Aqueous Stable Ti 3 C 2 MXene Membrane with Fast and Photoswitchable Nanofluidic Transport

Transcription

1 Supporting Information for Aqueous Stable Ti 3 C 2 MXene Membrane with Fast and Photoswitchable Nanofluidic Transport Junchao Lao, Ruijing Lv, Jun Gao, * Aoxuan Wang, Jinsong Wu, Jiayan Luo *,, Key Laboratory for Green Chemical Technology of Ministry of Education, State Key Laboratory of Chemical Engineering, School of Chemical Engineering and Technology, Tianjin University & Collaborative Innovation Center of Chemical Science and Engineering (Tianjin), Tianjin , China Physics of Complex Fluids, University of Twente, Enschede, 7500AE, The Netherlands State Key Laboratory of Advanced Technology for Materials Synthesis and Processing, Wuhan University of Technology, Wuhan , China Key Laboratory for Advanced Energy Materials Chemistry (Ministry of Education), Nankai University, Tianjin , China. * jun.gao@utwente.nl. * jluo@tju.edu.cn. 1

2 Table S1. Quantitative values of ionic conductivities in various situations. Concentration (M) Conductivity (μs/cm) 10-6 M 10-5 M 10-4 M 10-3 M 10-2 M 10-1 M HCl solution ,600 74,900 KCl solution ,240 14,000 NaCl solution ,410 13,200 AlCl 3 solution ,060 3,120 2,9400 Vertical model (KCl) ,010 21,000 Bending (KCl) ,210 12,200 Table S2. The ph values for the different concentrations of HCl, KCl, NaCl and AlCl 3. All electrolyte solutions were used immediately after preparation and stored in sealed bottles throughout experiment so as to minimize the dissolution of ambient CO 2. Concentration (M) ph value 10-6 M 10-5 M 10-4 M 10-3 M 10-2 M 10-1 M HCl solution KCl solution NaCl solution AlCl 3 solution

3 Figure S1. Transmission electron microscope (TEM) characterization of the MXene nanosheets. a) TEM image of Ti 3 C 2 T x nanosheets. b) TEM image of Ti 3 C 2 T x at high magnification. c) Selected-area electron diffraction pattern suggests the hexagonal structure and high crystallinity of the Ti 3 C 2 T x nanosheet. 1 Figure S2. X-ray photoelectron spectroscopy (XPS) characterization of MXene nanosheets. a) XPS survey spectrum and b) high-resolution Ti2p spectrum of MXene, which clearly show the presence of oxygen and fluorine, indicating the existence of -OH and -F surface groups. 3

4 Figure S3. Zeta potential of MXene colloid at various ph, suggesting that MXene nanosheets are negatively charged in the whole ph range from 1 to 14. Origin of Current To confirm that the measured current originates from the ionic solution, a control experiment was carried out. A MXene strip was first fully hydrated in water and then freeze-dried, such that it kept the same structure as the hydrated samples (compare Figure S4a, b). Then the MXene strip was embedded in PDMS matrix, and two reservoirs were carved out at the two ends of the strip. Immediately after filling the reservoirs with electrolyte (1 mm KCl), the ionic current (source voltage: 1 V) was recorded and is shown in Figure S4c (data for 0 min). After that, the electrolyte solution would gradually permeate into the nanochannels and hydrate the membrane. After 20 h, full hydration was reached, and a stable current was obtained. The only difference between the fully hydrated state and non-hydrated state is their origin of current. In fully hydrated state Figure S4b), both the MXene nanosheets and confined electrolyte solution contribute to the conductivity, while in the non-hydrated state (Figure S4a), only MXene nanosheets contribute to the conductivity. The orders of magnitude current difference between fully hydrated state (20 h and 24 h) and non-hydrated state (0 min), and the gradual increase of current during the hydration (from 0 min to 20 h) clearly suggest that the current mainly arises from the electrolyte solution, instead of from the nanosheets. Note that at lower concentrations, the conductivity of the membrane is lower (about one order of magnitude lower at 1 μm), but is still much larger than that of the freeze-dried membrane. Although it is known that usual dry MXene membrane are conductive, they did not contribute significantly to the conductivity possibly because the nanosheets are separated from each other with the hydrated structure. 4

5 Figure S4. Measured current is due to the ionic conductance in electrolyte solution. a, b) The hydrated and freeze-dried sample a) and hydrated sample b) have the same membrane structure, but their sources of conductance are different. The hydrated and freeze-dried sample conducts current through the MXene nanosheets, while the hydrated sample conducts current through both the nanosheets and the electrolyte solutions. c) The hydrated and freeze-dried sample is tested with 10-3 M KCl immediately after device fabrication, which showed negligible current (0 min, the first data point). After that, the membrane will be gradually hydrated, causing the current to increase by more than two orders of magnitude, until a saturation is reached at 20 h. These results confirm that the current is mainly conducted by the electrolyte solution instead of the nanosheets. Figure S5. Independence of membrane conductivity to thickness. a) SEM images of MXene membranes with thicknesses of 7.9 μm (top) and 17.4 μm (bottom), respectively. b) I-V curves for the two devices, showing that the current is approximately proportional to thickness. 5

6 Figure S6. Structural stability of clay membranes in KCl solutions. Clay membrane is not stable in dilue (10-4 M) solutions. Figure S7. Optical images of stable MXene membrane. a) pristine sample. b) MXene membrane soaked in DI water still remained its own structure after one month. 6

7 Figure S8. Ion transport through MXene membrane. (a) for horizontal configuration; (b) for vertical configuration when the membrane is compact and has regular channels; (c) for vertical configuration when the membrane is loosely packed with irregular channels. For the first two cases, the ion transport is dominated by plane surface charge. For the last case, the ion transport should be dominated by edge surface charge. Figure S9. Strong photothermal effect of MXene membrane. A piece of MXene membrane is placed on a polymeric membrane (Celgard 2400). Upon laser irradiation, the polymer beneath the MXene membrane was quickly burnt and thermally deformed. The polymers which is not covered by the MXene, however, is not affected by laser. This result suggests the strong photothermal property of MXene. 7

8 Figure S10. Weight changes of MXene membrane under continuous laser irradiation. a) For comparison, three 2-μm-thick MXene membranes were tested, including a dry one with the laser irradiated area exposed to air (left scheme), a hydrated one with the laser irradiated area exposed to air (middle scheme), and a hydrated one that is completed sealed with PDMS (right scheme). Apparently, according to the photothermal vaporization mechanism, only the hydrated one will lose weight. b) The results are indeed in agreement with the hypothesis. It is noteworthy that the weight loss rate is about 5*10-3 mg/s. The water mass in the irradiated area is about 10-4 mg. This suggests that all water can be vaporized in about 0.2 second, close to the simulation. 8

9 Figure S11. Ionic current of MXene membrane with various thickness upon laser irradiation, which suggests that the laser irradiation is able to completely shut off the ionic flow for a large range of MXene thickness, from 1 μm to 30 μm, as shown by the noisy currents. Figure S12. I-V curves of graphene film in various concentrations of KCl electrolytes. The ionic conductivity was not detected in this device. All currents remain at the same level in various tests. 9

10 Supplementary Note 1. Estimation of Surface Charge. The ionic conductance (G) consists of bulk conductance (G bulk ), which dominates at high concentration and a surface charge governed conductance (G surface ), which dominates at low concentration. G G G ( ) cn ewd / l 2 w/ l (S1) bulk surface cation anion A cation where μ cation and μ anion are the mobilities of cation and anion, respectively; N A the Avogadro s number, c the bulk electrolyte concentration, e the elementary charge, w the width of the membrane, d=0.67 nm the channel size which can be determined by XRD spectrum, l the length of the membrane, and σ the surface charge density of MXene. While G =2G bulk, the G bulk is equal to G surface. From Equation (S1), cation where t= cation ( cation anion ) cn Aed cn Aed = (S2) 2 2t anion cation (S3) In the tested solutions, the concentrations where G bulk =G surface is typically around c 1 mm, while t H(HCl) = 0.85, t K(KCl) = 0.5, t Na(NaCl) = Based on the Equation (S1), Equation (S2), Equation (S3) and above-mentioned data, we estimated the mean surface charge density to be about 0.06 mc/m 2 for these electrolytes. The surface charge density for AlCl 3 is not estimated since it was reported that multivalent cations strongly interact with the nanochannel walls 2, which generally does not affect surface-charge-governed transport behavior but makes the surface charge density calculation highly complicated. Supplementary Note 2. Numerical Simulation of the Photothermal Vaporization A two-dimensional model was established (Scheme S1) to simulate the photothermal vaporization process with Comsol Multiphysics. For simplicity, only the behavior of interlayer water was simulated. Light-to-heat conversion efficiency of MXene was assumed to be 100%. For the 2-μm-thick MXene membrane, the thickness of water h is ~0.8 μm. The total length of the membrane L is μm and the irradiated length d in the middle is 500 μm. Since h<<d<<l, the vaporization time is governed by h, and thereby the quantitative values of d and L are not important. The temperature of water T is determined by, T CP ( k T) Q t where ρ is the density of water, C P the specific heat capacity of water, k the thermal conductivity, and Q e the external heat source. Boundaries 1, 5, and 6 are set to be thermally insulative. Boundary 3 has an inward heat flux Q e of 2 W/cm 2. The heat transfer at boundaries 2, 3, and 4 are set to, 10 e (S4)

11 Q h ( T T) (S5) transfer b where Q is the heat flux, T b the boiling temperature ( K), and h transfer the heat transfer coefficient which is zero for T<T b and infinitely large for T>T b. This guarantees that the water temperature can not support boiling temperature. To address the vaporization, the mesh at boundaries 2, 3, and 4 are set to erode at a rate of v Q / H (S6) mesh where ΔH V the enthalpy of vaporization of water. Neglecting the effect of temperature on water conductivity, the ionic current I during vaporization is calculated to be, h L 0 1 I I0 / ( dx) L 0 h( x) (S7) where I 0 is the current before laser irradiation, h 0 the initial thickness of water (0.8 μm), h(x) the thickness of water at length x during laser irradiation. V Scheme S1. Two-dimensional model to simulate photothermal vaporization process. References 1. Ding, L.; Wei, Y.; Li, L.; Zhang, T.; Wang, H.; Xue, J.; Ding, L. X.; Wang, S.; Caro, J.; Gogotsi, Y. MXene Molecular Sieving Membranes for Highly Efficient Gas Separation. Nat. Commun. 2018, 9, F. H. van der Heyden, D. Stein, K. Besteman, S. G. Lemay, C. Dekker. Charge Inversion at High Ionic Strength Studied by Streaming Currents. Phys. Rev. Lett. 2006, 96,