【作者】 申淼；

【导师】 李冰；

【作者基本信息】 华东理工大学 ， 材料科学与工程， 2012， 博士

【摘要】 当前金属锂唯一的工业生产方法是氯化锂-氯化钾熔盐电解法。电解得到的产品纯度约为95-99wt%,含有K, Na, Al, Ca, Si, Fe, Mg, Ni等杂质。目前生产纯度在99.9wt%以上的金属锂是将电解得到的初级金属锂经两次真空蒸馏。真空蒸馏是在600-800℃下不锈钢蒸馏炉内将锂蒸发,其电能消耗大,蒸馏效率低,对设备产生严重腐蚀,增大了高纯金属锂的生产成本。本文针对现有技术的缺陷,拟采用电化学方法去除LiCl-KCl体系的微量杂质K,Al, Mg, Ca, Na等。主要研究内容包括：(1)采用电化学方法研究了LiCl-KCl(1:lmol)体系Li+与K+的还原电位差为0.3V,当电解质摩尔配比增加至3：7时,还原电位差减小至0.16V,当温度由450℃升高到600℃,Li+的还原电位正移0.15-0.2V。在电解过程中,通过控制电解质KCl与LiCl的比例、电解温度、电流密度可以降低金属Li中的杂质K含量。(2)通过电化学方法研究450℃时LiCl-KCl-AlCl3体系中,微量A13+的还原电位为-1.55Vvs.C12/C1-,与Li+的还原电位差为1.8V,满足电化学分离两种离子的条件。本文采用固态金属Fe和液态金属Zn作为阴极,在-1.6V恒电位电解10h后,Al3+的实际去除率分别为96.11%和99.9%,结果显示液态金属Zn更有利于去除微量杂质Al。通过Miedema模型计算,Al在Al-Fe合金中的活度系数小于1,在Al-Zn合金中的活度系数稍大于1,参照合金活度系数判断固态金属Fe更有利于去除杂质Al。但实验过程中,杂质Al的去除受合金动力学及热力学共同影响。由于Al3+的还原过程为受扩散控制的可逆过程,因此,在电解过程中添加氩气搅拌,在Fe阴极上恒电位电解4h后去除了99%以上的Al3+,这表明搅拌可以促进Al3+的扩散,加速杂质的去除。(3)通过电化学方法研究了LiCl-KCl-MgCl2体系中,Mg2+开始还原的电位为-2.7V,Li+欠电位还原从-2.95V开始,大量金属Li析出的电位为-3.5V,Mg2+与Li+(欠电位还原)的还原电位差为0.25V,与大量金属Li析出的电位相差0.8V,满足电化学分离的条件。结合二元合金相图及镁合金中组元镁的活度系数计算结果,提出采用液态金属Pb作为阴极去除微量杂质Mg。通过恒电位电解,LiCl-KCl体系中MgCl2含量越低,在相同时间内Mg2+的理论去除率越高；当体系中MgCl2含量相同,随着电解时间的增加,杂质Mg2+的理论去除率也相应增大。但阴极产物中有少量的元素Li。这说明在去除微量杂质Mg2+的过程中会有少量金属Li的析出。(4)通过能斯特方程计算了450℃时Ca2+与Li+的还原电位差仅为0.078V,不满足电化学分离Ca2+与Li+的条件；通过循环伏安测试了-LiCl(18mol%)-KCl(64mol%)体系中含有CaCl2(18mol%)时,Ca2+和Li+在W电极上的还原电位差为0.3V,与理论计算结果存在的差异。通过循环伏安法和方法伏安法研究得到,450℃时LiCl-KCl-CaCl2(0.5wt%)体系中W电极上Ca2+与Li+的还原电位差为0.11V。温度升高至680℃,还原电位差增加至0.20V；CaCl2含量增加至2.0wt%,该值没有明显改变。通过计时电位法研究,450和680℃时Li+和Ca2+在W电极上的还原电位差大于0.2V,满足W电极上电化学分离Ca2+和Li+的条件。采用Miedema模型计算得到Ca-Sn合金中Ca的活度系数较小。通过循环伏安法得到Ca2+与Li+在Sn电极上的还原电位差为0.6V,满足电化学分离条件。分别采用惰性电极W和活性电极Sn在-3.4V和-3.0V恒电位电解3h后,Ca2+的实际去除率分别为13.58%和55.48%,因此,Sn电极可以加速杂质Ca2+的去除。在搅拌条件下,Sn电极上在-3.0V电解3h和6h后,杂质Ca2+的实际去除率分别为83.65%和94.66%,搅拌可以明显提高Ca2+的电化学反应速率。-3.0V电解6h后,消耗的LiCl含量为杂质CaCl2的32.58%,在电解过程中消耗少量的LiCl是可以接受的。(5)通过能斯特方程计算,Na+还原电位较Li+负0.050V,理论上无法在金属Li电解之前去除杂质Na+。当LiCl-KCl体系中杂质NaCl的含量达到18mol%时,循环伏安法测得Na+和Li+的还原电位差大于0.2V,为电化学分离Na+和Li+提供了可能。通过电化学方法研究,当LiCl-KCl体系含有0.5-2.0wt%NaCl时,W电极上Li+和Na+的还原电位差为0.16-0.2V左右,不满足电化学分离的条件。采用Miedema模型计算得到,采用液态金属Pb作为阴极,Na在合金中的活度系数小于1,且形成合金的温度较低,有利于杂质Na的去除。通过循环伏安法研究,Pb电极上Na+和Li+的还原电位差达到0.2V,小于Ca2+和Li+在电极Sn上的还原电位差0.6V。在去除杂质Na+时,当电解电位由-2.6V负移至-3.0V,Na+的实际去除率由47.67%增大至75.71%。在-3.0V电解时,除了Na+的还原,Li+也会还原为相应的金属Li,通过消耗LiCl的方式可以提高杂质Na+的去除率。(6) LiCl-KCl-NaCl(2.0wt%)-CaCl2(2.0wt%)-MgCl2(2.0wt%)-AlCl3(2.0wt%)体系中比较了三叶桨、四叶桨、未添加搅拌桨的条件下,Pb电极上电解10h后,A13+、Mg2+、Ca2+、Na+各种杂质离子的去除率,实验结果表明,四叶桨更有利于杂质离子的均匀混合,提高杂质离子的去除率。同时也验证了FLUENT的数值计算结果。(7)通过Fluent数值计算,杂质离子在三叶桨和四叶桨搅拌槽内自由扩散的混合时间分别为13.764s和14.394s。与自由扩散相比,采用桨叶搅拌后,杂质离子混合均匀所需要的时间缩短了3个数量级。采用四叶桨搅拌所需要的混合时间在不同转速下均比三叶桨小,且混合时间的差别在转速为40rpm时最为明显,随着转速的增大,混合时间的差异减小。因此,采用四叶桨转速为40rpm,在极短的时间内就实现了杂质离子浓度的均匀化,这与实验得到的结论一致。这也说明了采用FLUENT中的组分运输模型可以对扩散现象进行较好的模拟。更多还原

【Abstract】 Currently, the only way in industry to produce metal Li is by electrowinning from the molten LiCl-KCl melt at about450℃. The obtained metal Li has the purity of98-99wt%with1-2wt%impurities including Na, K, Al, Ca, Mg, Si, etc, which are derived from the corresponding metal chloride compounds or metal oxide compounds existing in the LiCl and/or KC1raw materials. These impurities have to be removed by vacuum distillation in a stainless steel reactor under600-800℃, and sometimes followed by zone smelting. It consumes52kWh/kg-Li to refine the primary lithium from a purity of98.5to99.9wt%, accompanying with a serious corrosion to the reactor. It is necessary to explore a low-cost method to remove these impurities from the primary lithium.In this dissertation, the electrochemical methods have been employed to investigate the reduction mechanism and removal rate of race impurity ions in LiCl-KCl. Several obtained results including:(1) During the investigation of trace impurity K, the0.3V of reduction potential difference between Li+and K+in LiCl-KCl (1:1mol) have been obtained by electrochemical methods, which decreased to0.16V with the LiCl-KCl mol ratio increased to3:7. And the reduction potential of Li+was positively shifted0.15-0.2V with the temperature increasing from450to600℃. The content of impurity K can be reduced after electrolysis from lower mol ratio LiCl-KCl, lower temperature and current density.(2) In order to remove trace impurity AlCl3from LiCl-KCl melts before Li electrolysis, the Al+reduction potential on a tungsten electrode was determined by cyclic voltammetry (CV) and square wave voltammetry (SWV). The reduction potential difference between Al3+and Li+was1.8V, which satisfied the electrochemical removal of trace Al3+from LiCl-KCl melts. The constant potential electrolysis at-1.6V on both solid Fe and liquid Zn cathodes was performed to remove Al3+impurity from the LiCl-KCl-AlCl3melts. The results showed that96.11%of Al3+were removed on a Fe cathode and99.90%on a Zn cathode through10h electrolysis, respectively. The activity coefficient of Al in Al-Fe alloy was lower than1, while it was slightly higher than1in Al-Zn alloy, which means the Fe cathode may remove more impurity Al. The difference between theoretical calculation and the experiment was attributed to the thermodynamics and kinetics of alloy formation. While stirring the melts by argon gas,99.21%of Al3+was separated from the melts by4h electrolysis at450℃, which effectively expedited the Al3+electrochemical reduction rate and shortened the electrolysis time.(3) In order to remove trace impurity MgCl2from LiCl-KCl melts before Li electrolysis, MgCl2reduction processes in LiCl-KCl-MgCl2melt were investigated by cyclic voltammetry (CV), square wave voltammetry (SWV), chronoamperometry (CP) and chronopotentiometry (CA). The results showed that Mg2+was reduced in one step with two-electron transfer and its reduction potential was well defined. The reduction potential difference of Mg2+with Li+underpotential deposition was0.25V; therefore, in order to prevent Li+from codeposition with Mg, accurately controlling Mg2+reduction potential during the potentiostatic electrolysis was necessary. The liquid lead was recommended as cathode to remove trace impurity Mg2+according to the binary alloy phase diagram and magnesium activity coefficient calculation. Then potentiostatic electrolysis was carried out and Mg+was reduced onto a liquid lead cathode from the LiCl-KCl-MgCl2melt at the Mg deposition potential, which showed the Mg2+theoretical removal rates increased with the decreasing of MgCl2concentration and the extending of electrolysis time. About93-99%of removal rate of MgCl2was achieved after8-12h electrolysis when liquid lead was used as cathode. But a small amount of Li was found in cathode, which means the Li+codeposition with Mg during electrolysis.(4) In order to investigate the possibility of electroseparation of CaCl2from CaCl2-LiCl-KCl melts prior to Li+reduction, the theoretical reduction potential difference of0.078V between Ca2+and Li+ions was calculated by Nernst equation. But this potential difference was0.3V by cyclic voltammetry (CV) study in CaCl2(18mol%)-LiCl(18mol%)-KCl(64mol%) system. The potential difference between Ca2+and Li+in CaCl2(0.5wt%)-LiCl-KCl melts was0.2V at680℃analyzed by CV. More than0.2V reduction potential difference between Ca2+and Li+ions was observed on chronopotentiograms (CP) recorded on a tungsten electrode in CaCl2(2.0wt%)-LiCl-KCl melts either at450-680℃, which met the requirements of total electroseparation. Tin electrode can shift the reduction potentials of Ca2+and Li+to more positive ones and can improve their reduction potential difference to0.6V. After3h electrolysis on solid tungsten at-3.4V and liquid tin cathode at-3.0V, the Ca/Li (weight ratio) in the melts changed from3/25before electrolysis to2.6/25and0.76/25respectively, which tin cathode improved. Ca2+removal rate. Under stirring, even higher removal rates of83.65%and94.66%, corresponding to0.54/25and0.16/25of Ca/Li have been obtained for3h and6h electrolysis at-3.0V on a tin electrode. Clearly, stirring availably increases the Ca2+ions diffusion rate and its electrochemical reaction rate.(5) In order to investigate the possibility of electroseparation of NaCl from NaCl-LiCl-KCl melts prior to Li+reduction,-0.05V of theoretical reduction potential difference between Na+and Li+ions was calculated by Nemst equation. But this potential difference was higher than0.2V in LiCl(18mol%)-KCl(64mol%) with NaCl(18mol%) system studied by cyclic voltammetry (CV), which provided the possibility for electrochemical removal. The potential difference between Na+and Li+was0.16-0.2V on a tungsten electrode in LiCl-KCl with0.5-2.0wt%NaCl at450℃analyzed by cyclic voltammetry (CV), square wave voltammetry (SWV) and chronopotentiograms (CP). The liquid lead was recommended as cathode to remove trace impurity Na+. About0.2V reduction potential difference between Na+and Li+was observed on CV recorded on a lead electrode at450℃in LiCl-KCl-NaCl(2.0wt%), which was smaller than Ca2+and Li+potential difference on tin electrode. During the process of Na+impurity removal, the practical removal rates of Na+were increased from47.67%to75.71%with the electrolysis potential shifted from-2.6V to-3.0V. Except for Na+reduction, some Li+were reduced to Li at-3.0V for consumption of LiCl. With the electrolysis time increased to6h, the Na+impurity removal rate was reached to90.12%, meanwhile, the Na/Li (mass ratio) increased to4.25/1, which further proved the metal Li promoting Na+removal rate.(6) Comparing with the condition of three leaf blade impeller and without impeller electrolysis, more impurity removal rates have been achieved by electrolysis in LiCl-KCl-NaCl(2.0wt%)-CaCl2(2.0wt%)-MgCl2(2.0wt%)-AlCl3(2.0wt%) system on a lead cathode for10h with the four leaf blade impeller. The uniform mixing time of impurity ions free diffusion were13.764s and14.394s respectively in three and four leaf blade electrolytic cell by Fluent numerical simulation. Comparing with impurity ions free diffusion, the impellers could improve the ions diffusion and shorten mixing time to3orders of magnitude. And the mixing time for four leaf blade impeller was smaller than the three leaf blade impeller; but this superiority gradually decreased with the rotating speed increasing. Therefore, the four leaf blade with40rpm can shorten the mixing time for impurity ions from bulk solution to the electrode surface solution, which was in accordance with experimental results. The Fluent numerical simulation also showed the components transportation model can be employed to simulate the ions diffusion during electrochemical reaction.更多还原