United States
Environmental Protection
Agency
Environmental Sciences Research
Laboratory
Research Triangle Park NC 27711
EPA-600 2-79 095
May 1979
Research and Development
Determination of
Trace Quantities of
Sulfate Ion

New Approaches to
an Old Problem

-------
                RESEARCH REPORTING SERIES

Research reports of the Office of Research and Development, US. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology.  Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:

      1.  Environmental  Health  Effects Research
      2.  Environmental  Protection Technology
      3.  Ecological Research
      4.  Environmental  Monitoring
      5.  Socioeconomic Environmental Studies
      6.  Scientific and Technical Assessment Reports (STAR)
      7,  Interagency Energy-Environment Research and Development
      8.  "Special" Reports
      9.  Miscellaneous Reports

This report has been assigned  to the ENVIRONMENTAL PROTECTION TECH-
NOLOGY series. This series describes research performed to develop and dem-
onstrate instrumentation, equipment, and methodology to repair or prevent en-
vironmental degradation from point and non-point sources of pollution. This work
provides the new or improved technology required for the control and treatment
of pollution sources to meet environmental quality standards.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.

-------
                                                    EPA-600/2-79-095
                                                    May 1979
          DETERMINATION OF TRACE QUANTITIES OF SULFATE ION
                  New Approaches to an Old Problem
                            D. T.  Bostick
                    Analytical Chemistry Division

                                 and

                            W. D.  Bostick
                    Chemical Technology Division

                   Oak Ridge National Laboratory*
                     Oak Ridge, Tennessee 37830
                       Contract No. 40-S10-75
                          Project Officers

                          Eva Wittgenstein
                           James D. Mulik
             Atmospheric Chemistry and Physics Division
             Environmental Sciences Research Laboratory
            Research Triangle Park, North Carolina 27711
             ENVIRONMENTAL SCIENCES RESEARCH LABORATORY
                 OFFICE OF RESEARCH AND DEVELOPMENT
                U.S. ENVIRONMENTAL PROTECTION AGENCY
            RESEARCH TRIANGLE PARK, NORTH CAROLINA 27711
Operated by UNION CARBIDE CORPORATION for the DEPARTMENT OF ENERGY,
 Contract No. W-7405-eng-26.

-------
                            DISCLAIMER
     This report has been reviewed by the Environmental Sciences Research
Laboratory, U.S. Environmental Protection Agency, and approved for pub-
lication.  Approval does not signify that the contents necessarily
reflect the views and policies of the U.S. Environmental Protection
Agency, nor does mention of trade names or commercial products con-
stitute endorsement or recommendation for use.

     The Oak Ridge National Laboratory is operated for the U.S. Depart-
ment of Energy by Union Carbide Corporation under Contract W-7405-eng-26.
This article was supported by the Basic Energy Sciences Division.
                                 11

-------
                              ABSTRACT

     Several  analytical  methods for possible use in measuring trace
amounts of water soluble sulfate anions were reviewed and evaluated.
Enzymatic sulfate determination does not appear to be a  viable approach
until the required enzymes can be obtained commercially.   Gas chromato-
graphic analysis of bis(trimethylsilyl)sulfate was studied and,  with
further development, may be a selective method for the determination  of
sulfate as well as other oxy-anions.  A direct kinetic method, based  on
the ability of sulfate to catalyze the depolymerization  of zirconyl
species, was investigated.  The method can detect 0.2-20  ppm sulfate
and, with a minimum of sample preparation, is relatively  selective.
                                  m

-------
                              CONTENTS
Abstract	 iii
Fi gures	  vi
Tabl es	 yi i
Acknowledgement	viii

1.  I ntroducti on	   1

2.  Conclusions and Recommendations	   4

3.  Proposed Enzymatic Schemes for Sulfate Analysis	   7
         The sulfite oxidase reaction	   7
         The sul fite transferri ng enzymes	   8

4.  The Gas Chromatographic Analysis of Bis(trimethylsiyl)-
    Sulfate...	  12
         Previous investigations...	  12
         Preparation of bis(trimethylsilyl)sulfate standards.  12
         GC separation and analysis of TMS-SO.	  15

5.  A Spectrokinetic Procedure for Sulfate Analysis	  21
         Previous investigations	  21
         Reagents and instrumentati on	  21
         Experimental procedure and data analysis	  23
         Optimization of reaction conditions	  30
         Reaction interferences	  34
         Precision and accuracy of results	  39

References	  44

-------
                                FIGURES
Number                                                         Page
  1        Sulfate Transferring System	   9
  2        Effect of Solvent on Silylation of
           Ammoniurn Sul fate	  14
  3        Separation of TMS-SO^ on Various GC Columns	  18
  4        ZR-MTB Reaction	  22
  5        Optical Transmission of Sulfate Standards
           and Sampl es in Mul ticuvet Rotor	  25
  6        Reaction Progress of Sulfate Ion Standards	  26
  7        Li near Search Computer Subrouti ne	  27
  8        Data Output Format for Linear Search
           Computer Subrouti ne	  29
  9        Sulfate Ion Calibration Curve	  31
 10        Effect of ZR Polymer Concentration on
           Sulfate Ion Reaction Rate	  32
                                   VI

-------
                            TABLES

Number                                                     Page

   1       Comparison of GC Columns  for the
          Separation and Detection  of TMS-S04	    16

   2       Interferring Ions and Methods for
          Thei r Removal	    35

   3       Effect of Sample Pretreatment on
          Sulfate Analysis	    36

   4       Run-to-Run Reproducibility in Sample
          Analysis	    40

   5       Day-to-Day Reproducibility in Sample  Analysis...    41

   6       Correspondence of the ZR-MTB with  a Reference
          Method for the Determination of Sulfate  Ion	    43
                              vn

-------
                          ACKNOWLEDGEMENT
     The authors wish to thank Dr. Kenneth Monty for his advice and
for providing the Salmonella typhimurium culture used in our studies.
                                   vm

-------
                             SECTION 1
                           INTRODUCTION

     There are many methods currently available for the determination
of water soluble sulfate.  The majority of these techniques are based
on the reaction of sulfate with a barium salt to form the precipitate,
barium sulfate.  Direct sulfate analysis has been achieved by quanti-
tatively determining the amount of precipitate formed either turbidime-
trically, gravimetrically (1), or nephelometrically (2).  Sulfate con-
centration has also been determined indirectly by measuring excess
barium ion, or its inorganic counter ion, in solution using atomic
absorption (3), ion selective electrode (4, 5), or titrimetric (6)
procedures.  The above techniques usually exhibit only a modest sensi-
tivity and reproducibility, and are often not applicable to the analysis
of samples containing less than 2 ppm sulfate.  Alternatively, several
colorimetric procedures have been developed which employ barium salts
composed of organic bases to slightly improve the sensitivity of the
barium sulfate reaction (1).  In the latter case, the change in free
counter ion concentration is measured.
     Even with this and numerous other modifications in the above methods,
sulfate analysis is still subject to a number of difficulties which are
inherent to the barium sulfate reaction.  The sensitivity of the
analysis based on this reaction is limited by the solubility of barium
sulfate in solution.  Additionally, other divalent anions and cations
react with either barium or sulfate, respectively, and bias results.
These interferents therefore must be removed from the sample prior to

                                  1

-------
analysis.  Aqueous samples often contain sulfite and sulfide, in
addition to sulfate.  These species also react with the barium salts,
and thus total sulfur anion rather than sulfate is actually being
measured.
     Several novel reaction schemes have been evaluated in order to
overcome the deficiences encountered with the barium sulfate raction.
Sulfate analysis using an enzymatic reaction was proposed in an effort
to establish a more selective technique.  Enzymes characteristically
react specifically with micromolar concentrations of their substrates.
Hence, the anions and cations which interfere with the barium sulfate
reaction should not affect the enzymatic sulfate reactions.  Additionally,
the enzyme will react with only the single sulfur anion for which it is
specific.  Two enzymatic reaction schemes were investigated; one pro-
cedure used sulfite oxidase to indirectly determine sulfate and the
other required a three-enzyme "sulfate transferring system" to form a
chromogen which is directly related to sulfate concentration.
     Reaction selectivity can also be achieved by using chromatographic
techniques to separate sulfate from other anions present in a sample.
A gas chromatographic procedure for sulfate was evaluated which is
based on the formation of a volatile trimethylsilyl (TMS) derivatives
of oxy-anions.  Such a technique provides a separation of sulfate
from other potentially interfering species, thereby permitting the
selective analysis of sulfate and possibly other separated anions
contained  in  the  sample  as well.

-------
     A direct kinetic method for sulfate analysis was evaluated in a
third reaction scheme.  The kinetic method might be more sensitive and
reproducible than those methods based on the barium sulfate reaction,
and is amenable to the use of masking agents which might improve the
selectivity of the analysis.  A colorimetric procedure for sulfate was
previously reported, which was based on the catalytic properties of
sulfate to depolymerize a zirconyl species (7).  Sulfate concentration
was found to be linearly proportional to the rate of formation of a
free zirconyl-methylthymol blue chromogen.  This kinetic procedure has
been adapted to the miniature centrifugal fast analyzer, and masking
agents have been added to the reaction mixture to improve the selectivity
of the kinetic reaction.
     This report reviews in detail the three proposed reaction schemes,
summarizes the results obtained to date with each procedure, and eval-
uates their potential for providing a sensitive, and selective means for
determining sulfate concentration.

-------
                              SECTION 2
                   CONCLUSIONS AND RECOMMENDATIONS

     Enzymatic analysis of sulfate ion may represent a selective and
sensitive approach for the determination of micromolar concentrations.
The enzymes are reported to react specifically with sulfate rather than
total sulfur anion, as in the barium sulfate procedures.  If sufficient
enzyme activity is present in the reaction mixture, and if the reaction
proceeds toward equilibrium any enzymatic inhibition by other am'ons
or cations in the sample will be negligible.  The enzymatic analysis
of sulfate appears feasible since the same reaction sequences have
been previously used in the laboratory, although for other applications.
However, if the enzymatic analysis of sulfate is to have wide applica-
tion in the general chemistry laboratory, a reliable and stable source
of the enzymes must be available commercially.  Until such time, addi-
tional development of this procedure does not appear fruitful.
     Selective sulfate analysis can also be achieved by using chromato-
graphic procedures to separate sulfate from the interfering anions.  The
gc analysis of TMS-SO. may be easily adapted to the analysis of sulfate
and  be easily performed in a general chemical laboratory.  To date,
0.2  yg ammonium sulfate in aqueous solutions can be analyzed repro-
ducibly if the sample is evaporated to dryness and silylated in the
presence of bis(trimethylsilyl)trifluoroactamide and methylene chloride.
For  more dilute environmental samples, sulfate ion may be concentrated
on an anion exchange column and subsequently eluted by a concentrated
ammonium salt.  Thus, a concentration of sulfate and conversion to its
                                  4

-------
ammonium salt might be performed in a single step with an ion exchange
column to analyze less-concentrated environmental samples.  Precipitating
agents might also be useful to concentrate sulfate and remove water
prior to the silylation reaction.  The sensitivity of the gc procedure
might also be improved by investigating other gc column supports and
liquid phases.  A silylized glass capillary column would probably improve
the detection limit by eliminating the need for a solid support.  Finally,
either selective or non-selective gc detectors may be used to analyze
the TMS-SO, or all volatile oxy-anion derivatives, respectively, in a
sample.  Further studies in the gc analysis of sulfate appear profitable,
since the technique promises to provide a procedure which is signifi-
cantly more selective than the barium sulfate methods and is applicable
to several other anions as well.
     The kinetic Zr-MTB procedure is both more sensitive and selective
than the barium sulfate methods.  A detection limit of 0.3 ppm sulfate
can be measured and reproducibilities of 2% are obtainable if care is
taken in loading the rotor.  The major cation interferences are removed
with batch cation exchange treatment and fluoride and sulfide can easily
be masked with Al(III) and Hg(II), respectively.  Highly colored samples
may be accurately analyzed with a minimum of interference.  The instru-
mentation permits the simultaneous processing of standards and samples
in the 0.3-20 ppm sulfate range.  The linear reaction data is located
and the reaction rate calculated by a linear regression computer routine,
limiting the computing time required of the analyst.  Further studies
should be directed toward locating an appropriate masking agent for
phosphate and arsenate ions.  This would eliminate the need to use
                                  5

-------
magnesium oxide adsorption in samples containing these interferents.
By removing cations and adding a mixture of masking agents to both
samples and standards, the kinetic Zr-MTB technique would provide a
procedurally simple, semi-automated and selective analysis of sulfate
in environmental samples.

-------
                              SECTION 3
           PROPOSED ENZYMATIC SCHEMES FOR SULFATE ANALYSIS
                     THE SULFITE OXIDASE REACTION

     An indirect method for sulfate analysis can be achieved by
chemically reducing sulfate to sulfite and coupling this reaction
with the following enzymatic reaction:
          SQ2- + ^ Sulfite Oxidase, SQ2-
The sulfite concentration is then proportional to the production of
hydrogen peroxide, which can be measured by either colorimetric (8),
fluorimetric (8b), or chemiluminescent (9) procedures, all of which
would provide great sensitivity.  A sample would be analyzed by first
determining the sulfite concentration initially present in the sample.
A second sample aliquot would then be reduced and the total sulfite con-
tent determined.  The difference between the two measurements would
represent the sulfate concentration in the sample.  Such a reaction
scheme would provide a specific and sensitive means for analyzing both
sulfate and sulfite.  The procedure may be automated by use of reduction
and immobilized enzyme columns in a flow system.
     The feasibility of this enzymatic method is dependent upon the
reproducibly of the sulfate reduction, and therefore a study of several
reducing agents was initiated.  The reducing characteristics of various
granulated metal reactors were investigated first with the expectation
that no undesirable side products would be formed in the reduction

-------
that could interfere with the subsequent enzymatic  reaction.   Nitrate
reduction to nitrite and ammonia has been accomplished  by  using  such
reducing columns containing copper-coated cadmium granules (10),
Devarda's alloy (11, 12), a Jone's reductor (13), and copper granules
(14).  Similar columns and conditions were used in an attempt to
reduce sulfate to sulfite.  Ethylenediaminetetraacetic acid was added
to sulfate and sulfite standards to prevent formation of insoluble
metal sulfites and sulfides, as well as  to protect against subsequent
metal poisoning of the enzyme.  Although no reduction of sulfate was
observed with any of the metal  reducing  columns, 95% of sulfite standard
introduced into the columns was recovered.
     Sulfate reduction to  sulfite and sulfide was observed using
sodium borohydride pellets.  Optimizing  pH and borohydride concen-
tration may permit the reproducible reduction of sulfate to sulfite;
however, unreacted borohydride would have to be removed prior to the
enzymatic reaction.
     Because sulfate reduction has met with only limited success,
further studies in this direction were postponed in order to review
the  potential of a second enzymatic method for sulfate analysis.

THE  SULFATE TRANSFERRING ENZYMES
     A direct procedure for sulfate determination was proposed using
the  three-enzyme "sulfate transferring system."  The colorimetric
method, outlined in Figure 1, takes place in the presence of excess
adenosine triphosphate (ATP).  Sulfate is transferred to ATP in the
                                 8

-------
                                           ORNL-DWG. 76-3434A
             SULFATE  TRANSFERRING SYSTEM
(1,    ATP + SO
                     Sulfate
              .  Adenyltransferase
                 -

                            0  0
                                           NH
                                 QSOPOH-C  0
                                                   +  Pp
                                              'H

                                           OH OH

                                           APS
                                       Adenylsulfate
(2)  ATP -I- APS
                       APS Kinase
                                               "
                                 0 0
                                 II II
                                OSOPOH?C
                                 II II
                                 0 0
                                         P04 OH
                                                     ADP
                                          PAPS
                          3'-phosphoadenosine-5'-phosphosulfate
                          Phenol
                         Sulfurylase
      (3)   PAPS + Phenol  '           PAP + Phenol Sulfate
°r
                           PAPS
(3)  PARS-, 2e
                                   PAP
               Figure  1.  Sulfate Transferring System

-------
 presence  of  sulfate adenyltransferase  to  form adenylsulfate  (APS) and
 inorganic pyrophosphate  (PP.).  APS  kinase  catalyzes the second  reaction
 in which  APS and ATP  react to form 3'-phosphoadenosine-5'-phosphosulfate
 (PAPS)  and adenosine  diphosphate  (ADP).   Although  it is feasible to mon-
 itor reactions  1 and  2 by the consumption of ATP,  this would not be
 specific  for sulfate  since sulfate adenyltransferase (reaction 1) also
 reacts  with  molybdate, tungstate, selenate, chromate, and sulfite.
 However,  none of these other anions  produce PAP  (reaction 3).  Thus,
 reaction  3 can  be  used as a specific indicator reaction for  the  sul-
 fate reaction by monitoring the consumption of the phenol substrate.
      Sulfate analysis using the three-enzyme "sulfate transferring
 system" appears to be a  feasible  approach, since various individual
 aspects of the  reaction  scheme have  been  successfully employed for
 other purposes.  The  first two enzymes are currently being used  to
 prepare both APS and  PAPS.  The third  reaction has been used in  com-
 bination  with the  two ATP-dependent  reactions to quantitatively  deter-
 mine PAPS and PAP  in  clinical samples.  Literature sources describe
 the  use of various phenol substrates,  including p-nitrophenol and
 m-aminophenol,  as  well as methylumbilliferone, as spectrometric  indi-
 cators  for the  reaction  sequence.  By  appropriate selection  of reaction
 conditions,  the overall  reaction may be made dependent upon  sulfate ion
 concentration and  be  used as the basis for a novel approach  to sulfate
 analysis.
     To implement this reaction scheme, it was necessary to  locate a
source of  the enzymes.  Sulfate adenyltransferase is the only enzyme
of the three which is currently available commercially.  Several  bio-
chemical companies are presently attempting to isolate APS kinase, but
                                  10

-------
have not, as yet, obtained a stable preparation.  We have made several
attempts to prepare active APS kinase from yeast (15) and phenol sul-
furylase from guinia pig liver (16, 17).  No measurable enzyme activity
could be demonstrated in any of the preparations isolated from these
sources.  An attenuated strain of Salmonella typhimurium mutant was
also investigated as an enzyme source.  The bacteria possesses sulfate
adenyltransferase and APS kinase but completes the reaction sequence
by reducing the transferred sulfate to free sulfite, as shown in
reaction 3'.  A method for sulfate analysis was studied in which intact
cells are used as the enzyme source to microbiologically reduce sulfate
to sulfite-  The liberated sulfite is then measured via the West-Gaeke
colorimetric procedure (18) and is proportional to the original sulfate
concentration in the sample.  The feasibility of such a procedure is
demonstrated by the fact that an identical procedure has previously
been used to study genetic coding in Salmonella mutants (19).  A culture
of the mutant was obtained and initial studies were made to demonstrate
proof of principle.  Although the culture could be grown and cells
harvested,  second generation cells did not exhibit the ability to
reduce sulfate to sulfite.  Until a suitable source can be obtained for
the three enzymes, further assessment of sulfate analysis based on the
sulfate transferring system can not be made.
                                  11

-------
                              SECTION 4
   THE GAS CHROMATOGRAPHIC ANALYSIS OF BIS(TRIMETHYLSILYL)SULFATE
    PREVIOUS  INVESTIGATIONS

     A gas chromatographic (gc) technique was previously developed for
 the  determination of milligram quantities of oxy-anions contained in
 aqueous  samples  (20, 21).  The procedure required the conversion of
 anions to their  ammonium salts, followed by reaction with a silylating
 reagent  to form  the corresponding trimethylsilyl (TMS) derivatives of
 the  anions.   The volatile derivatives were subsequently separated and
 detected by  gc.  This  procedure was modified for the analysis of micro-
 gram levels  of sulfate in environmental samples.

 PREPARATION  OF BIS(TRIMETHYLSILYL)SULFATE STANDARDS
     Preliminary studies were concerned with the preparation of reliable
 bis(trimethylsilyl)sulfate (TMS-S04) standards.  Butts and Matthews pre-
 pared TMS-SO, samples  by adding 5-10 mg sulfate as either solid or
 aqueous  ammonium sulfate to 200 yl bis(trimethylsilyl)trifluoroacetamide
 (BSTFA).  Butts, et al. also included 200 pi dimethylformamide to the
 reaction mixture.  The resulting solution was shaken and then allowed
 to react overnight at  25°C.  The TMS-SO. yield was often irreproducible;
 occasionally, no TMS-SO. could be detected.  Several variables, including
water content, organic solvent, reaction time and reaction temperature,
were investigated in order to optimize the chemical reaction conditions
for maximum,  reproducible TMS-SO. yields at the microgram sulfate con-
centration range.
                                 12

-------
     We found that aqueous sulfate samples must be evaporated to dryness
prior to silylation.  If greater than 10 yl of water is present in the
reaction mixture, a precipitate will form and no TMS-SO, will be pro-
duced.  The presence of smaller quantities of water will result in
less than maximum yield of the derivative, as well as severe gc peak
tailing and irreproducible peak heights.
     The reaction yield is also affected by the presence of organic
solvents in the reaction mixture.  The percent yield of TMS-SO. and the
production of secondary products were compared when using BSTFA alone,
or in combination with dimethylformamide, acetonitrile, pyridine
or methylene chloride.  We found that the presence of an organic
solvent, in addition to BSTFA, was required to form TMS-SCL.  The
greater the solvent-to-BSTFA volume ratio, the more quantitative the
yield and the faster the reaction progressed.  A ratio of 2.5 was found
to be optimum.  Methylene chloride was found to be the solvent of choice.
In addition to producing the greatest yield of TMS-SO., this volatile
solvent elutes the gc column rapidly, resulting in a low gc background
(Figure 2).  Methylene chloride does not produce secondary products,
as do both dimethylformamide and pyridine.  No side reactions were
formed in the presence of acetonitrile; however, the TMS-SO, yield
was relatively low.  Optimum sample composition therefore includes
the addition of 250 yl methylene chloride and 100 yl BSTFA to an
evaporated ammonium sulfate standard.
     The effect of reaction temperature on the percent-yield of TMS-SO,
was also investigated.  No TMS-SO, is produced if the temperature is
less than 25°C.  Fifty percent of the maximum yield is obtained at 45°C,
                                 13

-------
                                                          ORNL DWG 78-13950
DIWETHYLFORMIMIDE
ACETONITRILE
PYRIDINE
METHYLENE CHLORIDE
 l    i    i    t    I   i    I    i
   Figure  2.   Effect of Solvent  on Silylation of Ammonium Sulfate
                                  14

-------
whereas maximum yield is obtained above 55°C.  A 60°C reaction tempera-
ture is used to prepare sulfate standards.  At reaction temperatures
above this value, solvent loss due to leakage during sample preparation
becomes a problem.  Even at elevated temperatures, an overnight reaction
time is required for maximum yield.  Using the above optimized reaction
conditions, reproducibility in the preparation of standard replicates
is +2% and day-to-day reproducibility in standard preparation is +5%.

GC SEPARATION  AND ANALYSIS OF TMS-S04
     The gas chromatographic conditions were also optimized for micro-
gram quantities of TMS-sulfate.  The derivatized oxy-anions were
previously separated using a 5% SE-30 coating on Chromosorb G(HP)  (21).
The analysis of sulfate on this support was found to be very non-
reproducible,  due to irreversible adsorption and/or decomposition  of
TMS-sulfate on the column.  The detection limit of the analysis was
reported as 2  ug sulfate.
     Several columns have been prepared to determine the gc parameters
which contribute to the loss of TMS-SO..  Table I describes five
columns which  either differ in the solid support used or in the percent
loading of the liquid phase, SE-30.  The performance of these columns
was compared with respect to the amount of tailing of the TMS-SO.  peak,
the detection  limit, and the linearity of the calibration curve.  Peak
tailing is described using a tailing factor (T), defined as

          T =  100 x W]/W2
                                 15

-------
                                                  TABLE  1
                                      COMPARISON OF GC COLUMNS FOR THE
                                     SEPARATION AND DETECTION OF TMS-SO,
GC r0SE-30 Optimum Detection
Column Loading Type of Column Retention Tailing Limit Relative
# (w/w) Support* Temperature Time (sec) Factor (yg SOff) Sensitivity
1 5 Gas Chrom Q 130°C 93 15 1 0.54
2 5+0.5% Gas Chrom Q 140°C 135 8 1 4.74
Carbowax
20M
3 5 Chromosorb 140°C 103 33 0.8 33.1
W(-HP)
4 1 Chromosorb 95°C 102 80 0.25 32.5
W(HP)
5 1 Chromosorb 90°C 93 100 0.2 37.3
750
Correlation
Coefficient*
0.973
0.992


0.990

0.997

0.998

*80/100 mesh support, 6' glass column, 0.25 in OD, 60 ml  argon flow

-------
where W-j is the base width of the first half of the gc peak and W2 is
the base width of the second half.  The base widths are measured at
10% of peak height.  Since the gc column tends to absorb TMS-SO,, the
practical limit of detection was estimated by calculating the x-inter-
cept of the TMS-SCL calibration curve.  The relative sensitivity was
determined by comparing the slope of the calibration plots at identical
amplifier attenuations.  The linearity of the calibration curves was
estimated by comparing the correlation coefficients obtained from a
linear regression analysis of sulfate standards.  Results of such a
comparison should suggest alternate columns which would improve the
reproducibility and detection limit of TMS-SO, analysis.
     Gas Chrom Q is similar to Chromosorb G(HP), used in the original
study  (20), except that it is distributed by another manufacturer.
Product specifications state that the support should be used to minimize
tailing and to eliminate catalytic decomposition of sample constituents.
Gas Chrom Q, with 5% SE-30 (column 1), performed similarly to the ini-
tial column using Chromosorb G(HP).  The column had to be "loaded" with
several injections of microgram quantities of TMS-SO. before any peak
could  be observed.  Once the column was conditioned, 4 yg sulfate and
greater could be analyzed reproducibly if samples were injected at
equal  time intervals.  The TMS-SO. peak width broadened considerably
and the retention time increased "by at least 15 seconds as the amount
of TMS-SO. injected was decreased.  The least concentrated standards
gave no  response unless at least 4 ug sulfate was injected just prior
to their analysis.  The 1 ug detection limit observed was slightly
improved over that reported for the Chromosorb G(HP).  Peak tailing
was relatively severe  on this support  (see Figure 3).
                                  17

-------
                                                                                                 ORNL DWG 78-13949
                60
                 50
                40
                          tR-93S
00
                 2O
                 10
                                                                                        tR • 102 S
                                                                                                       tR-93S
                         COLUMN l<
                       5% SE-3O ON
                       GAS CHROM 0
                       ATTN-64x10*
       COLUMN 2
5% SE-30t0.5% CARBOWAX
  20M ON GAS CHROM Q
     ATTN-64xlO2
     COLUMN 3<
   5% SE-30 ON
CHROMOSORB W(HP)
   ATTN-I6XI03
                                                                                                  l_
                                                                                                     J_
                                                                                                         I
                                                                                                             I
    COLUMN 4*
   1% SE-30 ON
CHROMOSORB W(HP)
  ATTN • I6XI03
   COLUMN 5'
  1% SE-30 ON
CHROMOSORB 750
  ATTN -I6x 10s
                                  Figure 3.   Separation of  TMS-SO^ on  Various GC  Col
                                                     umns

-------
     A 0.5% Carbowax 20M coating was used on Column 2 to cover any
active sites present on Gas Chrom Q, prior to loading the support with
5% SE-30.  This procedure has frequently been used to reduce peak
tailing and to increase the inertness of a column.  Peak broadening
at low IMS-SO. concentration was not as severe on this column, although
the retention time increased by as much as 50 seconds for dilute samples.
Tailing of the sulfate peak was more severe than without the Carbowax
coating.  It was still necessary to "load" the column before elution
of the anion could be observed.  The detection limit was not improved
with the 0.5% Carbowax 20M coating.
     Chromosorb W(HP) was used as the support in column 3.  Tailing of
the sulfate peak was reduced and the sensitivity of the analysis was
markedly improved.  The column exhibited fewer "loading" effects and
the detection limit was somewhat better (0.8 yg sulfate).
     A 1.0% SE-30 loading was used on Chromosorb W(HP) to determine the
effect of the liquid phase.  The tailing factor improved to 80 and the
retention time of the lower concentrated samples differed by only a few
seconds  from that of the more concentrated TMS-SO, standards.  The
column still required "loading" if a sample containing less than 0.5
yg SO, was to be analyzed.
     The final column support investigated was Chromosorb 750, which
has recently been placed on the market.  The support is reportedly
more inert than Chromosorb W(HP).  No peak tailing was observed as
indicated by a tailing factor of 100 (Table 1), and retention times
were identical regardless of sample concentration.  The sensitivity is
the greatest of all columns investigated and the calibration curve is
the most linear, whether based on peak area or peak height.
                                   19

-------
     In summary, these results suggest that the percent loading of
the liquid phase is the primary factor affecting detection limit,
peak tailing and linearity of the calibration curve.   Minimizing
the amount of SE-30 will reduce the TMS-SO. which is  adsorbed or
decomposed on the gc column.   Reducing the reactivity of the solid
supports appears to increase  the relative sensitivity of the analysis.
Under optimum conditions, a detection limit of 0.2 yg sulfate in
aqueous solution can be achieved using 1% SE-30 on Chromosorb 750.
This detection limit will have to be reduced by at least a factor of
10 if gc analysis of sulfate  is to be conveniently used for the
analysis of environmental samples.  Further assessment of this pro-
cedure should include the investigation of columns prepared with a
few-tenths percent of SE-30,  as well as other silicon oils, on sily-
lated glass beads or in a glass capillary column.  Using Chromosorb
750 with a larger mesh size may reduce the reactivity of the gc column
further without adversely affecting the gc separation of TMS-SO,.
                                 20

-------
                              SECTION 5
            A SPECTROKINETIC PROCEDURE FOR SULFATE ANALYSIS
PREVIOUS INVESTIGATIONS

     Hems, et al. (7) introduced a direct kinetic method for sulfate
determination based on the reactions presented in Figure 4.  Sulfate
increases the depolymerization of the aged zirconyl species in acidic
              i
media, and the rate of formation of the free zirconium ion-methyl thymol
blue chromophore is proportional to sulfate concentration.  Hems followed
the reaction of macro-volumes of sample at 586 nm and related the absor-
bance at a single 60 minute observation interval to sulfate concentra-
tion.
     This procedure has been modified to provide a true kinetic analysis
for use with a miniature centrifugal fast analyzer.  Such a technique
permits the relatively rapid, semi-automated analysis of micro volumes
of samples and standards, all of which can be analyzed simultaneously.
Sub-ppm detection limits are noted and, with appropriate sample pre-
treatmentj can be made relatively specific for sulfate ion.

 REAGENTS AND INSTRUMENTATION
     Stock methy thynol blue (MTB) is prepared by dissolving 0.0378 g
MTB in 100 ml water.  Although the basic form of this indicator is
relatively unstable, we found that this reagent is stable indefinitely
if it is acidified with one drop of 1M HC1.  The working MTB solution
is prepared by adding 1 ml of 6f1 HC1 to 3.45 ml of the stock MTB solu-
tion and diluting the mixture to 5 ml with water.
                                  21

-------
                   SO/,2"
ZROClo  POLYMER	a—>  FRE ZROC12             CD
         METHYLTHYMOL          BLUE CHROMOPHORE
                      - >  ,
             BUJE              (
           Figure 4.   ZR-MTB  Reaction
                      22

-------
     The zirconium polymer (ZR) is formed by dissolving 0.0805 g
ZrOCl2 (A. D. Mackay) in 25 ml of 0.0125 M HC1.  Two days are required
to adequately polymerize the zirconium.  This stock solution may then
be used for one week.  Fresh working zirconium polymer is prepared
daily by diluting 1 ml of the stock solution to 10 ml  with 0.05M.HC1.
     The instrumentation of the miniature fast analyzer and the auto-
matic rotor loading station are described elsewhere (22, 23).


EXPERIMENTAL  PROCEDURE AND DATA ANALYSIS
     A 17-place multicuvet rotor is used to simultaneously analyze a
combination of 16 sulfate standards and samples and a  water blank.  A
two-step procedure is used to load the ZR polymer, MTB, and sulfate
sample into the rotor.  The ORNL Sample-Reagent Loader is used to pre-
load 20 yl of the working ZR solution (+ 25 yl water)  into one of the
two reagent wells associated with each cuvet.  The ZR  polymer  is  then
transferred to each of the 16 cuvets by placing the rotor on the
Analyzer, accelerating it to 4000 rpm and allowing the rotor to coast
to rest.  The rotor is then carefully replaced on the  Loader and  20 pi
of sulfate standard or sample (+ 25 yl water) and 20 yl MTB working
solution (+ 25 yl water) are loaded into the two reagent wells.  This
two step loading scheme prevents premixing of the ZR polymer with
either the sulfate or the MTB prior to reaction initiation.
     The rotor and its contents are subsequently placed on the Analyzer
and brought to 30.0 j^ 0.2°C by radiant heating with an infrared lamp.
                                23

-------
The kinetic reaction is initiated when the sulfate and MTB are trans-
ferred into the cuvets containing the ZR by rapidly accelerating the
rotor to 4000 rpm.  The three reaction components are then thoroughly
mixed by instantly braking the rotor.  The rotor speed is then adjusted
from rest to 1000 rpm for the remainder of the reaction.
     After a 100-s delay period, 30 consecutive absorbance measurements
are taken at 30-s intervals.  Each measurement represents the average
of 25 consecutive passes of the appropriate cuvet past the stationary
phototube.  Figure 5 shows the oscilliscope trace of the optical trans-
mission of each cuvet at one such observation interval.  The percent
transmission (and thus absorbance) is reproducible within replicates
of the standards and samples.  However, contrary to Hems' report,
the absorbance measurement at a single reaction time was not found
to be linear with sulfate concentration.  Sequential absorbance
measurements are displayed in Figure 6 for five sulfate standards.
After an approximately 4-minute induction period, the absorbance
change for each standard is linear with time and proportional to
sulfate ion concentration.
     The kinetic data for all 17 cuvets are stored on a laboratory
computer.  A computer routine (Figure 7) has been written to retrieve
the data from computer memory, locate the linear portion of the reaction
data for each cuvet, and calculate the reaction rate.  The linear least
squares regression analysis subroutine locates the linear portion of
the data by computing the rate and linearity criterion (correlation
coefficient) for the first 6 data intervals of a given cuvet.  The
correlation coefficient computed is compared to a "minimum acceptable
                                 24

-------
                        ORNL-DWG.76-3435
      0%
%T
              0    S04
               STANDARDS
                   (ppm)
SAMPLES
   100%^
  Figure 5.  Optical Transmission of Sulfate Standards and Samples
         in Multicuvet Rotor
                     25

-------
                                               ORNL-DWG 76-15187
   0.70




   .0.63




   0.56
E  0.49
c
8  0.42
B  0.35
z


I  0.28

O
(fl

9  0.21
   0.14




   0.07



      0
+ 20
     CM*
        0     220     370    520    670     820     970

                            TIME (sec)


       Figure 6.   Reaction Progress of Sulfate Ion Standards
                              26

-------
                                                           ORNL-DWG 78-15653

C-FOCAL 69-GEMSAEC 6/81/71

91.01 C LINEAR SEARCH ROUTINE
HI.03 E
91.93
01.04
01.05
01.06
01.07
01.08
01.09
01.10
  "NlMBER OF SETS OF OBSERVATIONS". * I
  "DELAY INTERVAL".01
  "OBSERVATION 1NTE*VAL"«SE
  "RLM NUMBER".RN
  "GEMSAEC UN1T",UM
  "WHICH CUVET CONTAINS 8LANK?"»RC
!  "NINIHiH ACCEPTABLE CORRELATION COEFF. ?",CX
!  "NJMBE* OF COVETS UNDER CONSIDERATION",NK« I
03.01 S 1D*FITRm/2>
02.07 T »."CJVET      INTERVAL  *l0»0.S)-l >*50
03.01 F J*RC+I,I»NKI S  1*11 S '1*01 D 10
03.99 3UIT
07.10 S S3MNC*I )*IS S2*F6ET-SI
07.30 S SN»FGETs.4343*FLOG(S2/CSN*FSTR(ST+J>»
07.60 I (1-RO7. 7. 7.99
07.70 S SB»FGET-Sl
07.75 S AD=.4343»FLOG»
07.80 S A9(J)«AB(J)-AD
07.99 R

10.01 S SX»0I  S SY'9l S  XY«0I S  X2=9l  S Y2"0
10.05 S H*I+XD-I
10.06 F N=I.l>Hl  0 71 D  II
10.07 S M=«*l
10.10 S MM*XY-SX»SY)/FS3T<«NO)»X2-SX*SX)*-NI>I0. 91*10. 011  D  12
10.40 R

11.05 S SY*SY+AB(J>
11.10 S X*DI*( (M-l )*SE»
11.13 S SX»SX»X
I I. IS S XY*XY+X«AB
11.30 R

12.01 SMAX>MM(l)lS   CC*COCI>IS   IN«I
12.05 F M«2.1» (NI-1 )l  D 13
12.08 S TO**SE
12.30 I (CC-CXM2. 35.12.5,12. 5
12.35 I CMI-I )-IN)l2. 4,12.45

                     t2'91
 12.50  T I.X4.0.J,"    ".TO," - ".TF,X20.06.MAX.CCIR

 13.01  I > 13-1.13.1 .13.2
 13.10  S MAX-MM(«>I S CC«CO(H)I S
 13.20  R
 *

            Figure 7.   Linear Search Computer Subroutine
                                    27

-------
correlation coefficient" selected by the operator.  If the computed
value is less than this criterion, the slope is set to zero; otherwise,
the computed value is retained.  The interval used, slope, and correla-
tion coefficient are stored in computer memory.  The process is re-
peated for intervals 2-18, 3-19, ..., 14-30.  The list of computed
slopes is searched for the maximum rate, and the result is presented
on a teletype.  If none of the intervals meet the linearity criterion
selected by the operator, the teletype flags the result by typing
"DATA DO NOT MEET LINEARITY CRITERION."  The computer program
then proceeds to calculate the reaction rate of the next cuvet.
     The format of the program output is illustrated in Figure 8.  The
value of the minimum acceptable correlation coefficient is typically
selected between the values 0.9990 to 0.9995; this permits discrimina-
tion from the initial rapid, but relatively nonlinear, reaction "induc-
tion" phase (see Figure 6).  This initial sulfate-independent reaction
phase is presumably due to the presence of residual free Zr in the
polymeric reagent, and its duration is a function of the age (hence
extent of polymerization) of this reagent.  For a given polymeric
reagent, it is found that the initiation of the sulfate-dependent
reaction phase occurs at shorter time for higher sulfate concentration
in the sample.  The interval selected by the linear search program for
compution of reaction rate is indicated in Figure 6 by the joined
data points, and illustrates the versatility of the linear search sub-
routine.   By its use, a single set of observation conditions may be used
for a wide range of sulfate ion concentration and degree of zirconyl
polymerization in the reagent, since the computer automatically selects
the data  interval  used for computation of sulfate-dependent rate.
                                 28

-------
                                                            ORNL-DWG. 76-14949
             RUN NUMBER:  3*
             NUMBER OF CUVETS  UNDER CONSIDERATIONt  17*
             GEMSAEC UNIT:  17.2*
             DELAY INTERVAL:  100*
             OBSERVATION  INTERVAL: 30*
             NUMBER OF SETS OF OBSERVATIONS: 30»
             MINIMUM ACCEPTABLE CORRELATION COEFFICIENT:
0.9993*
ro
CUVET
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
INTERVAL
280 -
280 -
250 -
250 -
220 -
220 -
220 -
220 -
190 -
190 -
250 -
250 -
250 -
220 -
220 -
220 -
CSEC)
730
730
700
700
670
670
670
670
640
640
700
700
700
670
670
670
S0j~ Cone.
0 ppm
99
5 ppm
99
10 ppm
IB ppm
99
20 ppm
99

Sample. «1


Sample. #?.

RATE  (A/MIM)
_/ 0.020339
   0.020260
_{ 0.024149
   0.024459
_/ 0.028198
   0.028117
./ 0.031905
   0.032066
 r0.036377
   0.036486
   10.024886
   0.024344
   0.024938
 (0.03848B
-j 0.038599
 V0.037926
                 CC
             0.999620
             0.999530
             0.999428
             0.999391
             0.999355
             0.999367
             0.999488
             0.999407
             0.999476
             0.999581
             0*999544
             0.999585
             0.999488
             0.999478
             0.999571
             0.999471
                  Figure 8.   Data Output Format for Linear Search Computer Subroutine

-------
     The results of the computer output are plotted in Figure 9.  The
method is linear over a range of 0-20 ppm (0-400 ng) sulfate.  The
slope of the calibration curve is 0.010 absorbance units/min/ppm
sulfate.  The detection limit, equivalent to twice the standard
deviation of the lowest standard divided by the slope, is 0.3 ppm
sulfate.

OPTIMIZATION OF REACTION CONDITIONS
     The reaction concentration of the Zr polymer was found to affect
both the linearity and sensitivity of the sulfate calibration curve.
                                                 -4
At a zirconium reaction concentration of 1.1 x 10   M, the reaction
blank exhibits little change in absorbance (Figure 10).  As the zir-
conium reaction concentration increases, the reaction rate of the
reagent blank becomes significant and the sensitivity of the sulfate
calibration curve increases.  Above 1.8 x 10"  M, zirconium, the cali-
bration curve is no longer linear.  A reaction concentration of 1.5 x
  -4
10   hi zirconium is presently used for sulfate analysis.
     The age of the stock ZR solution also affects the sensitivity of
the analysis.  As the zirconium ages, it becomes more polymerized and
subsequently longer times are required to form free zirconium during
the reaction.  The sensitivity decreases by approximately 30% as the
stock ZR solution ages over a period of a week.  The working Zr
polymer exhibits no significant influence on sensitivity as it
ages during the day.
                                30

-------
 0.036
 0.032
 c

 6
\

<

UJ


a:

z
g
H-
o

UJ
oc
 0.028
0.024
                               ORNL-DWG 76-14950
0.020
                 5         10        15        20

           SULFATE   ION  CONCENTRATION (ppm)

         Figure 9.  Sulfate Ion Calibration Curve
                       31

-------
                                    ORNL  DWG. 75-8717
      0.04
  (0
  00
  in
 yj
 H
0.03
     0.02
                               [Zr] x I04m
Figure 10.
       2   4  6   8  10  12  14  16   18

     SULFATE  CONCENTRATION  (ppm)

     Effect of ZR Polymer Concentration on Sulfate  Ion
     Reaction Rate
                         32

-------
      The concentration  of  HC1  in  the  zirconium  reagent  and  in the
 final zirconium-methyl thymol  blue (Zr-MTB)  reaction mixture determines
 the degree of polymerization  of the zirconyl species.   Consequently,
 the reaction acidity  also  affects the sensitivity and linearity of the
 kinetic data for the  sulfate  analysis.  Optimum acid concentration was
 determined for the 0.05-0.36  M HC1 concentration range.  The sensitivit
 of the Zr-MTB procedure approaches zero outside the 0.1-0.3 M HC1
 range.  At low HC1 concentration  the  absorbance of the reagent blank
increases, as do both  the reaction rate and sensitivity of the analysis.
The reaction rate of a given sample is more linear for higher HC1  concen
trations.  An optimum  concentation of  0.2 M HC1  was selected to obtain a
modest sensitivity with  a minimal  reaction time and an adequate repro-
ducibility for the reaction rate of a  given sample.
     The sensitivity of  the reaction was also studied in the presence
of several organic solvents.   Previous techniques using methylthymol
blue dye as an indicator have  employed the addition of organics to
increase the absorption  coefficient of MTB, thereby improving the sensi-
tivity of the procedure. We  examined  the  effect of formaldehyde,
ethanol and methanol on  the kinetic procedure for sulfate ion determina-
tion.  Formaldehyde was  selected because it  is a common masking agent
for sulfite, and therefore, its presence might also make the Zr-MTB
method selective for sulfate.  However, calibration curves prepared
in the presence of 5%  formaldehyde were not  proportional to sulfate
concentration.  This response  remained non-linear even when the MTB
concentration was varied.   Ethanol  and methanol  are the most common
                                  33

-------
organic solvents used to improve sensitivity.   Reaction concentrations
of 5-50% of these solvents served only to decrease the sensitivity
of the Zr-MTB method, although the calibration curves remained propor-
tional to sulfate concentration.  Presently, a reaction concentration
of 0.52 M MTB is used in the absence of any organic solvents.

REACTION  INTERFERENCES
      Table  II summarizes the major cation and anion interferences in
the Zr-MTB  reaction  reported by Hems, et al.  They proposed the use
of Amberlite IR-120  (H) for the removal of cation interferents.  Because
acidity influences the sensitivity of the Zr-MTB reaction, the sodium
form  of the resin is presently used.  Batch-wise ion exchange with
approximately 200 mg resin was found to remove cation interferences
from  2 ml of sulfate sample.  Results from a study comparing sul-
fate  determination in the presence and absence of ionic interferences
are presented in Table III.  Sulfate analysis using the Zr-MTB pro-
cedure was compared with results obtained using a Technicon Autoanalyzer.
The Autoanalyzer system removes cation interferences with a flow-through
cation exchange column and indirectly determines sulfate concentration
colorimetrically by reacting excess barium with methyl thymol blue.
The untreated samples in Table III all exhibit a positive bias compared
to the analysis performed with the Autoanalyzer.  Samples in which cations
have been removed and analyzed with the Zr-MTB procedure are in much
closer agreement with those results obtained with the Autoanalyzer.
Recovery of added sulfate in these treated samples averaged 102% using
the kinetic procedure.
                                 34

-------
                      TABLE 2
   INTERFERING  IONS AND METHODS FOR THEIR REMOVAL
       Cations
Bi, Ce, Fe, Mn, Sb,
Se, Sn, Sr, Te, Th, V
        Removal
cation-exchange
on Amber! He IR-120 (Na)
       An ions
S03~
Oxalate, tartrate
                             mask with 30 ppm Al+
                             centrifuge with 30 ppm Hg
                             adsorb with MgO
                         2+
                        35

-------
                 TABLE 3



COMPARISON OF TREATED AND UNTREATED SAMPLES
Sample
1
2
3
4
Zr-MTB Procedure
Untreated
8.9
4.7
5.2
9.3
Cations
Removed
8.1
2.9
3.0
7.3
Cations &
An ions
Removed
7.7
2.7
2.9
7.6
Technicon
Autoanalyzer
10.4
2.8
2.8
6.7
                     36

-------
     Anions, including fluoride,  phosphate, and arsenate, interfere in
Hems' method at even trace  levels.  A 1 ppm concentration of fluoride
or arsenate has been found  to  interfere in the kinetic Zr-MTB procedure
by producing reaction rates comparable to 15 and 2 ppm sulfate,
respectively.  Phosphate varies in its effect from 0-9 ppm sulfate,
depending on the age of the zirconyl polymer.  Hems, et al.  suggested
the use of magnesium oxide  (24, 25) to remove the above anions from
sulfate samples.  We have investigated several masking agents which
might be added to the reaction mixture to remove the anions  and thus
eliminate lengthy pretreatment of samples.  Boric acid, La(III), or
Al(III) were added  to sulfate  samples in an attempt to mask  up to 2 ppm
concentrations of the anions.  Neither boric acid nor La(III) decreases
the interference of any of  the three anions.  However, 2 ppm fluoride
can be masked if at least 30 ppm  Al(III) is present in the sample.
Because this quantity of Al(III)  slightly increases the reaction rate
of sulfate, a similar concentration of the masking agent should be
added to standards  for accurate sulfate analysis.
     Sulfite and sulfide also  produce a positive interference in the
ZR-MTB procedure.   The effect  of  these two anions must be removed if
the procedure is to be selective  for sulfate anion.  The interference
from sulfite and sulfide is not due to these anions fier se_,  but was
found to be the result of anion auto-oxidation to sulfate prior to
sample analysis.  Mercuric'ion was used to stabilize the anions.  A
30 ppm concentration of the cation was added to samples containing
5 ppm sulfide and 10 ppm sulfate. The presence of the resulting HgS
precipitate and color formation did not interfere in the  kinetic analys
                                   37

-------
of sulfate ion in the samples, since the precipitate is centrifuged
out of the light path during the induction period of the reaction.
Mercuric  ion at this concentration level is sufficient to remove sul-
fite  interference if the samples are analyzed within 1-2 days  after
the masking agent is added.  If samples are to be stored, the  HgS
precipitate should be removed by centrifugation before sample  storage
to prevent the masked sulfide from slowly oxidizing during the elapsed
time.
    Sulfite anion was not stabilized in the presence of 30 ppm Hg(II),
even  for  a relatively short period of time.  A thousand-fold increase
in mercuric ion concentration is used to prevent the oxidation of sul-
fite  in the West-Gaeke sulfite technique (18).  However, mercuric ion
concentrations of this magnitude would interfere with the Zr-MTB pro-
cedure.
    Table III also includes the analysis of the four samples in which
cations were removed by batch ion exchange and fluoride and sulfide
interferences were removed by adding 30 ppm each of Al(III) and Hg(II)
to the samples.  The results are generally somewhat lower than when
just the  cations are removed.  However, the analysis indicates that
the major interference in the water samples are primarily due to the
presence  of cations.   There is a discrepancy between results obtained
for sample #1  using the Zr-MTB and Autoanalyzer procedures.  The sample
was highly colored with organic material.  Because the Autoanalyzer
used a single absorbance measurement to determine sulfate concentration,
any absorbing  species will  also interfere with the analysis.  However,
                                   38

-------
since the kinetic Zr-MTB  procedure  is  based  on  a  change  in absorbance,
the presence of a constantly  absorbing background does not interfere
with the determination.   This  represents  a significant improvement
over Hems' single measurement  method.

PRECISION AND  ACCURACY OF RESULTS
    The precision and  accuracy of the  Zr-MTB procedure was evaluated
by analyzing rain water samples and comparing the results with those
obtained with  the Technicon Autoanalyzer.  A within-run precision of
better than 2%, as demonstrated in Figure 8, can be achieved  with the
present method if the  reagents are reproducibly loaded into the rotor.
The major source of error appears to be in the loading of the MTB,
which acts as  a wetting agent  and forms beads of the reagent  on the
outside of the pipet of the loader.  This beading is minimized if the
pipet exterior is cleaned with methanol prior to loading the  rotor.
    Run-to-run reproducibility was demonstrated by analyzing  a series
of standards and samples  in five runs  made in a single day.  Results
of the five runs are given in Table IV.  The standard deviations
observed, both within-run and  run-to-run, are typically lower than the
                                         p
detection limit of the method  (0.3 ppm S04~).
    The same four samples were analyzed on four separate days to
determine day-to-day reproducibility in sample analysis.  Results for
this study are given in Table  V.  Results for a given day were cal-
culated from the average  of 3-5 runs,  performed on that day.   Again,
the standard deviations for day-to-day sample analysis do not exceed
the detection  limit of the Zr-MTB procedure.
                                 39

-------
                   TABLE 4



RUN-TO-RUN REPRODUCIBILITY IN SAMPLE ANALYSIS
Sample #
1
2
3
4
Run 1
2.9 +_ 0.1
3.1 + 0.1
5.3 +_ 0.1
20.9 + 0.0
Run 2
2.6 + 0.4
3.0 + 0.3
5.3
21.0 + 0.2
Run 3
2.8 +0.1
3.1 + 0.1
5.3 +0.3
20.1 + 0.2
Run 4
2.3 + 0.1
2.7 + 0.1
5.1
21.4 + 0.2
Run 5
2.2
3.6 +_ 0.0
5.5 + 0.3
21.3 + 0.4
Average
2.6 +_ 0.3
3.1 + 0.3
5.3 + 0.1
29.9 + 0.5

-------
                     TABLE 5




DAY-TO-DAY REPRODUCIBILITY IN SAMPLE ANALYSIS
Sample #
Day #
1
2
3
4
Average
1
2.7 +0.3
2.9 + 0.7
2.6 + 0.3
3.2 + 0.4
2.8 + 0.3
2
2.9 + 0.2
3.1 + 0.3
3.1 + 0.3
3.4 ±0.3
3.2+0.2
3
5.6 +0.1
5.4 + 0.3
5.3 +0.2
5.5 + 0.3
5.4 +0.1
4
21.1 + 0.2
21.4 + 1.3
20.9 + 0.5
21.3 +_ 0.4
21.2 +0.2
                          41

-------
    The accuracy of the method was  estimated by comparing results
obtained by the Zr-MTB procedure and the Technicon Autoanalyzer.
Ten rain water samples and a double-blind control  sample were
analyzed in the study.  The samples were analyzed independently
by the Environmental Analysis Laboratory, Analytical Chemistry Division,
ORNL, using the Technicon system.  Table VI lists the results of the
correspondence study.  Except for sample #10, results from the two
techniques are in close agreement.   A linear regression analysis,
based on all sample data, calculated a slope of 1.01 +_ 0.02, a y
intercept of -0.15 +_ 0.5 ppm and a correlation coefficient of 0.998.
The standard error of the estimate was equivalent to the detection
limit.  The results of the two methods agreed within a confidence
limit greater than 99.5% for 9 degrees of freedom and students'
T = 46.
                                 42

-------
                        TABLE 6
   CORRESPONDENCE OF THE ZR-MTB WITH A REFERENCE METHOD
          FOR THE DETERMINATION OF SULFATE ION
                     S0£~ Found, ppm
                        Technicon       CFA
Sample
Number
1
2
3
4
5
6
7
8
9
10
11
Autoanalyzer
(X)
2.8
2.8
21.0
2.8
1.4
3.5
4.0
2.5
1.4
3.5
6.3
Kinetic
(Y)
2.8
3.1
21.2
3.1
1.6
3.3
4.2
2.2
1.0
2.7
5.8
Linear Least Squares Regression Analysis:
Slope = 1.01 +0.02
Y intercept = -0.15 + 0.47 ppm
Correlation coefficient = 0.9979
Std error of estimate = 0.35 ppm
Student's t = 46
                            43

-------
                             REFERENCES

 1.  American Public Health Association,  Standard Methods for the
     Examination of Hater and Haste Hater,  14 ed. (A.P.H.A..  New
     York, 1976), p. 493-98.

 2.  6. Toennies and B.  Bakay, Anal.  Chem.. 2|,  160 (1953).

 3.  F. Y. Borden, L. H. McCorwick, Soil  Sci. Soc.  Amer.  Proc.,
     1970, p. 705-6.

 4.  R. Jasinski and I.  Trachtenberg, Analyst.  Chem.,  44, p.  2373-76
     (1972).                                          ~

 5.  T. A. Kokina, M. N. Petrikova, I. P. Alimarin, J. Analyt.  Chem..
     USSR. 2g, p. 2003-4 (1971).

 6.  N. H. Furman (ed.), Scotts'  Standard Methods of Chemical  Analysis,
     Vol. I (D. Van Nostrand Co., Inc., Princeton,  New Jersey,  1966),
     p. 1011-15.

 7.  R. V. Hems, G. F. Kirkbright and T.  S. West, Talanta. 16,  p.  789-96
     (1969).                                              —

8a.  E. Bernt and H. Bergmeyer, "Inorganic  Peroxides"  in  Methods in
     Enzymatic Analysis, H. Bergmeyer,  Ed.  (Academic Press,  New York,
     1965), p. 633-35.

8b.  G. G. Guilbault, P. Brignac and M. Zimmer,  Anal.  Chem.,  40, p. 190-96
     (1968).                                                 —

 9.  D. T. Bostick and D. M. Hercules, Anal.  Chem.. 47, p. 447-52 (1975).

10.  Manual of Methods for Chemical Analysis  of Water and Waste Mater,
     USEPA, 1974, p. 201-14.

11.  W. H. Stroud, Soil  Sci.. |,  p. 333-42  (1920).

12.  L. R. McKenzie and  P. N. W.  Young, Analyst. 100.  p.  620-28 (1975).

13.  K. Grasshof, Kiel.  Meeresforsch. 20, p.  5-11 (1964).

14.  F. Chapeville and P. Fromageot, Biochem. Biopnys. Acta,  14, p. 415-
     20 (1954).                                              —

15.  P. W. Robbins, "Sulfate Activating Enzymes" in Methods  in
     Enzymology, Vol. 5, p. 964-76 (1962).

16.  R. K. Banerjee and  A. B. Roy, Mol. Pharmacol., 2, p. 56-86 (1966).


                                   44

-------
17.  A. B. Roy, Biochem. J.. 74, p. 49-56 (1960).

18.  P. W. West and G. C. Gaeke, Anal. Chem.. 28,  p.  1816-19  (1956).

19.  J. Dreyfuss and K. J. Monty, J. Biol. Chem..  238,  p.  1019-24
     (1963).                      	  =

20.  W. C. Butts, Anal. Lett.. 3, p. 29-34 (1970).

21.  D. R. Matthews, "The Gas Chromatographic Determination of Trace
     Anions in Aqueous Media," University of Tennessee  Dissertation,
     1972.

22.  C. A. Burtis, J. C. Mailen, W. F. Johnson, et al., Clin. Chem..
     J|,  p. 753-61 (1972).

23.  C. A. Burtis, W. F. Johnson, J. C. Mailen, et al., Clin. Chem..
     2|,  p. 895-903  (1973).

24.  R. McGillivary  and S. C. Woodger, Analyst, JL p.  611-20 (1966).

25.  B. Belcher, A.  D. Campbell, P. Gouverneur and A. M. G. Macdonald,
     J. Chem.  Soc..  1962. p. 3033-37.
                                   45

-------
                                    TECHNICAL REPORT DATA
                            (Please read Instructions on the reverse before completing)
 1. REPORT NO.
   EPA-600/2-79-095
                              2.
3. RECIPIENT'S ACCESSIOWNO.
 4. TITLE AND SUBTITLE
  DETERMINATION OF TRACE QUANTITIES OF SULFATE  ION
  New Approaches to  an  Old Problem
                                                            5. REPORT DATE
                                                              Ma.y 1979
6. PERFORMING ORGANIZATION CODE
 7. AUTHOR(S)

        D. Bostick
                                                            8. PERFORMING ORGANIZATION REPORT NO
 9. PERFORMING ORGANIZATION NAME AND ADDRESS

  Oak Ridge National Laboratory
  Oak Ridge, Tennessee   37830
                                                            10. PROGRAM ELEMENT NO.
  1AD712
11. CONTRACT/GRANT NO.

  Contract  No.  40-S10-75
 12. SPONSORING AGENCY NAME AND ADDRESS
  Environmental Sciences  Research Laboratory--RTF,  NC
 j Office of Research  and  Development
 ! U.S. Environmental  Protection Agency
  Research Triangle Park. NC 27711            	
13. TYPE OF REPORT AND PERIOD COVERED
  Final
14. SPONSORING AGENCY CODE
  EPA/600/09
 !S. SUPPLEMENTARY NOTES
 16. ABSTRACT
       Several analytical  methods for possible  use  in  measuring trace amounts  of water
  soluble sulfate anions were reviewed and evaluated.   Enzymatic sulfate  determination
  does not appear to  be a  viable approach until  the required enzymes can  be  obtained
  commercially.  Gas  chromatographic analysis of bis(trimethylsilyl}sulfate  was studied
  and, with further development, may be a selective method for the determination of
  sulfate as well as  other oxy-anions.  A direct kinetic method, based  on the  ability
  of sulfate to catalyze the depolymerization of zirconyl species, was  investigated.
  The method can detect 0.2-20 ppm sulfate and,  with a minimum of sample  preparation,
  is relatively selective.

17. KEY WORDS AND DOCUMENT ANALYSIS
a. DESCRIPTORS
Air Pollution Reviews
* Sul fates Evaluation
Chemical Analysis
* Enzymes
* Gas Chroma tography
* Depolymerization
13. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
b.lDENTIFIERS/OPEN ENDED TERMS






19. SECURITY CLASS (This Report/
UNCLASSIFIED
20USNCLASSIFCIEDSrrtop^
c. COSATI Field/Group
13B
07B
07D
06A
07C
05B
21. NO. OF PAGES
54
22. PRICE
EPA Form 2220-1 (9-73)
                                            46

-------