vo L U M E
103
25, NO. 1, J A N U A R Y 1 9 5 3 LITERATURE CITED
Barieau, R. E., and Giauque, W,F., J . Am. Chem. Soc., 72, 5676 (1950). Bonner, TV. D., and Titus, A. C., Ibid., 52, 633 (1930). Brand, J. C. D., J . Chem. Soc., 1946,II, 585. Foulk, C. W,, and HollingsR-orth, M.,J . Am. Chem. Soc., 45, 1220 (1923). Gibbs, J. W , 3 Trans. Conla. Acad. Arts Sci., 3 , 155 (1876). Gillespie. R. J., J . Chem. Soc., 1950, 2493. Hammett, L. P.. and Deyiup, A. J., J . Am. Chem. Soc., 55, 1900 (1933). Hantasch, A. 2. p h y s . Chem., 61, 257 (1907). Hulett. G. A , . and Bonner. I\', D.. J . Am. Chem. Soc., 31, 390 (1909). Koch, C. W., personal communication. Kolthoff, I. M., and Sandell, E. B., "Textbook of Quantitative Inorganic Analysis," p. 309, Kex Tork, Jlacmillan Co., 1936. (13) Ibid., p. 419.
(13) Kunsler, J. E.. arid Giauque, IT. F., J . -4m. Chem. Soc., (1 . 952). . (14) Ibid., p. 3472.
74,
804
(15) Ibid., p. 5 2 i l . (16) Lichty, D. 3I.,Ihid., 30, 1845 (1908). (17) Mellor, J. IT., "Inorganic and Theoretical Chemistry," T-01. 10. P. 402, S e w York, Longmans, Green and Co., 1930. (18) Parker, R. G., J . SOC.Chem. Ind., 36, 692 (1917). (19) Reinhardt, R. d.,J . Ani. Cheni. Soc., 72, 3359 (1950). (20) Reinhardt, R. -I.,thesis, University of California, Berkeley,
1947. (21) Richards, T. IT-., and Hoover, C. R., J . Ani. Chem. Soc., 37, 95 (1915). (22) Seidell, .I.,"Solubilities of Inorganic and Metal Organic Compounds," pp. 1019, 1355, To1. I, New York, D. Van Xostrand Co., 1940. (23) Setlik, B., Chon. Ztg., 13, 1690 (1889). (24) Somiya, T., Proc. Imp. Acud. ( J a p a n ) , 3 , 76 (1927). RE:CEIVED for review Soyember 5 , 1951.
Accepted September 28. 1 9 j 2 .
Analytical Procedures Based on Complex Formation z. G. SZABO AND XI. T. BECK Znstitir te f o r I n o r g a n i c and Analytical C h e m i s t r y , Znirersity of Sueged, H u n g a r y
The paper discusses analytical procedures based on the precise quantitative elaboration of the detectable differences in the stability conditions of complex compounds. These procedures deal with the measurements of the changes occurring in the other parameters (pH, extinction, etc.) associated with the formation of complexes. The rules governing these changes can be deduced on the basis of simple assumptions. The theoretically obtained formulas can be practically illustrated by applying them for the determination of aluminum and fluoride ions, as well as of phosphoric, tartaric, citric, and oxalic acids forming normal or internal complexes.
E
since the beginning of analytical chemistry, complex compounds have played a role in chemical analyses. Considering the appreciable differences in the formation, dissociation, and stability conditions of complex compounds, their application can also be manifold. The specific and sensitive separation of the precipitates, their dissolution and masking, etc., areall based on the reactions of the complex compounds long used for this purpose. This paper-without giving a historical review of the problem-deals with the analytical procedures based on the exact elaboration of the existing differences in the stability conditions of c o m p l e ~compounds. These procedures are due to the measurements of the changes of other parameters (pH, extinction, rtc.) associated with the formation of complexes. The stability of the complex compounds is characterized by the dissociation equilibrium constant of the complex. If in a n ,4S,,complex the central ion is denoted by A and the coordinatively bound ions or radicals, the ligands, by X , then the equilihiium constant i c :
greater the smaller the radii of the central atom and the ligands. Thus, also taking the number of the charge of the central atom into account, i t can be predicted which of the analogous complex compounds will be more stable. Particularly stable complexes can be expected if the ligands are not only ions but also dipoles, which may be the case in ion radicals composed of several atoms such as cyanide, thiocyanate, hydroxyl, etc. (3). These statements concerning normal complexes can also be applied t o internal complexes, increasing to a great extent the number of reactions which can lie employed for analyt,ical purposes. Hitherto the differences in the stability conditions of comples compounds have been generally used for masking purposes in the field of qualitative and quantitative chemical analyses. I n these operations the component in question is introduced as a complex into the reaction. In the nonmasked reaction the complex decomposes and dissolvea into simple compounds. A good example of this is the separation of copper and cadmium in their cyanide complex by means of hydrogen sulfide: &Cd(CS)r
The smaller the value of the fraction, the more stable is the complex. The strength of the coordinative bond is a function of the quality of the central atom as aell as of the coordinatively bound radicals (3). Consequently a monotone series of variable stability can be set up either bl- changing the coordinatively bound radicals and retaining the same central atom, or by eychanging the central atoms within an identical coordinative sphere. As the formation of the complexes is an elertrostatic phenomenon, the evolved heat of formation-the measure of the stability-is
+ H!S
=
PHCS
+ 2 K C S + CdS
New possibilities open up on making the complex compounds react with a simple substance, so that on the addition of the simple substance a new coniples also forms. I n general this can be denoted as follows: Type I
AX,
+ B*
BX,
+ A*
(2)
in which the askrisk denotes a n ion-Le., a n electric chargewithout taking its sign and number into account. T h e reaction
ANALYTICAL CHEMISTRY
104 proceeds toward the right if complex B X , is more stable than AX,,, because (3)
and this can in turn be composed of the complex equilibriums as follows :
KAX,
k'=KBX,
(4)
If B is an easily determinable and B on the other hand is a difficultly determinable ion, then the procedure described above can be employed for the estimation of ion B-e.g., on the basis of the equation Fe(C204)3---
+ -4l+++$41(C20&--- + F e + + +
(5)
the determination of aluminum can be deduced from that of the iron complex, AI( CzO,)--- being, owing to the smaller diameter of the aluminum ion, far more stable than complex Fe(CZ0&---. Exchanging the ligands instead of the central ion during the reaction, the double 'decomposition Type I1 AX,
+ nY* e AY, + nX*
If only one of the complexes is sensitive to acid, it is sufficient to take only this dissociation into arcount. Axin-")- + A m -
+ nX-
(12)
and its equilibrium constant
(6)
will take place if KAX, > KAY,,
On the basis of this principle, experiments were started in order to elaborate the colorimetric determination of phosphoric, oxalic, tartaric, and citric acids. The more precise investigations immediately revealed that the color intensity of the thiocyanate complex is not merely the function of the concentration of the ferric and thiocyanate ions, but the extinction also depends to a great extent upon the Concentration of the hydrogen ion. This great dependence upon the p H can be interpreted by the dissociation of the complex corresponding to its equilibrium constant. Among these dissociation products the anion of the weak acid combines with the hydrogen ions of the solution, forming acid molecules. Therefore the acid sensitivity of the complex will be a function of the hydrogen ion concentration and of the dissociation constant of the acid formed from the coordinatively bound radical. The relation can be deduced as follows:
(7)
The dissociated X - anion combinm with the hydrogen ions of the solution H+
as t'he equilibrium constant of the process
+ 9-
$
HX
(14)
Let the equilibrium constant of this process be
KA
is again in relation
[H+l [X-I [HX]
I n this case the equilibrium value of X - is according to both
with the stability constants of the individual complexes. On the basis of this reaction type the following process can take place
Na3AI(OH)8
+ GSaF
$ Sa&F6
+ 6NaOH
(10)
enabling the alkalimetric determination of fluoride. However, within this type of reaction the exchange is not only suitable for the determination of ions bound coordinatively, but if the appropriate experimental conditions are maintained] conclusions can be drawn regarding the concentration of the central ion on the basis of the amount of the ligand exchanged. On the basis of Equation 10, the alkalimetric estimation of aluminum is elaborated ( 1 ) . I n the case so far discussed the complex formation was associated with the change of the concentration of the hydrogen ion; thus the shift of the equilibrium could be traced by measuring the pH. If one of the complexes is colored and the other colorless, the extinction of the solution may be a parameter which can be used for the measurement of the equilibrium shift. This type of reaction renders possible the colorimetric measurement of anions forming, for instance, more stable complexes with the ferric ion than thiocyanate; thus for measuring fluoride, phosphate, tartrate, oxalate, etc., the general reaction scheme is: Fe(CSN)c---
+ 6X-
FeX6---
+ 6CSS-
Hence the expression of the hydrogen ion concentration characterizing the state of equilibrium
(11)
where X is one equivalent of the complex-forming anion to be determined. Of course, Equation 11 alone does not reflect all the possibilities and reactions which really occur; however, these cannot be treated theoretically, as the derivations would hccome too complex. The significance of the coordination number and its change will become evident in the following. The authors' investigations were already in progress when the paper (4)of Ingols and his coworkers appeared, making use of this reaction for the determination of the fluoride ion.
shows that it may be a t given concentrations, the greater the dissociation constant of the acid and the smaller the complex constant. The less the resulting simple acid dissociates, the more sensitive is the complex against acid. From the practical point of vieiy it seems advisable to define the extent of the dissociation of the complex by determining the degree of decomposition as follows:
Khen on the action of the acid just half of the complex ion is decomposed, the left side nil1 be 1:
-(+
1 = K," [HX] KT [ H I On forming its logarithm and dividing it by n, and furthermore substituting the well-known terms -log K A = PKA and -log K T = PKT, we obtain: 0 = -1 n
KT
+ log [HX] + pH - PKA
from which the pH value causing 50% decomposition will be:
pH
=
1
~ K A - pKi1 - log [HX] n
The concentration of the resulting acid is unknown. However, if this acid dissociates to only a very small extent, which always occurs in the cases of interest here, then its concentra-
V O L U M E 25, NO. 1, J A N U A R Y 1 9 5 3
10s
tion can be assumed to be proportional in the first approximation to the original amount of the hydrogen ions added to the neutral complex salt solution (it is in no case equal thus to the hydrogen ion value obtained after the setting in of the equilibrium). If this is denoted by c, (22) in which the proportionality factor, a, is a quantity depending on the concentrations (mainly of the complex) prevailing in the solution. Equation 22 expresses the afore-mentioned fact that the pH value corresponding to the definite decomposition of the complex is the smaller (thus the comples can be produced and utilized a t larger hydrogen ion concentrations) the smaller the arid eyponent, t'he larger the complex exponent, and the more acid h:itl to be added to the solution to bring about adefinitedecomL)ositioii. The amount of this last term is therefore proportional to the I)uff(.i. capacity of the complex system. If, on the ot.her hand, both complexes are sensitive to xcitl, the pH dependence can be deduced on the basis of the folloxving considerations:
tate, as well as that of the concentration condition governing the system. The correlation can be deduced as follows: At the complete dissociation of the complex AX;n-m)-
e A"+ + n X -
(35) whereas the solubility product of the precipit:itr fo! rned on the basis of equation
+
ilm+ m O H - =
A[OH],
(36)
L
(37)
I*
[ A m - ] [OH-Im Thii,
.\-(I
=
(':in u-Iite that
m t t from this
T,et us assume the folloiying over-all reartion:
AYp-m)-
(39)
+ kx- $ L I L y i k - i n ) - + n]'-
whrre k and n represent the coordination numbers and valency of the cent.ral atom.
(23) vi
the
The dissociation of H X and HE' can be expressed by the miw action law and from these X and 1- can be substituted in (24)
H+
+ S-
$
HX
(34)
the value of the equilibrium constant is
(25)
Let us again form the quotient drtcrmining the extent of the conversion
This equation indicates that the decomposition of the comples occurs a t lower hydrogen ion concentration-thus, proceeding from the acid region, takes place the quicker-the smaller the solubility of the precipitate and the greater the concentration and complete dissociation constant of the complex. On the other hand, the presence of the X - displaces the limit of the precipitation. On forming the negative logarithm of the preceding equation and introducing the terms -log L = pL, -log KT = ~ K Tp ,H = 14-pOH, and considering that nearlv approximates the concentration of the complex c , whereas [X-] that of the concentration of the added X- ion, c', n e have
Thus on the basis of this equation the upper limit of our measuring region can easily be estimated. On the other hand, this equation can also be used for the determination of the equilibrium constant of complex compounds. EXPERIMENTAL
If [.i'yik-"'-]
=
['4E""-'"," I
then
forming the negative logarithm of the equation:
or
pII =
n pK3 - pKl
-k
+
p K n k log [HX] - n log [HY] (33) (n - k)
The above equations lead to statement that if n = k , Reaction 23 is independent of the pH, and if n < k , the increasing of the pH shifts Reaction 23 to the right, whereas if n > k it shifts it toward the left. In this case the experimental results have shown that n < k ; therefore a high p H value provides optimal experimental conditions. A further raising of the alkalinity is limited by the circumstance that a t higher p H values the precipitation of A(OH), sets in. This upper limit is aside from being a function of the equilibrium constant of the complete dissociation of the complex, also that of thesolubilityof the formed precipi-
In studying the conditions governing the complex 3ystem arid also their analytical application, the hydrogen ion concentration of the solution must be kept a t a known constant value. In order to effect this the different buffer solutions cannot be used because they usually contain only the ions to be determined, or a t least those interfering to a great extent with the measurements. The p H of the solutions should be adjusted with perchloric acid by means of a pH meter provided rrith an electronic voltmeter: Radiometer P H M Type 22. The colorimetric measurements were carried out with a Pulfrich photometer, and color filter 5 4 7 was chosen as most suitable for the purpose, Colorimetric Determination of Aluminum. On the basis of Type I (Equation 5 ) , the colorimetric determination of aluminum can be carried out as follows: The addition of the oxalate diminishes the extinction of the ferric thiocyanate solution. I n the presence of aluminum ions the decrease of the extinction does not reach such an extent as in their absence, because instead of ferric ion an equivalent or somewhat smaller aluminum ion proceeds into the complex. Thus a relative increase in extinction can be observed. If the other experimental conditions are kept precisely constant, the increase of the extinction is merely a function of the aluminum content, thus rendering possible the colorimetric estimation of aluminum. The maintenance of the pH a t a constant level is the indispensable condition of such procedures.
ANALYTICAL CHEMlSTAY
106 Solutions required are 4.822 grams of C.P. ferric ammonium sulfate dodecahydrate dissolved in distilled water and treated with enough perchloric acid to adjust the p H to 3 after mixing with 5 ml. of the ammonium thiocyanate solution final volume is 1000 ml. Three grams of C.P. ammonium thiocyanate are dissolved in 1000 ml. of water. C.P. potassium oxalate (2.672 grams) is dissolved in 1000 ml. of t r i c e distilled water and the pH is adjusted Kith perchloric acid to 3. C.P. (0.8962 gram) aluminum chloride hexahydrate is dissolved in twice distilled water and its pH is adjusted M-ith perchloric acid t o 3. Final volume is 1000 ml. The calibration curve rras recorded as follows: The reagents were introduced into the measuring flask (50 nil.) in definite sequences and the flask was filled up to the mark with perrhloric acid adjusted to suitable pH. The solution prepared in this manner was determined colorimetrically, 0.01 -V perchloric acid being used as a comparison solution. The cuvette was 0.997 cm. long. I n the first place the effect of the sequence of the mixing of the solutions on the extinction was examined.
D
60.
The results of the measurements are listed in Table I.
Table I.
Effect of Sequence of Mixing
Sequence of hlixine of Solutions Fe, CSN, Al, oxalate Fe, SCN, oxalate, -41 Same after 0.5 hour AI, oxalate, Fe, SCN
Extinction 0.824 0.690 0.796 0.813
Volumes of reagents were constant in each case: 10 ml. of AlCla. 5 ml. of KHaSCN, 10 rnl. of RzCzO~,
Table 11.
Transmittance, 0% 15.0 20.4 16.0 15.2 5 ml. of Fe(NHd)(SOr)g,
Figure 1. Calibration Curve of Aluminum
-
Effect of aluminum on 70 transmittance (-) and extinction (- -) of femc-thioc) anate-oxalate systems a t pH 3 Volume of reagents. 5 ml. of Fe(NHa)(SOdz, 5 ml. of NH‘SCN, 10 ml. of K&04
Change of Extinction with Time Time, A h . 15 30 60
E
D
0 458 0 460
34.8 34 7 34 6 34 6
0 461 0 461 Volunies of reagents were constant In each case: 5 ml. of S H a S C N , 10 ml. of KzCzOd, 5 ml. of .41Clr 120
5 ml. of Fe(NHd(S04)2,
The change of extinction in time in the order ferric ion, thiocyanate, oxalate, and aluminum can be attributed to the fact that the reartions between complexes present were not instantaneous. In later measurements the solutions were mixed in the following order: ferric ammonium sulfate, thiocyanate, aluminum chloride, oxalate. I n this case the extinction of the solution is reasonably constant, with practically no change during the period of the measurement as can be seen from Table 11. Consequently, under experimental conditions the reduction of ferric ion by means of oxalate does not take place. The calibration curve is shown in Figure 1. Considering that in the above procedure the determination of aluminum is deduced from the colorimetric estimation of ferric ion, all ions interfering with the determination of the latter naturally also interfere in this case, whereas, for instance, magnesium and zinc do not even interfere with the estimation in 100-fold excess. Alkalimetric Determination of Aluminum. On the basis of Type I1 (Equation 10) the alkalimetric determination of aluminum can be carried out as follow: The aluminum salt is treated with an excess of sodium hydroxide and equal amounts of the “aluminate” solution obtained in this manner are titrated with 0.1 N hydrochloric acid, n i t h or without alkali fluoride. I n both cases neutral red \+as used as indicator, changing its color a t the pH value of the precipitation of aluminum hydroxide. The ions most frequently associated with aluminum do not interfere with this procedure and the disturbance due to carbonates and silicates can easily be eliminated, the latter, for instance,
Figure 2.
Calibration Curve of Phosphate
Effect of phosphate on 70 transmittance of ferric-thiocyanate systems at pH 2.7
by coating the vessels 11ith paraffin. The detailed experimental procedure and data are given by Beck and Szab6 ( 1 ) . The determination can, of course, be carried out also by means of other anions involving complex formation with aluminume.g., by means of oxalate ions (6). With fluoride, the titration was accomplished from the alkaline side, The procedures can be written according t o the following equations :
+ 3H’ = hl(0H)a + 3H20 + 3 c ~ O a - - = .AI(CLO,)g--- + 6 0 H -
AI(OH)e---Al(OH)6---
(Ila) (41b)
The solutions required are the same as those used in the fluoride
V O L U M E 25, NO. 1, J A N U A R Y 1 9 5 3
107
Table 111. Batch -4nalyses -41
Calcd., Mg.
Consumed I
0.1 4 HC1
I1
-41 Found, Jig.
8.04 8.06 16.11 16.09
13.88 13 .93 30.85 30.89
5.376 5,421 10.770 10.812
5.394 5.394
10.788 10.788
Difference Mg.
70
-0.018 +0.027 -0.018
-0.33 +0.50 -0.16
+0.034
+0.32
procedure, but instead of the alkali fluoride a 20% potassium oxalate solution neutral t o neutral red as used. The determination \vas carried out as follows:
A solution containing 20 to 100 mg. of aluminum ion is introduced into a measuring flask of 100-ml. capacity. Enough 0.2 N carbonate-free sodium hydroxide is added to redissolve the initially formed aluminum hydroxide precipitate. The solution is made up to 100 ml. and aliquots of 20 ml. are titrated. The first sample is titrated to the color change of the neutral red indicator. I n titration of the second sample a solution of 20 ml. of potassium oxalate is previously added and then it is titrated also in the presence of neutral red nq indicator to the same color a'
the first one. The difference in the milliliters of 0.1 3' hydrochloric acid consumed in the two measurements is proportional to the aluminum content: 1 ml. 0.1 -V hydrochloric acid is equivalent to 0.899 mg. of aluminum ion. The results of a few batch analyses are listed in Table 111. Colorimetric Determination of Phosphoric, .Tartaric, Citric, and Oxalic Acids. Experiments were carried out a t different pH values in order to determine the diminution of the extinction. The extinction curve \vas recorded as follows: The ferric thiocyanate solution was adjusted to the appropriate pH and 5 ml. was introduced into a suitable measuring flask of 50-ml. capacity; then phosphoric, tartaric, citric, or oxalic acid solutions adjusted to the same p H were pipetted into the vessel in different quant'ities. Subsequently the flask was filled up to the mark with perchloric acid adjusted to the corresponding pH. The estinction of the solution prepared in t,his manner was determined; 0.01 A- perchloric acid was used as comparison solution. The Iwgth of the cuvet,te was 0.997 cm. Solutions used for the determination of phosphoric acid were: C.P. ferric ammonium sulfate dodecahydrate 9.644 grams, and ('.r.ammonium thiocyanate 3.6 grams, dissolved in 1000 ml. of water; the pH n-as adjusted with perchloric acid to 2.7 and 1.63, rcspectivcly. One gram of C.P. potassium dihydrogen phosphate was dissolved in 1000 ml. of water and the pH was adjusted with perchloric acid to 2.7 and 1.63. The extinction values obtained are illustrated in Figure 2. The solutions applied for the determination of organic acid n.ere: C.P. ferric ammonium sulfate dodecahydrate 4.822 grams and C.P. ammonium thiocyanate 1.8 grams dissolved in water; the pH was adjusted x i t h perchloric acid to 3.1, 2.55, and 1.6 units, respectively. Final volume was 1000 ml. One gram each of oxalic, t,artaric, and citric acid was dissolved in 1000 ml. of water. The extinction values are shown in Figures 3,4, and 5 ,
Figure 3.
Calibration Curve of Phosphate
Effect of phosphate on %transmittance of ferric-thiocyanate systems at p11 1.63
I
Figure 5.
Calibration Curves of Oxalic, Tartaric, and Citric Acids
-
Effect of oxalic (-) tartaric (-) and citric acids on % transmittance'of ferric-thiocyakate systems at pH 2.55 I
U
2
4
6
--mg
pH.37
Figure 4.
8 f0 of orqanrc a m o n
Calibration Curve of Oxalic, Tartaric, and Citric l c i d s
-
Effect of oxalic (-), tartaric (-) and citric (-.-.-.) acids o n % transmittance of ferric-thioc) &at, systems at pH 3.1
(-*-.-e)
Figures 2 to 5 demonstrate unequivocally that an increase of the concentration of the hydrogen ion does not favor the shift of reaction to the right. Thus on the basis of above theoretical considerations it can be stated that the coordination number of the complexes formed is higher than that of the thiocyanate complex present. .is thr estinction-diminishing effect of the different
ANALYTICAL CHEMISTRY
108 complexes depends to a different extent upon the concentration of the hydrogen ion, there opens up the possibility of the parallel colorimetric determination of the complexes by means of properly adjusted p H values. On changing the initial ferric ion concentration and the dimensions of the cuvette, the amount of the ions to be determined may vary widely.
D”.
i
f L ’
25
PH
Figure 7.
-
75
5 16
-mg
i0
SCN-
Effect of Thiocyanate
On % transmittance of ferric-oxalate (-) and ferric-citrate (-.-.-.) systems a t pH 1.6 Volume of reagent. 5 ml. of ferric thiocyanate, 25 ml. of citrate, 10 ml. of oxalate
Figure 6.
Calibration Curves of Oxalic, Tartaric, and Citric Acids
-
Effect of oxalic (-), tartaric (- -), and citric (-----) acids on 70 transmittance of ferric-thiocyanate systems a t pH 1.6
Another q-ay of extending the limits of the measurements is the changing of the thiocyanate concentration. This can be seen from the data demonstrated in Figure 6. CONCLUSION
The circumstance that the decrease of the extinction of the different complexes varies to a great extent with the pH enables the different ions to be determined independently, which is particularly significant in the case of certain organic acids.
I n the literature the determination of the stability constants of the complexes is gaining more prominence ( 2 ) . The preconditions for their analytical application are discussed above. ACKNOWLEDGMENT
Thanks are due to S. Iisldor for his hclpful assist.ance in the extinction measurements. LITERATURE CITED
(1) Beck, >I.,and Seab6, Z. G., Anal. Chim. A c t a , 6, 316 3,1952). (2) Chem. Eng. News, 29, 40T8 (1951). (3) Hein, Fr., “Chemische Koordinationslehre,” p. 229. Leipzig,
Hirzel Verlag, 1950. (4) Ingols, R. S., et al., AXAI..CHEM.,22, 799 (1933). ( 5 ) Lacroix, S., A n d . Chi7n. A c t a , 1, 3 (1947). RECEIVED for review February 4, 1952. .4coepted September 13. 1952.
Detection and Determination of Thallium J. R. -4. ANDERSON New South Wales University of Technology, Sydney, Australia
T
HE element thallium was discovered by Crookes in 1861. It is one of the rarer metals and occurs in small amounts in the minerals crookesite, lorandite, hutchinsonite, and vrbaite. It also exists in very small quantities in sea water and in certain mineral waters, and is widely distributed throughout the vegetable kingdom, being found in traces in wine, chicory, tobacco, beet, and beechwood. I t s chief industrial source is in the chamber mud from sulfuric acid plants in which thalliferous pyrites are burnt, and in the flue dusts of blast furnaces and zinc refineries. Alloys of thallium have many industrial applications and soluble thallium salts, on account of their extreme toxicity, are finding increasing use as economic poisons for the extermination of rodents and other pests.
DETECTION OF THALLIUM
Because of its toxicity it is often necessary to detect minute amounts of thallium, This is best done ~pectrographically, but wet methods or sensitive microchemical tests may be eniployed. Thallium gives a broad green line a t wave length 5350.7 A. and is, therefore, readily detected by means of the spectroscope. De l l e n t and Dake (16) report that univalent thallium salts form a double salt with uranyl carbonate, which emits fluorescent radiation in the short-wave region of the visible spectrum. The fluorescence detection of thallium was carried out by Got; (@), who found that thallium(1) was like silver in preventing the fluorescence of uranyl sulfate solution. Thallium(III), however, changes the fluorescence of rhodamine
