Srinivas
Gadipelli
*a,
Yue
Lu
a,
Neal T.
Skipper
b,
Taner
Yildirim
c and
Zhengxiao
Guo
*a
aDepartment of Chemistry, University College London, 20 Gordon Street, London, WC1H 0AJ, UK. E-mail: gsrinivasphys@gmail.com; z.x.guo@ucl.ac.uk
bDepartment of Physics & Astronomy, University College London, Gower Street, London, WC1E 6BT, UK
cNIST Centre for Neutron Research, National Institute of Standards and Technology, Gaithersburg, Maryland 20899, USA
First published on 4th August 2017
We demonstrate a simple and fully scalable method for obtaining hierarchical hyperporous graphene networks of ultrahigh total pore volume by thermal-shock exfoliation of graphene-oxide (exfGO) at a relatively mild temperature of 300 °C. Such pore volume per unit mass has not previously been achieved in any type of porous solid. We find that the amount of oxidation of starting graphene-oxide is the key factor that determines the pore volume and surface area of the final material after thermal shock. Specifically, we emphasize that the development of the hyperporosity is directly proportional to the enhanced oxidation of sp2 CC to form CO/COO. Using our method, we reproducibly synthesized remarkable meso-/macro-porous graphene networks with exceptionally high total pore volumes, exceeding 6 cm3 g−1. This is a step change compared to ≤3 cm3 g−1 in conventional GO under similar synthetic conditions. Moreover, a record high amine impregnation of >6 g g−1 is readily attained in exfGO samples (solid-amine@exfGO), where amine loading is directly controlled by the pore-structure and volume of the host materials. Such solid-amine@exfGO samples exhibit an ultrahigh selective flue-gas CO2 capture of 30–40 wt% at 75 °C with a working capacity of ≈25 wt% and a very long cycling stability under simulated flue-gas stream conditions. To the best of our knowledge, this is the first report where a graphene-oxide based hyperporous carbon network is used to host amines for carbon capture application with exceptionally high storage capacity and stability.
Since their discovery, graphene based materials have been widely regarded as promising and scalable materials to obtain desirable structures.2,10,15–17 In most cases, such structures are synthesized by a top-down approach of graphite oxidation followed by chemical or thermal reduction. Thus, the achievable conductivity and porosity in the final graphene based networks is directly related to the strengths of the oxidation of the starting graphite (to yield graphene-oxide, GO), and the reduction (of GO) processes. In this work, we report a simple and well-controlled method for obtaining highly hierarchical meso- and macro-porous graphene networks, prepared by thermal-shock exfoliation of GO at moderate temperatures for a short time of about 5 minutes. Specifically, we demonstrate that the degree of oxidation of the GO controls the strength of exfoliation to yield extraordinary porosity: both the specific surface area (SSA, ≈800 m2 g−1) and ultrahigh total pore volume (≥6 cm3 g−1). Such hyperporous exfGO samples in turn exhibit remarkable cyclic working flue-gas CO2 capacities in solid-amine-impregnated systems (solid-amine@exfGO).
Firstly, precursor GO samples with increasing degree of oxidation (named GO-A, -B, -C & -D, in the ascending order of oxidation state) were synthesized by oxidation of graphite powder using a modified Hummers' method (see the Experimental and ESI, Fig. 1 and S1†). Note that there are environmentally green approaches for the synthesis of graphene-oxide at industrial quantities.18 The increased oxidation is confirmed by an enhanced layer spacing (by powder X-ray diffraction, PXRD), increased concentration of epoxy (–C–O–C–), hydroxyl (–C–OH) and carboxylic (–C–O–OH) groups on the graphene basal plane and edge (by X-ray photoemission spectroscopy, XPS and Raman spectroscopy), and excess mass-loss and exothermicity during the decomposition in thermal reduction (by combined thermogravimetry-differential scanning calorimetry-mass spectrometry, TG-DSC-MS) (Fig. 1a, b, S2–S7 and Table S1†). Specifically the CO/COO content is increased from ≈7.5% to ≥15.0% at the expense of sp2 CC carbon, which is reduced from ≈54.0% to ≤44.0% in the samples GO-A to GO-D. In addition to this, the TG reveals about 9 wt% of excess decomposition in GO-D with respect to the GO-A sample. PXRD also shows a larger interlayer distance of >0.92 nm in GO-D compared to ≈0.79 nm in GO-A. All these characteristics are consistent with a greater degree of oxidation in the GO-D sample.
Next, the exfoliation of these GO samples was carried out by direct thermal-shock at a relatively mild temperature of 300 °C, and for less than 5 minutes to achieve complete exfoliation (Fig. 1c–g and S6–S10†). The exfoliation is associated with an exothermic reaction, where an increase in the heating rate leads to a highly enhanced exothermicity at the decomposition point of GO (Fig. S6 and S7†). A relatively high exothermicity is observed for the highly oxidized GO-D sample compared to the less oxidized GO-A sample. Since the GO structure incorporates surface oxy-functional groups and hydrogen bonded inter-lamellar water molecules, the sudden volatility of these labile species leads to a pressure build-up between the graphene sheets. Therefore, by this process, a highly oxidized GO sample always leads to a high degree of exfoliation (Fig. S9 and S10†). This is exactly what is observed through the continuous increase in the volume of the exfGO samples, packed by tapping in a 4 ml vial at the constant mass of 32.5 mg (Fig. 1d and S9†). The exfoliation is identified with the loss of the layered structure and the reduction of oxygen content, to 12.5 at% from 32.0 at% in precursor GO (Fig. 1, S1–S12, Tables S1 and S2†).
N2 adsorption–desorption isotherms (type II or S-type) indicate a considerable enhancement in the porosity of the samples from exfGO-A to exfGO-D (Fig. 2, S13–S15 and Table S3†). For instance, the additional oxidization of the starting GO increases the SSA from 360 m2 g−1 to 830 m2 g−1, the total N2 uptake (at P/P0 of ≤0.994) from 1748 cm3 g−1 to 3953 cm3 g−1, and the corresponding total pore volume from 2.71 cm3 g−1 to 6.10 cm3 g−1. A total pore volume of over 6.00 cm3 g−1 in the exfGO-D samples is consistently achieved when the synthesis and exfoliation were repeated for four different batches of the samples. Such a total pore volume of ≥6 cm3 g−1 is extremely high considering that our synthesis route is very simple and scalable. We note that the highest pore volume reported in the literature is around 4 cm3 g−1, which is only obtained by rather complex and lab scale processing for a range of graphene based materials (see the ESI,† we compare over 300 samples summarized in Tables S4, S5, Fig. 2c, d and S16,† including exfoliated graphenes from GO by ultrasonication, electrochemical, microwave and thermal-shock, template and chemical activation), and other family of mesoporous carbons, MOFs, silicas and zeolites.3–20
Fig. 2 Development of porosity against increasing degree of oxidation of the GO samples: (a and b) N2 adsorption–desorption isotherms at 77 K (a) and their derived pore size (line data on the left Y-axis) and pore volume (dotted data on the right Y-axis) distributions (b). The inset in (a) shows the pressure region near P/P0 ≈ 1 and the clear maximum uptake differences between samples. (c and d) The comparison of the total pore volume (c) and SSA (d) of our exfGO and a large number of samples (over 250) from the literature. (e and f) Porosity (pore volume and SSA) generation as a function of oxidation state of the precursor GO samples. The development of the porosity is directly proportional to the oxidation strength of GO, where a drastic enhancement in the porosity is observed with an increased degree of oxidation, i.e., conversion of sp2 CC to C–O/CO/COO. Insets (e and f) showing the linear increase of porosity with respect to the carboxylic group CO/COO formation (see the ESI†). |
Furthermore, the highly hierarchical pore sizes and their distributions across the micro-, meso- and macro-porous regimes of the samples can be understood from the qualitative behaviour of N2 adsorption–desorption isotherms (Fig. 2 and S13†). The type II or S-type isotherms with a sharp hysteresis at high relative pressures indicate highly hierarchical pore sizes and unsaturated condensation of N2 in macroporous networks of a plate-like geometry in the exfGO samples. We note that the majority of the meso-porosity is distributed in the range 3 to 50 nm that accounts for more than 75% of the total pore volume and yields ultrahigh meso-pore volume of up to 4.65 cm3 g−1.
In a controlled experiment, slow heating of GO samples to 300 °C without exfoliation (GO-300C) leads to negligible porosity, similar to the as-synthesized precursor GO (asGO) samples (Fig. S9, S10, S14 and Table S3†). Note that the exfoliation of GO-D at a very high temperature, 600 °C to 1000 °C, induces no further enhanced porosity compared to the samples exfoliated at 300 °C (Fig. S17†).
Moreover, our critical analysis shows, for the first time, that the development of porosity is directly proportional to the strength of the oxidation of the precursor (Fig. 2e and f). Both the SSA and the total pore volume are linearly increased with respect to the oxidation of sp2 (CC) carbon into C–O/CO/COO. Here a high degree of reduction in exfGO-D samples with an oxygen content of ≈12.5 at% compared to the ≈14.0 at% in exfGO-A is noticed, which further supports the high exothermic heat generation during the thermal-shock of GO-D (Fig. S7, S11 and Table S2†). Thus, a high degree of oxidation of the GO with an increased concentration of carboxylic groups, and their induced large exothermic heat generation during the thermal decomposition are the desirable parameters to achieve high enough porosities under simple thermal-shock conditions.
Such porosity characteristics of the samples are further well supported by SEM and TEM micrographs (Fig. 3 and S18–S22†). A high degree of exfoliation and large pore networked graphenes can be observed for the exfGO-D compared to exfGO-A. The exfGO-D samples show highly interconnected graphene networks resembling a kirigami structure. A closer inspection of the SEM micrographs reveals a characteristic pore-size development, directly relevant to the exfoliation with respect to the increased degree of oxidation of the precursor GO sample. The highly oxidized GO samples exhibit a larger pore structure after exfoliation. The structures consist of interconnected pores, with the macro-pores leading to meso-pores, and the meso-pores leading to micro-pores.
The exceptional meso- and macro-porosity in exfGO samples, in conjunction with considerable (≈12 at%) amounts of surface oxygen groups, can facilitate a large amount of amine impregnation to achieve an extremely efficient solid-amine based CO2 capture system.4,7,8,13–15,21–28 Here, triethylenetetramine-impregnated exfGO (TETA@exfGO) samples were obtained by methanol solution infiltration (Fig. 4 and S23†). Note that TETA is a low-cost short-amine and exhibits low thermal stability compared to long-chain poly-amines.8,22 In a controlled synthesis, the asGO sample showed a wet surface amine coating when the TETA loading was increased over 1 g g−1 of sample, due to a limited layer/pore space (Fig. S24 and S25†). Due to their ultrahigh pore volume, exfGO-D samples can accept a record high TETA loading of ≈10 g g−1 of sample, without any sign of surface wetting (Table S6†).4,7,8,13–15,21–28 Indeed, SEM micrographs still show void spaces for a TETA loading as high as 7.0 g g−1 in the exfGO-D sample (Fig. S26†). A systematic study of amine-impregnation confirms that the pore volume of the sample is the controlling factor for determining the amine loading (see SEM, TG, XPS and 77 K N2 uptake porosity, Fig. 4a, b, S26–S30, Tables S7 and S8†). The amine impregnation results in definite shifts in the XPS binding energy of the C 1s and N 1s peaks due to the pore confinement and/or interaction of surface/functionality.8 Relative changes in the O 1s peaks show amine interaction with the surface oxygen functional groups. Amine impregnation is also confirmed from the N2 adsorption isotherms of TETA-loaded samples, which show a negligible uptake.
The CO2 capture capacities of the TETA@exfGO samples have been determined by both volumetric and gravimetric methods for single-component CO2 and a flue-gas stream (consisting of 15% CO2 in 85% N2 and humidified by bubbling through water at a flow rate of 100 ml min−1), respectively (Fig. 4c, d, S30–S34 and Table S8†). Undoubtedly, the CO2 capture capacity is proportional to the pore volume defined by TETA loading. Exceptionally high CO2 (100% dry) capture capacities of 5.7–7.5 mmol g−1 are obtained at a practicable temperature of 75 °C under a CO2 pressure of 0.15 bar (Fig. 4c, S30, Tables S6 and S8†).4,5,7,8,10,11,13–15,20–33 These uptake capacities are further confirmed under a simulated flue-gas stream of ≈100 ml min−1, consisting of only 15% of CO2 in N2, bubbled through water. A CO2 capture of 35–40 wt% from the flue-gas is attained at 75 °C (Fig. 4d and S31–S34†). Under similar experimental conditions, the samples show negligible uptake capacity (≈5 wt%) for humidified N2 without CO2. Therefore, the actual CO2 uptake content under realistic flue gas conditions is >30 wt% (≈7.0 mmol g−1), in good agreement with the volumetric measurements of CO2 (100% pure) uptake capacities. Our CO2 capture capacities at 75 °C under simulated flue-gas stream conditions are currently the best among the literature values reported for any type of porous solid or solid-amine system (Table S6 and Fig. S35†).4,5,7,8,10,11,13–15,20–33 Here, it is worth noting that the pore volume of our exfGO samples is much higher than those reported in the literature for amine impregnation.
For the first time, we also show that the TETA@exfGO samples exhibit high room temperature flue-gas capture capacities of ≈30 wt% at 30 °C after humidification (Fig. S34†). It is interesting to note that the same samples show reduced CO2 uptake capacities for pure CO2 without humidification (Fig. S36 and S37†).8 Dry and humidified state CO2 interactions in the TETA@exfGO samples after uptake runs are shown by XPS analysis (Fig. S38†). It is worth pointing out here that the exfGO samples alone without amine impregnation show very low CO2 uptake capacities of ≤1 mmol g−1 (Fig. S39†). Direct amine loading on asGO without exfoliation (TETA@asGO) and bulk TETA samples shows a maximum CO2 capture capacity of only 12 wt% and 10 wt% respectively with very slow uptake kinetics at 75 °C under similar flue-gas conditions (Fig. 4d and S40†).
Although the TETA@exfGO samples show exceptional CO2 working capture capacities for both flue-gas CO2 and single-component CO2, the amine-group stability/volatility of the TETA is a concern during repetitive temperature swing cycles (Fig. S31–S33†). This problem can be easily overcome with more stable, higher molecular/branched poly-amines (Table S6†).8,13,14,21–23,28,32,33 Such amines can be more easily stabilized with a strong confinement effect in the large meso-pores of the exfGO consisting of considerable surface oxygen. This is exactly the case with the TEPA (tetraethylenepentamine)@exfGO (Fig. 5 and S41†). For instance, a 6.0 g g−1 TEPA impregnated exfGO-D sample shows a record-high CO2 working capacity of over 28 wt% for flue-gas when investigated using cyclic temperature swing measurements, at 75 °C and 100 °C for uptake and desorption respectively. After 50 cycles of continuous operation for 120 h, the sample still exhibits an impressive working CO2 capture of 22 wt%. Note that such working capacity and cycling stability of TEPA@exfGO are considerably higher compared to any solid-amine system reported to date (Fig. 5b).4,7,8,13–15,21–28,32,33 Here it is worth noting that the capacity loss of ≈21% in our TEPA@exfGO sample is arguably smaller than the ≈40% capacity loss from TEPA@mesoporous silica capsules, under similar synthetic and experimental conditions.14 A notable capacity loss of up to 23% within the first 5–10 uptake cycles is also observed for TEPA@silica samples at an extremely low loading, 1 g g−1, of TEPA (see the references in Table S6†).8 Similarly, TEPA@activated carbon shows a capacity loss of up to 36% within 20 uptake cycles. A capacity loss of 14% is estimated in TEPA@clay, within the first 10 uptake cycles, for a given loading, 1.5 g g−1, of TEPA and at a much reduced desorption temperature of 90 °C.26 Such capacity losses are primarily ascribed to the leaching of amines in successive temperature swing desorption runs at high temperature, around 100 °C.
Fig. 5 (a) Temperature swing cyclic uptake capacity plots of exfGO-D × 6.0TEPA. Uptake and desorption runs (left Y-axis) were carried out at 75 °C and 100 °C (right Y-axis), respectively. (b) Working CO2 uptake capacity and stability against the number of uptake-desorption cycles; comparison is made between TEPA@exfGO and the best literature samples of TEPA@carbon,8 TEPA@silica,14 TEPA@clay,26 and also PEI@silica,13,28 and diamine-appended-MOFs.4,24,25 (c) Volumetric single-component CO2 (100% dry) uptake isotherms of exfGO-D × 6.0TEPA and exfGO-A × 3.0TEPA samples at 75 °C. |
We also note that the reversible CO2 capacity of our TEPA@exfGO is comparable or superior to high molecular weight poly- and/or branched amines, e.g., PEI (polyethylenimine)@silica samples at similar working temperatures (Fig. 5b). In amine@silica systems, most of the amine is grafted on the surface of the particles. Due to their relatively small pore sizes and pore volumes of ≤1.0 cm3 g−1, they invariably show unfavourable CO2 capture performances (Table S6†). A similar problem exists for solid-amine@mesoporous carbons. The other main drawback with these carbon based systems is the negligible surface oxygen, required to obtain a strong amine–surface interaction or amine grafting. Therefore, the enhanced working capacity and stability of our solid-amine@exfGO are directly attributed to the well-impregnated amine within the oxygen-rich pores of ultrahigh pore volume and less oxidation due to the rapid dissipation of exothermic heat of reaction (Fig. 5 and S41†).28 Furthermore, a volumetric CO2 uptake capacity of 6.4 mmol g−1 (≈28 wt%) in the exfGO-D × 6.0TEPA sample, measured at 75 °C and 0.15 bar, is in good agreement with the gravimetric flue-gas uptake (Fig. 5b and c). Similar to the TETA@exfGO samples, the pore volume directed TEPA loading and CO2 uptake capacity can be understood from the comparative CO2 uptake isotherms of the exfGO-D × 6.0TEPA and exfGO-A × 3.0TEPA samples. Moreover, as demonstrated earlier, these stable amines impregnated in the highly exfoliated GO based structures in their pre-humidified state can be well-considered for room temperature CO2 scrubbers in on-board and confined living places (Fig. S34†).
Finally, our results show that the graphene based materials can be tuned for high performance energy storage and guest molecular capture/storage via a simplified synthesis route by simply enhancing the pore population in the meso-/macro-porous region (see the summary of the synthesis methods of literature samples in Tables S4–S6†). Most importantly, our solid-amine@exfGO samples can be favourably considered for an efficient column separation in a breakthrough method. These samples can exhibit tuneable volumetric densities, between 0.3 and 0.6 g cm−3, depending on the strength of exfoliation and amine loading to achieve desirable permeation/diffusion selectivities. The long-term cyclic uptake and further high temperature stabilities of solid-amine@exfGO can be easily achieved by choosing more stable polyamines. Graphene networks of ultrahigh meso-/macro-pore volumes can be effectively applied to other specific energy storage and conversion technologies. For example, these highly open-pore networked graphene samples can be ideal substrates for further functionalization.
A controlled sample, asGO-300C (without exfoliation), was also prepared as follows: the sample was placed in a tube furnace at room temperature and heated slowly at a heating rate of 3 °C per minute to prevent exfoliation, and the furnace was cooled immediately after reaching the temperature of 300 °C.
The gravimetric and temperature swing cyclic CO2 uptake tests were performed by TG under a constant gas flow at 100 ml min−1 around 1 bar. The tests were carried out under 3 different conditions of the test gas: (1) humidified (bubbled through a water bubbler) 15% CO2 balanced with 85% N2, (2) humidified 100% N2, and (3) 15% dry CO2, as a reference. In each case the CO2 desorption cycle was obtained with a dry N2 flow (100 ml min−1) at ≈100 °C. The sorption tests were conducted at different temperatures of 30, 50, 65, and 75 °C. As the experiment involved switching between the gases for each sorption and desorption run, the measurements were carried out in the daytime and left with CO2 overnight at room temperature to start the subsequent cycling the next day.
Note that the volumetric CO2 uptakes are shown in “mmol g−1”, whereas the gravimetric capacities are in “wt%”.
Footnote |
† Electronic supplementary information (ESI) available. See DOI: 10.1039/c7ta05789j |
This journal is © The Royal Society of Chemistry 2017 |