Research on Freeze Drying and High-temperature Molding in Medium-temperature Fluorocarbon Board

ZHANG Dong, ZHANG Pan, YAN Xuefang, CHAN Mingyang, YANG Lin＊

(1.Chemical Science and Engineering College, Sichuan University, Chengdu 610065, China; 2.Engineering Research Center of Comprehensive Utilization and Clean Processing of Phosphorus, Chengdu 610065, China)

Abstract: A research and test system for the carbon anode plate preparation technology was established to optimize the physical and chemical indicators of carbon anode plates, such as bulk density, resistivity, and compressive strength, and improve the operating cycle. In this study, a carbon plate was prepared via a combination of high-temperature molding and freeze drying using a formulation with asphalt content much lower than the industry standard. The experimental results show that the density of the carbon plate is increased by 0.02-0.04 g/cm3 by improving the drying method. The carbon plate prepared in the laboratory has a bulk density of 1.814 g/cm3, resistivity of 29.8 μΩ·m, and compressive strength of 89.27 MPa. The bulk density, room-temperature resistivity, compressive strength, graphitization, and other key indices of the carbon plates made in the laboratory and those procured from a factory in Shanxi, Datong, were tested. Additionally, the specimens were analyzed using thermogravimetry-differential scanning calorimetry, scanning electron microscopy, and X-ray fluorescence. The laboratory-made carbon plates were superior to the factory specimens in terms of all the indicators tested. The process in this study improves the performance of the carbon anode plate and is used to provide technical support for electrolytic fluorine production in enterprises. The carbon plates prepared in the laboratory fully meet the process requirements of a medium-temperature electrolytic fluorine production line, which indicates the possibility of its use in the stable production of fluorine gas.

Key words: fluorocarbon anode; high density; freeze drying; high-temperature molding

1 Introduction

Elemental fluorine is a very important auxiliary material for the extraction of uranium raw materials closely related to the nuclear industry and is additionally the most basic raw material for the preparation of various special fluoride gases[1]. In 1886, French chemist Moisson invented the electrolytic fluorine technology, wherein he used a platinum electrode to electrolyze a hydrogen fluoride solution containing a small amount of potassium fluoride to obtain elemental fluorine[2,3]. The fluoride electrolysis technology subsequently underwent several years of development. In 1940, the United States and parts of Europe established the earliest industrial production system for fluorine to meet the production needs of uranium hexafluoride. After continuous tests by researchers from the former Soviet Union, the United States, and other countries, the optimum technological parameters of a medium-temperature fluorine electrolytic cell were determined as follows: potassium diacid fluoride (KF-2HF) as the electrolyte, in which the mass fraction of hydrogen fluoride was approximately 40.8%, and tank temperature of approximately 95 ℃[4]. Several problems persist in the production of fluorine at medium temperature even after decades of development. Nevertheless, medium-temperature electrolysis is presently the only industrial technology whereby mass production of fluorine can be realized.

Nickel anodes were used in the early stages but were gradually replaced by carbon anodes owing to serious corrosion and low current efficiency. To date, carbon remains the preferred anode material in medium-temperature fluorine electrolysis. The carbon anode selected should have low electrical resistance, high strength, high corrosion resistance, anti-polarization ability, and good electrical contact performance. At present, the requirements of carbon anodes are as follows[5,6]: high conductivity and low resistivity, good corrosion resistance to fluorine and hydrogen fluoride, good infiltration of electrolyte and low overvoltage, good mechanical properties of carbon plates to facilitate long-term usage under high current densities, high density and low porosity so that in the case of local rupture, the fragments can sink to the bottom of the electrolytic cell to avoid a short circuit between the electrodes, manufacturing convenience and low price. Simultaneously, the density of carbon plates significantly impacts their resistivity, mechanical strength, wettability, corrosion resistance, and other properties. Therefore, the importance of high-density carbon plates is self-evident.

Deng He proposed a method for the preparation of carbon anodes for fluorine production wherein a mixture containing copper powder was added into the carbon anode material[7]. The copper-containing carbon plate thus fabricated had the advantages of uniform pores, high mechanical strength, and low resistivity. T. Teruhisa proposed a method for inhibiting the polarization of carbon electrodes by impregnating them with lithium fluoride, sodium fluoride, aluminum fluoride, magnesium fluoride, and other metal fluorides[8]. Huanzhi Wang modified the surface of commercial carbon boards via high-pressure impregnation and fabricated carbon plates with low porosity, low ash content, high mechanical strength, and high density[9]. The addition of transition metals such as nickel, cobalt, and vanadium to the electrode material increases the rate of charge transfer while the metal fluoride generated during electrolysis improves the wettability of the electrode surface so that the fluoride ions are adsorbed and the gas generated is desorbed more easily[10]. Spraying a metallic layer of nickel or copper on the contact surface of the carbon electrode can increase the effective contact area. Furthermore, dense coatings can diminish electrolyte penetration and corrosion of the contact surface, improve electrolysis efficiency, and extend the life of the carbon anode[11,12]. Although the above methods can improve the performance of carbon plates, the higher costs of metal and metal fluoride as well as the additional equipment required for high-pressure impregnation significantly increase the production costs of carbon plates. These methods are not conducive to mass production nor do they sufficiently address the current difficulties faced by enterprises. Hence, it is critical to develop a high-performance carbon anode plate, which can be produced in large quantities.

At present, the traditional formula for preparing carbon anodes entails an asphalt content of 16%-18%. Significant amounts of volatile matter are produced upon calcination, and the density of the carbon plate decreases accordingly. The carbon plate fabricated needs to be impregnated and roasted many times to meet standards, which increases its production cost. In the present study, the raw material formula and drying method were optimized, and a low-asphalt formula (13.3%) with asphalt content far lower than the standard was adopted. Compared with the traditional formula, our low-asphalt formula yielded fluorocarbon anodes with similar bulk density but fewer pores caused by the release of volatile matter from asphalt upon roasting[13]. The resultant carbon plate was freezedried, and the blank of a porous carbon plate with a smooth pore wall and a large aperture was formed, which favored the infiltration of asphalt into the carbon plate. After the carbon plate was freeze-dried, it was impregnated twice and roasted thrice. Its final density was 1.814 g/cm3.

2 Experimental

2.1 Materials

Petroleum coke was supplied by Tianjin Youhao New Material Technology Co., Ltd (China). Its trace element composition is shown in Table 1.

Table 1 Composition of petroleum coke trace elements

Table 2 Modified coal bitumen properties

The modified asphalt was provided by Shanghai Baosteel Chemical Co., Ltd. (China) and characterized using thermogravimetric analysis. The thermogravimetric analysis curve of asphalt is shown in Fig.1. The sample was ground to a fine powder and passed through a 200-mesh sieve. A synchronous thermal analyzer (STA 449 F3, NETZSCH, Germany) was used to heat the sample to 1 250 ℃ at a heating rate of 10 ℃/min in an N2atmosphere, and the change in sample weight with time was recorded. The properties of the modified asphalt are shown in Table 2.

Fig.1 Thermal gravimetric analysis curve of asphalt

Table 3 Particle size distribution of petroleum coke

The petroleum coke was pulverized and then placed on an electric vibrating screen for particle size classification. Particles with diameters in the range 0.0385-0.054 mm were ball-milled to obtain a powder. Table 3 shows the size distribution of the petroleum coke particles used in the preparation of carbon plates. The proportion of petroleum coke to asphalt in the formula was 13.3:100.

Table 4 Test block densities and maximum difference between densities of fluorocarbon anodes

Process flow: A crusher was used to obtain raw material aggregates of different particle sizes. These were mixed with the pulverized asphalt in a certain ratio, and the resultant compound was mixed and kneaded in a kneader to form a paste. After the paste was cooled to room temperature(22±2 ℃), it was transferred to a special mold for compression molding to yield a green carbon anode called embryo. The embryos were cooled with water and then freeze-dried. The dried raw embryos were placed in a high-temperature vacuum tube sintering furnace and roasted for the first time in accordance with a predetermined schedule of temperature rise. The green body after the first firing was subjected to the first impregnation, second firing, second impregnation, and third firing to obtain the finished carbon anode plate base material. After drilling, tapping, and installation of copper screws, the finished product was obtained[14]. The process flow chart of the preparation is shown in Fig.2.

Fig.2 Process flow chart

2.2 Carbon anode performance test

The internal morphology of the sample was observed via scanning electron microscopy (SEM, VGEA3-SBH). The degree of graphitization was measured using laser Raman spectroscopy (Raman, inVia Reflex, Reinshaw, UK) and X-ray diffraction (XRD, PANalytical BV, Netherlands). The thermal stability of the samples was analyzed via thermogravimetric differential scanning calorimetry (TG-DSC, STA 449F3, NETZSCH, Germany). A four-probe tester (KDB-1,Kun De, China) was used to analyze the resistivity of the carbon plate at different stages. Its compressive performance was tested using a universal testing machine (Instron, Norwood, Massachusetts). X-ray fluorescence spectrometry (XRF, S8 TIGER, Bruker) was used to analyze the trace elements of the carbon plate.

3 Results and discussion

3.1 Research on preparation process

High-temperature molding: As temperature rises, asphalt changes from solid to liquid. Under the action of pressure, the mold transfers potential energy to the material such that the asphalt is immersed in the aggregate micropores, the solid particles come into closer contact, porosity decreases, and a green carbon anode is formed.

Freeze drying: The water in the fluorocarbon board is cooled to a temperature below its eutectic point, and then solidified and crystallized to obtain a solid with a certain shape. The water crystals compress the petroleum coke particles in the carbon plate. The solidified body of the carbon plate is placed under pressure below the eutectic point of water to sublime the water crystals, which leave holes in the space they had occupied. Thus, the carbon plate becomes a porous body with the same crystal structure. After sintering, a porous carbon plate with the same pore and crystal structures is obtained[15,16].

In the present study, the green body obtained via high-temperature molding was placed in water for 2 h to ensure that it was completely wetted. A part of the green body was directly dried in an oven for 0.5 h. The other part was placed in a refrigerator at -20 ℃ for 2 h until the water in the carbon plate had completely frozen. The green body was then placed in a vacuum freeze dryer for 5 h of drying.

3.1.1 Effects of drying method and high-temperature molding on bulk density of green body

High-temperature molding is an indispensable process in green carbon plate production. Temperature and pressure are the two most critical factors affecting the density of the green body.

Fig.3 shows the effects of temperature and pressure on the densities of green bodies in two different drying methods. When the pressure was less than 30 MPa, the density of the carbon plate increased with pressure but the amplitude of increase declined. The increase in density as pressure increased from 10 MPa to 20 MPa was significantly greater than that observed with increase in pressure from 20 MPa to 30 MPa. At high temperatures, as density continued to increase, the solid particles were more closely bound. The friction and surface tension that the asphalt needed to overcome in the pores of the carbon plate additionally increased, which was not conducive to the continuous filling of the petroleum coke particle pores and the penetration of asphalt into the aggregate particles. Therefore, as pressure continued to increase, the increase in the density of the green bodies declined.

Fig.3 Effects of temperature and pressure on densities of green bodies under different drying conditions

With increase in temperature, the volume of the carbon plate first increased and then decreased. As temperature increased, asphalt changed from a solid into a liquid. Its viscosity decreased, which enhanced its fluidity and wettability. This was conducive to the infiltration of aggregate particles and filling of pores by asphalt during molding. As temperature increased further, the viscosity of asphalt decreased only slightly until it ceased to change. However, the overflow of several volatile gases produced at high temperatures caused an increase in pores and volume expansion of the carbon plate. The density of the green body decreased with increase in temperature.

Compared with direct drying, freeze drying can cause the density of the green body to increase by 0.02-0.04 g/cm3. Freeze drying needs to be carried out at low temperatures. Therefore, the green body shrank at low temperatures and its particles were more closely combined.

3.1.2 Effects of drying methods and calcination times on density of carbon plates

Fig.4 shows the relationship between temperature and density for different calcination durations. The density of the carbon plate gradually increased with the number of firings. In addition, density increased by 0.08-0.14 g/cm3between the first and second firings, and by 0.05-0.1 g/cm3between the second and third firings. After carbonization at a high temperature of 800 ℃ or more, the volatile gases in asphalt overflowed, and the pores of the matrix became open again. In the subsequent dipping, the asphalt achieved a better wetting effect. When asphalt was used as an impregnant, liquid-phase carbonization occurred. As impregnation progressed, the pores of the carbon plate were gradually filled with asphalt, porosity decreased, and resistance to impregnation increased.

Fig.4 Relationship between temperature and density at different roasting stages

Across roasting processes, the density of the carbon plate obtained via freeze drying was greater than that attained via direct drying. As water crystallized, the crystals squeezed the particles of the carbon plate together, rendering them more closely combined, while the porosity of small holes in the carbon plate declined and the pores became smaller. The carbon plate was placed under pressure below the eutectic point of water and the water crystals sublimated. This yielded a carbon plate with much larger pores than that before freeze drying. After freeze drying, a porous carbon plate with smooth pore walls and large pores was formed. Asphalt has minimum viscosity at 210 ℃ and can achieve good fluidity. When asphalt was impregnated into the porous carbon plates with smooth pore walls and large pore diameters, the friction and surface tension that need to be overcome were low,i e, it had low viscosity. Therefore, the impregnant could smoothly enter the carbon plate and increase its weight.

3.2 Density uniformity

The performance of the carbon plate at its weakest point is often below average. However, corrosion of the carbon plate in an electrolytic cell starts from the weakest point. Therefore, uniformity of carbon plates is significant to their performance. Among carbon plates with similar densities, those with uniform density distributions may be more resistant to corrosion and operate in electrolytic cells for longer durations than those with uneven density distributions. In the present study, a uniformity index was incorporated to evaluate the performance of the laboratory-made carbon plates.

The test block was prepared using the symmetrical thickness thinning method, and its density was calculated using the volumetric weight method. The density uniformity of the carbon plate was characterized by the difference between the highest and lowest densities of the test block (ΔDmax=Dmax-Dmin)[17].

Specific method: As shown in Fig.8, the test block was first selected from the carbon slab. The 0#test block was taken as the starting block, and thinning was carried out symmetrically from both sides in the thickness direction. A series of test blocks, labeled as 1#, 2#, 3#,etc, with successively lower thicknesses, were obtained. The density of each test block,i e,D0,D1,D2, andD3,etc, and the maximum difference between these values were calculated.

Fig.5 Symmetrical thickness thinning method used to obtain test blocks to detect density uniformity

Fig.8 XRD patterns of (a) Shanxi carbon plates and (b) laboratory-made carbon plates

Table 4 shows the density distributions of the test blocks with different densities. As the test block became thinner, the maximum difference in the densities of the preceding three test blocks was less than or equal to 0.011 g/cm3, except in the case of carbon plate “a”. This showed that the carbon plate fabricated in the laboratory had better density uniformity along the thickness direction than that of its factory counterpart.3.2.1 Resistivity test

Fig.6 shows the relationship between density and resistivity under different calcination times. With increase in calcination time, the resistivity of the laboratory-made carbon plate decreased gradually. From the first firing to the second, the resistivity decreased by 13-53 µΩ·m. From the second firing to the third, the resistivity decreased by 17-25 µΩ·m. When the samples were roasted to 750 ℃, the asphalt coking process was completed. To optimize the coking process and improve the physical and chemical performance indices further, the carbonization temperature should be increased to 1 000 ℃[18]. In the present study, carbonization was carried out twice at 1 250 ℃, which was primarily why the resistivity of the carbon plate decreased gradually with increase in roasting degree.

Fig.6 Relationship between density and resistivity at different roasting stages

The laboratory-made carbon plate had a resistivity of 29.8 µΩ·m and a density of 1.814 g/cm3. The resistivity of the Shanxi carbon plate was 51.718 µΩ·m and its density was 1.70 g/cm3. The pores of the carbon anode contained air, and its resistivity was much higher than that of a carbon matrix anode. The resistivity of the carbon anode increased with the increase of porosity[19]. As bulk density increased, the pores of the carbon plate were completely filled with asphalt, which decreased its porosity and, hence, its resistivity.

3.2.2 Compressive strength test

A universal testing machine was used to test the compressive strength of the laboratory-made and Shanxi carbon plates. The compressive strength of the carbon plate manufactured in the laboratory was 89.27 MPa while that of the carbon plate sourced from Datong, Shanxi, was 68.23 MPa. Fig.7 shows the relationship between compressive strength and displacement of the carbon plates. Curves 2-6 are the compressive strength-displacement curves of the laboratory-made carbon plates, and curve 1 is that of the Shanxi carbon plate. Curves 1-6 exhibit the characteristics of brittle fracture,i e, they change almost linearly before the limit load is reached. When the limit load is exceeded, the curves exhibit rapid and abrupt declines. The formulations used to manufacture carbon plates in the laboratory contained less asphalt. The carbon plates were mostly made of petroleum coke, and only small portions were filled with macerated asphalt. However, the carbon plates manufactured in the factory had higher asphalt contents. After roasting, the significant overflow of volatile matter generated several pores, which caused the density to decline considerably. The interior of the carbon plate was further ablated under electrolytic conditions. When subjected to a strong impact, these pores became the main source of fracture and eventually caused the carbon plate to fracture. The laboratory-made carbon plates were composed of petroleum coke particles of various sizes. As larger granular aggregates support the mechanical properties of the entire carbon plate, a network structure with good mechanical properties was formed. A comparison of the mechanical properties of the different carbon plates revealed that high-density carbon plates had greater compressive strength.

Fig.7 Compressive strength-displacement curves

The preparation process of the laboratory-made carbon plates was analyzed and the reasons for their high compressive strength are deduced as follows:

They contained petroleum coke of higher quality.Defects arising from calcined coke were decreased owing to the controlled calcination of petroleum coke. The processes of kneading and cooling had significant smoothening effects on the pores of the granular material.

3.2.3 Test of graphitization degree

The non-graphitization of carbon anodes used in fluorine production is the key characteristic that helps avoid anode polarization in fluorine cells. Research has shown that when a carbon anode mixed with graphite pieces is used in electrolytic fluorine production, the “passivation layer” and swelling phenomenon on its surface lead to the anode effect and cause the electrolytic current to change sharply. The graphite area is perforated and slag is dropped. Thus, the anode effect can seriously affect normal operations. The microstructure and porosity of the graphite anode surface as well as the passivation layer are not uniform. Owing to differences in density, stress, current, and voltage, the passivation layer in the area rich in fluorine-carbon bonds may fall off the graphite anode and be suspended on the surface of the electrolyte. Therefore, the determination of the graphitization degree of the fluorine-carbon anode and its detection method are very important. The former Soviet Union and China’s Jilin Carbon Co., Ltd. included a graphitization degree of 0 in the physicochemical index of finished carbon anode products[20,21].

Fig.8 shows the XRD patterns of the laboratory-made and Shanxi carbon plates. In pattern (a), a small, sharp diffraction peak near 2θ = 26.228° and two small diffraction peaks at 2θ values of 54.2° and 77.3° corresponded to the characteristic diffraction peaks of graphite microcrystals. These peaks did not appear in the XRD patterns of the laboratory-made carbon plates. In pattern (b), the broad diffraction peak at 2θ= 25.45° corresponded to the characteristic diffraction peak of amorphous carbon. The wider diffraction peak indicated low crystallinity and a mostly amorphous state of existence. Thus, it was established that the laboratory-made carbon plates comprised amorphous carbon and the Shanxi carbon plates contained graphite. The following reference patterns are employed: JCPDS card No. 26-1080 for amorphous carbon, and JCPDS card No. 41-1487 for graphite.

Fig.9 shows the Raman spectra of the laboratory-made and Shanxi carbon plates. The peak at 1 580 cm-1corresponded to the graphite structure (sp2), while that at 1 360 cm-1corresponded to the amorphous and polyphase carbon materials. The ratio of the two peak strengths,R=ID/IG, indicated the degree of graphitization of the material[22]. The smaller the ratio, the larger the microcrystal, the more complete the crystallization, and the higher the graphitization degree of the material. Generally, a value ofRlower than 0.8 indicates the presence of graphite. The Raman test results showed thatR= 1.010 for the laboratory-made carbon plate andR= 0.795 for the Shanxi carbon plate. These results were consistent with the XRD test results, which proved that the laboratory-made carbon plates did not contain graphite fragments, while the Shanxi carbon plate contained graphite.

Fig.9 Raman spectra of (a) Shanxi carbon plates and (b) laboratory-made carbon plates

3.2.4 TG-DSC test

The carbon anode was qualitatively analyzed using the method described by Sofronov[23,24]. The TG analysis was carried out using a STA 449 f3 thermal analyzer (Germany, Netzsch). The sample was heated to 900 ℃ in air at a rate of 10 ℃/min. The results of differential thermal decomposition are shown in Fig.10. The oxidation indices are defined as follows:

Fig.10 TG-DSC spectra of different carbon plates

ΔA(%) = mass loss rate when the derivative of the DTG curve reaches its maximum value;

ΔB(%) = ratio of the second extreme value (in the temperature range 650-800 ℃) to the sum of the strengths of the two extremes on the DTG curve;

ΔT(℃) = difference between the first and second extreme values of temperature on the DTG curve.

The grade of the carbon plate was then assessed using the following criteria.

Quality grade 1: ΔA= 50%-55%, ΔB= 0%, ΔTreaches 40 ℃, and number of extremes (n) = 1, 2;

Quality grade 2: ΔA= 56%-65%, ΔB= 51%-56%, ΔT= 41-60 ℃, and number of extremes (n) = 2;

Quality grade 3: ΔA= 66%-70%, ΔB= 57%-60%, ΔT= 61-70 ℃, and number of extremes (n) = 2;

Quality grade 4: ΔA= 70%, ΔB= 61%-70%, ΔT= 71-80 ℃, and number of extremes (n) = 3.

Extreme temperature values of 724.83 ℃ and 748.38 ℃ were observed for the laboratory-made and Shanxi carbon plates, respectively. Table 5 shows the oxidation indices of shanxi carbon plates and laboratory-made carbon plate. The weight loss of the laboratory-made carbon plate at the extreme temperature value was 51.97%, and it should be classified as quality grade 1. The weight loss of the Shanxi carbon plate at the extreme temperature value was 48.99%. It did not meet any of the quality criteria in the classification method. This showed that the laboratory-made carbon plate was superior in actual industrial production.

Table 5 Oxidation indices of (a) Shanxi and (b) laboratorymade carbon plates

Table 6 Trace element contents of (a) laboratory-made and (b) Shanxi carbon carbon plates

3.2.5 SEM test

Fig.11 shows the fracture surface morphologies of the Shanxi and laboratory-made carbon plates. Fig.11(a) presents a coral-like morphology for the Shanxi carbon plate. Several small and large pores were unevenly distributed on the surface of the fracture. Additionally, small holes were interconnected with large holes. This indicated that the internal structure of the carbon plate was loose and porous and not very dense. Fig.11(b) presents a gully-shaped morphology. There were only a few pores on the surface of the fracture, which indicated that the carbon plate was inlaid well with various particles and formed a certain framework. The pores of various particles were fully filled with asphalt, forming a good mechanical structure.

Fig.11 SEM images of (a) Shanxi and (b) laboratory-made carbon plates

3.2.6 XRF test

Table 6 shows the contents of each trace element in the laboratory-made and Shanxi carbon plates. The contents of trace elements in carbon plates should be strictly controlled. Most trace elements adversely impact the electrolytic production of fluorine.

In the electrochemical reaction, nickel has a catalytic effect on the anode, which intensifies the selective oxidation of the anode, causing it to easily drop slag and impede the process.

An increase in sulfur content decreases corrosion resistance.

Increased contents of iron, calcium, and other elements in ash cause a significant increase in resistivity.

4 Conclusions

a) Carbon plates were prepared via high-temperature molding and freeze drying. The finished carbon anode had a density of 1.814 g/cm3, resistivity of 29.8 μΩ·m, and compressive strength of 89.28 MPa.

b) The laboratory-made carbon plates were superior to the Shanxi carbon plates in terms of density, resistivity, compressive strength, and degree of graphitization, as well as in terms of the results of TG-DSC, SEM, and other analyses. The laboratory-made carbon plates exhibited improved corrosion resistance and polarization resistance than those of the Shanxi carbon plates.

Acknowledgements

This study is nancially supported by the National key researchand development plan: Electrochemical preparation technologyof high value-added ne chemicals (Project No.2017YFB0307504). We sincerely acknowledge the experimentsite support of China National Nuclear 272 Uranium Industry Co. Ltd.