- Email: [email protected]
Impurity distribution in metallic dysprosium during distillation purification
JOURNAL OF RARE EARTHS, Vol. 34, No. 9, Sep. 2016, P. 924
Impurity distribution in metallic dysprosium during distillation purification ZHANG Xiaowei (张小伟)1,*, MIAO Ruiying (苗睿瑛)1, LI Chuanjun (李传军)2, WU Daogao (吴道高)1, YAN Huan (闫 缓)1, WANG Zhiqiang (王志强)1, CHEN Dehong (陈德宏)1, YAN Shihong (颜世宏)1, LI Zongan (李宗安)1 (1. National Engineering Research Center for Rare Earth Materials, General Research Institute for Nonferrous Metals, and Grirem Advanced Materials Co. Ltd., Beijing 100088, China; 2. Otto Schott Institute of Materials Research, Friedrich Schiller University Jena, Löbdergraben 32, D-07743 Jena, Germany) Received 29 January 2016; revised 20 May 2016
Abstract: The distribution rules of impurities contents in distilled metallic dysprosium were researched, and a theoretical analysis was carried out. The research results indicated that, the content of impurity in distilled metal, such as Al and Fe, was lower in the initial stage, increased slowly in the middle stage, and increased rapidly in the last stage during the process of distillation purification. The calculated method of separation coefficient of impurity in crude metal by content of impurity in distilled metal was not suitable for high pure metals, and the modified separation coefficient was proposed, and it equaled 1/6.1 and 1/16.9 for impurity Al and Fe. The physical process of distillation was coincident with that of solidification essentially, and solute re-distribution theory in solidifying front could be used to describe the impurity distribution near evaporating surface. In the former stage of distillation purification, the diffusion of impurity in liquid metal could reach a quasi-equilibrium state, the calculated result of impurity content in distilled metal agreed well with experiments. In the latter stage of distillation process, the diffusion rate of impurity in liquid metal decreased, and the content in distilled metal was larger than the calculated result. Keywords: vacuum distillation purification; distilled dysprosium; impurity distribution; modified separation coefficient; rare earths
High-purity rare earth metals have been widely used in high technique field, which is a key raw material of high-performance magnetostrictive material[1], magnetic refrigerant materials[2,3], superconducting materials[4–6], etc.. Vacuum distillation is an important method for purifying rare earth metals. Vacuum distillation is an important method for purifying rare earth metals, in this purification process, the impurities, with lower saturated vapor pressure than the matrix metal, can be separated at high temperature environment in vacuum distillation purification process. The impurities, with lower saturated vapor pressure than the matrix metal, can be separated at high temperature environment in vacuum distillation purification process. Up to now, most previous studies focused on preparation of rare earth metals by vacuum distillation technology[7–14], measurement of distillation velocity[1,7,15,16], the effects of experimental conditions on purification[17,18], and few researchers used the separation coefficient to judge whether the impurity can be separated from the matrix metal[18–22]. Xi et al.[19] discussed the behavior of impurities in distillation purification of metallic terbium, the impurities with the separation coefficient βi>1 (defined as βi≡γipiθ/γmpmθ) can be evaporated preferentially, the impurities with the separation coefficient βi<1 will be remained mostly in the crucible, and the impurities with
βi≈1 cannot be separated by vacuum distillation; Li et al.[18,20] considered that it is credible to use separation coefficient to judge whether the impurity can be separated or not, but due to the lack of the activity coefficient of impurity, βi cannot be calculated accurately by the definition formula, but actually it can be obtained by another equation of βi = α i M i / M m , αi is volatilization coefficient of impurity, determined by volatilization rate of impurity and matrix metal based experiment, and it is found that the impurities of Al, Cu, Cr and Co present a negative deviation in liquid scandium, and Ni and Si present a positive deviation; Pang et al.[21] calculated the volatile quantity of impurities in metallic neodymium, the theoretical quantity equated with the experimental results for the impurities with βi>1, and there is a large error for the impurities with βi>1; Zaiour et al.[22] discussed behavior of impurities in distillation process and the removal efficiency of submicron major impurities in tellurium can be characterized by studying the effective separation coefficient α, which is affected by both the evaporation rate and particle size. The above researches are all related on the separation coefficient to judge whether the impurities can be removed, however, the removal rate of the impurities and the distribution of impurities in distilled metal have not been studied. In present study, a vacuum distillation purification ex-
Foundation item: Project supported by National Natural Science Foundation of China (51504036), National Basic Research Program of China (2012CBA01207), National High Technology Research and Development Program of China (2011AA03A409) * Corresponding author: ZHANG Xiaowei (E-mail: [email protected]; Tel.: +86-10-82241180) DOI: 10.1016/S1002-0721(16)60116-3
ZHANG Xiaowei et al., Impurity distribution in metallic dysprosium during distillation purification
periment of metallic dysprosium was carried out, and the content of impurity was determined by Inductively Coupled Plasma Atomic Emission Spectrometry (ICP-AES), and the distribution rules of impurities were obtained, and the theoretical analysis of the distribution rules was discussed.
1 Experimental The self-preparation metallic dysprosium (Dy) was used in the present experiment, the starting material consists of 20 casting ingots, every ingot is about 250 g, and the total amount of metal is 4.99 kg; 6 samples were sampled and analyzed by ICP-AES, and the average contents of impurities are listed in Table 1, and the purity of metallic dysprosium is about 99.87 wt.%. The schematic of distillation equipment is shown in Fig. 1. The starting material was placed into a tungsten crucible with dimensions of 115 mm in external diameter, 95 mm in inner diameter and 225 mm in height; considering that the melting point of Dy is 1407 ºC, a soaking temperature of 1500 ºC was maintained for 8 h under a pressure of 10–5 Pa, the temperatures of crucible were measured by an infrared radiation thermometer (RATMR1SBSF, Reytek, USA) through the sight hole in the furnace. In this period, the metallic dysprosium evaporated from liquid metal surface and condensed at the cooled collector (tantalum sheet, consists of a cylinder and a cover, with thickness of 0.1 mm, about 50 g), leaving high melting point and low vapor pressure impurities at the bottom of tungsten crucible. In order to be convenient for discussing the distribution rule of the impurities, the height of collector was de-
925
termined by the total amount of starting material, the height of collector was calculated by following equation H=mDy/(ρDyπ(dEx/2)2), where mDy is mass of metallic dysprosium, kg; ρDy is density, 8550 kg/m3; dEx is external diameter of crucible, m; which could guarantee the lower surface of the distilled metal to be at tungsten upper edge level.
2 Results After distillation experiment, about 120 g grey powder was left in crucible bottom, and the total amount of deposited Dy and collector (tantalum sheet) was 4.92 kg, after machining off the collector, distilled metal was 4.86 kg, the longitudinal section of distilled metal is shown in Fig. 2, and the schematic diagrams in various distillation periods are indicated in Fig. 3. In initial stage of distillation purification, as seen in Fig. 3(a), the metallic vapor will condense on the inner surface of the cylinder and the cover sheet to form a thin metallic film; and then metallic atom will deposit on the surface of metallic film, seen in Fig. 3(b) and (c); as the distillation process continues, the evaporated metallic atom deposited on the cover sheet presents a truncated cone and deposited on the cylinder sheet presents a circular ring with a triangular section, as seen in Fig. 3(d). In order to discuss the distribution rules of impurities, five samples were taken along the axial line in distilled metal successively (as seen in Fig. 2) after the experiment finished, and the size of every sample is 10 mm×5 mm×5 mm; 3 samples were taken for each position, the sample were dissolved by nitric acid, the impurities were determined by ICP-AES (PerkinElmer Optima 8100), the
Table 1 Content of 24 impurities in metallic dysprosium (mg/kg) Y
La
Ce
Pr
Nd
Sm
Eu
Gd
Tb
<10
19
<10
<10
<10
<10
Yb
Lu
Mg
Al
Si
Ca
10
30
23
86
116
211
Ho
Er
<10
36
Cr
Mn
15
22
Tm
117
68
<10
<10
Fe
Co
Ni
Cu
45
29
18
189
Fig. 1 Schematic of distillation equipment of metallic dysprosium before (a) and after (b) distillation
926
JOURNAL OF RARE EARTHS, Vol. 34, No. 9, Sep. 2016
Fig. 2 Distilled metal in present experiment
mean contents of impurities are plotted in Fig. 4, and the standard deviation is listed in Table 2. It is found that the contents of impurities of Al and Fe present the same tendency. Taking Al impurity for example, in initial stage of distillation experiment, the content of Al in distilled metal is 13 mg/kg, as the distillation process continues, the content of Al increases to 20, 45 and 83 mg/kg, in last stage of purification experiment, the content of Al increases to 170 mg/kg. It is concluded that, in initial stage the content of impurity is lower, and then increases slowly, and in last stage of distillation experiment, the content of impurity increases rapidly.
3 Theoretical analysis and discussion 3.1 Modified separation coefficient In the distillation purification process, the separation coefficient of the impurity is considered as a criterion to judge whether the impurity can be removed, and it is de-
fined as the following form[13]: γ pθ (1) β i = i θi γ m pm where γi and γm are activity coefficients of impurity and matrix metal respectively, pθi and pθm are saturated vapor pressures of impurity and matrix metal respectively. For the distillation purification of rare earth metals, the purity of starting material is over 99.9 wt.% generally, and the activity coefficient of matrix metal is approximately 1, and the activity coefficient of impurity is usually determined by distillation experiment. The separation coefficient can be calculated by another formula in distillation separation of crude metal[18]: (2) βi = αi M i / M m where αi is volatilization coefficient, obtained by Yi= 100–100(1–Xm/100)αi [18]; Yi is volatilization rate of impurity, calculated by Yi=cidmd/cisms, cid and cis are contents of impurity in distilled metal and starting material, md and ms are mass of distilled metal and starting material, Xm is volatilization rate of matrix metal, and the above parameters are measured by experiment. Throught Combining Eqs. (1) and (2), the activity coefficient of impurity γi can be written: γ i = αi ⋅ γ m ⋅
pmθ piθ
Mi Mm
(3)
In the distillation separation of crude metal, the conTable 2 Standard deviation of the measurement results of content of impurities (mg/kg) Impurity
Sample 1
Sample 2
Sample 3
Sample 4
Sample 5
Al
0.82
1.25
1.63
1.70
2.16
Fe
1.25
0.94
1.70
1.25
2.49
Fig. 3 Schematic diagrams of distillation process
Fig. 4 Content of impurities Al (a) and Fe (b) in distilled dysprosium
ZHANG Xiaowei et al., Impurity distribution in metallic dysprosium during distillation purification
tent of impurity or component is about several percents to several tens percents, and the content of such impurity or component in distilled metal is almost unchanged throughout the whole distillation process; however, in the distillation purification for preparation of high purity metal, the purity of starting material is over 99 wt.% generally, the amount of impurities is very small, and each impurity is about hundreds of mg/kg, due to the difference of saturated vapor pressure between impurity and matrix metal, the content of impurity in distilled metal has great difference during the whole distillation process, several times or several tens of times difference between initial stage and last stage. In the present distillation experiment, the content of impurity is changed throughout the distillation process, taking Al impurity for example, it is 13 mg/kg in initial stage, and it is up to 170 mg/kg in last stage, and the difference is about 13 times. Consequently, the separation coefficient of impurity cannot be calculated accurately by above method through the content of impurity in distilled metal, and the above calculation method is not suitable for preparation of high-pure metal by vacuum distillation. In order to describe the separation degree of impurity to the matrix metal, a modified separation coefficient is introduced, which is the dimensionless ratio of distillation velocity of the impurity element (pure metal) to the matrix metal, and defined as: β' =
Ei pθ = θi E m pm
Mi Mm
(4)
where Ei and Em are the distillation velocity of impurity and matrix metal respectively, g/cm2/s, defined as θ θ E = 0.0583 ⋅ α ⋅ p M / T ; pi and pm are vapor pressure of impurity and matrix metal respectively, mm Hg; Mi and Mm are the atomic mass of impurity and matrix metal. 3.2 Theoretical analysis Compared with the physical process of solidification, in which process metal is transformed from liquid to solid state, and in distillation purification process, metal is transformed from liquid via gas to solid state, and there is no mass accumulation or loss in metal vapor, so that the metal vapor transforms to solid metal completely; therefore, it is concluded that the distillation purification process is coincident with that of the solidification essentially. In solidification process, the solute with equilibrium partition coefficient of k<1, its content in solid metal will be less than that in liquid metal; and in distillation purification process, the distillation velocity of impurity is smaller than matrix metal, the impurity content in distilled metal will be less than that in liquid metal, the impurity will be separated from matrix metal, and then the matrix metal can be purified. Therefore, the solidification theory can be used in distillation purification field to
927
discuss the behavior of impurity. In present experiment, the modified separation coefficients (β′) of Al and Fe in metallic dysprosium are 1/6.1 and 1/16.9[23,24] respectively. Under above conditions, if content of impurity at evaporating surface is C0, and it will equal to β′C0 in distilled metal, the remained impurity atoms can accumulate at evaporating surface, and the content of such impurity will increase gradually throughout the whole purification run, at the same time, a concentration gradient is formed in the vicinity of evaporating surface, impurity atoms will be driven to diffuse into the liquid metal. When the impurity in liquid metal is homogeneous resulted by convection mass transfer, and there is no concentration gradient of impurity near evaporating surface, which is equivalent to the condition of that soaking temperature is relatively lower, along with a small distillation velocity, and a longer distillation period is needed; and the impurity accumulated at evaporating surface has enough time to diffuse into liquid metal, and reaches an equilibrium state, the content in distilled metal can be described as follows: (5) C = β ' ⋅ C0 (1 − x / H ) (1/ β −1) '
The content of impurity in distilled metal calculated by Eq. (5) is demonstrated in Fig. 5. In initial stage of distillation, the content of impurity in distilled metal is lower, and then increases slowly, at last stage of distillation, it increases rapidly, even larger than the initial content of starting material. With modified separation coefficient decreasing, the content of impurity in distilled metal increases correspondingly, and the slope of content curve also presents an decreasing tendency, it means that the content of impurity in distilled metal is very lower in initial and middle stage, and the impurity can be separated more effectively. Compared with the impurities in metallic dysprosium, as seen in Fig. 4, the content curves of Al and Fe present a same tendency, the content of impurity in distilled metal is lower in initial stage, increases slowly in middle stage, and increases rapidly in last stage. Fig. 6 compares the dimensionless contents of impurities Al and Fe in distilled metal between experiment and
Fig. 5 Distribution of impurity in distilled metal
928
JOURNAL OF RARE EARTHS, Vol. 34, No. 9, Sep. 2016
Fig. 6 Comparison of dimensionless content of impurities Al (a) and Fe (b) between experiment and calculated
calculated results. As seen in Fig. 6(a) and (b), in the former stage of distillation experiment, the calculated content agrees well with experiment results, and in the latter stage of distillation experiment, the calculated result is less than experiment result. In fact, the vacuum distillation purification is a complex process, and many factors affect the distribution of impurity in distilled metal. The first influence factor is evaporating surface temperature, during the distillation process, a part of heat energy is dissipated at crucible bottom to the inner wall of the furnace, and it results in a temperature drop at evaporating surface, the thinner liquid metal in crucible, the bigger temperature drop at evaporating surface; because saturated vapor pressure is a function of temperature, the distillation velocity of matrix metal and impurities will drop at the same time, and the dimensionless parameter of 1/β′ will increase (as seen in Fig. 7 for impurity Al). In this circumstance, the experiment result of contents of Al and Fe should be less than the calculated results, however, the experiment result is larger than the calculated. It is summarized that the temperature drop at evaporating surface is the reason that the calculated result is higher than experiment, but not the reason that the calculated result is lower than experiment. The second influence factor is activity coefficient of impurity, because the elemental electro-negativity
Fig. 7 Modified separation coefficient Al impurity with matrix metal of dysprosium
values of Dy, Al and Fe are equal to 1.1, 1.5 and 1.8, there is an inter-atomic force between the impurity atom and matrix metallic atom, and the activity coefficient will be less than 1, and the calculated result will be less than experiment result; on the other hand, the larger electro-negativity difference, the stronger inter-atomic forces, so that the activity coefficient of Fe in liquid dysprosium will be smaller than Al, the dimensionless content (C/C0) of Fe is less than Al, which is consistent with Fig. 6(a) and (b). The third influence factor is impurity diffusion, diffusion coefficient of impurity in liquid dysprosium decreases with temperature decreasing, and the impurity diffusion cannot reach an equilibrium state or a quasi-equilibrium state, the content of impurity at evaporating surface is larger than that in the liquid metal. Consequently, the experimental result is larger than that by theoretical analyzed result. The forth influence factor is viscosity of liquid metal, with heat energy dissipated at crucible bottom, the temperature of liquid metal decreases correspondingly, the viscosity of liquid metal increases greatly and the mobility of liquid metal turns weaker (metallic dysprosium is heated by resistance heating in present experiment, rather than electromagnetic induction heating), therefore, the diffusion resistance of impurities in the liquid metal increases correspondingly, it also results in the accumulation of the impurities near the evaporating surface, and if the heat loss is very serious, the liquid metallic dysprosium will be transformed into solid state. In addition, there are other influence factors, for example, entrainment, density, etc. In the vacuum distillation purification, if the soaking temperature is very high, and distillation velocity is very large, the impurity atom will be entrained by matrix metallic vapor, however, in present experiment, the soaking temperature is 93 ºC over melting point of dysprosium, and as the distillation continues, the evaporating surface temperature decreases gradually, so that the entrainment of impurity may not happen. The involved metal densities are following: Dy, 8550 kg/m3; Al, 2261 kg/m3 (1100 ºC); Fe, 6900 kg/m3
ZHANG Xiaowei et al., Impurity distribution in metallic dysprosium during distillation purification
(1535 ºC); Cu, 8300 kg/m3 (liquid). In liquid dysprosium, the buoyancy of Al atom is 3.57 times than that of Fe atom, so that more Al atoms may float up to evaporating surface, and the content of Al increases correspondingly, but the dimensionless content (C/C0) of Al impurity is not significantly greater than Fe (as seen in Fig. 6(a) and (b)), consequently, the density of impurity and matrix metal is not a significant influence factor. Based on above discussion, it is concluded that, in the former stage of distillation process, the diffusion of impurity in liquid metal can reach a quasi-equilibrium state, and the impurity is almost homogeneous, the content of impurity in distilled metal can be described by Eq. (5); in latter stage of distillation process, the diffusion rate of impurity in liquid metal decrease greatly due to temperature decrease, and the content of impurity at evaporating surface is larger than that in the liquid, consequently, the content in distilled metal is larger than the calculated result.
4 Conclusions (1) During the distillation purification of metallic dysprosium, the content of impurity in distilled metal, such as Al and Fe, was lower in initial stage, increased slowly in middle stage, and increased rapidly in last stage. (2) The separation coefficient of impurity could not be calculated accurately by the content of impurity in distilled metal because the content of impurity was changed during the distillation process, a modified separation coefficient was proposed, and the modified separation coefficient of impurity Al and Fe was 1/6.1 and 1/16.9 respectively, and these impurities, with modified separation coefficient less than 1, could be removed in distillation purification; (3) The physical process of the distillation was coincident with that of the solidification essentially. In the former stage of distillation purification, the diffusion of impurity in liquid metal could reach a quasi-equilibrium state, the calculated result of impurity content in distilled metal by solidification theory agreed well with experiments, and in latter stage of distillation process, the diffusion rate of impurity in liquid metal decreased, and the impurity content in distilled metal was larger than the calculated result.
References: [1] Hao Z Z, Jiang Y J, Yang J Y, Dai Q, Zhang X Q, Sun Z G, Gao J M, Han Z. Measurement of average distillation rate of Dy, Ho and Er metals. Chin. Rare Earths (in Chin.), 1999, 20(5): 65. [2] Gschneidner K A Jr, Pecharsky A O, Pecharsky V K. Method of Making Active Magentic Refrigerant, Colossal Magnetostriction and Giant Magentoresistive Materials
929
Based on Gd-Si-Ge Alloys. US Patent: US 6589366 B1, 2003. [3] Zhang T B, Chen Y G, Teng B H, Tang Y B, Fu H, Tu M J. The structure and magnetocaloric effect of rapidly quenched Gd5Si2Ge2 alloy with low-purity gadolinium. Mater. Lett., 2007, 61(2): 440. [4] Macaluso R, Nakatsuji S, Kuga K, Thomas E L, Machida Y, Maeno Y, Fisk Z, Chan J Y. Crystal structure and physical properties of polymorphs of LnAlB4 (Ln=Yb, Lu). Chem. Mater., 2007, 19(8): 1918. [5] Matsumoto Y, Nakatsuji S, Kuga K, Karaki Y, Horie N, Shimura Y, Sakakibara T, Nevidomskyy A H, Coleman P. Quantum criticality without tuning in the mixed valence compound β-YbAlB4. Science, 2011, 331: 316. [6] Nakatsuji S, Kuga K, Machida Y, Tayama T, Sakakibara T, Karaki Y, Ishimoto H, Yonezawa S, Maeno Y, Pearson E, Lonzarich G G, Balicas L, Lee H, Fisk Z. Superconductivity and quantum criticality in the heavy-fermion system β-YbAlB4. Nat. Phys., 2008, 4: 603. [7] Pang S M, Wang Z Q, Zhou L, Chen B Y, Xu L H, Zhao B, Yan S L, Li Z A. Study on preparation of high-purified terbium and dysprosium metals used for rare earth giant magnetostrictive materials. Chin. Rare Earths (in Chin.), 2008, 29(6): 31. [8] Yang Q S, Liu S P, Ouyang Y P, Liu S Z, Su Z F. Study on production technology of high-purity terbium metal. Jiangxi Nonferrous Metals (in Chin.), 2005, 19(1): 31. [9] Cheng W, Li Z A, Pang S M, Wang Z Q, Wang X S. Preparation of high purity lanthanum by combined method of lithium-thermal reduction and vacuum distillation. Chin. J. Rare Metals (in Chin.), 2011, 35(5): 781. [10] Su Z F. Purifying Sc metal in the method of vacuum distilling. Jiangxi Nonferrous Metals (in Chin.), 2002, 16(2): 13. [11] Liu B Z. Preparation of high purity metal dysprosium. Rare Metals Lett. (in Chin.), 2008, 27(7): 39. [12] Zhang Z Q, Wang Z Q, Miao R Y, Zhu Q, Chen D H, Zhang X W, Zhou L, Li Z A, Yan S L. Purification of yttrium to 4N5+ purity. Vacuum, 2014, 107: 77. [13] Zhang Z Q, Wang Z Q, Chen D H, Miao R Y, Zhu Q, Zhang X W, Zhou L, Li Z A. Purification of praseodymium to 4N5+ purity. Vacuum, 2014, 102: 67. [14] Li Z A, Zhang W, Xu J, Zhao B, Xiao F, Zhang Y. Study on the preparing process of high purity terbium metal. Chin. Rare Earths (in Chin.), 2002, 23(6): 36. [15] Jiang Y J, Hao Z Z, Dai Q, Yang J Y, Zhang X Q, Qiu D M. Measurement and theoretical computation of the distillation rate of metallic terbium. Chin. Rare Earths (in Chin.), 1999, 20(6): 11. [16] Zhang X Q, Jiang Y J, Dai Q. Preparation of highpurity matallic gadolinium by the vacuum distillation method. Science & Technology of Baotou Steel (Group) Corporation (in Chin.), 2007, 33(3): 17. [17] Jiang Y J, Hao Z Z, Zhang X Q, Dai Q. Discussion on the factors affecting purification of rare earth metals by vacuum distillation. Chin. Rare Earths (in Chin.), 2003, 24(4): 60. [18] Li G D, Liu Y L. Study on technology and optimization of purifying scandium metal by vacuum distillation method. J.
930 Chin. Soc. Rare Earths (in Chin.), 2000, 18(2): 183. [19] Xi B, Yu L P. The impure behavior of terbium in the distillation process. Jiangxi Metallurgy (in Chin.), 2001, 21(1): 27. [20] Li G D, Liu J L. The study on the new technology and purification mechanism in the preparation high purity scandium metal. Acta Scientiarum Naturalium Universitatis Neimongol (in Chin.), 1997, 28(5): 614. [21] Pang S M, Chen D H, Li Z A, Yan S H, Zhou L, Xu L H, Wang Z Q, Xu J L. Theory and technology of vacuum dis-
JOURNAL OF RARE EARTHS, Vol. 34, No. 9, Sep. 2016 tillation method for preparing high-purity metal neodymium. J. Chin. Soc. Rare Earths (in Chin.), 2013, 31(1): 14. [22] Zaiour A, Zahraman K, Roumie M, Charara J, Fawaz A, Lmai F, Hage-Ali M. Purification of tellurium to nearly 7N purity. Mater. Sci. Eng. B, 2006, 131: 54. [23] Hansen S, Getchius J, Steward R, Brumleve T. Vapor Pressure of Metal Bromides and Iodides, 2nd Edition. Urbana, Illinois: APL Engineered Materials, Inc. 2006. [24] David R L. CRC Handbook of Chemistry and Physics 85th. Boca Raton: CRC Press, 2000.
Impurity distribution in metallic dysprosium during distillation purification ZHANG Xiaowei (张小伟)1,*, MIAO Ruiying (苗睿瑛)1, LI Chuanjun (李传军)2, WU Daogao (吴道高)1, YAN Huan (闫 缓)1, WANG Zhiqiang (王志强)1, CHEN Dehong (陈德宏)1, YAN Shihong (颜世宏)1, LI Zongan (李宗安)1 (1. National Engineering Research Center for Rare Earth Materials, General Research Institute for Nonferrous Metals, and Grirem Advanced Materials Co. Ltd., Beijing 100088, China; 2. Otto Schott Institute of Materials Research, Friedrich Schiller University Jena, Löbdergraben 32, D-07743 Jena, Germany) Received 29 January 2016; revised 20 May 2016
Abstract: The distribution rules of impurities contents in distilled metallic dysprosium were researched, and a theoretical analysis was carried out. The research results indicated that, the content of impurity in distilled metal, such as Al and Fe, was lower in the initial stage, increased slowly in the middle stage, and increased rapidly in the last stage during the process of distillation purification. The calculated method of separation coefficient of impurity in crude metal by content of impurity in distilled metal was not suitable for high pure metals, and the modified separation coefficient was proposed, and it equaled 1/6.1 and 1/16.9 for impurity Al and Fe. The physical process of distillation was coincident with that of solidification essentially, and solute re-distribution theory in solidifying front could be used to describe the impurity distribution near evaporating surface. In the former stage of distillation purification, the diffusion of impurity in liquid metal could reach a quasi-equilibrium state, the calculated result of impurity content in distilled metal agreed well with experiments. In the latter stage of distillation process, the diffusion rate of impurity in liquid metal decreased, and the content in distilled metal was larger than the calculated result. Keywords: vacuum distillation purification; distilled dysprosium; impurity distribution; modified separation coefficient; rare earths
High-purity rare earth metals have been widely used in high technique field, which is a key raw material of high-performance magnetostrictive material[1], magnetic refrigerant materials[2,3], superconducting materials[4–6], etc.. Vacuum distillation is an important method for purifying rare earth metals. Vacuum distillation is an important method for purifying rare earth metals, in this purification process, the impurities, with lower saturated vapor pressure than the matrix metal, can be separated at high temperature environment in vacuum distillation purification process. The impurities, with lower saturated vapor pressure than the matrix metal, can be separated at high temperature environment in vacuum distillation purification process. Up to now, most previous studies focused on preparation of rare earth metals by vacuum distillation technology[7–14], measurement of distillation velocity[1,7,15,16], the effects of experimental conditions on purification[17,18], and few researchers used the separation coefficient to judge whether the impurity can be separated from the matrix metal[18–22]. Xi et al.[19] discussed the behavior of impurities in distillation purification of metallic terbium, the impurities with the separation coefficient βi>1 (defined as βi≡γipiθ/γmpmθ) can be evaporated preferentially, the impurities with the separation coefficient βi<1 will be remained mostly in the crucible, and the impurities with
βi≈1 cannot be separated by vacuum distillation; Li et al.[18,20] considered that it is credible to use separation coefficient to judge whether the impurity can be separated or not, but due to the lack of the activity coefficient of impurity, βi cannot be calculated accurately by the definition formula, but actually it can be obtained by another equation of βi = α i M i / M m , αi is volatilization coefficient of impurity, determined by volatilization rate of impurity and matrix metal based experiment, and it is found that the impurities of Al, Cu, Cr and Co present a negative deviation in liquid scandium, and Ni and Si present a positive deviation; Pang et al.[21] calculated the volatile quantity of impurities in metallic neodymium, the theoretical quantity equated with the experimental results for the impurities with βi>1, and there is a large error for the impurities with βi>1; Zaiour et al.[22] discussed behavior of impurities in distillation process and the removal efficiency of submicron major impurities in tellurium can be characterized by studying the effective separation coefficient α, which is affected by both the evaporation rate and particle size. The above researches are all related on the separation coefficient to judge whether the impurities can be removed, however, the removal rate of the impurities and the distribution of impurities in distilled metal have not been studied. In present study, a vacuum distillation purification ex-
Foundation item: Project supported by National Natural Science Foundation of China (51504036), National Basic Research Program of China (2012CBA01207), National High Technology Research and Development Program of China (2011AA03A409) * Corresponding author: ZHANG Xiaowei (E-mail: [email protected]; Tel.: +86-10-82241180) DOI: 10.1016/S1002-0721(16)60116-3
ZHANG Xiaowei et al., Impurity distribution in metallic dysprosium during distillation purification
periment of metallic dysprosium was carried out, and the content of impurity was determined by Inductively Coupled Plasma Atomic Emission Spectrometry (ICP-AES), and the distribution rules of impurities were obtained, and the theoretical analysis of the distribution rules was discussed.
1 Experimental The self-preparation metallic dysprosium (Dy) was used in the present experiment, the starting material consists of 20 casting ingots, every ingot is about 250 g, and the total amount of metal is 4.99 kg; 6 samples were sampled and analyzed by ICP-AES, and the average contents of impurities are listed in Table 1, and the purity of metallic dysprosium is about 99.87 wt.%. The schematic of distillation equipment is shown in Fig. 1. The starting material was placed into a tungsten crucible with dimensions of 115 mm in external diameter, 95 mm in inner diameter and 225 mm in height; considering that the melting point of Dy is 1407 ºC, a soaking temperature of 1500 ºC was maintained for 8 h under a pressure of 10–5 Pa, the temperatures of crucible were measured by an infrared radiation thermometer (RATMR1SBSF, Reytek, USA) through the sight hole in the furnace. In this period, the metallic dysprosium evaporated from liquid metal surface and condensed at the cooled collector (tantalum sheet, consists of a cylinder and a cover, with thickness of 0.1 mm, about 50 g), leaving high melting point and low vapor pressure impurities at the bottom of tungsten crucible. In order to be convenient for discussing the distribution rule of the impurities, the height of collector was de-
925
termined by the total amount of starting material, the height of collector was calculated by following equation H=mDy/(ρDyπ(dEx/2)2), where mDy is mass of metallic dysprosium, kg; ρDy is density, 8550 kg/m3; dEx is external diameter of crucible, m; which could guarantee the lower surface of the distilled metal to be at tungsten upper edge level.
2 Results After distillation experiment, about 120 g grey powder was left in crucible bottom, and the total amount of deposited Dy and collector (tantalum sheet) was 4.92 kg, after machining off the collector, distilled metal was 4.86 kg, the longitudinal section of distilled metal is shown in Fig. 2, and the schematic diagrams in various distillation periods are indicated in Fig. 3. In initial stage of distillation purification, as seen in Fig. 3(a), the metallic vapor will condense on the inner surface of the cylinder and the cover sheet to form a thin metallic film; and then metallic atom will deposit on the surface of metallic film, seen in Fig. 3(b) and (c); as the distillation process continues, the evaporated metallic atom deposited on the cover sheet presents a truncated cone and deposited on the cylinder sheet presents a circular ring with a triangular section, as seen in Fig. 3(d). In order to discuss the distribution rules of impurities, five samples were taken along the axial line in distilled metal successively (as seen in Fig. 2) after the experiment finished, and the size of every sample is 10 mm×5 mm×5 mm; 3 samples were taken for each position, the sample were dissolved by nitric acid, the impurities were determined by ICP-AES (PerkinElmer Optima 8100), the
Table 1 Content of 24 impurities in metallic dysprosium (mg/kg) Y
La
Ce
Pr
Nd
Sm
Eu
Gd
Tb
<10
19
<10
<10
<10
<10
Yb
Lu
Mg
Al
Si
Ca
10
30
23
86
116
211
Ho
Er
<10
36
Cr
Mn
15
22
Tm
117
68
<10
<10
Fe
Co
Ni
Cu
45
29
18
189
Fig. 1 Schematic of distillation equipment of metallic dysprosium before (a) and after (b) distillation
926
JOURNAL OF RARE EARTHS, Vol. 34, No. 9, Sep. 2016
Fig. 2 Distilled metal in present experiment
mean contents of impurities are plotted in Fig. 4, and the standard deviation is listed in Table 2. It is found that the contents of impurities of Al and Fe present the same tendency. Taking Al impurity for example, in initial stage of distillation experiment, the content of Al in distilled metal is 13 mg/kg, as the distillation process continues, the content of Al increases to 20, 45 and 83 mg/kg, in last stage of purification experiment, the content of Al increases to 170 mg/kg. It is concluded that, in initial stage the content of impurity is lower, and then increases slowly, and in last stage of distillation experiment, the content of impurity increases rapidly.
3 Theoretical analysis and discussion 3.1 Modified separation coefficient In the distillation purification process, the separation coefficient of the impurity is considered as a criterion to judge whether the impurity can be removed, and it is de-
fined as the following form[13]: γ pθ (1) β i = i θi γ m pm where γi and γm are activity coefficients of impurity and matrix metal respectively, pθi and pθm are saturated vapor pressures of impurity and matrix metal respectively. For the distillation purification of rare earth metals, the purity of starting material is over 99.9 wt.% generally, and the activity coefficient of matrix metal is approximately 1, and the activity coefficient of impurity is usually determined by distillation experiment. The separation coefficient can be calculated by another formula in distillation separation of crude metal[18]: (2) βi = αi M i / M m where αi is volatilization coefficient, obtained by Yi= 100–100(1–Xm/100)αi [18]; Yi is volatilization rate of impurity, calculated by Yi=cidmd/cisms, cid and cis are contents of impurity in distilled metal and starting material, md and ms are mass of distilled metal and starting material, Xm is volatilization rate of matrix metal, and the above parameters are measured by experiment. Throught Combining Eqs. (1) and (2), the activity coefficient of impurity γi can be written: γ i = αi ⋅ γ m ⋅
pmθ piθ
Mi Mm
(3)
In the distillation separation of crude metal, the conTable 2 Standard deviation of the measurement results of content of impurities (mg/kg) Impurity
Sample 1
Sample 2
Sample 3
Sample 4
Sample 5
Al
0.82
1.25
1.63
1.70
2.16
Fe
1.25
0.94
1.70
1.25
2.49
Fig. 3 Schematic diagrams of distillation process
Fig. 4 Content of impurities Al (a) and Fe (b) in distilled dysprosium
ZHANG Xiaowei et al., Impurity distribution in metallic dysprosium during distillation purification
tent of impurity or component is about several percents to several tens percents, and the content of such impurity or component in distilled metal is almost unchanged throughout the whole distillation process; however, in the distillation purification for preparation of high purity metal, the purity of starting material is over 99 wt.% generally, the amount of impurities is very small, and each impurity is about hundreds of mg/kg, due to the difference of saturated vapor pressure between impurity and matrix metal, the content of impurity in distilled metal has great difference during the whole distillation process, several times or several tens of times difference between initial stage and last stage. In the present distillation experiment, the content of impurity is changed throughout the distillation process, taking Al impurity for example, it is 13 mg/kg in initial stage, and it is up to 170 mg/kg in last stage, and the difference is about 13 times. Consequently, the separation coefficient of impurity cannot be calculated accurately by above method through the content of impurity in distilled metal, and the above calculation method is not suitable for preparation of high-pure metal by vacuum distillation. In order to describe the separation degree of impurity to the matrix metal, a modified separation coefficient is introduced, which is the dimensionless ratio of distillation velocity of the impurity element (pure metal) to the matrix metal, and defined as: β' =
Ei pθ = θi E m pm
Mi Mm
(4)
where Ei and Em are the distillation velocity of impurity and matrix metal respectively, g/cm2/s, defined as θ θ E = 0.0583 ⋅ α ⋅ p M / T ; pi and pm are vapor pressure of impurity and matrix metal respectively, mm Hg; Mi and Mm are the atomic mass of impurity and matrix metal. 3.2 Theoretical analysis Compared with the physical process of solidification, in which process metal is transformed from liquid to solid state, and in distillation purification process, metal is transformed from liquid via gas to solid state, and there is no mass accumulation or loss in metal vapor, so that the metal vapor transforms to solid metal completely; therefore, it is concluded that the distillation purification process is coincident with that of the solidification essentially. In solidification process, the solute with equilibrium partition coefficient of k<1, its content in solid metal will be less than that in liquid metal; and in distillation purification process, the distillation velocity of impurity is smaller than matrix metal, the impurity content in distilled metal will be less than that in liquid metal, the impurity will be separated from matrix metal, and then the matrix metal can be purified. Therefore, the solidification theory can be used in distillation purification field to
927
discuss the behavior of impurity. In present experiment, the modified separation coefficients (β′) of Al and Fe in metallic dysprosium are 1/6.1 and 1/16.9[23,24] respectively. Under above conditions, if content of impurity at evaporating surface is C0, and it will equal to β′C0 in distilled metal, the remained impurity atoms can accumulate at evaporating surface, and the content of such impurity will increase gradually throughout the whole purification run, at the same time, a concentration gradient is formed in the vicinity of evaporating surface, impurity atoms will be driven to diffuse into the liquid metal. When the impurity in liquid metal is homogeneous resulted by convection mass transfer, and there is no concentration gradient of impurity near evaporating surface, which is equivalent to the condition of that soaking temperature is relatively lower, along with a small distillation velocity, and a longer distillation period is needed; and the impurity accumulated at evaporating surface has enough time to diffuse into liquid metal, and reaches an equilibrium state, the content in distilled metal can be described as follows: (5) C = β ' ⋅ C0 (1 − x / H ) (1/ β −1) '
The content of impurity in distilled metal calculated by Eq. (5) is demonstrated in Fig. 5. In initial stage of distillation, the content of impurity in distilled metal is lower, and then increases slowly, at last stage of distillation, it increases rapidly, even larger than the initial content of starting material. With modified separation coefficient decreasing, the content of impurity in distilled metal increases correspondingly, and the slope of content curve also presents an decreasing tendency, it means that the content of impurity in distilled metal is very lower in initial and middle stage, and the impurity can be separated more effectively. Compared with the impurities in metallic dysprosium, as seen in Fig. 4, the content curves of Al and Fe present a same tendency, the content of impurity in distilled metal is lower in initial stage, increases slowly in middle stage, and increases rapidly in last stage. Fig. 6 compares the dimensionless contents of impurities Al and Fe in distilled metal between experiment and
Fig. 5 Distribution of impurity in distilled metal
928
JOURNAL OF RARE EARTHS, Vol. 34, No. 9, Sep. 2016
Fig. 6 Comparison of dimensionless content of impurities Al (a) and Fe (b) between experiment and calculated
calculated results. As seen in Fig. 6(a) and (b), in the former stage of distillation experiment, the calculated content agrees well with experiment results, and in the latter stage of distillation experiment, the calculated result is less than experiment result. In fact, the vacuum distillation purification is a complex process, and many factors affect the distribution of impurity in distilled metal. The first influence factor is evaporating surface temperature, during the distillation process, a part of heat energy is dissipated at crucible bottom to the inner wall of the furnace, and it results in a temperature drop at evaporating surface, the thinner liquid metal in crucible, the bigger temperature drop at evaporating surface; because saturated vapor pressure is a function of temperature, the distillation velocity of matrix metal and impurities will drop at the same time, and the dimensionless parameter of 1/β′ will increase (as seen in Fig. 7 for impurity Al). In this circumstance, the experiment result of contents of Al and Fe should be less than the calculated results, however, the experiment result is larger than the calculated. It is summarized that the temperature drop at evaporating surface is the reason that the calculated result is higher than experiment, but not the reason that the calculated result is lower than experiment. The second influence factor is activity coefficient of impurity, because the elemental electro-negativity
Fig. 7 Modified separation coefficient Al impurity with matrix metal of dysprosium
values of Dy, Al and Fe are equal to 1.1, 1.5 and 1.8, there is an inter-atomic force between the impurity atom and matrix metallic atom, and the activity coefficient will be less than 1, and the calculated result will be less than experiment result; on the other hand, the larger electro-negativity difference, the stronger inter-atomic forces, so that the activity coefficient of Fe in liquid dysprosium will be smaller than Al, the dimensionless content (C/C0) of Fe is less than Al, which is consistent with Fig. 6(a) and (b). The third influence factor is impurity diffusion, diffusion coefficient of impurity in liquid dysprosium decreases with temperature decreasing, and the impurity diffusion cannot reach an equilibrium state or a quasi-equilibrium state, the content of impurity at evaporating surface is larger than that in the liquid metal. Consequently, the experimental result is larger than that by theoretical analyzed result. The forth influence factor is viscosity of liquid metal, with heat energy dissipated at crucible bottom, the temperature of liquid metal decreases correspondingly, the viscosity of liquid metal increases greatly and the mobility of liquid metal turns weaker (metallic dysprosium is heated by resistance heating in present experiment, rather than electromagnetic induction heating), therefore, the diffusion resistance of impurities in the liquid metal increases correspondingly, it also results in the accumulation of the impurities near the evaporating surface, and if the heat loss is very serious, the liquid metallic dysprosium will be transformed into solid state. In addition, there are other influence factors, for example, entrainment, density, etc. In the vacuum distillation purification, if the soaking temperature is very high, and distillation velocity is very large, the impurity atom will be entrained by matrix metallic vapor, however, in present experiment, the soaking temperature is 93 ºC over melting point of dysprosium, and as the distillation continues, the evaporating surface temperature decreases gradually, so that the entrainment of impurity may not happen. The involved metal densities are following: Dy, 8550 kg/m3; Al, 2261 kg/m3 (1100 ºC); Fe, 6900 kg/m3
ZHANG Xiaowei et al., Impurity distribution in metallic dysprosium during distillation purification
(1535 ºC); Cu, 8300 kg/m3 (liquid). In liquid dysprosium, the buoyancy of Al atom is 3.57 times than that of Fe atom, so that more Al atoms may float up to evaporating surface, and the content of Al increases correspondingly, but the dimensionless content (C/C0) of Al impurity is not significantly greater than Fe (as seen in Fig. 6(a) and (b)), consequently, the density of impurity and matrix metal is not a significant influence factor. Based on above discussion, it is concluded that, in the former stage of distillation process, the diffusion of impurity in liquid metal can reach a quasi-equilibrium state, and the impurity is almost homogeneous, the content of impurity in distilled metal can be described by Eq. (5); in latter stage of distillation process, the diffusion rate of impurity in liquid metal decrease greatly due to temperature decrease, and the content of impurity at evaporating surface is larger than that in the liquid, consequently, the content in distilled metal is larger than the calculated result.
4 Conclusions (1) During the distillation purification of metallic dysprosium, the content of impurity in distilled metal, such as Al and Fe, was lower in initial stage, increased slowly in middle stage, and increased rapidly in last stage. (2) The separation coefficient of impurity could not be calculated accurately by the content of impurity in distilled metal because the content of impurity was changed during the distillation process, a modified separation coefficient was proposed, and the modified separation coefficient of impurity Al and Fe was 1/6.1 and 1/16.9 respectively, and these impurities, with modified separation coefficient less than 1, could be removed in distillation purification; (3) The physical process of the distillation was coincident with that of the solidification essentially. In the former stage of distillation purification, the diffusion of impurity in liquid metal could reach a quasi-equilibrium state, the calculated result of impurity content in distilled metal by solidification theory agreed well with experiments, and in latter stage of distillation process, the diffusion rate of impurity in liquid metal decreased, and the impurity content in distilled metal was larger than the calculated result.
References: [1] Hao Z Z, Jiang Y J, Yang J Y, Dai Q, Zhang X Q, Sun Z G, Gao J M, Han Z. Measurement of average distillation rate of Dy, Ho and Er metals. Chin. Rare Earths (in Chin.), 1999, 20(5): 65. [2] Gschneidner K A Jr, Pecharsky A O, Pecharsky V K. Method of Making Active Magentic Refrigerant, Colossal Magnetostriction and Giant Magentoresistive Materials
929
Based on Gd-Si-Ge Alloys. US Patent: US 6589366 B1, 2003. [3] Zhang T B, Chen Y G, Teng B H, Tang Y B, Fu H, Tu M J. The structure and magnetocaloric effect of rapidly quenched Gd5Si2Ge2 alloy with low-purity gadolinium. Mater. Lett., 2007, 61(2): 440. [4] Macaluso R, Nakatsuji S, Kuga K, Thomas E L, Machida Y, Maeno Y, Fisk Z, Chan J Y. Crystal structure and physical properties of polymorphs of LnAlB4 (Ln=Yb, Lu). Chem. Mater., 2007, 19(8): 1918. [5] Matsumoto Y, Nakatsuji S, Kuga K, Karaki Y, Horie N, Shimura Y, Sakakibara T, Nevidomskyy A H, Coleman P. Quantum criticality without tuning in the mixed valence compound β-YbAlB4. Science, 2011, 331: 316. [6] Nakatsuji S, Kuga K, Machida Y, Tayama T, Sakakibara T, Karaki Y, Ishimoto H, Yonezawa S, Maeno Y, Pearson E, Lonzarich G G, Balicas L, Lee H, Fisk Z. Superconductivity and quantum criticality in the heavy-fermion system β-YbAlB4. Nat. Phys., 2008, 4: 603. [7] Pang S M, Wang Z Q, Zhou L, Chen B Y, Xu L H, Zhao B, Yan S L, Li Z A. Study on preparation of high-purified terbium and dysprosium metals used for rare earth giant magnetostrictive materials. Chin. Rare Earths (in Chin.), 2008, 29(6): 31. [8] Yang Q S, Liu S P, Ouyang Y P, Liu S Z, Su Z F. Study on production technology of high-purity terbium metal. Jiangxi Nonferrous Metals (in Chin.), 2005, 19(1): 31. [9] Cheng W, Li Z A, Pang S M, Wang Z Q, Wang X S. Preparation of high purity lanthanum by combined method of lithium-thermal reduction and vacuum distillation. Chin. J. Rare Metals (in Chin.), 2011, 35(5): 781. [10] Su Z F. Purifying Sc metal in the method of vacuum distilling. Jiangxi Nonferrous Metals (in Chin.), 2002, 16(2): 13. [11] Liu B Z. Preparation of high purity metal dysprosium. Rare Metals Lett. (in Chin.), 2008, 27(7): 39. [12] Zhang Z Q, Wang Z Q, Miao R Y, Zhu Q, Chen D H, Zhang X W, Zhou L, Li Z A, Yan S L. Purification of yttrium to 4N5+ purity. Vacuum, 2014, 107: 77. [13] Zhang Z Q, Wang Z Q, Chen D H, Miao R Y, Zhu Q, Zhang X W, Zhou L, Li Z A. Purification of praseodymium to 4N5+ purity. Vacuum, 2014, 102: 67. [14] Li Z A, Zhang W, Xu J, Zhao B, Xiao F, Zhang Y. Study on the preparing process of high purity terbium metal. Chin. Rare Earths (in Chin.), 2002, 23(6): 36. [15] Jiang Y J, Hao Z Z, Dai Q, Yang J Y, Zhang X Q, Qiu D M. Measurement and theoretical computation of the distillation rate of metallic terbium. Chin. Rare Earths (in Chin.), 1999, 20(6): 11. [16] Zhang X Q, Jiang Y J, Dai Q. Preparation of highpurity matallic gadolinium by the vacuum distillation method. Science & Technology of Baotou Steel (Group) Corporation (in Chin.), 2007, 33(3): 17. [17] Jiang Y J, Hao Z Z, Zhang X Q, Dai Q. Discussion on the factors affecting purification of rare earth metals by vacuum distillation. Chin. Rare Earths (in Chin.), 2003, 24(4): 60. [18] Li G D, Liu Y L. Study on technology and optimization of purifying scandium metal by vacuum distillation method. J.
930 Chin. Soc. Rare Earths (in Chin.), 2000, 18(2): 183. [19] Xi B, Yu L P. The impure behavior of terbium in the distillation process. Jiangxi Metallurgy (in Chin.), 2001, 21(1): 27. [20] Li G D, Liu J L. The study on the new technology and purification mechanism in the preparation high purity scandium metal. Acta Scientiarum Naturalium Universitatis Neimongol (in Chin.), 1997, 28(5): 614. [21] Pang S M, Chen D H, Li Z A, Yan S H, Zhou L, Xu L H, Wang Z Q, Xu J L. Theory and technology of vacuum dis-
JOURNAL OF RARE EARTHS, Vol. 34, No. 9, Sep. 2016 tillation method for preparing high-purity metal neodymium. J. Chin. Soc. Rare Earths (in Chin.), 2013, 31(1): 14. [22] Zaiour A, Zahraman K, Roumie M, Charara J, Fawaz A, Lmai F, Hage-Ali M. Purification of tellurium to nearly 7N purity. Mater. Sci. Eng. B, 2006, 131: 54. [23] Hansen S, Getchius J, Steward R, Brumleve T. Vapor Pressure of Metal Bromides and Iodides, 2nd Edition. Urbana, Illinois: APL Engineered Materials, Inc. 2006. [24] David R L. CRC Handbook of Chemistry and Physics 85th. Boca Raton: CRC Press, 2000.