Colorimetric Determination of Dissolved Oxygen at Low Concentrations

Author(s): William W. Broenkow and Joel D. Cline

Source: Limnology and Oceanography, Vol. 14, No. 3 (May, 1969), pp. 450-454
Published by: American Society of Limnology and Oceanography
Stable URL: http://www.jstor.org/stable/2833812 .
Accessed: 09/06/2014 21:05

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp

.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact [email protected].

American Society of Limnology and Oceanography is collaborating with JSTOR to digitize, preserve and
extend access to Limnology and Oceanography.

http://www.jstor.org

This content downloaded from 188.72.127.42 on Mon, 9 Jun 2014 21:05:45 PM

All use subject to JSTOR Terms and Conditions
450 NOTES AND COMMENT

COLORIMETRIC DETERMINATION OF DISSOLVED OXYGEN AT

Low CONCENTRATIONS'

The Winkler (1888) method for the de- released in proportion to the dissolved
terminationof dissolved oxygen is used oxygen originallyin the sample.
almost exclusively in limnological and
oceanographic work. Divalent manganese REAGENTS
is oxidized by molecular oxygen in basic 1. Manganous sulfatesolution. Dissolve
solution and, upon subsequent acidifica- 25 g of reagent grade MnSO4 4H2O or 20
tion, iodide is oxidized to free iodine by g of MnSO4*H20 in 500 ml of distilled
the oxidized manganese; the free iodine water. This reagent is stable indefinitely.
concentrationis then determinedby titra- 2. Alkaline iodide solution. Dissolve 14
tionwiththiosulfate.Althoughthe method g of reagent grade KOH and 60 g of
has been subject to some criticismand reagent grade KI in 200 ml of distilled
modificationsince its inception(Thompson water. This reagenttends to discolor after
and Robinson 1939; Wheatland and Smith a time because of the oxidation of iodide
1955; Carpenter1965; Carrittand Carpen- to iodine, and only small quantitiesshould
ter 1966), it is probablythe most-usedana- be prepared. Afterthe calibration factor
lytical method in the chemistryof natural has been determined,add 0.1 g of reagent
waters. grade sodium azide (NaN3) to the alka-
When oxygen concentrationsare less line iodide solution. Sodium azide is used
than 10% of the saturation value the to eliminatenitriteinterference(Alsterberg
Winklermethod is imprecise,atmospheric 1925); however,it also interfereswith the
contaminationcan be serious, and special oxidationof iodide by iodate.
analytical and sampling procedures must 3. Phosphoricacid solution. Dilute 100
be used. Our work in oxygen-deficient en- ml of 85% phosphoricacid with 400 ml of
vironmentshas led us to devise a method distilled water. This reagent is stable in-
to measure preciselyoxygenconcentrations definitely.If the pH of the treatedsample
between zero and 50 ,ug-atoms/liter. The is less than about 2, the rate of air oxida-
method is not intended to replace the tion of the iodide is increased, and high
Winklermethod,and it is useful only with apparentoxygenconcentrations result.The
low oxygen concentrations. use of phosphoricacid permitsbetter pH
control, and phosphoric acid is less cor-
MATERIALS AND METHODS
rosive than the sulfuricacid normallyused
Oxygen concentrationscan be deter- in the Winklermethod.
mined colorimetricallyby measuring the 4. Primarybiiodate solution. Dissolve
absorption of ultravioletlight by the tri- exactly 0.1300 g of dry, reagent grade
iodide ion that is produced when Winkler KH(IO3)2 in distilledwater and dilute to
reagents are added to the sample (Custer exactly1 liter (0.3333x 10-3 M). It is
and Natelson 1949). Unless the sampling stable indefinitely.
procedure is precise, atmosphericoxygen 5. Secondary biiodate solution. Dilute
contaminationwill invalidate the results. 25.00 ml of the primarybiiodate solution
Samples are thereforedrawn into syringes to exactly 1 liter. This potassium biiodate
to eliminate contact with the atmosphere solution is 0.8333 x 10-5 M and is equiva-
(Broenkow 1969). Reagents similar to lent to a solutionwith an oxygen concen-
those used in the Winkler method are trationof 50 pg-atoms/liter.
injected into the syringes,and iodine is 6. Iodine solution. Dissolve 25 g of
reagent grade KI and 0.1 g (one pellet)
1 Contribution
No. 494 fromthe Departmentof of KOH in about 200 ml of distilledwater.
Oceanography,University of Washington,Seattle.
This researchwas supportedby NationalScience Dissolve 0.32 g of reagent grade I2 in this
FoundationGrantGA-644 to Dr. F. A. Richards. solutionand dilute to 1 liter. This solution

This content downloaded from 188.72.127.42 on Mon, 9 Jun 2014 21:05:45 PM

All use subject to JSTOR Terms and Conditions
 NOTES AND COMMENT 451
should be weakly basic to retardair oxida- fromlightto preventphotochemicalreduc-
tion of the iodide ion. It should be stored tion of the iodine.
in an amber bottle and protected from Measure the absorbance of the triiodide
intenselight. The solutionis about 1.25 x ion at 352 mjt in a 1-cm glass cell. A
10-3 M in I2, and the concentrationin- spectrophotometer having either a tung-
creases slowly as the iodide oxidizes. sten or a hydrogenlamp can be used.
The oxygen concentrationsin the man- In some oxygen-deficient environments,
ganous sulfate and alkaline iodide nor- compoundsare presentthatcan produce or
mally used in the Winkler determination remove iodine. If this is the case, a sepa-
are about 30% of the saturationvalue for rate analysis must be made to determine
distilled water. The errorassociated with the iodine equivalence of the interfering
the introductionof oxygen with the re- substances. To do this, add to 25 ml of
agents in the Winkler method is about 1 the sample 0.5 ml of the phosphoricacid,
jtg-atom/liter when 1 ml of each reagent shake well, then add 0.5 ml of the alkaline
is added to 300-ml samples. Oxygen con- iodide solution and shake. Add 0.500 ml
taminationfromthe reagentsin the method of the iodine solution and shake. The
described here would be about 20 ug- iodine solutioncan be added accuratelyto
atoms/literif the oxygenwere not removed the sample syringewith a micrometersy-
from the manganous sulfate and alkaline ringe fittedwith a small-borecannula. If
iodide. To minimize this contamination, no interferingsubstances are present,the
the manganous sulfateand alkaline iodide resultingsolutionshould contain25 umoles
are stored in aspiratorbottles fittedwith I2/liter. If the absorbance of the sample
rubber septums, and nitrogenis bubbled treatedfor interfering substancesindicates
throughthe reagentsfor30 min beforeand more or less than this concentration,a
duringthe analyses. Because the reagents correctionmust be applied.
are added to the samples with syringes,
the reagent viscosity must be kept low, STANDARDIZATIONS
so the reagents are about one-tenththe Standardizations preferably should be
concentrationof those used in the Winkler carried out in the syringesso that dilution
method. effects (about 4%) are duplicated in the
PROCEDURE standardsand in the samples; standardiza-
Samples (25 ml) are drawn into glass tions can be done in suitable glassware
syringes(Broenkow 1969) fromsampling and a correctionapplied to account for
bottles fitted with rubber septums using dilution. As in the Winkler analysis, the
large-bore (16 gauge) cannulas. Air con- iodate ion is the primary standard, and
taminationmust be avoided, and the sy- oxygen contaminationduring standardiza-
ringes must be flushed with the sample tion is not a factor. For standardization,
water. draw a syringefulof secondary biiodate
Immediatelyaftersampling,add 0.5 ml solution. To this add 0.5 ml of phosphoric
of manganous sulfate and 0.5 ml of alka- acid, shake well, and add 0.5 ml of alkaline
line iodide solutionsthroughthe tip of the iodide. Mix well and add 0.5 ml of man-
sample syringeusing calibrated1-mltuber- ganous sulfate solution. The iodine lib-
culin syringesfitted with small-bore (20 erated is equivalent to an oxygen concen-
gauge) cannulas. After these reagents trationof 50 jtg-atoms/liter.
have been added, cover the tip of the The reagentblank is determinedby add-
sample syringeand shake. After 30 sec, ing 0.5 ml of phosphoricacid and 0.5 ml
add 0.5 ml of phosphoricacid solutionand of alkaline iodide to a syringefulof dis-
shake. The reactionis sufficientlyfastthat tilled water. Shake well and add 0.5 ml
it is unnecessaryto allow the precipitate of manganous sulfate solution. A small
to settle before acidification. absorbance results from the presence of
The treatedsamples should be protected reducing compounds in the reagents and

This content downloaded from 188.72.127.42 on Mon, 9 Jun 2014 21:05:45 PM

All use subject to JSTOR Terms and Conditions
452 NOTES AND COMMENT

TABLE 1. Reproducibility
studyof the colorimet- TABLE 2. Effectof sodiumazide in eliminating
ric oxygenmethod nitriteinterference

Oxygenconcn Iodine concn

Sample Absorbance (ALg-atoms/liter) (ALmoles/liter)
(oxygenequiv-
1 0.484 20.1 Absorb- alence gg-
ance atoms/liter)
2 0.486 20.2
3 0.497 20.7 Withoutsodiumazide added to alkaline-iodide
4 0.484 20.1 0.0 Ag-atoms/liter
NO2- 0.458 19.0
5 0.492 20.4 21.4
5.0 ug-atoms/liter
NO2- 0.515
6 0.492 20.4
7 0.494 20.5 Withsodiumazide added to alkaline-iodide
8 0.481 20.0 0.0 Ag-atoms/liter
NO2- 0.460 19.1
9 0.481 20.0
5.0 Ag-atoms/liter
NO2- 0.462 19.1
10 0.489 20.3
Avg 0.488 20.3
SD ?+0.23 AR is the absorbance of the test solution
treated for reducing compounds.

DISCUSSION
fromsmall amounts of I2 in the KOH-KI
An estimate of the overall precision of
solution.
the colorimetricoxygenanalysiswas made
The calibrationfactor,F, is determined
fromthe expression by determiningthe oxygen content of 10
replicate samples of low oxygen water.
F = 5O/(As-Ab), The sample water was withdrawnwith the
sample syringesfrom an airtight,piston-
whereAs is the average absorbance of three fittedchamber(Cline 1968) thatprevented
standards,and Ab is the average absorb- atmospheric contamination. The average
ance of two reagent blanks. The value of oxygen concentrationwas 20.3 ,ug-atoms/
F should be about 43. Oxygen concentra- liter with +0.5 ,ug-atoms/liter (2 SD) pre-
tions are computed fromthe expression cision (Table 1).
02 (,ug-atoms/liter)= F ( As - Ab), The efficiencyof sodium azide in elim-
where An,is the absorbanceof the unknown inating nitriteinterferencewas shown by
sample. adding free iodine to samples containing
The iodine solutionmustbe standardized nitrite.When azide was not present,nitrite
daily by adding 0.5 ml of the alkaline oxidized iodide to iodine giving higher
iodide solution and 0.5 ml of phosphoric apparent oxygen values. When azide was
acid solutionto 25 ml of distilledwater in added to the alkaline-iodidesolution,how-
a sample syringe. Shake well, and add ever, the nitrite interferencewas com-
0.500 ml of iodine solutionwith a microm- pletely eliminated (Table 2).
eter syringeand shake again. The absorb- Certain reducingsubstancessuch as sul-
ance of the triiodideion is determinedas fite, thiosulfate,or sulfide are sometimes
before. presentin polluted waters or reducing en-
The concentration of reducing sub-
stances is computed fromthe relation:
TABLE 3. Effectof thiosulfate
interference
reducing compounds
(Ftmoles 12/liter) = F (AI -AR,), Iodine concn
(,umoles/liter)
(oxygenequiv-
where the concentrationof reducing sub- Absorb- alence ,ug-
ance atoms/liter)
stances is reported in iodine equivalents,
F is the calibrationfactordeterminedfrom 0.0 umoles/liter S2032- 0.458 19.0
the biiodate standardization,A, is the ab- 10.7 umoles/liter
S2032- 0.325 13.3
sorbance of the iodine standardization,and 21.4 umoles/liter
S2032- 0.203 8.2

This content downloaded from 188.72.127.42 on Mon, 9 Jun 2014 21:05:45 PM

All use subject to JSTOR Terms and Conditions
 NOTES AND COMMENT 453

DISSOLVED OXYGEN(/4g-atoms/liter) COLORIMETRIC

OXYGEN
(ml/liter)
O 10 20 30 40 50 60 0 0., 0.2 0.3 0.4 0.5
0 70 I

0.6
a) 60 *-g,? /

E 50 - * * X
200 -

4 ~~~~~~0.3
g30 *u;7K/ x
400 -
o o
20* 0: ':

zL o
0/
lo 20* Q050 6-0.
CL

600
0 10 20 30 40 50 60
OXYGEN
COLORIMETRIC (/.g-atoms/liter)
set
FIG.1: 2. orlton ahdln
Comparisonbetween rpeet
colorimetrich
and
Winkleranalysesof dissolvedoxygenduringRV
800 _ ThomasG. Thompsoncruise28. Solid line repre-
sents 1: 1 correlation.Dashed line represents
t-he
least squares regressionwith a regressioncoeffi-
cient, b - 1.06, with standarddeviationon the
slope, Sb = +0.13.
1000 , I x,
0 0.1 0.2 0.3 0.4 0.5
DISSOLVED OXYGEN (mI/liter)
1968 (University of Washington,unpub-
lished data), oxygen concentrationsin the
FIG. 1. Verticaldistribution of oxygendeter- low oxygenwaterwere determinedby both
minedcolorimetrically in the easternPacificOcean the Winkler and colorimetric methods.
(RV Thomas G. Thompsoncruise26, station67,
8? 50' N lat, 900 21' W long,21 February1968).
Colorimetric samples were drawn from
conventionalsamplingbottlesthroughrub-
ber septums. A comparisonof the Winkler
vironments. These substances reduce io- and colorimetric values (Fig. 2) shows that
dine and give lower apparent oxygen Winkler values are about 10 ytg-atoms/
values, and their concentrationsmust be liter higher than the colorimetricvalues,
determined by the described procedure. and the data suggest that oxygen contam-
The addition of 19 )umoles/literof free ination in the reagents and atmospheric
iodine to a sample 21.4 )uM in thiosulfate contaminationof the samples are impor-
decreased the iodine concentrationby 10.8 tantsourcesof errorin the Winklermethod
ytmoles/liter, in agreementwith the iodine- at these low concentrationlevels. The
thiosulfatestoichiometry(Table 3). modificationof the Winkler method used
An extensive body of oxygen-deficient here employsreagentsas recommendedby
water occurs in the eastern Pacific Ocean Carrittand Carpenter (1966), but differs
along the coasts of Mexico and Central by titrating100-ml subsamples. The re-
America between about 100 and 200 N gression coefficientbetween the Winkler
latitude. In much of this region, oxygen and colorimetricvalues is 1.06 and is not
concentrations below 10% of the saturation significantly differentfrom1.00. The stan-
value occur between depths of 100 and dard deviationof the Winklervalues about
1,000m (Fig. 1). DuringRV ThomasG. the regressionline is Sz,I,= ?7 )ug-atoms/
Thompson cruise 26, January to March literand representsan estimateof the pre-

This content downloaded from 188.72.127.42 on Mon, 9 Jun 2014 21:05:45 PM

All use subject to JSTOR Terms and Conditions
454 NOTES AND COMMENT

cision of the Winkler analysis including Winklermethodfor dissolvedoxygenanaly-

sampling errors. Thus, at low concentra- sis. Limnol. Oceanog., 10: 135-140.
CARRITT, D. E., AND J. H. CARPENTER. 1966.
tion levels, the Winklermethod can be in Comparisonand evaluationof currently em-
error by 10 ug-atoms/literand has ten ployed modifications of the Winklermethod
times poorer precision than the colori- fordetermining dissolvedoxygenin seawater;
metricmethod. a NASCO report. J. MarineRes., 24: 286-
318.
WILLIAM W. BROENKOW CLINE, J. D. 1968. Kineticsof the sulfide-oxy-
gen reactionin seawater;An investigation at
JOEL D. CLINE
constanttemperature and salinity. M.S. The-
sis, Univ. Washington,Seattle. 68 p.
Department of Oceanography, CUSTER, J. J., AND S. NATELSON. 1949. Spectro-
Universityof Washington, photometric determination of microquantities
Seattle 98105. of iodine. Anal. Chem., 21: 1005-1009.
THOMPSON, T. G., AND R. J. ROBINSON. 1939.
REFERENCES
Notes on the determination of dissolvedoxy-
gen in sea water. J. MarineRes., 2: 1-8.
ALSTERBERG, G. 1925. Methoden zur Bestim- WHEATLAND, A. B., AND L. J. SMITH. 1955.
mung von in Wasser gelosten elementaren Gasometricdetermination of dissolvedoxygen
Sauerstoff bei gegenwart von salpetriger in pure and saline wateras a check of titri-
Saure. Biochem. Z., 159: 36-47. metricmethods. J. Appl. Chem. (London),
BROENKOW, W. W. 1969. An interface sam- 5: 144-148.
pler using spring-actuated syringes. Limnol. WINKLER, L. W. 1888. Die Bestimmung des im
Oceanog., 14: 288-291. Wasser gelostenSauerstoffes. Chem. Ber.,
CARPENTER, J. H. 1965. The accuracy of the 21: 2843-2855.

SPECTROPHOTOMETRIC DETERMINATION OF HYDROGEN SULFIDE IN

NATURAL WATERS'

The determinationof small concentra- investigationsof anoxic marine basins in

tions of hydrogensulfideby the methylene which hydrogensulfide occurs as a respi-
blue method has been widely used since ration product of sulfate reduction.
its introductionby Fischer (1883). The A procedureusing unsubstitutedp-phen-
methodis sensitiveand specificforsulfide- ylenediamineto produce Lauth's violet in
sulfur and is readily adaptable for rou- the determinationof sulfide-sulfur in sea-
tine analyses. Many modificationsof the water was described recently (Strickland
method have been made for specific pur- and Parsons 1968). The method described
poses (Almy 1925; Sheppard and Hudson here has several advantages over the
1930; Sands et al. 1949; Fogo and Popow- Lauth's violet procedure,including 1) the
sky 1949; Budd and Bewick 1952). The use of a single reagent,containingN,N-di-
procedurehas also been used in determin- methyl-p-phenylenediamine sulfate, which
ing sulfide-sulfurin sewage (Pomeroy need not be recrystallizedbefore use; 2)
1936) and seawater (Fonselius 1962). superiorcolor and reagent stabilityat low
In our experience, the methyleneblue sulfideconcentrations;3) applicabilityto a
method was somewhaterratic,and studies wide range of sulfide concentrations;4)
were undertaken to define the optimum approximatelya 10% increase in sensitivity,
conditionsfor determiningsulfide-sulfur
in and 5) a simplifiedprocedure for stan-
naturalwaters. The studyevolved during dardization. The method is applicable to
natural waters containing 1-1,000 ug-at-
1ContributionNo. 493 fromthe Departmentof oms/litersulfide-sulfur(0.03-32 ppm) and
Oceanography,University of Washington,Seattle.
This researchwas supportedby National Science is free of salt effects and temperature
FoundationGrantGA-644 to Dr. F. A. Richards. dependence.

This content downloaded from 188.72.127.42 on Mon, 9 Jun 2014 21:05:45 PM

All use subject to JSTOR Terms and Conditions