Investigation of li diffusion and intercalation processes in nanoporous carbon materials produced from carbides

INVESTIGATION OF Li DIFFUSION AND INTERCALATION PROCESSES IN NANOPOROUS
CARBON MATERIALS PRODUCED FROM CARBIDES
I.M. Kotina, V.M. Lebedev, A.G. Ilves, G.V. Patsekina, L.M. Tuhkonen, S.K. Gordeev1, M.A.
Yagovkina2, and Thommy Ekström3
1 Central Research Institute for Materials, 191014, St. Petersburg, Russia 2 Mehanobr-Analyt Co, 199026, St. Petersburg, Russia 3 Skeleton Technologies Group, SE-12653, Stockholm, Sweden INTRODUCTION
We first became interested in investigating Li behavior in solid solution due to the production of Ge(Li) detectors. In the manufacture of wide depletion layer of these detectors, the lithium, in addition to being thecompensating donor impurity, is also used to form the n+-layer of the n-i junction. Since the lithiumconcentration in diffused layer greatly exceeds the drift and room temperatures solid solubility, the lithiumtends to precipitate and the junction gradually deteriorates. We have investigated the precipitation kinetics ofLi in supersaturated solid solution in Ge crystals [1−2]. It was found that the nucleation time, the mechanismof nucleation, and Li resulting concentration is strongly related to the degree of supersaturation. We haveshowed that during nucleation time the structure formation of precipitates occurs. To draw on our experiencewe have carried out the investigation solid solution of Li in nanoporous carbons. Lithium can be inserted reversibly within most carbonaceous materials. Therefore, in the past decade, significant research efforts have been focused on the search for suitable carbon materials as an alternativeanode(s) for lithium rechargeable batteries. The main requirement to these materials is a high quantity ofreversible lithium ion insertion. This property appears to be a function of the carbon network structure [3−4].
The structural variation of carbon materials also plays an important role in the stoichiometry and the phasecomposition of the lithiated carbons.
In this paper, the process of lithium insertion in nanoporous carbons (NPC) obtained from carbides (SiC, TiC, Mo2C) by chlorination is studied. These materials have high-developed surface (total porosity is up to 70 %) on which lithium deposition can take place and high open nanoporosity that makes it possible tohave enough value of lithium diffusion coefficient. Thanks to this they look promising for the use in lithiumrechargeable batteries. Nanopore sizes in these samples depend on the used initial carbide and they are in therange 0.8-2.0 nm. Because the sizes of carbon bridges in NPC samples are of the order of pore sizes it isinteresting to determine if the formation of intercalation phases is possible in this system. It is important alsoto define lithium diffusion mechanism and the factors that affect the stoichiometry and phase composition oflithium insertion in NPC. These questions have been studied in this report.
EXPERIMENTAL
Bulk nanoporous carbon (NPC) samples used in this study were obtained by a high temperature chlorinating process from an intermediate product prepared on the base of carbide powders (SiC, TiC, Mo2C)of different content and crystal symmetry [5]. Therewith two type bulk NPC samples obtained fromSiC powder were available: C/SiC/ B and C/SiC/A which are characterized by the content ofpyrocarbon (PC) [6 ]. All bulk NPC materials have high total porosity - up to 70% and nanoporosityabout 50%. Their specific surface area is up to 1300 cm2 /g. Pore sizes and value of graphitised fragmentsin bulk NPC materials under study depend on initial carbide powder [7]. So pore sizes are the largest in theC/ Mo2C/B samples and are the smallest in the C/ SiC/ ones. At the same time, total volume of ordered graphite fragments is the largest in the C/SiC/B samples. The samples of C/ SiC/ A have large volume ofnanopore and do not almost contain PC.
Lithium insertion in the samples being studied was carried out by vacuum evaporation and subsequent diffusion at the temperatures ranging from 30 to 200 oC. The nuclear reaction 7Li(p,α)4He was used tomeasure the concentration profiles of lithium. Protons with energy Ep = 1.1 MeV were used and α-particles were registered with Si surface barrier detector [8]. The total quantity of the evaporated lithium wasmonitored with Si sample. It is known that in this temperature interval the diffusion of lithium into silicon does not take place in practice [9]. In such a situation the total quantity of evaporated Li is calculated byintegration of the Li concentration profile. After cooling the lithiated samples were kept in an atmosphere ofdry nitrogen.
The investigation of phase composition of the lithiated samples was performed by means of X-ray diffraction measurements using a “ Geigerflex” D/max-Rc Rigaku diffractometer with a Co X-ray tube.
Samples were plane-parallel plates much smaller in thickness than in cross-section. So the solution of the 1D diffusion equation was used to estimate the Li diffusion coefficient. For the case of an infinite sourceand not-interacting diffusing atoms, the solution is: where erfc denotes the complementary error function, N(x) is the Li concentration at any point x, t -diffusion time, D – the diffusion coefficient, No - the surface concentration of Li.
In some cases experimental lithium concentration profiles could be described by the equation (1) with a rather high accuracy, fig. 1. This made it possible to determine lithium coefficient diffusion DLi by fitting one erfc function the experimental profile. If fitting with one erfc was impossible, the DLi, values were estimated from eq. (1), in ten points of the experimental profile and then averaged <DLi>was found. Experimental results show that the value of <DLi> may be used as the characteristic parameter for the lithium diffusion process in the different NPC materials being studied at the same experimental conditions.
First we have carried out an investigation of the action of preliminary vacuum annealingon Li diffusion process. Table 1 shows <DLi> values in the samples under investigation inthe temperature range of 30 – 100 oC beforeand after annealing at 200 oC during 4 h.
Diffusion process was taken 5 min. One cansee that annealing leads to increase of <DLi> values. Effect of preliminary annealingdepends on the type of samples under studyand it is more marked in the NPC with largerpore sizes. Moreover, a clear correlation isobserved between the values of <DLi> andpore sizes:<DLi> values are maximum in the Fig. 1. Lithium diffusion profile in the C<TiC>B sample.
Temperature and time of diffusion are 800C and 10 min C/SiC/B ones. It is known that disordered respectively. Straight line is the best fit to experimental results carbon structure does not change at annealingtemperature up to 1100 oC , and the main result of low-temperature preliminary annealing NPC is the removal of water adsorbed on the surface ofpores [10]. Based on this we believe that obtained data is indicative of a diffusion of lithium along the porewalls. The validity this conclusion is also suggested by observed decrease of the influence of preliminaryvacuum annealing on the value of <DLi> with an increase of diffusion temperature (Table 1). A fraction of absorbed water is removed below 100 oC by heating of the sample to the diffusion temperature. As a result,the difference in the values of <DLi> before and after preliminary annealing decreases. As one would expect this effect is more marked in the samples with the larger pore sizes (Table 1) Values of diffusion coefficients in the samples of npC prior to and after preliminary annealing.
It should be noted that the following processes determining the value of DLi occur during the insertion of the lithium in carbons through the vacuum deposition and diffusion: solid-statediffusion, intercalation and accumulation reactions and phase transitions among intercalation stages.
Therewith there are two possible solid state diffusion mechanisms: diffusion on the pores anddiffusion within the Van der Waals space between every pair of carbon sheets in the quasigraphiteclusters. A more pronounced effect of preliminary annealing on the value of <DLi> in the samples with larger pore sizes indicates that the main mechanism of the Li diffusion in investigated NPC isthe diffusion along pore walls. The dependence of the diffusion coefficient on pore sizes canprobably be explained by taking into account that there is the interaction between the Li diffusingatoms and pore walls. In the case of the smaller pore the total wall surface is larger. It defines largerresistance to the flux of Li atoms and thereby smaller diffusion coefficient. In separate experiments we studied Li insertion in the annealed samples at different diffusionprocess durations. Lithium diffusion processes have been performed at temperature 80oC during 5,10 and 20 min. It turned out that the variation in the value of lithium diffusion coefficient, takingplace as a function of diffusion time, depends on NPC type (Table 2). From Table 2 we can see thatthe increase of diffusion time from 10 to 20 min results in the decrease of <DLi> in the C/SiC/B sample approximately by a factor of approximately 6 and in the C/SiC/A sample the <DLi> - less than by a factor of 2. At the same time in the C/TiC/B sample lithium diffusion coefficient isconstant in this time interval. Besides, for example, if at diffusion time of 20 min, the diffusioncoefficient varied from 4.1×10-9 cm2/sec for the C/SiC/B sample to 3.5×10-8 cm2/sec for theC/TiC/B sample, almost by a factor of 9; for diffusion time of 5 min the values of diffusioncoefficients of these samples differ by a factor of only 1.5. Diffusion coefficients at different durations of time of process. Diffusion temperature is 80oC We believe that there are several factors that retard the diffuse motion of the lithium and decrease the value of <DLi> with the increase of diffusion process duration. These factors are: the attractive interaction between lithium atoms located in the intercalation sites inside of Van derWaals spaces followed by the formation of staged phases and the accumulation of Li inside the pores followed by the formation of small lithium metal clusters. To regret in our experiments wewere not able to separate these factors.
XRD phase composition examination was carried out in the samples preliminary treated at 300 oC for 70 h to minimize the effects due to water. Note that the formation of intercalation phase isobserved mainly in the samples in which sufficient volume of the graphite-like fragments isavailable. This fact is in full accordance to results described in literature [11]. To follow the structural changes during diffusion process, XRD study has been performed from both sides of the sample under investigation. It should be emphasized that Li insertion in the samples does notgive a pure single-stage compound as determined by XRD patterns. Coexistence of several phases wasalways observed. Some of the results are shown in fig.2 for the C/SiC/B sample. In this case the durationof lithium diffusion process was 30 h. As one can see, from the front side pattern intercalation phasesLiC6, LiC12, LiC24 and large volume of carbonate Li2CO3 and carbide Li2C2 phases are registered. At the same time from the back of the sample only small quantity of Li2CO3 and Li2C2 is seen. We observed no evidence of Li2CO3 and Li2C2 existence, when sample was polished from the front side at the depth of 0.4 mm. However, the intercalation phases LiC6, LiC12. LiC24 are observed. Lithium concentration profiles measured on both sides indicate that in this case high lithium gradient is available in the sample (fig. 3).
From this fact it transpires that the formation of Li2CO3 and Li2C2 phases strongly depends on Li concentration. We suppose that at first Li clusters arise when high concentration of lithium is available inpores and then clusters react with atmospheric CO2 that is the cause of Li2CO3 formation. As for Li2C2 , additional experiments are required to understand the mechanism of its formation.
It is evident lithium filling pores are dependent on the lithium deposition rate. We have examined the phase composition of the lithiated samples at different rates of Li deposition and diffusion. The total volumeof carbonate and carbide phases is decreased with the decrease of Li deposition rate and the increase ofdiffusion rate. As a result, it was stated that ratio of deposition to diffusion rates of lithium plays animportant role in the phase composition forming of lithiated samples. Based on these results we haveprepared samples up to a thickness of about 0.6 mm uniformly impregnated by lithium and containingintercalation phases only. Fig. 2. X-ray diffractional patterns of lithiated C<SiC>B sample for front side (top) and back side (down).
Fig. 3. Concentration profiles measured on different
To estimate the graphite interlayer distance and the lithium intercalate sandwich thicknesses weused the expansion of complex peaks on the pattern in the angle interval 2θ =16 – 35 ° . It should benoted that peaks resulting from this expansion were broad. The relative content of the intercalationand graphite phases were determined by the integrating of the corresponding peaks. Therewith, thetotal content of these phases in the the sample under study was taken as 1. Obtained data are shown in Tabl.3. One can see from the table that the relative content ofintercalation phase is more in the C/SiC/B sample than one in the C/SiC/A sample (75% and 25 %respectivly). This is consistent with a large volume of graphite-like clusters in the C/SiC/B samplecompared to C/SiC/A one. As this takes place, the intercalation phases of different stage index (1and 3) coexist in the C/SiC/B:Li composite.
Table 3
Phase composition lithiated samples. Temperature of diffusion 80 oC, diffusion duration 40 min.
. Let us also note that the graphite interlayer distance as well as the third-stage interlayer distance inthe lithiated C/SiC/B sample is more than one in graphite, as the first-stage interlayer distance isless. As the greater part of carbon clusters in C<SiC>A is disordered they can not be intercalated byLi. However, there is a small share of graphitized clusters, which are intercalated as LiC6. The attention has been given to the fact that in this case the first- stage interlayer distance is more thanone in the intercalated C/SiC/B . Obtained values of the interlayer graphite distances indicate thatgraphite-like clusters is more ordered in the C/SiC/A sample, but in both samples clusters are lessordered than ones in graphite. As the content of intercalation phases is larger in the C/SiC/B:Li,afterwards we used the C/SiC/B material to create NPC-Li composites.
We have examined the phase composition of the lithiated samples at different Li deposition ratesand the durations of diffusion. The total content of carbonate and carbide phases decreased with thedeacrease of Li deposition rate and the increase of diffusion time. As a result, it was stated that theratio between deposition and diffusion rates of lithium plays an important in the phase compositionof lithiated samples. We used pulse evaporation process with a duration of the evaporation pulsemuch less than the interval between pulses in order to avoid of Li accumulation in pores. Thisprovided an opportunity to produce the samples uniformly impregnated by lithium and containingmainly intercalation phases. CONCLUSION
We have investigated the process of Li insertion in nanoporous carbons by vacuum evaporation and
subsequent diffusion at the temperatures ranging from 30 to 200 oC.
Given method of Li introduction with the use of XRD phase analysis makes it possible to conduct a rapid meaningful comparison between the various NPCMs, to reveal if there are ordered graphitized clasters inthem and thereby to evaluate suitablitity of NPCM samples for Li containers and for rechargable Li batteries.
Lithium diffusion in NPCM occurs along pore walls and within the Van der Waals space between every paircarbon sheets in the quasi-graphite clusters. One or another mechanism of diffusion effects on the value of<DLi> according to the diffusion duration. The value of <DLi> is big enough (10-9 –10-7 cm2/s). Diffusion process was studied at different deposition and diffusion rates .Different intercalation phasesappear depending on NPCM carbon skeleton structure, duration of diffusion process and concentration ofintroduced Li. The technology process conditions for producing of lithiated samples without lithium carbidand carbonate phases were found. The samples of a thickness up to 0.6 mm, uniformly impregnated bylithium and containing intercalation phases only were made.
Currently carbon materials, especialy with nanostructure have recently took a great deal of attention ashydrogen storage materials. Taking this into account the studies concerning to hydrogenadsorption/desorption in NPC materials and their hydrogen storage capacity are projected.
The studies were carried out on NPC samples developed and produced in collaboration with Skeleton REFERENCES
[1] I.M.Kotina, V.V.Kuryatkov, G.N.Mosina, S.R.Novikov, L.M.Sorokin, FTT (Soviet Physics), 26, 436
[2] I.M.Kotina, V.V.Kuryatkov, G.N.Mosina, S.R.Novikov, L.M.Sorokin, Preprint LNPI, 813, (1982).
[3] A.M.Wilson and J.R.Dahn, J.Electrochem.Soc. 142, 326 (1995).
[4] J.R.Dahn, Tao Zheng, Yinghu Liu, J.S.Xue, Science, 270, 590 (1995).
[5]Gordeev S.K,.Vartanova A.V, Zh. Prikl. Khim., 66 , 1080 (in Russian, 1994)
[6] I.M.Kotina, V.M.Lebedev, A.G.Ilves at el. VIII Intern. Conf. “Hydrogen Material Science and Chemistry of
Carbon Nano-Materials”, Sudak, Ukraine, September 14-20, 2003, p .866-67.
[7] R.N.Kyutt, E.A.Smorgonskaya.A.M.Danishevskii, S.K.Gordeev, A.V.Grechinskaya, Phys. Solid State, 41,
808 (1999).
[8] I.M.Kotina, V.M.Lebedev, G.V.Patsekina, (50 Meeting on Nuclear Spectroscopy and NuclearStructure, June 14-17, 2000, St-Petersburg, Russia), St-Petersburg, 396 (2000).
[9] Semiconductor Detectors, ed. by G.Bertolini and A.Coche (1968), North-Holland Publishing [10] Edward Buiel, J.R.Dahn, Electrochim. Acta, 45, 121 (1999).
[11] J.R.Dahn, A.K.Sleigh, H.Shi, J.N.Reimers, Q.Zhong, and B.M.Way, Electcrochim.Acta, 38, 1179
(1993).

Source: http://nrd.pnpi.spb.ru/pdf/kotina31.pdf