Corrosion behaviour of carbon materials exposed to molten lithium chloride–potassium chloride salt

CARBON

6 7 ( 2 0 1 4 ) 6 4 3 –6 5 5

Available at www.sciencedirect.com

ScienceDirect journal homepage: www.elsevier.com/locate/carbon

Corrosion behaviour of carbon materials exposed to molten lithium chloride–potassium chloride salt Jagadeesh Sure, A. Ravi Shankar, S. Ramya, C. Mallika, U. Kamachi Mudali

*

Corrosion Science and Technology Group, Indira Gandhi Centre for Atomic Research, Kalpakkam 603 102, India

A R T I C L E I N F O

A B S T R A C T

Article history:

The corrosion behaviour of four carbon materials namely low density graphite, high den-

Received 28 April 2013

sity graphite, glassy carbon and pyrolytic graphite were investigated in molten LiCl–KCl

Accepted 15 October 2013

electrolyte medium at 600 C for 2000 h under high pure argon atmosphere. Structural

Available online 23 October 2013

and microstructural changes in the carbon materials after exposure to molten chloride salt were investigated from the weight change and using scanning electron microscopy, atomic force microscopy, X-ray diffraction and laser Raman spectroscopic techniques. Microstructural analysis of the samples revealed the poor corrosion resistance of high density and low density graphite and severe attack was observed at several places on the surface. On the other hand, glassy carbon and pyrolytic graphite were relatively inert, while pyrolytic graphite showed the best corrosion resistance to molten salt attack. In the order of increasing corrosion resistance to molten salt, the carbon materials were found to follow the sequence: low density graphite < high density graphite < glassy carbon < pyrolytic graphite.  2013 Elsevier Ltd. All rights reserved.

1.

Introduction

Metallic fuel has been chosen for future fast breeder reactors in India and reprocessing of the spent metallic fuel through pyrochemical route involving eutectic mixture of LiCl–KCl molten salt medium is considered to be the best option [1– 4]. Development of molten salt based technologies are widely increasing in different fields viz. pyrochemical reprocessing of spent metallic fuels, accelerator driven transmutation technology and in molten salt reactors [1,4–7]. Selection of materials is one of the critical issues in molten salt based technologies. The material should be highly resistant to corrosion at high temperatures in molten salt environment. Carbon based materials are considered as candidates for the fabrication of containers (crucibles/liners) and electrodes in pyrochemical reprocessing [8–10]. The key advantages of graphite materials are the ease with which they can be fabricated into various shapes, their mechanical integrity at high

operating temperatures, high temperature strength, thermal shock resistance and inertness towards chlorine atmosphere [11]. The reaction of chlorine gas with graphite is reported to occur not below the electric arc temperature [12]. Different carbon based materials are used for various applications in molten salt environment. Pyrolytic graphite (PyG) is used as coating material on graphite,1 crucible in the Research Institute of Atomic Reactor’s (RIAR) process [11] and as electrode material [13]. Glassy carbon (GC) is widely employed as electrode and crucible material for various processes [14]. Glassy carbon was also used as anode crucible and working electrode for electrochemical measurements in pure LiF–NaK–KF, LiF–NaK–KF–UO2 and UF4 melts [15]. High density (HD) and low density (LD) graphite materials had been used as cathode processor, casting furnace crucibles [16,17], liners [16], electrode and tubes for chlorination at 700 C. 1 As the operating conditions for pyrochemical reprocessing in molten salt environment create an

* Corresponding author: Fax: +91 44 27480121. E-mail address: [email protected] (U. Kamachi Mudali). 1 www.nea.fr/pt/iempt8/abstracts/Abstracts/Sakamura.pdf. 0008-6223/$ - see front matter  2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.carbon.2013.10.040

644

CARBON

6 7 ( 2 0 1 4 ) 6 4 3 –6 5 5

aggressive corrosive environment for the process crucibles, liners and electrodes, these materials undergo drastic degradation, resulting in the replacement of components and accumulation of solid waste becomes a critical issue. Hence, there is considerable interest in the selection of materials and coatings for pyrochemical reprocessing application and it is important to understand the behaviour of materials in molten salt environment [9–11,18–22]. It should be noted that pyrochemical reprocessing operations are carried out in high purity argon atmosphere containing less than 1 ppm oxygen and water vapour and less than 15 ppm of nitrogen [4]. The corrosion behaviour of structural materials in molten salt under high purity inert atmosphere or in the presence of oxygen and the mechanisms involved have been investigated and reported in the literature [9,11,18–22]. The PyG and GC tested in molten fluoride environment at 540 C for 1–10 h and at 995 C for 1 h showed that the weight loss in PyG was nearly zero when compared to that of GC (0–0.5 wt.%) [5]. Magdziarz [23] studied the thermal oxidation behaviour of general grade graphite, GC, pre-baked and pyrolytic carbon in air and nitrogen atmospheres at 560 C for application in molten salt based FFC Cambridge electro reduction process and had recommended GC and pyrolytic carbon as anode materials for electro reduction at lower temperatures in an eutectic salt mixture of NaCl–CaCl2 bath, because of their high resistance to oxidation compared to graphite and prebaked carbon. Corrosion studies on ceramic materials and PyG in molten NaCl–KCl salt (for oxide fuel reprocessing) under Cl2 bubbling at 750 C revealed high corrosion resistance; however, PyG had undergone oxidation under Cl2–O2 (1:1) atmosphere in molten salt [11]. Graphite was used as the crucible for cathode processor in consolidating uranium from the cathode deposit comprising uranium, cadmium and 30 wt.% of salt [24]. Graphite crucible along with liner was used for melting and purification of molten salt in chlorine atmosphere [25]. Carbon materials viz. nuclear grade graphite, reactor graphite, GC and PyG are proposed as structural materials for molten salt reactors (families of fourth generation nuclear plant) in which molten fluoride salt is used both as fuel and coolant [6,26]. Hence, it is important to evaluate the corrosion behaviour of carbon materials in molten salts. Kamali et al. [27,28] found the molten salt corrosion of graphite to be advantageous for the production of carbon nanomaterials (nanotubes, nanoparticles and nanorods) either by electrolysis of molten alkali metal chlorides with graphite electrodes [27] or by eroding graphite in the presence of molten LiCl [28]. Various molten electrolytes (KCl, LiBr, NaCl and LiCl) are employed for electrochemical production of carbon nanostructures. Also, it was found that carbon nanostructure materials can be produced by heating a mixture of graphite and LiCl to 1250 C and cooling down to room temperature in a stream of air [28]. For the production of carbon nanostructures, graphite needs to be corroded in the molten salt; on the contrary, for molten salt reactors and pyrochemical reprocessing applications graphite materials should exhibit good corrosion resistance in molten salt. Characterisation techniques like Raman spectroscopy, Xray diffraction (XRD), scanning electron microscopy (SEM) and atomic force microscopy (AFM) have a key role in

understanding the microstructural features of carbon materials exposed to different environmental conditions [6,29]. Donnet and Wang [30] studied the surface morphology of four commercially available carbon materials (highly oriented pyrolytic graphite, synthetic graphite powder, carbon black and carbon fiber) using AFM and had compared the results with those obtained using scanning tunneling microscopy. Though there are several studies on the surface characterisation of carbon materials using SEM, AFM and Raman spectroscopy, the surface morphological changes associated with molten LiCl–KCl salt corrosion have not been reported. In the present study, the performance of the four carbon materials low density graphite (LDG), high density graphite (HDG), GC and PyG in molten chloride at 600 C was investigated and the morphological changes in the samples induced by molten LiCl–KCl salt after continuous exposure to 2000 h had been evaluated at micrometer scale using SEM, AFM, XRD and Raman spectroscopy. The possible mechanism by which surface degradation of carbon materials occurred in molten salt has also been discussed.

2.

Experimental

2.1.

Materials

Four different types of commercially available carbon materials namely LDG, HDG, GC and PyG were selected for the molten salt corrosion experiments. The properties of the four carbon materials are compared in Table 1. The variation in the physical properties of LDG, HDG, GC and PyG is mainly due to the method of preparation of these four carbon materials. Even though the method of preparation for LDG and HDG is same, they differ significantly in density, porosity, grain size and specific surface area because of variation in the baking temperature and the parameters used for compaction. LDG and HDG materials were of origin from Japan and supplied by M/s. Nickunj Pvt Ltd, Mumbai. GC was procured from M/s. SIGRADUR HTW GmbH, Germany. The open porosity of GC was negligible and its bulk density was moderate. PyG material used in the present study was received from Poland. The temperature of formation of PyG is the main controlling parameter in obtaining very fine graphite crystallite size, high density, negligible open porosity and low impurity content compared to other carbon materials. LDG, HDG and PyG were crystalline materials, whereas GC used in the present study was amorphous. For corrosion testing, cylindrical samples of dimensions 6 mm diameter and 50 mm length were fabricated from rods of LDG, HDG, GC and PyG by machining without disturbing the as-received surface. PyG, HDG and LDG flat samples (20 · 16 mm) machined out from a plate were also tested in molten salt. Prior to immersion testing in the molten salt, the weight of the carbon materials was measured. The weight recorded for the tested samples were LDG: 2.45 g (rod) and 1.42 g (flat); HDG: 2.52 g (rod) and 1.45 g (flat); GC: 2.42 g (rod) and PyG: 2.55 g (rod) and 1.48 g (flat). The eutectic composition of the molten salt was prepared from commercial LiCl and KCl (greater than 99% purity; procured from Sigma–Aldrich). The eutectic salt consisted of 44.48% LiCl and 55.52% KCl by weight. Corrosion activity in

CARBON

645

6 7 (2 0 1 4) 6 4 3–65 5

Table 1 – Properties of the carbon materials LDG, HDG, GC and PyG. Property

LDG

HDG

GC

PyG

Bulk density (g/cm ) Open porosity (%) Average grain size (lm) Specific surface area (m2/g) Impurity content (ppm)

1.66 20–22 32 25.2 Si-13.824, Ca-4.206, Al-0.134, Fe-0.175, Ni-0.292, Cr-0.007, Cu-0.042, Na-0.247, K-0.024, Mg-0.068

1.85 11–12 20 18.6 Si-8.253, Ca-3.475, Al-0.046, Fe-0.047, Ni-0.282, Cr-0.001, Cu-0.008, Na-0.004, K-0.003, Mg-0.051

1.54 Negligible – – Si-3.873, Al-0.274, Fe-0.925, Ni-0.127, Cr-0.092, Cu-0.034, Na-0.785, K-0.346, Ti-0.092, Zr-0.063

Flexural strength (MPa) Hardness (Shore)

40–43 67

78 83

210 48

2.22 Negligible 0.2 4.7 Si-0.927, Fe-0.0256, impurities of La, Nd, Ta, W, Na, Mg, A1, Sc, V and Mn in the order of 105% and 2–3% soot 120 37

3

molten salts depends on the major salt constituents and impurities in the salt. The molten salt can be prepared and maintained in such a way that moisture and hydroxides do not control the corrosion response. Hence special molten salt preparation procedure is required for commercial salts. These salts were dried at 150 C for 48 h under a vacuum of 103 mbar to eliminate moisture from them. After drying, chlorination was performed by purging high purity chlorine gas (99.99%) into the salt mixture at 500 C for 1 h to remove residual hydroxide ions and moisture content [31]. Excess chlorine and other gaseous impurities were removed by purging ultra high pure (UHP) argon gas through the salt. Purified LiCl–KCl salt only can be used as the processing medium for all pyrochemical reprocessing activities.

2.2.

Molten salt experiment

Since LiCl and KCl are highly hygroscopic in nature, all the operations were carried out in a double modular argon atmo-

sphere glove box. The corrosion studies were conducted in the molten salt test assembly (MOSTA), the schematic of which is shown in Fig. 1. MOSTA consists of an alumina crucible in a stainless steel cell and a thermocouple with provision for argon gas inlet and outlet. The stainless steel cell was sealed inside the glove box along with the thermocouple and was transferred to the MOSTA setup. The cell was connected to the argon lines of the furnace. The surface of the specimen was cleaned with distilled water and dried before testing. For each experiment 100 g of the purified salt mixture was loaded in the alumina crucible along with samples of carbon materials for corrosion testing. The test samples of LDG, HDG, GC and PyG were immersed in chlorinated molten LiCl– KCl eutectic salt at 600 C for 2000 h with continuous purging of UHP argon. Molten salt corrosion testing was carried out for 2000 h in order to understand the long term corrosion behaviour of carbon materials as well as to generate data for predicting the service life of structural materials by modelling. After corrosion testing in the molten salt, the samples were

Fig. 1 – Schematic of MOSTA for corrosion testing of carbon materials in molten LiCl–KCl medium.

646

CARBON

6 7 ( 2 0 1 4 ) 6 4 3 –6 5 5

cleaned ultrasonically in distilled water for a few minutes and using fresh distilled water each time to remove the salt present on the surface, followed by drying the samples. The weight of the carbon materials were recorded before and after corrosion test.

2.3.

Microstructural characterisation

The surface of the as-received and corrosion tested carbon materials were examined visually to find out the extent of surface degradation and variation in dimensions. The relative weight changes in the specimens after molten salt corrosion test were measured using an analytical balance (A&DHM202, Japan). The microstructural changes in the carbon materials exposed to molten salt were investigated by SEM (FE-SEM of FEI, Quanta 200F and ESEM of Philips XL-30) attached with energy dispersive X-ray spectroscopy (EDX). Topographical changes on the carbon surfaces were investigated using AFM mode of electrochemical scanning probe microscope (NT-MDT make Solver Pro EC). The samples were analysed in contact mode using the standard conical silicon tip attached to a cantilever under ambient conditions. The average roughness (Eq. (1)) [32] and Root Mean Square (RMS) (Eq. (2)) [32,33] values of the surfaces were calculated by NOVA software to evaluate the extent of corrugation present on the surface. Ra ¼

Xn i¼1

Zi  Z

sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Pn 2 i¼1 ðZi  ZÞ Rq ¼ n

3.

Results and discussion

3.1.

Visual examination and corrosion behaviour

Visual examination of the carbon samples after 2000 h of exposure to LiCl–KCl salt revealed the formation of pores on the surfaces of LDG and HDG and uniform attack was observed on the entire surface of these samples. The surface of as-received GC exhibited shiny and mirror-like appearance; however, after corrosion test the surface became marginally dull without any attack on the surface. The corrosion tested PyG surface exhibited a similar appearance as that of as-received PyG. Dimensional variations after corrosion testing was significant in the case of LDG, less pronounced in HDG and no change could be observed for GC and PyG samples. After exposure to molten salt at 600 C for 2000 h, weight gain of about 0.1875% and 0.0335% was observed in the case of LDG and HDG respectively. However, GC (9.2 · 104%) and PyG (3.3 · 104%) showed negligible weight loss after exposure to LiCl–KCl salt, indicating the possibility of penetration of molten salt into porous LDG and HDG compared to GC and PyG. LDG and HDG samples lost weight by degradation (Fig. 2) and gained weight by absorption of salt. More pronounced weight gain was observed for LDG when compared to HDG. The changes in weight indicated that GC and PyG materials had limited interaction and possess excellent compatibility with molten LiCl–KCl salt.

ð1Þ

3.2.

Microstructural studies and surface morphology

ð2Þ

where, Zi is the height of each single point, Z is the mean of all the height values and n is the number of data points within the image. XRD results obtained for LDG, HDG and PyG using a Philips X’pert MPD diffractometer (40 kV, 30 mA, Cu Ka radiation, k = 0.1542 nm) were analysed for phase identification by matching with standard JCPDS-ICDD values. The full width at half-maximum (FWHM) was determined by Lorentzian fitting of (0 0 2) peak. Raman microscopic analysis was performed on LDG, HDG and PyG using a HR 800 (Jobin Yvon) Raman spectrometer equipped with 1800 grooves/mm holographic grating. The samples were placed under an Olympus BXFM-ILHS optical microscope mounted at the entrance of the Raman spectrograph. Low energy He–Ne laser of wavelength 633 nm served as the excitation source since higher energy argon line was reported to affect the structure of graphitic materials due to absorption effect [34]. The laser spot size of 3 lm diameter was focused on the surface of the sample using a diffraction limited 10· (NA = 0.25) long distance objective and the laser power at the sample was 5 mW. The slit width of the monochromator was 300 lm, corresponding to a resolution of 4 cm1. The back scattered Raman spectra were recorded using super cooled (<110 C) 1024 · 256 pixels chargecoupled device (CCD) detector, over the range 80–2800 cm1 with an exposure of 5 s time and 20 CCD accumulations. All the spectral data were baseline corrected and deconvoluted. The data for Raman maps were collected over a selected area of interest and controlled by associated software.

The surface morphology of as-received and corrosion tested LDG and HDG samples are shown in Figs. 2 and 3 respectively. The microstructure of as-received LDG (Fig. 2a) and HDG (Fig. 3a) revealed pores, defects and micro cracks which are the active sites for molten salt to penetrate through the surface, causing internal corrosion either by loss of material or by absorption of salt. According to Vacik et al. [5] the penetration of molten salt depends on the exposure time and the temperature of the molten salt system. It is well known that graphite materials contain significant porosity (as shown in Figs. 2 and 3), and according to the mass transport theory reported by Castro and McEnaney [35], the salt penetration into the graphite structure is practically controlled by in-pore diffusion and by boundary layer diffusion at high temperatures. The corrosion behaviour of LDG and HDG was influenced by pore diffusion of molten salt into the graphitic structures. Owing to the availability of significant amount of open pores in the LDG when compared to HDG (Table 1), degradation of LDG (Fig. 2b) occurred faster than that of HDG (Fig. 3b). Generally molten salt corrosion is caused by the dissolution of the constituents of the material, selective leaching/attack, pitting, uniform surface corrosion and chemical reactions [36]. Any one or more of the above corrosion processes can take place depending on the nature of materials, testing medium, temperature and environment. Thermodynamic data for the reaction of carbon with LiCl and KCl are not available in the literature, and hence, the values of Gibbs free energy of these reactions could not be

CARBON

647

6 7 (2 0 1 4) 6 4 3–65 5

Pores

Wide pores Wide pore

Wide pore

Attacked region

2 Fig. 2 – Surface morphology of (a) as-received and (b) corrosion tested LDG in molten LiCl–KCl salt for 2000 h at 600 C. (A colour version of this figure can be viewed online.)

calculated. However, the reactions that can occur at 600 C are assumed to be as follows: C þ 4LiCl ! 4Li þ CCl4 ; C þ 4KCl ! 4K þ CCl4 ;

DG ¼ 1372 kJ=mol DG ¼ 1438 kJ=mol

ð3Þ ð4Þ

The DG values at 600 C for reactions (3) and (4) were calculated using FactSage Version 6.2 software. The positive values of DG of reactions (3) and (4) reveal that spontaneous chemical reaction between carbon and the molten salt is not possible at 600 C. In the case of LDG and HDG materials, carbon particles detaching from the surface of graphite and dispersing into the molten salt, thereby resulting in uniform surface attack (Figs. 2b and 3b) could be clearly observed. The pores were found to widen after exposure to molten chloride salt

at high temperatures, leaving behind a relatively more porous graphitic structure. The other possibility for significant corrosion of LDG and HDG materials is due to the presence of trace impurities in the chloride salt [5,19], hetero-elements bound to the edges and impurities (Table 1) present in the specimen. These impurities may form corrosive and gaseous products at high temperature and attack the carbon materials. The carbon particles flaked off from LDG and HDG after immersion test as observed in the magnified SEM images (Figs. 2b and 3b). Degradation of carbon materials in molten salt could be attributed to one or more of the following mechanisms: (1) adhesion of salt to graphite, (2) diffusion and filling the pores in graphite, (3) formation of intercalation compounds [37] and (4) removal of carbon particles [6]. The major difference between LDG and HDG with respect to corrosion behaviour is

648

CARBON

6 7 ( 2 0 1 4 ) 6 4 3 –6 5 5

Pores

Wide pores

Wide Pore

Attacked region

Attacked region

Fig. 3 – Surface morphology of (a) as-received and (b) corrosion tested HDG in molten LiCl–KCl salt for 2000 h at 600 C. (A colour version of this figure can be viewed online.)

their density and macro porosity. The molten salt attack on LDG and HDG appeared to be similar; however, variation in severity of attack was noticed. Molten salt can penetrate or can be absorbed into LDG compared to HDG, because of high specific surface area, lower density and high porosity features which accelerate the degradation of LDG by uniform surface attack, pore widening and dislodging carbon particles into the salt. The salt penetrates through the pores available on the surface into graphitic structure and thus, carbon particles are removed by the molten salt. This leads to enhancement in the absorption of molten salt and showing more weight gain. Figs. 2 and 3 depict the porous nature of the surfaces. The number and the pore size enlargement were clearly observed in both LDG (Fig. 2b) and HDG (Fig. 3b). The X-ray microanalysis (using EDX) revealed the presence of molten salt (K: 4.53 wt.% and Cl: 3.47 wt.%) along with carbon (C: 92.04 wt.%) in the microstructure of corrosion tested LDG; likewise, salt was present

in the microstructure of corrosion tested HDG. It was reported that as the exposure time increased the degradation increased considerably [9,10]. The surface morphologies of asreceived and corrosion tested GC are shown in Fig. 4a and b respectively. The lines in the micrograph of the as-received GC (Fig. 4a) sample may be due to preparation of the surface by milling/machining of the material. After corrosion test in LiCl–KCl salt the surface morphology appeared to be smooth (Fig. 4b) and there was no significant attack. Initiation of pores was not observed on the surface even after corrosion testing for 2000 h because of the non-wetting characteristic of GC by molten salts. Even though the density of GC was quite low, it was non-porous due to the absence of large open pores and the presence of small closed pores of a few nanometer in size [38]. Owing to closed porosity, penetration of molten salt into GC was not evident. High temperature stability and absence of long-range order in the structure also make GC chemically more inert

CARBON

6 7 (2 0 1 4) 6 4 3–65 5

649

Fig. 4 – Surface morphology of (a) as-received and (b) corrosion tested GC in molten LiCl–KCl salt for 2000 h at 600 C.

Fig. 5 – Surface morphology (back scattered electron) of (a) as-received, (b) corrosion tested PyG; secondary electron micrographs and EDX spectra of (c) as-received and (d) corrosion tested PyG in molten LiCl–KCl salt for 2000 h at 600 C. (A colour version of this figure can be viewed online.)

towards molten salt than LDG and HDG materials [5]. As-received PyG surface exhibited convex feature and pore morphology was not observed in the back scattered SEM micrographs (Fig. 5a). PyG immersed in molten LiCl–KCl salt for 2000 h did not show any evidence of degradation and attack (Fig. 5b). Magnified SEM images and corresponding EDX spectra of PyG surface before and after corrosion test in molten LiCl–KCl salt are shown in Fig. 5c and d. EDX analysis confirmed the absence of change in elemental composition after corrosion test in molten salt. The surface of PyG showed only carbon peak (100 wt.%) before and after corrosion test and no impure element was identified. The preparation method for PyG at very high temperatures causes high graphitization behaviour, availability of less active sites and hetero-elements in its structure leading to lower corrosion rate than GC, HDG

and LDG. Vacik et al. [5] reported that the salt penetration depths in the case of molten LiF tested GC and PyG samples were very less and negligible. Thus, GC and PyG are called as impermeable grades of carbon materials [7]. The present study also suggests that the penetration of molten salt into carbon materials can be avoided if impermeable graphite grades like GC and PyG are used. The penetration and absorption of molten salt into graphitic structures can proceed in different modes depending on the arrangement of graphite planes in the structure. The observation of restricted penetration of molten salt into non-porous material and accelerated salt penetration into porous structures is upheld by the results reported by Vacik et al. [5]. The degradation of graphite depends on the degree of structural order, surface uniformity, preparation method

650

CARBON

6 7 ( 2 0 1 4 ) 6 4 3 –6 5 5

and density of the material. Though GC showed excellent corrosion resistance to molten LiCl–KCl salt, for the intended structural application only LDG, HDG and PyG were considered and further investigation was made with these three materials in the present work. AFM is a powerful tool for accurate evaluation of surface morphology of carbon based materials [29,30]. Fig. 6 shows the typical AFM morphology of as-received and corrosion tested LDG, HDG and PyG materials in molten LiCl–KCl salt for 2000 h. Aggressive attack by molten salt was observed on the surface of LDG (Fig. 6a and b) and HDG (Fig. 6c and d). The topography of LDG and HDG surfaces revealed imperfections owing to surface attack induced by LiCl–KCl salt. The

topography also showed that the surface layer was removed in the form of flakes (Fig. 6b and d) and lots of pores were formed. The surface roughness values of carbon materials calculated using AFM results are listed in Table 2. As-received LDG and HDG surfaces became rough after exposure to molten salt. The increase in the surface roughness was due to the increase in the size of pores as well as creation of pores over the surface because of the removal of carbon particles from HDG and LDG by molten salt. The deposition and penetration of salt through the surfaces of LDG and HDG resulted in increased roughness. The surface topography of as-received and corrosion tested PyG is shown in Fig. 6e and f respectively. The AFM

a

b

c

d

e

f

Fig. 6 – AFM topography of as-received and molten salt corrosion tested LDG (a) and (b), HDG (c) and (d) and PyG (e) and (f). (A colour version of this figure can be viewed online.)

CARBON

651

6 7 (2 0 1 4) 6 4 3–65 5

Table 2 – Surface roughness of as-received and corrosion tested carbon materials using AFM. Carbon material

LDG HDG PyG

Average roughness (nm)

RMS roughness (nm)

As-received

Corrosion tested

As-received

Corrosion tested

294.3 104.5 25.5

520.5 252.1 22.4

400.1 135.9 31.9

640.8 321.4 27.2

topography showed nodular morphology, which did not change even after immersion in molten salt for 2000 h. This nodular morphology is characteristic of the deposition method of PyG [39]. The surface roughness of as-received PyG was low compared to LDG and HDG materials. After immersion in molten salt the surface roughness of PyG decreased (Table 2) and surface became smoothened. The decrease in surface roughness could be attributed to the removal of soot formed during deposition on the surface. This feature was also observed in the SEM images of PyG (Fig. 5a and b). AFM observations were in accordance with the SEM microstructures of LDG, HDG and PyG. Structural disorder increased in LDG and HDG owing to high corrosion in molten salt. The degradation and corrosion behaviour of carbon materials depend upon the nature of hetero-elements present at the edges of carbon planes, the nature of precursors, defects and heat treatment conditions at the time of manufacturing [35]. The major differences among LDG, HDG, GC and PyG are in the graphite structure, macro porosity and the impurity content. The structural features and absence of macro porosity in GC and PyG contributed to their inertness towards molten LiCl– KCl salt at 600 C under UHP argon atmosphere when compared to LDG and HDG.

3.3.

X-ray diffraction and Raman spectroscopy studies

XRD patterns of the as-received and immersion tested carbon materials in molten LiCl–KCl salt are shown in Fig. 7. These patterns were analysed by matching with the data in JCPDS diffraction file number: 41-1487 [40]. XRD patterns of LDG (Fig. 7a) and HDG (Fig. 7b) showed standard graphite diffraction peaks before and after corrosion test. The lattice planes of LDG and HDG were (0 0 2), (1 0 0), (1 0 1) (0 0 4), and (1 1 0) while that of PyG (Fig. 7c) showed (0 0 2) and (0 0 4) with highly preferred orientation along (0 0 2) plane before and after corrosion test. The XRD data neither revealed any major change in the diffraction planes nor the formation of any new compound after the corrosion test. Salt phase could not be identified in the diffraction patterns because the amount of absorbed salt into the porous structure might be within the detection limit of XRD. However, the salt present in the LDG and HDG was identified by EDX point analysis. The diffraction patterns of LDG and HDG were observed to be similar because of no structural difference and use of same processing method; with only variation in the physical properties (Table 1). The FWHM values of the (0 0 2) peak of the LDG, HDG and PyG were calculated to understand the disorder produced in the carbon materials from as-received to corrosion tested ones. The FWHM of the as-received samples were

0.299, 0.478, and 0.831; while in case of corrosion tested materials were 0.391, 0.492 and 0.835 for LDG, HDG and PyG respectively. Although the differences are small, an increase in FWHM of (0 0 2) peak of LDG and HDG indicate the increasing structural disorder of graphite after the corrosion test in molten salt. The increase in FWHM was quite low from as-received to corrosion tested PyG; hence, significant disorder was not observed in the PyG after exposure to molten LiCl–KCl salt. The arrangement of crystallites in turbostratic structure is reported to have high degree of preferred orientation in PyG [41]. The graphitic planes are oriented towards c-axis perpendicular to the plane of deposition in PyG. The turbostratic structure, graphitization, high preferred orientation and high density of PyG prevent the permeability of molten salt into its structure. The XRD patterns confirmed the absence of reaction of graphite materials with molten LiCl– KCl salt at 600 C. The peaks corresponding to intercalation compounds were not observed in all the carbon materials after exposure to molten salt for 2000 h. Laser Raman spectroscopy is a promising tool to acquire additional in-depth knowledge of microstructural information of carbon based materials, especially in the wave number region between 1000 and 1800 cm1 for providing interesting information related to microstructural features [34,42]. Raman investigation of different carbon materials was conducted elaborately by different research groups in the past [34,42]. However, very little information is available on the Raman analysis of carbon materials after molten salt corrosion test and no Raman mapping (imaging) study is reported after molten salt test. The Raman spectra recorded for the carbon materials before and after molten salt immersion are shown in Fig. 8. In order to quantify the effect of molten salt exposure, Raman spectra of unexposed graphite specimens were compared with that of molten salt exposed samples. Raman spectra of as-received and corrosion tested LDG, HDG and PyG samples showed peaks around 1337, 1582 and 2700 cm1, which were attributed to D, G and G 0 bands respectively. All Raman spectra recorded looked alike and had similar spectral signatures around 1337 and 1582 cm1, implying that the fraction of disorder in the microstructures of all three types of graphite is more-or-less the same. The position of the D band shifted by several wave numbers than expected. There are literature reports for the dependency of D band on laser wavelengths [43–46]. The D band shifted from 1360 to 1330 cm1 when the excitation wavelength was increased from 488 to 647 nm. The wavelength of 633 nm was used in the present Raman investigation and the observed D band values were found to be in good agreement with the reported values [43].

652

CARBON

6 7 ( 2 0 1 4 ) 6 4 3 –6 5 5

Fig. 7 – XRD patterns of (a) LDG, (b) HDG and (c) PyG materials before and after corrosion testing in molten salt at 600 C for 2000 h.

The D band frequency was found to be consistent in all types of carbon materials whether exposed or unexposed to molten salt. Wang et al. [43] reported that the intensity of D band depends on edge density and does not depend on the size of the microcrystallites; the D band frequency is dependent only on the laser wavelength. The Raman spectral analysis clearly indicated that all the carbon materials under

Normalised Intensity (a.u)

(f)

(e)

(d) (c) (b) (a)

1200

1600

2000

2400

2800

-1

Raman Shift (cm )

Fig. 8 – Raman spectra of as-received and molten salt exposed carbon materials; (a) LDG – as received, (b) LDG – molten salt exposed, (c) HDG – as received, (d) HDG – molten salt exposed, (e) PyG – as received and (f) PyG – molten salt exposed.

investigation were partially disordered polycrystalline in nature. Analysis of intensity, peak position and band width also gave useful information on the microstructural features of carbon materials. The integrated intensity ratio, ID/IG for the D and G bands is widely used for quantifying the defect in graphitic materials [34]. The band width and ID/IG ratio calculated from Raman spectra of as-received and corrosion tested LDG, HDG and PyG are listed in Table 3. The ID/IG ratio for LDG was comparatively higher than that for HDG and PyG materials, which indicates that the disorder in LDG was more than HDG and PyG. The corrosion tested LDG samples were also observed to possess higher ID/IG ratio than that of the tested HDG and PyG specimens. It is also observed that the band width of the corrosion tested LDG was higher than that of other carbon materials. Generally, greater the band width, higher is the degree of disorderness of carbon materials [43]. Raman investigation on different carbon materials after their exposure to molten salt revealed that PyG and HDG samples had ordered structure and free from interstitial defects when compared to LDG sample and the degree of disorderness was the least for PyG. To investigate the D band variation (semi quantitative) and the effect of molten salt on the degree of disorder in the molten salt exposed specimens, Raman mapping was carried out over the surface area of 80 · 80 lm2. Generally, Raman map gives the intensity distribution of one particular spectral range from which, changes in the composition of a specific component can be determined.

CARBON

Table 3 – Raman spectroscopic parameters derived for carbon materials. Sample

Peak position t (cm1)

LDG As-received D 1336 G 1588 Corrosion tested D 1334 G 1584 HDG As-received D 1330 G 1584 Corrosion tested D 1333 G 1586 PyG As-received D 1342 G 1592 Corrosion tested D 1340 G 1590

D Band width x1/2 (cm1)

ID/IG

48

1.03

59

1.36

43

0.87

56

1.26

The degree of disorder (high D band distribution) was higher in LDG (Fig. 9a and d) compared to HDG (Fig. 9b and e) and PyG (Fig. 9c and f) which is indicated as white colour in the mapping image. Raman mapping analysis performed on the carbon materials after immersion in molten salt revealed that high disorder (indicated by white colour intensity) was introduced in the graphitic structure of LDG and HDG and the modification in the D band of PyG was negligible. Raman study on graphite tested in molten salt by Bernardet et al. [6] confirmed the molten salt induced disorderness in the structure of the three carbon materials. Analysis of integrated intensity ratio suggested the introduction of minor defect in HDG and PyG after molten salt corrosion test. However, the Raman mapping ascertained that the quantum of disorder introduced in PyG was less than that in HDG.

4. 40

0.83

42

1.21

The Raman map also represents the image distribution of one particular wave number range (corresponding to one component). Average Raman spectra and imaging (mapping) were recorded before and after the testing of carbon materials in molten salt.

LDG

653

6 7 (2 0 1 4) 6 4 3–65 5

HDG

Summary

Among the carbon materials LDG, HDG, GC and PyG, excellent corrosion resistance was exhibited by PyG and GC when exposed to molten LiCl–KCl salt medium for 2000 h at 600 C under UHP argon atmosphere. Morphology of carbon materials by SEM and AFM analysis revealed that the surface of LDG and HDG were severely attacked in molten LiCl–KCl salt and the penetration of salt into GC and PyG was insignificant. Xray diffraction analysis of carbon materials after immersion test showed no evidence of salt phase on the surface because the absorption of salt by the carbon materials was less than the limit of detection by XRD. Intercalation of molten salt between the graphite layers did not occur as evident by XRD. The integrated intensity (ID/IG) ratio and band width analysis

PyG

Fig. 9 – Raman mapping of D band of carbon materials (a) LDG – as received, (b) HDG – as received, (c) PyG – as received, (d) LDG – molten salt exposed, (e) HDG – molten salt exposed, (f) PyG – molten salt exposed. (A colour version of this figure can be viewed online.)

654

CARBON

6 7 ( 2 0 1 4 ) 6 4 3 –6 5 5

by Raman mapping confirmed the extent of disorder created by molten salt was least in PyG. The corrosion results derived from this study indicated that the resistance of carbon materials to molten LiCl–KCl salt under UHP argon atmosphere at 600 C is in the following order: LDG < HDG < GC < PyG.

Acknowledgements The authors are indebted to Dr. B. Prabhakara Reddy, Chemistry Group, IGCAR for his support in the preparation of molten salt and to Shri K. Thyagarajan and Dr. S.C. Vanitha Kumari of CSTG, IGCAR for their help in the MOSTA operation and AFM characterization respectively. The authors also acknowledge Shri. Shayamala Rao Polaki and Ms. Kalavathy, Materials Science Group, IGCAR for their help in FE-SEM and XRD studies respectively. The authors are thankful to Dr. Ch. Jagadeeswara Rao and Mr. Pradeep Kumar Samantaroy for useful discussions.

R E F E R E N C E S

[1] Simpson MF. Introduction to the pyroprocessing special issue. Nucl Technol 2008;62(2):117. [2] Raj B, Kamath HS, Natarajan R, Rao PRV. A perspective on fast reactor fuel cycle in India. Prog Nucl Energ 2005;47(1– 4):369–79. [3] Nishimura T, Koyama T, Iizuka M, Tanaka H. Development of an environmentally benign reprocessing technology– pyrometallurgical reprocessing technology. Prog Nucl Energ 1998;32(3–4):381–7. [4] Laidler JJ, Battles JE, Miller WE, Ackerman JP, Carls EL. Development of pyroprocessing technology. Prog Nucl Energy 1997;31(1–2):131–40. [5] Vacik J, Naramoto H, Cervena J, Hnatowicz V, Peka I, Fink D. Absorption of molten fluoride in glassy carbon, pyrographite and hastealloy. J Nucl Mater 2001;289(3):308–14. [6] Bernardet V, Gomes S, Delpeux S, Dobois M, Guerin K, Avignant D, et al. Protection of nuclear graphite toward fluoride molten salt by glassy carbon deposit. J Nucl Mater 2009;384(3):292–302. [7] Manly WD, Allen JW, Cook WH, DeVan JH, Douglas DA, Inouye H, et al. Construction materials for molten salt reactors. In: MacPherson HG, editor. Molten salt reactors: Part-II. Tennessee: Oak Ridge National Laboratory Press; 1958. p. 595–625. [8] Song KC, Lee H, Hur JM, Kim JG, Ahin DH, Cho YZ. Status of pyroprocessing technology development in Korea. Nucl Eng Technol 2010;42(2):131–44. [9] Sure Jagadeesh, Ravi Shankar A, Ramya S, Kamachi Mudali U. Molten salt corrosion of high density graphite and partially stabilized zirconia coated high density graphite in molten LiCl–KCl salt. Ceram Int 2012;38(4):2803–12. [10] Sure Jagadeesh, Ravi Shankar A, Ramya S, Kamachi Mudali U. Evaluation of corrosion behaviour of plasma sprayed partially stabilized zirconia coated and uncoated high density graphite in molten LiCl–KCl salt. In: Sudarshan TS, Nyle´n Per, editors. Surface modification technologiesXXV. VALAR Docs Publications; 2012. p. 235–44. [11] Takeuchi M, Kato T, Hanada K, Koizumi T, Aose S. Corrosion resistance of ceramic materials in pyrochemical reprocessing condition by using molten salt for spent nuclear oxide fuel. J Phys Chem Solids 2005;66(2–4):521–5.

[12] Pierson HO. Handbook of carbon, diamond and fullerenes. Park Ridge NJ: Noyes; 1993. p. 67. [13] Inoue T, Koch L. Development of pyroprocessing and its future direction. Nucl Eng Technol 2008;40(3): 183–90. [14] Nakamura K, Morooka H, Tanabe Y, Yasuda E, Akatsu T, Shindo H. Surface oxidation and/or corrosion behavior of glass-like carbon in sulfuric and nitric acids, and in aqueous hydrogen peroxide. Corros Sci 2011;53(12):4010–3. [15] Boussie H. Pyrometallurgical processing research program ‘‘PYROREP’’-final report; 2003. [16] Committee on Electrometallurgical Techniques for DOE Spent Fuel Treatment, Board on Chemical Sciences and Technology. Electrometallurgical techniques for DOE spent fuel treat. Washington, DC: National Academy Press; 2000. p. 1–116. [17] Benedict RW, Henslee SP. EBR-II spent treatment demonstration project status. In: 5th international nuclear conference on recycling, conditioning and disposal, Nice, France; 1998. p. 149–55. [18] Ravi Shankar A, Kamachi Mudali U. Corrosion of type 316L stainless steel in molten LiCl–KCl salt. Mater Corros 2008;59(11):878–82. [19] Indacochea JE, Smith JL, Litko KR, Karell EJ. Corrosion performance of ferrous and refractory metals in molten salts under reducing conditions. J Mater Res 1999;14(5):1990–5. [20] Kamachi Mudali U, Ravi Shankar A, Ningshen S, Girija S, Ravi Kumar S, Thyagarajan K. Materials development and corrosion issues in the back end of fuel cycle. Energy Procedia 2011;7:463–73. [21] Ravi Shankar A, Kamachi Mudali U, Ravi Kumar S, Khatak HS, Raj Baldev. Plasma sprayed yttria stabilized zirconia coatings on type 316L stainless steel for pyrochemical reprocessing plant. J Nucl Mater 2008;372(2–3):226–32. [22] Ravi Shankar A, Mathiya S, Thyagarajan K, Kamachi Mudali U. Corrosion and microstructure correlation in molten LiCl– KCl medium. Metall Mater Trans A 2010;41(7):1815–25. [23] Magdziarz A. Thermal gravimetric analysis of general-grade graphite, glassy prebaked and pyrolytic carbons. Arch Metall Mater 2006;51(3):471–4. [24] Lee H, Park G-I, Kang K-H, Hur J-M, Kim J-G, Ahn D-H, et al. Pyroprocessing technology development at KAERI. Nucl Eng Technol 2011;43(4):317–28. [25] Kamachi Mudali U, Ravi Shankar A, Mallika C, Thyagarajan K, Ningshen S, Reddy BP et al. Materials and coating technologies for pyrochemical reprocessing of spent metallic fuel from fast breeder reactors. Kalpakkam India, Internal IGCAR Report, Number:IGC/MMG/CSTG/2011. [26] Feng S, Xu L, Li L, Bai S, Yang X, Zhou X. Sealing nuclear graphite with pyrolytic carbon. J Nucl Mater 2013;441(1– 3):449–54. [27] Kamali AR, Schwandt C, Fray DJ. Effect of the graphite electrode material on the characteristics of molten salt electrolytically produced carbon nanomaterials. Mater Charact 2011;62(10):987–94. [28] Kamali AR, Fray DJ. Molten salt corrosion of graphite as a possible way to make carbon nanostructures. Carbon 2013;56:121–31. [29] Vakar Z, Denisov EA, Kompaniets TN, Makarenko IV, Marushchak VA, Titkov AN. The surface morphology of pyrolytic graphite irradiated by hydrogen atoms. Tech Phys 2001;46:773–7. [30] Donnet JB, Wang TK. AFM observation of carbon materials. Analysis 1994;22:M24–6. [31] Maricle DL, Hume DN. A new method for preparing hydroxide-free alkali chloride melts. J Electrochem Soc 1960;107(4):354–6.

CARBON

6 7 (2 0 1 4) 6 4 3–65 5

[32] Padhy N. Role of passivity and surface modification on the corrosion behavior of AISI 304L stainless steel in nitric acid medium. [Ph.D. thesis] Mumbai, India: Homi Bhabha National Institute; 2010. [33] Khulbe KC, Feng CY, Matsuura T. Synthetic polymeric membranes-characterization by atomic force microscopy. Springer Lab Publications; 2008. p. 19–45. [34] Jawhari T, Roid A, Casado J. Raman spectroscopic characterization of some commercially available carbon black materials. Carbon 1995;33(11):1561–5. [35] Castro LDD, McEnaney B. The control of high temperature corrosion of engineering carbons and graphites. Corros Sci 1992;33(4):527–43. [36] Kane R. Molten salt corrosion. In: Cramer S, Covino B, editors. ASM handbook: corrosion–fundamentals, testing and protection, vol. 13A. Ohio: ASM International; 2003. p. 216–9. [37] Tanaike O, Inagaki M. Degradation of carbon materials by intercalation. Carbon 1999;37(11):1759–69. [38] Fitzer E, Schaefer W, Yamada S. The formation of glasslike carbon by pyrolysis of polyfurfuryl alcohol and phenolic resin. Carbon 1969;7(6):643–6.

655

[39] Lewis JC, Floyd IJ. Reorientation effects in vitreous carbon and pyrolytic graphite. J Mater Sci 1966;1(2):154–9. [40] Powder Diffraction File-2 (PDF-2). JCPDS-ICDD, PCPDFWIN Version 2.02; May 1999. [41] Fialkov AS, Baver AL, Sidorov NM, Chalkun ML, Rabinovich SM. Pyrographite – preparation, structure, properties. Russ Chem Rev 1965;34(1):46–58. [42] Nakamizo M, Kammereck R, Walker PL. Laser Raman studies on carbons. Carbon 1974;12(3):259–67. [43] Wang Y, Alsmeyer DC, McCreery RL. Raman spectroscopy of carbon materials: structural basis of observed spectra. Chem Mater 1990;2(5):557–63. [44] Vidano RP, Fischbach DB, Willis LJ, Loehr TM. Observation of Raman band shifting with excitation wavelength for carbons and graphites. Solid State Commun 1981;39(2):341–4. [45] Baranov AV, Bekhterev AN, Bobovich YS, Petrov VI. Interpretation of certain characteristics in Raman spectra of graphite and glassy carbon. Opt Spectrosc 1987;62:612–6. [46] Mernagh TP, Cooney RP, Johnson RA. Raman spectra of graphon carbon black. Carbon 1984;22(1):39–42.