ABSTRACT
It is clear from the equations that each atom of oxygen in solution finally liberates one molecule of iodine. The iodine is titrated against sodium thiosulphate solution using starch as indicator; 1 c.c. of normal sodium thiosulphate solution corresponds to 8 mg. of oxygen, or 5·6 c.c. of oxygen at N.T.P.
A number of modifications of the original technique of Winkler has been described. The accuracy both of the original method and also of any of its modifications is necessarily limited by the accuracy of the sampling technique. This factor is especially important when dealing with water of low oxygen concentration, for in order to obtain accurate analyses it is necessary to keep the water sample out of contact with the air. Krogh (1935) described a syringe pipette which could be used in the estimation of the oxygen content of water ; by the use of such a pipette contamination of the water sample with atmospheric air was completely avoided. Van Dam (1933, 1935) modified Krogh’s syringe pipette, and in the course of our studies of the respiration of aquatic animals we have further modified the syringe pipette of van Dam by replacing the metal collar with two locknuts. By our method it is possible to obtain an accuracy within 2%, even at low oxygen concentrations, using 1–2 c.c. of water as a sample (Fox, 1936 ; Fox, Wingfield and Simmonds, 1937 ; Fox & Wingfield, 1937).
In our syringe pipette the glass syringe with its thick-walled capillary glass nozzle is fixed in a metal and ebonite frame (Fig. 1).1 The head-screw, which can be rotated, moves in a sleeve rigidly attached to the frame. The end of the glass plunger of the syringe is kept in contact with the end of the head-screw shaft by a spring attached at one end to the plunger and at the other end to the frame. By means of the locknuts, which can be secured in any required position on the headscrew, the volume of the sample taken for analysis can be varied. In all our work the locknuts have been adjusted to give a pipette volume of about 1·5 c.c.
The volume of the syringe (barrel + dead space in nozzle) is determined chemically as follows. The head-screw is screwed down until the locknuts are in contact with the sleeve and then the syringe is completely filled with a standard solution (N/40) of potassium iodate, which is subsequently delivered from the syringe into a titration vessel, the syringe being rinsed out twice with distilled water. One c.c. of 1 % potassium iodide solution and 3 drops of o-phosphoric acid are then added to the titration vessel. The iodine liberated, which is quantitatively equivalent to the amount of potassium iodate originally present, is titrated against a standard solution (N/40) of sodium thiosulphate, using starch as indicator. The titration is carried out with a 10 c.c. blue-line burette. From the result of the titration the volume of the syringe is calculated. The volume of the dead space in the nozzle of the syringe is determined in a similar way, the dead space alone being filled with potassium iodate solution. This solution is then drawn up into the barrel of the syringe together with some distilled water, washed out, and the amount of iodate present determined by titration with the micrometer syringe burette described below. The volume of the barrel alone is found by subtracting the volume of the dead space from the total volume of the syringe.
For the determination of the oxygen content of fresh water the procedure is as follows. The head-screw is screwed down until the locknuts touch the sleeve. The dead space of the syringe pipette is then filled with aqueous manganous chloride solution (40 g: in 100 c.c. of solution). To do this some of the solution is drawn into the syringe, which is then turned so that the nozzle points upwards. The plunger is raised, expelling the air bubble, so that only the dead space remains filled with the solution. The outside of the nozzle is then washed in water, and the water sample to be analysed is drawn into the syringe, the plunger being withdrawn until its head comes into contact with the shaft of the head-screw. Alkaline iodide solution in water (32 g. sodium hydroxide and 10 g. potassium iodide in 100 c.c. solution) is next drawn into the syringe by unscrewing the head-screw. The quantity of this solution drawn in is made equal to about twice the volume of manganous chloride solution in the dead space. In practice this was found to be given by one turn of the head-screw. The approximate volume corresponding to one turn of the headscrew is previously found by titration in the way described above. After the alkaline iodide has been taken in, the syringe pipette is shaken a number of times until the precipitate of manganous hydroxide is evenly distributed, and then laid on its side for 3 min. to complete the absorption of oxygen by the precipitate. After further shaking of the syringe pipette, o-phosphoric acid is taken in by three turns of the head-screw. The syringe pipette is again shaken until all the precipitate has disappeared and iodine is liberated. The solution is finally ejected into a titration vessel.
With sea water the precipitate of manganous and manganic hydroxides obtained with the concentrations of reagents given above is much coarser than that obtained with fresh water. This leads to considerable inaccuracy in results because the coarse precipitate of manganous hydroxide does not quickly absorb all the dissolved oxygen: instead of obtaining values within 2 % of one another, analyses of a given sample of water were found to differ by as much as 10 %. We have rectified this error with sea water by using more dilute reagents, which produce a finer precipitate. For analyses of the oxygen in sea water the manganous chloride solution is half of the strength used for fresh water (i.e. 20 g. in 100 c.c. of solution), while the alkaline iodide solution is a quarter of its previous strength (i.e. 8 g. sodium hydroxide and 2·5 potassium iodide in 100 c.c. of solution). Two turns instead of one turn of the head-screw are now necessary to take in the diluted alkaline iodide in order that the proportions of reagents used shall remain the same.
In subsequently calculating the oxygen content of the water sample analysed, allowance must be made for the oxygen content of the reagents drawn into the pipette. The oxygen content of the reagents used for fresh-water analysis1 is 3·4 c.c. per 1. at 20° C. (Krogh, 1935), and that of the reagents for sea water 5·4 c.c. per 1. (Fox & Wingfield, 1937). From these data the necessary correction factors can be deduced.
The iodine in the titration vessel is titrated against standard sodium thiosulphate solution. In our earlier titrations we used the Rehberg (1925) micro-burette of 0·1 c.c. capacity. Now, however, we use a special micrometer injection syringe as a burette, because it is not only equally accurate but is portable. The apparatus is shown in Fig. 2.2 It consists of a small-bore glass injection syringe, the plunger of which is actuated by a steel micrometer head. The glass syringe is held in a metal holder which is attached to the micrometer head by a metal clamp. The head of the glass plunger is held in contact with the shaft of the micrometer head by means of a rubber band. The syringe is filled by screwing the shaft of the micrometer head downwards until the plunger of the syringe reaches nearly to the end of the syringe barrel; standard sodium thiosulphate solution is then drawn into the syringe by screwing the micrometer head in the opposite direction. Any air bubbles which have been introduced are expelled by inverting the syringe and turning the screw. For a titration the syringe is held in a burette stand. The capacity of the syringe is about 500 c.mm. and readings can be taken accurately to 0·1 c.mm. The syringe is calibrated by weighing the amounts of water ejected by screwing the micrometer head downwards from one position to another. This procedure is repeated over different regions of the barrel of the syringe. We have found that the upper third of the syringe is of variable bore, and have therefore never used this part of the syringe for titrating.
For titration of the iodine liberated in a water analysis the solution is ejected and washed out from the syringe pipette into a titration vessel, which is an ordinary colourless glass specimen tube of about 10 c.c. capacity with a flat bottom. The micrometer syringe burette is held in a burette clamp so that the tip of the needle dips just below the surface of the solution in the titration tube. The solution is stirred throughout the titration with a glass rod bent so that its end forms a ring at right angles to its stem. One drop of a dilute starch solution (0·5 g. British Drug Houses’ Analar soluble starch per 100 c.c. of saturated salt solution) is used as indicator. The starch solution used is half the concentration of that recommended by Krogh. We have found that the weaker solution gives a more definite end-point. The colour of the liquid in the titration tube is compared with that of distilled water containing 1 drop of starch solution in a precisely similar glass specimen tube.1 The total time required for one oxygen determination by our method is less than 10 min.
The solution of sodium thiosulphate used in the titrations must be standardized at regular intervals, for its strength does not remain altogether constant. The standardization may be carried out in one of two ways, as follows. (1) Two c.c. of N/40 solution of potassium iodate are placed in a titration vessel together with 1 c.c. of 1 % potassium iodide solution and about two to three c.c. of distilled water. 5 drops of o-phosphoric acid are then added, and the iodine liberated is titrated against the solution of sodium thiosulphate using a 10 c.c. blue-line burette. From the result of the titration the strength of the sodium thiosulphate solution can be calculated. (2) The dead space of the syringe pipette is filled with N/40 solution of potassium iodate. This is drawn into the syringe barrel together with some distilled water and the solution ejected into a titration vessel. Three drops of 1 % solution of potassium iodide are added together with 1 drop of 0-phosphoric acid. The iodine liberated is titrated against the sodium thiosulphate solution using the micrometer syringe burette described above. As the volume of the dead space is known, the strength of the sodium thiosulphate solution can be calculated. The second method is useful when travelling as it obviates the necessity of carrying a 10 c.c. burette.
Using the method described above, a series of estimations was made of the oxygen content of a single water sample. This procedure was repeated with a number of water samples of different oxygen concentrations. The results, which include estimations both with fresh water and sea water, are shown in Tables I and II. It will be seen that in no case was the variation in any one series of estimations greater than 2 %, even at low oxygen concentrations.
The results obtained with the above method were also compared with those got by using the ordinary Winkler technique with water samples of approximately 160 c.c. The ordinary Winkler method was that described in Standard Methods of
Water Analysis (1925), but with the following modification. The amounts of Winkler reagents used were approximately two-thirds of the amounts recommended in the standard method. This modification was made necessary by the fact that we used water samples of 160 c.c. instead of 250 c.c. The results are summarized in Table III. It will be seen that the two methods give closely comparable results at high oxygen concentrations, but that at low oxygen concentrations the new method gives constantly lower results. This is probably due to atmospheric contamination of the water sample in the ordinary Winkler method.
Krogh (1935) states that a constant discrepancy of 1 % was found to exist between the results which he obtained, using a modified Winkler method, and those of C. J. J. Fox (1909) who used a physical method of analysis.
The calculations involved in the method of analysis described above are as follows :
Standardization of the sodium thiosulphate solution.
Two c.c. of N/4.0 solution of potassium iodate used.
Amount of sodium thiosulphate used in titration = 1·93 c.c.
Therefore normality of sodium thiosulphate solution =
.Determination of the volume of the dead space in the nozzle of the syringe pipette.
Dead space filled with N/40 solution of potassium iodate.
Amount of sodium thiosulphate solution used in the titration = 0·0185 c.c.
Therefore volume of dead space =
Determination of the volume of the syringe pipette (dead space and barrel).
Syringe filled with N/40 solution of potassium iodate.
Amount of sodium thiosulphate solution used in titration = 1·59 c.c.
Therefore volume of syringe =
And volume of barrel of syringe = 1·65—0·02 = 1·63 c.c.
Determination of the volume of liquid introduced with one turn of the head-screw.
Syringe filled with distilled water; N/40 potassium iodate solution then introduced by one turn of the head-screw.
Amount of sodium thiosulphate solution used in titration = 0·0425 c.c.
Therefore volume of liquid introduced with one turn of the head-screw =
Derivation of a formula for the calculation of the oxygen content of the water sample from the experimental data.
1 c.c. N solution sodium thiosulphate = 0·008 g. oxygen (see p. 437).
Let h = normality of the sodium thiosulphate used.
Then 1 c.c. of a solution of sodium thiosulphate of normality h
But
, where i = normality of potassium iodate used in the standardization of the sodium thiosulphate solution and t = titre of the sodium thiosulphate (2 c.c. of potassium iodate solution being used as described above).Therefore 1 c.c. of hN solution of sodium thiosulphate ≡
oxygen at N.T.P.Let the volume of the water sample be V c.c. and the volume of sodium thiosulphate solution required in the estimation be n c.c.
Then oxygen concentration of the water sample expressed in c.c./1.
- Numerical example.
SUMMARY
A method for the estimation of the amount of oxygen dissolved in water is described. The method requires only 1-2 c.c. of water and is accurate to 2%, even at low oxygen concentrations. The apparatus is portable.
REFERENCES
ADDENDUM
Van Dam (Zool. Anz. 118, 1937, 122) states that if the alkaline iodide solution comes into contact with concentrated sulphuric acid in the nozzle of the syringe pipette, free iodine may be liberated in solution, and he recommends that the nozzle of the syringe pipette be filled first with the alkaline iodide solution instead of the maganous chloride solution, and that the latter reagent be taken in subsequently. In this way direct contact of the concentrated alkaline iodide solution with the concentrated acid is avoided. We have tested the validity of van Dam’s statement as follows. The syringe pipette was filled with distilled water and the usual amounts of concentrated alkaline iodide solution and concentrated acid were drawn in ; the resultant solution was then tested for free iodine. When concentrated o-phosphoric acid is used no free iodine can be detected but if concentrated sulphuric acid is employed an easily detectable amount of free iodine is liberated, even if the concentrated acid does not come into direct contact with the concentrated alkaline iodide solution. It is clear that the use of sulphuric acid must be avoided ; we have always used o-phosphoric acid. The use of sulphuric acid in the ordinary Winkler method does not cause liberation of iodine from the alkaline iodide solution, presumably owing to the greater volume of the water sample and consequent greater dilution of the reagents.
The syringe pipette, with stainless steel metal parts and a spare glass part, fitted in a wooden case for travelling, can be purchased from Messrs Philip Harris and Co., 144 Edmund Street, Birmingham.
It was stated by Fox, Wingfield and Simmonds (1937, p. 213) that the oxygen content of the reagents is negligible. This is a mistake.
This is the “Agla” micrometer syringe apparatus of Messrs Burroughs, Wellcome and Co., Snow Hill Buildings, London, E.C. It can be purchased from this firm with the metal parts made of Stainless steel, fitted in a wooden case for travelling, with a spare glass part and spare needle.
This mode of comparison is preferable to that recommended by van Dam, who used the fluid decolorized in a former titration. Such a procedure is unsatisfactory because the sodium tetrathionate formed in the reaction is unstable and the blue colour often reappears some time after the titration has been completed.
