materials

Article Preparation of 6N,7N High-Purity Gallium by : Process Optimization

Jianfeng Hou 1,2, Kefeng Pan 2,* and Xihan Tan 2 1 Department of Chemistry and Chemical Engineering, Lvliang University, Lvliang 033001, China 2 School of Metallurgy, Northeastern University, Shenyang 110819, China * Correspondence: [email protected]; Tel.: +86-1307-063-9976

 Received: 16 July 2019; Accepted: 8 August 2019; Published: 10 August 2019 

Abstract: In this study, radial crystallization purification method under induction was proposed for preparing 6N,7N ultra-high purity gallium seed. The effect of cooling temperature on the morphology of the crystal seed, as well as the cooling water temperature, flow rate, and the addition amount of crystal seed on the crystallization process was explored, and the best purification process parameters were obtained as follows: temperature of the crystal seed preparation, 278 K; temperature and flow rate of the cooling water, 293 K and 40 L h 1, respectively; and number of added crystal · − seed, six. The effects of temperature and flow rate of the cooling water on the crystallization rate were investigated. The crystallization rate decreased linearly with increasing cooling water temperature, but increased exponentially with increasing cooling water flow. The governing equation of the crystallization rate was experimentally determined, and three purification schemes were proposed. When 4N crude gallium was purified by Scheme I, 6N high-purity gallium was obtained, and 7N high-purity gallium was obtained by Schemes II and III. The purity of high-purity gallium prepared by the three Schemes I, II, and III was 99.999987%, 99.9999958%, and 99.9999958%, respectively.

Keywords: gallium; high-purity; crystallization; process optimization

1. Introduction In the 1970s, compounds comprising gallium and Group IIIA elements were discovered to have excellent semiconductor properties. Since then, gallium (Ga) has been gradually used in the semiconductor industry as a raw material [1]. In recent years, with the continuous development of science and technology and people’s pursuit of low-carbon economy and green energy, the application of Ga has been fully developed and it has become one of the important raw materials in the fields of modern semiconductors (approximately 80% of the total consumption of gallium), solar energy (approximately 10% of the total gallium consumption, magnetic materials (~5% of the total gallium consumption), and catalysts, and has been widely applied in defense, optical fiber communication, aerospace, and other fields [2,3]. At present, the production technology of low-grade gallium (purity 99.99%) has been perfected ≤ day by day [4–6]. The statistics from the US Geological Survey (USGS) in 2018 showed that [7] the amount of low-grade primary gallium production around the world in 2017 was ~315 tons, which increased by 15% compared to the amount in 2016 with 274 tons. According to the survey report of USGC in 2015 [8], the global demand for Ga was estimated to increase by 20 times in 2030, whereas with the development of semiconductor devices with higher performance, the demand of high impurity gallium has been increasing, because even very small amounts of impurities such as Cu, Pb, Fe, Mg, Zn, and Cr, which are always present in current large-scale commercial-quality gallium, can degrade or limit the electrical properties [9]. Conventional refining methods such as electrolytic refining [10,11], regional melting [12], vacuum distillation [13], and drawn method [14] have been applied

Materials 2019, 12, 2549; doi:10.3390/ma12162549 www.mdpi.com/journal/materials Materials 2019, 12, 2549 2 of 11 for the preparation of high-purity gallium, and the electrolytic refining method is the most widely used high-purity gallium production technology in industry at present. However, these traditional methods had many problems such as high energy consumption, lack of environmental friendliness, low production efficiency, and inconvenient automation control. Therefore, developing an advanced purification technology is of great significance to the development of the contemporary semiconductor and the solar industry. The purification and refining of gallium have been systematically investigated by us [15–17]. Based on the traditional crystallization purification method, a radial crystallization purification method was proposed by the seed crystal induced crystallization. The method had the advantages of low energy consumption, simple equipment, convenient operation, and short production cycle. In this study, the crystallization experiment was used to explore the effect of cooling temperature on the morphology of the crystal and the effect of cooling water temperature, flow rate, and the amount of seed crystal added on the crystallization process. The parameters of the purification process were explored in order to optimize the best purification process, determine the crystallization rate control equation, and prepare high-purity (6N and 7N) metal gallium under the process conditions.

2. Materials and Methods

2.1. Process Design Figure1a shows the preparation process of 6N,7N high-purity gallium by radial crystal purification method through seed induction. The main steps and operation methods are as follows: (1) Cleaning the crystallizer and assembling the purification device First, the crystallizer was rinsed with ultrapure water (resistivity 16 MΩ cm) to remove the dust ≥ · on the surface. Then, it was cleaned using an ultrasonic cleaning device containing ultrapure water for 2 h to remove residual contaminants on the surface. The purification device was assembled, as shown in Figure1b. (2) Pretreatment of 4N crude gallium 4N crude gallium along with the package bottle was placed on a hot plate, and the heating temperature was set to 335 K. After the gallium was melted, the molten crude gallium was transferred to a polytetrafluoroethylene beaker and mixed with 200 mL of 3 mol/L HCl at 335 K for 2 h. The hydrochloric acid was suctioned out using a plastic pipette, and then 200 mL of 3 mol/L HNO3 was added to the beaker, followed by mixing and stirring for 2 h. The crude gallium was washed with acid, followed by washing with ultrapure water three times. The hydrochloric acid and nitric acid used in the acid treatment were all high purity grades, and ultrapure water was used for preparation of acid solution. (3) The pretreated molten crude gallium (4N) was transferred into a clean crystallizer. (4) Circulating cooling water was introduced into the water jacket of the crystallizer. The cooling water was supplied using a low constant temperature water tank with a built-in circulating water pump. The temperature range was in the range 263–373 K, the temperature control accuracy was 0.1 K, and the flow rate of the cooling water was controlled using a glass rotor flow meter. ± (5) When the temperature of the liquid gallium dropped to the critical point of crystallization, crystal seed was added thereto, and the cooling water was circulated. The seed crystal was prepared before the start of the purification experiment, using 7N gallium as the raw material. The method used is as follows: a polytetrafluoroethylene beaker containing molten 7N gallium was placed in a low constant temperature water bath to cool and crystallize. Liquid gallium was continuously stirred with a Teflon rod to disperse the crystal nucleus and improve the nucleation rate. The crystallization of liquid gallium was observed in the stirring process. When the grain with the desired size (0.5 cm) was formed, PTFE tweezers were used to pick out the crystal for later use. (6) When the liquid gallium was crystallized to a preset crystallization ratio, the introduction of cooling water was stopped, and the residual liquid gallium was discharged out of the crystallizer. MaterialsMaterials 20192019,, 1212,, x 2549 FOR PEER REVIEW 3 3of of 11 11

(7) A three-way switch was switched, and the circulating hot water was introduced into the cooling/melting(7) A three-way zone switchof the crystallizer. was switched, After and th thee solid circulating gallium hotcompletely water was melted, introduced the three-way into the switchcooling was/melting switched, zone ofand the the crystallizer. circulating Aftercooling the water solid was gallium reintroduced completely into melted, the cooling/melting the three-way zone;switch was switched, and the circulating cooling water was reintroduced into the cooling/melting zone; TheThe hot hot water water was was supplied supplied using using a a constant constant temperature temperature water water tank tank with with a a built-in built-in circulating circulating water pump. The temperature range was 278–373 K, and the temperature control accuracy was 1 K. water pump. The temperature range was 278–373 K, and the temperature control accuracy was ±1± K. (8)(8) The The steps steps (4)–(7) (4)–(7) (as (as shown shown in in Figure Figure 11c)c) werewere repeatedrepeated toto aa predeterminedpredetermined numbernumber ofof ,crystallizations, and after the completion of purification, purification, the productproduct quality was detected.

FigureFigure 1. 1. SchematicSchematic diagram diagram of of the the purification purification process: process: ( (aa)) Process Process flow flow for for purification, purification, ( (bb)) assembly assembly ofof the the purification purification device, and ( c) repeated crystallization process.

2.2.2.2. Detection Detection Method Method InIn the the experiment, experiment, the the impurity impurity contents contents of of the the 4N 4N raw raw gallium gallium material material and and the the purified purified 6N,7N 6N,7N high-purityhigh-purity gallium gallium were were detected by high-resolut high-resolutionion glow discharge mass spectrometry (Evans MaterialsMaterials TechnologyTechnology (Shanghai) (Shanghai) Co., Co., China, China, HR-GDMS), HR-GDMS), and theand purity the ofpurity the product of the wasproduct calculated was calculatedby the diff erenceby the method.difference Argon method. was Argon used as was the dischargeused as the gas discharge for detection. gas for The detection. mass spectrometry The mass spectrometryparameters are parameters as follows: are discharge as follows: current, discharge 1.9 mA; current, discharge 1.9 voltage, mA; discharge 1 kV; beam voltage, current 1 ofkV; gallium beam ion, 1 10 6 mA; insulating layer, aluminum; and resolution 3600. Before data collection, the current× of gallium− ion, 1 × 10−6 mA; insulating layer, aluminum; ≥and resolution ≥ 3600. Before data collection,ion source the of theion HR-GDMSsource of the was HR-GDMS cooled to was the cooled temperature to the temperature of liquid nitrogen of liquid (90 nitrogen K) to reduce (90 K) the to ion interference in the background gas. Then, the surface of the tested sample (0.2 2 mm2) was reduce the ion interference in the background gas. Then, the surface of the tested sample× (0.2 × 2 mm2) waspre-sputtered pre-sputtered for 5 for min 5 atmin liquid at liquid nitrogen nitrogen temperature temperature to remove to remove the contaminants the contaminants from the from sample the samplesurface. surface. The pre-sputtering The pre-sputtering conditions conditions were kept were constant, kept andconstant, data collectionand data wascollection started. was During started. the Duringdata collection the data process, collection the process, integration the integration time was set time as 80 was ms. set as 80 ms. 3. Results and Discussion 3. Results and Discussion 3.1. Effect of Cooling Temperature on Seed Morphology 3.1. Effect of Cooling Temperature on Seed Morphology The appearance morphology of the seed crystal prepared at the cooling temperature in the The appearance morphology of the seed crystal prepared at the cooling temperature in the range range 265–295 K is shown in Figure2, indicating that at 265 K, the solidified structure comprised 265–295 K is shown in Figure 2, indicating that at 265 K, the solidified structure comprised many fine many fine crystal grains, and the grains were interspersed with a large amount of liquid gallium. crystal grains, and the grains were interspersed with a large amount of liquid gallium. When the When the solidified structure was removed, a large amount of liquid gallium was attached to the solidified structure was removed, a large amount of liquid gallium was attached to the surface, surface, resulting in an extremely irregular shape of crystal seed, because at 265 K, the growth rate resulting in an extremely irregular shape of crystal seed, because at 265 K, the growth rate of crystal of crystal nucleus increased after nucleation due to the large degree of supercooling, leading to the nucleus increased after nucleation due to the large degree of supercooling, leading to the emergence emergence of a large number of dendrites. The rapid growth of dendrites not only mixed with of a large number of dendrites. The rapid growth of dendrites not only mixed with a lot of liquid phase inside the solidified structure, but also caused a lot of hollow surface of the solidified structure. Materials 2019, 12, 2549 4 of 11

Materials 2019, 12, x FOR PEER REVIEW 4 of 11 a lot of liquid phase inside the solidified structure, but also caused a lot of hollow surface of the Whensolidified the structure.preparation When temperature the preparation was 273 temperature K, the solidified was 273 K,structure the solidified exhibited structure the geometric exhibited polyhedronthe geometric shape polyhedron characteristics, shape characteristics,indicating that with indicating decreasing that withsubcooling decreasing degree, subcooling the growth degree, rate ofthe nucleus rate of decreased, crystal nucleus and its decreased,growth mode and tr itsansited growth from mode dendrite transited growth from to lamellar dendrite growth. growth Whento lamellar the preparation growth. When temperature the preparation was 278 temperature K, the solidified was 278structure K, the showed solidified a regular structure hexahedral showed a shape,regular suggesting hexahedral that shape, with suggesting increasing thattemperatur with increasinge, the subcooling temperature, degree the of subcoolingthe growth degreefront further of the reducedgrowth frontafter furtherthe formation reduced ofafter crystal the nucleus, formation and of the crystal growth nucleus, mode and changed the growth into vertical mode changedlayered growth.into vertical When layered the preparation growth. When temperature the preparation was 295 temperature K, the supercooling was 295 K, degree the supercooling of the solid–liquid degree of interfacethe solid–liquid further interfacedecreased further after the decreased crystal nucleus after the fo crystalrmed, nucleushindering formed, the release hindering of latent the heat release from of crystallization.latent heat from At crystallization. this time, in order At this to time,release in the order latent to release heat of the crystallization latent heat of at crystallization a faster rate, atthe a growthfaster rate, direction the growth of crystal direction nucleus of crystal changed nucleus to side changed growth, to side distorting growth, its distorting geometric its shape. geometric By comparingshape. By comparingthe morphological the morphological characteristics characteristics of crystal of seed crystal prepared seed prepared at four at temperatures, four temperatures, the optimalthe optimal preparation preparation temperature temperature of the of crystal the crystal seed seed was was finally finally determined determined to be to 278 be 278K. K.

FigureFigure 2. 2. MorphologyMorphology of ofthe the seed seed crystals prepared prepared at different at different cooling cooling temperatures temperatures ((a) ((265a) 265K, (b K,) 273(b) 273K, (c K,) 278 (c) 278K, and K, and(d) 295 (d) 295K). K). 3.2. Effect of Process Parameters on the Crystallization Process 3.2. Effect of Process Parameters on the Crystallization Process 3.2.1. Effect of Temperature of Cooling Water on Crystallization Process 3.2.1. Effect of Temperature of Cooling Water on Crystallization Process 1 When the flow rate of cooling water was 40 L h− and the temperature was in the range When the flow rate of cooling water was 40 L·h−1 and· the temperature was in the range 288–298 288–298 K, 2.9774 kg crude gallium obtained by pickling pretreatment was cooled to the critical point K, 2.9774 kg crude gallium obtained by pickling pretreatment was cooled to the critical point of of crystallization, followed by adding crystal seeds for 15 min. The corresponding crystal growth crystallization, followed by adding crystal seeds for 15 min. The corresponding crystal growth morphology is shown in Figure3. morphology is shown in Figure 3. Figure3 shows that when the temperature of the cooling water were 288 and 290 K, the crystal growth mode of liquid gallium was mainly dendrite growth after adding the crystal seed, and the crystal branches bridged with each other, exhibiting liquid gallium trapped inside the crystal. This was because at lower cooling water temperature, the temperature gradient inside the liquid gallium was higher, and the growth rate of the crystal was faster after adding the seed crystal. Although the positive temperature gradient was formed at this time, the temperature at the front of the solid–liquid interface was higher in the radial direction of the crystallizer, hindering the release of latent heat of crystallization in this direction, and thus decreasing the crystal growth in this direction. However, in order to facilitate the release of latent heat of crystallization, the growth orientation of the crystal changed and grew rapidly in the form of dendrites, eventually bridging the crystal branches and Figure 3. Crystalline morphology of liquid gallium after adding crystal seeds for 15 min at different temperatures of cooling water ((a) 288 K, (b) 290 K, and (c) 293 K). Materials 2019, 12, x FOR PEER REVIEW 4 of 11

When the preparation temperature was 273 K, the solidified structure exhibited the geometric polyhedron shape characteristics, indicating that with decreasing subcooling degree, the growth rate of crystal nucleus decreased, and its growth mode transited from dendrite growth to lamellar growth. When the preparation temperature was 278 K, the solidified structure showed a regular hexahedral shape, suggesting that with increasing temperature, the subcooling degree of the growth front further reduced after the formation of crystal nucleus, and the growth mode changed into vertical layered growth. When the preparation temperature was 295 K, the supercooling degree of the solid–liquid interface further decreased after the crystal nucleus formed, hindering the release of latent heat from crystallization. At this time, in order to release the latent heat of crystallization at a faster rate, the growth direction of crystal nucleus changed to side growth, distorting its geometric shape. By comparing the morphological characteristics of crystal seed prepared at four temperatures, the optimal preparation temperature of the crystal seed was finally determined to be 278 K.

Figure 2. Morphology of the seed crystals prepared at different cooling temperatures ((a) 265 K, (b) 273 K, (c) 278 K, and (d) 295 K).

3.2. Effect of Process Parameters on the Crystallization Process

3.2.1.Materials Effect2019, of12, Temperature 2549 of Cooling Water on Crystallization Process 5 of 11

When the flow rate of cooling water was 40 L·h−1 and the temperature was in the range 288–298 K,entrapment 2.9774 kg of crude the liquid gallium phase. obtained The entrained by pickling liquid pretreatment phase impurities was cooled cannot to be the removed critical because point of Materials 2019, 12, x FOR PEER REVIEW 5 of 11 crystallization,the growth mode followed of crystal, by thusadding affecting crystal the seeds purification for 15 process.min. The When corresponding the temperature crystal of growth cooling watermorphology was 293 is K,shown the liquid in Figure gallium 3. grew into a single crystal after adding the seed crystals. Figure 3 shows that when the temperature of the cooling water were 288 and 290 K, the crystal growth mode of liquid gallium was mainly dendrite growth after adding the crystal seed, and the crystal branches bridged with each other, exhibiting liquid gallium trapped inside the crystal. This was because at lower cooling water temperature, the temperature gradient inside the liquid gallium was higher, and the growth rate of the crystal was faster after adding the seed crystal. Although the positive temperature gradient was formed at this time, the temperature at the front of the solid–liquid interface was higher in the radial direction of the crystallizer, hindering the release of latent heat of crystallization in this direction, and thus decreasing the crystal growth in this direction. However, in order to facilitate the release of latent heat of crystallization, the growth orientation of the crystal changed and grew rapidly in the form of dendrites, eventually bridging the crystal branches and entrapment of the liquid phase. The entrained liquid phase impurities cannot be removed because of Figure 3. Crystalline morphology of of liquid gallium after after adding adding crystal crystal seeds seeds for for 15 15 min at different different the growth mode of crystal, thus affecting the purification process. When the temperature of cooling temperatures of cooling water ((a) 288 K, (b) 290 K, andand ((c)) 293293 K).K). water was 293 K, the liquid gallium grew into a single crystal after adding the seed crystals. In order order to further further analyze analyze the the growth growth law law of of the the liquid liquid gallium gallium during during crystallization, crystallization, the the crystal crystal morphology at different different times after adding the crystal seeds was investigated by the dynamic timing −1 observation method, method, when when the the cooling cooling water water flow flow was was 40 40L·h L h and1 and the temperature the temperature was was293 K. 293 The K. · − resultThe result is shown is shown in Figure in Figure 4. 4.

Figure 4. Morphological images of crystal blocks at different times after adding crystal seed ((a) 15 min, Figure 4. Morphological images of crystal blocks at different times after adding crystal seed ((a) 15 (b) 30 min, and (c) 60 min). min, (b) 30 min, and (c) 60 min). Figure4 shows that after the addition of the crystal seed, the gallium crystal block gradually grew Figure 4 shows that after the addition of the crystal seed, the gallium crystal block gradually with increasing crystallization time, and the crystal growth mode of liquid gallium exhibited typical grew with increasing crystallization time, and the crystal growth mode of liquid gallium exhibited layer-by-layer push growth after adding the seed crystals, indicating that the temperature gradient typical layer-by-layer push growth after adding the seed crystals, indicating that the temperature environment formed by the cooling water at 293 K could release the latent heat of crystallization gradient environment formed by the cooling water at 293 K could release the latent heat of generated during the crystal growth to the front of the solid–liquid interface, and transferred and crystallization generated during the crystal growth to the front of the solid–liquid interface, and released it outwardly along the direction of temperature gradient. This kind of layer-by-layer crystal transferred and released it outwardly along the direction of temperature gradient. This kind of layer- growth mode was beneficial to the enrichment of impurity elements from the solid–liquid interface by-layer crystal growth mode was beneficial to the enrichment of impurity elements from the solid– to the liquid phase, thereby affording the solid Ga metal with higher purity. The supercooling liquid interface to the liquid phase, thereby affording the solid Ga metal with higher purity. The degree of the growth tip was the largest when the crystal grew, and the liquid gallium atoms at the supercooling degree of the growth tip was the largest when the crystal grew, and the liquid gallium solid–liquid interface preferentially attached to the growth tip, and the heat was transferred outward atoms at the solid–liquid interface preferentially attached to the growth tip, and the heat was from the crystallized solid gallium in the direction of positive temperature gradient in the crystallizer. transferred outward from the crystallized solid gallium in the direction of positive temperature Therefore, in the crystallization process, crystal growth was always in the form of a pyramind-shaped gradient in the crystallizer. Therefore, in the crystallization process, crystal growth was always in the step-by-layer advancement. According to the crystal growth kinetics and thermodynamics, the form of a pyramind-shaped step-by-layer advancement. According to the crystal growth kinetics and layer-by-layer growth method was conducive to increase the surface area of crystals, facilitating the thermodynamics, the layer-by-layer growth method was conducive to increase the surface area of release of latent heat of crystallization and ensuring the continuous and steady growth of crystals crystals, facilitating the release of latent heat of crystallization and ensuring the continuous and during the crystallization process. Moreover, according to the separation and coagulation theory steady growth of crystals during the crystallization process. Moreover, according to the separation of impurities in the crystallization process, the layer-by-layer growth mode was favorable to the and coagulation theory of impurities in the crystallization process, the layer-by-layer growth mode was favorable to the enrichment of impurity elements from the solid–liquid boundary to the liquid state and can avoid the inclusion impurities of liquid-phase envelopment caused by irregular crystal growth direction. Figure 4 shows that with increasing crystallization time, the pyramind tip of crystal became more and more obvious, and the layered step of crystal growth was also more and more obvious, Materials 2019, 12, 2549 6 of 11 enrichment of impurity elements from the solid–liquid boundary to the liquid state and can avoid the inclusion impurities of liquid-phase envelopment caused by irregular crystal growth direction. MaterialsFigure 2019, 412 shows, x FOR PEER that withREVIEW increasing crystallization time, the pyramind tip of crystal became6 moreof 11 and more obvious, and the layered step of crystal growth was also more and more obvious, attributed attributedto the fact to that the as fact the that crystallization as the crystallization continued, co impurityntinued, elements impurity constantly elements accumulatedconstantly accumulated in the liquid inphase, the liquid and the phase, impurity and the content impurity at the content solid–liquid at the interfacesolid–liquid increased, interface which increased, enhanced which the probabilityenhanced theof impurityprobability elements of impurity attaching elements to the attaching crystal growth to thetip. crystal Owing growth to the tip. di ffOwingerence to in the atomicdifference radius in theand atomic electronegativity radius and between electronegativity Ga and impurity between elements, Ga and theimpurity impurity elements, atoms attached the impurity to the growthatoms attachedtip entered to the into growth the Ga tip lattice entered or latticeinto the gap, Ga causinglattice or the lattice growth gap, defect causing of the Ga growth crystal [defect18–20 ].of ThisGa crystalindicated [18–20]. that theThis removal indicated of impurity that the removal elements of decreased impurity with elements the progress decreased of crystallization with the progress and was of crystallizationconsistent with and the was literature consistent data with [16]. the literature data [16].

3.2.2.3.2.2. Effect Effect of of Cooling Cooling Water Water Flow Flow on on Crystallization Crystallization Process Process InIn a a previous previous study, study, the the effect effect of of cooling cooling water water flow flow on on the the crystallization crystallization process process was was primarily primarily 1 investigated [16]. The results revealed that when the cooling water flow rate was 30 L h−1 , the growth investigated [16]. The results revealed that when the cooling water flow rate was 30 L·h· −, the growth raterate of of gallium gallium crystal crystal near near the the outl outletet of of the the crystallizer crystallizer was was slightly slightly lower lower than than that that in in other other regions. regions. 1 When the cooling water flow rate was 50 L−1h , the growth rate of the gallium crystal in the lower When the cooling water flow rate was 50 L·h ·, the− growth rate of the gallium crystal in the lower part ofpart the of crystallizer the crystallizer was wasslightly slightly larger larger than than that that of the of theupper upper part, part, and and the the growth growth rate rate near near the the 1 crystallizercrystallizer inlet inlet was was the the largest. largest. When When the the cooling cooling water water flow flow rate rate was was 40 40 L·h L h−−1, ,the the growth growth rate rate of of · galliumgallium crystals crystals in in all all the the regions regions of of the the crystallizer crystallizer was was basically basically th thee same, same, and and too too fast fast or or too too slow slow locallocal growth growth phenomenon phenomenon was was not not observed. observed. In In orde orderr to to further further explore explore the the effect effect of of this this process process parameterparameter on on the the crystallization crystallization process, process, the the crysta crystall morphology morphology of of liquid liquid gallium gallium at at different different cooling cooling waterwater flow flow rates rates was was observed, observed, and and the the results results are are shown shown in in Figure Figure 55..

FigureFigure 5. 5. (a()a )At At 30 30 L·h L −h1, 1near, near the the crystallizer crystallizer outlet, outlet, (b) ( bat) 40 at 40L·h L−1,h near1, near the thecrystallizer crystallizer inlet, inlet, and and(c) · − · − at(c 50) at L·h 50− L1, detailh 1, detail morphology morphology of gallium of gallium crystal crystal near nearthe crystallizer the crystallizer inlet. inlet. · − FigureFigure 5 shows that at a cooling water flowflow raterate ofof 4040 LL·hh−11, ,the the crystal crystal morphology morphology of of gallium gallium · − exhibitedexhibited a a distinct “shell“shell pattern”pattern” with with the the uniform uniform grain grain spacing. spacing. This This indicated indicated that that under under that that flow flowrate, rate, the gallium the gallium crystal crystal grew grew in the in layer-by-layer the layer-by-layer manner manner and was and favorable was favorable to remove to theremove impurity. the impurity.At a cooling At a water cooling flow water rate offlow 30 Lrateh 1of, the 30 crystalL·h−1, the growth crystal rate growth at the rate outlet at sidethe outlet of the side crystallizer of the · − crystallizerwas slightly was slower slightly than slower that inthan other that areas, in other and areas, its crystal and its morphology crystal morphology was the samewas the as same that at as a thatcooling at a cooling water flow water of flow 40 L hof 140, also L·h− displaying1, also displaying a distinct a di “shellstinct pattern”.“shell pattern”. This suggested This suggested that under that · − underthis flow this condition, flow condition, the gallium the gallium crystals crystals also grew also in grew the layer-by-layer in the layer-by-layer manner, manner, which was which favorable was favorablefor the removal for the ofremoval impurity; of however,impurity; thehowever, crystal th growthe crystal rate growth here was rate slower here was than slower its surrounding than its surroundingregion, and thusregion, the and possibility thus the of envelopingpossibility of the enveloping liquid phase the with liquid the phase progress with of crystallizationthe progress of at crystallizationthis point cannot at this be point ruled cannot out. However, be ruled atout. a coolingHowever, water at a flowcooling rate water of 50 flow L h rate1, due of 50 to theL·h− higher1, due · − toheat thetransfer higher heat efficiency transfer at theefficiency bottom at of the the bottom outlet side of the of theoutlet crystallizer, side of th thee crystallizer, driving force the of driving gallium force of gallium crystal growth was larger and the crystal growth rate was faster, changing the crystal morphology and presence of a large number of irregular growth steps. It can be deduced that the crystals at the site were not completely in the way of layer-by-layer promotion growth, and the crystal growth process may be accompanied by dendrite or peritectic formation, leading to envelope liquid phase, entrap impurities, and reduce purification effect in solid gallium. Materials 2019, 12, 2549 7 of 11 crystal growth was larger and the crystal growth rate was faster, changing the crystal morphology and presence of a large number of irregular growth steps. It can be deduced that the crystals at the site were not completely in the way of layer-by-layer promotion growth, and the crystal growth process Materialsmay be 2019 accompanied, 12, x FOR PEER by REVIEW dendrite or peritectic formation, leading to envelope liquid phase, entrap7 of 11 impurities, and reduce purification effect in solid gallium. 3.2.3. Effect of Seed Number on Crystallization Process 3.2.3. Effect of Seed Number on Crystallization Process When the cooling water flow rate was 40 L·h−1 and the temperature was 293 K, the liquid gallium cooledWhen to the the critical cooling point water of flowcrystallization rate was 40 and L h 3,1 and4, 5, the and temperature 6 crystal seeds was 293were K, added. the liquid When gallium the · − crystallizationcooled to the criticalreached point a certain of crystallization proportion, its and morp 3,hological 4, 5, and 6image crystal is seedsshown were in Figure added. 6, indicating When the thatcrystallization the number reached of seed aadded certain determined proportion, the its shape morphological of the uncrystallized image is shown region. in When Figure 6three, indicating crystal seedsthat the were number added, of the seed uncrystallized added determined region theexhibite shaped ofa triangle the uncrystallized shape. When region. four crystal When threeseeds crystal were added,seeds were the uncrystallized added, the uncrystallized region exhibited region exhibiteda quadrilateral a triangle shape. shape. However, When fourwhen crystal the number seeds were of addedadded, crystal the uncrystallized seed was 3 regionor 4, the exhibited shape aand quadrilateral size of the shape. uncrystallized However, whenarea were the number not consistent, of added showingcrystal seed a funnel was 3shape or 4, with the shape a large and top size and ofa small the uncrystallized bottom. This easily area were led to not the consistent, intersection showing of the crystala funnel growth shape at with the bottom a large topof the and crystallizer a small bottom. with the This continuous easily led progress to the intersection of crystallization, of the which crystal causedgrowth the at the envelopment bottom of the of crystallizer the liquid with phase the and continuous entrapment progress of impurities, of crystallization, thus affecting which caused the purificationthe envelopment effect. of theWhen liquid five phase seed and crystals entrapment were ofadded, impurities, the uncrystallized thus affecting theregion purification presented effect. a pentagonWhen five shape, seed crystalsand the wereproblem added, of shape the uncrystallized with a large regiontop and presented small bottom a pentagon in the uncrystallized shape, and the regionproblem improved. of shape In with the a largecase of top adding and small six seed bottom crystals, in the uncrystallizedthe uncrystallized region region improved. displayed In the a hexagonalcase of adding shape six with seed a crystals,regular shape the uncrystallized and uniform regionsize and displayed was most a beneficial hexagonal to shape control with the a overall regular directionshape and of uniform the crystal size during and was the most purification beneficial of to crude control gallium. the overall Therefore, direction the ofoptimal the crystal number during of seedthe purification addition was of determined crude gallium. to be Therefore, six, when the th optimale 4N crude number gallium of seed was additionpurified wasusing determined a self-made to crystallizer.be six, when the 4N crude gallium was purified using a self-made crystallizer.

FigureFigure 6. 6. PhotographsPhotographs of of crystal crystal morphology morphology when adding didifferentfferent numbersnumbers of of seeds seeds (( ((aa)) 3, 3, ( b(b)) 4, 4, ( c()c) 5, 5,and and (d ()d 6).) 6).

3.3. Effect of Process Parameters on Crystallization Rate In the actual crystallization solidification process of liquid gallium, the crystallization rate (i.e., the crystal growth rate of gallium after adding crystal seed) depended on the supercooling degree of the solid–liquid interface. The supercooling degree of solid–liquid interface was a function of temperature and cooling water flow keeping other process conditions constant. In the experiment, Materials 2019, 12, 2549 8 of 11

3.3. Effect of Process Parameters on Crystallization Rate In the actual crystallization solidification process of liquid gallium, the crystallization rate (i.e., the crystal growth rate of gallium after adding crystal seed) depended on the supercooling degree of the solid–liquid interface. The supercooling degree of solid–liquid interface was a function of Materials 2019, 12, x FOR PEER REVIEW 8 of 11 temperature and cooling water flow keeping other process conditions constant. In the experiment, the relationships between the crystallization rate and cooling water temperature as well as the flow were the relationships between the crystallization rate and cooling water temperature as well as the flow measured by the control variable method, and the empirical control formula of the crystallization rate were measured by the control variable method, and the empirical control formula of the was obtained by analyzing the experimental data. In order to reduce the experimental error, improve crystallization rate was obtained by analyzing the experimental data. In order to reduce the the accuracy of empirical control formula and its adaptability to the actual production process, each experimental error, improve the accuracy of empirical control formula and its adaptability to the group of measurement experiment was repeated four times, and the average value was taken. The actual production process, each group of measurement experiment was repeated four times, and the crystallization rate measured in the experiment was the average rate during the complete solidification average value was taken. The crystallization rate measured in the experiment was the average rate process of liquid gallium after adding the crystal seed, and the calculation formula is as follows: during the complete solidification process of liquid gallium after adding the crystal seed, and the calculation formula is as follows: m v = (1) m t v = (1) where v is the average rate, kg/h; m is the total masst of liquid gallium, kg; t is the time required for the wherecomplete v is solidification the average ofrate, liquid kg/h; gallium, m is the hour total (h). mass of liquid gallium, kg; t is the time required for the completeThe effect solidification of the temperature of liquid gallium, and flow hour rate (h). of the cooling water on the crystallization rate determinedThe effect by theof the test temperature is shown in and Figure flow7. Figurerate of7 athe shows cooling that water with increasingon the crystallization cooling water rate determinedtemperature, by the the crystallization test is shown rate in Figure gradually 7. Fi decreased,gure 7a shows and anthat obvious with increasing linear relationship cooling water was temperature,observed between the crystallization the two. The empiricalrate gradually control decreased, formula ofand the an cooling obvious water linear temperature relationship on was the observedcrystallization between rate the was two. obtained The empirical by Origin control software fo fitting.rmula of the cooling water temperature on the crystallization rate was obtained by Origin software fitting. v(T) = 0.09T + 27 (2) v(T) = −0.09− T + 27 (2) where TT isis thethe temperaturetemperature ofof the the cooling cooling water, water, K; K; and and the the linear linear correlation correlation coe coefficientfficient of of the the data data fit wasfit was R2 R=2 0.997.= 0.997. Figure7 7bb showsshows thatthat withwith increasingincreasing flowflow raterate of cooling water, the crystallization rate increases, and aa significantsignificant exponential exponential function function relationship relationsh wasip was observed observed between between the two. the The two. empirical The empirical control controlformula formula of the cooling of the cooling water flow water to theflow crystallization to the crystallization rate was rate obtained was obtained by Origin by software Origin software fitting. fitting. Q v(Q) = 96.73e 4.94− +0.66 (3) − -Q v(Q) = -96.73e4.94 + 0.66 (3) where Q is the flow rate of the cooling water, L/h; and the standard deviation of the data fit was where Q is the flow rate of the cooling water, L/h; and the standard deviation of the data fit was R2 = R2 = 0.997. 0.997.

Figure 7. (a)Effect of cooling water temperature on the crystallization rate at a flow rate of 40 L h 1; Figure 7. (a) Effect of cooling water temperature on the crystallization rate at a flow rate of 40 L·h· −−1; (b) effect of cooling water flow rate on the crystallization rate at 293 K. (b) effect of cooling water flow rate on the crystallization rate at 293 K.

3.4. Analysis of Purification Results Based on the above studies, the optimum technological parameters for the crystal purification of 4N raw material gallium were determined as follows: the temperature of seed preparation, 278 K; cooling water temperature, 293 K, cooling water flow, 40 L·h−1, and the number of seed crystals added was six. Combined with our previous research results [16], three purification schemes were determined under the optimized process parameters, as listed in Table 1. Materials 2019, 12, 2549 9 of 11

3.4. Analysis of Purification Results Based on the above studies, the optimum technological parameters for the crystal purification of 4N raw material gallium were determined as follows: the temperature of seed preparation, 278 K; cooling water temperature, 293 K, cooling water flow, 40 L h 1, and the number of seed crystals added · − was six. Combined with our previous research results [16], three purification schemes were determined under the optimized process parameters, as listed in Table1.

Table 1. Scheme of preparing 6N,7N high-purity gallium by the crystal method.

Scheme Recrystallization Crystal Proportion % Number Number First Time Second Time Third Time Forth Time I 2 70 70 // II 3 70 70 85 / III 4 70 70 85 85

The impurity contents in high-purity gallium prepared by the three purification schemes were tested and compared to the raw gallium material, and the removal rate of impurities was calculated. The results are shown in Table2.

Table 2. Detection of impurity contents in raw gallium material and high purity purified gallium.

Mass Fraction ( 10 6) Removal Rate (%) Element × − Raw Material I II III I II III Mg 76 1.59 0.34 0.07 97.9 99.6 99.9 Al 1 <0.1 <0.1 <0.1 100 100 100 ≈ ≈ ≈ Cr 40 2.66 0.94 0.34 93.3 97.6 99.2 Fe 15 1.94 0.93 0.45 87.1 93.8 97.0 Cu 107 1.32 0.22 0.04 98.8 99.8 >99.9 Zn 24 2.42 1.05 0.45 89.9 95.6 98.2 Pb 56 2.32 0.68 0.20 95.9 98.8 99.6

Table2 shows that for Scheme I, after the purification, the mass fraction of Al impurities contained in the raw materials reduced below the detection limit of HR-GDMS, and the other six main impurities were also well removed. The removal rates were: Fe-87.1%, Pb-95.9%, Zn-89.9%, Mg-97.9%, Cu-98.8%, and Cr-93.3%, and the mass fraction of the main gallium metal in the refined high-purity Ga calculated by the difference method was 99.999987%. For Scheme II, the removal rates of the six main impurities were Fe-93.8%, Pb-98.8%, Zn-95.6%, Mg-99.6%, Cu-99.8%, and Cr-97.6%, and the mass fraction of the main Ga metal was 99.9999958%. For Scheme III, the removal rates of the six major impurities further increased, and the removal rates of Mg and Cu exceeded 99.9%. In contrast, although the removal rate of Fe was the lowest, it also reached 97%. The mass fraction of main Ga metal was 99.9999958%.

4. Conclusions In conclusion, the impurity removal of Ga was investigated in detail, and the radial crystallization purification method under the seed crystal induction is proposed. The effect of cooling temperature on the morphology of the crystal as well as the cooling water temperature, flow rate, and the number of crystal seeds added on the crystallization process were investigated. The optimum purification process was obtained; the control equation of the crystallization rate was determined; and the high-purity (6N and 7N) gallium was prepared under the technological conditions. The main conclusions of this study are as follows: Materials 2019, 12, 2549 10 of 11

(1) The optimum process parameters for the crystallization purification of 4N raw material gallium are as follows: temperature of the seed preparation, 278 K; cooling water temperature, 293 K; cooling water flow, 40 L h 1; the number of seed crystals added six 6; · − (2) The crystallization rate decreased linearly with increasing cooling water temperature and increased exponentially with increasing cooling water flow. The control formulas of the cooling water temperature T and flow Q on the crystallization rate v are, v(T) = 0.09T + 27 and Q − v(Q) = 96.73e 4.94− +0.66, respectively; − (3) The three proposed purification schemes effectively removed the impurity elements. When using Scheme I to purify the 4N crude gallium, high-purity gallium with a purity of 6N was obtained. When adopting Schemes II and III, 7N high-purity gallium was obtained. The purities of the high-purity gallium prepared by Schemes I, II, and III were 99.999987%, 99.9999958%, and 99.9999958%, respectively.

The seed crystal induced radial crystallization purification method proposed in the study has the advantages of simple operation, convenient process flow, low energy consumption, environmentally friendly, and easy to realize automatic control of the purification process, providing a new idea for the large-scale industrial production of ultra-high purity gallium.

Author Contributions: Conceptualization, J.H. and K.P.; methodology, K.P.; validation, X.T.; formal analysis, J.H. and X.T.; investigation, K.P.; data curation, J.H.; writing—original draft preparation, X.T.; writing—review and editing, K.P.; supervision, K.P.; project administration, K.P. Funding: This research received no external funding Conflicts of Interest: The authors declare no conflict of interest.

References

1. Chou, W.L.; Wang, C.T.; Yang, K.C.; Huang, Y.H. Removal of gallium (III) ions from acidic aqueous solution by supercritical carbon dioxide extraction in the green separation process. J. Hazard. Mater. 2008, 160, 6–12. [CrossRef][PubMed] 2. Critical Mineral Resources of the United States—Economic and Environmental Geology and Prospects for Future Supply; Schulz, K.J.; DeYoung, J.H.J.; Seal, R.R., II; Bradley, D.C. (Eds.) US Geological Survey Reston: Reston, VA, USA, 2017. 3. Tolcin, A.C. (Ed.) Mineral Commodity Summaries 2019; US Geological Survey Reston: Reston, VA, USA, 2019. 4. Zhao, Z.; Yang, Y.; Xiao, Y.; Fan, Y. Recovery of gallium from Bayer liquor: A review. Hydrometallurgy 2012, 125–126, 115–124. [CrossRef] 5. Swain, B.; Mishra, C.; Kang, L.; Park, K.S.; Lee, C.G.; Hong, H.S. Recycling process for recovery of gallium from GaN an e-waste of LED industry through ball milling, annealing and leaching. Environ. Res. 2015, 138, 401–408. [CrossRef][PubMed] 6. Liu, F.; Liu, Z.; Li, Y.; Wilson, B.P.; Lundström, M. Recovery and separation of gallium(III) and germanium(IV) from zinc refinery residues: Part I: Leaching and iron(III) removal. Hydrometallurgy 2017, 169, 564–570. [CrossRef] 7. Ober, J.A. Mineral Commodity Summaries 2018; US Geological Survey Reston: Reston, VA, USA, 2018. 8. U.S. Geological Survey. Mineral Commodity Summaries 2015; US Geological Survey Reston: Reston, VA, USA, 2015. 9. Meyer, S.; Wahl, S.; Molchanov, A.; Neckermann, K.; Möller, C.; Lauer, K.; Hagendorf, C. Influence of the feedstock purity on the solar cell efficiency. Sol. Energy Mater. Sol. Cells 2014, 130, 668–672. [CrossRef] 10. Kenichi, T.; Nagayasu, Y.; Kishio, T. Electrolytic Refining Method for Gallium and Apparatus for Use in the Method. U.S. Patent 19,990,428,476, 2001. 11. Kenichi, T.; Nagayasu, Y.; Kishio, T. Method for Electrolytically Refining Gallium and Deveice Therefor. JP10310331, 2000. 12. Bollong Alan, B.I.; Bult Roelof, P.; Proux Gary, T. Method for the Zone Refining of Gallium. U.S. Patent 23371088A, 1989. Materials 2019, 12, 2549 11 of 11

13. Lee, M.S.; Ahn, J.G.; Oh, Y.J. Production of high-purity indium and gallium metals by vacuum refining. Mater. Trans. 2002, 43, 3195–3198. [CrossRef] 14. Harper, J.G. Method of Purifying Galliun by Recrystallization. U.S. Patent 3,088,853, 7 May 1963. 15. Pan, K.; Li, Y.; Zhao, Q.; Zhang, S. Simulation of Solidification Process of Metallic Gallium and Its Application in Preparing 99.99999% Pure Gallium. JOM 2019, 71, 737–743. [CrossRef] 16. Pan, K.; Li, Y.; Zhang, J.; Zhao, Q. A Facile and Low-Cost Method to Produce Ultrapure 99.99999% Gallium. Materials 2018, 11, 2308. [CrossRef][PubMed] 17. Li, Y.; Pan, K.; Li, Z.; Yan, C. Preparation of high-purity gallium by crystallization method. Xiyou Jinshu Chin. J. Rare Met. 2015.[CrossRef] 18. Kozlov, S.A.; Potolokov, N.A.; Gusev, A.V.; Fedorov, V.A. Determination of the Effective Segregation Coefficients of Metallic Impurities in Gallium during Directional Solidification. Inorg. Mater. 2002, 12, 1212–1215. [CrossRef] 19. Kozlov, S.A.; Potolokov, N.A.; Gusev, A.V.; Fedorov, V.A. Effect of Impurities on Crystal Growth of Gallium. Inorg. Mater. 2003, 39, 1267–1270. [CrossRef] 20. Kozlov, S.A.; Potolokov, N.A.; Gusev, A.V.; Fedorov, V.A. Segregation Behavior of Impurities in Gallium during Directional Solidification. Inorg. Mater. 2003, 5, 452–454. [CrossRef]

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).