Journal of Thermodynamics & Catalysis

Journal of Thermodynamics & Catalysis
Open Access

ISSN: 2157-7544

+44 1300 500008

Research Article - (2016) Volume 0, Issue 0

Thermodynamic Calculation of Melting Heat

Sobko AA*
Department of Physics and Chemistry of New Materials AM Prokhorov Academy of Engineering Sciences 19 Presnensky Val, Moscow 123557, Russia
*Corresponding Author: Sobko AA, Department of Physics and Chemistry of New Materials AM Prokhorov Academy of Engineering Sciences 19 Presnensky Val, Moscow 123557, Russia Email:

Abstract

In the present paper, the method for approximate calculation of phase space volumes proposed in Ref. [1] is extended. Expressions for the melting heat of materials with f.c.c., b.c.c. and h.c.p. lattices are derived. Experimental results are compared to numerical calculations for a number of elements.

Keywords: Thermodynamic; Heat; First-order phase transitions

Calculation of a Phase Volume For Solids

A principally new approach to the calculation of the heat of firstorder phase transitions was proposed in Ref. [1]. It is based on an approximate calculation of the phase volumes of a system in different phases: solid, liquid or gaseous and calculation of the change of entropy given as a logarithmic ratio of the volumes of old and new phases thermodynamics-catalysis. The volumes of phase space for liquid and gaseous phases are also calculated therein:

thermodynamics-catalysis(1)

thermodynamics-catalysis (2)

Where VL and VG are the volumes of liquid and gas, respectively. The expression for the evaporation heat is obtained in Ref. [1] in terms of other thermodynamic parameters of the process. A good agreement of numerical results and experimental data suggests that the method is versatile and can be used for calculating the transition heat of other first-order phase transitions. As it was shown in Ref. [1], the calculation should take into account the work performed against external forces during a phase transition with employment of the free volume. The expression for the phase volume of liquid was obtained in Ref. [1], whereas the calculation of a melting heat requires the phase space volume of solid state. The following model is used to find this volume. Every atom (molecule, ion) of a solid is assumed to vibrate near the equilibrium state, and its energy in a self-consisted periodic field is thermodynamics-catalysis.

Quadratic expansion of energy into powers of thermodynamics-catalysisgives:

thermodynamics-catalysis

thermodynamics-catalysiscan be interpreted as an effective potential energy thermodynamics-catalysis, thermodynamics-catalysis.

By diagonalizing the quadratic form according to the standard technique, we obtain:

thermodynamics-catalysis(3)

Where mxx, myy, mzz are components of the effective mass tensor. thermodynamics-catalysis. This approach is utilized in solid state physics, for example, in Ref. [2]. Equation (3) is the elliptic equation with the axes thermodynamics-catalysis , the corresponding volume of the ellipsoid is

thermodynamics-catalysis

Integration over thermodynamics-catalysis in the phase space gives:

thermodynamics-catalysis

It was assumed in Ref. [1] that near a point of first-order phasetransition, the majority of atoms are in the state with the average kinetic energy. This assumption is proved by a good agreement between numerical results and experimental data on evaporation heat. Consequently, according to the equipartition theorem for kinetic energy [3], one can substitute 3KT/2 for thermodynamics-catalysis. Thus, the approximate volume of phase space for solid is given by:

thermodynamics-catalysis(3)

Calculation of the Melting Heat

The volume of phase space for solid is given by (3). In the first part of the present work [1] it was shown that the employment of the free volume (for unit mole, the free volume is thermodynamics-catalysis, where V is the geometrical volume, NA is the Avogadro constant, and Va is the volume of atom (molecule, ion)) gives a substantially better agreement between numerical and experimental results. The phase space volume for solid state is expressed as:

thermodynamics-catalysis (4)

and the phase space volume for liquid state is, respectively, (1):

thermodynamics-catalysis

Correspondingly, the logarithm of the ratio thermodynamics-catalysis is equal to:

thermodynamics-catalysis

By using the Stirling formula thermodynamics-catalysis and taking into account R=kNA, we obtain:

thermodynamics-catalysis

Thus, the expression for the melting heat has the form:

thermodynamics-catalysis(5)

Under normal pressure, the term PΔV is on the order of 105*10-6 ~ 10-1, and the value of the first term is ~104, hence, up to pressures of ~ 1010 Pa the term PΔV can be neglected. thermodynamics-catalysis. Finally, the expression for the melting heat at pressures below 1010 Pa has the form:

thermodynamics-catalysis (8)

Where dT/dP is the initial slope of the melting curve [6]. The obtained equation is simply solved numerically by the iterative method and provides the value of the melting heat. The results of calculations by formula (6) and the solutions of equation (8) are given in Tables 1-3. Experimental data for the calculations are taken from Ref. [4,5]. In Tables 1-3 the number of polymorphic phase transitions before melting is marked with asterisks *. Spaces in the tables mean that there is no available data on the jumps of volume in melting.

Melting heat for the f.c.c. lattice

To the f.c.c. lattice in the thermodynamics-catalysis –space, in which the expansion was performed, corresponds the b.c.c. lattice. In view of the symmetry of b.c.c. lattice, we take: mxx, myy, mzz Then, the melting heat for the f.c.c. lattice takes the form:

thermodynamics-catalysis(9)

The structural melting constant for the f.c.c. lattice is S0=1.1765.

For low-melting elements with weak interaction between atoms the effective masses are approximately equal to atomic masses, whereas for hard-melting materials the difference can be substantial. Results of calculations of the melting heat for f.c.c. lattices by formula (9) and solutions of equation (8) are given in Table 1.

El-t T λex VS ΔV ra VSf λT1 δ1 dT/dP λT2 δ2
Ne 24,48 324 1,397 0,22 1,60 0,365 335,4 -3,5 13,16 317 2,0
Ar 83,87 1210 2,412 0,35 1,91 0,656 1118 7,6 24,89 1102 8,9
Kr 116,6 1640 2,968 0,45 1,98 1,012 1497 8,7 30,13 1444 11,9
Xe 161,3 2290 3,709 0,56 2,18 1,098 2129 7,0 38,9 2082 9,1
Pb 600,7 4770 1,826 0,07 1,32 1,246 6146 -28,8 7,73 6182 -30,0
Al 933,5 10670 1,000 0,07 0,57 0,953 9676 9,3 6,41 9648 9,6
Ar 1235 11300 1,027 0,06 0,89 0,849 12775 -13,1 4,9 12664 -12,1
Au 1338 12700 1,019 0,05 0,91 0,829 13732 -8,1 6,2 13914 -9,6
Cu 1357 13000 0,709 0,04 0,72 0,615 13978 -7,5 4,69 14129 -8,7
Ac 1370 14200 2,260 1,18 1,846 19,4 14603 -2,8
Ni 1726 17600 0,659 0,05 0,78 0,539 18146 -3,1 3,7 17858 -1,5
Pd 1825 17200 0,885 0,64 0,819 6,4 19030 -10,6
Pt 2045 19700 0,910 0,06 0,70 0,824 21188 -7,6 6,2 21275 -8,0
Rh 2239 22600 0,829 0,86 0,669 5,9 23540 -4,2
Ir 2683 26400 0,857 0,89 0,679 6,2 28280 -7,1

Table 1: Calculation results of melting heat for f.c.c. lattices.

T is the melting temperature in K, λex [J/mol] is the experimental value of the melting heat, VS*105[m3/mol] is the molar volume of the solid state, ΔV*105[m3/mol] is the volume jump in melting, r*1010[m] is the radius of atom (ion), VSf*105[m3/mol] is the free volume, (dT/ dP)*108[K/Pa] is the initial slope of the melting curve, λT1 [J/mol] is the molar melting heat calculated by formula (9), λT2 [J/mol] is the molar melting heat obtained by solving equation (8), δ1 and δ2[%] are inaccuracies of λT1 and λT2, respectively.

Experimental values of the melting heat and the line of a simple regression are shown in Figure 1. The slope of the regression line is 1.1670, which well coincides (with a good accuracy of ~ 0.85%) with the structural melting constant of the f.c.c. lattice S0=1.1765.

Experimental-reduced-melting-heat

Figure 1: Experimental values of the reduced melting heat for elements with the f.c.c. lattice and the line of simple regression.

Melting heat for the b.c.c. lattice

To the b.c.c. lattice in the thermodynamics-catalysis –space corresponds the f.c.c. lattice. Hence, taking into account geometrical factors we take thermodynamics-catalysis Then the expression for the melting heat in the case of b.c.c. lattice takes the form

thermodynamics-catalysis

or

thermodynamics-catalysis (10)

The structural melting constant for the b.c.c. lattice is S0=1.0032.

Results of calculations of the melting heat for b.c.c. lattices by formula (10) are given in Table 2 along with the results of solving equation (8).

El-t T λex VS ΔV ra VSf λT1 δ1 dT/dP λT2 δ2
Cs 301,5 2090 7,096 0,169  1,65 5,964 2150 -2,8 25,9 2156 -3,2
Rb 312,2 2200 5,579 0,193 1,49 4,745 2257 -2,6 21,1 2234 -1,6
K 336,8 2400 4,536 0,117 1,33 3,943 2405 -0,2 17,7 2411 -0,5
Na 371 2640 2,368 0,064 0,98 2,131 2650 -0,4 8,77 2648 -0,3
Li 453,7 3130 1,300 0,032 0,78 1,180 3230 -3,2 3,41 3205 -2,4
Pb 600,7 4770 1,826 0,070 1,32 1,246 4416 7,4 7,73 4363 8,5
Ba 983 7530 3,821 0,030 1,43 3,084 6859 8,9 4,3 6990 8,7
Sr** 1042 8000 3,450 0,177 1,27 2,934 7694 3,8 16,5 7531 5,9
Eu 1095 9200 2,898 0,140 1,12 2,544 8040 12,6 12,8 7875 14,4
Yb* 1097 7700 2,484 0,100 1,07 2,175 7976 -3,6 17,0 8075 -4,9
La** 1194 10040 2,260 0,012 1,14 1,887 8298 17,3 2,5 8325 17,1
Pr* 1204 8200 2,080   … 1,06 1,780    …   … 2,2 8389 -2,3
Nd* 1294 10680 2,059 0,019 0,99 1,814 9037 15,4 2,6 9031 15,4
Sm* 1350 10900 2,000 0,072 1,20 1,564 9816 9,9 6,0 9612 11,8
U** 1405 9200 1,256 0,029 0,97 1,026 10016 -8,9 2,8 9920 -7,8
Mn*** 1517 12070 0,738 0,042 0,91 0,548 11392 5,6 3,3 11000 8,8
Be* 1551 13000 0,488 0,056 0,35 0,477 12128 6,7 5,0 11675 10,2
Gd* 1586 10200 1,990 0,041 0,97 1,760 11243 -10,2 5,7 11237 -10,2
Tb* 1629 10800 1,931 0,061 0,89 1,753 11699 -8,3 7,4 11636 -7,7
Dy* 1685 10900 1,900 0,093 0,89 1,722 12358 -13,4 12,0 12316 -13,0
Ho* 1747 14160 1,875 0,140 0,95 1,659 13226 6,6 14,3 12947 8,5
Y* 1795 11390 1,989   … 1,06 1,689   …   … 16,5 13432 -17,9
Er* 1802 17200 1,844 0,166 0,93 1,641 13872 19,3 15,0 13413 22,0
Fe** 1808 13800 0,709 0,030 0,83 0,565 13248 4,0 3,0 13033 5,5
Sc* 1814 15900 1,504   … 0,83 1,360   …   … 15,1 13726 13,7
Ti* 1933 15000 1,055   … 0,90 0,871   …   … 15,4 15454 -3,0
Th* 2023 15600 1,980   … 1,11 1,635   …   … 18,6 15342 1,6
Cr 2110 21000 0,723   … 0,84 0,574   …   … 15,9 18336 12,7
Zr* 2125 20000 1,402   … 1,09 1,076   …   … 16,3 16635 16,8
V 2160 23100 0,834   … 0,88 0,662   …   … 15,7 18158 21,4
Hf* 2503 20935 1,341   … 0,84 1,192   …   … 17,3 19489 6,9
Nb 2741 27200 1,084   … 0,77 0,969   …   … 16,2 21739 20,1
Mo 2890 27600 0,939 0,010 0,92 0,743 20254 26,6 0,96 20146 27,0
Ta 3269 31400 1,087 0,056 0,77 0,972 24069 23,3 5,4 23613 24,8
W 3680 35200 0,953 0,125 0,77 0,838 29634 15,8 7,8 27430 22,1

Table 2: Calculation results of melting heat for b.c.c. lattices.

T is the melting temperature in K, λex[J/mol] is the experimental value of the melting heat, VS*105[m3/mol] is the molar volume of the solid state, ΔV*105[m3/mol] is the volume jump in melting, r*1010[m] is the radius of atom (ion), VSf*105[m3/mol] is the free volume, (dT/ dP)*108[K/Pa] is the initial slope of melting curve, λT1[J/mol] is the molar melting heat calculated by formula (10), λT2[J/mol] is the molar melting heat obtained by solving equation (8), δ1 and δ2[%] are the inaccuracies of λT1 and λT2, respectively.

Experimental values of the melting heat and the line of simple regression are presented in Figure 2 for the most low-melt elements. First seven metals were chosen with the lowest melting points for which the most relevant experimental data on melting temperature and melting heat are available. The slope of the regression line is 0.967 and matches the structural melting constant of the b.c.c. lattice S0=1.0032 with a good accuracy of ~ 3.6%.

Experimental-line-simple

Figure 2: Experimental values of melting heat and the line of simple regression for elements with b.c.c. lattice and the lowest melting points.

Melting heat for the h.c.p. lattice

h.c.p lattice transforms into itself in the thermodynamics-catalysis – space. The effective masses are, respectively, thermodynamics-catalysis. Thus, the expression for the melting heat for h.c.p. lattice has the form

thermodynamics-catalysis(11)

where a and c are parameters of the h.c.p. lattice. Since the ratio c/a for the elements does not differ much from the ideal case thermodynamics-catalysis, we may replace real values of the ratio a/c in the expression for S0 with the ideal value and obtain the relationship for the melting heat for the h.c.p. lattice:

thermodynamics-catalysis(12)

The structural melting constant for the h.c.p. lattice is S0=1.075.

Results of calculations by formula (12) are given in Table 3. The results of solving equation (8) for h.c.p. lattices are also presented in the table.

El-t T λex VS ΔV ra VSf λT1 δ1 dT/dP λT2 δ2
Cd 594,1 6110 1,300 0,04 1,14 0,927 5516 9,7 5,3 5565 8,9
Zn 692,7 6670 0,917 0,04 0,83 0,773 6479 2,9 4,8 6516 2,3
Mg 992 9040 1,398 0,04 0,78 1,278 9116 -0,8 7,5 9303 -2,9
Ca** 1112 9330 2,586 0,13 1,06 2,286 10445 -11,9 14,9 10485 -12,4
Pm 1441 12600 2,010 1,06 1,710 31,3 14956 -18,7
Co* 1768 15200 0,662 0,01 0,82 0,523 15014 1,2 3,5 15578 -2,5
Tm 1818 18400 1,812 0,13 0,87 1,646 17389 5,5 12,0 17251 6,2
Lu 1936 19200 1,778 0,06 0,85 1,623 17879 6,9 15,9 18753 2,3
Tc 2445 23810 0,860 0,95 0,644 5,5 23441 1,5
Ru 2583 23700 0,814 0,77 0,699 6,1 24802 -4,6
Os 3327 29300 0,843 0,89 0,665 6,5 32220 -10,0
Re 3453 33100 0,886 0,03 0,72 0,792 31913 3,6 3,2 31898 3,6

Table 3: Calculation results of melting heat for h.c.p. lattices.

T is the melting temperature in K, λex [J/mol] is the experimental value of the melting heat, VS*105[m3/mol] is the molar volume of the solid state, ΔV*105 [m3/mol] is the volume jump in melting, r*1010[m] is the radius of atom (ion), VSf*105 [m3/mol] is the free volume, (dT/ dP)*108 [K/Pa] is the initial slope of the melting curve, λT1[J/mol] is the molar melting heat calculated by formula (12), λT2 [J/mol] is the molar melting heat obtained by solving equation (8), δ1 and δ2 [%] are the inaccuracies of λT1 and λT2, respectively.

Experimental values of the melting heat are presented in Figure 3 along with the line of a simple regression. The slope of the regression line is 1.1186 and matches the structural melting constant of h.c.p. lattice S0=1.0075 with a good accuracy of ~ 4%.

Experimental-melting-heat

Figure 3: Experimental values of melting heat for elements with the h.c.p. lattice and the line of simple regression.

Analysis and Conclusions

As one can see from Tables 1-3, the proposed model gives a satisfactory description of the melting heat for the elements considered. In Ref. [7], experimental data on temperatures and melting heat values for eight well characterized elements are given. As one can see from Table 4, even for these elements the experimental results noticeably differ.

El-t Tmin Tmax λmin λmax El-t Tmin Tmax λmin λmax
Al 930 933.74 7928.93 11300.9 Nb 2523 2773 26756.9 26849.8
Cu 1356 1357.8 7968.67 13453.3 Mo 2853 2901 20051.5 41254.2
Ni 1721 1744 15984.2 18663.4 Ta 3053 3275 24699.4 32027.8
Ti 1878 2093 15474.3 20923.6 W 3523 3683 33782.4 54971.1

Table 4: Spread of experimental values on temperatures and melting heat.

Tmin and Tmax are the minimal and maximal values of experimental melting temperature, λmin and λmax are the lowest and highest values of experimental melting heat. The same trend holds true for the values given in various handbooks. The results only coincide if data are taken from a single source. Consequently, it would not be correct to hope for a good agreement between experimental data and the numerical results calculated by formulae (9-10, 12).

The substantial difference between the experimental value of the melting heat for lead and the corresponding value calculated for the f.c.c. lattice along with the satisfactory agreement with the value of the melting heat calculated for the b.c.c. lattice suggest that a polymorphic transition from the f.c.c. lattice to b.c.c. lattice occurs near the melting point with the phase transition heat equal to λ ≈1700 −1800 J/mol. The phase transition heat is calculated in the frameworks of the proposed model.

Thus, one can assert that the molar melting heat at pressures below 1010 Pa is given by the expression

thermodynamics-catalysis

or by a solution of the equation

thermodynamics-catalysis ,

Where S0=1.1765 for the f.c.c. lattice, S0=1.0032 for the b.c.c. lattice and S0=1.075 for the h.c.p. lattice. The rest values have been defined earlier.

References

  1. Sobko AA (2016) Thermodynamic calculation of a heat of first-order phase transitions. Part I.J Thermodyn Catal 7.
  2. JM Ziman (1960) Electrons and Phonons.The Theory of Transport Phenomena in Solids.Oxford at the Clarendon Press, UK.
  3. Huang K (1963) Statistical mechanics. 2nd edn. John Wiley & Sons, New York.
  4. Elementov S (1976) Properties of Elements. 1st edn. Samsonov GV, Moscow M (eds.).
  5. Tonkov EU (1979) Fazovye diagrammy elementov pri vysikom davlenii (Phase diagrams of the elements under high pressure), Handbook, Main Editorial Office of Physico-Mathematical Literature.
  6. Stolovich NN, Minitskaya NS (1975) Temperaturnye zavisimosti svoistsv nekotorykh metallov (Tempereture dependencies of the properties for certain metals).
Citation: Sobko AA (2016) Thermodynamic Calculation of Melting Heat. J Thermodyn Catal 7: 165.

Copyright: © 2016 Sobko AA. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Top