Supplementary Figure S1. XPS survey spectra of graphite, GO and RGO Na-NH3 powder samples.

Transcription

1 Supplementary Figure S1. XPS survey spectra of graphite, GO and RGO Na-NH3 powder samples. Supplementary Figure S2. FTIR spectra for GO, RGO Na-NH3 and RGO NH2-NH2. 1

2 Supplementary Figure S3. FTIR spectra of GO and RGO obtained by immersing in Na-NH 3 solution for different reduction time. In general, the GO samples consist of covalently attached oxygen-containing groups such as hydroxyl, epoxy, carbonyl, and carboxyl groups. The FTIR spectrum of GO illustrates the characteristic features including stretching vibration peak of O-H groups (carboxylic) centered at 3400 cm -1, the C=O stretching vibration peak in carboxyl and carbonyl at 1726 cm -1, deformation peaks of O H groups at 1405 cm -1, the C O-C (epoxy) stretching vibration peak at 1224 cm -1, and the C O (alkoxy) stretching vibration peak at 1052 cm The peak at 1620 cm -1 was assigned to C=C skeletal vibrations of un-oxidized graphitic domains or contribution from the stretching deformation vibration of intercalated water 33. We have prepared the RGO samples at different reduction stages by controlling the reduction time in the Na-NH 3 solution, and performed FTIR studies to find out how the functional groups evolve on GO when exposed to the solvated electrons. FTIR spectra clearly show the progressively removal of the oxygen groups by solvated electrons. (1) The carboxylic group initially shows its presence throughout the spectra of GO in the form of a very broad -OH peak centered at 3400 cm -1. However, just after 1 min reduction in Na-NH 3 solution, the O-H peak decreased dramatically and almost entirely disappeared. Therefore, the carboxylic groups are particularly affected by the solvated electrons. (2) The appearance of the weak shoulders at cm -1 corresponds to C-H stretching vibration, which may be due to the open-circle of the epoxide and protonation by ammonia. In addition, with 2

3 the increase of the reduction time, it is found that the intensity of C-H stretching vibration at 2970 cm -1 decreases and finally disappears, indicating the dehydrogenation and the recovery of the C=C backbone. The carbon backbone C=C stretch can be seen at 1562 cm -1 as in graphite. (3) The decreasing peak at 1726 cm 1 was a strong indication of carbonyl reduction in GO, suggesting that C=O group in products can be easily converted. (4) As the GO films are reduced by solvated electrons, peaks for oxygen functional groups were significantly reduced and perhaps entirely removed, and two broad peaks at 1562 cm -1 and 1186 cm -1 were found for the RGO. The peak at 1562 cm -1 could be assigned to the aromatic C=C stretch. The peak at 1186 cm -1 could be assigned to the C-O stretch; however, it is not clear which kind of C-O groups it is at this. These observations confirmed that most oxygen functionalities in the GO were removed. Supplementary Figure S4. Possible reduction mechanism and procedure for preparing the RGO Na-NH3 sheets with solvated electrons in ammonia. We thoroughly explored the possible de-oxygenation process of oxygen-containing groups in GO including epoxide, carbonyl, and hydroxyl and carboxyl. 16,24,45 The unique properties of Na-NH 3 solutions present unusual opportunities for the study of reduction of GO sheets. The sodium metal is dissolved in liquid ammonia (at -33 C) by forming the metal cation and a solvated electron. In contrast to a solution of sodium amide in liquid ammonia, which is a strong base, the deep blue solution of sodium in liquid ammonia is a very strong reducing agent The de-oxygenation and formation of pie bond 3

4 conjugation are explained as the following: (a) Possible mechanism proposed for de-epoxide of GO. As shown in Figure S4a, the solvated electron is so strong to attack the oxygen and open the epoxy (step 1). This one-electron transfer yields a C-O radical anion and an alkyl radical (step 2). The radical anion is then stabilized by protonation with an ammonia molecule to form a hydroxyl (step 3). Subsequently, the alkyl radical is also added by a solvated electron (step 4) and the formed alkyl anion, which is then protonated by ammonia to form C-H (step 5). Finally, the hydroxyl is attacked to form radical anion (step 6) and the dehydration (step 7) occurs, which might be further driven re-establishment of the conjugated graphene network in GO. The direct evidence for the protonation of alkyl anion has been demonstrated by the appearance of the C-H vibration at 2970 cm -1 in FTIR spectrum in 1min (Figure S3). However, the C-H peak disappears after 15 min reduction maybe due to dehydration process and C=C recovery. (b) Possible mechanism proposed for de-carbonyl of GO. As shown in Figure S4b, the carbonyl group (step 1) can readily accept a solvated electron, and its bond is cleaved to form radical anion and carbon radical (step 2). And then the radical anion is protonated by the ammonia to form hydroxyl (step 3). Meanwhile, the carbon radical would be reduced to the strongly basic carbanion (step 4), which would then be rapidly protonated by ammonia to form C-H (step 5). Bond fission or bond saturation may take place to produce a radical species and an anionic species, while bond saturation results in a radical-anion intermediate, which have been suggested previously in the metal-ammonia reductions 46,49,50. (c) Possible mechanism proposed for de-hydroxyl of GO. As shown Figure S4c, the hydroxyl can be attacked by the solvated electrons and readily form a radical anion. The fission of polar single bonds (C-O) occurs. Bond cleavage reaction mechanisms have been suggested previously in the metal-ammonia reductions 51. A general scheme is presented in this paper where A and B can represent either individual atoms or radical groups [AB + e s (solvated electrons) A + B - ]. (d) Possible mechanism proposed for de-carboxyl of GO. As shown in Figure S4d, we speculate that the possible mechanism for the reduction of carboxyl by solvated electron is due to the fission of polar single bonds (C-O) and the de-co 2 (step 2) to form a radical species (step 3). The carboxylic acids are presumably capable of forming radical anions paired with the sodium ion (step 2) prior to bond cleavage. 4

5 Supplementary Figure S5. Possible mechanism proposed for the de-oxygenation and the recovery of GO with solvated electrons in ammonia. In order to understand how the pi-bonds or the conjugation are repaired when oxygen is pulled out from the basal plane of GO, we have illustrated the possible mechanism for the de-oxygenation and the recovery of C=C in the GO sheet. From the outset, the de-oxygenation process can generate carbon radicals or carbanion, which, with the presence of solvated electrons in ammonia, can be more stable than usual and have a longer lifetime with the presence of solvated electrons in ammonia 50,52. Moreover, these radicals can be partially delocalized, it is possible for the adjacent radicals on the plane of GO to communicate and rearrange synergistically into a lower energy state. Especially, the adjacent carbon radicals or carbanion might be rearranged in GO to re-establish the conjugated graphene network to achieve the lowest energy state. As a result, numerous new graphitic (sp 2 ) domains can be restored, which is consistent with the analysis of the change of Raman spectra 32. 5

6 Supplementary Figure S6. RGO Na-NH3 film was achieved by reduction of the GO film in Na-NH 3 solution for just 1 min. Supplementary Figure S7. XPS analysis of the C 1s regions in GO film and RGO Na-NH3 film. 6

7 Supplementary Figure S8. Tapping mode AFM height images of vacuum-filtrated GO films on the alumina membrane before and after reductive treatments. (a,b) Continuous GO film shows the rough, wrinkled morphology. R q are 21.4 nm and 10.8 nm, respectively. (c,d) GO film reduced by Na-NH 3 solution for 3 min. R q is 18.3 nm and 9.9 nm, respectively. The GO sample was vacuum-filtrated on the alumina membrane (pore diameter: nm) and dried at room temperature. The roughness of GO films deposited on the alumina membrane may be slightly higher than the GO films spin coated on mica due to different deposited substrates 54. The RGO film was prepared by dipped GO film into the Na-NH 3 solution for 5 min, washed with deionized water, and dried at room temperature. 7

8 Supplementary Figure S9. (a) I-V characteristics of rgo multiple layers film transferred on a PET. (b) The linear I-V characteristics of the rgo film on a PET substrate is also observed between V=-5~5 V. Our measurement was conducted on large area RGO film with the distance between electrodes > 1 mm using two-point measurement as well as standard four-point probe approach using 4 point surface resistivity meter (RC2175). Supplementary Figure S10. Transmission spectrum of thin rgo film deposited on a PET substrate shows an average of ~80% transmittance in the visible region. Both measurements demonstrate that the transparent multi-layer rgo film on PET substrate with ~80% optical transmittance displays the sheet resistance around ohms/square, which is the lowest among solution processed graphene transparent conductors. The high electrical conductivity can be attributed to more complete reduction and restoration of C=C π-conjugation in the graphene basal plane due to the highly reductive nature of the solvated electrons, and clearly 8

9 demonstrates the advantage of our approach. Supplementary Figure S11 a) I d -V d curves of reduction-go FETs reduced by Na-NH 3 solution. Insert shows the optical images of a single rgo sheet FET device, and the scale bar is 5 µm. b) The I d -V g curves. Field effect transistors was fabricated on Si/SiO 2 (300 nm) substrate using single-layer rgo sheets as the semiconducting channel and the silicon substrate the back gate. The electrical transport studies on single layer single flake RGO shows a sheet resistance (R s ) of ~3.3 kω/sq. If we assume the resistance scales linearly with the number of layers, we expect a 10 layer rgo film could produce a sheet resistance of ~330 Ω/square with an ideal transmittance of 77% rgo film, which is in agreement with our experimental results obtained with bulk rgo film (~350 Ω/sq at 80% transmittance). One may argue that graphene film is made of multiple rgo flakes and there could be contact resistance from flake to flake. On the other hand, it is important to note that our single flake measurement is based on two-terminal measurement without excluding the contact resistance between the metal electrodes and the rgo flake. For the bulk rgo film, one would expect the homo-contact between the same rgo materials should be no worse than the rgo-metal hetero-contact. To further verify the high quality of the rgo sheets, the carrier mobility was also determined. The I d -V g shows a typical hole transport characteristics with no Diract pointed observed up to +80 V gate, indicating the rgo sample is relatively highly doped with holes. The hole carrier mobility was calculated from the linear region of the I d -V g curves using the following equation: ΔIds L µ =Δσ /( CgΔ Vg) = /( CgΔVg) Vd W 9

10 Where L and W are channel length and width, respectively; C g is the specific capacitance of the gate dielectric; I d, V d and V g are drain-source current, drain-source voltage and gate voltage, respectively. From the I d -V g curve, we derive a maximum hole-mobility of 123 cm 2 V -1 s -1. It should also be noted that the contact resistance is not excluded in the mobility determination, which could lead to an underestimation. Nonetheless, the observed hole-mobility represents a very high value reported in rgo samples, further demonstrating the unusually high quality of the rgo sample obtained with our approach. Supplementary Figure S12. Nitrogen adsorption and desorption isotherms of our RGO samples by Na-NH 3 reduced. We have checked the BET of freeze-dried RGO samples by Na-NH 3 solution and found the specific surface area 648 m 2 /g. 10

11 Supplementary Figure S13. Nyquist plot of the EIS of RGO-based ultracapacitors. The inset is equivalent circuit diagram of different elements from the EIS analysis. Nyquist plot for RGO electrode using a sinusoidal signal of 5 mv over the frequency range from 100 khz to 1 mhz. Z is real impedance. Z is imaginary impedance. The EIS data can be fitted by an equivalent circuit consisting of a bulk solution resistance R s, a charge-transfer R ct, a pseudocapacitive element C p from redox process, and a constant phase element (CPE) to account for the double-layer capacitance. The bulk solution resistance Rs and charge-transfer resistance R ct can be obtained from the Nyquist plots, where the high frequency semicircle intercepts the real axis at R s and (R s + R ct ), respectively. The solution resistance R s of these two composites was measured to be 8.5 Ω, while the charge-transfer resistance R ct was calculated to be 13.5 Ω. 11

12 Supplementary Table S1. Comparison of elemental analysis results of graphene using different methods. 12

13 Supplementary Table S2. Comparison of sheet resistance and transparency of graphene and related materials. 13

14 Supplementary Table S3. Comparison of mobility values for rgo FETs in other reports. Supplementary methods Oxidation of graphite Graphite oxide (GO) was synthesized from graphite powder (325 mesh) using a modified Hummer s method 40. In the pre-oxidized step, concentrated H 2 SO 4 (12 ml) was heated to 80 C in a 300 ml beaker. K 2 S 2 O 8 (2.5 g) and P 2 O 5 (2.5 g) were added with stirring until the reactants were completely dissolved. The mixture was kept at 80 C for 4.5 hours using a hotplate. Successively, the mixture was cooled to room temperature and diluted with 0.5 L of de-ionized (DI) water and left overnight. The following day the mixture was filtered and washed with DI water using a 0.2 micron Nylon Millipore filter to remove the residual acid. The product was dried under ambient condition overnight. For the oxidation step of the synthesis, the pretreated graphite powder was then put into cold (0 C) concentrated H 2 SO 4 (120 ml) using an ice bath. Then, KMnO 4 (15 g) was added 14

15 slowly under stirring and the temperature of the mixture was kept to be below 20 C by cooling. Successively, this mixture was then allowed to react at 35 C for 2 hours and then diluted with DI water (250 ml). Because the addition of water caused the temperature of the mixture to rise rapidly, it was done in an ice bath to keep the temperature below 50 C. After adding all of the 250 ml of DI water, the mixture was stirred for 2 h, and then additional 0.7 L of DI water was added. Shortly after the dilution with 0.7 L of water, 20 ml of 30% H 2 O 2 was added to the mixture, and the color of mixture changed into brilliant yellow along with bubbling. The mixture was filtered and washed with 1 L of 10% HCl aqueous solution to remove metal ions followed by 1 L of DI water to remove the acid. The resulting solid was dried in air and diluted to make a 0.5% w/w GO dispersion that was put through dialysis for 2 weeks to remove any remaining metal species. Exfoliation was carried out by sonicating 0.1 mg ml -1 GO dispersion under ambient condition for 20 min. The resulting homogeneous yellow-brown dispersion was tested to be stable for several months and used for reduction. Sodium metals in liquid ammonia: electronic solution Our results provide the deoxygenation of GO via solvated electrons treatment. For this purpose, we used the sodium-ammonia solution as a very potent electron source. Dissolution of the sodium metal in the liquid ammonia resulted in ionization of the metal to form a sodium cation and a solvated electron strongly associated with the solvent ammonia. The deep colour of solutions of sodium metal in liquid ammonia arises from the presence of solvated electrons. A solvated electron is a free electron in a solution, which can be used as a highly reducing agent. Sodium metal is too active to donate an electron in most circumstances. This case is no exception, although the location to which the electron is donated is a little unusual. Ammonia s liquid lattice has a recurring gap with a radius of about 3.3 A. Through a fortunate orientation of the dipole moments within the lattice, this space can actually accommodate an electron. Sodium donates the electron to the gap, even though the electron doesn't bond to the ammonia. In this manner, sodium can donate an electron and liquid ammonia doesn t have to pick it up. This leaves an uncoupled electron in solution, the solvated electron. Eventually, the surplus electronic solution will lose H 2 to form sodium amide: 2 Na (s) + 2 NH 3 (l) 2 NaNH 2 (s) + H 2 (g) It should be note that keeping low temperature is a relatively simple and safe process. Actually, solvated electrons in other solvent (such as 1,2-dimethoxyethane (DME)) 54 can be used at room temperature. The 15

16 operation of reactive metal sodium is actually very safe in our method as well compared with other reducing agent such as hydrazine. Vacuum filtration The GO suspension obtained was vacuum-filtrated through a mixed cellulose ester (MCE) membrane (Millipore) with 25 nm pores. We can achieve the controllable deposition of uniform layers by either varying the filtration volume or the concentration of GO in the suspension. Doubling the concentration has the same effect as doubling the filtration volume. Since individual sheets of GO were sufficiently larger than the pore dimension, growth occurred by creating a uniform continuous thin film of a single layer first and then by building additional layers on top. The GO suspension was also vacuum-filtrated through a polyvinylidene fluoride (PVDF) membrane and alumina membrane with nm pores. The calculation of the specific capacitance of the RGO based ultracapacitors The value of specific capacitance of the RGO ultracapacitor can be calculated by the following formula C sp I = ( dv / dt) m Where, I is discharge current, dv/dt is slope of discharge curve, and m is mass of each electrode of double layer capacitor. Here, we knew the mass of the GO film before reduction. We got the atomic composition of the GO and RGO film by XPS analysis. Because of the same carbon content before and after reduction, the amount of RGO can be estimated via the GO amount. Supplementary References 42. Guo, H., Wang, X., Wang, F., Qian, Q., Xia, X. A green approach to synthesis of graphene sheets using electrochemical technique. ACS Nano 3, (2009). 43. Zhang, J. et al. Reduction of graphene oxide vial-ascorbic acid. Chem. Commun. 46, (2010). 44. Lin, Z. et al. Solvent-assisted thermal reduction of graphite oxide. J. Phys. Chem. C 115, (2011). 45. Gao, X., Jang, J., & Nagase, S. Hydrazine and thermal reduction of graphene oxide: reaction mechanisms, product structures, and reaction design. J. Phys. Chem. C. 114, (2010) 16

17 46. Sinclair, S., & Jorgensen, W. L. Computer assisted mechanistic evaluation of organic reactions. 23. dissolving metal reductions with lithium in liquid ammonia including the Birch reduction. J. Org. Chrm. 59, (1994). 47. Huffman, J. W. Metal-ammonia reduction of cyclic aliphatic ketones. Acc. Chem. Res. 16, (1983). 48. Kharasch, M. S., Sternfeld, E., & Mayo, F. R. The reaction of esters with sodium in ammonia. J. Am. Chem. Soc. 61, 215 (1939). 49. Huffman, J. W. & McWhorter, W. W. Dissolving metal reduction of cyclic ketones. J. Org. Chrm. 44, (1979). 50. Corey, E. J., Ensley, H. E., Suggs, J. W. Convenient synthesis of (S)-(-)-pulegone from (-)-citronellol. J. Org. Chem. 5, 362 (1940). 51. Dewald, R. R. Mechanistic studies of metal-ammonia reductions. J. Phys. Chem. 79, (1975). 52. Kerr, J. A. Bond dissociation energies by kinetic methods. Chem. Rev. 66, (1966). 53. Becerril, H. A., Mao, J., Liu, Z., Stoltenberg, R., Bao, Z., Chen, Y. S. Evaluation of Solution processed functionalized graphene films as transparent conductors. ACS Nano 2, (2008). 54. Englert, J. M. et al. Covalent bulk functionalization of graphene. Nature Chem. 3, (2011). 17