Anal. Chem. 1994,66, 2921-2925
H202 Determination by the Is- Method and by KMn04
Norman V. Kiassen,' David Marchington, and Heather C. E. McGowant
Institute for National Measurement Standards, Ionizing Radiation Standards, National Research Council
Canada, Ottawa K1A OR6 Canada
The analysis of aqueous H202 at concentrations as low as 1
pM is conveniently done by the 13- method, which is based on
the spectrophotometric determination of 13- formed when
Hz02 is added to a concentrated solution of I -. At 351 nm,
was measured to be 26 450 M-l cm-l. By contrast,
an apparent value of 25 800 M-' cm-l was determined from
a calibration of the 13- method against titration by perman-
ganate. The difference could only be partially accounted for
by the equilibrium between 1 3 ;
correction of -1% was required and was traced to a side
reaction between HzOz and the buffer normally used in the
13- method. A simple spectrophotometric procedure was
developed which improves the sensitivity of the permanganate
titration to 0.3 pM HzO2. Measurements of H202 using the
oxidation of ferrous ions (Fricke solution) and permanganate
titration differed by less than 1%.
12, and I-. A further
Schumb, Satterfield, and Wentworthl considered the
permanganate titration to be one of the most exact and reliable
analytical procedures for HzO2. The reaction follows the
2Mn0; + 5H20z + 6H' - 2Mn2+ + 8Hz0 + 50,
which was confirmed to -1 part in 5000 by Huckaba and
Keyes2 The sensitivity of the permanganate end point, when
observedvisually, is - 3 pM HzO2.l We now present a simple
way to improve the sensitivity to 0.3 pM H202 by titrating
to the visible end point and determining the excess KMn04
by optical absorption. However, this new method does not
remedy the problem that the analysis of <lo0 pM HzOz
requires several hundred milliliters of H202 if commercially
available 0.1000 N KMn04 is used for the titration.
The analysis of H202 at concentrations as low as 1 pM is
conveniently done by determining the yield of 13- formed
when H202 reacts with KI in a buffered solution containing
ammonium molybdate as a catalyst. In the radiation chemistry
literature, the 13- method is commonly referred to as the
Ghormley meth~d.~-~ The useful range of the 13- method
matches the concentrations of HzOz often encountered in
irradiated aqueous solutions, and only small volumes, e.g., 1
mL, are required for the analysis. We wish to use the 13-
Faculty of Science, Physics Dept., McMaster University, Hamilton, ON,
Canada. WES student at NRC, May-August 1993.
(1) Schumb, W. C.; Satterfield, C. N.; Wentworth, R. L. Hydrogen Peroxide;
Reinhold: New York, 1955.
(2) Huckaba, C. E.; Keyes, F. G. J. Am. Chem. Soc. 1948, 70, 1640.
(3) Hochanadel, C. J. J. Phys. Chem. 1952, 56, 587.
(4) Hochanadel, C. J. Radiat. Res. 1962, 17, 286.
(5) Allen, A. 0.; Davis, T. W.; Elmore, G.; Ghormley, J. A,; Haines, B. M.;
Hochanadel, C. J. Oak Ridge Natl. Lab. Publ. 1949, ORNL 130.
Published 1994 by the American Chemical Society
method to measure theconcentration of H2Oz in the irradiated
aqueous solutions which form part of a primary standard for
absorbed dose to water.6 For this purpose, an accuracy of
199% is needed. No calibration of the If- method existed at
this level of accuracy. Moreover, the molar absorptivity of
I< at the maximum of its absorption band, tmax(I3-), quoted
in reports of the 13-method are actually apparent, rather than
true, values and include the effect of an equilibrium constant
and a, hitherto unsuspected, side reaction involving the buffer
In this study, we calibrated the 13- method against
titration of HzO2 by permangate. Using our measured values
of tmax(I3-), the equilibrium constant K = [I,-]/([I -1 [Iz]),
and a reaction of H202 with the buffer, we have accounted
for the difference between emax(I3-) and its apparent value in
the I< method. The measurement of H202 by the oxidation
of ferrous ions to ferric ions was also calibrated against titration
Materials. Standard solutions of -0.1000 N KMn04 were
obtained from Aldrich Chemical Co. and J. T. Baker. Aristar
concentrated sulfuric acid from BDH was used. Ammonium
molybdate tetrahydrate was obtained from Aldrich Chemical
Co. (ACS reagent) and from Aesar, Johnson Matthey
(Puratronic grade, 99.999%). Anachemia stabilizer-free,
analyzed, 30% H202 was used to make up dilute solutions of
H202. KI (99.99+%), NaOH (99.99%), and potassium
hydrogen phthalate (KHP; 99.95-100.05%) were obtained
from Aldrich Chemical Co. Iodine (99.999%), in the form
of 60-mesh spheres, was obtained from Aesar, Johnson
Matthey. Sodium chloride (Ultrex) came from J. T. Baker
and ammonium iron(I1) sulfate hexahydrate (99.999%) came
from Aldrich Chemical Co.
Water was purified by passage through a charcoal filter,
a Millipore ROlO (reverse osmosis) unit, and a Millipore
Milli-Q UV unit, in that order. The purified water was stored
in quartz containers.
Glassware was cleaned in an ultrasonic cleaner using a
solution of Alcanox in hot water. The spectrophotometer
cuvettes were cleaned each day with concentrated sulfuric
acid or chromic acid cleaning solution. The absorbance with
water in the cuvette was checked frequently during each
experiment. If the absorbance increased, the cuvette was
recleaned. This happened occasionally with the 13- method
but rarely for KMn04 titrations.
(6) Klassen, N. V.; Ross, C. K. Radiat. Phys. Chem. 1991, 38, 95.
AnalytlcalChemlstty, Vol. 66, No. 18, September 15, 1994 2921
Procedures. Many variations in the procedure for per-
manganate titrations have been rep~rted.',~,~-l~
Brights showed that modest variations only affect the results
at the 0.2% level. We adopted the procedure recommended
by Schumb et a1.l Four millimeters of concentrated sulfuric
acid per 100 mL of H202 solution was slowly added down the
side of the vessel, while stirring. The solution was then cooled
to room temperature by briefly immersing the flask in a bath
of cold water. The titration was done with 0.1000 N (0.020 00
M) KMn04. The KMn04 was added at less than 0.5 mL
min-I from a dropper bottle to the stirred solution. An acid
solution, and slow addition of KMn04, are needed to avoid
the formation of manganese dioxide, which can catalytically
decompose H202. The bottle plus dropper was weighed to
0.001 g before and after the titration. There was no weight
loss due to evaporation during the titration. Up to 500 mL
of H202 solution was used per titration so that most titrations
used - 1 g of KMn04 solution. After a permanent light pink
was apparent in the stirred solution, the [MnOL] was
calculated from the absorbance at 525.0 nm in a 4-cm cuvette,
the blank being a water/sulfuric acid solution.
Solutions A and B for the I< method were prepared
according to the usual recipe.14 Solution A for the I< method
consisted of 33 g of KI, 1 g of NaOH, and 0.1 g of ammonium
molybdate tetrahydrate diluted to 500 mL with water. The
solution was stirred for - 10 min to dissolve the molybdate.
Solution A was kept in the dark to inhibit the oxidation of
I -. Solution B, an aqueous buffer, contained 10 g of KHP
per 500 mL. The pHs, as measured using a Sentron pH meter,
were as follows: solution A, 12.8; solution B, 4.03; equiweight
mixture of A and B, 4.86; equiweight mixture of A and B and
The I< method consists of mixing equal weights of A and
B, followed by addition of the H202 solution. The absorbance
of the resulting solution was measured at 351.0 nm in a 1-,
2-, or 4-cm cuvette, depending on the [H202], to get an
absorbance of -0.9. The blank absorbance was determined
by subtracting the absorbance of water from the absorbance
of an equiweight mixture of A and B, assuming that the
difference was due to adventitious 13-, and calculating the
expected absorbance for dilution of the A/B mixture with
pure water. A measurement of 100 pM H202 by the 1 ;
method required only 1 mL of the H202 solution.
In order to measure emax(13-) at 351 nm, -0.002 30 g of
I2 was weighed, using a regularly maintained, semimicro-
balance, and dissolved in 300 mL of a solution of equal weights
of A, B, and water. Stirred in the dark, dissolution of the I2
took 10-25 min, after which the absorbance was measured.
The blank consisted of an aliquot of solution treated in exactly
the same way except that no 12 was added to it.
(7) Kolthoff, I. M. Quantitative Chemical Anulysis, 4th ed.; Macmillan: London,
(8) Fowler, R. M.; Bright, H. A. J. Res. Null. Bur. Std. 1935, I S , 493.
(9) Bellinger, F.; Friedman, H. B.; Eastes, J. W.;
Edmonds, S. M. Ind. Eng. Chem.
(IO) Reichert, J. S.; McNeight, S. A.; Rudel, H. W. Ind. Eng. Chem. Anal. Ed.
1939, Ii, 194.
(1 I) Duval, C. Chim. Anal. 1953. 35, 265.
(12) Vogefs Texfbook of Quanrifatiue Chemical Analysis, 5th ed.; Jeffery, G. H.,
Bassett, J., Mendham, J., Denney, R. C., Eds.; Longman Scientific and
Technical, Harlow, England, and Wiley, New York, 1989.
(13) Allen, N. Ind. Eng. Chem. 1930, 2, 55.
(14) Allen, A. 0.;
Hochanadel, C. J.; Ghormley, J. A,; Davis, T. W. J. Phys. Chem.
1952, 56. 575.
For the measurement of H202 using a ferrous ammonium
sulfate solution (Fricke solution), an aliquot of H202 solution
was added to an equal volume of Fricke solution which had
been made up to double the usual concentration. In this way,
the final solution had the same concentrations as the usual
Fricke dosimeter ( l P 3 M ferrous ammonium sulfate,
NaCl, and 0.4 M sulfuric acid).
The most convenient way to carry out the analysis of H202
to the accuracy and precision we desired was to weigh the
solutions to bemixed. For this reason, the densities of solutions
were determined by assuming water to have a density of 0.997
g mL-l at room temperature (21.8-22.8 "C). A few of the
commonly used densities (g mL-') were as follows: 0,1000 N
KMn04, 1.002; concentrated sulfuric acid, 1.84; dilute
sulfuric acid (4 mL added to 100 mL of water),
0.997 + [(0.0404)(wt % acid/6.38)]; solution A of the 1 ;
method, 1.054; solution B of the 13- method, 1.014; solution
of equal weights of A, B, and H20, 1.017; le3 M H202,
0.997; and Fricke solution, 1.023. Care was taken when
weighing flasks which contained magnetic stir bars to add a
spacer between the balance pan and the flask to eliminate
interference by the magnet.
Optical Measurements. A Cary 2 10 spectrophotometer,
which measures to 0.0001 absorbance unit, was used. The
wavelength was calibrated to 0.02 nm using the 485.99-nm
emission line of the deuterium lamp. Spectra were measured
using bandpasses of both 0.25 and 1.0 nm, a 0.25-s time
constant, and an appropriately slow wavelength sweep. In all
cases, identical values for molar absorptivity were found using
either bandpass, and routine measurements were made using
1 .O nm, which was less noisy. The spectra shown in this report
have had the blank subtracted. Absorbance measurements
were made -5 min after placing the cuvette into the
spectrophotometer, in order to let solutions reach the spec-
trophotometer temperature. All absorbance measurements
were made within the range 23.9-24.9 "C.
RESULTS AND DISCUSSION
KMn04 Method. We found, as stated by Schumb et al.,l
that visual detection of excess MnOL by its pink color limited
the sensitivity of the method to -3 pM H202. We were able
to extend the sensitivity to 0.3 pM HzO2 and, as well, avoid
the task of adding just the right amount of KMn04. Our
method consists of adding KMn04 to a slight excess and
measuring the absorbance of the slightly pink solution at Amax
to determine the excess concentration of KMn04. In order
to do this, it was necessary to (i) confirm Amax, (ii) measure
the molar absorptivity at A,,,,,,
law is obeyed. Other, less desirable, methods which have
been suggested in the past are to visually color match the
excess KMn04 to standard solutions of KMn04, to determine
the excess iodometrically, or to measure H202 by the decrease
it produces in the absorbance of a KMn04 solution at 525
KMnOd Spectrum and Molar Absorptivity. Solutions of
3.5 X lo4 M KMn04, made by diluting 0.0200 M KMn04
with either water or a solution of 4 mL of concentrated
H2SO4 per 100 mL of water, gave the same absorption spectra
(Figure 1) with A,,,,,
= 525.3 f 0.2 nm. The spectrum is
similar to that reported by Waterbury and Martinis and very
similar to the better resolved spectrum shown on page 7 13 of
and (iii) confirm that Beer's
2922 AnalyticalChemistry, Vol. 66, No. 18, September 15, 1994
Flgure 1. Optical absorption spectrum of 3.5 X lo4 M KMn0, In
water or sulfuric acid/water using a l t m cuvette and a 0.25-nm
ref 12. The molar absorptivity of KMn04 was determined
from the slope of absorbance vs [KMnOd] for solutions of
KMn04 added to water and to water/HzS04 (three trials
each). In each trial, measurements were made at about 1,3,
10, 20, and 45 pM KMn04. A molar absorptivity of 2457
M-l cm-l was determined for the water solution and 2444
M-1 cm-1 for the water/HzS04 solution. The combined results
led to a value of 2450(f22) M-' cm-l for tmax(Mn04). The
same value of tmax(Mn04-) was found for both 0.25- and 1 .O-
nm bandpasses, and Beer's law was obeyed to the highest
[KMn04] used. Straight line plots of absorbancevs [KMnOd]
intercepted zero absorbance at the origin for water but at 0.6
f 0.1 pM KMn04 for acidified water. Evidently, the H2SO4
added to the solution consumed -0.6 pM KMn04. A
correction of 0.6 pM KMn04 was applied in the present study.
The 1% uncertainty in our value of emax(Mn04-) was minor
because emax(Mn04-) was only used to make a correction of
- 5% to the titration. Waterbury and Martin15 reported emax-
= 2240 M-l cm-l. The lower resolution of their
spectrum might account for their value being 9% lower than
ours. Three separate lots of standard KMn04 from two sources
were compared by the titration of a solution of H202 (a total
of 12 titrations). The standard deviation of the measured
values of [H202] was 0.5% with no dependence on the source
of the KMn04 solution. Hence, we have accepted the
concentrations of standard KMn04 as specified by the
suppliers. The method described here can detect changes of
0.3 pM H202.
During this study, many KMn04 titrations were done in
the normal way, making use only of the visible end point. A
fine dropper was used from which w 1 / 4 of a drop could be
left on the inside of the flask to be picked up by the peroxide
solution. Done in this way, five trials involving H202
concentrations above 200 pM showed an average standard
deviation of 0.3%, whereas an average standard deviation of
1.3% was achieved for 13 trials involving HzOz concentrations
from 41 to 200 pM. The precision possible using our new
method, in which the excess KMn04 is measured spectro-
photometrically, was assessed for a 90 pM solution of H202.
Six aliquots of 500 mL were titrated. The standard deviation
(15) Waterbury, G. R.; Martin, D. S. J. Am. Chem. Soc. 1953, 75, 4162.
Flgurr 2. Optical absorption spectrum of 2.6 X
cwette and a 0.25-nm bandpass.
IJ- uslng a l-cm
was 0.6%. At H202 concentrations lower than 100 pM, the
blank correction (see above) is nonnegligible and will differ
from one laboratory to another.
13- Method. The reaction of HzO2 with acidified KI to
produce 12, followed by titration of the 12 with standard sodium
thiosulfate, is called Kingzett's method.1°J6J7 At high
concentrations of KI, the equilibrium between 12, I -, and
13- is strongly in favor of 13-. The measurement of 13- by its
optical absorption forms the basis of the Is- method.
H202 + 21- + 2H' - I, + 2H20
1, + I- + 1,-
The large value of tmx(13-) makes the Is- method conve-
nient at less than 1 0 pM HzOz. By using a 10-cm cuvette and
a procedure sensitive to a change in absorbance of 0.0002,
Bielski and Allen18 could measure changes in [HzOz] of
M. The reaction between H202 and I -to form 12 is slow but
is accelerated by the catalyst, ammonium molybdate.5J9-23
Ovenston and ReesZ4 showed that an approximately neutral
solution is needed for accurate measurements using the Is-
method, because acid solutions showed an increase in absor-
bance with time.
The Is- method is less sensitive to the presence of organic
impurities than the KMn04 titration. The solutions needed
for the 13- method are simple to prepare, they require no
standardization, and they last for months if solution A is kept
in the dark.
Z3- Spectrum and tmpx(Z3-).
in Figure 2, was measured for a solution of 26 pM 12 in the
mixture of A, B, and water used in the 13- method. The peak
is broad and A ,
= 35 1 f 0.5 nm. The spectrum was identical
for 0.25- and 1 .O-nm bandpasses. Other reported values for
Xmax range from 350 to 354 nm.25-29
The spectrum of 13-, shown
(16) Scott, W . W.
N. H., Ed.; Van Nostrand New York, 1939; Vol. 2, p 2180.
In Standard Methods of Chemical Analysis, 5th ed.; Furman,
(17) Kingzctt, C. T. J. Chem. Soc. 1880, 37,792.
(18) Bielski, B. H. J.; Allen, A. 0. In?. J. Radiar. Phys. Chem. 1969, 1, 153.
(19) Brode, J. Z. Phys. Chem. 1901.37, 257.
(20) Kolthoff, I. M . Z. AMI. Chem. 1921,60, 400.
(21) Rothmund, V.; Burgstaller, A. Monarsh. Chem. 1913, 31, 693.
(22) Deleted in proof.
(23) Ramana Rao, D. V. J. Sci. Ind. Res. 1956, 158, 668.
(24) Ovenston, T. C. J.; Res, W . T. A ~ l y ~ r
1950, 75, 204.
Analytcai Chemistw, Voi. 68, No. 18, September 15, 1994
The value of emax(13-) was measured at 24.4 f 0.5 OC.
Three determinations led to an apparent value of 26 120 f
20 M-I cm-l for cmax(Ij-) in the solution used in the 1 ;
method. We found no dependence of emax(If) on the
presence or absence of ammonium molybdate in the solution
or the use of 0.25- or 1.0-nm bandpasses. Using our
determination of 600 (or the more commonly accepted value
of 714) for the equilibrium constant, K = [Ij-]/([I-] [I2]), we
calculate that If represented 98.76% (98.95%) of the sum of
I2 plus 1 ; . We calculate a value of 26 450 M-1 cm-I for the
true emax(IG) using K = 600 (26 400 when using K = 714).
Thesevaluesagree with thevalueof 26 400reported by Awtrey
and C ~ n n i c k ~ ~ and by Daniele.29 Hence, we conclude that
is little affected by the KHP present in the If
method. Awtrey and Connick claimed an uncertainty of 1%
for their value of tmax(13-). They also reported that 412) at
351 nm is - 16 M-l cm-l and t(I-) is much less. Hence, both
are insignificant insofar as the 13- method is concerned.
Equilibrium Constant. A measurement was made of the
equilibrium constant, K = [13-]/([1-] [I2]). The preponder-
ance of reported values of K for low concentrations of 12 in
water with no added salts puts K in the range 650-760.28-35
No measurement of K has been reported for the solution used
in the 13- method. We measured K by preparing a stock
solution of 12 in a solution of KHP. This was then mixed with
solutions of A containing various concentrations of KI so that
the ratio [13-]/( + [I3-]), at equilibrium, ranged from 0.4
to greater than 0.99. The [I21 in the stock solution was first
estimated by assuming complete conversion of I2 into 13- in
the most concentrated solutions of I -. Then, K was calculated
for less concentrated solutions of I -. The calculation was
then iterated. The value of K determined in this way was 600
f 30 M-l. When the usual recipe for the If method is used,
the final [KI] is 0.1325 M, which means that the 13- produced
is 1.24% less than the H202 added. For K = 714 (the most
generally accepted value), the If should be 1.05% less than
the H202. Under our conditions, the formation of 162- or
1 ; is in~ignificant.~~
Calibration of the 1; Method by KMn04 Titration. In
two trials, 250 and 1000 pM H202 solutions were measured
by KMn04 titration and by the 13- method. Agreement
between the KMn04 titration and the 13- method required a
value of 25 800 M-1 cm-1 for the apparent value of
emaX(I3-). Note that use of the apparent value of emax(I3-)
assumes the quantitative conversion of H202 into I<. This
value of 25 800 M-l cm-1 agrees with the values of 25 700
M-1 cm-1 by Buxton and Sellers37 and 25 900 M-' cm-I by
Anderson and Hart,38 where the authors used three significant
figures, suggesting that some sort of unspecified calibration
(25) Awtrey, A. D.; Connick, R. E. J. Am. Chem. SOC. 1951, 73, 1842.
(26) Okada, T.; Hata, J. Mol. Phys. 1981, 43, 1151.
(27) Job, P. Ann. Chim. 1928, 9, 8.
(28) Gutman, V. Hulogen Chemistry; Academic Press: London, 1967.
(29) Daniele, G. Gazz. Chim. Itul. 1960, 90, 1068.
(30) LaMer, V. K.; Lewinsohn, M. H. J. Phys. Chem. 1934, 38, 171.
(31) Davies, M.; Gwynne, E. J. Am. Chem. SOC. 1952, 74, 2748.
(32) Katzin, L. I.; Gebert, E. J. Am. Chem. SOC. 1955, 77, 5814.
(33) Jones, G.; Kaplan, 8. B. J. Am. Chem. Soe. 1928, 50, 1845.
(34) Bray, W. C.; MacKay, G. M. J. Am. Chem. Soc. 1910, 32, 914.
(35) Rengevich, E. H.; Shilov, E. A. Ukr. Khim. Zh. 1962, 28, 1080.
(36) Ramette, R. W.; Sandford, R. W., Jr. J. Am. Chem. SOC. 1965, 87, 5001.
(37) Buxton, G. V.; Sellers, R. M. J. Chem. Soc., Furuday Trans. I 1985,81,449.
(38) Anderson, A. R.; Hart, E. J. J. Phys. Chem. 1962, 66, 70.
had been done. Hence, the apparent value of tmax(1;)
appropriate to the If method is 1.2% less than the apparent
value of 26 120 M-' cm-I, which we determined by dissolving
12 in a solutionof A, B, and water (see above). For this reason,
we investigated the possibility of a competing reaction which
consumes 1.2% of the H202 in the 13- method.
Kinetics. A possible explanation for loss of peroxide in the
If method emerged when it was observed that, if H202 were
mixed with B before adding A a minute or two later, the yield
of 13- was substantially lower than if H202 were added to a
previously mixed solution of A and B. The extent of the
reduction ranged from 5 to 33%. The average, and median,
reduction was 25%. Out of 20 trials (an average of three
measurements per trial), 18 gave a reduction in the range
17-3376 and 11 fell into the range 20-30%. The reaction
which causes this reduction is unknown. Solution B was needed
to cause the reduction. Kinetic measurements revealed that
the magnitude of the reduction increased with an increase in
the delay in adding A but became constant for delays exceeding
2 min. The reason for the constant reduction after 2 min is
not understood. To achieve half the maximum reduction,
H202 had to remain in contact with B for 30-90 s before
adding A. The calculations below are based on a half-life of
60 s for the reaction of H202 with KHP (or an impurity in
solution B) and a maximum loss of HzO2 of 25%.
An estimate of the rate of conversion of H202 into Ij- was
made by rapidly mixing a solution of H202 with a mixture of
A and B already in a cuvette. Absorbance readings at 351
nm on the chart recorder began at 5 s. A semilogarithmic
plot of ((final absorbance) - (absorbance at time t)) vs t was
a straight line, giving a half-life of -2.5 s for the catalyzed
production of I2 from H202 and I -. The equilibrium between
12, I -, and 13- was found to be established much too quickly
to be measured by our crude method, in agreement with other
report^.^^,^^ Using a solution of A which did not contain any
ammonium molybdate, the uncatalyzed production of 12 was
found to be pseudo first order with a half-life of 8 min.
The relative half-lives for the reactions of H202 with KHP
and with I - mean that 4% of the H202 reacts with solution
B in some manner when H202 is added to premixed A and
B. Combining this with the average maximum reduction in
If of 25% when H202 is first added to solution B (see above)
results in a reduction of 1.0% in the 13- yield when the 13-
method is used in the normal way. Using this percent reduction
and the values of tmax(I<) and K determined in this study, we
are able to account for, within 0.296, the apparent value of
emax(If) appropriate for the 13- method.
Values of Apparent tmax(l;). Hochande13 determined
tmax(If) to be 25 840 M- cm-I based on a calibration using
H202 solutions which had been standardized with ceric sulfate
solution. His final solutions contained 0.0994 M KI. His
value for emaX(I3-), and the values of 25 700 M-l cm-I
reported by Buxton and Sellers37 and 25 800 M-' cm-l reported
by Schwarz and Biel~ki~~ are consistent with the 1% effect of
the buffer found in the present study. Anderson and Hart38
reported a value of 25 900 M-1 cm-l for, apparently, the same
[KI] as used by Hochanadel. Their value would indicate only
a 0.4% effect by the buffer.
(39) Schwarz, H. A,; Bielski, B. H. J. Phys. Chem. 1986, 90, 1445.
(40) Turner, D. H.; Flynn, G. W.; Sutin, N.; Beitz, J. V. J. Am. Chem. Soc. 1972,
2924 AnalflicalChemistty, Vol. 66, No. 18, September 15, 1994
Table 1. Estlmate of the Standard Uncertainty In the
Mearurement of 300 pM H202 Udng the I;
the Callbratlon by KMnO,
std Uncertainty in [HzOz],
concn of KMn04
weighing H 2 0 2
0.6 pM blank corr
absorbance(Mn04-) / c,,,(MnO()
e m & - )
(includes K )
absorbance(I,-) (includes temp)
reaction with buffer
combined standard uncertainty
The independence of error sources is assumed.
Schwarz and Salzman41 and Bielski and Allen18 used an
A solution which contained only 20% as much KI as normal.
We measured a half-life of 7.5 s for the appearance of 13-
using their [KI]. Hence, the reaction of H202 with solution
B using the Schwarz and Salzman recipe should lead to a
2.8% reduction in 13-. This 2.8% reduction and K = 600 leads
to a value of 24 220 M-l cm-l for the apparent value of
emax(13-) appropriate to their solution, a value which is 0.6%
greater than the average value of 24 045 given by Schwarz
and Salzman41 and Bielski and Allen.18 This indicates that
the reaction of H202 with the buffer solution occurred in those
studies but the value the authors used for the apparent
tmax(13-) gave the [H202] correct to 1%.
We conclude that the literature values of apparent
fmax(13-) that seem to have the best precision support the
occurrence of a loss of 13- due to a reaction of the buffer
solution of about the magnitude measured in the present study.
Temperature Coefficient of the I< Method. The tem-
perature coefficient of the 13- method will be due to the
combined temperature coefficients of the equilibrium constant
and of fmax(13-). A 2.7% increase in K per degree drop in
temperature has been rep~rted,~’.~~.~~
increase in absorbance per degree drop in temperature at the
[KI] used in the I< method. Awtrey and C ~ n n i c k ~ ~
that emax(I3-) did not change with temperature.
An estimate of the temperature coefficient of the 13-
method was made by equilibrating a cuvette containing 13-
solution to 4-7 OC higher or lower than the spectrophotometer
temperature and reading the absorbance immediately after
insertion into the spectrophotometer and later when it had
reached the spectrophotometer temperature. We measured
a 0.2% increase in absorbance per degree drop in temperature,
7-fold greater than measured by Awtrey and C~nnick.~~
our value, we expect a variation of 0.1 pM in the measured
concentration of H202 due to the 0.5 OC range of our
Accuracy o f the Zy Method. The uncertainty in the 13-
method is detailed in Table 1 for 300 pM H202 and the
calibration procedure used in this study. The reaction of H202
with the buffer solution is the largest uncertainty. This reaction
which means a 0.03%
(41) Schwarz, H. A.; Salzman, A . J. Radiat. Res. 1958, 9, 502.
(42) Ross, C. K.; Klassen, N. V.; Shortt, K. R.; Smith, G. D. Phys. Med. Eiol. 1989,
(43) Bjergbakke, E.; Navaratnam, S.; Parsons, B. J.; Swallow, A. J. Radiat. Phys.
Chem. 1987, 30, 59.
(44) Peyton, G. R.; Glaze, W. H. Ewiron. Sci. Technol. 1988, 22, 761.
(45) Littman, F. E.; Benoliel, R. W. Anal. Chem. 1953, 25, 1480.
could be reduced by lowering the concentrations of KHP and
NaOH or by the use of another buffer.
Ferrous Ion Method. The measurement of 200 pM HzO2
by permanganate titration was compared to the measurement
by the ferrous sulfate method in which Fe2+ ions are oxidized
to Fe3+ ions by H202. The [Fe3+] was measured spectro-
photometrically at 303 nm. The ferrous sulfate solution
employed was the usual Fricke dosimeter solution which is
used as a dosimeter for ionizing radiation. The ferrous sulfate
solution was made up so that, after addition of an equal volume
of H202 solution, the final concentrations of ferrous ammonium
sulfate, NaCl, and sulfuric acid matched those of the usual
Fricke dosimeter. The value of (e(Fe3+) - €(Fez+)) at 303 nm
was taken to be 2174 M-I cm-l at 25 OC (corrected to 2159
at 24 0C).42 A computer simulation using a reaction scheme
based largely on the report by Bjergbakke, Navaratnam,
Parsons, and Swallow43 agreed that the final [Fe3+] should
equal 2.000 times the [H202] in the mixed solution, before
reaction. The ratio of [H202] measured by KMn04 to that
measured by Fricke was 0.997 f 0.004, based on four
measurements each. This confirmed that the ferrous ion
method is a reliable method for measuring H202 with an
uncertainty of less than 1%. While the value of c(Fe3+) makes
the ferrous ion method most convenient for H202 concentra-
tions above 50 pM, it should be noted that Bielski and Allen18
reported measuring the concentration of Fe3+ ions with a
reproducibility of f 9 X
Water Quality. The lamp in the Milli-Q UV water
purification system emits 185- and 254-nm light and will
produce ozone from dissolved 0 2 in the water. Peyton and
Glaze44 showed that this results in the formation of H202.
Schumb et al. (in ref 1, p 55 1) stated that ozone does not react
with KMn04. The water used in this study was tested for
H202 by KMn04 titration and for ozone and H202 by the
From the dependence of the absorbance at 525 nm vs
[KMn04], we estimate that the water used in this study
contained 50.2 pM H202.
We also tested the water using the 13- method. In a
neutral or alkaline solution, 1 molecule of ozone produces 1
molecule of iodine, whereas in acid solution 1 molecule of
ozone produces 1.25 molecules of iodine. *3945
for H202 or 03,
the absorbance at 351 nm was measured for
(i) water, (ii) an equiweight mixture of A and B, and (iii) an
equiweight mixture of A, B, and water. The absorbance in
(ii) minus the absorbance in (i) was assumed to be due to the
presence of adventitious 13- in (ii). The absorbance expected
in (iii) for dilution of (ii) with absolutely pure water was
calculated. The difference between the calculated and
measured values of the absorbance in (iii) was taken to be due
to H202 or O3 in the water. Several such tests indicated that
the water contained 50.2 pM (H202 plus 0 3 ) .
To test the water
We are happy to acknowledge the excellent technical
assistance of Andrew Weber. N.V.K thanks Dr. John Elliot
for very useful discussions during this study.
Received for review February 16, 1994. Accepted May 9, 1994.’
Abstract published in Advance ACS Abstracrs, July 15, 1994.
Analytlcal Chemlstry, Vol. 66, No. 18, September 15, 1994