Determination of purity by differential scanning calorimetry (DSC

May 1, 1979 - An Undergraduate Experiment Using Differential Scanning Calorimetry: A Study of the Thermal Properties of a Binary Eutectic Alloy of Tin...
28 downloads 17 Views 3MB Size
M. E. Brown Rhodes University Grahamstown 6140. South Africa

Determination of Purity by Differential Scanning Calorimetry (DSC)

Measurement of the depression of the freezing point of a sample is often used to determine its purity. Some of the prohlems associated with the definition of the purity of a suhstance have been discussed by Smit ( I ) . He stresses that it is inadequate to specify only the fraction (usually mole percent) of the main component present, for identification of the impurities and investigation of their mode of incorporation in the solid and the nature of the melt formed are essential. Freezine-ooint denression calculations are usuallv based on . thr assumpticm that s d i d solutions are not formed and that the melt is an ideal solurion. The practical nim ot .vurity. d r terminations id usually to deride whrthrr the sample meets certain prcificarions umhirh arr determined by the intended iurther uses of the sample. Purity determination by rryoscopy was out un a mow quantitative ba& in thr method o i mea. suring the enthalpy o i premrlting (2) of n sul~sfnnwas u functiun of temprrature, introduced by Johnsron and Giauqur (3).Large sample masses were used and long periods were required for establishment of thermal equilihrium. Both factors have been dramatically reduced (4) by the development of differential scanning calorimetry (DSC). In differential scannine calorimetrv (5) the enerev -- rewired . to maintain both a s a m d e and an inert reference material a t the same temoerature throughout a . programmed temperature . change is recorded. During melting, energy has to he ;upplied to the sample. A melting endotherm for a pure substance is illustrated in Figure 1. To is the freezing point of the sample, and the area ABC is proportional to the enthalpy of melting, AH,.lt of the sample. T h e presence of an impurity in the sample (the solvent) generally lowers the freezing point of the solvent and also broadens the melting range, giving a broader DSC endotherm as illustrated in the inset in Figure 2. From endotherms such as illustrated in Figures 1 and 2, freezing points and enthalpies of melting may be determined readily. With more effort an estimate of the puritv of a comoound can he obtained, from analysis of the deiailedshape of %smelting endotherm, for example, Figure 2, without reference to compounds containing known amounts of impurities. The method ~

.~

Figme 1. idealized DSC recad of melting of 0.01 1-

of very pure indium metal (heatingrate. 4 = 1 K min-I). The slope of AB(= 1/R,) is used to correct for mermal lag. To is hmelting point. Area ABCrepesems me enUlllpy of melting, hHk,,,,,ZR is the maximum valueof the apparent heat capacity of this reference material (see text).

310 1 Journal of Chemical Education

is only satisfactorily applied for impurity levels below about 3 mole per cent. Such an exercise has several educational advantages. I t concentrates attention on some aspects of the thermodynamics of melting which are not often covered in detail in lecture courses. The experimental result, that is, lowering of the freezing point by impurities, is well known from introductory laboratory sessions, but is seen here in anew light with theorv. orovidina an "absolute" route to determination of . sample purity. Attention is also fnrusrd on m e of the important rechni~ruesof thermal analysis, TJSC, and its wrsati.its tor estimating thermodynamic properties. Sinw differential scannin: calorimrters may nor he widelv ;~rces;ible,data and referenre> are provided to enable the. rxerciie to be carried nut us a "drv-lah" experiment.

-

lhe Melting Process

The simplest system to consider is that in which the impurity does not go into solid solution with the solvent, hut forms an ideal solution in the melt, that is, aeutectic system. If the impurity is labelled, in the customary way, as component 2 and the solvent as comnonent 1,then freezing-point depression calculations are based (6)on the expressionfor the mole fraction of solute (impurity) in the equilihrium melt a t temperature, T

where ~ ~ , 1 , is, 1the standard molar enthalpy of melting of the pure solvent, assumed to he independent of temperature over the range, and Tois the freezing point of the pure solvent. At the unique temperature T = To the freezing point of the impure sample, the mole fraction of impurity is given by

Figure 2. idealized DSC record of melting of 0.01 mmal of an impure sample of phenacetin (heatingrate. 6 = 1 K min-'). The slope d YB (= slope of AB in Fig. 1) is used to correct the programmed temperature. Tp. to the sample temperature. Ts. (Ts = T, - YU). T, is the freezing point. The area ABC repThe fraction melted, resents the enthalpy of melting of phenacetin. TH~,,,,. F.. at temperature, T. is equal to (area ADEl(area ABC)). 2s is the maximum value of the apparent heat capacity d the sample (see text).

I t is assumed also, because of the restriction to consideration of ideal eutectic systems and to the formation of ideal solutions on melting, that .F H(F) = =Re,, therefore dHldF = mk.t,,l Combining these results

Ixi

1x2

0

x2

-

Figure 3. Low concentration region (xz x,. me fradion of sample melted F = ~ / ( n .+ n,) = ABI(AB+ BC) = ABIAC= x;/x2.

--

If the depression of the freezing point, ATf, is small, To Tr To2. Also x2 = n2/(n1 n2) = mMd1000, so that T,Tf where m is the molarity of solute and M1 the molar mass of the solvent. Hence

--

+

where K f is termed the cryoscopic constant. T h e ahove relationships apply only to low impurity levels (xa < 0.03), that is, to the extremity of the liquidus curve, PQ, of the simple eutectic phase diagram (Figure 3, inset), shown on a much enlarged scale in Figure 3. Only when the sample is completely melted, a t T > Tf, is the mole fraction of impurity in the liquid, xz, the same as that in the original sample, x2*. The value xz* is the minimum value which x2 attains. At T < Th when the fraction of sample that has melted, F, is less than unity, x2 > x2*. Assuming a linear initial segment of the liquidus Curve, PQ, and using eqn.

.

The lower limit of the melting process is T

If F can be determined a t various temperatures, T, a plot of L a g a i n s t 1IF should yield a straight line, provided that x k , ~ , is, ~independent of temperature. If the values of Toand are known, x2* can he determined from the measured slope of the line. The DSC curve is capable of providing values of F a t temperatures T for use in such a plot. The DSC Melting Curve

T h e DSC measures the rate of thermal energy transfer, dHldt, to or from the sample as the temperature of the sample holder, T, is changed a t a constant rate, d T l d t = 4. Thus the output from the DSC is directly proportional to the apparent heat capacity of the system, dHldT.

For an absolutely pure compound with zero melting range, d H l d T would become infinite a t the melting point, To. For a n impure compound, d H l d T is finite and is a function of T.

When the fraction melted, F, is zero, the apparent heat capacity of the sample is that of the solid mixture, and when F = 1the apparent heat capacity of the sample is that of the ideal solution. Intermediate behavior is obtained as follows