Advances in liquid metals for biomedical applications

Junjie Yan abcd, Yue Lu acde, Guojun Chen acd, Min Yang *b and Zhen Gu *acd
aJoint Department of Biomedical Engineering, University of North Carolina at Chapel Hill and North Carolina State University, Raleigh, North Carolina 27695, USA. E-mail: zgu@email.unc.edu
bMolecular Imaging Center, Key Laboratory of Nuclear Medicine, Ministry of Health, Jiangsu Key Laboratory of Molecular Nuclear Medicine, Jiangsu Institute of Nuclear Medicine, Wuxi 214063, P. R. China. E-mail: yangmin@jsinm.org
cDivision of Molecular Pharmaceutics and Center for Nanotechnology in Drug Delivery, Eshelman School of Pharmacy, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina 27599, USA
dDepartment of Medicine, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina 27599, USA
eNanoSystems Biology Cancer Center, Division of Chemistry and Chemical Engineering, California Institute of Technology, Pasadena, California 91125, USA

Received 29th December 2017

First published on 20th March 2018


To date, liquid metals have been widely applied in many fields such as electronics, mechanical engineering and energy. In the last decade, with a better understanding of the physicochemical properties such as low viscosity, good fluidity, high thermal/electrical conductivity and good biocompatibility, gallium and gallium-based low-melting-point (near or below physiological temperature) alloys have attracted considerable attention in bio-related applications. This tutorial review introduces the common performances of liquid metals, highlights their featured properties, as well as summarizes various state-of-the-art bio-applications involving carriers for drug delivery, molecular imaging, cancer therapy and biomedical devices. Challenges for the clinical translation of liquid metals are also discussed.


image file: c7cs00309a-p1.tif

Junjie Yan

Junjie Yan obtained his PhD degree from the University of Science and Technology of China (USTC) in 2014, in the Department of Polymer Science and Engineering. Then, he joined the Jiangsu Institute of Nuclear Medicine as a research associate. Currently, he is a visiting scholar in the joint Department of Biomedical Engineering at the University of North Carolina (UNC) at Chapel Hill and North Carolina State University, collaborating with Prof. Zhen Gu. His research interests focus on synthetic methodologies and applications of functional materials in molecular imaging, detection and theranostics of cancer and other diseases.

image file: c7cs00309a-p2.tif

Yue Lu

Yue Lu obtained her PhD degree in Biomedical Engineering at the University of North Carolina (UNC) at Chapel Hill and North Carolina State University, under the guidance of Prof. Zhen Gu in the Joint Department of Biomedical Engineering. She is currently a postdoctoral associate working with Prof. James R. Heath at Caltech. Her main research interests include controlled drug delivery, materials science and engineering, as well as nanotechnology towards enhanced cancer treatment. Recently, her research focus has shifted to the development of novel imaging methods and the applications of such methods to fundamental biology and translational medicine.

image file: c7cs00309a-p3.tif

Min Yang

Min Yang obtained her PhD degree at China Pharmaceutical University in 2005, under the guidance of Prof. Guangji Wang in the Department of Pharmacology. She was a visiting scholar working with Prof. Xiaoyuan Chen at the National Institutes of Health (NIH) in 2009 and with Prof. Zhen Cheng at Stanford University in 2013, respectively. She is currently a professor and the director of the Molecular Imaging Center in the Jiangsu Institute of Nuclear medicine. Her center focuses on new drug development, molecular imaging and nanomedicine, mainly for cancer and neuron disease theranostics.

image file: c7cs00309a-p4.tif

Zhen Gu

Zhen Gu obtained his PhD degree at the University of California, Los Angeles, under the guidance of Prof. Yi Tang in the Department of Chemical and Biomolecular Engineering. He was a postdoctoral associate working with Prof. Robert Langer and Prof. Daniel Anderson at MIT and Harvard Medical School. He is currently a Jackson Family Endowed Chair Associate Professor in the Joint Department of Biomedical Engineering at the University of North Carolina (UNC) at Chapel Hill and North Carolina State University. He also holds a joint position in the Eshelman School of Pharmacy and Department of Medicine at UNC. His group studies controlled drug delivery, bio-inspired materials, and nanobiotechnology, especially for cancer and diabetes treatment.



Key learning points

(1) Typical and featured properties of liquid metals.

(2) Strategies for the functionalization of liquid metals.

(3) Usefulness of liquid metal nanomaterials in vitro and in vivo studies.

(4) Biomedical applications of gallium and gallium-based liquid metal alloys.

(5) Key challenges and opportunities for clinical translation.


1. Introduction

From the atomic level, materials are composed of chemical elements, the majority of which (over 90) are metals when we refer to the periodic table of elements. Metals contribute crucially to many aspects of human life, ranging from manufacturing and construction industries to electronics and biomedical engineering, largely owing to their essential mechanical strength, high thermal conductivity and high electrical conductivity.

Liquid metal (LM) is a special family of materials that simultaneously possess both metallic and fluidic properties.1,2 Although there is a term “liquid” in the name of LMs, strictly speaking, they are amorphous solids, which is because the low melting points endow them with fluidic properties to perform as liquids. The pioneering research on LMs was initiated by Dr Duwez and coworkers at Caltech in the 1960s,3 with the generation of gold–silicon non-crystalline alloys upon a rapid quenching process. Compared to conventional metals, LMs are non-crystalline and lack grains, so their atoms are typically arranged in naturally ordered patterns, which provide them with very strong hardness similar to that of glass. Nowadays, LMs refer in particular to metal elements and low-melting-point alloys (LMPAs) that are liquids near or below room temperature. Up to now, there are five known metal elements whose melting points are near or below room temperature, namely mercury (Hg), gallium (Ga), rubidium (Ru), cesium (Cs) and francium (Fr). In the past decade, LM-mediated materials have drawn continuous attraction in diverse areas, such as materials science and engineering (including soft and stretchable electronics, microfluidics and functional composites),1,2 catalysis,4 energy (battery)5 and biomedicine (including reconfigurable medical devices, molecular imaging and phototherapy).6,7 LMs not only retain some metallic characteristics of metals, but also present low viscosity and good ductility.

In the past decades, the unique properties of LMs have been continuously disclosed and subsequently leveraged for various applications. For example, plenty of exciting progress has been achieved, as summarized by previous reviews regarding gallium-based LMs for soft electronics, composites and biomaterials.1,2,8–10 Considering that a comprehensive survey of LMs with high operating temperatures (generally >200 °C) can be found elsewhere,5 here we mainly focus on LMs that are liquids below physiological temperatures (Fig. 1). To promote further fundamental research and practices in this uprising leading edge, this tutorial review targets the basic fundamentals (categories and properties) of LMs and summarizes the recent advances in gallium and LMPA-based biomedical applications. The challenges and opportunities for clinical translation are also discussed.


image file: c7cs00309a-f1.tif
Fig. 1 Schematic of liquid metals (or alloys) and gallium-based LMs for biomedical applications.

2. Gallium and gallium-based alloys

Mercury is one of the most used LMs with a low melting point (mp) of −38.8 °C, and it has been used as thermometers, barometers, electrodes and dental fillings for a long time.9 However, mercury is toxic in any of its forms.11 Particularly, the low vapor pressure of mercury makes it dangerous to handle at room temperature in the absence of a hood. Beyond that, mercury also has large surface tension (>400 mN m−1), which gives it a spherical shape to minimize its surface energy. On the other hand, rubidium, cesium and francium are highly reactive alkali metals with high liability to autoignition in air. They can also interact vigorously with water below room temperature and release hydrogen. Moreover, rubidium, cesium and francium exhibit increasing radioactivity either from themselves or from their isotopes. The radioactive half-life of 87Rb is up to 49-billion years. These safety concerns significantly limit their potential applications, especially in areas involving physiological contact. In contrast, gallium and its eutectic alloys preserve the essential characteristics of LMs, such as great combination of conductivity and transformability, while presenting better biocompatibility.

No free gallium has ever been found in nature. It is mostly refined from a trace amount of various minerals including bauxite, diaspore and sphalerite. Gallium has a melting point of 29.8 °C, a low viscosity and no obvious vapor pressure. However, gallium is extremely sensitive to air and immediately generates a gallium oxide layer when the oxygen content in the surroundings is above the ppm level.9 The thickness of the gallium oxide layer is around 0.7 nm under vacuum conditions, but the environmental excitation and perturbation will help continue the oxidation process and increase the thickness of the oxide layer to about 3 nm. It should be noted that gallium oxide transforms into the form of gallium oxide monohydroxide (GaOOH) in water, which is less passivating and much softer than gallium oxide. Additionally, this oxide layer can reform immediately when it is broken.12 Such self-passivating behavior greatly stabilizes the inside liquid gallium and reduces its surface tension, acting as an “intrinsic” surfactant and facilitating the molding of gallium into various shapes. However, the oxide layer causes gallium to readily adhere to the surfaces of many kinds of materials, such as porcelain, glass and polymer, which sacrifices the fluidic properties of gallium to some extent, and also makes it challenging to quantify the accurate amount of gallium-based LMs. Theoretically, the oxide layer can be depleted via either an acidic condition (pH < 3) or a basic condition (pH > 10) according to the Pourbaix diagram,13 but the corrosion effect is rather restricted. According to the latest development, electrochemical reactions have been alternatively utilized to remove the oxide layer and a low voltage (<1 V) can efficiently decrease the surface tension from >500 mN m−1 to near zero.14

Since it can readily alloy with most metals, gallium has been widely used as an ingredient to prepare LMPAs. To further simplify the operability of gallium at room temperature, it is necessary to tune down the melting point of gallium by alloying with other metals. Currently, the two most used and commercially available gallium-based alloys are eutectic gallium indium (EGaIn, 75% Ga and 25% In by weight)12,15 and gallium indium tin (galinstan, which has several composition ratios, typically 68% Ga, 22% In and 10% Sn by weight),16 with melting points of 15.7 °C and −19 °C, respectively (Table 1). However, alloying does not decrease the oxidization rate of gallium.

Table 1 Physical properties of common liquid metals15,16
Liquid metals Density (g cm−3) Melting point (°C) Boiling point (°C) Viscosity (10−3 Pa s) Vapor pressure (Pa) Surface tension (10−3 N m−1) Thermal conductivity (W m−1 K−1) Electrical conductivity (106 S m−1)
Mercury 13.55 −38.8 356 1.53 1 (42 °C) 487 8.5 1.04
Gallium 6.09 29.8 2205 1.37 ∼10−35 (29.9 °C) 707 29.3 6.73
EGaIn 6.28 15.7 2000 1.99 N/A 624 26.6 3.4
Galinstan 6.44 −19 >1300 2.4 <1.33 × 10−6 (500 °C) 718 16.5 3.46


3. Featured properties

Generally, gallium-based LMs not only retain the basic attributes of conventional metals including high thermal conductivity, high electrical conductivity and excellent mechanical properties, but are also accompanied with typical liquid performances of good fluidity and low viscosity (nearly twice that of water at 20 °C in the case of gallium). Besides, gallium-based LMs also possess several featured properties, such as shape transformability, low toxicity, facile functionalization accessibility, catalytic properties, magnetic properties and self-healing capability. Usually, these combined characteristics can be only realized by hybridizing different kinds of materials. More significantly, more and more advancements have substantiated the promising potential of gallium-based LMs in various scopes of biomedical applications.

3.1 Shape transformation

Despite low viscosity and fluidity, high surface tension renders LMs with a spherical shape to minimize the surface free energy. Therefore, external stimuli are usually demanded to drive macroscopic LMs into a specific morphology. For example, Sheng et al. applied electricity (the voltage is 12 V direct current (DC)) onto galinstan that was immersed in or sprayed with water to achieve several transformation behaviors, such as self-organizing a two-dimensional (2D) sheet into a three-dimensional (3D) sphere, fast integration of different metal droplets into big spheres, self-rotation and planar locomotion.17 This electricity triggered transformation can aid in LM recycling and more importantly provides opportunities for manufacturing a series of biomedical devices, such as soft electronics, microfluidic pumps and artificial robots.

At the nanoscale, nanoparticles with transformable shapes are not only important for the understanding of colloidal self-assembly, but also play essential roles in catalysis and drug delivery. Recently, Lu et al. discovered that EGaIn-based nanodroplets could experience great morphological transformation in aqueous solution upon simple light irradiation (Fig. 2a).18 To this end, sonicating a mixture of EGaIn and graphene quantum dots (GQDs) produced stimuli-responsive LM nanodroplets. The GQD shell increased the water-solubility and stability of LM nanodroplets in aqueous solution and provided LM nanodroplets with an efficient photothermal transduction effect. Moreover, photoirradiation simultaneously produced a high concentration of local reactive oxygen species (ROS), which is a paramount index in photodynamic therapy. Upon light irradiation (635 nm, 100 mW cm−2, 20 min), the initial spherical nanoparticles dispersed into small nanofragments within the first 5 min (Fig. 2b). Then, the random-shaped nanofragments gradually assembled into oval-shaped nanosheets during the following 10 min. Finally, continuous irradiation quickly shifted nanosheets into hollow nanorods, and the energy-dispersive X-ray spectroscopy (EDS) results verified the nanorod GaOOH nature rather than EGaIn, while indium was dealloyed as precipitated nanospheres. Therefore, this series of morphology transformations is a dynamic process, and both heat and ROS are critical to driving the simultaneous phase separation and morphological transformation of the LM segments. The effects of shape transformation of LMs toward drug delivery are discussed in Section 4.1.


image file: c7cs00309a-f2.tif
Fig. 2 Shape transformation of LMs in aqueous solutions. (a) GQDs generate local heating effects and ROS upon light irradiation, followed by the morphological transformation from nanospheres to nanorods. (b) Dynamic changes of the aspect ratio of the GQD coated LMNPs. Gray balls in M2 and M3 represent the remaining nanospheres. (c) Transmission electron microscopy (TEM) images of GQD coated LMNPs. Scale bars: 100 nm. Reprinted with permission from ref. 18. Copyright 2017 American Chemical Society.

3.2 Toxicology profiles

The toxicity of metals and metal compounds is determined by the degree to which they are bioavailable, such as the degree to which they are absorbed through cell membranes, are dispersed within the cells and bound to cellular macromolecules.19 To date, toxicology profiles of gallium-based LMs have not been thoroughly tested and remain elusive. A previous report showed that inhalation of gallium–arsenium (GaAs) particulates was toxic in laboratory animals, inducing liver and kidney dystrophy, as well as pneumosclerosis.20 Generally, gallium and gallium compounds are often considered less toxic. For example, gallium (67Ga, III) citrate injection is an approved scanning formulation by the United States Food and Drug Administration (FDA) for theranostics of lymphoma, bronchogenic carcinoma and Hodgkin's disease, without any obvious side effects.21 Usually, low doses of soluble gallium compounds can be excreted mostly through urine and thus would not arouse acute toxicity. However, the solubility of gallium salts is rather limited and a large dosage utilization will cause insoluble gallium hydroxide precipitation and lead to renal toxicity. Particularly, exposure to gallium chloride has been evidenced to result in acute toxicity symptoms including dermatitis, tachycardia, dyspnea and chest pain.22

Recently, Lu et al. evaluated the in vivo toxicity of EGaIn nanoparticles. The toxicology of LM nanoparticles (LMNPs) was monitored for over three months in a female Balb/c mouse model with an injection dosage of 45 mg kg−1.7 Important liver function markers such as alanine transaminase (ALT), aspartate transaminase (AST) and alkaline phosphatase (ALP) were all within normal levels, indicating no essential toxicity of the EGaIn nanoparticles. Also, no damage to the kidney was detected by the measurement of urea levels in the blood. Meanwhile, blood indexes including erythrocyte, leukocyte, haemoglobin (Hb), mean cell volume (MCV), mean corpuscular haemoglobin (MCH), mean corpuscular haemoglobin concentration (MCHC) and platelet count did not show significant difference to that of the control group. After sacrifice and necropsy of the mice, no abnormalities in normal tissues, including heart, liver, spleen, lungs, kidneys, brain and muscles, were found. In addition, the researchers have investigated the single-dose, acute toxicity of LMNPs and identified its maximum tolerated dose (MTD) to be 700 mg kg−1, suggesting the low toxicity of EGaIn nanoparticles. Lastly, no apparent upregulation of immunoglobulin E (IgE) was observed in LMNP-treated mice, revealing no indication of anaphylaxis.

3.3 Functionalization

Generally, the functionalization of LMs is mainly based on the formation of a metal thiolate complex, which is a classical strategy used in synthesizing transition metal-based nanoparticles.23 Recently, Yu et al. attached 16-mercaptohexadecanoic acid (16-MHDA) as a ligand to the surface of acid-pretreated macroscopic EGaIn, leaving it covered with peripheral carboxyl groups.24 Afterwards, biomolecules such as bovine serum albumin (BSA) and epidermal growth factor receptor (EGFR) antibody (anti-EGFR) were covalently bioconjugated to the outlayer of EGaIn via a standard N-hydroxysuccinimide (NHS)/1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC) procedure. This two-step versatile method well preserved the bioactivities of biomolecules according to the further molecular binary fusion test.

Small-sized LM-based droplets are preferred when applied in bioareas, especially in biomedical applications. Hohman et al. found that sonication could efficiently emulsify LMs and assisted LM particle formation by molecular self-assembly (Fig. 3).25 Once the surface areas of bulk EGaIn increase during the sonication, fast alkanethiolate self-assembly at the interface shields EGaIn from oxidation. Moreover, diverse ligands differ in stabilizing and size controlling of the final EGaIn droplets, and ligands with strong intermolecular forces facilitate particle cleavage to the nanoscaled droplets. Yamaguchi et al. further comprehensively studied the factors to control their size and concluded that temperature, sonication power, sonication time and the concentration of the ligand all have influence on controlling the size of LMNPs (Fig. 3f–j).26 Briefly, low temperature, strong sonication power, long sonication time and high concentration of the ligand generate small-sized gallium nanoparticles. Moreover, this size control process via sonication is reversible in the presence of an acid (hydrochloric acid (HCl)). This sonication approach is also applicable to binary and ternary LM-based alloys.27,28 Meanwhile, Lin et al. discovered that a short sonication period (10 min) without a ligand could also yield EGaIn nanoparticles (∼105 nm in diameter) in ethanol.27 However, either with or without a thiolated ligand, LMNPs synthesized in aqueous solutions were not stable, and quickly precipitated within 10 min after fresh preparation. Recently, both Dickey's and Gu's groups found that the utilization of a surfactant (cetrimonium bromide (CTAB), poly(4-vinyl-1-methylpyridinium bromide), lysozyme and GQD) could greatly stabilize the LMNPs in water.18,29


image file: c7cs00309a-f3.tif
Fig. 3 (a) EGaIn nanodroplets synthesized via sonication in ethanol with 1-dodecanethiol or 3-mercapto-N-nonylpropionamide as ligands. (b) SEM image of 1-dodecanethiol-stablized GaNPs. (c) TEM image of 3-mercapto-N-nonylpropionamide-stabilized GaNPs after filtration (100 nm filter). (d) TEM image of 3-mercapto-N-nonylpropionamide-stabilized GaNPs. (e) TEM image of 1-dodecanethiol-stabilized GaNPs. Reprinted with permission from ref. 25. Copyright 2011, American Chemical Society. (f) Size distribution of GaNPs prepared. Sonication parameter: 40% sonication power, 20 °C, 120 min. (g–j) Factors that control the size of GaNPs. (g) Sonication power. (h) Temperature. (i) Sonication time. (j) 1-Dodecanethiol and HCl. Reprinted with permission from ref. 26. Copyright 2015 Wiley-VCH Verlag GmbH & Co. KGaA. (k) EGaIn in ethanol before sonication and EGaIn suspension after sonication. (m) Size distribution of EGaIn droplets. Thumbnail is a TEM image. Adapted with permission from ref. 27. Copyright 2015 Wiley-VCH Verlag GmbH & Co. KGaA. (n) Preparation of LM droplets via sonication. (p) Size distribution and field emission scanning electron microscopy (FESEM) image of EGaInSn nanodroplets. (q) High resolution transmission electron microscopy (HRTEM) image of EGaInSn nanodroplets. The inset shows the electron diffraction (SAED) pattern of the selected area. Adapted with permission from ref. 28. Copyright 2016 Wiley-VCH Verlag GmbH & Co. KGaA.

3.4 Magnetic properties

None of the LMs have magnetic elements, however, some specific properties often endow LMs with unexpected behavior. For instance, when immersed in an electrolyte solution, LMs could oscillate between the electrodes because of the different surface tensions. On the other hand, some kinds of conventional metal alloys could still preserve magnetic properties when they are in liquid states.30 Therefore, Liu and coworkers constructed a galinstan-based electric motor, which is composed of concentric ring graphite electrodes, electrolyte solution (sodium hydroxide (NaOH), potassium hydroxide (KOH) or HCl), and a magnet.31 The oxide layer of galinstan was removed by the electrolyte and induced to rotate circularly around the electrodes. The rotating speed of galinstan was voltage-dependent and was not closely related to the volumes. The magnetic field effect of galinstan was so strong that even a larger volume required a lower voltage, suggesting that galinstan could be potentially made into motors for medical devices.

3.5 Self-healing capability

Currently, the majority of self-healing materials are polymers. However, self-healing metals frequently demand requirements including high temperature, appropriate pressure and alloy preparation due to their high melting points and poor atom mobility. To circumvent these issues, LMs have attracted extensive interest and have been employed for fabricating self-healing materials. The unavoidable oxide layer not only prevents LMs from leaking out or reflowing into the microchannel, but also keeps the cut interface steady. Upon the cut interfaces merging together, the inside metal bridges probably form and result in subsequent self-healing electrically.32 If the LM microchannels are simultaneously blended with self-healing polymers, the LM-based device could self-heal both mechanically and electrically.33 Currently, self-healing LM-based materials are mainly used as electronics such as wires, antennae and electrodes, which can potentially serve as primary components over conventional metals in many useful biomedical devices. LMs with self-healing capability not only can satisfy the continuous sensing and monitoring requirement in emergent biomedical circumstances, but also can enhance the clinical efficiency and decrease the medical cost.

4. Biomedical applications

Metals have been regarded as one significant category of biomaterials for a long time. Owing to their outstanding mechanical strength, metals are mainly used as implants to replace damaged hard tissues, but their applications as soft implantable/injectable materials are limited. By contrast, LMs offer more possibilities because of their unique properties and have expanded the utilization of metals in many bio-applications (Table 2).
Table 2 LMs for different biomedical applications introduced in this review
Bioapplication Modality/component Liquid metal Physiological contact Ref.
Drug delivery Anticancer delivery EGaIn nanosphere/nanorod Intravenous injection 6, 7 and 18
Molecular imaging X-ray Bulk Ga, EGaIn nanocapsule Targeted vessel injection 6 and 36
CT Bulk Ga Targeted vessel injection 37
Photoacoustic imaging EGaIn nanocapsules Intravenous injection 6
Cancer therapy Chemotherapy EGaIn nanosphere/nanorod/nanocapsule Intravenous injection 7 and 18
Photothermal therapy EGaIn nanocapsules Intravenous injection 6
Electrochemical treatment Bulk EGaIn Tissue injection 40
Medical devices Biosensor Bulk galinstan Skin contact 41 and 42
Microfluidic pumps Bulk EGaIn/galinstan N/A 43–46
Nerve connector Bulk galinstan Tissue contact 47
Implantable electrodes Bulk galinstan Tissue injection 48


4.1 Carriers for drug delivery

Except for the oxide layer, there are no functional groups on the surface of LMs or LMNPs. Moreover, LMNPs alone precipitate quickly in aqueous solutions without the aid of a surfactant, which is typically a thiol-containing molecule. This thiol–LM interaction greatly facilitates the surface functionalization of LMNPs. Briefly, an “emulsion”-like ligand-mediated procedure is simply applied through sonication at room temperature. As a bulk LM is ultrasonically dispersed, the multifunctional thiolated ligands can be readily covalently conjugated onto the surface of LMNPs. Based on this strategy, Lu et al. utilized thiolated (2-hydroxypropyl)-β-cyclodextrin (CD) and thiolated hyaluronic acid (HA) as the drug loading matrix and active targeting moiety, respectively (Fig. 4).7 Of note, unlike conventional “static” metal particles incapable of undergoing dramatic morphological changes, this formulation is transformable and capable of fusing nanoparticles with each other under mildly acidic endosomal/lysosomal microenvironments for the promoted Dox release. When subjected to in vivo evaluation, this nanoplatform demonstrated much better inhibition efficacy towards tumor growth than the free Dox. This probably stems from the HA moiety functionalization, which is one of the main components of the tumor extracellular matrix (ECM) and is capable of binding over-expressed receptor CD44 on solid tumors.34
image file: c7cs00309a-f4.tif
Fig. 4 (a) Schematic design of the transformable EGaIn NPs delivery system. (b) Intracellular delivery of LMNP/Dox-L towards HeLa cells at different times monitored by confocal laser scanning microscopy. The cells were incubated with LMNP/Dox-L at 37 °C for 1 h and 4 h, respectively. The late endosomes and lysosomes were stained with LysoTracker Green, and the nuclei were stained with Hoechst 33342. Scale bar, 10 mm. (c) Ex vivo fluorescence imaging of the tumor and normal tissues from the HeLa tumor-bearing nude mice that were killed at 48 h post injection. The numeric label for each organ is as follows: (1), heart; (2) liver; (3) spleen; (4) lung; (5) kidney; and (6) tumor. (d) Region-of-interest (ROI) analysis of fluorescent signals from the tumors and normal tissues. Error bars indicated s.d. (n = 3). *P < 0.05 (two-tailed Student's t-test). (e) Optical images of the HeLa xenograft tumors of the mice after treatment at day 14. The numeric label for each mouse is as follows: (1) saline; (2) Dox; (3) LM-NP/Dox; and (4) LM-NP/Dox-L. Arrows indicate the sites of tumors. Scale bar, 1 cm. (f) Haematoxylin and eosin (H&E) results of the tumor tissues after treatment. Scale bar: 100 μm. Reprinted with permission from ref. 7. Copyright 2015 Macmillan Publishers Limited.

In a more recent attempt, Lu et al. equipped EGaIn nanoparticles with a photosensitizer GQD to mechanically disrupt cellular structures upon a remote trigger (Fig. 5).18 The zero-dimensional nanospheres transformed into one-dimensional nanorods upon light irradiation, accompanied with a dramatic increase in both the aspect ratio and the volume of the metallic structures. The morphological transformation from nanospheres to hollow nanorods with a remarkable change of aspect ratio can physically disrupt the endosomal membrane to promote the endosomal escape of payloads. The researchers loaded Dox and rhodamine-labeled BSA onto LMNPs via π–π stacking and thiol–metal interaction, respectively. Upon near infrared (NIR) light irradiation, the enhanced endosomal escape of both payloads was verified by both fluorescence quantification analysis and fluorescence colocalization technique. Meanwhile, Chechetka et al. also confirmed the shape-transformable LMNPs as a potential carrier for drug delivery.6 In their research, they encapsulated carmofur within the EGaIn nanocapsules via sonication first. After 30 min of NIR laser irradiation, the structures of EGaIn nanocapsules were totally destroyed and a cumulative concentration of 38 μg mL−1 carmofur was detected. Both examples indicate that LM nanotransformers offer a new means to achieve spatiotemporally controlled intracellular drug delivery.


image file: c7cs00309a-f5.tif
Fig. 5 (a) Preparation of GQDs-coated EGaIn NPs. (b) Light-fueled transformation from nanospheres to nanorods. (c) GQD-coated LMNPs can be used for light-triggered endosomal escape. (i) EGaIn NPs enter the cell via endocytosis. (ii) EGaIn NPs undergo intracellular morphological transformation to physically disrupt endosomal membrane upon light irradiation. (d) Intracellular ROS generation and physically driven Dox release. Scale bar: 20 μm. (e) Promoted intracellular protein delivery. BSA was labeled with rhodamine. Scale bar: 10 μm. (f) EGaIn NPs for live cell imaging. Scale bar: 10 μM. Reprinted with permission from ref. 18. Copyright 2017 American Chemical Society.

4.2 Molecular imaging

Metal compounds have been widely used either as contrast agents (e.g., Gd3+, Mn2+, and iron oxides) for magnetic resonance imaging (MRI), computerized tomography (CT) or as radiotracers (e.g., 67Ga3+, 68Ga3+, 89Zr3+, 64Cu2+, and 99Tc7+) for positron emission tomography (PET). But many of these metal ions would detach from the chelating agents and these free metal ions show various degrees of acute toxicity towards organs such as kidneys, livers and the brain. By contrast, the good biocompatibility and biodegradability of gallium-based LMs present several advantages over these conventional metal compounds. Until now, LMs have been applied to X-ray imaging, CT and photoacoustic imaging (PAI).

X-ray is a good imaging modality for hollow and fluid-filled body structures (angiography). However, its overall effect is often limited without the use of contrast agents, which are typically iodinated compounds. Alternatively, a high energy X-ray source can be applied instead of searching for various chemicals. For instance, Larsson et al. leveraged heated EGaIn-alloy anodes as X-ray sources and a maximum brightness of 24 keV could be obtained by increasing the indium content (65 wt%).35 Compared to the existing sources, EGaIn-based high photo energy not only increases the penetration depth through thick tissues, but also provides high resolution and absorption-contrast imaging in angiography and mammography. Considering the poor imaging quality of iodinated contrast agents in small capillaries, Wang et al. directly injected gallium into the targeted organ (kidney, heart) vessels of pigs due to the low viscosity, good fluidity and high radiographic densities of gallium-based LMs (Fig. 6a and b).36 Remarkably, the imaging resolution can be optimized to visualize small capillaries (∼100 μm) upon increasing the irradiation intensity. Also, a good signal contrast was achieved in the CT scan.37 However, the toxicity and the degradability of a single dose of the macroscopic gallium were not systematically evaluated. On the other hand, LMNPs have also been used for X-ray imaging, but the X-ray signal was much weaker than that of macroscopic gallium due to their lower density in a restricted nanospace. However, assembling LMNPs to large precipitates can restore the enhanced X-ray signal (Fig. 6f).6


image file: c7cs00309a-f6.tif
Fig. 6 (a) X-ray angiograms of hearts filled with gallium (left) and iohexol (right). Plots (down) of the gray scale along the horizontal lines at five different heights. (b) X-ray imaging of a pig kidney perfused with gallium. Reprinted with permission from ref. 36. Copyright 2014 IEEE. (c) CT scan of a mouse perfused with gallium. Reprinted with permission from ref. 37. Copyright 2014 Wang et al. (d) X-ray imaging intensities of EGaIn nanocapsule solutions at various concentrations ((i) 0 mg mL−1; (ii) 0.01 mg mL−1; (iii) 0.1 mg mL−1; (iv) 0.25 mg mL−1; (v) 0.5 mg mL−1; (vi) 1 mg mL−1; and (vii) 1 mg mL−1). Scale bar, 1 cm. (e) Photos (up) and 3D-X-ray images and sectional views (down) of EGaIn nanocapsule-injected rabbit organs (from left to right: brain, heart and eyeball). Orange circles show the injected parts of EGaIn nanocapsules. Red arrows show the laser-irradiated sites. Scale bars, 1 cm. (f) 3D-X-ray image of an EGaIn nanocapsule-injected living mouse. Red arrow indicates the laser-irradiated site. Inlet image shows the magnified view of the laser-irradiated part. Scale bar, 3 cm (inlet: 2 mm). Adapted with permission from ref. 6. Copyright 2017 Macmillan Publishers Limited.

Additionally, Chechetka et al. constructed a multifunctional EGaIn nanocapsule, which was stabilized by a phospholipid layer via photopolymerization (Fig. 7).6 This nanoplatform did not show any characteristic peaks in the UV-Vis-NIR absorption spectrum from 350 to 800 nm, but it exhibited a maximum PA signal at 680 nm and the PA intensity linearly correlated with the concentration of EGaIn nanocapsules. Moreover, further functionalization with EGFR antibody significantly augmented the PA signal (∼22–102%) of EGaIn nanocapsules in the tumors as verified by 3D imaging.


image file: c7cs00309a-f7.tif
Fig. 7 Structure and properties of EGaIn nanocapsules. (a) Chemical components. (b) TEM image. Scale bar, 200 nm. (c) Absorbance spectrum (500 mg mL−1). (d) PA intensity of EGaIn nanocapsules (100 mg mL−1) at different wavelengths. (e) Enhancement of PA intensity in tumor by EGFR antibody-functionalized EGaIn nanocapsules (100 mg mL−1). (f) Ultrasound (US) (grey) and PA (red) imaging of a living mouse injected with antibody-functionalized EGaIn nanocapsules (100 mg mL−1). Ex = 750 nm. Scale bars, 2 mm. (g) 3D imaging of tumor treated by EGFR antibody-functionalized EGaIn nanocapsules (Ex = 750 nm). Blue circle shows the part for construction of the 3D structure. Adapted with permission from ref. 6. Copyright 2017 Macmillan Publishers Limited.

4.3 Enhanced cancer therapy

The most conventional approaches of cancer therapies that are applied in clinics predominantly comprise surgeries, chemotherapy and radiotherapy. Although many lives have been saved and persistent progress has been achieved via these methods, either invasive manners or severe side effects always stump both clinicians and medical researchers. Owing to the remarkable thermal and electrical properties, LMs have also been exploited for medical therapies.

Liquid sodium–potassium (NaK) alloy (m.p. is ∼12.8 °C) was once trialed as a tumor ablation material, which was directly injected into the target site. The mechanism of this method is mainly based on the heat generation between the liquid NaK alloy and water. The local high temperature (ΔT > 80 °C from 0.35 mL NaK) can result in the complete necrosis of the tumor.38 However, alkali metals are highly corrosive and it is difficult to accurately control the local temperature of the irradiated area, which may lead to the dispersive necrotic region volume (usually three times of the injection volume). To overcome this problem, Miyako and coworkers utilized the EGFR antibody functionalized EGaIn nanocapsules for photothermal therapy. The tumor was completely removed at the third day after laser irradiation (Fig. 8).6 This phototherapy effect was much better than that of nanocarbons or gold nanomaterials, which usually take at least one week. Superior to the liquid NaK alloy, the temperature change of EGaIn nanocapsules can be readily controlled by the irradiation power, irradiation distance and the dosage of the formulation. Moreover, this EGaIn-based formulation did not evoke obvious side effects.


image file: c7cs00309a-f8.tif
Fig. 8 (a) Schematic of EGFR antibody-functionalized EGaIn nanocapsules. (b) A mouse with tumors on the left side of its back 7 days after tumor cell transplantation is being irradiated with a 785 nm laser. (c) A mouse after 3 days of treatment (day 10) with anti-EGFR–biotin–avidin–DSPE–PEG2000–amine–DC(8,9)PC–LM and laser irradiation of the tumor on the left side of its back. (d) Relative volumes of tumors on the left side back of the mouse after laser irradiation. All formulations were intratumorally injected and treated with a 785 nm laser. Red, blue, and black arrows represent laser irradiation periods for HEPES dispersions of anti-EGFR–biotin–avidin–DSPE–PEG2000–amine–DC(8,9)PC–LM-, HEPES dispersions of avidin–DSPE–PEG2000–amine–DC(8,9)PC–LM-, and HEPES-injected mice, respectively. aP < 0.005; bP < 0.05; cP < 0.5; dP < 0.0005. Reprinted with permission from ref. 6. Copyright 2017 Macmillan Publishers Limited. (e) Schematic design of EChT with platinum (left) or EGaIn electrodes (right) in vitro. Reprinted with permission from ref. 40. Copyright 2017 Elsevier Ltd.

Electrochemical treatment (EchT) is a historical modality for treating tumors and has recently been considered as a non-invasive, effective and low-cost approach for local tumor therapy.39 In general, DC is imposed when electrodes are placed within the tumors or in the nearby tumoral surroundings, subsequently inducing various biochemical and physiological responses, such as the apoptotic induction, pH and potential changes in tumor, toxic products generation, tumor tissue ionization and immune system stimulation (Fig. 8e). However, hindered by the abnormal shapes of tumors and non-deformability of electrodes, the effective therapy area is often restricted unless multiple electrodes are employed, which complicates the treatment operation and increases patients’ suffering as well. The use of LMs as electrodes greatly overcomes these challenges because of their amorphous nature, which facilitates electrodes to fill complex physiological conditions. According to a recent study, EchT equipped with EGaIn electrodes produced more toxic products during electrochemical reactions and remarkably enhanced the therapeutic efficacy.40

4.4 Medical devices

Conventional metals are widely employed in various components of devices, such as sensors, electrodes, pumps, interconnections, and antennas. Replacing traditional solid-state metals with LMs enables the formation of soft, flexible, and stretchable versions of biomedical devices. The aforementioned properties of LMs are highly appealing for the development of wearable electronics and soft robots.
4.4.1 Biosensors. The largest organ of the human body is the skin, which can generate subtle signals in response to the physiological changes including temperature, strain and pressure. Therefore, one of the ultimate applications for soft electronics is to be utilized on the human skin. Since LMs are low-viscosity liquids under room or body temperature, multiple electrical contacts with tunable geometry could be realized simply by injecting these metals into microchannels that are placed on the substrates of interest. For example, Li et al. fabricated a galinstan-based pulse sensor for heartbeat monitoring via a selective LM plating process (SLMP), which briefly embedded electronic LM circuits within polydimethylsiloxane (PDMS) films (Fig. 9a).41 Futhermore, a wireless pulse measurement system was constructed, which was mainly composed of a LM-based sensor, a bluetooth module, a computer, and a self-programmed visualized software. However, the acid-pretreated galinstan would corrode the chromium (Cr) layer and decrease the patterning resolution, thus a chemically inert copper (Cu) layer coverage is demanded. Recently, Jeong et al. expanded the utilization scope of LMs and developed a galinstan-integrated system for sensing physical motion (Fig. 9j).42 Different from the previous devices, active components such as interconnections, strain sensors and antenna were all made of LMs, which provided more complex deformation and more sensitive sensing of body movements.
image file: c7cs00309a-f9.tif
Fig. 9 Galinstan circuit embedded in the PDMS film with (a) crimping and (b) crumpling deformation, (c) 600 μm-thickness, and (d) stretching deformation. (e) The circuit attached onto the skin of the arm for demonstration. (f) Wireless pulse-measurement system. (g) The wireless pulse-measurement system consists of a galinstan pulse sensor, a bluetooth module, a computer and a self-programmed visualization software program. (h) Measurement results of the human heartbeat when attached to different parts of the human body. Adapted with permission from ref. 41. Copyright 2017 The Royal Society of Chemistry. (j) Schematic of the LM NFC device. (k) Illustration of the operating principle. Power is transferred to, and data are transmitted from the LM antenna via inductive coupling using a commercial NFC reader. (l) Exploded view illustration of the three-layered structure. (m) An optical image of the device. Scale bar, 4 mm. (n) Schematic and motion detection of the index, middle and ring fingers with a simultaneous measurement of skin temperature. Scale bar, 5 cm. Adapted with permission from ref. 42. Copyright 2017 Macmillan Publishers Limited.
4.4.2 Microfluidic pumps. Microfluidics is known as an interdisciplinary field focused on precise control of fluids on the sub-millimeter length scale. The common components include channels, pumps, electrodes, valves, sensors, heaters and coolers. Recently, it has been identified as a significant tool in both fundamental research and clinical applications. Among them, microfluidic pumps are closely related with medical devices. In a recent study, Ga66In20.5Sn13.5 (m.p. ∼ 10.6 °C) was used as a non-contact electrode in an electroosmotic flow (EOF) pump to fill PDMS microchannels.43 When the gap between the pumping channel and the electrode is 20 μm, the velocity of the fluid can reach 5.93 μm s−1 at a voltage of 1.6 V. The employment of LM as microelectrodes realized the miniaturization of microfluidic pumps, and the design of non-contact electrodes prohibited the bubble formation because of reduced heat generation, which presents promise in the liquid delivery of small drugs and biomacromolecules. Tang et al. also built a microfluidic pump with the incorporation galinstan droplets (Fig. 10a).44 Based on the strategy of electrowetting/deelectrowetting on the metal surface, a low power input (<15 mW) readily achieved a high liquid flow rate (>5000 μL min−1). To overcome the situation that most existing micro-pumps demand an external energy as a driving force, Liu's group proposed a self-promoted LM (EGaIn or galinstan) motor in the presence of a small amount of aluminum (Al), which was based on the Rebinder effect.45 Once LMs are exposed to the surface Al2O3 oxide layer, fresh Al will be activated and further initiate the redox reaction (Fig. 10f). Alternatively, modification of the electrolyte surrounding LMs can also be the energy for self-propulsion. For example, Zavabeti et al. found that pH or ionic imbalance around the galinstan surroundings could stimulate both the deformation and Marangoni flow of LM droplets (Fig. 10h).46
image file: c7cs00309a-f10.tif
Fig. 10 (a) Schematic of the experimental setup. (b) Schematic of the galinstan droplet surface charge distribution when placed in the droplet chamber filled with NaOH solution. (c) Schematic of the galinstan droplet surface charge distribution when an electric field is applied between the graphite electrodes. (d) Sequential snapshots for the pumping effect of a galinstan droplet in the PMMA channel filled with NaOH solution (0.3 mol L−1), while a square wave signal (200 Hz, 5 Vp–p, 2.5 V DC offset and 50% duty cycle) is applied between the two graphite electrodes. A droplet of dye is used to demonstrate the pumping effect. Reprinted with permission from ref. 44. Copyright 2014 The National Academy of Sciences. (e) Schematic of forces affecting the velocity of the motor. (f) Spatiotemporal evolution of EGaIn (60 μL) in a Petri dish containing 0.25 mol L−1 NaOH solution. (g) Sequential snapshots of the motion of a 130 μL EGaIn droplet in a U-shaped open-top channel. Reprinted with permission from ref. 45. Copyright 2015 Wiley-VCH Verlag GmbH & Co. KGaA. (h) Droplet moves from the 1.2 mol L−1 HCl to 0.6 mol L−1 NaOH reservoir. Reprinted with permission from ref. 46. Copyright 2016 Macmillan Publishers Limited.
4.4.3 Nerve connector. In the past decade, the use of stem cells to regenerate nerves has been one of the most cutting-edge methods, and autografting has been an effective approach for peripheral nerve repair in clinical practice. However, the high cost and limited donor graft still call for a new strategy for repairing nerve injury. Zhang et al. utilized the Ga67In20.5Sn12.5 alloy to reconnect a severed nerve on a dead bullfrog (Fig. 11).47 Interestingly, as the connection, LM not only physically connected the separate nerves but also revived the electrical signals, which were very close to those generated from the healthy nerves. Furthermore, the in vivo efficacy of LMs in repairing the peripheral nerve damage of the mouse was tested. After the epineuriums were sutured by gallium filling in a silicon rubber, the nerve signals recovered to the normal level of the complete nerves. In addition, the typical atrophy tendency that occurs in nerve-injured mice was significantly reduced in gallium-repaired mice according to the pathological results. These positive results indicate that LM is a promising alternative as the peripheral nerve-repairing and function-healing material.
image file: c7cs00309a-f11.tif
Fig. 11 (a) Schematic of target nerves of a bullfrog. (b) The transected sciatic nerve reconnected by the GaInSn alloy. (c) The X-ray image of a bullfrog after injecting the GaInSn alloy. (d–f) Three kinds of nerve conduits to repair the injured peripheral nerves. (d) Nerve conduit with microchannels. (e) Nerve conduit in the shape of a thin slice. (f) Concentric microchannels. (g) The electric-stimuli signal and the excitement signal from the intact nerve, the transected nerve reconnected by Ga67In20.5Sn12.5 or Ringer's solution (10 mm × 1 mm × 0.5 mm), respectively. (h) The resistance and reactance curves of the LM. Reprinted with permission from ref. 47. Copyright 2014 Zhang et al.
4.4.4 Implantable electrodes. Implantable medical devices (IMDs) are specific equipment for monitoring and treating various physiological conditions. Although these intelligent devices remarkably improve a patient's life, they also bring about several severe concerns including invasive implantation surgery, high cost, discomfort and long-lasting maintenance. Typically, a general implantation surgery involves incision, device implantation and suture, so the resulting fear and pain are inevitable, accompanied by related complications. Liu's group prepared a 3D medical device by first integrating a galinstan (Ga67In20.5Sn12.5) electrode and a packaging material (biodegradable gelatin) and subsequently injecting it inside the biological tissues (Fig. 12).48 This strategy is preferred to conventional surgical principles because of its simplicity, biosafety and minimum invasiveness. In the in vitro experiment, the composite electrode was injected into porcine tissues and it exhibited uniform distribution. The electronic performance test indicated that the employment of LM as a medium would not sacrifice the features of input signals. Electrocardiogram (ECG) signals and the conductivity of the electrical-stimuli signals to the nerves were measured in mouse and frogs, respectively. The results proved the excellent nerve connection property and good electrical properties of LM-based composite electrodes. Additionally, the direct contact area between LMs and the tissue was obviously decreased by the introduction of allied gelatin, which could be possibly replaced with other functional materials.
image file: c7cs00309a-f12.tif
Fig. 12 (a) Injectable Ga67In20.5Sn12.5 electrode within the porcine tissues. (b) The electronic test results for the Ga67In20.5Sn12.5 electrode. (c and d) Injectable Ga67In20.5Sn12.5 electrode utilized within the sciatic nerve of a frog. (e) X-ray image of the frog's leg with injectable Ga67In20.5Sn12.5 electrode. (f) Schematic of ECG measurements on mouse using injectable Ga67In20.5Sn12.5 electrode. (g) Illustration of ECG measurements using injectable Ga67In20.5Sn12.5 electrode upon a 10 Hz electrical stimulation with different magnitudes of 0.6, 1.2 and 1.3 mV. Adapted with permission from ref. 48. Copyright 2013 Macmillan Publishers Limited.

5. Summary and outlook

LMs provide an unparalleled combination of fluidity, deformability and biocompatibility, rendering them promising candidates for a number of biomedical applications, as surveyed above. In addition, LMs could be applied for creating soft robots, which are mainly made of deformable materials that can intelligently transform their shapes in light of environmental conditions. At present, capsule robots have been used in endoscopy for gastric and colonic theranostics.49 Their painless and non-invasive working style significantly improves the efficiency of examination. However, when the tissue is as small as micro-scale vessels, the preparation of such soft robots becomes more difficult. Although hydraulic pressure and magnetic force have been introduced to achieve this goal, these micro-scale vessel robots could not adapt their shapes to different surroundings. Based on the latest findings of LMs such as high velocity, self-fuelled behavior and lasting power, accompanied by their unique properties of both good fluidity and deformability, there should be plenty of room for LMs as potential candidates for vessel robots.

Despite the exciting advances in this field, many challenges still remain for further translation. The biggest concern of leveraging LMs in bio-applications is their controversial toxicity due to their direct or indirect contact with physiological conditions. In general, gallium and gallium-based alloys have presented good biocompatibility based on the existing findings and have been widely used as an alternative to toxic mercury. However, a comprehensive study of acute and chronic toxicities on LMs is still limited, since the human body is a complex environment involving various dynamically biological and chemical changes. In particular when LMs are fabricated into small dimensions with various morphologies, the physiological behavior of LMs may be significantly different. Therefore, further efforts should be devoted to thoroughly evaluate the toxicology of LMs.

In addition, the oxide layer of LMs is a double-edged sword. It has led to the generation of diverse LM-mediated soft electronics, microfluidics and reconfigurable devices. On the other hand, the oxide layer sticks to the surfaces of various materials and increases the viscosity of LMs, which would sacrifice their intrinsic fluidity. Selective thiolation can partially suppress the oxide growth and provide LMs with simultaneous stabilization and additional properties (electrical, magnetic, etc.), but it will not completely stop the fluidic reduction of LMs.50 Additionally, the sticky surface results in the loss of LMs in every utilization via a syringe or a pipette tip, so the quantification of LMs is challenging based on the present technique. Therefore, a simple reagent (anti-oxidative agents, neither acids nor bases) or a strategy that can readily control the existence of the oxide layer is highly urgent.

The melting points of LMs described in this review are all below physiological temperatures, so their mechanical strength can hardly satisfy the requirement for hard biomedical devices. For this purpose, LMs need to alloy with high melting-point metals such as bismuth (Bi) and zinc (Zn), or to form hybrids with engineering polymers. For example, LMPA Bi35In48.6Sn16Zn0.4 (m.p. ∼58.3 °C) has been used as bone cement for bone disease repair. Compared to the two representative calcium phosphate and acrylic cements, Bi35In48.6Sn16Zn0.4 cement exhibits high contrast images due to the metallic properties. More importantly, the reversible phase transition of this alloy cement facilitates both the implantation into the target and the removal from the bone bed. In another recent report, Chen et al. improved the mechanical characteristics of galinstan droplets by surface-coating multilayer polytetrafluoroethylene (PTFE) particles.51 The non-sticky PTFE shell not only enhances the mechanical robustness of LMs, but also retains their intrinsic mobility and elasticity.

Last but not least, cost is always a factor to be considered in the practical use of materials and devices. The price of gallium is about one-third of silver. However, there is no free gallium in nature and it usually needs to be extracted from different minerals, which indirectly increases the cost. Another factor that should not be ignored is the residue LM loss due to their sticky oxide layer. Particularly, the employment of LMNPs will cause the loss of a large portion of pristine LMs in the sonication/centrifugation procedure. Based on these facts, the cost of LMs cannot be reduced significantly. Therefore, the exploration of novel materials with similar properties of LMs is highly desirable in the long run.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was supported by the grants from the Alfred P. Sloan Foundation (Sloan Research Fellowship), the National Key Research and Development Program of China (No. 2017ZX09304021), the National Natural Science Foundation of China (21504034, 31671035), the Jiangsu Provincial Medical Innovation Team (CXTDA2017024) and the Jiangsu Province Foundation (BK20161137, BK20170204, BE2016632).

Notes and references

  1. M. D. Dickey, Adv. Mater., 2017, 29, 1606425 CrossRef PubMed .
  2. N. Kazem, T. Hellebrekers and C. Majidi, Adv. Mater., 2017, 29, 1605985 CrossRef PubMed .
  3. P. Duwez, R. H. Willens and W. Klement, J. Appl. Phys., 1960, 31, 1137 CrossRef CAS .
  4. N. Taccardi, M. Grabau, J. Debuschewitz, M. Distaso, M. Brandl, R. Hock, F. Maier, C. Papp, J. Erhard, C. Neiss, W. Peukert, A. Gorling, H. P. Steinruck and P. Wasserscheid, Nat. Chem., 2017, 9, 862–867 CrossRef CAS PubMed .
  5. H. Kim, D. A. Boysen, J. M. Newhouse, B. L. Spatocco, B. Chung, P. J. Burke, D. J. Bradwell, K. Jiang, A. A. Tomaszowska, K. L. Wang, W. F. Wei, L. A. Ortiz, S. A. Barriga, S. M. Poizeau and D. R. Sadoway, Chem. Rev., 2013, 113, 2075–2099 CrossRef CAS PubMed .
  6. S. A. Chechetka, Y. Yu, X. Zhen, M. Pramanik, K. Y. Pu and E. Miyako, Nat. Commun., 2017, 8, 15432 CrossRef CAS PubMed .
  7. Y. Lu, Q. Y. Hu, Y. L. Lin, D. B. Pacardo, C. Wang, W. J. Sun, F. S. Ligler, M. D. Dickey and Z. Gu, Nat. Commun., 2015, 6, 10066 CrossRef CAS PubMed .
  8. Q. Wang, Y. Yu and J. Liu, Adv. Eng. Mater., 2017, 1700781 CrossRef .
  9. M. D. Dickey, ACS Appl. Mater. Interfaces, 2014, 6, 18369–18379 CAS .
  10. L. T. Yi and J. Liu, Int. Mater. Rev., 2017, 62, 415–440 CrossRef CAS .
  11. S. Bose-O'Reilly, K. M. McCarty, N. Steckling and B. Lettmeier, Curr. Prob. Pediatr. Ad., 2010, 40, 186–215 Search PubMed .
  12. R. C. Chiechi, E. A. Weiss, M. D. Dickey and G. M. Whitesides, Angew. Chem., Int. Ed., 2008, 47, 142–144 CrossRef CAS PubMed .
  13. M. Pourbaix, National Association of Corrosion, Atlas of Electrochemical Equilibria in Aqueous Solutions, Houston, Texas, 1974 Search PubMed .
  14. M. R. Khan, C. B. Eaker, E. F. Bowden and M. D. Dickey, Proc. Natl. Acad. Sci. U. S. A., 2014, 111, 14047–14051 CrossRef CAS PubMed .
  15. M. D. Dickey, R. C. Chiechi, R. J. Larsen, E. A. Weiss, D. A. Weitz and G. M. Whitesides, Adv. Funct. Mater., 2008, 18, 1097–1104 CrossRef CAS .
  16. T. Y. Liu, P. Sen and C. J. C. J. Kim, J. Microelectromech. Syst., 2012, 21, 443–450 CrossRef CAS .
  17. L. Sheng, J. Zhang and J. Liu, Adv. Mater., 2014, 26, 6036–6042 CrossRef CAS PubMed .
  18. Y. Lu, Y. L. Lin, Z. W. Chen, Q. Y. Hu, Y. Liu, S. J. Yu, W. Gao, M. D. Dickey and Z. Gu, Nano Lett., 2017, 17, 2138–2145 CrossRef CAS PubMed .
  19. D. Beyersmann and A. Hartwig, Arch. Toxicol., 2008, 82, 493–512 CrossRef CAS PubMed .
  20. R. J. Lewis and N. I. Sax, Sax's dangerous properties of industrial materials, Van Nostrand Reinhold, New York, 9th edn, 1996, vol. 1–3, p. 1717 Search PubMed .
  21. G. E. Jackson and M. J. Byrne, J. Nucl. Med., 1996, 37, 379–386 CAS .
  22. C. S. Ivanoff, A. E. Ivanoff and T. L. Hottel, Food Chem. Toxicol., 2012, 50, 212–215 CrossRef CAS PubMed .
  23. Y. Negishi, K. Nobusada and T. Tsukuda, J. Am. Chem. Soc., 2005, 127, 5261–5270 CrossRef CAS PubMed .
  24. Y. Yu and E. Miyako, Angew. Chem., Int. Ed., 2017, 56, 13606–13611 CrossRef CAS PubMed .
  25. J. N. Hohman, M. Kim, G. A. Wadsworth, H. R. Bednar, J. Jiang, M. A. LeThai and P. S. Weiss, Nano Lett., 2011, 11, 5104–5110 CrossRef CAS PubMed .
  26. A. Yamaguchi, Y. Mashima and T. Iyoda, Angew. Chem., Int. Ed., 2015, 54, 12809–12813 CrossRef CAS PubMed .
  27. Y. L. Lin, C. Cooper, M. Wang, J. J. Adams, J. Genzer and M. D. Dickey, Small, 2015, 11, 6397–6403 CrossRef CAS PubMed .
  28. L. Ren, J. C. Zhuang, G. Casillas, H. F. Feng, Y. Q. Liu, X. Xu, Y. D. Liu, J. Chen, Y. Du, L. Jiang and S. X. Dou, Adv. Funct. Mater., 2016, 26, 8111–8118 CrossRef CAS .
  29. Y. L. Lin, Y. Liu, J. Genzer and M. D. Dickey, Chem. Sci., 2017, 8, 3832–3837 RSC .
  30. N. H. March and C. M. Sayers, Adv. Phys., 1979, 28, 1–47 CrossRef CAS .
  31. L. Wang and J. Liu, Proc. R. Soc. A, 2015, 471, 20150177 CrossRef .
  32. J. H. So, J. Thelen, A. Qusba, G. J. Hayes, G. Lazzi and M. D. Dickey, Adv. Funct. Mater., 2009, 19, 3632–3637 CrossRef CAS .
  33. E. Palleau, S. Reece, S. C. Desai, M. E. Smith and M. D. Dickey, Adv. Mater., 2013, 25, 1589–1592 CrossRef CAS PubMed .
  34. S. Misra, P. Heldin, V. C. Hascall, N. K. Karamanos, S. S. Skandalis, R. R. Markwald and S. Ghatak, FEBS J., 2011, 278, 1429–1443 CrossRef CAS PubMed .
  35. D. H. Larsson, P. A. C. Takman, U. Lundstrom, A. Burvall and H. M. Hertz, Rev. Sci. Instrum., 2011, 82 Search PubMed .
  36. Q. Wang, Y. Yu, K. Q. Pan and J. Liu, IEEE Trans. Biomed. Eng., 2014, 61, 2161–2166 CrossRef PubMed .
  37. Q. Wang, Y. Yu and J. Liu, 2014, arXiv:1408.0989 0989, arXiv preprint.
  38. W. Rao and J. Liu, Minim. Invasive Ther. Allied Technol., 2009, 18, 30–35 CrossRef PubMed .
  39. H. M. C. Ciria, M. M. Gonzalez, L. O. Zamora, L. E. B. Cabrales, G. V. S. Gonzalez, L. O. de Oliveira, R. Zanella, A. C. Buzaid, O. Parise, L. M. Brito, C. A. A. Teixeira, M. D. Gomes, G. Moreno, V. F. da Veiga, M. Tello and C. Holandino, Chin. J. Cancer Res., 2013, 25, 223–234 Search PubMed .
  40. X. Y. Sun, B. Yuan, W. Rao and J. Liu, Biomaterials, 2017, 146, 156–167 CrossRef CAS PubMed .
  41. G. Y. Li and D. W. Lee, Lab Chip, 2017, 17, 3415–3421 RSC .
  42. Y. R. Jeong, J. Kim, Z. Q. Xie, Y. G. Xue, S. M. Won, G. Lee, S. W. Jin, S. Y. Hong, X. Feng, Y. G. Huang, J. A. Rogers and J. S. Ha, NPG Asia Mater., 2017, 9, e443 CrossRef CAS .
  43. M. Gao and L. Gui, Lab Chip, 2014, 14, 1866–1872 RSC .
  44. S. Y. Tang, K. Khoshmanesh, V. Sivan, P. Petersen, A. P. O'Mullane, D. Abbott, A. Mitchell and K. Kalantar-zadeh, Proc. Natl. Acad. Sci. U. S. A., 2014, 111, 3304–3309 CrossRef CAS PubMed .
  45. J. Zhang, Y. Y. Yao, L. Sheng and J. Liu, Adv. Mater., 2015, 27, 2648–2655 CrossRef CAS PubMed .
  46. A. Zavabeti, T. Daeneke, A. F. Chrimes, A. P. O'Mullane, J. Z. Ou, A. Mitchell, K. Khoshmanesh and K. Kalantar-zadeh, Nat. Commun., 2016, 7, 12402 CrossRef CAS PubMed .
  47. J. Zhang, L. Sheng, C. Jin and J. Liu, 2014, arXiv preprint, arXiv: 1404.5931.
  48. C. Jin, J. Zhang, X. K. Li, X. Y. Yang, J. J. Li and J. Liu, Sci. Rep., 2013, 3, 3442 CrossRef PubMed .
  49. J. Li, B. Esteban-Fernández de Ávila, W. Gao, L. Zhang and J. Wang, Science Robotics, 2017, 2, eaam6431 CrossRef .
  50. Z. J. Farrell and C. Tabor, Langmuir, 2018, 34, 234–240 CrossRef CAS PubMed .
  51. Y. Z. Chen, Z. Liu, D. Y. Zhu, S. Handschuh-Wang, S. Q. Liang, J. B. Yang, T. T. Kong, X. H. Zhou, Y. Z. Liu and X. C. Zhou, Mater. Horiz., 2017, 4, 591–597 RSC .

This journal is © The Royal Society of Chemistry 2018