development
     U.S.  Environmental Protection Agency
     Research Triangle Park,  North Carolina  27711
             13. TYPE OF REPORT AND PERIOD COVERED

             	Final  (10/82 -  5/8.6)	
             14. SPONSOFIIN3 AGENCY CODE
                 EPA/600/09
15. SUPPLEMENTARY NOTES
 16 ABSTRACT
          A new analytical  method was developed for  the  determination of oxygen.,
     isotope ratios in  peroxides  in rainwater.  In the method,  rainwater samples
     were quantitatively  degassed of dissolved air by a  combined treatment of
     evacuation, ultrasonic agitation, and nelium sparging (VUS), followed
     by a permanganate  oxidation  of the dissolved peroxide to 02.  The 02 was
     then quantitatively  removed  fron the rainwater  by the VUS  treatment and
     converted to C02 for  mass  spectrometric analysis.   Using this method,
     14 rainwater samples  collected at four sites (Argonne,  IL, Research
     Triangle Park, NC, Uhiteface Mountain, NY and Dearborn,  MI) were analyzed
     to determinne the  fraction of the sulfate in the samples that was produced
     as a result of the aqueous-phase reaction of dissolved S02 with hydrogen
     peroxide.  It was  concluded  that 40% or more of the sulfate in the samples
     was formed by peroxide oxidation.
17.
                                KEY WORDS AND DOCUMENT ANALYSIS
                  DESCRIPTORS
                                              b.lDENTIFIERS/OPEN ENDED TERMS  C.  COSATI Field/Group
is. DISTRIBUTION STATEMENT
                                               19. SECURITY CLASS (Tna Report)

                                                     IINP.I A.SSTFIFD
                                                                          21. NO. OF PAGES
                              43
       RELEASE TO PUBLIC
30. SECURITY CLASS (This page)

      illsiri ASSTFTFil
                                                                          22. PRIC
                                                                            t
EPA Form 222C-1 (1-7))

-------
                    DISCLAIMER


The information in this document has been funded by the
United States Environmental Protection Agency under In-
teragency Agreement DW 030060-01-0 to Argonne National
Laboratory.  It has been subject to the Agency's peer and
administrative review, and it has been approved for pub-
lication as an EPA document. Mention of trade names or
commercial products does  not  constitute endorsement or
recommendation for use.

-------
                                    ABSTRACT




    Oxygen isotope ratio measurements were used to examine the importance of peroxide



oxidation of SO3 to sulfate in the atmosphere. A new analytical method was developed for



the determination  of oxygen isotope ratios in peroxides (concentrations in the ppb range)



in rainwater.  According to  the method,  20-L samples were quantitatively  degassed  of



dissolved air by a combined treatment of evacuation, ultrasonic agitation,  and helium



sparging (VUS), followed  by permanganate oxidation  of  the dissolved  peroxide to O2;



the C>2  was then quantitatively removed from the rainwater by  the \TJS treatment and



converted to COj  for  mass spectrometric analysis.  Stock solutions of HjOa of various



18O enrichments were  prepared by a high-voltage discharge method; and, by using these



solutions to oxidize SOa to SO^~, the following .sotopic relationship was established:

                      _  =  0.57 «51802- + 0.435180H0  + 8.4%.
This relationship was used to calculate (he <518Ocr.2- formed  by peroxide oxidation and
                                             ^U4


a similar previously established relationship was used to calculate the 618O<-,.~ 2— formed



by metal-catalyzed aqueous oxidation.  These sets of calculated values were compared to



the measured values for evaluation of the estimated  fraction of sulfate in rainwater that



was formed by peroxide oxidation, assuming that metal (or carbon) catalysis was the only



other major sulfate formation mechanism. It  wns concluded that ~40C7> or more of the



sulfate in summer rains in the northeastern U. S. was formed by peroxide oxidation.
                                        ill

-------
                              CONTENTS
   ABSTRACT    	    iii
   TABLES	    vi
   FIGURES	vii
   ACKNOWLEDGMENT	viii
1.  INTRODUCTION     	       	    1
2.  EXPERIMENTAL      	       	       .    3
       Development of Analytical Method  ...      	    3
       Conversion of O2 to CO2	      •  .       ....      7
       Removal of Dissolved O2 from Water	      9
       Oxidation of H2O2 to O2 in Water      	      .  .     13
       Synthesis of 18O-enriched H2O2	13
       Collection of Rain Samples	   14
3.  RESULTS	   16
       Recovery of Added H2O2   .      	   16
       Test for Isotopic Interference    	   16
       Effects of Autodecomposition of H2O2   	       16
       Effects of KMnO4 Decomposition	   18
       Procedure Blank	      ...    .18
       Isotopic Relationships	     18
       <518O and Concentrations of H2O2 in Rainwater	     22
4.  DISCUSSION   	   26
       6-8O of H2O2 in Rainwater   	   26
       618O of Atmospheric Sulfates: Measured vs. Calculated	   27
       Fraction of Atmospheric Sulfate Formed by Peroxide Oxidation  ....   29
5.  SUMMARY	   32
   REFERENCES   	   33

-------
                                 TABLES

                                                                        Page

1.  Concentration and isotopic results for rainwater collected at Argonne
   National Laboratory (ANL), Dearborn, MI (DEA), Whiteface Moun-
   tain,  NY (WFM), and Research Triangle Park, NC (RTP), May-
   September, 1985	        	23
                                    VI

-------
                                    FIGURES

Number                                                                      Page


   1.  Four-step method for conversion  of oxygen  in H2O2 (ppb range in
      water) to CO2 for mass spectrometric  analysis   ............    4

   2.  Analytical train  ...........................    5

   3.  Reaction chamber .......................      6
   4.  Recovery of O2  as CO2 after reaction with charcoal in  contact with
      platinum at various temperatures   .................    8

   5.  Recovery of CO2 after exposure to charcoal and platinum at various
      temperatures  .........................   10

   6.  Removal of O2 from water by various  degassing techniques ........   12

   7.  Isotopic effects of water solvent on the <518O of CO2 originating from
      H202  .............................   17
   8.  Dependence of £l8OjjoQ  on  518Oj| Q in a high-voltage discharge
      process of formation ........................   19

   9.  Dependence of 6 18OC,^.2_ on <518O|j Q  in oxidation of SO2 to sulfate .  .  .     21

  10.  Measured isotopic data for II2O2, SO^~, and H2O in the samples of   —
      rainwater   .............................   24

  11.  618O of sulfate in rainwater: calculated compared to measured       ....   28

  12.  Percent of sulfate formed by peroxide oxidation,  assuming negligible
      primary sulfates   .........................   30
                                       vn

-------
                              ACKNOWLEDGMENT





    The authors wish to acknowledge that some of the laboratory studies involving the



determination of isotopic relationships using solutions of known 18O/16O ratios were con-



ducted under a related  program for  the U.S. Department of Energy's Office of Energy



Research, Basic Energy Sciences, Division of Materials Science.  Also, the  authors are



grateful to Dr.  Mark Dubois, Whiteface Mountain Field  Station, Atmospheric Sciences



Research Center, Wilmington, NY; to Dr.  Gary Eaton, Research Triangle Institute, Re-



search Triangle Park, NC; and to Dr. William Pierson and Ms. Wanda Brachaczek, Ford



Research Engineering Staff, Dearborn, MI, for collection and shipment of rainwater samples



to Argonne for analysis.
                                       vin

-------
                                    SECTION 1






                                 INTRODUCTION





    Considerable interest has been  shown in the possibility of using oxygen isotopy to



elucidate the role of H2O2 in the oxidation of SO2 to sulfates  in the atmosphere.  The



potential importance of II2O2  lies in the view  that aqueous-phase oxidation of SO2 to



II2SO4- probably accounts for a major fraction of the observed SO^" in the precipitation



occurring in the northeastern United States.  The  key  reactants responsible  for this ox-



idation are not well known,  although a large number  of possible catalysts and  oxidants



exist in the atmosphere, including carbon, transition rnetal ions, hydroxyl and  organic free



radicals, hydrogen peroxide (including organic peroxides) and ozone (Penkett et a/., 1979).



It has been postulated that the atmospheric oxidation of SO2 is  limited by  the availability



of these oxidants and catalysts,  rather than by the availability of SO2 itself.  This is the



so-called phenomenon of non-linearity between the concentration (and, by implication, the



emission) of SO2 in the atmosphere  and the conversion of this  SO2 to H2SO4.  If such is



indeed the case, it is important to determine if peroxides have  a pivotal role in  acid for-



mation and deposition from the atmosphere [oxidation by ozone decreases in significance



with decreasing pH of the rainwater].





    Oxygen isotopic studies  are useful  in distinguishing between the different oxidation



mechanisms effective in the atmosphere. Laboratory simulation of several  different atmo-



spheric reaction sequences has shown that the oxygen  isotope ratio in the product SO^"



is uniquely related to the reaction pathway followed in its formation (Holt et al., 1982).



It may therefore be possible to determine if the atmospheric hydrogen peroxide is respon-



sible for significant oxidation of SO2 to sulfate,  and if it is this oxidant  that limits the



aqueous-phase formation of sulfuric  acid. The results of these studies could  have signifi-



cant implications for energy technology, particularly if they indicate that, because of the




non-linearity in SO2 conversion discussed  above, it  may be more important io reduce the

-------
ambient concentrations of peroxides than of SO2.




    In the 1981 JASON Committee Report to the U.S. Department of Energy (Chamberlain



et al.,  1981), some recommendations for further research were based, at least in part, on



our earlier work on SO2 oxidation  by H2O2 (Holt et al., 1981). According to our earlier



work, the 518O [deviation in parts  per thousand (%o) of the !8O/lflO ratio of the sample



from that of the standard reference material,  Standard Mean Ocean Water (SMOW)] of



sulfates produced by H2O2 oxidation were  significantly lower than the <5I8O of  sulfates



found  in rainwater. However,  the  518O of the reagent-grade H2O2 used in those exper-



iments was not known.  The results suggested the need for isotopic analysis of H2O3 in



dilute solutions, and for a methodology whereby the <518O values of H2O2, H2O, and SO^~



in rainwater could be compared, in order to assess the importance of H2O2 in the formation



of sulfate-constituted acid rain.




    The plan  of this investigation  was  to develop a method for the determination of the



<518O of H2O3 in dilute aqueous solutions (simulating rainwater);  to prepare solutions of



H2O2 of various 18O enrichments;  to use the freshly prepared solutions of H2O2  to oxidize



SO2 to SO^"  for evaluation of the  relationship between 618O~,-,2_ and 618Opj Q  ; and to



apply this relationship to the measured 618Ojj Q , 518Oo/-wa-, and £18Ojj  Q in  precipi-



tation  water, for assessment of  the importance of H2O2 in the atmospheric transformation



of SO2 to sulfate.

-------
                                   SECTION 2





                                 EXPERIMENTAL




Development of Analytical Method




    No suitable analytical method was available for the quantitative extraction of the



oxygen in H2O2, (dissolved in water, ppb range), for isotopic analysis. Figure I shows a



4-step method which we developed for this purpose. It consists of the removal of dissolved



O2 from 20-liter samples of water by a combination of evacuation, ultrasonic agitation, and



sparging with helium; oxidation of the  dissolved H2O2 to O2 in the water with KMnOv,



removal of the newly formed O2 from the water in a carrier-gas stream; and conversion of



the O2 to CO2 by reaction with platinum-catalyzed carbon at 600°C. The CO2 is then



mass spectrometrically analyzed to determine its 618O (identical to that of the oxygen in



the original  H2Oi).




    Analytical Train.   Diagrams of the apparatus are  shown in  Figs. 2 and 3. Fig. 2



shows the all-glass analytical train. It consists of a bed of molecular sieve at —196°C for



the removal of traces of O2 from helium carrier gas, a gas pipet for  the injection  of known



amounts of O2 or CC>2 into the carrier gas stream during standardization, the water sample



chamber detailed in Fig.  3, a 20-mm o.d. cold trap at  —78°C for the removal of residual



water vapor from the gas stream, a cold trap at -196°C for the removal of CO2 that is



scrubbed from the water sample by the helium stream, a bed of activated charcoal (3 g,



8-10 mesh, coconut grade, in a vertically mounted quartz  tube, lined with platinum gauze)



at 600°C for the conversion of O2  to CO2 in the carrier  gas stream, another cold trap at



— 196°C for  the collection of the newly formed CO? from  the gas stream, a capillary open-



well mercury manometer for measurement of the CO2, and a gas sample bulb, attached to




the  train for transfer of the CO2 to a mass spectrometer  for isotopic analysis. A cold trap



at —19G°C, not shown in the diagram, is in (he vacuum manifold to protect the analytical



train from vapor contamination from the mechanical vacuum pump.

-------
SAMPLE
H?0
Dissolved 02
H202
Dissolved 02
OXYGEN
REMOVAL
H2
H2
- Low prssaura H* spnrg«
•f utiraionic ogtlation
     co2
To Mass Spec
   Analysis
02 -» C02
                        Chorcool/Pt b«d
                        600°C
                                                   KMnO,
                                                    H202
                                                 OXIDATION
                                                       H20
 OXYGEN
RECOVERY
                      Low prassurt H« iporg*
                       •+• ultrasonic agitation
    Figure 1.   Four-step method for conversion of oxygen in
               (ppb range in water) to COa for mass spectrometric
               analysis

-------
              REACTION
              CHAMBER
FLOW
METER
HELIUM
         MOLECULAR
         SIEVE
         (-I96-C)
                                COLO
                                TRAP
                              (-I96-C)
                                               COLO   GAS
                                               TRAP   SAMPLE
                                             (-!96-C5 BULB
                                      CHARCOAL
                                      PLATINUM
                                      FURNACE
                                      (600'C)
                     COLO TRAP
                      (-78-C)
                                                      CAPILLARY
                                                      MANOMETER
                                                                    COLO TRAP
                                                                     (-I96-C)
                               Figure 2.  Analytical train

-------
             ANALYTICAL
                TRAIN
                      COLO WATER
                      CONDENSER
KMn04
CRYSTALS
                       20-LITER FLASK

                         ULTRASONIC BATH
   Figure 3. Reaction chamber

-------
    Reaction Chamber.  The reaction chamber assembly, Fig. 3, consists of a 20-L round



bottom  flask,  a  rotatable side-arm  tube  for the addition of crystalline KMnO4, a cold



water condenser to limit  the amount of water vapor swept by the helium stream into the



—78°C cold trap, three stopcocks for manipulation of the helium stream, and a funnel



for  the addition of solutions of acid and  oxidants (used only during the development of




the method). The round-bottom flask was supported by a stainless sleel rack in an 83-L



stainlfss-stcel tank, interior 40.6 x 50.8 x 40.6 cm, (Model ATI! 1620-21), of an ultrasonic



cleaning system  (Model EMa 50-24) manufactured  by  Branson  Cleaning Equipment Co.,



Shelton, CT. The water in the ultrasonic tank was maintained  at about 10°C by circulation



through a refrigeration unit.






    The components of the glass analytical  train were connected by No.  18 ball joints,



sealed with solidified black wax.  The hollow  plug of stopcock 22 was modified to provide



extra volume in  the capillary manometer  (Holt, 1955).










    Conversion  of O2 to  CO2. The  first part of the method to be developed was the con-



version of O2  to CO2 in  the carrier gas stream (4th box in the flow sheet, Fig.  1). The




procedure is a modification of an established  vacuum technique for the conversion of oxy-



gen in  air to CO2 for mass spectrometric analysis.  By the vacuum procedure,  the air is



circulated at low pressure over graphite and  platinum  at 600°C and the resulting CO2 is



cryogenically removed  from the closed system (Horibe et al., 1973; Holt,  1977).  By our



new method, in which helium is conducted through a bed of 8-10 mesh activated coconut




charcoal at 600°C and in contact with platinum gauze, the yield of CO2 was found to be



affected by both the bed temperature and  the axial temperature gradient. With a suitably



long furnace to  obtain an essentially  isothermal bed,  the recovery of O2 (as CO2) as a



function of bed  temperature is shown in  Fig.  4.  The data show that the optimum bed



temperature for maximum yield  of  CO2  was ~600°C, the same as that used by Horibe

-------
 (SI
o
o
    100-
     90-
     80-
70-
 (f>
 o
 CM   60
O
 CD
 >
 O
 o
     50-
40-
C£   30-



     20-



     10-



      0
        0    100  200  300  400  500  600  700  800

             Temperature of CC-Pt (°C)
   Figure 4.  Recovery of Oj as COa after reaction with charcoal

            in contact with platinum at various temperatures

-------
et al., (1973) and Holt (1977) in vacuum-line applications.




    The data in Fig. 4 also indicate that the maximum recovery of the added O2, measured



as CO2, was ~Q7%. Recoveries higher than 97% were apparently unattainable because



of competing reactions. At  temperatures  lower than 600°C, the formation of CO2  was



decreased due to the incomplete reaction,






                                  02 + C->C02                               (1)






and at higher temperatures,  the yield of CO2 was decreased by the competing reaction,






                                      + C-»2CO.                              (2)
The effect of reaction (2) was demonstrated by injecting CO2 into the carrier gas instead



of O2. The results, Fig. 5, confirm that about 3% of the CO2 decomposed at 600°C.




     The 97cc recovery is adequate for isotopic studies; and, at 600°C, oxygen isotope



fractionation within  the equilibrated system of CO2, CO,  unconverted  CO2, and  fixed



oxygen on the charcoal is negligible. The  demonstrated reliability of the 979o recovery at



600°C allows the applicability of the technique to the quantitative determination of O2 in



gas streams, as well as to isotopic studies.








Removal of Dissolved Oq from Water





     After it vas experimentally demonstrated that O2 could be reliably converted to CO2



in a carrier-gas  stream, the technique for the quantitative removal of dissolved oxygen



from multi-liter  quantities of water was developed (1st and 3rd boxes in Fig.  1).  Various



combinations of sparging with  helium, vacuum pumping,  and ultrasonic agitation were



tested. The best results were obtained by  a combination of all three. (The commonly used



degassing method of alternately freezing and thawing the water under vacuum was much




too impractical and time-consuming for large samples of water.)

-------
o
O
O
100



 90H



 80



 70-
 O
 O
     50-
 40-
 30-
     20-
     10-1
      0
                    v-
                             -V-
                                          v.
          A First run after increase of temperature
          V First run after decrease of temperature
       0   100  200  300  400  500  600  700  800

             Temperature of CC-Pt (°C)
   Figure 5.  Recovery of COj after exposure to charcoal and
           platinum at various temperatures
                         10

-------
    Helium Sparge.  Attempts were made at first to remove dissolved oxygen from water



(saturated with air) by only sparging with helium. Quantitative removal of air oxygen from



the water by this technique was intolerably slow. Curve SP-1 in Fig. 6 shows the percent



removal of O2 (estimated to be dissolved in the air-saturated water at room temperature)



plotted vs. minutes of sparging. After 50 min of sparging, only ~l/3 of the dissolved O2



had been removed.





    T. G.  Holt (1983) cited  a comparison of various degassing techniques in Technical



Brief No.  101 of Waters Associates, Milford, Massachusetts. The techniques, listed in the



order of decreasing effectiveness for the removal of dissolved gases from liquids, were: (1)



vacuum pumping combined with ultrasonic agitation,  (2) vacuum filtration using a Waters



Solvent Clarification Kit,  (3) ultrasonic  agitation,  (4) vacuum pumping, (5) boiling, and



(6) sparging.





    Vacuum-Ultrasonic Agitation  (VU).  The apparatus was arranged to accommodate



treatment of the water sample by  a combination of vacuum pumping and ultrasonic agi-



tation. The results of two experiments, VU-1 and VU-2, are plotted in Fig. 6. The rate of



removal of dissolved oxygen was found to be greatly enhanced by "rinsing" the water with



helium. That is, the evacuated sample chamber was filled with pure helium to atmospheric



pressure by allowing the gas to bubble through the water; then it was exhausted through



the analytical train. The sequential dissolution and removal of the helium from the water



had the effect of "rinsing" the  O2 from the water.  Curve VUR-1 in Fig.  6 shows the



improved results of a third experiment in which  the water sample was rinsed with He at



regular intervals throughout the O2-removal treatment.





    Vacuum-Ultrasonication-Sparging (VUS).  The improved rate of O2 removal by fre-



quently rinsing the O2 from the water with helium suggested that a  low-pressure sparge



by helium should be combined with the vacuum-ultrasonic agitation treatment. This tech-



nique, tested in runs VUS-1, VUS-2, VUS-3, and VUS-4 yielded the best O2 removal  rates.





                                        11

-------
 X
O

 CD
~0
 to
O
«*—
 O
"o
 o
 E
 CD
100-
 80-
 60-
 40-
 20-
       vus-i,?
 S = Sporgtng
 VU
    •^ Ul trosonlcal Ion
VUR B VociA^>
    + UlIrasoolcatlon
    + HjIiun "rInso"
   = VQCULITI
    4- Ul tr aiootcat Ion
    + H« I iun Sparging
                  40    60    80    100    120    140
                  Degassing Treatment Time (min)
            160
                                                                      180
Figure G.   Removal of O3 from water by  various degassing
            techniques
                               12

-------
The results show removal of ~99% of the dissolved oxygen in 1 h of VUS for 3-L water



samples.  (20-liter samples were later found to require ~5 h of the treatment for complete



air  removal.)
Oxidation cf H2O2 to O2 in Water




    Bromine water was tested as an oxidant of H2O2 to O2 in the degassed water sample.



Although it was very effective in converting a known amount of H2O2 to O2, its use was



complicated by its exposure to stopcock grease and by contamination of the analytical



train.  Curve "VU-1 + H2O2  + Br2" in Fig.  6 shows the results of cine experiment in



which  a  degassed water sample (VU-1) was spiked with H2O2 and treated with bromine



water, yielding  a recovery of about 97% in 50 min of VU treatment.




    Potassium  permanganate  was used in all subsequent experiments.  Curve  "VUS-2 +



H2O2  + KMnO4" in Fig.   6  shows the  results of an experiment made to  compare its



effectiveness to that  of bromine.  The recovery of oxygen,  added as HjO2 to a degassed



water sample (see curve VUS-2), was also ~979& after 1 h of vacuum ultrasonic sparging.



In earlier experiments with KMnC-4, it was added as a O.lN solution through the funnel,



Fig. 3;  later, the rotatable side  arm was made a part of  the apparatus to permit  the



addition of KMnC"4 crystals, with correspondingly lower procedure blanks.
Synthesis of 18O-enriched H2O2




    Hydrogen peroxides of  various <518O  were not commercially  available.  A suitable



method of synthesis  (Vol'nov et al., 1964) was identified and successfully  applied to the



laboratory preparations of four stock solutions of H2O2 of different 618O. By this method,



H2O2 is formed by exposure of supplies of water vapor, each differing in 518O, to a high-



voltage (~1.4 kV) discharge in ~100 cm of 10-mm o.d. glass tubing between two water-





                                        13

-------
cooled aluminum electrodes. Some of the HO radicals, formed by the dissociative reaction,






                                2H20 «=i 2110 +  2H                             (3)






are condensed in a cold trap (-196°C) where they combine to form H2O3, leaving the H



radicals to combine in formation of H2 and be pumped away through the vacuum line.




    Other techniques which we experimentally found to give inadequate yields of HaO2



were  conduction of an  electric arc across a stream of aerosolized water droplets (Kok,



1982), excitation of water vapor by a radio-frequency silent discharge in a glass chamber



(7.5 cm dia x  20 cm long) in a commercially available plasma cleaner unit,  and excitation



by  a  glow-discharge unit  (4.8 cm dia x 70  cm long)  that had uncooled aluminum  disk



electrodes (Jarnagin and Wang, 1958).










Collection of Rain Samples




    Each rain sampling station consisted of four 1-m2  plastic skylights,  inverted to form



funnels and fitted with  plastic nipples and hoses to allow the water to flow into two 25-L



plastic collection bottles. From each collected sample of >29-L, a 4-L bottle was filled for




subsequent determination of 618Opj Q  anc'  ^18O^-/~,2-, and a 25-L bottle was filled for



subsequent determination  of <518Oo Q . The 4-L sample was treated with 20 mg CuCl to



prevent bacterial alteration of the sulfate during storage before analysis; the 25-L sample



was treated with 25 ml concentrated H2SO4  and was refrigerated at  ~5°C to prevent




autodecomposition of the  H2O2 during storage before analysis.




    During the period  June-August  1985, rainwater collection stations,  in addition to



the one at Argonne, IL, were operated by the Research Engineering Staff, Ford  Motor



Company,  Dearborn, MI;  Research Triangle Institute,  Research Triangle Park, NC; and



Whitcface Mountain Field Station,  State University of New  York, Albany, NY. Approxi-




mately two rain evt-iits per month were sampled at each site  during the 3-month period.





                                        14

-------
    The rain collectors were installed by GP Engineering, Downers Grove, EL, at each of



the three sites.  Insulated containers were used for shipment of 25-L bottles of acidified,



chilled rainwater by air express to Argonne for analyses.
                                          15

-------
                                   SECTION 3





                                    RESULTS




Recovery of added H2O2




    The recovery of reagent grade H:O2) added to 20 liters of distilled water, was about



96%.  This recovery is adequate for determining the oxygen isotope ratio and the concen-



tration of HoOa in rainwater.











Test for Isotopic Interference




    Schumb et al. (1955) cite references showing that dissolved HaO2 does not exchange



oxygen atoms with the solvent water, dissolved molecular oxygen, or oxygenated products.



In decomposition of H2O2 by oxidation to molecular oxygen, the O-O bond is not broken



and no fractionation occurs. The MnO^"  ion does not cause isotopic interference (Cahill



and Taube, 1952). However,  in decomposition of H2O^ by reduction (not applicable to our



experimentation), the bond  is severed and fractionation may occur.




    To confirm the absence of appreciable isotopic interference in our procedure by oxygen



exchange between the H2O and either the H2O2 or the O2, before, during, or after the



oxidation reactions, the oxidation  was carried out in the presence of three different water



supplies of various 518O. The results in Fig. 7 show that the 619O of the CO2 product



was unaffected by the 518O of the water solvent.











Effects of Autodecomposition of H2O2
    Results from two sets of experiments on the autodecomoosition of H2O2 (spikes added



to 20-L samples of rainwater), over storage times of up to 11 days, showed  that if the



rainwater is stored unacidified  at  room temperature, the H2O2 concentration declines



rapidly  (~30% depletion in 2 days and ~9S% in 11 days);  and that the <518O  of the





                                        16

-------
        0
 
-------
undecomposed H2O2 undergoes a corresponding increase (~5%, during the first 5 days).
However, if the rainwater is first acidified (~20 ml cone. H3SO4 in 20-L)  and  kept cold
(1-9°C) during storage, the changes in the concentration and in the <518O of the  H2O2 are
negligible.
Effects of KMnO4 Decomposition
    Decomposition of excess KMnO4 by organic matter in rainwater was shown to product-
only CO2, which is cryogenically removed in the analytical train before the H3O3-derived
O2 is converted to  CO2.  The reaction between KMnO4 and organic matter does not
produce O2 and is th°refore not a source of interference in the method.
Procedure Blank
    The blank of the analytical procedure was reduced from 8.5 to 3.5 /imoles of O2 by
adding the KMnO4 to the 20-L sample as pulverized crystals rather than as a pre-boilod
KMnO4 solution. (In 20-L of water, 3.5 /^moles is ~3 ppb.)
Isotopic Relationships
    High-Voltage Preparation of H3O2 from H2O. The dependence of the 518O of M3O3
on the 518O of the H2O from which it was prepared by the high-vol*age discharge method
described above is shown  in Fig.  8. The equation of the  best fit  regression curve is

                       *18°H203  = 1.03<5'80H20  +29.4%,.                    (4)

    The results indicate that the <5'8Ojj2Q3 is controlled directly by the <518O of the H3O
and that the 518O of the H2O2  is substantially  higher (~29%o) than that of  the water
from which  it was formed. Although the high-voltage process used in these preparations

                                        18

-------
      60
      50-
  CM  4°-
 O
  CM
 ^

 *s   30:
 o
00

 40   20-
      10
       0
                          0
10
-20      -10
      1R
    6  0 of H20 (precursor)
H
 20
Figure 8.  Dependence of 518Oj|3Q,j on S16O^Q in a high-

         voltage discharge process ol formation
                       19

-------
of H2O2 may simulate lightning in the atmosphere, the importance of lightning, relative


to other sources of H2O2 in the atmosphere, is open to question (Kok, 1982).


    H2O2 Oxidation of SO2 to SO^". Using the four stock solutions of hydrogen peroxide,


each of different 618O, sulfate solutions of correspondingly different <518O were prepared by


oxidation of SO2 which was in isotopic equilibrium with water of constant <518O = —7.9%,.


In Fig. 9 the 618O of each resulting sulfate is plotted versus the 618O of the H2O2 and the


equation of the best-fit  regression  curve is




                       5l8°SOr  =  °'43 518°H203 + 3-5V                   (5)



    The regression curve of the  previously  determined relationship (Holt et al., 1981)


between <518Oc/~)2- and <518Oj^ Q in aqueous-phase oxidation of SO3 by H2O3 was
                               *- = 0-57 5180H20- 2.4%,.                    (6)




Assuming that all significant effects of the <518O of the SO2 on the 618O of the SO^~


are lost by rapid isotopic exchange between the SO2 and the large excess of water, prior


to appreciable oxidation (Holt et al., 1981), 618Ojj  Q an(^ ^'8^H O  rema'n as ^e on'y


two complementary variables in the equation for <518OQn2-; therefore, the comprehensive
                                                 ^»J4

regression curve for <518Ocr.2_ is
                        )2_  =  0.57 5180H20 + 0.43 5i80Ha02 + C             (7)
The constant, C, was evaluated at 8.4%, from the data given in Fig. 9 by substituting the


corresponding measured values for (518OgQ2- and <$18OH2O2' and ~7-9%> for <5l8°H2O-
    The comprehensive equation then becomes
                6180S()2_  = 0.57 <5180H30 + 0.43 5180H202+ 8.4%,             (9)




                                        20

-------
   30
   •20-
CM -
-------
and can be used to calculate 61BO^^- from 518Ojj Q  and £18Opj Q of rainwater for




comparison with corresponding measured values of $18Oq~a_. This comparison may prove



to be uniquely useful in the assessment of the importance of HaO2 in the oxidation of SO2



to SO^" in the atmosphere.




    The slope of 0.43 (approximately 2/5) in Eq. (5) of the regression curve  through the



data of Fig. 9 confirms the evidence of the intermediate species, I^CvSOg", which was



previously proposed (Holt et a/.,  1981). Apparently, the 618O of the sulfate product is



2/5-controlled by the two oxygens in the H2O2 of the adduct, and 3/5-controlled  by the



HSO^~, which, in turn, is isotopically controlled by rapid oxygen exchange with the large



excess of water with which it is associated.






6IBO and Concentration of II2O2  in Rainwater
    Forty-four samples of rainwater, collected at Argonne from September 1984 through



November 1985 and at the other three sites (in Michigan, New York, and North Carolina)



during the summer months of 1985, were analyzed by the new method. Results obtained



for 14 of these samples arc given in Table 1 and in Fig.  10. Isotopic data for H2O, SO^",



and Il2O2 in rainwater from the four sites are plotted vs.  time in Fig. 10; in addition, Table



1 gives concentrations of H2O2  and SO^" in the rainwater, and the elapsed time between



collection and analysis of each sample. Our analytical method d >es not discriminate H2O2



from organic peroxides in rainwater; consequently, in the report of field results and in the



discussions that follow, "H2O2" will refer to combined H2O2 and other active peroxides.




    Results for 25 of the samples are not reported because the concentrations of H2O2 were



too low (<15 ppb) to yield sufficient amounts of CO2  to give reliable isotopic data by mass



spectrometric  analysis.  Discarding of these samples of low H2O2  concentration  has the




effect of estimating a lower than actual proportion of sulfate formed by peroxide oxidation



of SO2 discussed later in the  report.  [The cause(s) of very low peroxide concentrations in




rainwater samples may be one or more of the following: (1) during early use of the method,





                                       22

-------
                            Table I.

Concentration and isotopic results for rainwater collected at Argonne
National Laboratory, IL, (ANL); Dearborn, MI, (DBA); Whiteface
Mountain,  NY, (WFM);  and Research Triangle Paik,  NC, (RTF);
May-September, 1985.
. Concentration 518O
Sample No.
ANL-5/27/85
WFM-6/07/85
RTP-6/ 12/85
RTP-6/ 18/85
ANL-7/02/85
DEA-7/02/85
ANL-7/09/85
RTP -7/10/85
WFM-7/ 15/85
RTP-7/ 16/85
WFM-7/31/85
DEA-8/ 14/85
RTP-8/20/85
ANL-9/06/85
Flotage
(days)
2
6
6
9
7
9
7
8
17
20
64
127
79
76
H202
(Ppb)
330
124
147
374
316
23
90
433
201
224
160
16
184
33
so*-
(mgL-1)
2.4
1.5
2.4
2.1
2.1
4.4
4.4
3.6
2.0
3.6
1.2
2.9
4.1
1.2
H202
44.8
50.5
47.5
45.0
41.4
58.3
54.2
53.8
48.4
58.0
41.0
47.9
49.6
44.8
H20
-3.1
-3.7
-3.5
-2.5
-3.1
-6.6
-1.0
-2.6
-3.9
-1.5
-7.9
-1.3
-3.5
+0.4
^ V^y >i
15.8
15.0
14.4
15.8
12.6
12.9
16.9
14.3
15.3
16.4
14.1
16.8
12.8
16.1
                              23

-------
  o
  o
 O
00
       60-
       50-
       40-1
       30-
       20-
       10-
        0-
      -10
O Argonne   A Whiteface Mountain


D Dearborn  V Research Triangle Park
           so:
            H20  O
                              1985
     Figure 10.  Measured isotopic data for H2C>2, SO4 , and H2O

               in the samples of rainwater
                            24

-------
storage procedures were inadequate for prevention of autodecomposition of peroxides; (2)



several samples were collected  during  winter months of 1984-85 when concentrations of



peroxide? in precipitation were very low, and (3) high  concentrations of SOj,  relative to



peroxides,  in the atmosphere depletes the peroxide that  might otherwise  be  present in



collected samples of rainwater.] Similarly, three samples were disqualified because of very



low sulfate concentrations (<0.3  mg L—1), and the data for two other samples were not



included because of analytical difficulties (back diffusion of CO2 from the charcoal furnace



into cold trap 2, when the line was inadvertently evacuated between the furnace and the



water sample with no helium flow).
                                         25

-------
                                    SECTION 4





                                   DISCUSSION




»518O of H2O2 in Rainwater




    A unique characteristic of atmospheric peroxides, as shown for the first time by ap-



plication of our newly developed analytical method, is the very high <518O, relative to the



618O of atmospheric water,  sulfates, and air. As demonstrated in Fig.  10, the measured



<518O of atmospheric peroxides ranged from 45 to 60%,; sulfates, IS to 17%,; and rainwater



0 to — G%O. (The <518O of air oxygen is constant at 23.5%0.)  In contrast to the very high



<518O of atmospheric peroxides, the <518O of reagent-bottle H3O3 (Fisher Scientific, H325)



was found to be —6%,.




    The oxygen isotopy  of peroxide formation  in  the atmosphere, resulting in such high



S1SO, is intriguing because it is necessarily related to the mechanism(s) by which the per-



oxide is formed. For the various sources of atmospheric peroxides (Lee, 1985), each possible



mechanism of formation is characterized by its own oxygen isotopy. Consequently, our re-



sults suggest  that oxygen isotopic studies  may well be uniquely applicable to investigations



of the origin  of peroxides in the atmosphere.




    As mentioned earlier, the 618O of peroxides in rainwater was higher than might be



expected for H2O2 that is formed by high-voltage discharge, such as lightning. For example,



application of Eq. 4 to the  range of 618O in  water (Fig.  10) would lead to an estimated



range for H3Oa of ~24%> instead of the observed 45-60%,.




    A phenomenon that  may contribute substantially to the relatively  high <518O of per-



oxides in rainwater (depending on the conditions of the rainfall, the sample collection, and



the storage)  is autodecomposition of the peroxide. As  the  peroxides decompose  (under



favorable conditions of catalyzing contaminates, temperature and pH), the <518O  by  the



residual peroxides is expected to increase. Therefore, depending on the extent of autode-



composition between the  time of liquid-phase oxidation of SOa in the  atmosphere and  the





                                        26

-------
time of collection, acidification, and refrigeration on the ground, the 61BO of the residual



H2O2 in the analyzed sample may be significantly higher than that of the H3O3 involved



in the  SO2 oxidation.  Further experimentation to establish the relationship between the



618O of peroxide in cloud water and that of concurrently collected rainwater on the ground




might provide a correction factor which would improve the significance of the isotope data



that is conveniently available from the analysis of water in  ground-based collectors.




618O of Atmospheric Sulfates: Measured vs.  Calculated




     The measured 618O values for sulfates, Fig.  10, are co-plotted  in Fig. 11 with corre-



spondingly calculated values for HjC^ oxidation, metal-catalyzed aqueous oxidation, and



primary sulfates.  The measured values are clearly higher than those calculated for metal-



catalyzed oxidation and clearly lower than those calculated cither for H2O2 oxidation or



for primary sulfates. Calculated values for H2O2 oxidation were obtained by use of Eq. (9)



for metal-catalyzed aqueous oxidation by





                 6180so2_  = 0.86180H20 + 10%, (Holt et al., 1982)            (10)






and for primary sulfates by





               <518OgO2_  = 0.015<518OH2o +45%, (Holt et d., 1984),           (11)






     In 1982, Holt et al proposed that since their measured values for 618Oc~2-  in precip-



itation water ranged consistently between the calculated values for primary sul.'-tes and



secondary sulfate? produced by metal-catalyzed  oxidations,  the fraction of atmospheric



sulfates at a given site originating as primary sulfates, could be estimated.  At that time,



however, the isotopic qualities of atmospheric peroxides and of the sulfates which they



might produce were unknown. Since then field experiments were performed near a strong




source of primary sulfates, the isotopic results of which indicated that scavengement was



~300 times more efficient for sulfates than for SO2 (Holt et d.,  1983).  Consequently, we



expect that, during precipitation, scavengement of primary  sulfates is essentially complete





                                         27

-------
 I

-------
within a few kilometers of the source, whereas beyond a few kilometers,  the scavenged



sulfates are mainly secondary sulfates, formed earlier or within the storm system by one



or more mechanisms of SO2 oxidation.



Fraction of Atmospheric Sulfates Formed by Peroxide Oxidation



    Assuming that primary sulfates of characteristically high 618O (Holt et al., 1984) are



effectively scavenged by rain within a short distance from their sources, an estimation of the



fraction of sulfates in rainwater several kilometers from a strong source can be made from



relative deviations of the measured <518Oc;r.2- values  from the corresponding calculated
                                       oU4


<518O<-,,^2- values for peroxide oxidation and metal-catalyzed O2 oxidation, respectively.



    Using the following relationship for calculation of the percentage of sulfate in rainwater



formed by H2O2 oxidation,
                         4  (meaa.)          4  (calc.,  O2 oxdn.)

                                                                        100'
                   4  (calc., H2O2 oxdn.)       b(J4  (calc., O2 oxdn.)



the percentages for our four sampling sites during 1985 are plotted vs.  time in Fig.  12.



The average for all of the values at all of the sites is 37 ± 8% sulfate, formed by peroxide



oxidation.



     Equation (12) assumes that  all sulfate measured in the rainwater samples was pro-



duced within the raining cloud and that only H2O2 and metal-catalyzed aqueous oxidation



contributed to the oxid if ion of the dissolved SO2. Other potential sources of sulfate in the



rainwater Camples were aqueous-phase ozone oxidation of SO2, dissolution after gas-phase



reaction of SO2 with OH radicals, and dissolution  of sulfate nuclei previously formed in



non-precipitating clouds by any of the above oxidation processes.



     In reviewing our work, Dodge (1986) pointed out that the results of other researchers



(Scott, 1982; Hegg and Hobbs, 1984; Scire and Venkatram, 1985) variously indicate that 65



to 85% of sulfate in oloudwater is due to in situ aqueous phase reactions and the remainder




                                         29

-------
    100
    90-
    80-
    70
    60-
•z.
UJ
O
C£
UJ
Q_
    40-
    30
    20-
     10-
O Argonne   A Whiteface Mountain

D Dearborn  V Research Triangle Park
     O
                     V  _ „ , -A
                         ~
                                         •-- o
                            1985
  Figure 12.  Percent of sulfate formed by peroxide oxidation,

            assuming negligible primary sulfates
                          30

-------
to cloud scavenging of pre-existing sulfates. However, since the residence time of aerosol



sulfate in the atmosphere is rather short, any aerosol sulfate scavenged could be expected



to be about the same in isotopic quality as that in the precipitating cloud system.  Further,



ozone is expected to play a minor role in the oxidation of SO2 in cloudwater of pll lower



than about 5 (Penkett et a/., 1970).




    At best, these data establish  a lower limit for peroxide oxidation. As discussed above,



the <518O of atmospheric peroxides  may  be  somewhat  lower than  that which is finally




measured from the ground-based rain collector.  It follows that the calculated 618O^~2-



would be correspondingly lower, and the *>o of sulfate formed by peroxide oxidation would



be correspondingly higher.




    The technological significance  of  our findings (that  ~40% or more of acid sulfate



in rainwater originates from peroxide oxidation  of SOa) is that further investigations  of




the origin(s) and of possible methods of control of atmospheric peroxides are of prime



importance.
                                         31

-------
                                    SECTION 5




                                    SUMMARY



    A new method was  developed for the determination of the 618O of peroxide (ppb



range) in rainwater.  Experimental results showed the method to be reliable with respect



to recovery of added HaO^, blanks, and freedom from isotopic interference.  A procedure



for collection  and treatment  of 25-L samples of rainwater with minimum  peroxide de-



composition between collection  and analysis was developed. An isotopic relationship was



established between the oxygen  in sulfate and the oxygen in the peroxide and in the water



involved in the peroxide oxidation of SO2 to SO^". By using this relationship for peroxide



oxidation  and a similar relationship previously determined for aqueous metal-catalyzed



oxidation, the fraction of sulfates formed by peroxide oxidation ;n rainwater from  four



sampling sites in Illinois, North Carolina, New York, and Michigan was estimated to be



~\Q%.  This fraction could be substantially  higher  if a  substantial amount of peroxide



undergoes decomposition between the time of oxidation  of SOa in the atmosphere and



the time of collection and peroxide-stabilization treatment of the rainwater sample on the



ground. Our results show that  atmospheric peroxides play a major role in the formation



of sulfates in  the atmosphere,  and, therefore,  that  it is essential  to  obtain all  possible



information on their origin and control.
                                        32

-------
                                 REFERENCES





Cahili, A. E. and Taube, H.  The use of heavy oxygen in the study of reactions of hydrogen



    peroxide. J. Amer. Chem. Soc., 74, 2312 (1952).




Chamberlain, J., Foley, H., Hammer, D., MacDonald, G., Rothaus, O., and Ruderman, M.



    The Physics and Chemistry of Acid Precipitation, Technical Report JSR-81-25, pre-



    pared for U.S. Department of Energy, September 1981.




Dodge, M. C.  Private communication, U.S. Environmental Protection Agency, Research



    Triangle Park, NC, 1988.




Hegg, D. A., Hobbs, P. V.,  and Radke, L. F.  Measurements of the scavenging of sulfate



    and nitrate in clouds. Atmos. Environ., 18, 1939 (1084).




Holt, B. D.  Use of capillary  trap in microdetermination of carbon. Anal. Chem., 27, 1500



    (1955).




Holt, B. D.  Preparation of carbon dioxide from sulfates,  sulfur dioxide, air, and water for



    determination of oxygen isotope ratio.  Anal. Chem., 49, 1664 (1977).




Holt, B. D., Kumar, R., and Cunningham, P. T. Primary sulfates in atmospheric sulfates:



    estimation  by oxygen isotope ratio measurements. Science, 217, 51 (1982).




Holt, B. D.,  Kumar, R., and Cunningham, P. T.  Oxygen-18 study of the aqueous-phase



    oxidation of sulfur dioxide. Atmoa.  Environ., 15, 557 (1981).




Holt, B. D. and Kumar, R.  Oxygen-18 study  of high-temperature air oxidation of SOj.



    Atmos.  Environ., 18, 2089 (1984).




Holt, B. D,, Nielsen, E., and Kumar, R.  Oxygen-18 estimation of primary sulfate in total



    sulfate scavenged by rain from a power plant plume. "Precipitation Scavenging, Dry



    Deposition, and Resuspension," Vol. 1.  "Precipitation Scavenging," H. R. Pruppacher,



    R. G. Semonin, and W. G. N. Slinn, cds., pp. 357-368, Elsevier, New York, 1983.




Holt, T. G.  Private communication, University of Illinois, Urbana, 1983.
                                       33

-------
Horibe, Y., Shigehara, K., and Takakawa, Y.  Isotope separation factor of carbon dioxide-



    water system and isotopic composition of atmospheric oxygen. J. Geophys. Res., 78,



    2625 (1973).




Jarnagin, R. C. and Wang, J. H.  Investigation of the catalytic mechanism of catalase and



    other ferric compounds with doubly 18O-Iabeled hydrogen peroxide. J. Amer.  Chem,



    Soc., 80, 786 (1958).




Kok, G. L.  The  production of hydrogen peroxide by lightning.  Paper presented at the



    American Geophysical Union Meeting, San Francisco, 1982.




Lee,  Y.  N.  Chemical transformations in acid rain; Vol.  DL Investigation  of kinetics



    and mechanism of aqueous-phase peroxide formation.  U.S. Environment Protection



    Agency Technical Report EPA/600/09 (1985).




Penkett, S. A., Jones, B. M. R., Brice, K. A., and Eggleston, A.  E.  J.  The importance



    of atmospheric ozone and hydrogen peroxide in oxidizing sulfur dioxide in cloud and



    rainwater. Atmos. Environ., 13, 123 (1979).




Schumb, W. C., Satterfield, C. N., and VVentworth, R. L. "Hydrogen Peroxide," Reinhold



    Publishing Corporation, New York, (1955).




Scire, J. S., and Venkatram, A. The contribution of in-cloud oxidation of SO3  to wet



    scavenging of sulfur in convective clouds. Atmos. Environ., 19, 637 (1985).




Scott, B. C. Predictions of in-cloud conversion rates of SOa to SO4 based upon a simple



    chemical and kinematic storm model. Atmos. Environ., 16, 1735 (1982).




Vol'nov, I. I., Tsentsiper, A. B., Chamova, V. N., Latysheva, E. L, and Kuznetsova, Z. I.



    Synthesis  of hydrogen [18O] peroxide from the dissociated vapor of [18Oj water in a



    glow discharge. Russian J. Phys. Chem., 88, 645 (1964).
                                       34

-------
Page Intentionally Blank

-------