- Email: [email protected]
Investigation of the high, stable electrical conductivity in graphite intercalation compounds prepared from flexible graphite sheets
Synthetic Metals 198 (2014) 107–112
Contents lists available at ScienceDirect
Synthetic Metals journal homepage: www.elsevier.com/locate/synmet
Investigation of the high, stable electrical conductivity in graphite intercalation compounds prepared from flexible graphite sheets Rika Matsumoto * Faculty of Engineering, Tokyo Polytechnic University, 1583 Iiyama, Atsugi, Kanagawa 243-0297, Japan
A R T I C L E I N F O
A B S T R A C T
Article history: Received 13 July 2014 Received in revised form 12 September 2014 Accepted 30 September 2014 Available online xxx
Potassium–graphite intercalation compounds (K–GICs) prepared from flexible graphite sheets are unusually stable in air and their high electrical conductivity can be maintained for long periods. To investigate the reasons for this air-stability, we evaluated the changes in the electrical conductivity as a function of electrical carrier densities and mobilities, upon exposure of the GICs to air. These electrical carrier parameters were estimated using the measured electrical conductivity, magnetoresistance, and Hall coefficient. The decomposition processes varied depending on the characteristics of both the host graphite and the intercalated materials. In the case of K–GICs prepared from flexible graphite sheets, the electron density was decreased by the de-intercalation of K atoms. However, the electrical conductivity hardly changed, because the electron mobility was increased sufficiently to compensate for the decrease in the density. The recovery of the electron mobility was thought to be caused by the flexibility of the graphite layers. Therefore, it was found that the air-stability of GICs could be determined by considering the degree of de-intercalation as well as the flexibility of the graphite layers. Furthermore, the perfect graphite crystal is not suitable to prepare air-stable and high conductive GICs. ã 2014 Elsevier B.V. All rights reserved.
Keywords: Graphite intercalation compound Electrical conductivity Air-stability Flexible graphite sheet Galvanomagnetic properties
1. Introduction Chemical species such as alkali metal atoms, metal halides, and halogen molecules can intercalate into the interlayers of lamellar graphite and form graphite intercalation compounds (GICs). A result of this intercalation is the occurrence of carrier transfer between the intercalated materials and adjacent graphite planes (graphene), causing the electrical conductivities of GICs to become approximately 10 times higher than those of the host graphite materials. However, GICs are known to be generally unstable in air. Most GICs are immediately decomposed by oxidation or the deintercalation of intercalated materials when exposed to the atmosphere [1]. Historically, GICs have attracted much attention as light-weight conductive materials. The highest conductivities ever reported for GICs exceeded those of copper and silver [1]. However, the air-instability of GICs has obstructed their practical use. Improving the air-stability of GICs is considered to be profitable. For example, nanocarbons, i.e., carbon-based nanosized materials such as graphene and carbon nanotubes, have been studied for practical use in electrodes [2,3], large-scale integrated circuits [4],
* Tel.: +81 46 242 9575; fax: +81 46 242 9575. E-mail address: [email protected] (R. Matsumoto). http://dx.doi.org/10.1016/j.synthmet.2014.09.037 0379-6779/ ã 2014 Elsevier B.V. All rights reserved.
and other applications. However, in order for nanocarbons to be used more effectively, the enhancement of their electrical conductivities is necessary. One technique that can increase the electrical conductivity of nanocarbons is intercalation [5]. However, the decomposition rates of GICs prepared from smaller graphite particles are faster, and they decompose easily in air [6]. Therefore, when carrying out the intercalation of nanocarbons, the lower airstability of the intercalated products is a problem. In another example, we have studied the thermoelectric properties of GICs, and found that they have possible applications as thermoelectric materials [7–9]. Although, both n- and p-type GICs are necessary in thermoelectric applications, there are very few air-stable GICs, especially n-type GICs, reported in the literature. In our investigations of GICs stability, we have succeeded in forming air-stable GICs that exhibit high electrical conductivity. In previous papers [10,11], we reported that potassium–graphite intercalation compounds (K–GICs) prepared from commercially available flexible graphite sheets (manufactured by the pyrolysis of polyimide films), exhibited high stability and high electrical conductivity for an extended period of time in air. However, we were unable to thoroughly explain the stabilization mechanism in the previous paper. We also found that Li–GICs could not be stabilized even when using the same flexible graphite sheet. Our ultimate goal is to establish an intercalation technique for carbon/ graphite materials, which produces intercalation compounds with
108
R. Matsumoto / Synthetic Metals 198 (2014) 107–112
high electrical conductivity and air-stability. Therefore, it is important to first understand the stabilization mechanism in the previously studied K–GICs. The air-stabilities of GICs are usually investigated by examining changes in their structure using X-ray diffraction (XRD), or by observing changes in their electrical conductivity. However, since we wish to use the intercalation technique for practical applications, we consider the long-term maintenance of high electrical conductivity in air as a particularly important parameter. Therefore, in this study, “stability” refers to the retention of high electrical conductivity. In our experience, the change in electrical conductivity is a more sensitive indicator of GICs decomposition than changes in the XRD patterns [12]. In this study, the causes for the stabilization of K–GICs prepared from flexible graphite sheets were investigated by monitoring the changes in electrical carrier densities and mobilities. Electrical conductivity is proportional to the product of the electrical carrier density and mobility. The electrical carrier density provides information about the amount of carrier transfer between the graphene and intercalated materials. In other words, it is thought that the electrical carrier density provides an indication of the amount of intercalated materials in the interlayers. Conversely, the electrical carrier mobility provides information about the number of scattering factors of the phonon. It is thought that the mobility also indicates the degree of perfection of the graphite structure. Therefore, an investigation of the carrier density and mobility can lead to an understanding of the amount of intercalated materials and the degree of perfection of the graphite crystalline structure. 2. Experimental 2.1. Preparation and measurements PGS1 graphite sheets with a thickness of 0.1 mm (Panasonic Co., EYGS 182310), GRAFOIL1 sheets with thicknesses of 0.1 mm and 0.3 mm (GrafTech Co., GTA grade), and highly ordered pyrolytic graphite plates (HOPG) (NT-MDT Co., grades ZYH and ZYA) were used as the host graphite materials. PGS is produced by the thermal decomposition of a polyimide film [13–15]. GRAFOIL is made from H2SO4–GICs, which are prepared from pure graphite flakes. After the exfoliation of the GICs, the resulting exfoliated graphite is sheeted. The PGS and GRAFOIL sheets were cut into 3–5 20– 25 mm2 rectangles. HOPG (10 10 2 mm3) was cleaved with the edge of a cutter and an adhesive tape, and cut into 10 5 0.3 mm3 rectangles. GRAFOIL was heat treated under vacuum at 900 C for 4 h before use to remove S, which is a common impurity. Potassium, lithium, and anhydrous copper(II) chloride (CuCl2) were used as the intercalate species without further purification. The GICs were prepared by allowing graphite to react with the vapor of the intercalate species at specific temperatures under vacuum. The reaction temperatures and durations for K, Li, and CuCl2 intercalation were 473 K for 3 days, 723 K for 7 days, and 753 K for 45 days, respectively. The structures formed were saturated stage-1 structures, where the intercalated layers existed in all the graphite interlayers. The intercalated layers are inserted in an orderly manner into the graphene layers. In a stage1 structure, there are intercalated layers between all the graphene layers, whereas in stage-2 and stage-n structures, the intercalated layers occur at intervals of two layers and n layers, respectively. Measurements of the in-plane electrical conductivity (s ), magnetoresistance (Dr/r), and the Hall coefficient (RH) were performed by the five-terminal method after the GICs specimens were exposed to air, as shown in Fig. 1, with magnetic fields (B) up to 0.5 T at room temperature. To measure the changes in the values over time, the specimens were kept on a measurement holder in
Fig. 1. Measurement set-up for the electrical conductivity (s ), magnetoresistance (Dr/r), and Hall coefficient (RH).
air. The average temperature and humidity during the measurements, and storage were 293–300 K and 50%, respectively. 2.2. Estimation of carrier density and mobility The electrical carrier densities (ne and nh) and mobilities (me and mh) were estimated from the values of s , Dr/r, and, RH [16,17]. In cases where Dr/r was not detected, the transport was considered to occur through one-carrier conduction by either electrons or holes. Therefore, the carrier density (n) and mobility (m) were estimated from Eqs. (1) and (2), where e is the electron charge.
s ¼ enm
(1)
RH ¼ ð1=neÞ
(2)
In cases, where Dr/r was detected, the transport was considered to occur via two-carrier conduction by both electrons and holes. The carrier densities and mobilities were then estimated using Eqs. (3)–(5).
s ¼ eðne me þ nh mh Þ
(3)
1 ne me 2 nh mh 2 RH ¼ e ðne me þ nh mh Þ2
(4)
Dr=r ¼
ne nh me mh ðme þ mh Þ ðne nh þ me mh Þ2
B
(5)
Furthermore, the approximations shown in Eqs. (6) and (7) were adopted for simplicity. Eq. (6) was used for graphite, whereas Eq. (7) was used for GICs.
m ¼ me ¼ mh
(6)
ðne nh ÞGraphite ¼ ðne nh ÞGIC
(7)
For practical calculations, we adopted Newton’s method for a quick and correct analysis [18]. In cases where the Newton’s method of analysis was not adopted, the trial and error methods were used. The set of parameter values (ne,nh, me, and mh) to minimize the error was searched by scanning. 3. Results and discussion 3.1. Measured and estimated electrical properties of host graphite and K–GICs In our previous work [10], the changes in the electrical conductivity of K–GICs specimens (with a composition of KC8)
R. Matsumoto / Synthetic Metals 198 (2014) 107–112
prepared from PGS, GRAFOIL, and HOPG after exposure to air were measured. For the specimens prepared from GRAFOIL and HOPG, a clear reduction in the electrical conductivity occurred immediately after the specimens were exposed to air. In contrast, for the specimens prepared from PGS, the electrical conductivity values decreased only slightly and gradually, and were maintained at a value that was six times higher than that of host PGS, even after 30 days in air. In this study, to investigate the cause of the airstability of the K–GICs prepared from PGS, we assessed the electrical conductivity in terms of its constituent factors, namely electrical carrier densities and mobilities. Table 1 summarizes the measured and estimated electrical properties for the host graphite materials. Since, the measurements were conducted several times for each host and the values were somewhat scattered, the exhibited maximum and minimum electrical conductivity values are listed. Although, thick (0.3 mm) and thin (0.1 mm) GRAFOIL specimens were used, no significant differences were noted between them. Similarly, there were no significant differences between specimens prepared from good (ZYA) and poor (ZYH) grade HOPG. The electron density (ne) was slightly larger than the hole density (nh) for all of the graphite materials. Further, both the carrier densities (ne, nh) and the mobilities (m) exhibited the same trend, i.e., HOPG > PGS > GRAFOIL. However, it should be noted that the accuracy of values for HOPG is low. While the manufacturer specification for the HOPG electrical conductivity is 2.5 104 S cm1, the values that we measured were significantly smaller and scattered in the range of 0.7 104 S cm1–1.5 104 S cm1. This may be a result of the specimen’s imperfect rectangular form. Furthermore, the structural perfection of HOPG might have been damaged, because we roughly cut and peeled off the HOPG plate with the edge of a cutter and adhesive tape. However, since the changes in the values of electrical conductivity were more important in this study than the absolute values themselves, we concluded that a slight inaccuracy in the absolute values would not significantly impact our interpretations. The measured electrical conductivities (s ) and Hall coefficients (RH), as well as the estimated electron densities and mobilities (ne and me), for K–GICs prepared from PGS, GRAFOIL, and HOPG (denoted as K–PGS, K–GRAFOIL, and, K–HOPG, respectively), are summarized in Table 2. All the K–GICs specimens failed to demonstrate magnetoresistance (Dr/r) and the RH values were negative, indicating that the K–GICs exhibit one-carrier conduction using only electrons. In a K–GICs, the intercalated K atoms provide electrons to the adjacent graphene planes, which increase the ne in the graphene planes, resulting in n-type conduction. On the other hand, me becomes lower than that of the host graphite because the number of scattering factors of the phonon is increased by Kintercalation. A few K–GICs specimens for each host were prepared and their electrical properties were measured (data for replicate samples are
109
indicated as 1 or 2). The values for K–HOPG, especially s, varied widely, likely due to the irregularity of the specimen shape and the degradation of the graphite structure as mentioned above. The measurements were started within 1 h after exposure of the specimen to air. Although, each specimen was prepared as a stage1 K–GICs with a KC8 composition, the structure of the specimen at the onset of the measurement was probably not pure stage-1 because the K–GICs are likely to have decomposed to some degree after exposure to air. Based on our previous work [10], the structures at the time of measurement were supposed to be stage3 (KC36) or stage-4 (KC48) for K-PGS and K-HOPG, a mixture of a small amount of stage-1 and higher stages for K–GRAFOlL. 3.2. Changes in the electrical properties of K–GICs prepared from PGS, GRAFOIL, and HOPG in air Fig. 2 shows the changes in electrical conductivities (s /s 0), electron densities (ne/ne0), and, electron mobilities (me/me0) for K– PGS-1, K–GRAFOIL-1, and K–HOPG-1 after exposure to air. The vertical axes are normalized by the initial values (s 0, ne0, me0), which were obtained after 1 h of exposure to air. Measurements for a second set of specimens, K–PGS-2, K–GRAFOIL-2, and K– HOPG-2, were also conducted, with similar results. The K–PGS specimens were stable in air, as reported. The s value was maintained at 78% of the initial value after 6 days (144 h) of exposure to air. However, as a result of the exposure to air for 6 days, ne decreased to 17% and me increased to 460% of the initial values. As mentioned above, the K–PGS structure transitioned from stage-1 to stage-4 immediately upon exposure to air. Moreover, it was observed that the K–PGS specimens were dampened with a strong basic solution (KOH) after exposure to air. This indicates that the K atoms in the interlayers were partially de-intercalated and reacted with H2O in the air. From these observations, it was concluded that the de-intercalation of K atoms occurred within a short duration after the K–PGS specimens were exposed to air. After reaching the stage-4 structure, the de-intercalation in K–PGS was abated, This is thought to be due to the coating of the graphite crystallite edges with air-stable substances such as oxides. In the case of K–GRAFOIL, ne decreased and me increased upon exposure to air. However, the change in the electron density (ne/ ne0) was larger than that in K–PGS; ne decreased to 12% of the initial value within 3 days, while me increased to only 230%. Consequently, s decreased by a significant amount, to 27% of the initial value. In the case of K–HOPG, ne/ne0 was relatively low, reaching 34% of the initial value after 4 days in air. Although this value is twice as large as that of K–PGS, the observation is easily understood. Since, the graphite crystallites in HOPG are larger than those in PGS, the diffusion of K atoms in the interlayers is expected to be difficult. In fact, we have found that the structural change of K–HOPG in air is slow and similar to that of K–PGS and the near-stage-4 structure is
Table 1 Measured electrical conductivity (s ), magnetoresistance (Dr/r), and Hall coefficient (RH), and estimated carrier densities (ne and nh) and mobility (m) of PGS, GRAFOIL, and HOPG at room temperature. Specimens
s /S cm1
Dr/r
RH (at 0.4 T)/cm3 C1
ne/cm3
nh/cm3
m/cm2 V1 s1
(at 0.4 T)/ PGS
4.6 103 4.3 103
6.2 102 6.5 102
1.3 101 1.5 101
2.5 1018 2.3 1018
2.1 1018 1.9 1018
6.3 103 6.4 103
GRAFOIL
1.2 103 1.1 103
1.6 102 9.6 103
9.6 102 9.8 102
1.2 1018 1.4 1018
1.1 1018 1.3 1018
3.2 103 2.4 103
HOPG
1.5 104 7.0 103
1.6 101 6.9 102
5.7 102 7.2 102
5.2 1018 3.6 1018
4.3 1018 3.1 1018
1.0 104 6.6 103
110
R. Matsumoto / Synthetic Metals 198 (2014) 107–112
Table 2 Measured electrical conductivity (s ), magnetoresistance (Dr/r), and Hall coefficient (RH), and estimated electron density (ne) and mobility (me) of K–GICs prepared from PGS, GRAFOIL, and HOPG after exposure to air at room temperature. Host
Specimens
Time in air
s/S cm1
RH (at 0.4 T)/cm3 C1
ne/cm3
me/cm2V1 s1
PGS
K–PGS-1 K–PGS-2
1h 1h
3.8 104 3.9 104
3.5 103 7.2 103
1.8 1021 8.7 1020
1.3 102 2.8 102
GRAFOIL
K–GRAFOIL-1 K–GRAFOIL-2
1h 1h
1.7 104 1.9 104
2.4 103 1.7 103
2.7 1021 3.6 1021
3.9 101 3.3 101
HOPG
K–HOPG-1 K–HOPG-2
1h 4h
3.7 104 3.8 105
3.5 103 5.9 103
1.8 1021 1.1 1021
1.3 102 2.2 103
(b) K-GRAFOIL-1
(c) K-HOPG-1
5
5
5
4
4
3
μe
2
1
0
0
20
40
60
80
100
120
μe
2
1
0
20
40
60
80
σ ne
4
3
0
140
σ ne
ne/ne0, μe/μe0, σ/σ0
σ ne
ne/ne0, μe/μe0, σ/σ0
ne/ne0, μe/μe0, σ/σ0
(a) K-PGS-1
100
μe
3
2
1
0
0
20
t/h
t/h
40
60
80
100
t/h
Fig. 2. Changes in the electrical conductivities (s ), electron densities (ne), and electron mobilities (me) for K–GICs prepared from (a) PGS, (b) GRAFOIL, and (c) HOPG, after exposure to air at room temperature.
maintained for a long period in air [10]. However, me of K–HOPG decreased to 56% of the initial value, unlike the cases of K–PGS and K–GRAFOIL, where me increased relative to the initial values. The HOPG has a higher mobility (m = me = mh) compared to PGS and GRAFOIL, because the graphite crystallites in HOPG are more perfect and larger. It was hypothesized that the damage to the crystallite structure by the de-intercalation of K atoms in K–HOPG was greater than that in K–PGS and K–GRAFOIL. As a result, me was not recovered and was decreased. The high stability of K–PGS is caused by the recovery of me during the de-intercalation. This recovery is considered to be related in part to the unique structure of PGS. PGS graphite sheets are characteristically flexible and strong. They are composed of nearly ideal thin graphite layers that are 6–7 nm thick. The PGS sheet is thought to contain minutes vesicles, which impart flexibility and excellent mechanical properties [14,15]. We believe that these flexible layers are also effective in improving the airstability of K–GICs. The flexible layer structure is expected to buffer
disturbances by the de-intercalation of K atoms, as a result of which me can be recovered. In other words, the perfect graphite crystalline structure (like HOPG) is not suitable for the synthesis of GICs that are air-stable and have high electrical conductivity. 3.3. Changes in the electrical properties of Li–GIC and CuCl2–GIC in air The air-stability of K–GICs could be improved greatly by using flexible graphite sheets, such as PGS, as host materials. However, some GICs were not stabilized even if PGS was used. In our previous work [10], the changes in the electrical conductivities of Li–GIC (LiC6), K–GICs (KC8), and Cs–GIC (CsC8) specimens prepared from PGS, after exposure to air were measured. The electrical conductivities of the K–GICs and Cs–GIC specimens remained high for long periods, whereas the electrical conductivity of the Li–GIC specimen decreased immediately upon coming into contact with air and decreased below that of the host PGS within 1 day. Cs–GIC and K–GICs specimens exhibited stage-3 (CsC36) and stage-4 (KC48)
Table 3 Measured electrical conductivity (s ), magnetoresistance (Dr/r), and Hall coefficient (RH), and estimated carrier densities (ne and nh) and mobilities (me and mh) of Li–GIC prepared from PGS and CuCl2–GIC prepared from GRAFOIL, after exposure to air at room temperature. Specimens Li–PGS CuCl2–GRAFOIL
s/S cm1 4.2 104 5.4 103
Dr/r
RH (at 0.4 T)/cm3 C1
ne/cm3
nh/cm3
me/cm2 V1 s1
mh/cm2 V1 s1
(at 0.4 T)/– 2.8 102 7.3 104
1.2 102 1.9 102
3.4 1020 7.6 1015
1.5 1016 2.4 1020
7.6 102 1.3 104
6.7 104 1.4 102
R. Matsumoto / Synthetic Metals 198 (2014) 107–112
111
2.5
5 σ nh
2.0
ne/ne0, μe/μe0, σ/σ0
nh/nh0, μh/μh0, σ/σ0
σ ne
4
μe
3 2 1 0
μh
1.5
1.0
0.5
0.0
0
20
40
60
80
100
0
50
structures, after 2 months of exposure to air, respectively. However, for the Li–GIC specimen, it appeared that almost all of the Li atoms were extracted from the interlayers of the graphite. The electrical properties for Li–PGS are summarized in Table 3. Because Dr/r was observed in this case, the calculations were done based on a two-carrier system with electrons and holes. Although ne was not very large, s was higher than in the case of K–PGS because me was higher. It is possible that the Li atoms, being smaller than K, slightly damage the graphite layers during intercalation. However, it may be noted that as the decomposition of Li–GIC is rapid, these measurement values are probably affected by the decomposition of the specimens. Fig. 3 shows the changes in the electrical properties for Li–GICs prepared from PGS after exposure to air. The vertical axes are normalized by the initial values (s 0, ne0, me0). In this case, ne and me were estimated by a trial and error method, because Newton’s method could not be applied. After exposure to air for 1 day, ne rapidly decreased to 0.85% of the initial value. The final value of 2.9 1018 cm3 was only 1.2 times that of the host PGS. The extensive reduction in ne rendered the approximation in Eq. (7) unsuitable and this was the reason why the Newton’s method could not be applied for Li–PGS. Although me was increased to 510% of the initial value, the resulting s significantly decreased to 5.8%. In this case, it was considered that the de-intercalation was too large and immediate to retain the high electrical conductivity. In Table 3, the electrical properties for CuCl2–GRAFOIL are also summarized. CuCl2–GIC is known to be air-stable [19]. For this substrate, as the intercalated CuCl2 molecules extract electrons from the adjacent graphene planes, the hole density (nh) in the graphene planes becomes higher and results in p-type conduction, therefore, RH is positive. As the Dr/r observed for CuCl2–GRAFOIL was smaller than that for Li–PGS, it was found that the electrons contributed to conduction only slightly. In the case of CuCl2– GRAFOIL, s was not high at only 4.9 times that of the host GRAFOIL because nh was low (2.4 1020 cm3). Fig. 4 shows the changes in the electrical properties of CuCl2– GIC prepared from GRAFOIL after exposure to air. The vertical axes are normalized by the initial values (s 0, nh0, mh0). For this intercalated material, the changes in the hole density and mobility (nh/nh0 and nh/nh0) after 8 days (192 h) of exposure to air were very small at 91% and 120%, respectively. Consequently, the change in the electrical conductivity (s /s 0) was also small (110%). This
150
200
250
300
350
t/h
t/h Fig. 3. Changes in the electrical conductivities (s ), electron densities (ne), and electron mobilities (me) for the Li–GIC prepared from PGS, after exposure to air at room temperature.
100
Fig. 4. Changes in the electrical conductivities (s ), hole densities (ne), and hole mobilities (me) for the CuCl2–GIC prepared from GRAFOIL, after exposure to air at room temperature.
indicates that CuCl2 molecules remained intercalated in the graphite interlayers, and de-intercalation did not occur. 4. Conclusions The causes for the air-stabilization of GICs were investigated. The electrical conductivity was assessed in terms of electrical carrier densities and mobilities and, their changes upon exposure to air were measured. The decomposition processes for the GICs upon exposure to air varied depending upon the characteristics of the host graphite materials and the intercalated species. The reasons behind the high and stable electrical conductivities of K–GICs prepared from flexible graphite sheets such as PGS could be understood from the changes in the electron density and mobility. Whereas, the electron density was decreased by the de-intercalation of K atoms after exposure to air, the electron mobility was increased sufficiently. As a result, the electrical conductivity was almost unchanged. Two factors were found to be responsible for the retention of high electrical conductivity in air for the GICs. The first is that deintercalation is limited, and the second is the flexible nature of the graphite layer structures, which makes possible the increase in the carrier mobility, which compensates for the decrease in the carrier density. Furthermore, a perfect graphite structure (like HOPG) is not suitable for synthesizing GICs that are required to exhibit high and stable electrical conductivity. Acknowledgement This work was supported by KAKENHI Grant Number 24560830. References [1] M.S. Dresselhaus, G. Dresselhaus, Adv. Phys. 51 (2002) 1–186. [2] I. Khrapach, F. Withers, T.H. Bointon, D.K. Polyushkin, W.L. Barnes, S. Russo, M.F. Craciun, Adv. Mater. 24 (2012) 2844–2849. [3] F. Gunes, H.-J. Shin, C. Biswas, G.H. Han, E.S. Kim, S.J. Chae, J.-Y. Choi, Y.H. Lee, ACS Nano 4 (2010) 4595–4600. [4] R. Murali, K. Brenner, Y. Yinxiao, T. Beck, J.D. Meindl, IEEE Electron Device Lett. 30 (2009) 611–613. [5] I. Khrapach, F. Withers, T.H. Bointon, D.K. Polyushkin, W.L. Barnes, S. Russo, M.F. Craciun, Adv. Mater. 24 (2012) 2844–2849. [6] M. Inagaki, G. Watanabe, Synth. Met. 94 (1998) 235–238. [7] R. Matsumoto, Y. Hoshina, N. Akuzawa, Mater. Trans. 50 (2010) 1607–1611. [8] R. Matsumoto, N. Akuzawa, Y. Takahashi, Mater. Trans. 47 (2006) 1458–1463.
112
R. Matsumoto / Synthetic Metals 198 (2014) 107–112
[9] R. Matsumoto, Y., Okabe, N. Akuzawa, J. Electron. Mater., accepted. [10] R. Matsumoto, M. Arakawa, H. Yoshida, N. Akuzawa, Synth. Met. 162 (2012) 2149–2154. [11] Y. Gotoh, K. Tamada, N. Akuzawa, M. Fujishige, K. Takeuchi, M. Endo, R. Matsumoto, Y. Soneda, T. Takaichi, J. Phys. Chem. Solids 74 (2013) 1482– 1486. [12] R. Matsumoto, Y. Takahashi, N. Akuzawa, Mol. Cryst. Liq. Cryst. 340 (2000) 43–48. [13] M. Murakami, N. Nishiki, K. Nakamura, J. Ehara, H. Okada, T. Kouzaki, K. Watanabe, T. Hoshi, S. Yoshimura, Carbon 30 (1992) 255–262.
[14] N. Nishiki, H. Hake, M. Murakami, S. Yohimura, K. Yoshino, IEEJ Trans. FM 123 (2003) 1115–1123 (in Japanese). [15] N. Nishiki, H. Hake, K. Watanabe, M. Murakami, S. Yohimura, K. Yoshino, IEEJ Trans. FM 124 (2004) 812–816 (in Japanese). [16] A. Marchand, R. Mathur, Carbon 27 (1989) 349–357. [17] N. Akuzawa, S. Takei, M. Yoshioka, Y. Takahashi, Carbon 29 (1991) 899–903. [18] R. Matsumoto, TANSO, 2003 No. 209 2003 174–178 (in Japanese). [19] J.R. Gaier, M.E. Slabe, N. Shaffer, Carbon 26 (1988) 381–387.
Contents lists available at ScienceDirect
Synthetic Metals journal homepage: www.elsevier.com/locate/synmet
Investigation of the high, stable electrical conductivity in graphite intercalation compounds prepared from flexible graphite sheets Rika Matsumoto * Faculty of Engineering, Tokyo Polytechnic University, 1583 Iiyama, Atsugi, Kanagawa 243-0297, Japan
A R T I C L E I N F O
A B S T R A C T
Article history: Received 13 July 2014 Received in revised form 12 September 2014 Accepted 30 September 2014 Available online xxx
Potassium–graphite intercalation compounds (K–GICs) prepared from flexible graphite sheets are unusually stable in air and their high electrical conductivity can be maintained for long periods. To investigate the reasons for this air-stability, we evaluated the changes in the electrical conductivity as a function of electrical carrier densities and mobilities, upon exposure of the GICs to air. These electrical carrier parameters were estimated using the measured electrical conductivity, magnetoresistance, and Hall coefficient. The decomposition processes varied depending on the characteristics of both the host graphite and the intercalated materials. In the case of K–GICs prepared from flexible graphite sheets, the electron density was decreased by the de-intercalation of K atoms. However, the electrical conductivity hardly changed, because the electron mobility was increased sufficiently to compensate for the decrease in the density. The recovery of the electron mobility was thought to be caused by the flexibility of the graphite layers. Therefore, it was found that the air-stability of GICs could be determined by considering the degree of de-intercalation as well as the flexibility of the graphite layers. Furthermore, the perfect graphite crystal is not suitable to prepare air-stable and high conductive GICs. ã 2014 Elsevier B.V. All rights reserved.
Keywords: Graphite intercalation compound Electrical conductivity Air-stability Flexible graphite sheet Galvanomagnetic properties
1. Introduction Chemical species such as alkali metal atoms, metal halides, and halogen molecules can intercalate into the interlayers of lamellar graphite and form graphite intercalation compounds (GICs). A result of this intercalation is the occurrence of carrier transfer between the intercalated materials and adjacent graphite planes (graphene), causing the electrical conductivities of GICs to become approximately 10 times higher than those of the host graphite materials. However, GICs are known to be generally unstable in air. Most GICs are immediately decomposed by oxidation or the deintercalation of intercalated materials when exposed to the atmosphere [1]. Historically, GICs have attracted much attention as light-weight conductive materials. The highest conductivities ever reported for GICs exceeded those of copper and silver [1]. However, the air-instability of GICs has obstructed their practical use. Improving the air-stability of GICs is considered to be profitable. For example, nanocarbons, i.e., carbon-based nanosized materials such as graphene and carbon nanotubes, have been studied for practical use in electrodes [2,3], large-scale integrated circuits [4],
* Tel.: +81 46 242 9575; fax: +81 46 242 9575. E-mail address: [email protected] (R. Matsumoto). http://dx.doi.org/10.1016/j.synthmet.2014.09.037 0379-6779/ ã 2014 Elsevier B.V. All rights reserved.
and other applications. However, in order for nanocarbons to be used more effectively, the enhancement of their electrical conductivities is necessary. One technique that can increase the electrical conductivity of nanocarbons is intercalation [5]. However, the decomposition rates of GICs prepared from smaller graphite particles are faster, and they decompose easily in air [6]. Therefore, when carrying out the intercalation of nanocarbons, the lower airstability of the intercalated products is a problem. In another example, we have studied the thermoelectric properties of GICs, and found that they have possible applications as thermoelectric materials [7–9]. Although, both n- and p-type GICs are necessary in thermoelectric applications, there are very few air-stable GICs, especially n-type GICs, reported in the literature. In our investigations of GICs stability, we have succeeded in forming air-stable GICs that exhibit high electrical conductivity. In previous papers [10,11], we reported that potassium–graphite intercalation compounds (K–GICs) prepared from commercially available flexible graphite sheets (manufactured by the pyrolysis of polyimide films), exhibited high stability and high electrical conductivity for an extended period of time in air. However, we were unable to thoroughly explain the stabilization mechanism in the previous paper. We also found that Li–GICs could not be stabilized even when using the same flexible graphite sheet. Our ultimate goal is to establish an intercalation technique for carbon/ graphite materials, which produces intercalation compounds with
108
R. Matsumoto / Synthetic Metals 198 (2014) 107–112
high electrical conductivity and air-stability. Therefore, it is important to first understand the stabilization mechanism in the previously studied K–GICs. The air-stabilities of GICs are usually investigated by examining changes in their structure using X-ray diffraction (XRD), or by observing changes in their electrical conductivity. However, since we wish to use the intercalation technique for practical applications, we consider the long-term maintenance of high electrical conductivity in air as a particularly important parameter. Therefore, in this study, “stability” refers to the retention of high electrical conductivity. In our experience, the change in electrical conductivity is a more sensitive indicator of GICs decomposition than changes in the XRD patterns [12]. In this study, the causes for the stabilization of K–GICs prepared from flexible graphite sheets were investigated by monitoring the changes in electrical carrier densities and mobilities. Electrical conductivity is proportional to the product of the electrical carrier density and mobility. The electrical carrier density provides information about the amount of carrier transfer between the graphene and intercalated materials. In other words, it is thought that the electrical carrier density provides an indication of the amount of intercalated materials in the interlayers. Conversely, the electrical carrier mobility provides information about the number of scattering factors of the phonon. It is thought that the mobility also indicates the degree of perfection of the graphite structure. Therefore, an investigation of the carrier density and mobility can lead to an understanding of the amount of intercalated materials and the degree of perfection of the graphite crystalline structure. 2. Experimental 2.1. Preparation and measurements PGS1 graphite sheets with a thickness of 0.1 mm (Panasonic Co., EYGS 182310), GRAFOIL1 sheets with thicknesses of 0.1 mm and 0.3 mm (GrafTech Co., GTA grade), and highly ordered pyrolytic graphite plates (HOPG) (NT-MDT Co., grades ZYH and ZYA) were used as the host graphite materials. PGS is produced by the thermal decomposition of a polyimide film [13–15]. GRAFOIL is made from H2SO4–GICs, which are prepared from pure graphite flakes. After the exfoliation of the GICs, the resulting exfoliated graphite is sheeted. The PGS and GRAFOIL sheets were cut into 3–5 20– 25 mm2 rectangles. HOPG (10 10 2 mm3) was cleaved with the edge of a cutter and an adhesive tape, and cut into 10 5 0.3 mm3 rectangles. GRAFOIL was heat treated under vacuum at 900 C for 4 h before use to remove S, which is a common impurity. Potassium, lithium, and anhydrous copper(II) chloride (CuCl2) were used as the intercalate species without further purification. The GICs were prepared by allowing graphite to react with the vapor of the intercalate species at specific temperatures under vacuum. The reaction temperatures and durations for K, Li, and CuCl2 intercalation were 473 K for 3 days, 723 K for 7 days, and 753 K for 45 days, respectively. The structures formed were saturated stage-1 structures, where the intercalated layers existed in all the graphite interlayers. The intercalated layers are inserted in an orderly manner into the graphene layers. In a stage1 structure, there are intercalated layers between all the graphene layers, whereas in stage-2 and stage-n structures, the intercalated layers occur at intervals of two layers and n layers, respectively. Measurements of the in-plane electrical conductivity (s ), magnetoresistance (Dr/r), and the Hall coefficient (RH) were performed by the five-terminal method after the GICs specimens were exposed to air, as shown in Fig. 1, with magnetic fields (B) up to 0.5 T at room temperature. To measure the changes in the values over time, the specimens were kept on a measurement holder in
Fig. 1. Measurement set-up for the electrical conductivity (s ), magnetoresistance (Dr/r), and Hall coefficient (RH).
air. The average temperature and humidity during the measurements, and storage were 293–300 K and 50%, respectively. 2.2. Estimation of carrier density and mobility The electrical carrier densities (ne and nh) and mobilities (me and mh) were estimated from the values of s , Dr/r, and, RH [16,17]. In cases where Dr/r was not detected, the transport was considered to occur through one-carrier conduction by either electrons or holes. Therefore, the carrier density (n) and mobility (m) were estimated from Eqs. (1) and (2), where e is the electron charge.
s ¼ enm
(1)
RH ¼ ð1=neÞ
(2)
In cases, where Dr/r was detected, the transport was considered to occur via two-carrier conduction by both electrons and holes. The carrier densities and mobilities were then estimated using Eqs. (3)–(5).
s ¼ eðne me þ nh mh Þ
(3)
1 ne me 2 nh mh 2 RH ¼ e ðne me þ nh mh Þ2
(4)
Dr=r ¼
ne nh me mh ðme þ mh Þ ðne nh þ me mh Þ2
B
(5)
Furthermore, the approximations shown in Eqs. (6) and (7) were adopted for simplicity. Eq. (6) was used for graphite, whereas Eq. (7) was used for GICs.
m ¼ me ¼ mh
(6)
ðne nh ÞGraphite ¼ ðne nh ÞGIC
(7)
For practical calculations, we adopted Newton’s method for a quick and correct analysis [18]. In cases where the Newton’s method of analysis was not adopted, the trial and error methods were used. The set of parameter values (ne,nh, me, and mh) to minimize the error was searched by scanning. 3. Results and discussion 3.1. Measured and estimated electrical properties of host graphite and K–GICs In our previous work [10], the changes in the electrical conductivity of K–GICs specimens (with a composition of KC8)
R. Matsumoto / Synthetic Metals 198 (2014) 107–112
prepared from PGS, GRAFOIL, and HOPG after exposure to air were measured. For the specimens prepared from GRAFOIL and HOPG, a clear reduction in the electrical conductivity occurred immediately after the specimens were exposed to air. In contrast, for the specimens prepared from PGS, the electrical conductivity values decreased only slightly and gradually, and were maintained at a value that was six times higher than that of host PGS, even after 30 days in air. In this study, to investigate the cause of the airstability of the K–GICs prepared from PGS, we assessed the electrical conductivity in terms of its constituent factors, namely electrical carrier densities and mobilities. Table 1 summarizes the measured and estimated electrical properties for the host graphite materials. Since, the measurements were conducted several times for each host and the values were somewhat scattered, the exhibited maximum and minimum electrical conductivity values are listed. Although, thick (0.3 mm) and thin (0.1 mm) GRAFOIL specimens were used, no significant differences were noted between them. Similarly, there were no significant differences between specimens prepared from good (ZYA) and poor (ZYH) grade HOPG. The electron density (ne) was slightly larger than the hole density (nh) for all of the graphite materials. Further, both the carrier densities (ne, nh) and the mobilities (m) exhibited the same trend, i.e., HOPG > PGS > GRAFOIL. However, it should be noted that the accuracy of values for HOPG is low. While the manufacturer specification for the HOPG electrical conductivity is 2.5 104 S cm1, the values that we measured were significantly smaller and scattered in the range of 0.7 104 S cm1–1.5 104 S cm1. This may be a result of the specimen’s imperfect rectangular form. Furthermore, the structural perfection of HOPG might have been damaged, because we roughly cut and peeled off the HOPG plate with the edge of a cutter and adhesive tape. However, since the changes in the values of electrical conductivity were more important in this study than the absolute values themselves, we concluded that a slight inaccuracy in the absolute values would not significantly impact our interpretations. The measured electrical conductivities (s ) and Hall coefficients (RH), as well as the estimated electron densities and mobilities (ne and me), for K–GICs prepared from PGS, GRAFOIL, and HOPG (denoted as K–PGS, K–GRAFOIL, and, K–HOPG, respectively), are summarized in Table 2. All the K–GICs specimens failed to demonstrate magnetoresistance (Dr/r) and the RH values were negative, indicating that the K–GICs exhibit one-carrier conduction using only electrons. In a K–GICs, the intercalated K atoms provide electrons to the adjacent graphene planes, which increase the ne in the graphene planes, resulting in n-type conduction. On the other hand, me becomes lower than that of the host graphite because the number of scattering factors of the phonon is increased by Kintercalation. A few K–GICs specimens for each host were prepared and their electrical properties were measured (data for replicate samples are
109
indicated as 1 or 2). The values for K–HOPG, especially s, varied widely, likely due to the irregularity of the specimen shape and the degradation of the graphite structure as mentioned above. The measurements were started within 1 h after exposure of the specimen to air. Although, each specimen was prepared as a stage1 K–GICs with a KC8 composition, the structure of the specimen at the onset of the measurement was probably not pure stage-1 because the K–GICs are likely to have decomposed to some degree after exposure to air. Based on our previous work [10], the structures at the time of measurement were supposed to be stage3 (KC36) or stage-4 (KC48) for K-PGS and K-HOPG, a mixture of a small amount of stage-1 and higher stages for K–GRAFOlL. 3.2. Changes in the electrical properties of K–GICs prepared from PGS, GRAFOIL, and HOPG in air Fig. 2 shows the changes in electrical conductivities (s /s 0), electron densities (ne/ne0), and, electron mobilities (me/me0) for K– PGS-1, K–GRAFOIL-1, and K–HOPG-1 after exposure to air. The vertical axes are normalized by the initial values (s 0, ne0, me0), which were obtained after 1 h of exposure to air. Measurements for a second set of specimens, K–PGS-2, K–GRAFOIL-2, and K– HOPG-2, were also conducted, with similar results. The K–PGS specimens were stable in air, as reported. The s value was maintained at 78% of the initial value after 6 days (144 h) of exposure to air. However, as a result of the exposure to air for 6 days, ne decreased to 17% and me increased to 460% of the initial values. As mentioned above, the K–PGS structure transitioned from stage-1 to stage-4 immediately upon exposure to air. Moreover, it was observed that the K–PGS specimens were dampened with a strong basic solution (KOH) after exposure to air. This indicates that the K atoms in the interlayers were partially de-intercalated and reacted with H2O in the air. From these observations, it was concluded that the de-intercalation of K atoms occurred within a short duration after the K–PGS specimens were exposed to air. After reaching the stage-4 structure, the de-intercalation in K–PGS was abated, This is thought to be due to the coating of the graphite crystallite edges with air-stable substances such as oxides. In the case of K–GRAFOIL, ne decreased and me increased upon exposure to air. However, the change in the electron density (ne/ ne0) was larger than that in K–PGS; ne decreased to 12% of the initial value within 3 days, while me increased to only 230%. Consequently, s decreased by a significant amount, to 27% of the initial value. In the case of K–HOPG, ne/ne0 was relatively low, reaching 34% of the initial value after 4 days in air. Although this value is twice as large as that of K–PGS, the observation is easily understood. Since, the graphite crystallites in HOPG are larger than those in PGS, the diffusion of K atoms in the interlayers is expected to be difficult. In fact, we have found that the structural change of K–HOPG in air is slow and similar to that of K–PGS and the near-stage-4 structure is
Table 1 Measured electrical conductivity (s ), magnetoresistance (Dr/r), and Hall coefficient (RH), and estimated carrier densities (ne and nh) and mobility (m) of PGS, GRAFOIL, and HOPG at room temperature. Specimens
s /S cm1
Dr/r
RH (at 0.4 T)/cm3 C1
ne/cm3
nh/cm3
m/cm2 V1 s1
(at 0.4 T)/ PGS
4.6 103 4.3 103
6.2 102 6.5 102
1.3 101 1.5 101
2.5 1018 2.3 1018
2.1 1018 1.9 1018
6.3 103 6.4 103
GRAFOIL
1.2 103 1.1 103
1.6 102 9.6 103
9.6 102 9.8 102
1.2 1018 1.4 1018
1.1 1018 1.3 1018
3.2 103 2.4 103
HOPG
1.5 104 7.0 103
1.6 101 6.9 102
5.7 102 7.2 102
5.2 1018 3.6 1018
4.3 1018 3.1 1018
1.0 104 6.6 103
110
R. Matsumoto / Synthetic Metals 198 (2014) 107–112
Table 2 Measured electrical conductivity (s ), magnetoresistance (Dr/r), and Hall coefficient (RH), and estimated electron density (ne) and mobility (me) of K–GICs prepared from PGS, GRAFOIL, and HOPG after exposure to air at room temperature. Host
Specimens
Time in air
s/S cm1
RH (at 0.4 T)/cm3 C1
ne/cm3
me/cm2V1 s1
PGS
K–PGS-1 K–PGS-2
1h 1h
3.8 104 3.9 104
3.5 103 7.2 103
1.8 1021 8.7 1020
1.3 102 2.8 102
GRAFOIL
K–GRAFOIL-1 K–GRAFOIL-2
1h 1h
1.7 104 1.9 104
2.4 103 1.7 103
2.7 1021 3.6 1021
3.9 101 3.3 101
HOPG
K–HOPG-1 K–HOPG-2
1h 4h
3.7 104 3.8 105
3.5 103 5.9 103
1.8 1021 1.1 1021
1.3 102 2.2 103
(b) K-GRAFOIL-1
(c) K-HOPG-1
5
5
5
4
4
3
μe
2
1
0
0
20
40
60
80
100
120
μe
2
1
0
20
40
60
80
σ ne
4
3
0
140
σ ne
ne/ne0, μe/μe0, σ/σ0
σ ne
ne/ne0, μe/μe0, σ/σ0
ne/ne0, μe/μe0, σ/σ0
(a) K-PGS-1
100
μe
3
2
1
0
0
20
t/h
t/h
40
60
80
100
t/h
Fig. 2. Changes in the electrical conductivities (s ), electron densities (ne), and electron mobilities (me) for K–GICs prepared from (a) PGS, (b) GRAFOIL, and (c) HOPG, after exposure to air at room temperature.
maintained for a long period in air [10]. However, me of K–HOPG decreased to 56% of the initial value, unlike the cases of K–PGS and K–GRAFOIL, where me increased relative to the initial values. The HOPG has a higher mobility (m = me = mh) compared to PGS and GRAFOIL, because the graphite crystallites in HOPG are more perfect and larger. It was hypothesized that the damage to the crystallite structure by the de-intercalation of K atoms in K–HOPG was greater than that in K–PGS and K–GRAFOIL. As a result, me was not recovered and was decreased. The high stability of K–PGS is caused by the recovery of me during the de-intercalation. This recovery is considered to be related in part to the unique structure of PGS. PGS graphite sheets are characteristically flexible and strong. They are composed of nearly ideal thin graphite layers that are 6–7 nm thick. The PGS sheet is thought to contain minutes vesicles, which impart flexibility and excellent mechanical properties [14,15]. We believe that these flexible layers are also effective in improving the airstability of K–GICs. The flexible layer structure is expected to buffer
disturbances by the de-intercalation of K atoms, as a result of which me can be recovered. In other words, the perfect graphite crystalline structure (like HOPG) is not suitable for the synthesis of GICs that are air-stable and have high electrical conductivity. 3.3. Changes in the electrical properties of Li–GIC and CuCl2–GIC in air The air-stability of K–GICs could be improved greatly by using flexible graphite sheets, such as PGS, as host materials. However, some GICs were not stabilized even if PGS was used. In our previous work [10], the changes in the electrical conductivities of Li–GIC (LiC6), K–GICs (KC8), and Cs–GIC (CsC8) specimens prepared from PGS, after exposure to air were measured. The electrical conductivities of the K–GICs and Cs–GIC specimens remained high for long periods, whereas the electrical conductivity of the Li–GIC specimen decreased immediately upon coming into contact with air and decreased below that of the host PGS within 1 day. Cs–GIC and K–GICs specimens exhibited stage-3 (CsC36) and stage-4 (KC48)
Table 3 Measured electrical conductivity (s ), magnetoresistance (Dr/r), and Hall coefficient (RH), and estimated carrier densities (ne and nh) and mobilities (me and mh) of Li–GIC prepared from PGS and CuCl2–GIC prepared from GRAFOIL, after exposure to air at room temperature. Specimens Li–PGS CuCl2–GRAFOIL
s/S cm1 4.2 104 5.4 103
Dr/r
RH (at 0.4 T)/cm3 C1
ne/cm3
nh/cm3
me/cm2 V1 s1
mh/cm2 V1 s1
(at 0.4 T)/– 2.8 102 7.3 104
1.2 102 1.9 102
3.4 1020 7.6 1015
1.5 1016 2.4 1020
7.6 102 1.3 104
6.7 104 1.4 102
R. Matsumoto / Synthetic Metals 198 (2014) 107–112
111
2.5
5 σ nh
2.0
ne/ne0, μe/μe0, σ/σ0
nh/nh0, μh/μh0, σ/σ0
σ ne
4
μe
3 2 1 0
μh
1.5
1.0
0.5
0.0
0
20
40
60
80
100
0
50
structures, after 2 months of exposure to air, respectively. However, for the Li–GIC specimen, it appeared that almost all of the Li atoms were extracted from the interlayers of the graphite. The electrical properties for Li–PGS are summarized in Table 3. Because Dr/r was observed in this case, the calculations were done based on a two-carrier system with electrons and holes. Although ne was not very large, s was higher than in the case of K–PGS because me was higher. It is possible that the Li atoms, being smaller than K, slightly damage the graphite layers during intercalation. However, it may be noted that as the decomposition of Li–GIC is rapid, these measurement values are probably affected by the decomposition of the specimens. Fig. 3 shows the changes in the electrical properties for Li–GICs prepared from PGS after exposure to air. The vertical axes are normalized by the initial values (s 0, ne0, me0). In this case, ne and me were estimated by a trial and error method, because Newton’s method could not be applied. After exposure to air for 1 day, ne rapidly decreased to 0.85% of the initial value. The final value of 2.9 1018 cm3 was only 1.2 times that of the host PGS. The extensive reduction in ne rendered the approximation in Eq. (7) unsuitable and this was the reason why the Newton’s method could not be applied for Li–PGS. Although me was increased to 510% of the initial value, the resulting s significantly decreased to 5.8%. In this case, it was considered that the de-intercalation was too large and immediate to retain the high electrical conductivity. In Table 3, the electrical properties for CuCl2–GRAFOIL are also summarized. CuCl2–GIC is known to be air-stable [19]. For this substrate, as the intercalated CuCl2 molecules extract electrons from the adjacent graphene planes, the hole density (nh) in the graphene planes becomes higher and results in p-type conduction, therefore, RH is positive. As the Dr/r observed for CuCl2–GRAFOIL was smaller than that for Li–PGS, it was found that the electrons contributed to conduction only slightly. In the case of CuCl2– GRAFOIL, s was not high at only 4.9 times that of the host GRAFOIL because nh was low (2.4 1020 cm3). Fig. 4 shows the changes in the electrical properties of CuCl2– GIC prepared from GRAFOIL after exposure to air. The vertical axes are normalized by the initial values (s 0, nh0, mh0). For this intercalated material, the changes in the hole density and mobility (nh/nh0 and nh/nh0) after 8 days (192 h) of exposure to air were very small at 91% and 120%, respectively. Consequently, the change in the electrical conductivity (s /s 0) was also small (110%). This
150
200
250
300
350
t/h
t/h Fig. 3. Changes in the electrical conductivities (s ), electron densities (ne), and electron mobilities (me) for the Li–GIC prepared from PGS, after exposure to air at room temperature.
100
Fig. 4. Changes in the electrical conductivities (s ), hole densities (ne), and hole mobilities (me) for the CuCl2–GIC prepared from GRAFOIL, after exposure to air at room temperature.
indicates that CuCl2 molecules remained intercalated in the graphite interlayers, and de-intercalation did not occur. 4. Conclusions The causes for the air-stabilization of GICs were investigated. The electrical conductivity was assessed in terms of electrical carrier densities and mobilities and, their changes upon exposure to air were measured. The decomposition processes for the GICs upon exposure to air varied depending upon the characteristics of the host graphite materials and the intercalated species. The reasons behind the high and stable electrical conductivities of K–GICs prepared from flexible graphite sheets such as PGS could be understood from the changes in the electron density and mobility. Whereas, the electron density was decreased by the de-intercalation of K atoms after exposure to air, the electron mobility was increased sufficiently. As a result, the electrical conductivity was almost unchanged. Two factors were found to be responsible for the retention of high electrical conductivity in air for the GICs. The first is that deintercalation is limited, and the second is the flexible nature of the graphite layer structures, which makes possible the increase in the carrier mobility, which compensates for the decrease in the carrier density. Furthermore, a perfect graphite structure (like HOPG) is not suitable for synthesizing GICs that are required to exhibit high and stable electrical conductivity. Acknowledgement This work was supported by KAKENHI Grant Number 24560830. References [1] M.S. Dresselhaus, G. Dresselhaus, Adv. Phys. 51 (2002) 1–186. [2] I. Khrapach, F. Withers, T.H. Bointon, D.K. Polyushkin, W.L. Barnes, S. Russo, M.F. Craciun, Adv. Mater. 24 (2012) 2844–2849. [3] F. Gunes, H.-J. Shin, C. Biswas, G.H. Han, E.S. Kim, S.J. Chae, J.-Y. Choi, Y.H. Lee, ACS Nano 4 (2010) 4595–4600. [4] R. Murali, K. Brenner, Y. Yinxiao, T. Beck, J.D. Meindl, IEEE Electron Device Lett. 30 (2009) 611–613. [5] I. Khrapach, F. Withers, T.H. Bointon, D.K. Polyushkin, W.L. Barnes, S. Russo, M.F. Craciun, Adv. Mater. 24 (2012) 2844–2849. [6] M. Inagaki, G. Watanabe, Synth. Met. 94 (1998) 235–238. [7] R. Matsumoto, Y. Hoshina, N. Akuzawa, Mater. Trans. 50 (2010) 1607–1611. [8] R. Matsumoto, N. Akuzawa, Y. Takahashi, Mater. Trans. 47 (2006) 1458–1463.
112
R. Matsumoto / Synthetic Metals 198 (2014) 107–112
[9] R. Matsumoto, Y., Okabe, N. Akuzawa, J. Electron. Mater., accepted. [10] R. Matsumoto, M. Arakawa, H. Yoshida, N. Akuzawa, Synth. Met. 162 (2012) 2149–2154. [11] Y. Gotoh, K. Tamada, N. Akuzawa, M. Fujishige, K. Takeuchi, M. Endo, R. Matsumoto, Y. Soneda, T. Takaichi, J. Phys. Chem. Solids 74 (2013) 1482– 1486. [12] R. Matsumoto, Y. Takahashi, N. Akuzawa, Mol. Cryst. Liq. Cryst. 340 (2000) 43–48. [13] M. Murakami, N. Nishiki, K. Nakamura, J. Ehara, H. Okada, T. Kouzaki, K. Watanabe, T. Hoshi, S. Yoshimura, Carbon 30 (1992) 255–262.
[14] N. Nishiki, H. Hake, M. Murakami, S. Yohimura, K. Yoshino, IEEJ Trans. FM 123 (2003) 1115–1123 (in Japanese). [15] N. Nishiki, H. Hake, K. Watanabe, M. Murakami, S. Yohimura, K. Yoshino, IEEJ Trans. FM 124 (2004) 812–816 (in Japanese). [16] A. Marchand, R. Mathur, Carbon 27 (1989) 349–357. [17] N. Akuzawa, S. Takei, M. Yoshioka, Y. Takahashi, Carbon 29 (1991) 899–903. [18] R. Matsumoto, TANSO, 2003 No. 209 2003 174–178 (in Japanese). [19] J.R. Gaier, M.E. Slabe, N. Shaffer, Carbon 26 (1988) 381–387.