Journal of NUCLEAR SCIENCE and TECHNOLOGY, Vol. 44, No. 1, p. 73–80 (2007)
ARTICLE
Lithium Isotope Effects Accompanying Electrochemical Insertion of Lithium into Metal Oxides Masahiro MOURI, Kei ASANO, Satoshi YANASE and Takao OI Department of Chemistry, Sophia University, 7-1 Kioicho, Chiyoda-ku, Tokyo 102-8554, Japan (Received April 27, 2006 and accepted in revised form September 25, 2006) Lithium was electrochemically inserted into SnO2 and Fe2 O3 –SiO2 binary oxide from the mixed solution of ethylene carbonate and methyl ethyl carbonate containing 1 M LiClO4 , and the lithium isotope effect accompanying the insertion was investigated. Tin(IV) oxide showed a slight selectivity towards the heavier isotope of lithium with the separation factor (S) ranging from 0.982 to 0.999 at 25 C. Fe2 O3 –SiO2 binary oxide, on the other hand, showed no specific lithium isotope selectivity with S scattered around unity at the same temperature. These results were reasonably understood by assuming the reduction of the metals in the host materials from Sn(IV) to Sn(0) and from Fe(III) to Fe(II) and the formation of lithium oxide upon Li insertion. KEYWORDS: lithium, lithium isotopes, isotope separation, isotope effects, electrochemical insertion, metal oxides, Fe2 O3 –SiO2 binary oxide, tin(IV) oxide, separation factor
isotope enrichment in the future. The reaction of Eq. (1) is in principle a redox reaction between the neutral lithium atom and the monovalent lithium ion. There may be some non- or less toxic material that can be a medium for the redox reaction of lithium. In this context, we paid attention to electrode materials for lithium ion secondary batteries. Those materials can take up lithium from lithium ion-bearing electrolyte solution through the redox reaction. Thus, if they show large lithium isotope effects, it may be possible to establish a system in which lithium isotopes are enriched by utilizing the charge-discharge cycles of lithium ion secondary batteries. In previous papers,5,6) we investigated the lithium isotope effects accompanying electrochemical insertion of lithium from organic electrolyte solution to graphite and tin. As is well known, graphite is used as the cathode of lithium ion secondary batteries and tin is investigated as possible electrode material.7) Both the materials preferentially took up 6 Li like mercury, although lithium isotope effects on them are smaller than that of mercury. In this paper, we report lithium isotope effects on metal oxides, more specifically, crystalline tin(IV) oxide and amorphous iron(III) oxide-silicon oxide binary oxide.8–10) Both oxides have also been studied as possible electrode materials of lithium ion secondary batteries.
I. Introduction Lithium has been and still is a main target of isotope separation study. This is because in part of the importance of the isolated and enriched isotopes of lithium. A large demand for the lighter isotope of lithium, 6 Li, is expected in DT fusion power reactors in the future where lithium compounds rich in 6 Li will be required for the tritium breeder blanket. The only method that was applied to a large-scale lithium isotope separation is so called amalgam method.1) In this method, lithium is distributed between the amalgam phase and the aqueous or organic electrolyte solution phase and lithium isotope separation is practiced based on the lithium isotope exchange reaction between the two phases, 7
Li(Hg) þ 6 Liþ (SOL) ¼ 6 Li(Hg) þ 7 Liþ (SOL);
ð1Þ
where A Li(Hg) and A Liþ (SOL) denote the isotope A surrounded by mercury atoms in the amalgam phase and solvent molecules in the electrolyte solution phase, respectively. The reported value of the 7 Li-to-6 Li single-stage separation factor, S, defined as, S ¼ (7 Li/6 Li)sol =(7 Li/6 Li)Hg ; where (7 Li/6 Li)sol and (7 Li/6 Li)Hg denote the 7 Li/6 Li isotopic ratios in the solution and amalgam phases, respectively, is 1.049–1.062,2) meaning that 6 Li is preferentially fractionated in the amalgam phase. The value of 1.049– 1.062 is large compared with those of ion exchange (up to about 1.003 with organic ion exchangers) and crown ethercryptand systems (1.002 to 1.047).3,4) Although a large lithium isotope effect is attractive, the use of toxic mercury conveys biological and environmental problems and makes the amalgam method difficult to be applied to large-scale lithium
II. Experimental 1. Host Materials of Lithium Insertion and Reagents SnO2 powder was purchased from Kojundo Chemical Lab Co. Ltd. Fe2 O3 –SiO2 binary oxide with the Fe2 to Si mole ratio of 3:7 was synthesized by the sol-gel method as follows.9) Fe(NO3 )3 9H2 O (28.0 g) was dissolved with a mixed solution of H2 O (10.0 cm3 ), C2 H5 OH (3.0 cm3 ) and CH3 COOH (0.24 g) to obtain an iron(III) nitrate solution. A mixed solution of Si(OC2 H5 )4 (TEOS; 6.0 cm3 ) and C2 H5 OH (1.0 cm3 ) was prepared and stirred with a magnetic stirrer for 30 min. To this TEOS solution slowly added was
Corresponding author, E-mail:
[email protected]
This article was received and accepted as ‘‘Original Paper’’.
Atomic Energy Society of Japan 73
74
M. MOURI et al.
(a)
(b)
SnO 2
copper foil
SnO 2 400 µm
400 µm
copper foil (c)
Fe 2 O3 −SiO2
(d)
Fe 2 O3 −SiO2
copper foil
200 µm
copper foil
200 µm
Fig. 1 Optical photos of the cross-sections of a SnO2 electrode (a) before and (b) after the lithium insertion (electrolysis) and of a Fe2 O3 –SiO2 electrode (c) before and (d) after the lithium insertion
the iron(III) nitrate solution to obtain a brown opaque sol. The sol was kept being stirred until it gelled. The obtained gel was dried at 100 C and crushed into small pieces with diameters of less than 32 mm. The crushed gel was heated at 600 C for 72 hours in an electric furnace and crushed again into fine particles with diameters of less than 32 mm. Fe2 O3 –SiO2 binary oxide thus prepared was amorphous and showed no hygroscopic property. Lithium foils, 1 mm thick and with a purity of 99.8%, were purchased from Honjo Metals Co., Ltd. A 1:2 v/v mixed solution of ethylene carbonate (EC) and methylethyl carbonate (MEC) containing 1 M (M = mol dm3 ) LiClO4 (LIPASTE-E2MEC/1), used as organic electrolyte solution, was purchased from Tomiyama Pure Chemical Industries Ltd. An N-methylpyrrolidone (NMP) (82 wt %) solution of polyvinylidene difluoride (PVDF) (18 wt %) (KF polymerL #1120) purchased from Kureha Chemical Industry Co., Ltd. was used as a gluing agent (binder). Copper foils were used as basal plates of the metal oxide electrodes. The other reagents were of analytical grade and were used without further purification except hexane, which was used after dehydration with molecular sieves. H2C O H2C O EC
O C O H CO C OCH CH 3 2 3 MEC
( CH2 CF 2 )n PVDF
2. Cathodes SnO2 cathodes were made as follows: Powders of SnO2 and the NMP solution of PVDF with weight ratios of 1:1 to 1:3 were first mixed to obtain a SnO2 paste. This paste was daubed manually on a copper foil as uniformly as possible. The copper foil with the paste on it was heated at 120 C for 20 min to remove the NMP component of the
paste and finally to obtain a cathode of tin(IV) oxide laminated on a copper foil. An optical photo of the cross-section of a SnO2 electrode with an optical microscope (a Keyence digital microscope V-H7000) is shown in Fig. 1(a). The size of laminated SnO2 cathode was about 16 mm 16 mm 80 mm. Starting from mixing of fine particles of Fe2 O3 –SiO2 and the NMP solution with weight ratios of 1:2, Fe2 O3 –SiO2 cathodes were prepared in a way similar to that for SnO2 electrodes. An optical photo of the cross-section of a Fe2 O3 –SiO2 electrode is shown in Fig. 1(c). The size of laminated Fe2 O3 –SiO2 cathode was about 10 mm 10 mm 150 mm. 3. Lithium Insertion Experiments The experimental apparatus used, schematically drawn in Fig. 2, is basically the same as the one used in experiments in the previous paper5) where lithium was electrochemically intercalated into graphite. It was composed of a power supply (a Hokuto Denko Corporation HJ-201B battery charge/ discharge unit), a three-electrode cell (electrolytic cell) equipped with water jacket and a data acquisition unit consisting of an A/D converter and a personal computer. The electrolytic cell was built up in a dry argon atmosphere. The volume of the electrolyte solution (EC/MEC solution) was about 10 cm3 and the cathode was placed so that the laminated metal oxide was wholly immersed in the electrolyte solution. The lithium insertion was performed in the constant current-constant voltage mode. That is, the lithium insertion (electrolysis) was at first carried out in a constant current mode. As the electrolysis proceeded, the electric potential of the cathode against the lithium reference electrode (cathode potential) swiftly decreased and reached pre-determined JOURNAL OF NUCLEAR SCIENCE AND TECHNOLOGY
Lithium Isotope Effects Accompanying Electrochemical Insertion of Lithium into Metal Oxides
1 2 3
10
6 4
5 8
7 9
Fig. 2 The experimental apparatus 1: personal computer, 2: A/D converter, 3: power supply, 4: lithium reference electrode, 5: lithium anode, 6: copper foil, 7: metal oxide cathode, 8: electrolyte solution, 9: stirrer tip, 10: electrolytic cell.
value. The electrolytic mode was then automatically changed to the constant voltage mode; the electrolysis was continued and the electric current gradually decreased while the cathode potential was kept constant at the pre-determined value. The electrolysis continued until the integrated quantity of electricity reached the predetermined value and was discontinued manually. During the electrolysis, the temperature of the electric cell was kept at 25 C by circulating the temperature-controlled water through the water jacket and the electrolyte solution was stirred with a magnetic stirrer. 4. Analyses After the electrolysis was finished, the metal oxide cathode was taken out of the cell in a dry argon atmosphere, washed thoroughly with dehydrated hexane and was allowed to stand to remove adhering hexane by evaporation. Lithium extraction and isolation from the Li-inserted SnO2 electrode and the sample preparation for the mass analysis for the 7 Li/6 Li isotopic ratio determination was carried out as follows: The lithium-inserted SnO2 was stoichiometrically removed from the copper foil and was decomposed, while heating, by fusion with ammonium iodide with the weight ratio of SnO2 and NH4 I of about 1:15 using a platinum wire as catalyst. The decomposition product was dissolved with 6 M HCl and insoluble matter (probably PVDF and its derivatives), if any, was separated by filtration. Lithium in the filtrate was separated from tin by solvent extraction using tributyl phosphate as the organic phase with the 1:1 volume ratio of the aqueous and organic phases. Tin was extracted into the organic phase, while lithium remained in the aqueous phase. Using part of the aqueous phase, the amount of lithium inserted in the SnO2 cathode was determined by flame photometry. The chemical form VOL. 44, NO. 1, JANUARY 2007
75
of lithium in another part of the aqueous phase was converted to lithium iodide through cation and anion exchanges and addition of HI and the lithium iodide thus obtained was subjected to mass analysis for the 7 Li/6 Li isotopic ratio determination. Lithium extraction and isolation from the Li-inserted Fe2 O3 –SiO2 binary oxide electrode and the sample preparation for the mass analysis was carried out as follows: The Liinserted binary oxide was stoichiometrically removed from the copper foil and lithium and iron in the binary oxide were completely dissolved out with 4 M HCl. The Li and Fe-bearing HCl solution was separated from the remaining solid by filtration and evaporated to dryness on a hot plate. The evaporation residue was dissolved with 4 M HCl. Iron in this aqueous HCl solution was solvent-extracted using a 1:1 v/v mixed solution of trioxtylamine and xicilene as the organic phase with the 1:1 volume ratio of the aqueous and organic phases. Using part of the aqueous phase, the amount of lithium inserted into the Fe2 O3 –SiO2 cathode was determined by flame photometry. The chemical form of lithium in another part of the aqueous phase was converted to lithium iodide through cation and anion exchanges and addition of HI and the lithium iodide thus obtained was subjected to mass analysis for the 7 Li/6 Li ratio determination. An aliquot of the electrolyte solution, after the lithium insertion was finished, was heated at 650 C for 4 hrs after the organic components were removed by evaporation, and the residue was dissolved with 4 M HCl. The chemical form of lithium in this HCl solution was converted to lithium iodide as in the cases of metal oxide electrodes and the lithium iodide thus obtained was subjected to mass analysis for the 7 Li/6 Li ratio determination in the electrolyte solution. Using the purified lithium (lithium iodide) in the ways mentioned above, the 7 Li/6 Li isotopic ratios of lithium-inserted metal oxide electrodes and the organic electrolyte solution were determined to estimate the 7 Li-to-6 Li singlestage separation factor, S, accompanying the lithium insertion, defined as: S ¼ (7 Li/6 Li)solution =(7 Li/6 Li)oxide ; where (7 Li/6 Li)solution and (7 Li/6 Li)oxide are the lithium isotopic ratios of the electrolyte solution and the electrode, respectively. By definition, S is larger than unity when 6 Li is preferentially taken up by the metal oxide electrode. The 7 Li/6 Li ratios of the mass samples were determined by the surface ionization technique with a Varian Mat CH-5 mass spectrometer. The procedure of the lithium isotopic measurements was described elsewhere.3) The lithium concentrations in solutions were determined by flame photometry with a Daini Seikosha SAS-727 atomic absorption spectrometer. The powder X-ray diffraction (XRD) patterns were recorded using a Rigaku RINT2100V/P X-ray diffractometer with the Cu K radiation. Scanning electron microscopy (SEM) photos were taken with a Hitachi scanning electron microscope S-4500, and optical photos with a Keyence digital microscope V-H7000. Electron spectroscopy for chemical analysis (ESCA) spectra were measured with a ULVAC-PHI ESCA 5800ci X-ray photoelectron spectrometer.
76
M. MOURI et al. Table 1 Results of the lithium uptake experiments with no electric current
SnO2
Fe2 O3 –SiO2
Graphite None
Immersing time (h) 15.0 20.0 22.5 25.0 27.3 30.8 32.5 35.0 40.0 50.0 2.0 4.0 8.0 16.0 18.0 22.0 24.0 25.0 25.0 20.0 24.0 25.0 30.0 35.0 40.0 49.8
Amount of oxide or graphite (mg)
Amount of PVDF (mg)
Amount of Li taken up (mg)
Amount of Li taken up per g PVDF (mg)
33.55 55.66 56.21 29.92 58.72 45.36 54.83 40.03 49.79 47.96 8.56 7.62 3.56 4.46 5.75 6.36 7.66 22.74 20.17
12.75 25.52 13.03 12.29 14.99 16.93 16.00 13.46 14.69 10.92 3.20 3.78 1.48 1.74 2.47 2.52 2.93 8.11 6.38 5.22 8.84 3.78 3.96 4.19 4.86 5.60
0.134 0.366 0.208 0.181 0.159 0.215 0.149 0.131 0.150 0.204 0.042 0.041 0.022 0.042 0.064 0.029 0.030 0.090 0.064 0.028 0.218 0.064 0.051 0.103 0.045 0.074
10.51 14.34 15.96 14.73 10.61 12.70 9.31 9.73 10.21 18.68 13.13 10.85 14.86 24.14 25.91 11.51 10.24 11.10 10.03 5.36 24.66 16.93 12.88 24.58 9.26 13.21
5. Lithium Uptake by Oxide Electrodes with No Electric Current During the course of above mentioned electrochemical lithium-insertion experiments, it became apparent that lithium might be taken up by metal oxide electrodes by simply placing them in the EC/MEC solution of 1 M LiClO4 in an argon atmosphere without applying voltage. To examine whether this phenomenon actually occurred, and if so, how much amount of lithium was taken up, additional experiments were carried out. They were basically the same as electrochemical lithium-insertion experiments except that no voltage was applied between the anode and the cathode.
S 1.002 1.003 0.994 0.995
1.005 0.993 0.995 1.002 1.002
30 Amount of Li per g PVDF (mg)
Electrode material
25 20 15 10 5 0 0
10
20
30
40
50
60
Immersing time ( h ) III. Results and Discussion 1. Results of the Experiments with No Electric Current Results of the experiments, in which oxide electrodes were simply immersed in electrolyte solution without passing the electric current, are summarized in Table 1. Table 1 also includes the results of other non-electric current experiments using graphite electrodes made in the similar way described in the previous paper5) and copper foils on which only the binder was daubed. In Fig. 3, the amount (mg) of inserted lithium per gram of PVDF is plotted against immersing time. Except for four plots at around 25 mg Li per 1 g PVDF, which are most probably due to insufficient washing of electrodes with hexane, the Li uptake is nearly constant at 10 to 15 mg lithium per 1 g PVDF, irrespective
Fig. 3 Plot of the amount of Li taken up per 1 g PVDF against the immersing time in no-electric-current experiments Electrode material; : SnO2 , : Fe2 O3 –SiO2 , +: graphite, : none.
of kind of electrode material and immersing time from 2 to 50 hrs. Another experiment confirmed that the SnO2 powder alone took up no lithium. The results shown in Table 1 and Fig. 3 thus strongly indicate that PVDF used as the binder takes up 10 to 15 mg lithium per gram of PVDF, with the average value of 12.1 mg. This value was used to correct the electric current efficiency and the separation factor obtained in the experiments in which the electric JOURNAL OF NUCLEAR SCIENCE AND TECHNOLOGY
77
Lithium Isotope Effects Accompanying Electrochemical Insertion of Lithium into Metal Oxides Table 2 Experimental results of the electrochemical insertion of Li
Run No.
Oxide
Sn–1 Sn–2 Sn–3 Sn–4 Sn–5 Sn–6 Sn–7 Sn–8 Sn–9 Sn–10 Sn–11 Sn–12 Sn–13 Sn–14 Sn–15 Sn–16 Sn–17 Sn–18
SnO2
Fe–1 Fe–2 Fe–3 Fe–4 Fe–5 Fe–6
Fe2 O3 –SiO2
Amount of Sn or Fe2 (mmol)
Predetermined cathode potential (V)
Insertion time (h)
Integrated quantity of electricity (mC)
Li uptake (mmol)
Curent efficiency (%)
x in SnO2 –Lix or Fe2 O3 –Lix
S
0.212 0.216 0.068 0.242 0.170 0.383 0.331 0.364 0.342 0.420 0.218 0.218 0.274 0.268 0.295 0.302 0.290 0.294
0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.03 0.03 0.03 0.03 0.05 0.05 0.05 0.10 0.10 0.10 0.10
5.00 9.00 1.52 7.18 4.67 7.22 1.48 9.30 3.08 1.26 5.83 6.15 0.90 0.60 1.97 2.13 0.87 10.17
2430 3350 820 6400 2020 29500 16000 52100 33000 13500 5460 16600 9700 6460 7120 29200 9340 14100
0.05 0.07 0.02 0.11 0.04 0.37 0.20 0.62 0.37 0.17 0.09 0.20 0.13 0.09 0.10 0.30 0.11 0.17
60 126 108 124 103 113 102 111 103 103 124 101 222 105 92 89 82 98
0.23 0.31 0.29 0.43 0.23 0.97 0.59 1.72 1.09 0.40 0.42 0.90 0.49 0.35 0.32 0.99 0.37 0.59
0.994 0.994 0.997 0.989 0.989 0.993 0.982 0.992 0.996 0.999 0.992 0.992 0.998 0.986 0.985 0.988 0.993 0.991
0.027 0.022 0.019 0.023 0.016 0.015
0.02 0.02 0.02 0.02 0.02 0.02
7.00 5.31 1.30 1.52 6.22 0.56
3900 3100 2440 445 2610 1430
0.04 0.04 0.03 0.01 0.03 0.02
71 110 109 136 112 124
1.31 1.81 1.68 0.48 2.08 1.43
0.999 1.001 1.002 0.999 0.997 1.008
voltage was applied between the anode and cathode (vide infra). The mechanism of Li taking-up by PVDF is not well understood. The existence of highly electronegative fluorine in PVDF may be concerned with it. S values were measured for some of the experiments and the results are listed on the last column of Table 1. The data are scattered around S ¼ 1 (S ¼ 0:999 0:005 (SD)), indicating that no appreciable lithium isotope effect accompanies the lithium uptake from the EC/MEC solution of LiClO4 by PVDF used as the binder. 2. Experiments with SnO2 Electrodes Results of the experiments, in which electric voltage was applied between the lithium foil anode and the SnO2 cathode, are summarized in the upper part of Table 2. The amount of Sn loaded on the copper foil ranged from 0.0679 to 0.420 mmol. The pre-determined cathode potential was set at 0.02, 0.03, 0.05 or 0.1 V, and the insertion time was between 0.60 and 10.17 hrs. The pre-determined integrated quantity of electricity was varied between 820 to 52100 mC, and the resultant Li uptake ranged from 0.0199 to 0.300 mmol. The electric current efficiency, defined as 100 times the amount of lithium analytically found per mole of metal oxide divided by the amount of lithium calculated from the integrated quantity of electricity, ranged from 60 to 222% after corrected for the amount of lithium taken up by PVDF. As mentioned above, the average value of VOL. 44, NO. 1, JANUARY 2007
12.1 mg-Li per g-PVDF was used for the correction, because the exact Li uptake by PVDF could not be measured in each experiment. Most values are reasonable, scattering around 100%. Exceptionally large value (Run Sn-13) may be due to incomplete washing of the electrode with hexane. Figure 1(b) shows an optical photo of the cross-section of the same SnO2 electrode that shown in Fig. 1(a) taken after the lithium insertion. The Sn to Li molar ratio for the electrode expected from the integrated quantity of electricity is 1:0.87. Comparison of the two figures shows, albeit not very clearly, that the lithium insertion accompanies the cubic expansion of the host material; the thickness of the electrode increased from its original value of ca. 80 mm to ca. 150 mm. The cubic expansion of SnO2 due to the lithium insertion is also evidenced in the ESCA spectra recorded before and after the lithium insertion. Figures 4(a) and 4(b) show the ESCA spectra of the surface of SnO2 electrode before and after the lithium insertion, respectively. In the former spectrum, the fluorine peak showing the existence of PVDF on the electrode surface is clearly seen along with the Sn peak, indicating that the sufficient amount of the binder exists on the surface of the electrode and SnO2 is not completely covered with PVDF. Contrary to this, the F peak is hardly seen in the spectrum of the electrode after the Li insertion, indicating the non-existence of the binder on the surface. This observation is consistent with that in Figs. 1(a) and 1(b). In Figs. 5(a) and 5(b) are shown SEM photos of the surface of a SnO2 electrode before and after the lithium
78
M. MOURI et al.
(b)
Intensity (au)
Intensity (au)
(a)
700
600
B.E./eV
F
700
500
600
500
B.E./eV
O Sn Sn
O Sn Sn
F
(d)
Intensity (au)
Intensity (au)
(c)
720
740
700
680
740
B.E./eV
Fe
Fe
720
Fe
F
700
Fe
680
B.E./eV
F
Fig. 4 ESCA spectra of the surfaces of a SnO2 electrode (a) before and (b) after the lithium insertion and of a Fe2 O3 – SiO2 electrode (c) before and (d) after the lithium insertion
(a)
1.01
(b)
1.01
S
1.00
crack
1.00 0.99
5 µm
(c)
5 µm
0.99 0.98
(d)
0.0
0.5
1.0
1.5
2.0
2.5
x Fig. 6 Plot of S against x in SnO2 -Lix (pre-determined cathode potential; : 0.02 V, : 0.03 V, : 0.05 V, : 0.1 V) or in Fe2 O3 SiO2 -Lix (pre-determined cathode potential; : 0.02 V) 50 µ m
50 µ m
Fig. 5 SEM photos of the cross-sections of a SnO2 electrode (a) before and (b) after the lithium insertion and of a Fe2 O3 –SiO2 electrode (c) before and (d) after the lithium insertion
insertion. While cracks are observed in the photo after the lithium insertion (Fig. 5(b)), there are not before the insertion, which is also consistent with the results in Figs. 1(a) and 1(b) and Figs. 4(a) and 4(b).
The S values are listed on the last column of Table 2, and plotted against x in SnO2 -Lix in Fig. 6. As is seen in the figure, S is smaller than unity and seems independent of x and the predetermined cathode potential. The average of the 18 data points is 0:992 0:005 (SD). An S value that is smaller than unity means that the heavier isotope 7 Li is preferentially fractionated into the host material. This is contrastive to the results of previous paper,6) where Sn metal was used as the host material of Li insertion from EC/ JOURNAL OF NUCLEAR SCIENCE AND TECHNOLOGY
Intensity
Lithium Isotope Effects Accompanying Electrochemical Insertion of Lithium into Metal Oxides
(a)
30
40
50
60
Intensity
20
70 (b)
20
30
40
50
60
70
2 θ (CuK α ) / degree Fig. 7 XRD patterns of a SnO2 electrode (a) before and (b) after the Li insertion : SnO2 , : Sn.
MEC solution of LiPF6 and the lighter isotope was preferentially fractionated into Sn. To clarify why the SnO2 -EC/MEC and Sn-EC/MEC systems are opposite in the direction of the Li isotope fractionation, XRD patterns were measured of a SnO2 electrode before and after the Li insertion, and are shown in Figs. 7(a) and 7(b), respectively. Whereas, in the XRD pattern of the SnO2 electrode before the Li insertion, only the peaks ascribable to SnO2 are observed, the peaks of Sn are observed in the electrode after the Li insertion, in addition to those of SnO2 . This means that a part of Sn(IV) of the electrode is reduced to Sn(0). Then the following series of reactions are expected to have occurred in the SnO2 electrode upon Li insertion: SnO2 þ 4e ! Sn þ 2O2 ; Li ! Liþ þ e ;
and
2Liþ þ O2 ! Li2 O: Lithium is thus expected to exist formally in the monovalent state rather than as a neutral atom, although Li2 O is not evidenced by the XRD measurement. It is highly probable that lithium inserted in SnO2 electrodes exists actually as lithium ion that interacts with surrounding O2 ions. A molecular orbital calculation11) showed that EC molecules were preferentially coordinated to lithium ions using their carboxyl oxygens in the EC/MEC mixed solvent. Then, the lithium isotope effect observed in the present SnO2 -EC/MEC system is the one between the Liþ ion interacting with O2 of Li2 O in the SnO2 electrode and the Liþ ion interacting with carbonyl oxygens of EC molecules in the electrolyte. The sum of the forces acting on the lithium ion is slightly stronger in the electrode than in the electrolyte, which leads to the slightly preferential insertion of the heavier lithium isotope into SnO2 . Contrary to this, lithium with the formal charge zero weakly interacts with tin, forming the VOL. 44, NO. 1, JANUARY 2007
79
metal-metal bond, in the Sn-EC/MEC system, which leads to preferential insertion of the lighter isotope into Sn metal from the EC/MEC mixed solvent. 3. Experiments with Fe2 O3 –SiO2 Binary Oxide Electrodes Results of the experiments where electric voltage was applied between the lithium foil anode and the Fe2 O3 –SiO2 binary oxide cathode are summarized in the lower part of Table 2. The amount of Fe2 loaded on the copper foil to prepare the electrodes ranged from 0.015 to 0.027 mmol. The cathode potential was set at 0.02 V in all the experiments and the insertion time was from 0.57 to 7 hrs. The pre-determined integrated quantity of electricity was between 440 and 3890 mC, and the resultant Li uptake ranged from 0.011 to 0.040 mmol. The electric current efficiency ranged from 71% to 136%. The value over 100% may mean insufficient electrode washing with hexane. Optical photos of the cross-sections of a Fe2 O3 –SiO2 electrode before and after the lithium insertion are shown in Figs. 2(c) and 2(d), respectively. Contrary to the case of the SnO2 electrode, no appreciable cubic expansion of Fe2 O3 –SiO2 upon Li insertion is observed due to the amorphous nature of the mixed oxide. The Fe to F regions of ESCA spectra of the surface of a Fe2 O3 –SiO2 electrode before and after the Li insertion are shown in Figs. 4(c) and 4(d), respectively. No substantial difference is observed between the two spectra. Whereas the peak ascribable to F is observed in both the figures, peaks of Fe are hardly seen. This means the surface of the Fe2 O3 –SiO2 electrode was almost completely covered with the binder before the Li insertion, and even after the insertion, the oxide did not come out to the surface. In Figs. 5(c) and 5(d) are shown SEM photos of the surface of a Fe2 O3 –SiO2 electrode before and after the lithium insertion. Again, no substantial difference is observed between the two. The S values are listed on the last column of Table 2, and plotted against x in Fe2 O3 -Lix in Fig. 6. As is seen in the figure, S is scattered around unity (1:001 0:004 (SD)), meaning that the Fe2 O3 –SiO2 binary oxide shows no specific affinity toward either of the lithium isotopes. Analogously to the SnO2 -EC/MEC system, it is expected that lithium exists as lithium oxide rather than neutral atomic lithium and part of iron as Fe(II) in Li-inserted Fe2 O3 –SiO2 binary oxide electrodes. Unfortunately, XRD evidence of LiO2 was not obtained. Also, the formation of Fe(II) by the expected reaction Li(0) þ Fe(III) ! Li(I) þ Fe(II) was not observed in the ESCA spectra of a Li-inserted Fe2 O3 – SiO2 electrode. Li-inserted Fe2 O3 –SiO2 is very unstable against water; lithium readily comes out to the surfaces of electrodes upon the contact with the moisture in the air. 4. Comparison with Past Results Values of the separation factor obtained in redox systems using different materials as hosts of the Li insertion are summarized in Table 3. While mercury, tin and graphite show selectivity for the lighter isotope of lithium, the oxides show selectivity for the heavier isotope or show no specific affinity. As mentioned above, this is most probably ascribable to
80
M. MOURI et al. Table 3 Comparison of Li isotope effects in the redox system Host material Hg Graphite Sn Zn SnO2 Fe2 O3 –SiO2
Solvent H2 O EC/MEC EC/MEC EC/MEC EC/MEC EC/MEC
Selectivity 6
Li Li 6 Li 6 Li 7 Li None 6
S
Ref.
1.049–1.062 1.007–1.025 1.002–1.015 1.005–1.023 0.982–0.999 0.997–1.008
2) 5) 6) Unpublished data This work This work
S ¼ (7 Li/6 Li)solvent =(7 Li/6 Li)host where (7 Li/6 Li)solvent is the 7 Li/6 Li isotopic ratio in the solvent and (7 Li/6 Li)host that in the host material.
the formation of lithium oxide in the oxide host materials. A general conclusion may be derived from the present experiments that metal oxides show no 6 Li selectivity due to the formation of lithium oxide, and may be poor materials for 6 Li enrichment utilizing the redox reaction of lithium due to the irreversible oxidation reaction of lithium. The absolute value of S-1, which is an index of the magnitude of isotope effects, observed for mercury is by far the largest among the host materials in Table 3. This may be related to the fact that only mercury is in the liquid state in the experimental temperature range.
IV. Conclusions To summarize, we make the following statements: (1) The heavier isotope of lithium was preferentially inserted into SnO2 from organic electrolyte solution of ethylene carbonate (EC) and methylethyl carbonate (MEC) containing 1 M LiClO4 . Upon Li insertion, the reduction of Sn(IV) to Sn(0) was observed, which presumably accompanied the formation of lithium oxide. This oxide formation was expected to account for the preferential uptake of 7 Li by SnO2 . (2) The Fe2 O3 –SiO2 binary oxide showed no specific selectivity towards either of the lithium isotopes upon the uptake from the EC/MEC electrolyte solution.
Acknowledgements Professor Y. Fujii, Tokyo Institute of Technology (Titech) kindly offered the use of a Varian MAT CH-5 mass spectrometer. We acknowledge Dr. M. Nomura, Titech, for his assistance in mass spectrometric measurements of lithium isotopic ratios.
References 1) A. A. Palko, J. S. Drury, G. M. Begun, ‘‘Lithium isotope separation factors of some two-phase equilibrium systems,’’ J. Chem. Phys., 64, 1828 (1976). 2) M. Fujie, Y. Fujii, M. Nomura et al., ‘‘Isotope effects in electrolytic formation of lithium amalgam,’’ J. Nucl. Sci. Technol., 23, 330 (1986). 3) T. Oi, K. Kawada, M. Hosoe et al., ‘‘Fractionation of lithium isotopes in cation-exchange chromatography,’’ Sep. Sci. Technol., 26, 1353 (1991), and references cited therein. 4) K. Nishizawa, H. Watanabe, S. Ishino et al., ‘‘Lithium isotope separation by cryptand (2B , 2, 1) polymer,’’ J. Nucl. Sci. Technol., 21, 133 (1984), and references cited therein. 5) S. Yanase, W. Hayama, T. Oi, ‘‘Lithium isotope effect accompanying electrochemical intercalation of lithium into graphite,’’ Z. Naturforsch., 58a, 306 (2003). 6) S. Yanase, T. Oi, S. Hashikawa, ‘‘Observation of lithium isotope effects accompanying electrochemical insertion of lithium into tin,’’ J. Nucl. Sci. Technol., 37, 919 (2000). 7) J. Yang, Y. Takeda, N. Imanishi et al., ‘‘Li-alloy materials for lithium ion cell,’’ Battery Technol., 11, 103 (1999). 8) S. Panero, G. Savo, B. Scrosati, ‘‘Tin oxide-based lithium-ion polymer-electrolyte cells,’’ Electrochem. Solid-State Lett., 2, 365 (1999). 9) T. Misao, ‘‘Lithium insertion to amorphous Fe2 O3 –SiO2 binary oxides,’’ Master thesis, Keio University, 1997 [in Japanese]. 10) C. Savii, M. Popovici, C. Enache et al., ‘‘Fe2 O3 –SiO2 composites obtained by sol-gel synthesis,’’ Solid State Ionics, 151, 219 (2002). 11) S. Yanase, T. Oi, ‘‘Solvation of lithium ion in organic electrolyte solutions and its isotopic reduced partition function ratios studied by ab initio molecular orbital method,’’ J. Nucl. Sci. Technol., 39, 1060 (2002).
JOURNAL OF NUCLEAR SCIENCE AND TECHNOLOGY
