Rapid Magnesium Determination in Cement Manufacture in Sweden

touched safely while in operation. The currents induced in the bottom of the crucible produce a temperature of about 1300° C\, which is generally hig...
0 downloads 0 Views 271KB Size
1028

ANALYTICAL CHEMISTRY

(ca. 100 ml.) had emerged, and with small deviations remained a t these values until the column was nearly exhausted. The total amount of sodium hydroxide collected before the normality of the effluent began its marked decrease was 70 milliequivalents. Forty grams of resin whose capacity when fully regenerated was 2.3 me. per gram were used in the column. Assuming that the resin was initially 100% regenerated, the maximum amount of sodium hydroxide obtainable was about 90 me. Therefore 78% of the theoretically obtainable sodium hydroxide was collected in a relatively chloride-free state. The normality of the effluent sodium hydroxide was 0.097 and the normality of the chloride solution was 0,0019. This represents a ratio of 0.02 equivalent of chloride per equivalent of hydroxyl or an exchange efficiency of 98’%. The exchange efficiency was increased by reducing the rate of flow. When weighed amounts of sodium chloride were passed through the column a t the lower flow rate and collected in 500-ml. volumetric flasks, the normality of the sodium hydroxide agreed with the theoretical normality within the experimental error. If proper precautions are exercised, the sodium hydroxide

solutions prepared in this manner are almost completely devoid of carbonate. To assure this the column should be washed occasionally with dilute hydrochloric acid before regeneration to prevent an accumulation of carbonate in the resin. The distilled water used should be freshly boiled. For most purposes the limited amount of chloride present should not be a detriment, since its effect on the ionic strength is negligible. If solutions of sodium hydroxide less than 0.1 N are desired, the amount of chloride n-ill also be correspondingly less. Sodium salts other than the chloride might also be used in the preparation of sodium hydroxide; an anion such as sulfate which has a high ionic potential might be affixed to the resin more easily and therefore appear in the effluent in lower concentrations. LITERATURE CITED

(1) Daviesand Nancollas, flattire, 165,237 (1950). R E C E I V Efor D review July 16, 1951. Accepted December 13, 1951. bution 832. Department of Chemistry, University of Pittsburgh.

Contri-

Rapid Magnesium Determination in Cement Manufacture in Sweden K. G . I ~ W E ANDERSSON AB Gullhagens Brrtk, Skaude, Sweden

iK T H E Swedish cement industry, a method of analysis is used

I which has proved to be dependably trol of the manufacturing process.

accurate for the ron-

One of the manufacturers concentrates limestone by flotation, which increases the calcium carbonate rontent and reduces the magnesium silicate content. To obtain fully satisfactory control of the concentration process, the author has developed a rapid colorimetric method which reveals both the carbonate and silicate compounds of magnesium.

For this purpose a hi h frequency generator was obtained from the Reeve Electronic 609 West Lake St., Chicago, 111. The output of the generator is 750 watts; the induction coil, in which a crucible of platinum was heated, has only a few turns. The voltage drop across the coil is therefore so low that it can be touched safely while in operation. The currents induced in the bottom of the crucible produce a temperature of about 1300’ C., which is generally high enough to render the raw mix and limestone soluble in hydrochloric acid. The material is then only sintered, so that for the determination of magnesium the sintered cake is transferred to a beaker for further treatment. If the Iimestone is low in calcium the material will melt; to prevent this (as i t takes considerable time to dissolve the melted substanrr) ralciuni carbonate is added.

80.0.;

MAGNESIUM DETERMLYATION WITH BRILLIANT Y E L L O I

The reagent brilliant yellon- ( d ) , produced by Coleman (9: Bell, Norwood, Ohio, has been used in America for determination of magnesium in water. The addition of aluminum prevents th(. presence of this element from deranging the analysis; neither is the method affected by the presence of iron in small quantities. With the use of organic reagents for magnesium such as hematoxylin, curcumin, Titan yellon-, hydroquinone, and the like, it has usually been the great sensitivity of the reagent to sesquiosides which has prevented the method from being used in practice. As it was thought that this disadvantage could be eliminated, and the method was believed to be rapid, there was reason for further investigation of this reagent for use in the concentration plant, where a rapid test for magnesium was urgently needed.

A Lumetron Model 402E colorimeter was available. The instrument vias equipped with a mercury vapor lamp, monochromatic filters, and circular cuvettes. Earlier work with colorimetric methods has shown that no reproducible values can he ohtairied unless the indicator and sta-

ldizer are introduced below a pH of 5 or 6, and then the pH is rapidly raised by adding sodium hydroxide, and the colloidal magnesium hydroxide which is t,hen instantly formed is absorbed by the indicator (1). The stabilizer revents precipitation. Examination of a number of materia% revealed that a cellulose product from hlo- and Domsjo in Sweden was an excellent st,abilizer with very good keeping properties. This product is eth?los~ethylcellulose,sold as Tylose SLilOO. Prelimiiiary tests to find the proper wave length and sensitivit shon.ed t’hat 550 mp was best. The series of experiments whicg followd revealed clearly the need of a stabilizer. Without the st,d)ilizer, the samples precipitated immediately, regardless of ivhether magnesium was present in large or in small quantities. 1Iere it proved, lion-ever, that the zone of measurement extends up to about 0.5 mg. of magnesium oxide; therefore the change in transmittancy would be t.oo small in relation to the variation in magnesium content to allox sufficient accuracy in the analysis. Test runs with variations in the calrium oxide and aluminum oxide content showed that b0t.h these compounds strengthen the dye, giving lower transmittance values. When 8 mg. of calcium oxide and 0.1 mg. of aluminum oxide are added per 100 nil. of the solution t o be analyzed, t,he niet,hod is rendered insensitive to variat,ions in these quantities. Variation of Tylose and Stabilization with Time. Table I shows that the color intensity reached its maximum after 5 minutes and that 2 ml. of 1% Tylose make the method insensitive to large errors in titration. Variation of Sodium Hydroxide. Five milliliters of 20% sodium hydroxide are used hereafter to make the method insensitive

Table I. Variation of Tylose and Variation of Stabilization with Time Additions.

8 mg. of of CaO 0.1 mg. A1203

diluted to 80 ml.

ml. Of 0.057 dye 5 ml. of 20%

R.~oH

hIgO. Mg. 0.000 0.500 0.500

0.500 0.500 0.500

Tylose, 1% M1. 1

0.1 0.5 1.0

2.0 5.0

5 min.

100

i7,4 44.6 43.0 44.0

,

diluted t o 100 ml.

Transmittancy 15 min. 30 min. 100.9 100.9 100.9

10 min.

FJ7.7 45.0 44.4

44 0

i8.0

i7.1

45.4 43.2 43.4

45.4 41.9 43.4

Remarks Precipitation

V O L U M E 2 4 , NO, 6, J U N E 1 9 5 2

1029 10-ml. automatic pi et. Because of the difficulty of obtaining other cuvettes, the &e had t o be transferred with an ordinary volumetric pipet.

Table 11. Variation of Sodium Hydroxide Additions.

MgO, Mg. 0.000 0.300 0.300 0.300 0,300 0.300

i

8 mg. of CaO 0.1 mg. of A1208 2 mi. of 1 % Tylose 2 mi. of 0.05% D y e NaOH 20% NaOH, PH 12.2 11.4 1 1 22 .. 21 12.2 12.4

hil. 5 0 0.1 0 1 .. 5 0 6.0

20.0

diluted t o 80 ml. diluted t o 100 ml. Transmittancy, 5 hiin. 100.0 57.3 5 7 5 86 ,. O 60.3 01.3

Table 111. Influence of Other Metals



Additions.

AlgO, Mg,

0.000 0.200 0.200

8 mg. 0.2 mg.of ofCaO A1203 2 ml. of 1% Tylose 5 mi. of 0.027c dye 5 ml. of 20% S a O H TransFeiOa, mittancy, 1190, Mg. A Min. Alg. 0.00 100 0.000 0.03 60 5 0.17 0.50 70.0 0.17 0.17

_ _..

diluted to 80 rnl.

i

ZnO, Mg. 0.00 0 00 0.01 0.10

This method has been in use for more than 2 years, and i t has been applied to limestone, the raw mix, clinker, and cement. It is more accurate than the oxine method. It yields values 0.1 or 0.2% lower than the pyrophosphate method, but it is the author’s opinion that the gravimetric method gives values that are somewhat too high. This appears to have been demonstrated in the analysis of the standard solutions. As regards speed, five or six specimens can be analyzed in 3 hours. I t takes 30 or 35 minutes to determine the magnesium content of a clinker or cement specimen.

diluted to 100 ml.

Transmittancy, 5 Min. 100 70.8 69.6 75.5

Ifgo, Mg.

0 000 0.17 0.17 0.17

PROCEDURE RECOMMENDED FOR LIMESTONE SPECIMENS

Trans>[no, F i t t a n c y , Jlg. o Min. 0.00 100 0.00 70.8 0.01 61.3 0.10 63.0

~~~

0 -

Grind the specimen in an agate mortar until it no longer feels gritty, Weigh out 0.200 gram and heat it for 2 minutes in the crucible by induction. When the specimens are low in calcium, add about 0.2 gram of calcium carbonate to raise the melting point of the mixture sufficiently so that a sintered cake will be produced in the crucible. After s i n k i n g , dissolve in 20 ml. of 1to 1 hydrochloric acid and add about 50 ml. of water. Boil and make sure that solution is complete. Filter through 9 cm. of Munktell OOR paper direct into a 500-ml. graduate. Wash Kith hot water, dilute to the 500-ml. line, and mix very carefully. 90

-

60 70

I, /

2

3

4

-

60 -

5

500.05

Figure 1 .

% D y e in m / .

Variation of Brilliant Yellow

0.j

0.2

0,3

0.4 0,5

my. MgO

to hydrochloric acid additions for dissolving the technical specimens, on the assumption that the method will not be deranged by relatively large quantities of sodium chloride. Variation of Brilliant Yellow. Additions.

8 m g . of CaO 0.1 mg. of AlzOa 2 ml. of 1% Tylose

diluted t o 80 ml.

Figure 1 shoa s that about 3 ml. of the indicator in 0.05% solution should be used to make the method insensitive to errors in titrating; however, to keep within the same zone of measurement with this concentration-i.e bet\\-een 0.0 and 0.5 mg. of niagnesium oxide-the thickness of the solution must be changed. This n as not possible here, and it was therefore necessary to FT ork on the steepest part of the curve. However, in order to reduce the titrating error somewhat, the dye concentration can be changed to 0.02% and 5 ml. can be added instead of 2, using a standard volumetiic pipet. Influence of Other Metals. I n the practical application of the method, the iron oxide content for such eases as are pertinent here should never exceed 0.5 mg. Zinc oxide n 111probably be present in the order of