Mechanical energy harvester based on cashmere fibers

Showcasing a study on the fabrication of a flexible, lightweight and cost-effective mechanical energy harvester based on cashmere fibers by Dr Walid Daoud’s group at City University of Hong Kong.

As featured in:

Mechanical energy harvester based on cashmere fibers Solvent treatment plays a significant role in the surface properties of cashmere, impacting the electric output of the device. The Tween 20-treated cashmere-based harvester can achieve 4.5-fold higher open-circuit voltage than a pristine counterpart, yielding an output power density of 41.7 mW m-2 with good stability over a wide range of relative humidity. See Walid A. Daoud et al., J. Mater. Chem. A, 2018, 6, 11198.

rsc.li/materials-a Registered charity number: 207890

Journal of

Materials Chemistry A View Article Online

PAPER

View Journal | View Issue

Cite this: J. Mater. Chem. A, 2018, 6, 11198

Mechanical energy harvester based on cashmere fibers† *

Lingyun Wang, Xiya Yang and Walid A. Daoud

Published on 04 April 2018. Downloaded on 6/24/2018 4:23:27 AM.

Fabrication of a flexible, lightweight and cost-effective mechanical energy harvester is a promising approach for next-generation wearable electronics. Herein, a triboelectric mechanical energy harvester based on cashmere fibers is developed for the first time using a facile fabrication process with polytetrafluoroethylene as the triboelectric counterpart. The surface properties of cashmere play a significant role in output performance. Remarkable power output can be achieved when subjected to Tween 20 treatment with an open-circuit voltage of 19.5 V, which is 4.5-fold higher than that of a pristine fiber based harvester. As a H-bond donor, the lowered ionization potential of cashmere significantly increases its tendency to lose electrons during contact electrification. The treated cashmere Received 28th January 2018 Accepted 4th April 2018

based harvester can achieve an output power density of 41.7 mW m2 at a load resistance of 14.1 MU. Additionally, the device displays good stability over a wide range of relative humidity. This study demonstrates a new approach for surface modification via mild solvent treatment. Successful

DOI: 10.1039/c8ta00909k

demonstration of a cashmere based harvester as a power source for light-emitting diodes shows

rsc.li/materials-a

potential application in self-powered wearable electronics.

Introduction Portable and wearable electronic devices have gained tremendous attention owing to their promising properties of exibility, light weight, and miniaturization.1 Traditionally, the energy required for power portable devices is supplied by rechargeable batteries. However, the limitations of battery technologies, such as lifetime, size and weight, prevent their practical application in wearable electronics. Harvesting energy from an ambient environment and human body movements has been considered a promising approach to replace rechargeable batteries.2 With the advantages of light weight and simple fabrication, a triboelectric nanogenerator (TENG), as a mechanical energy harvester for selfpowered devices, can convert mechanical energy from the environment into electricity through the coupling of contact electrication and electrostatic induction effects.3–5 For example: a wearable TENG based micromotion sensor could be triggered by eye motion.6 Chandrasekhar et al.7 have devised a self-powered smart puzzle that could drive a liquid crystal display upon one nger pressing. A kirigami patterned paper-based TENG has been demonstrated for a self-charging power package.8 Particularly, a textile based TENG is promising for wearable electronics owing to its unique properties of exibility, durability and cost-effectiveness.9,10 It can either be simply integrated as an add-on or

School of Energy and Environment, City University of Hong Kong, Tat Chee Avenue, Kowloon, Hong Kong. E-mail: [email protected] † Electronic supplementary 10.1039/c8ta00909k

information

(ESI)

11198 | J. Mater. Chem. A, 2018, 6, 11198–11204

available.

See

DOI:

built in portable electronics for applications, such as UV sensors,11 healthcare monitoring,12 motion sensors,13,14 and selfpowered pedometers.15 Previous studies of textile based triboelectric generators focused on surface modication of threads or textiles with the aim of increasing the surface contact area so as to improve the surface charge density using various approaches, such as nanopatterning,11,16,17 surface core–shell nanostructuring,18 and transfer printing.19 Besides, chemical modication was also used to change the tribopolarities of cellulose textiles via sulfate treatment with HNO3/H2SO4 and NaOH/dimethyl reaching an average voltage output of 8 V and a current output of 9 mA.20 Another reported approach was to dissolve silk21 to regenerate a lm structure through electrospinning, with a maximum power output of 4.3 mW m2. However, these processes involve complex procedures and specialized equipment, making them less effective in device fabrication. As a protein ber, cashmere is one of the most abundant naturally occurring biomaterials. It has a stronger tendency to lose electrons in comparison with other bers, such as cotton, polyester and silk according to the triboelectric series. In addition, its naturally inherent rough surface provides a large surface contact area making it an ideal triboelectric material. Herein, for the rst time, cashmere fabric is employed to fabricate a cashmere based mechanical energy harvester using a facile process by adopting a mild one-step solvent treatment. Typically, textiles are wet processed prior to utilization. The treatment medium plays a signicant role in the surface charges of textiles due to the electric “Double-layer” effect.22 It

This journal is © The Royal Society of Chemistry 2018

View Article Online


Paper

was reported that solvents with intermediate dielectric constants may likely result in enhancing electrostatic interactions and other polar interactions, such as hydrogen bonding and formation of ion pairs, when encountered with solid surfaces.23 So far, researchers have utilized this effect in solid– liquid triboelectric generators.24,25 However, to the best of our knowledge, there has been no report on the effect of electrostatic interaction between the textile as the solid surface and treatment solvent on the textile surface properties and its impact on the contact electrication output. Initially, an enhanced voltage output was unexpectedly observed in the cashmere based TENG aer treatment with water in comparison with pristine. Given that cashmere has a hydrophobic surface, it cannot fully interact with water. Thus, a non-ionic detergent Kieralon F-OLB, frequently used as an emulsifying wetting agent and detergent, was employed resulting in a much higher electric output. This triggered us to investigate the effect of the properties of the solvents on the contact electrication of the cashmere based TENG. Thus, in this communication, we investigated the effect of treatment solvents: ethanol, deionized (DI) water, poly(sodium 4-styrenesulfonate) (PSS), and Tween 20 (T20) (Fig. 1a), broadly classied into two groups (polar and anionic/non-ionic) representing a wide-ranging scope with regard to polarity, on the

Journal of Materials Chemistry A

electrication output. The cashmere TENG, constructed of cashmere fabric and polytetrauoroethylene (PTFE) as the triboelectric pair, was fabricated without prior micro/nano patterning surface modication or physical spacers. Strikingly, it was found that a 4.5-fold higher output voltage could be achieved using the T20 treated cashmere TENG as compared to a pristine TENG. Meanwhile, in comparison with a reported silk based TENG,21 an almost 10-fold higher maximum power output (41.7 mW m2) was achieved. Furthermore, an interesting nding is that the relative humidity of the environment, which conventionally has a signicant effect on the electrication output, had no impact on the cashmere based generator. In addition, the generator output was further investigated under resistive and capacitive loads. The use of cashmere TENG as the power source for light-emitting diodes (LEDs) was also investigated.

Results and discussion The structure and optical images (side, front, and back views) of the cashmere TENG with its charge generation mechanism is schematically illustrated in Fig. 1c–e. The dimension of the device is 3 2.5 cm with a 2.30 mm thickness (Fig. 1d). In Fig. 1e, the continuous external mechanical force was applied

(a) Molecular structures of PSS and T20. (b) Schematic diagram of cashmere in contact with PSS and T20. (c) Schematic diagram of the structure of the cashmere triboelectric generator. (d) Digital photos of the cashmere TENG showing the device thickness in front and back views. (e) Schematic illustration of the charge generation process in a press–release cycle. (f) Output open-circuit voltage signal.

Fig. 1


J. Mater. Chem. A, 2018, 6, 11198–11204 | 11199

View Article Online



on the TENG via a linear motor in contact-separation mode. Based on contact electrication, triboelectric charges were generated between cashmere and PTFE surfaces when the TENG is fully pressed, resulting in positively charged cashmere fabric and negatively charged PTFE. Due to the multiple airgaps between the cashmere bers, the bulk of the cashmere fabric has contact electrication with air in the fully pressed state and thus results in the generation of a small amount of negative charges that are distributed in the fabric. This leads to a 2 V potential difference between the two electrodes, which is demonstrated as a “stair” signal in Fig. 1f. Once the mechanical force was released, the fabric regained its original shape due to its resilience and numerous air gaps. A strong dipole moment was then formed accompanied by a large potential difference between the two electrodes.17 Thus, a peak voltage was observed as denoted in Fig. 1f. As a result, the electrons started to transfer from the top electrode through the external circuit to the bottom electrode in order to maintain an electrostatic charge equilibrium, generating a negative current signal. As soon as the cashmere fabric fully returned to its original state, the opposite charges on the electrodes reached an equilibrium state and the potential difference decreased to zero. Once the generator was re-pressed, the electrons on the bottom electrode owed back to the top electrode to neutralize the positive charges, resulting in a positive current. Thus, upon a periodic application of mechanical force, the cashmere based generator produced an alternating electric current, as shown in Fig. S1.† Before device fabrication, the cashmere fabric was treated with ethanol, DI water, PSS and T20, under ultrasonication for 15 min, and the treated cashmere is denoted as Ethanol-T, Water-T, PSS-T, and T20-T, respectively. As a keratinous ber, cashmere is composed of various amino acids with disulde cross-links from cysteine residues. The strong disulde bonds in cashmere contribute to its high elasticity, which enables cashmere to regain its shape aer being stretched out or compressed. The FTIR spectra (Fig. 2) exhibited the characteristic absorption bands of cashmere, which are attributed to the

Fig. 2 (a) FTIR spectra of cashmere before after solvent treatment. (b) XPS full scan spectrum of T20-T. (c) SEM image of T20-T. (d) EDX analysis of T20-T.

11200 | J. Mater. Chem. A, 2018, 6, 11198–11204

Paper

peptide bonds (–NH–CO–) and are labelled amide A, amide I, amide II, and amide III. An absorption band at 3284 cm1 is attributed to the N–H stretching band, while the absorption bands at 1637 cm1, 1520 cm1, and 1238–1240 cm1 are the characteristic peaks of amides, corresponding to C]O (amide I), C–N stretching and N–H bending vibrations (amide II), and C–N and C–O stretching vibrations (amide III), respectively.26,27 The silicone backbone absorption bands (Si–O–Si) appeared between 1130 and 1000 cm1. No residues of the treatment media could be found as veried by the FTIR spectra of the solvents (Fig. S2†). Noticeably, the intensity of the N–H band of Ethanol-T was stronger than those of others. This is because ethanol can denature protein by disrupting the hydrogen bonds of amide groups,28 which increases the intensity of N–H and C]O bands. The results are in accordance with previous studies.29 T20 is a mild detergent and the hydrogen bonds formed between T20 and amide groups brought a slightly increased intensity of peptide bonds. Additionally, it was reported that the application of mechanical stress on polymeric materials results in changes in infrared band intensity.29 The acoustic waves generated by ultrasonication may have induced conformational changes, leading to the increased intensity of the N–H band. In contrast, as an anionic surfactant, PSS disrupted noncovalent bonds within and between the ber proteins,30 which could be responsible for the decrease of N–H and C]O band intensity. Therefore, the formation of new bonds or disruption of original bonds affected the surface properties of cashmere. X-ray Photoelectron Spectroscopy (XPS) spectra (Fig. 2b and S3†) and Energy Dispersive X-ray spectroscopy (EDX) (Fig. 2d and S4†) showed no difference in elemental composition between pristine and solvent treated cashmere, where the predominant elements were C, O, N and S, while the presence of Si element indicated the presence of silicone on the bers. The surface morphology before and aer treatment was characterized by SEM (Fig. 2c and S5†). The characteristic scale morphology of cashmere could be observed on the treated samples as well as pristine cashmere. The coarse morphology of the cashmere bers was also observed in the SEM images. Besides, the mechanical properties of the cashmere fabrics before and aer solvent treatment were investigated (Table S1†). The change of tensile strength aer solvent treatment was within 8%, indicating that the solvent treatment has no signicant impact on the mechanical properties of cashmere. Subsequently, the treated cashmere fabrics were used to fabricate the corresponding triboelectric nanogenerators, namely Ethanol-T TENG, Water-T TENG, PSS-T TENG, and T20T TENG, and their output was compared with that of the generator based on pristine cashmere (pristine TENG). The output open circuit voltage (Voc) (Fig. 3a) and short circuit current (Isc) (Fig. 3b) revealed that treatment conditions play a signicant role in the output of cashmere TENGs. A summary of average Voc and Isc is shown in Fig. 3c (error bars were derived from ve cycles of peak voltage/current). While the pristine TENG recorded an average peak of 4.2 V, 4.0, 5.6, 11.5, and 19.5 V were recorded aer treatment with ethanol, water, PSS and T20, respectively. Meanwhile, the average Isc yielded by


View Article Online


Paper


Fig. 3 Output of cashmere TENGs fabricated after treatment. (a) Open circuit voltage (Voc). (b) Short-circuit current (Isc). (c) Summary of Voc and Isc. (d) Effect of frequency on Voc of T20-T TENG. (e) One cycle of Isc signal of T20-T TENG with the calculated surface charge.

Ethanol-T, Water-T, PSS-T and T20-T TENG was 0.19, 0.22, 0.5, and 0.53 mA, respectively, compared to 0.1 mA for the pristine TENG. In addition, polydimethylsiloxane (PDMS, thickness of 1 mm), a frequently used negative triboelectric material, was employed in place of PTFE as a counterpart for cashmere TENGs. The output voltage and current shown in Fig. S6,† demonstrate the similar trend and magnitude compared with the PTFE based device. This further veries the impact of solvent treatment of cashmere on the output performance of the TENG. To date, the fundamental mechanism of the triboelectric effect is still under debate.31 Nevertheless, the following reasons are proposed to explain the different responses of the TENGs. First, for pristine cashmere bers, surface impurities, such as grease, affect the electron affinity of cashmere and result in reducing its ability to lose electrons. Hence, the pristine TENG displayed a lower output Voc. Second, the higher the dielectric constant, the higher the dipole moment a solvent has. DI water, a polar solvent with a high dielectric constant of 80 (20 C), led to higher electrostatic charges when in contact with the cashmere surface under acoustic waves. While, ethanol being less polar reduced the polarity of water and enhanced the interaction with hydrophobic cashmere, which consequently lowered the output voltage and current. This phenomenon is consistent with previous ndings.24 On the other hand, one of the main characteristics of PSS (anionic) and T20 (non-ionic) is their strong adsorption tendency at the interfaces between water and the surroundings.32 Thus, the interfaces between hydrophobic cashmere and water were dominated by surfactant monolayers resulting in


reduced surface free energy,26 as schematically illustrated in Fig. 1b. Both PSS and T20 have a hydrophilic head and hydrophobic tail, and thus when dissolved in water these molecules form micelles with tails in the interior of the micelle and heads in the exterior. When cashmere fabric was in contact with these molecules, the hydrophobic tails adhered to the cashmere surface leaving the hydrophilic heads free in the solution. The main difference between these two surfactants is that PSS possesses negatively charged heads, so when cashmere was treated with PSS, an electrostatic potential difference was generated as soon as the monolayer was formed, leading to a homogeneously charged interface. This effect can be described through the electrical double layer models.27 In addition, the migration of Na+ ions increased the dipole moment of the monolayer,33 which was further enhanced with the aid of the shear forces of ultrasonication. Accordingly, this increased the electrostatic charges on the cashmere surface. While, T20 is electrically neutral in solution, the –OH groups and the oxygen in the polyether chain (denoted in red in Fig. 1a) formed H-bonds with –NH groups and other oxygen and nitrogen containing groups of cashmere. Particularly, the –OH groups can either serve as a H-bond donor or acceptor, while the oxygen in the polyether chain only serves as a H-bond acceptor. Therefore, as a H-bond acceptor, T20 increased the electron affinity and lowered the ionization potential of cashmere,34 leading to the cashmere surface losing electrons more easily. This explains the much higher Voc of T20-T TENG and is also in agreement with the FTIR results. The increased Isc for PSS-T TENG and T20-T TENG was not as obvious as Voc, implying that the total amount of charges owing between the two electrodes was limited.

J. Mater. Chem. A, 2018, 6, 11198–11204 | 11201

View Article Online



For body motion energy harvesting, the effect of dominant frequencies (#5 Hz) on the output Voc of T20-T TENG was investigated (Fig. 3d). It is noticed that the peak Voc slightly increased with the increasing frequency. Due to the absence of a physical spacer between the contact surfaces, the cashmere and PTFE surfaces are unable to get fully separated from each other during higher frequency deformation of cashmere, resulting in accumulation of un-neutralized charges on the electrodes. Additionally, the amount of transferred charges between T20-T cashmere and PTFE was calculated based on one cycle of Isc (Fig. 3e). The integration of current over time showed a similar trend for both pressing (12.6 nC) and releasing (12.3 nC), indicating that an equal number of electrons transfer in opposite directions. The maximum surface charge density (s) was 16.8 mC m2, which is 9-fold higher than that reported for a silk based TENG (1.86 mC m2).21 Also, the microstructure gap due to air cavities between T20-T cashmere and PTFE surfaces was calculated to be 10 mm, ESI.† To further evaluate the performance of T20-T TENG as a power source, a rectied circuit with external loads was set up (inset in Fig. 4a), where the TENG was connected to an external resistance (100 U to 150 MU) and the output voltage, current and power density were recorded as illustrated in Fig. 4a and b. The output current was found to decrease with increasing load resistance due to ohmic losses, while the voltage showed an opposite trend under the same conditions. Consequently, the power density (W ¼ Vpeak2/R) gradually increased with increasing load resistance reaching a maximum of 41.7 mW

Paper

m2 at an external resistance of 14.1 MU according to the maximum power transfer theorem. Further, at a 5 Hz mechanical frequency and an acceleration/deacceleration of 5 m s2, the output voltage was accumulative, reaching over 70 V, which simultaneously lit up 37 LEDs connected in series (inset in Fig. 4b). The energy conversion efficiency reached 3.8% according to our previously reported calculation method35,36 (more details in the ESI†). Additionally, a range of capacitors was used to investigate the charging behavior of T20-T TENG (Fig. 4c), where it took 120 s for the 1 mF capacitor to reach a saturation voltage of 3.0 V at a frequency of 0.25 Hz. A linear charging behavior was observed with larger capacitance. The stability of T20-T TENG was evaluated by recording the output Voc aer storage under ambient conditions for 24 days (Fig. 4d). The stable output indicates the good durability of T20-T TENG. In order to investigate the effect of the relative humidity (RH) of the environment, being a signicant factor that inuences the generated charges in contact surfaces,37 the normalized output Voc of T20-T TENG and T20-T TENG0 (with a 5 mm-gap between cashmere and PTFE) in a wide range of RH is shown in Fig. 5a. It is noted that the overall output Voc remained relatively stable with minor uctuations regardless of the presence/absence of a gap. The results reveal that T20-T exhibited good resistance to humidity. For comparison, cashmere TENGs subjected to the solvent treatment with a 5 mmgap were also fabricated and employed to investigate the inuence of the RH of the environment. Accordingly, the TENGs are termed Ethanol-T TENG0 , Water-T TENG0 , and PSS-T TENG0 .

Fig. 4 Output of T20-T TENG. (a) Output voltage and current with an external load resistance ranging from 100 U to 150 MU. The inset shows the rectifier circuit for output signals. (b) Output power density when connected with an external load resistance. The inset shows the optical image of 37 LEDs powered by the device at a frequency of 5 Hz. (c) Charging behavior of capacitors of 1 mF to 100 mF. (d) Stability of T20-T TENG.

11202 | J. Mater. Chem. A, 2018, 6, 11198–11204


View Article Online


Paper


ultrasonic bath for 15 min, followed by rinsing with DI water and drying at 60 C. The samples are denoted as Ethanol-T, Water-T, PSS-T and T20-T, respectively. The cashmere triboelectric nanogenerator (TENG), with the size of 3 2.5 cm, consisted of cashmere fabric (thickness 1.8–2.0 mm) as the positive triboelectric material, polytetrauoroethylene (PTFE, thickness 0.3 mm) as the negative counterpart, and aluminum and copper tape as the top and bottom electrodes, respectively (Fig. 1c). The adhesive side of Al tape on the cashmere side had the same electric resistance as its front side, ensuring no impact on charge transfer between the cashmere and Al electrode. Subsequently, plastic tape was used to package the triboelectric pair without spacers. The cashmere nanogenerators aer treatment are termed Ethanol-T TENG, Water-T TENG, PSS-T TENG, and T20-T TENG, respectively. Humidity effect. (a) Comparison of the normalized output Voc of T20-T TENG with and without a 5 mm-gap at various relative humidities. Normalized output Voc of (b) Ethanol-T TENG0 , (c) Water-T TENG0 , and (d) PSS-T TENG0 in a range of relative humidities. Fig. 5

It was observed that when the RH is lower than 45%, the normalized Voc of Ethanol-T TENG0 (Fig. 5b) slightly decreased, while Water-T TENG0 exhibited rst a slight increase and then a decreasing trend (Fig. 5c). However, the Voc decreased sharply in both TENGs when the RH was above 45%, with a decrease of 52% and over 60% for Ethanol-T TENG0 and Water-T TENG0 , respectively, at a RH of 70%. The small uctuation of Voc when the RH is lower than 45% may be attributed to physisorbed OH ions from ambient air moisture, which was proposed in a previous study.38 However, when the moisture in the air was much higher, although cashmere itself is hydrophobic, the presence of a water lm or water islands on these hydrophobic surfaces affected the electron affinity inhibiting the contact electrication.39 Further, when the RH of the testing chamber regained 16%, the output Voc was recovered as indicated in Fig. 5c. For PSS-T TENG0 (Fig. 5d), the output Voc was relatively stable regardless of RH, which is different from Ethanol-T TENG0 and Water-T TENG0 but similar to T20-T TENG0 (Fig. 5a). To date, the effect of water on contact electrication is still inconclusive in the sense that it could either promote or inhibit contact electrication, largely depending on RH and the properties of the contact materials.38 The results implied that the anionic and non-ionic solvent treated cashmere TENGs were not affected by environmental moisture. This could be due to the higher electrostatic charges accumulated on both cashmere surfaces where the ability of charge screening by adsorbed moisture was not strong enough. More work is needed to nd out the possible mechanism.

Experimental section Materials and methods Commercial 100% cashmere fabric was treated before device fabrication with four solvents: ethanol, deionized (DI) water (18.2 MU cm), poly(sodium 4-styrenesulfonate) (PSS, 2 g L1, Sigma-Aldrich) and Tween 20 (T20, 2 g L1, Promega) in an


Characterization Fourier transform infrared (FT-IR) spectra of cashmere fabric before and aer treatment were recorded on a Shimadzu spectrophotometer (IRAffinity-1, Japan) equipped with an attenuated total reectance (ATR) sampling accessory. Each spectrum was recorded in the range of 550–4000 cm1 at a resolution of 4 cm1 for 240 scans. The surface morphology was characterized using scanning electron microscopy (SEM, Carl Zeiss EVO 10, Germany) at an accelerating voltage of 20 kV. Elemental analysis was carried out using Energy Dispersive X-ray spectroscopy (EDX, INCA Energy 200). X-ray photoelectron spectroscopy (XPS, Physical Electronics PHI5802) scans were performed using a Mono Al K-a X-ray source. The mechanical properties of cashmere with and without solvent treatment were tested using material testing equipment (Lloyd LS1, AMETEK, USA) with an elongation speed of 50 mm min1 and a preload of 5 N at room temperature on a sample of size 50 15 mm. TENG output measurements An aluminum chamber equipped with a linear motor (LinMot) was used as the testing station for providing periodical mechanical compression input (force magnitude: 8 N, velocity of 1 m s1 and acceleration/deacceleration of 1 m s2). The relative humidity of the testing chamber was controlled using nitrogen purging/humidier and monitored with a hygrometer placed inside the chamber. The output open-circuit voltage (Voc) and short-circuit current (Isc) were recorded on a Keithley source meter (Keithley 2400) via a full-wave rectier bridge.

Conclusion In summary, we have successfully fabricated cashmere based triboelectric generators. T20-T TENG demonstrated the highest output compared with other solvent counterparts. Furthermore, a stable output of T20-T TENG was achieved with a Voc, Isc and power density of 19.5 V, 0.53 mA and 47.1 mW m2 at a load resistance of 14.1 MU, respectively. The ability to use the cashmere based TENG as a power source was demonstrated through lighting up of 37 LEDs. Interestingly, T20-T TENG shows remarkable resistance to humidity. The cashmere

J. Mater. Chem. A, 2018, 6, 11198–11204 | 11203

View Article Online


mechanical energy harvester is promising for self-powered wearable electronics. Further development should focus on nding effective integration into garments and further improving the output using effective power management.

Conflicts of interest There are no conicts to declare

Acknowledgements


This study was supported by grants from City University of Hong Kong (9667079, 7004005 and 7200299).

References 1 M. Ha, J. Park, Y. Lee and H. Ko, ACS Nano, 2015, 9, 3421– 3427. 2 Z. Wen, M. H. Yeh, H. Y. Guo, J. Wang, Y. L. Zi, W. D. Xu, J. N. Deng, L. Zhu, X. Wang, C. G. Hu, L. P. Zhu, X. H. Sun and Z. L. Wang, Sci. Adv., 2016, 2, e1600097. 3 S. Wang, L. Lin and Z. L. Wang, Nano Lett., 2012, 12, 6339– 6346. 4 F. R. Fan, W. Tang and Z. L. Wang, Adv. Mater., 2016, 28, 4283–4305. 5 A. Proto, M. Penhaker, S. Conforto and M. Schmid, Trends Biotechnol., 2017, 35, 610–624. 6 X. J. Pu, H. Y. Guo, J. Chen, X. Wang, Y. Xi, C. G. Hu and Z. L. Wang, Sci. Adv., 2017, 3, e1700015. 7 A. Chandrasekhar, N. R. Alluri, V. Vivekananthan, J. H. Park and S. J. Kim, ACS Sustainable Chem. Eng., 2017, 5, 7310– 7316. 8 C. S. Wu, X. Wang, L. Lin, H. Y. Guo and Z. L. Wang, ACS Nano, 2016, 10, 4652–4659. 9 S. Lee, W. Ko, Y. Oh, J. Lee, G. Baek, Y. Lee, J. Sohn, S. Cha, J. Kim, J. Park and J. Hong, Nano Energy, 2015, 12, 410–418. 10 X. Pu, W. X. Song, M. M. Liu, C. W. Sun, C. H. Du, C. Y. Jiang, X. Huang, D. C. Zou, W. G. Hu and Z. L. Wang, Adv. Energy Mater., 2016, 6, 1601048. 11 S. Bai, L. Zhang, Q. Xu, Y. B. Zheng, Y. Qin and Z. L. Wang, Nano Energy, 2013, 2, 749–753. 12 W. Zeng, L. Shu, Q. Li, S. Chen, F. Wang and X. M. Tao, Adv. Mater., 2014, 26, 5310–5336. 13 H. J. Sim, C. Choi, S. H. Kim, K. M. Kim, C. J. Lee, Y. T. Kim, X. Lepro, R. H. Baughman and S. J. Kim, Sci. Rep., 2016, 6, 35153. 14 X. S. Zhang, J. Brugger and B. Kim, Nano Energy, 2016, 20, 37–47. 15 A. Chandrasekhar, N. R. Alluri, M. S. P. Sudhakaran, Y. S. Mok and S. J. Kim, Nanoscale, 2017, 9, 9818–9824.

11204 | J. Mater. Chem. A, 2018, 6, 11198–11204

Paper

16 K. N. Kim, J. Chun, J. W. Kim, K. Y. Lee, J. U. Park, S. W. Kim, Z. L. Wang and J. M. Baik, ACS Nano, 2015, 9, 6394–6400. 17 W. Seung, M. K. Gupta, K. Y. Lee, K. S. Shin, J. H. Lee, T. Y. Kim, S. Kim, J. Lin, J. H. Kim and S. W. Kim, ACS Nano, 2015, 9, 3501–3509. 18 C. X. Wu, T. W. Kim, F. S. Li and T. L. Guo, ACS Nano, 2016, 10, 6449–6457. 19 J. Yoon, Y. Jeong, H. Kim, S. Yoo, H. S. Jung, Y. Kim, Y. Hwang, Y. Hyun, W. K. Hong, B. H. Lee, S. H. Choa and H. C. Ko, Nat. Commun., 2016, 7, 11477. 20 F. R. Fan, Z. Q. Tian and Z. L. Wang, Nano Energy, 2012, 1, 328–334. 21 H.-J. Kim, J.-H. Kim, K.-W. Jun, J.-H. Kim, W.-C. Seung, O. H. Kwon, J.-Y. Park, S.-W. Kim and I.-K. Oh, Adv. Energy Mater., 2016, 6, 1502329. 22 J. N. Israelachvili, in Intermolecular and Surface Forces, Academic Press, San Diego, 3rd edn, 2011, pp. 291–340, DOI: 10.1016/B978-0-12-375182-9.10014-4. 23 W. Peyre, V. Lair, V. Andre, G. le Maire, U. Kragh-Hansen, M. le Maire and J. V. Moller, Langmuir, 2005, 21, 8865–8875. 24 Z. H. Lin, G. Cheng, L. Lin, S. Lee and Z. L. Wang, Angew. Chem., Int. Ed., 2013, 52, 12545–12549. 25 H. K. Zhong, Z. Q. Wu, X. Q. Li, W. L. Xu, S. Xu, S. J. Zhang, Z. J. Xu, H. S. Chen and S. S. Lin, Carbon, 2016, 105, 199–204. 26 K. Wang, R. Li, J. H. Ma, Y. K. Jian and J. N. Che, Green Chem., 2016, 18, 476–481. 27 A. Idris, R. Vijayaraghavan, U. A. Rana, A. F. Patti and D. R. MacFarlane, Green Chem., 2014, 16, 2857–2864. 28 N. Booth, Biochem. J., 1930, 24, 1699–1705. 29 W. F. McDonald, H. Goettler and M. W. Urban, Appl. Spectrosc., 1989, 43, 1387–1393. 30 M. Johnson, Mater. Methods, 2013, 3, 163. 31 H. T. Baytekin, B. Baytekin, S. Soh and B. A. Grzybowski, Angew. Chem., Int. Ed., 2011, 50, 6766–6770. 32 K. Bernardino and A. F. de Moura, Langmuir, 2015, 31, 10995–11004. 33 S. Scheiner, Isr. J. Chem., 2017, 57, 385–392. 34 K. Li Han and G.-J. Zhao, Hydrogen Bonding and Transfer in the Excited State, 2010. 35 X. Yang and W. A. Daoud, Adv. Funct. Mater., 2016, 26, 8194– 8201. 36 X. Yang and W. A. Daoud, J. Mater. Chem. A, 2017, 5, 9113– 9121. 37 A. F. Diaz and R. M. Felix-Navarro, J. Electrost., 2004, 62, 277– 290. 38 H. T. Baytekin, B. Baytekin, S. Soh and B. A. Grzybowski, Angew. Chem., Int. Ed., 2011, 50, 6766. 39 L. S. McCarty and G. M. Whitesides, Angew. Chem., Int. Ed., 2008, 47, 2188.
