Functional liquid structures by emulsification of graphene and other two-dimensional nanomaterials

Matthew J. Large*a, Sean P. Ogilviea, Manuela Melonia, Aline Amorim Grafa, Giuseppe Frattaa, Jonathan Salvageb, Alice A. K. Kinga and Alan B. Dalton*a
aUniversity of Sussex, Falmer, Brighton, BN1 9RH, UK. E-mail:;
bImage Analysis Unit, University of Brighton, Brighton, BN2 4GJ, UK

Received 28th July 2017 , Accepted 4th October 2017

First published on 9th January 2018

Pickering emulsions stabilised with nanomaterials provide routes to a range of functional macroscopic assemblies. We demonstrate the formation and properties of water-in-oil emulsions prepared through liquid-phase exfoliation of graphene. Due to the functional nature of the stabiliser, the emulsions exhibit conductivity due to inter-particle tunnelling. We demonstrate a strain sensing application with a large gauge factor of ∼40; the highest reported in a liquid. Our methodology can be applied to other two-dimensional layered materials opening up applications such as energy storage materials, and flexible and printable electronics.

The interesting functional properties of many two-dimensional (2D) materials are often only observed when the particles are in their mono- or few-layer forms. Such properties include the widely reported, and extremely large, electrical and thermal conductivity of graphene,1 for example, or the direct-bandgap semiconductor behaviour of transition metal dichalcogenides (TMDs).2 Strong inter-particle dispersion forces must be overcome in order to exfoliate these nanomaterials, and they also drive aggregation of the isolated particles. Reaggregation has been a long-standing roadblock to utilisation of nanomaterials in macroscopic materials systems on industrially-relevant scales.

Emulsification of nanomaterials is an effective method for formation of macroscopic structures while preserving the degree of particle exfoliation, facilitating the formation of self-assembled liquid structures with functional properties. Pickering emulsions were first described in detail by S.U. Pickering in a 1907 paper3 on oil-in-water emulsions stabilised by precipitated salt nanoparticles. More recently, clays have been used to stabilise oil-in-water emulsions4 as well as polymerisation reactions.5 Such solid-stabilised emulsions are understood to have several advantages over typical surfactant-stabilised emulsions;6 the structures of these two systems are illustrated in Fig. 1a and b. Due to the much greater free energy decrease associated with adsorption of the solid particles at the liquid–liquid interface, Pickering emulsions are extremely stable (both over time and with elevated temperature) and are not susceptible to spontaneous phase separation.3

image file: c7nr05568d-f1.tif
Fig. 1 a: Standard oil-in-water emulsion stabilised by amphiphilic surfactant molecules. b: Pickering oil-in-water emulsion stabilised by solid particles. c: Illustration of the interfacial tension components at the three-phase boundary in a Pickering emulsion system. d and e: Illustration of the formation process for layered nanomaterial-stabilised water-in-oil Pickering emulsions. A dispersion of nanoparticles in the matrix phase is added to the droplet phase to be emulsified; the mixture is subsequently homogenised. f and g: Photograph and optical microscope image of the formed droplets in a water-in-oil emulsion stabilised with pristine graphene.

In this paper, we discuss the use of 2D nanomaterials as stabilisers for water–oil emulsions, where the nanomaterial of choice confers additional system functionality. We demonstrate a sensor where liquid-exfoliated graphene acts as both stabiliser and conductive filler. This approach may be built upon to produce functional structures for a range of other applications, such as energy storage materials for supercapacitors and batteries; water filtration, purification and pumping (by electro-osmosis); wearable, flexible and printable electronics, to name but a few.

The necessary condition for Pickering emulsion formation is that both immiscible phases partially wet the stabiliser surface.6 Equivalently, one must ensure the two spreading coefficients S, for liquids a and b, on the solid surface have the same sign;

image file: c7nr05568d-t1.tif(1)

The three interfacial tensions γa,b (liquid–liquid), γa,s and γb,s (solid–liquid) dictate which phase will form the matrix (more negative S, corresponding to a higher affinity for the stabiliser), and which will form the droplets (greater S). This interfacial force balance is illustrated in Fig. 1c. It should be noted that the surface tension of the stabilising solid must necessarily be between those of the liquids in order to form an emulsion. Conveniently, graphene and similar 2D materials are understood to have surface tensions in between water and water-immiscible liquids.

Fig. 1d and e schematically illustrate the formation of a nanomaterial-stabilised emulsion. To demonstrate the simplicity of the process, we use deionised water and a common household oil (baby oil, which is primarily paraffin or mineral oil) to prepare Pickering emulsions stabilised with layered materials. Initially graphite is ultrasonically exfoliated into the oil phase to yield a dispersion of multilayer graphene nanoparticles with a modal thickness of ∼10 nm (see ESI). This dispersion is then homogenised after addition of water, first by shaking and subsequently by brief ultrasonic mixing. Following coalescence of the initially formed droplets an equilibrium size distribution is rapidly reached by Ostwald ripening. Fig. 1f and g shows a photograph and optical micrograph (respectively) of such an emulsion. These systems remain stable over several months.

To understand the structures in Fig. 1f and g we calculate the spreading coefficients using an interfacial tension approximation;7

image file: c7nr05568d-t2.tif(2)
where γa and γb are the surface tensions of the respective phases. For the liquids the surface tensions are readily characterised experimentally, with γoil = 28.0 mN m−1 and γwater = 72.2 mN m−1. However, for 2D materials there is a larger degree of uncertainty surrounding the correct value;8 for example, the surface tension of graphene is estimated to be in the range of 40 to 50 mN m−1.9–11 Taking the suggested value γgraphene = 40 mN m−1 (ref. 12) with the given values for the liquids we find that Soil = −28 mN m−1 and Swater = −16 mN m−1; this model correctly describes that water forms the droplet phase in the stable Pickering emulsion. Calculations suggest that low surface tension organic liquids are capable of forming similar water-in-oil emulsions; such as n-hexane (γ = 18.8 mN m−1), ethyl acetate (γ = 23.2 mN m−1), and cyclopentanone (γ = 33.3 mN m−1).

The fact that the surface tension of graphene is between those of water and many common water-immiscible liquids is convenient. This is necessary for the formation of stable emulsions, and negates the need for surface modification (by functionalisation or the use of surfactants) to control the surface tension of the nanoparticles. Modifications of this kind are well known to compromise the properties of 2D nanomaterials and their assemblies.

Fig. 2a shows a representative measurement of the droplet diameter distribution, measured from the inset optical micrograph. The data are fitted using a multimodal Gaussian model. The multiple peaks are due to incomplete interfacial coverage of the initial droplets, which have a characteristic diameter di. By coalescence di increases (reducing the surface area-to-volume ratio) and thereby drives the system towards stability.13 The appearance of a multimodal distribution of sizes is therefore indicative that the stabiliser concentration is insufficient to fully stabilise the droplets as they are formed during the homogenisation process.

image file: c7nr05568d-f2.tif
Fig. 2 a: Representative measurement of the emulsion droplet size distribution, with a multimodal Gaussian fit. (Inset) Optical micrograph of emulsion droplets under bright field illumination (the scale bar is 500 μm). b: Photograph of a “liquid wire” formed by encapsulating a graphene emulsion in a silicone pipe. c: Measurements of the mean droplet diameter, and emulsion conductivity, as a function of graphene volume fraction. d: Fractional resistance change against time for a liquid wire system under repetitive manual strain.

Using a geometric argument about the interfacial area of the stabiliser (assuming full coverage of the droplets), it can be shown that;

image file: c7nr05568d-t3.tif(3)
where d [m] is the characteristic droplet diameter; VSρS [g] is the stabiliser mass (where VS and ρS are the volume and density); SSA [m2 g−1] is the stabiliser specific surface area; N is the nominal layer number, and the total droplet volume is VD [mL]. Clearly the droplet diameter decreases with increasing volume fraction (VS/VD) and decreasing layer number of the stabiliser.

We note that if the initially formed droplets (with diameter di) are fully covered by stabiliser then it is not possible for the droplet diameter to decrease further, and any excess stabiliser will remain un-emulsified. Equally, even if there is sufficient stabiliser present in the system the homogenisation procedure will affect the proportion of material that is trapped at the liquid–liquid interface. In the present system the proportion of graphene emulsified is estimated to be at least 90% based on UV-Visible spectroscopy measurements of the “supernatant” immediately after emulsification and sedimentation of the droplets in a representative sample. Using this measurement we estimate that the fractional surface coverage of the droplets is greater than 1 for all samples described by Fig. 2c. This suggests that the droplet surfaces are entirely coated with graphene, and that there is a degree of re-aggregation of the stabiliser during the emulsification process.

This type of system acts as a segregated network when the droplets pack together; due to the functional nature of the stabilising material we observe the emergence of system-scale functional properties. In the graphene-stabilised system, electron tunnelling at particle–particle contacts leads to electrical conduction across and between the droplet surfaces and gives rise to macroscopic conductivity across the system. The inherent conductivity of the graphene-oil dispersion prior to emulsification is negligible (∼40 fS m−1 at 11.3 mg ml−1 graphene, corresponding to 1 vol% at the oil–water ratio used to prepare the emulsions). Based on this we anticipate no significant contribution to the system conductivity as a result of un-emulsified graphene in the oil phase.

Fig. 2b shows a simple application whereby an emulsion is encapsulated in a silicone pipe with electrical end contacts. The mean droplet diameter, as well as the emulsion conductivity in such pipe structures, are characterised as a function of the volume fraction of the stabilising graphene nanoparticles in Fig. 2c. The mean droplet diameters are fitted using eqn (3).

We immediately notice from Fig. 2c that the conductivity is correlated with the droplet diameter and inversely correlated with the graphene volume fraction. This unexpected result contrasts with the typical percolation behaviour observed in composites such as latex-nanocarbon systems.14 This is due to the fact that, in essence, the spheres in the Pickering emulsion system can be considered to have an effective surface conductivity due to near-continuous coverage with the stabilising graphene particles. However, in the latex composite systems the conductive filler is localised exclusively at the interstitial voids between adjacent polymer particles. This means that in our Pickering system the conductivity is dominated by the inter-particle electron tunnelling and the droplet packing. Since the aspect ratio of the stabilising particles is high, a thin oil layer will be maintained between contacting droplets which prevents coalescence (to maintain the three-phase contact at the edges of the stabilising particles). As such electrons experience an additional tunnelling barrier (which is dependent on the dielectric properties of the oil phase) on top of the inter-layer tunnelling barrier present in graphitic multilayers. Since the bulk conductivity in graphite has an anisotropy of ∼1000 between the directions parallel and perpendicular to the basal plane, we anticipate that the system-scale conductivity is limited by inter-droplet contact resistance. Further, the packing fraction of the droplets (which affects the average coordination number; the average number of neighbours each droplet is in contact with) correlates with the overall conductivity. Both the packing fraction and coordination number increase with the polydispersity of the droplet diameter distribution.

Since tunnelling dominates the electron transport, such devices have a large resistance change with strain (in a similar way to elastomer-based composites15,16). Fig. 2d is a plot of relative resistance change ΔR/R0 against time under repeated manual strain for the liquid wire shown in Fig. 2b. The first range shows a strain of approximately 5% with a frequency of 0.7 Hz, followed by a similar strain with a frequency of 5.3 Hz. As the droplets are deformed under strain transfer from the pipe, their surface area-to-volume ratio increases. Increases in the inter-particle separation on the droplet surface lead to the rapid and reversible rise in resistance observed. There will also be a threshold where non-affine reordering of the droplets minimises the system stress. This may be irreversible and will contribute to drift in the measurements. Based on the approximate amplitudes of the strain and resistance change in Fig. 2d, the gauge factor of the system shown is ∼40; the highest reported for a liquid.17 This is comparable to high-sensitivity elastomer-graphene composites previously described, with gauge factors of up to 35.15 The system we describe is liquid-based with a simple preparation procedure, with potential for integration into a range of applications.

It is interesting to note that, at the highest graphene mass fraction studied (0.75 vol% or ∼1.6 wt%), the effective loading of the system relative to the droplet phase is 16 mg ml−1; significantly higher than the maximum stable concentration in many other liquid-phase exfoliation methodologies (in solvents such as NMP Cmax ≈ 1 mg ml−1;18 in surfactant-water systems Cmax ≈ 0.3 mg ml−1 (ref. 19)). We believe that tailored emulsion systems may be used to produce high-loading inks suitable for printing. In such applications it is important to understand and control the droplet size distribution and droplet volume fraction, since these parameters dominate the emulsion rheology.20,21

Micro-emulsification techniques can produce droplets with diameters of order 100 nm. Pickering emulsions can be formed with droplet-matrix ratios up to ∼50% (where droplet sedimentation/buoyancy can achieve higher volume ratios post-preparation). Therefore it is possible to stabilise 20 mg ml−1 of monolayer graphene provided the platelet sizes are small enough to allow conformation to the interface. Typically, liquid exfoliated and size-selected graphene is 3–5 layers thick, which places an upper bound on the mass concentration of 60 to 100 mg ml−1, which is comparable to microfluidisation of graphite.22 Since the surface tension of graphene is similar to graphite,10 we expect our results to be robust to the degree of material exfoliation.

The broad optical absorption of graphene also confers the capability to directly apply energy to the droplets in the form of laser irradiation. This facilitates direct, non-contact manipulation of the droplet contents in microfluidic applications. The ESI Multimedia files show laser-induced heating of graphene-stabilised emulsion droplets using a 405 nm laser, with approximately 4 μW incident radiation per droplet. The localised heat generation forms tight advective flows which drives the droplet movement in the matrix phase, and the emulsion remains stable under continuous illumination under these conditions. It is also possible to control the stability of emulsion droplets directly by applying laser radiation in a similar manner.23

The process used to form these emulsions is not limited to graphene. Fig. 3 demonstrates a range of material stabilisers with different properties. This makes use of the fact that the surface tensions of most layered nanomaterials are demonstrably similar,24 and so there is little variation in the spreading coefficients calculated using eqn (1) and (2). Fig. 3 demonstrates emulsions stabilised with hexagonal boron nitride (h-BN; Fig. 3a), molybdenum disulfide (MoS2, Fig. 3b), tungsten disulfide (WS2, Fig. 3c) and molybdenum diselenide (MoSe2, Fig. 3d). This set of materials, alongside graphene, exhibit conducting, semi-conducting, and insulating behaviours. As such, this system could be used to prepare emulsions with a range of functionalities for applications like high surface area electrode and catalyst structures, functional inks, and microfluidic devices.

image file: c7nr05568d-f3.tif
Fig. 3 Optical micrographs of (a) boron nitride, (b) molybdenum disulfide, (c) tungsten disulfide, and (d) molybdenum diselenide-stabilised emulsion droplets under bright field illumination. The scale bars in all cases are 1 mm.

In summary, we have demonstrated a simple method, using readily available materials, for preparing functional Pickering emulsion structures stabilised by graphene. The influence of composition on the droplet diameter distribution was studied. The droplets are fully coated with graphene particles, and a degree of re-aggregation is inferred from measurements of the surface coverage. We anticipate that the graphene quality could be improved through a careful choice of liquid phases, however the present results show the possibilities of the most simple realisation.

The electrical properties of this system appear to be dominated by contact resistance between adjacent droplets resulting from electron tunnelling. Making use of this effect, we have demonstrated a strain-sensing application for graphene-stabilised Pickering emulsions which has a gauge factor of ∼40. This is comparable to other state-of-the-art graphene-based devices.15

Since the mass concentration in these stabilised systems is very high relative to other liquid stabilisation techniques, Pickering emulsions are an ideal route to producing nanomaterial-carrying inkjet inks. A significant advantage is the ability to control the dispersion viscosity using only the droplet volume fraction;20,21 this eliminates the need for viscosity modifiers such as poly-(ethylene glycol) which hamper the performance of printed structures using conventionally formulated graphene inks.

We have further shown that the available range of functional properties may be expanded using materials such as boron nitride and transition metal dichalcogenides.

Author contributions

ABD, MJL developed the study concepts. MJL, MM, SO performed experiments. AAG, JS performed SEM imaging. MJL, GF performed Raman spectroscopy. MJL, SPO, ABD prepared the manuscript. SPO, AAKK, ABD reviewed the final manuscript.

Conflicts of interest

There are no conflicts to declare.


The authors would like to thank Zenyatta Ventures Ltd for supplying graphite powder. The authors are also grateful to Thomas Swan Ltd for supplying boron nitride powder.


  1. A. K. Geim, Science, 2009, 324, 1530–1534 CrossRef CAS PubMed.
  2. Q. H. Wang, K. Kalantar-Zadeh, A. Kis, J. N. Coleman and M. S. Strano, Nat. Nanotechnol., 2012, 7, 699–712 CrossRef CAS PubMed.
  3. S. U. Pickering, J. Chem. Soc., Trans., 1907, 91, 2001–2021 RSC.
  4. N. P. Ashby and B. P. Binks, Phys. Chem. Chem. Phys., 2000, 2, 5640–5646 RSC.
  5. S. Cauvin, P. J. Colver and S. A. F. Bon, Macromolecules, 2005, 38, 7887–7889 CrossRef CAS.
  6. Y. Chevalier and M.-A. Bolzinger, Colloids Surf., A, 2013, 439, 23–34 CrossRef CAS.
  7. S. Wu, J. Polym. Sci., Part C: Polym. Symp., 1971, 34, 19–30 CrossRef.
  8. A. Ferguson, A. Harvey, I. J. Godwin, S. D. Bergin and J. N. Coleman, 2D Mater., 2016, 4, 015040 CrossRef.
  9. Y. Hernandez, V. Nicolosi, M. Lotya, F. M. Blighe, Z. Sun, S. De, I. T. McGovern, B. Holland, M. Byrne, Y. K. Gun'Ko, J. J. Boland, P. Niraj, G. Duesberg, S. Krishnamurthy, R. Goodhue, J. Hutchison, V. Scardaci, A. C. Ferrari and J. N. Coleman, Nat. Nanotechnol., 2008, 3, 563–568 CrossRef CAS PubMed.
  10. S. Wang, Y. Zhang, N. Abidi and L. Cabrales, Langmuir, 2009, 25, 11078–11081 CrossRef CAS PubMed.
  11. A. Kozbial, Z. Li, C. Conaway, R. McGinley, S. Dhingra, V. Vahdat, F. Zhou, B. D'Urso, H. Liu and L. Li, Langmuir, 2014, 30, 8598–8606 CrossRef CAS PubMed.
  12. Y. Hernandez, M. Lotya, D. Rickard, S. D. Bergin and J. N. Coleman, Langmuir, 2010, 26, 3208–3213 CrossRef CAS PubMed.
  13. T. Wu, H. Wang, B. Jing, F. Liu, P. C. Burns and C. Na, Nat. Commun., 2015, 6, 5929 CrossRef PubMed.
  14. I. Jurewicz, P. Worajittiphon, A. A. K. King, P. J. Sellin, J. L. Keddie and A. B. Dalton, J. Phys. Chem. B, 2011, 115, 6395–6400 CrossRef CAS PubMed.
  15. C. S. Boland, U. Khan, C. Backes, A. O'Neill, J. McCauley, S. Duane, R. Shanker, Y. Liu, I. Jurewicz, A. B. Dalton and J. N. Coleman, ACS Nano, 2014, 8, 8819–8830 CrossRef CAS PubMed.
  16. C. S. Boland, U. Khan, G. Ryan, S. Barwich, R. Charifou, A. Harvey, C. Backes, Z. Li, M. S. Ferreira, M. E. Mobius, R. J. Young and J. N. Coleman, Science, 2016, 354, 1257–1260 CrossRef CAS PubMed.
  17. D. Y. Choi, M. H. Kim, Y. S. Oh, S.-H. Jung, J. H. Jung, H. J. Sung, H. W. Lee and H. M. Lee, ACS Appl. Mater. Interfaces, 2017, 9, 1770–1780 CAS.
  18. U. Khan, A. O'Neill, M. Lotya, S. De and J. N. Coleman, Small, 2010, 6, 864–871 CrossRef CAS PubMed.
  19. M. Lotya, P. J. King, U. Khan, S. De and J. N. Coleman, ACS Nano, 2010, 4, 3155–3162 CrossRef CAS PubMed.
  20. W. H. Cogill, Nature, 1965, 207, 742–743 CrossRef CAS.
  21. R. Pal, J. Rheol., 2001, 45, 509–520 CrossRef CAS.
  22. P. G. Karagiannidis, S. A. Hodge, L. Lombardi, F. Tomarchio, N. Decorde, S. Milana, I. Goykhman, Y. Su, S. V. Mesite, D. N. Johnstone, R. K. Leary, P. A. Midgley, N. M. Pugno, F. Torrisi and A. C. Ferrari, ACS Nano, 2017, 11, 2742–2755 CrossRef CAS PubMed.
  23. M. D. J. Quinn, K. Vu, S. Madden and S. M. Notley, ACS Appl. Mater. Interfaces, 2016, 8, 10609–10616 CAS.
  24. J. N. Coleman, M. Lotya, A. O'Neill, S. D. Bergin, P. J. King, U. Khan, K. Young, A. Gaucher, S. De, R. J. Smith, I. V. Shvets, S. K. Arora, G. Stanton, H.-Y. Kim, K. Lee, G. T. Kim, G. S. Duesberg, T. Hallam, J. J. Boland, J. J. Wang, J. F. Donegan, J. C. Grunlan, G. Moriarty, A. Shmeliov, R. J. Nicholls, J. M. Perkins, E. M. Grieveson, K. Theuwissen, D. W. McComb, P. D. Nellist and V. Nicolosi, Science, 2011, 331, 568–571 CrossRef CAS PubMed.


Electronic supplementary information (ESI) available: Materials, experimental methods, and characterisation of graphene platelets. See DOI: 10.1039/C7NR05568D

This journal is © The Royal Society of Chemistry 2018