WO2013029094A1

WO2013029094A1 - High performance graphene oxide electromechanical actuators

Info

Publication number: WO2013029094A1
Application number: PCT/AU2012/001003
Authority: WO
Inventors: Geoffrey William ROGERS; Jefferson Z. LIU
Original assignee: Monash University
Priority date: 2011-08-29
Filing date: 2012-08-29
Publication date: 2013-03-07

Abstract

An actuator comprising graphene oxide that elongates and/or contracts on charge injection. The graphene oxide can have a zig-zag configuration whereby the oxygen atoms of the graphene oxide are aligned along the zig-zag direction of the graphene lattice or an armchair configuration whereby the oxygen atoms of the graphene oxide are aligned along the armchair direction of the graphene lattice. The oxygen atoms can be assembled on only one side of the graphene lattice in a symmetrical configuration or can be assembled on both sides of the graphene lattice in an asymmetrical configuration. The one or more oxygen atoms can be assembled in such positions that the graphene oxide has an interchanging periodicity.

Description

"High performance graphene oxide electromechanical actuators"

Cross-Reference to Related Applications The present application claims priority from AU 2011903456 the content of which is incorporated herein by reference.

The present application relates to actuators, particularly actuators formed at least in part from graphene oxide.

Background

Since the first demonstration of a carbon nanotube (CNT) actuator in 1999, the actuation of covalently bonded carbon-based nanomaterials has attracted much interest. This is not surprising given that these materials have already been shown capable of generating higher stresses and strains than natural muscle and high-modulus ferroelectric materials, respectively. In addition, these materials possess other inherent strengths including the ability to operate under extreme temperature conditions. The chemical exfoliation of bulk graphite has become a popular method for synthesising graphene, due to its potential for economical large-scale production. This process involves the oxidation and reduction of crystalline graphite, which leads to the synthesis of graphene oxide (GO) as a pre-reduction product. Due to this ease of bulk manufacture, and thus its availability, graphene oxide has generated wide interest as to potential applications.

Studies investigating graphene oxide have found that different atomic structures are attainable, which give rise to differing electronic and mechanical properties. In a recent experimental investigation, local graphene oxide periodic structures, representative of the highly-ordered doping of single oxygen (O) atoms onto the hexagonal lattice of pristine graphene, were observed. Interestingly, this study revealed two distinct O-atom doping configurations: so-called clamped and unzipped. Both of these configurations fall within an overarching category, namely symmetrical graphene oxide. In the clamped case, the in-plane lattice constant of the doped graphene was found to be very similar to that of pristine graphene. For this to be possible, it is believed that each dopant O-atom binds to two carbon (C) atoms in the graphene lattice, without rupturing the adjoining C-C bond (see Fig. 1(a)). For the unzipped case, the in-plane lattice constant was much greater than that of pristine graphene, indicating that the C-C bond is ruptured in this case, unzipping the lattice into conjoined graphene nanoribbons (see Fig. 1(b)).

While the oxygen atoms can be aligned along the y-axis to achieve zig-zag alignment (Fig. 1), it is also possible to align the oxygen atoms along the x-axis (perpendicular to the zig-zag alignment). This x-axis alignment of the oxygen atoms is known as an armchair alignment.

An interesting feature of graphene oxide (clamped and unzipped) is that it exhibits a unique structural phenomenon, herein referred to as rippling. As shown in Fig. 1, this prevents the graphene oxide lattice from lying completely flat, in contrast to pristine graphene which relaxes into a near-perfect 2D plane. The extent of the rippling differs considerably between the clamped and unzipped configurations of graphene oxide.

Any discussion of documents, acts, materials, devices, articles or the like which has been included in the present specification is not to be taken as an admission that any or all of these matters form part of the prior art base or were common general knowledge in the field relevant to the present invention as it existed before the priority date of each claim of this application.

Summary

Throughout this specification the word "comprise", or variations such as "comprises" or "comprising", will be understood to imply the inclusion of a stated element, integer or step, or group of elements, integers or steps, but not the exclusion of any other element, integer or step, or group of elements, integers or steps.

In one aspect, there is provided an actuator comprising graphene oxide that elongates and/or contracts on charge injection. In one embodiment, the graphene oxide can have a zig-zag configuration whereby the oxygen atoms of the graphene oxide are aligned along the zig-zag direction of the graphene lattice. In another embodiment, the graphene oxide can have an armchair configuration whereby the oxygen atoms of the graphene oxide are aligned along the armchair direction of the graphene lattice.

The oxygen atoms can be assembled on only one side of the graphene lattice in a

<— symmetrical configuration. In another embodiment, the oxygen atoms can be assembled on both sides of the graphene lattice in an asymmetrical configuration. The one or more oxygen atoms can be assembled in such positions that the graphene oxide has an interchanging periodicity.

In o e embodiment of this aspect, the actuator comprises unzipped graphene oxide. The unzipped graphene oxide can elongate and/or contract on charge injection. The unzipped graphene oxide ban contract on electron injection. In another embodiment, the unzipped graphene oxide can elongate on electron injection. In yet another embodiment, the unzipped graphene oxide can elongate on hole injection.

In one embodiment of the above aspects, the contraction and/or elongation of the graphene oxide can be at least partly or fully reversible.

In this aspect, the contraction and/or elongation can be in the plane of the graphene oxide. In one embodiment, the contraction and/or elongation can increase with increasing charge (electron or hole) injection. By way of example, the contraction can increase up to at least about or exactly 0.25% for a -0.15 e/C-atom charge. By way of another example, the elongation can increase up to about or exactly 3.5% for a 0.15 e/C-atom charge. Different charges can lead to different levels of contraction and elongation.

In another aspect, there is provided an actuator comprising clamped graphene oxide. In this aspect, the actuator can elongate and/or contract upon electron and/or hole injection. In one embodiment of this aspect, the elongation of the graphene oxide can be at least partly or fully reversible. For example, the graphene oxide can contract from its elongated condition on removal of the charge (electron or hole) injection or the graphene oxide can elongate from its contracted condition on removal of the charge (electron or hole) injection. In another embodiment, the elongation or contraction can be irreversible on removal of the charge (electron or hole) injection. In another embodiment, the elongation can be only at least partly reversible once a certain percentage of elongation is exceeded.

In this aspect, the elongation and/or contraction can be in the plane of the graphene oxide.

In this aspect, the elongation and/or contraction can increase with increasing charge (electron or hole) injection. By way of example, the elongation can increase up to about or exactly 28.2% for a 0.15 e/C-atom charge. In one embodiment, if the elongation exceeds about 10%, the elongation can be at least partially irreversible. In another embodiment, if the elongation reaches 28.2%, the elongation can be at least only partially reversible.

In one embodiment, once the elongation has reached a defined level, the input power to the graphene oxide can be removed, resulting in the material only partly contracting from its maximum extent. In one embodiment, the input power is sufficient to modify the graphene oxide from the clamped configuration to the unzipped configuration.

In one embodiment, if the elongation reaches about or exactly 28.2%, upon removal of the charge, the elongation can reverse to about or exactly 23.8%. In another embodiment, the elongation can be fully reversible at or about an elongation of 6.3% or less.

1

The clamped graphene oxide can also elongate with increasing electron injection. In one embodiment and by way of example, the elongation can be up to 0.4% for a -0.15 e/C-atom charge. In other embodiments, the clamped graphene oxide can contract on electron injection. Different oxygen concentrations in the same form of graphene oxide can lead to different response properties of the graphene oxide and hence different forms of actuators formed therefrom. As defined herein, changes in the graphene oxide (clamped or unzipped) can be made dependent on the concentration of the oxygen doping. In one embodiment and by way of example, the clamped graphene oxide can contract if the oxygen doping concentration is such that the graphene-oxygen compound is given by C2O. In other embodiments of all aspects, other oxygen concentrations can be used including C₄O, CsO and Ci₆0. In one embodiment of all aspects, the actuator can comprise a layer of graphene oxide as defined herein or a composite of two or more layers of graphene oxide, with at least one layer being said graphene oxide that elongates and/or contracts on charge injection. In this embodiment, all layers of the composite can be graphene oxide, with at least one layer being said graphene oxide that elongates and/or contracts on charge injection. In another embodiment, the composite can comprise layers of graphene oxide, with at least one layer being said graphene oxide that elongates and/or contracts on charge injection, and one or more other materials. In a further embodiment, all of the graphene. oxide layers of the composite can have the same configuration. In another embodiment, some or all of the graphene oxide layers can have different configurations to that of other graphene oxide layers.

In yet another embodiment, the actuator can comprise one or more narrow strips or nanoribbons of graphene oxide that elongate and/or contract on charge injection. Such nanoribbons could be constructed, and used by cutting relatively narrow strips (or ribbons) from bulk graphene oxide. In one embodiment, the ribbons can have a width as small as a single unit cell of the material (eg. on the order of 0.2mn).

In a still further embodiment, the actuator can comprise graphene oxide that elongates and/or contracts on application of an electric field. Such an actuator can have one, some or all of the features of the actuator defined herein. _r

In yet another embodiment, the actuator can comprise graphene oxide that elongates and/or contracts on application of a magnetic field. Such an actuator can have one, some or all of the features of the actuator defined herein. The magnetic field can be created using electric charge (an electromagnet). In a still further embodiment, changes in temperature can be used alone or in combination with other influencers to generate changes in the response of the formed graphene oxide. In all aspects, the graphene oxide as defined can be useful for micro or nano- electro-mechanical systems (MEMS NEMS) actuators. In another embodiment, the actuator comprising single or multiple layers of graphene oxide as defined can be usable as artificial muscle. In another embodiment, the actuator as defined can comprise or be part of a switch, particularly switches used in relatively long-term, Iow- power switching applications or, for example, a switch used in fibre optic communications. Where the elongation is only at least partially reversible, the actuator can comprise part of a single-use switch. The actuator can be used as a resonator, including broadband resonators. The actuator can also be actuated by an alternating current electrical signal. In other embodiments, the actuator as defined can be used in optical zoom lenses and/or autofocus lenses, including lenses used in mobile telephones, including smartphones, and digital cameras). The actuator as defined can also comprise or be used as part of a robotic actuator including robotic actuators for use in minimally invasive surgery. In other embodiments, the actuator as defined can be used in electromechanical valves and controllable channels for use in micro and/or nanofluidic devices. In yet another embodiment, the actuator as defined can be used in devices such as electromechanical carbon nanotube extruders.

In yet another aspect, there is provided a method of delivering a force comprising actuating an actuator as defined herein.

Brief Description of the Drawings

By way of example only, embodiments are now described with reference to the accompanying drawings, in which:

Fig. 1 depicts symmetrically clamped (a) and unzipped (b) zig-zag graphene oxide configurations. The C4O unit cells of the graphene oxide are depicted by dotted lines. The Figure labelled as 1(c) also depicts asymmetrical, zig-zag, unzipped graphene oxide with the same oxygen concentration as those depicted by Fig. 1(a) and Fig. 1(b). The C- and O-atoms are represented by small-light-grey spheres 11 and large-dark-grey spheres 12, respectively. The corresponding in-plane lattice parameters for the unit cells (along with those of pristine graphene) are as follows:

In-plane lattice parameters (A)

a b

Clamped 4.332 2.545

Unzipped 5.359 2.476

Graphene 4.268 2.464 Fig. 2 depicts actuation of symmetrically clamped and unzipped zig-zag graphene oxide and pristine graphene due to positive (hole) and negative (electron) charge injection. The inset in this drawing is a close-up of the strain responses between -0.1 and 0.05 e/C-atom charge injection. Fig. 3 depicts the total strain response of symmetrically clamped (a) and unzipped (b) zig-zag graphene oxide upon electron injection. Unlike the clamped case, unzipped graphene oxide contracts upon electron injection due to structural rippling modulation, which can be quantified by the C-O-C bond angle, a. In each case the deformed cell is contrasted with the undeformed cell directly above. The excess charge density profiles for clamped (c) and unzipped (d) graphene oxide explain the observed actuation behaviour upon electron injection (regions 13 represent excess electron density). The following table sets out the a-axis resolved interatomic deformations, where the numbers correspond to the respective positions indicated in (a) and (b). The profiles and deformations shown are for a -0.15 e/C-atom electron injection. q-axis deformations (A)

Clamped Unzipped

1 -0.0016 -0.0050

2 0.0082 -0.0085

3 -0.0016 -0.0050

4 0.0122 0.0061

Total 0.0172 -0.0124

Fig. 4 depicts the total electromechanical strain of symmetrically clamped (a) and unzipped (b) zig-zag graphene oxide as a function of the interatomic bond-length and structural rippling changes. The extent of the contribution by each effect to the total clamped and unzipped graphene oxide responses can be seen to be distinctly different.

Fig. 5 depicts the relative total energy of the symmetrical zig-zag graphene oxide unit cell as a function of the in-plane lattice parameter a and injected charge. The metastable clamped and more stable unzipped configurations correspond to lattice parameters of 4.33 and 5.36A, respectively.

Fig. 6 depicts the non-folly reversible elongation of clamped graphene oxide to unzipped graphene oxide.

Fig. 7 depicts the actuation stress as a function of charge injection for symmetrical zig-zag graphene oxide (clamped and unzipped) and pristine monolayer graphene. The cross sectional areas used to calculate the stresses were based on the respective b lattice constants and the out-of-plane separation between layers in multilayer form (3.34 A for graphene oxide and 3.35 A for graphene). Stresses were calculated by fixing the respective unit cells at their charge-neutral equilibrium geometries and injecting the corresponding charges. Fig. 8 depicts electromechanical actuation of graphene oxide due to positive

(hole) and negative (electron) charge injection. The inset is a close-up of the charge- strain results between -0.15 and 0.075 e/C-atom charge injection.

Fig. 9(a) depicts the interatomic and a-axis resolved deformations of CiO-asym- unzip graphene oxide upon electron injection and shows that whilst the QM effect leads to interatomic bond length expansion, an increase in the degree of rippling results in an overall contraction of the unit cell along the a-axis. This is compared with the case of C_tO-sym-unzip graphene oxide depicted in Fig. 9(b). The excess charge density distributions of C^O-asym-unzip (Fig. 9(c)) and C-jO-sym-unzip (Fig. 9(d)) show the origin of the very high contraction of the former upon electron injection. All results correspond to an electron injection level of -0.05 e/C-atom.

Fig. 10 depicts partial DOS (Fig. 10(a) and Fig. 10(b)) of C₄0-asym-unzip graphene oxide for 0 e/C-atom, 0.025 e/C-atom, 0.075 e/C-atom and 0.15 e/C-atom hole injection. The partial DOS are decomposed into both orbital (s and p) and site (C and O atom) projections, where the C atom DOS shown are for the O atom's nearest neighbour. Excess charge density distributions are also shown for 0.025 (Fig. 10(c)), 0.075 (Fig. 10(d)) and 0.15 e/C-atom (Fig. 10(e)) hole injection. The excess charge density isosurfaces in Figs. 10(c) to 10(e) are equivalent (normalised against the injected charge).

Fig. 11 depicts the total strain, C-0 interatomic bond strain, and C-O-C bond angle change as a function of charge injection. All hole injection levels reduce the C-O bond length, while the C-O-C bond angle initially decreases for low (0.025 e C-atom) hole injection and then increases for higher (>0.05 e/C-atom) hole injection.

Fig. 12 is a simplified view of one embodiment of an actuator as defined herein.

Description of Embodiments An actuator (see Fig.12 for an example labelled 100) can be formed from one or more layers of the materials as described in the accompanying drawings.

One form of actuator can comprise unzipped graphene oxide that is formed so as to contract on electron injection. As defined herein, the unzipped graphene oxide can be formed so as to elongate on electron injection. Whether it be contraction or elongation, the change in the unzipped graphene oxide can be at least partly or fully reversible and/or in the plane of the graphene oxide.

The contraction can increase with increasing electron injection. The contraction can increase, for example, up to at least about or exactly 0.25% for a -0.15 e/C-atom charge.

The unzipped graphene oxide in the actuator can also elongate on hole injection. Such elongation can again be in the plane of the graphene oxide. The elongation can increase with increasing hole injection with, for example, the elongation increasing up to about or exactly 3.5% for a 0.15 e/C-atom charge. Different charges can lead to different levels of contraction and elongation.

In another form, the actuator can comprise clamped graphene oxide that elongates on hole injection. As defined herein, the clamped graphene oxide can be formed so as to contract on hole injection. In the case of elongation, the elongation of the graphene oxide is at least partially reversible and the graphene oxide can contract from its elongated condition on removal of the hole injection. In another embodiment, the elongation can be irreversible even on removal of the hole injection. The elongation can be at least partly irreversible once a certain percentage of elongation is exceeded.- As with the other actuator, the elongation can be in the plane of the graphene oxide.

In this other form, the elongation of the clamped graphene oxide can increase with increasing hole injection. The elongation can increase up to about or exactly 28.2% for a 0.15e/C-atom charge. If elongation exceeds about 10% (such as depicted in Fig. 6), the elongation can be at least partly irreversible.

Once the elongation has reached a defined level, the input power to the graphene oxide can be removed, resulting in the material partly contracting from its maximum extent. In one embodiment, if the elongation reaches about or exactly 28.2%, upon removal of the charge, the elongation can reverse to about or exactly 23.8% strain. In another embodiment, the elongation can be fully reversible. The clamped graphene oxide can also elongate with increasing electron injection. In one embodiment, the elongation can be up to 0.4% for a -0.15 e/C-atom charge. The clamped graphene oxide can also contract with increasing electron injection for some oxygen dopant concentrations. Studies have been undertake by the present inventors to determine the properties of the actuators as described hereon.

In a first study, first principle density functional computations of the clamped and unzipped graphene oxide materials were performed using the Vienna ab initio simulation package (VASP v.5.2.2). Projector augmented wave (PAW) pseudopotentials and the generalised gradient approximation (GGA) were used with a plane wave cutoff energy of 400eV. Fig. 1 shows the clamped and unzipped unit cells used to model graphene oxide in this study. For comparison, pristine monolayer graphene was also simulated using the same C₄ cell geometry as for the case of graphene oxide, albeit in the absence of the O-atoms. All structures were fully relaxed to their respective ground states prior to charge injection. A Monkhorst-Pack gamma- centred fc-point grid of dimensions 24x42x1 was adopted for the clamped graphene oxide and pristine graphene cells, with a 20x2x1 grid for the unzipped graphene oxide cell. Injected charges (electrons and holes) were compensated using a background jellium to maintain charge neutrality in the unit cell. As VASP employs periodic boundary conditions, very thick vacuum layers were included adjacent to the graphene oxide sheets in order to minimise interlayer electrostatic interactions. As shown previously by the inventors, this minimises the jellium self-energy contribution to the overall strain, and thus realistically predicts the true quantum-mechanical actuation. An interlayer spacing of 6θΑ was used throughout, which provided a good balance between computational accuracy and effort. To hold this interlayer spacing constant, the VASP source code Was modified to allow the cell to completely relax within the plane of the GO and graphene layers only, not perpendicular to the plane. In all cases, all C and O ions comprising the respective cells were allowed to relax freely in all directions.

The in-plane strains, measured as the change in lattice parameter a, for charge injection into both symmetrical zig-zag graphene oxide configurations (Fig. 1) and pristine monolayer graphene are shown in Fig. 2. The most pronounced feature is that very high strains are observed for hole (positive charge) injection into graphene oxide, especially for the clamped configuration, where a hole injection of 0.15 e/C-atom induces a 28.2% strain (see Fig. 6). The significance of such a determination for strain is that it acts along a covalently bonded axis of the material. From Fig. 2, it is evident that the same hole injection into unzipped graphene oxide and pristine graphene produces strains of only 3.6% and 4.7%, respectively, which are significantly less than the clamped graphene oxide value of 28.2%.

Further inspection of the results provided in Fig. 2 (inset) reveals that electron injection into clamped graphene oxide produces a charge-strain relationship very similar to that of pristine graphene, exhibiting expansions of up to 0.4% for a -0.15 e/C-atom charge. Conversely, the charge-strain dependency of unzipped graphene oxide is distinctly dissimilar to those of clamped GO and pristine graphene, contracting upon electron injection. This_r contraction persists right up to higher charges (-0.15 e/C- atom) and is not considered by the present inventors to be indicative of an intermediate quantum-mechanical behaviour. To explain this peculiar observation, the inventors propose that the rippling that is present in graphene oxide, being an out-of-plane structural phenomenon, allows it to be possible for the unzipped graphene oxide structure to undergo an interatomic bond-length expansion, whilst the unit cell experiences a net contraction along the -axis.

To investigate, Fig. 3 shows the expansion and contraction of symmetrical clamped (Fig. 3(a)) and unzipped (Fig. 3(b)) zig-zag graphene oxide, respectively, for a -0.15 e/C-atom charge injection. In Fig. 3(a), the interatomic bonds of clamped graphene oxide expand upon electron injection, whilst the rippling within the structure remains unchanged, leading to an expansion of the unit cell along the a-axis. This effect is identical to the actuation of pristine graphene upon electron injection in the absence of structural rippling. In Fig. 3(b), the interatomic bond-lengths of unzipped graphene oxide also change, but this time the extent of the structural rippling also changes. Fig. 3 depicts the extent of the rippling by the C-O-C bond angle a as indicated, which remains constant during pure interatomic bond-length expansion/contraction. For the unzipped case, an increase in the structural rippling effect, and thus a decrease in the C-O-C bond angle a, leads to a contraction of the unit cell along the a-axis. It is appropriate to consider and compare the clamped (Fig. 3(c)) and unzipped

(Fig. 3(d)) excess charge density distributions. For the unzipped case (Fig. 3(d)), the -0.15 e/C-atom injected charge aggregates atop the O-atom, and also atop and beneath the O-bonded C-atoms. Via inspection of the excess charge density contours, it is evident that the excess charge on the C-atoms is repelled by that on the O-atom, as the contours lean away from the O-atom. This repulsive force between the bonded O-and C-atoms results in a net out-of-plane force on the O-atom, which causes the bond angle (a) to decrease and the rippling to increase. This is contrasted with the clamped case (Fig. 3(c)), where the excess injected charge aggregates aside the O-atom, and atop and beneath the non-O-bonded C-atoms. This time, the excess charge contours indicate that there is a repulsive force acting between these two non-O-bonded C-atoms, which leads to an expansion of their C-C bond.

To further validate this, the table described above in relation to Fig. 3 reports the measured Λ-axis resolved individual-bond and total deformations. The most significant difference between the clamped and unzipped graphene oxide deformations occurs for dimension 2. Here, clamped graphene oxide experiences a significant expansion (0.0082A), whilst unzipped graphene oxide exhibits a significant contraction (- 0.0085A), with each respective deformation having an approximately equal magnitude. Another interesting observation for clamped and unzipped graphene oxide is that dimensions 1 and 3 contract for both configurations, but by varying amounts. For clamped graphene oxide the contraction is small (-0.0016A), but is more significant for unzipped graphene oxide (-0.005 A). In both cases the contraction of dimensions 1 and 3 is due to an increase in the C-C-C bond angle of the C-ring in response to the repulsive forces between the non-O-bonded C-atoms (Fig. 3(c)). Also worth noting is that both the clamped and unzipped configurations experience an expansion of dimension 4. The aforementioned large repulsive force between the non-O-bonded C- atoms for the clamped case gives rise to a large expansion (0.0122A), whilst the relatively smaller excess charge density located at the non-O-bonded C-atoms for the unzipped case causes a smaller expansion (0.0061 A). To quantify the extent of the rippling modulation contribution to the overall strain, Fig. 4 presents the breakdown of this and the interatomic bond-length changes. Here, the interatomic bond-length contribution was calculated by summing the individual interatomic bond-length changes along the -axis for a single unit cell, leading to a prediction of the strain that would be measured if the graphene oxide sheet was effectively flat (unrippled). Using these strains, it was then possible to isolate the rippling modulation strains by subtracting the interatomic bond-length strains from the total strains of Fig. 2.

Firstly, from Fig. 4(a) it is strikingly apparent that the rippling modulation effect contributes very little to the total strain for the clamped graphene oxide case. The interatomic bond-length and total strains exhibit a very close relationship across the entire charge injection window investigated, indicating that modulation of the structural rippling is a negligible effect in this structure. Recall from Fig. 2 that clamped graphene oxide behaved very similar to pristine graphene, which agrees with this result. The inventors expect that rippling modulation in clamped graphene oxide is not significant because of the C-C bond presence beneath the doped O-atoms, which results in the clamped graphene oxide structure retaining the predominant mechanical characteristics of the pristine graphene lattice. In contrast, for the case of unzipped graphene oxide (Fig. 4(b)), it is evident that rippling modulation does have a significant effect on the total strain. Interestingly, the interatomic bond-length charge-strain relationship in this case is very similar to that of pristine graphene, even displaying the same initial contraction for low hole injection levels (see the pristine graphene case in Fig. 2). Despite the similarity between the interatomic bond-length and pristine graphene strains, the total unzipped graphene oxide charge-strain relationship is far more representative of the rippling modulation strain, at all times having the same strain sign (elongation/contraction). This is also supported by an observed strong positive correlation between the total strain response of unzipped graphene oxide and the change in structural rippling, as defined by changes in the C— O-C bond angle. This demonstrates that it is possible to generate unique and high strain responses via modulation of the structural rippling for unzipped graphene oxide.

In order to explain the relatively high (eg 28.2%) strain performance of clamped graphene oxide, Fig. 5 shows the energy configuration of C4O symmetrical zig-zag graphene oxide as a function of the lattice parameter (a) and the extent of charge injection. The clamped structure represents a metastable phase of graphene oxide, whilst the unzipped case is the more stable one. Despite the higher stability of the unzipped configuration, it is nonetheless possible to synthesise clamped graphene oxide for use in practical MEMS NEMS actuators due to the relatively high (0.63eV) energy barrier that separates these two states. To understand the origin of the measured 28.2% strain, it is appropriate to consider this 0.63eV energy barrier between the clamped and unzipped phases in Fig. 5 (between a values of 4,33 and 4.7 A). Upon hole injection into clamped graphene oxide the energy profile is modified (Fig. 5), which results in an expansion of the unit cell along the α-axis (Fig. 2). This expansion is due to the concentration of excess holes at the O-bonded C-atom sites, which gives rise to a repulsive force between these C- atoms. With further hole injection, the energy profile continues to change until the energy barrier between the clamped and unzipped configurations disappears, and the clamped unit cell snaps from an initial lattice constant of 4.33A to 5.55A (whilst charged). This is the point at which the O-bonded C-atoms are repelled to such an extent that the C-C bond ruptures, which is the configurational origin of the predicted 28.2% strain. An interesting feature of this actuation mechanism is that it is possible to actuate the material by 28.2%, and then remove the input power without the strain returning to zero (a will relax to 5.36A upon charge removal, which corresponds to a 23.8% strain). This is since the material will maintain its new unzipped configuration once the C-C bonds beneath the O-atoms have ruptured and the new C— O— C bonds have formed. This feature would be particularly beneficial for long-term, low power switching applications.

Practically speaking, it is typically desirable that induced actuation be fully reversible. Due to the very high energy barrier on the approach side from the unzipped to clamped configurations in Fig. 5 (1.33eV between a values of 4.7 and 5.36A), it would seem difficult to reverse the 28.2% strain. To do so would require an increase in the extent of the unzipped graphene oxide rippling and thus a decrease in the C-O-C bond angle a, which can only be achieved via the injection of electrons. From Fig. 5, the lattice parameter (a) would need to be decreased from 5.36 to 4.75A, some -11.4%. It is evident that evert the maximum electron injection level (-0.15 e/C-atom) has little effect on the overall energy profile (Fig. 5), which agrees with the results of Fig. 2 (a strain of only -0.23% is attainable for -0.15 e/C-atom electron injection). Hence, reversal of the 28.2% strain would require some additional influence to charge injection alone. Nonetheless, even with only irreversible actuation of graphene oxide by 28% being potentially possible, this would still be extremely useful for select applications, such as legislated single-use industrial safety switches. In addition, graphene oxide can be used for the generation of high reversible strains in a more traditional actuation sense, with clamped graphene oxide being capable of generating peak strains of up to 6.3% prior to surpassing the energy barrier, and unzipped graphene oxide being capable of both large contraction (-0.25%) and expansion (3.6%).

The applicational significance of graphene oxide for use in practical actuators is further enhanced by its ability to generate very high stresses (in excess of 1 OOGPa), due to the characteristic high modulus of pristine monolayer graphene (~lTPa) and graphene oxide (~0.6TPa) along covalently-bonded directions (see Fig. 7). By Way of comparison with other actuation materials, the volumetric work density of graphene oxide^' for a 0.125 e C-atom charge injection, which equates to the highest reversible strain of 6.3%, is 52.9 J/cm³ (based on the computed strain energy). This is approximately double the value reported for the first CNT-based actuation material and some 53 times greater than the highest values presented for the widely used ferroelectric materials. The inventors also calculated the work density of pristine monolayer graphene for the highest strain predicted (4.7%), which they found to be 54.1 J/cm³. This value is slightly greater than that of graphene oxide, due to the small difference in the respective moduli. Another important practical consideration is the voltage required to inject charge into graphene and graphene oxide. Based on quantum capacitance considerations (estimated from Fermi-level shift measurements based on the integrated density of states) upon charge injection, the inventors estimate the voltages required to inject 0.15 e/C-atom charges (both electrons and holes) into graphene and graphene oxide to be 2— 4V. In a further study, the Vienna ab initio simulation package (VASP v.5.2.2) was again used to perform density functional computations on the electromechanical responses of the respective GO compounds.

V

Using density functional computations, predictions were made concerning the electromechanical responses of graphene oxide upon charge injection (see Fig. 8). The first noteworthy result is that all unzipped graphene oxide configurations exhibit the anomalous electron-induced contraction (see OO-sym-unzip, C.iO-asym-unzip and CgO-sym-unzip in Fig. 8). This effect is due to an aggregation of the excess electrons at the O atom and O-bonded C atom sites in such a manner that the C-O-C bond angle decreases, thereby resulting in a contraction of the unit cell within the basal plane. It was found that the same is true of the newly tested unzipped configurations (C₄O- asym-unzip and CgO-sym-unzip), confirming earlier preliminary trends for graphene oxide. A significant result in Fig. 8 is that of C-iO-asym-unzip graphene oxide, which has an extremely high strain sensitivity to charge injection compared to other compounds. This particular graphene oxide configuration generates a contraction of - 4.8% upon -0.15 e C-atom electron injection, a response which is more than five times greater in magnitude than any other graphene oxide configuration upon electron injection. In order to understand this unique response behaviour, computation of the interatomic and a-axis resolved deformations for GtO-asym-unzip upon -0.05 e C-atom electron injection (see Fig. 9(a)). The a-axis resolved deformations were calculated by considering only the a-axis interatomic expansion components, as these are the components responsible for net changes in the unit cell length along this direction of the basal plane. Due to basal plane symmetry, there are only three principal interatomic segments, labelled 1, 2 and 3 in Fig. 9(a), respectively. As shown in the inset table, C40-asym-unzip graphene oxide experiences a net interatomic expansion, with segment 2 expanding by a considerable amount (0.014 A). Segment 1 does exhibit a small contraction (-0.003 A) due to an increase in the associated C ring C-C-C bond angle, but this is more than offset by the expansion of segment 3. Despite this net interatomic expansion, examination of the a-axis resolved deformations makes it clear that all segments show large contractions (< -0.019 A). These individual contractions lead to an overall contraction of the unit cell along the a-axis of -0.192 A, which equates to -2.7%. This is in comparison to the behaviour of dO-sym-unzip graphene oxide (see Fig. 9(b)). Here; all segments (1, 2 and 3) experience an interatomic expansion, while the a-axis resolved changes are dominantly negative, resulting in a net contraction of the unit cell along the a-axis by -0.004 A.

Further insight into the high magnitude contraction of C40-asym-unzip graphene oxide can be garnered by considering the excess charge density distribution upon electron injection (see Fig. 9(c)). Here, the excess injected electrons aggregate atop the O atom, and atop and beneath the O-bonded C atoms. As most of the excess injected charge collects atop these atoms, a large repulsive interaction between these electronic regions exists, which causes the C-O-C bond angle to decrease in order to lower the total energy. This bond angle decrease leads to an increase in the degree of rippling within the graphene oxide layer, which is responsible for the shrinkin of the unit cell along the a-axis. To understand why the electromechanical strain sensitivity of C-jO-asym-unzip graphene oxide is so high, a plot was made of the excess charge density distribution of C40-sym-unzip graphene oxide (see Fig. 9(d)); the aim being to highlight that the only difference between these structures is the symmetry. Similar to the case of C-jO-asym-unzip graphene oxide, C₄0-sym-unzip appears to exhibit excess electron aggregation atop the O atoms, as well as atop and beneath the O-bonded C atoms. This leads to the question of why the sensitivity of C-_tO-asym-unzip graphene oxide is more than five times higher. In Fig. 9(c), there exists regions of excess hole concentrations. This indicates that additional electrons to those injected could in fact be contributing to the overall response, such that there would in effect be two contributions to the total contraction: (1) an "extrinsic" contribution (due to the injection of foreign electrons), and (2) an "intrinsic" contribution (due to the redistribution of domestic valence electrons). This is understood to be the first reported observation of a possible intrinsic electromechanical actuation in graphene-based materials. In addition, for C₄0-sym-unzip graphene oxide (see Fig. 9(b)), a reduction in the C-O-C bond angle would tend to flex each graphene nanoribbon between adjacent epoxy groups (increasing the elastic energy), which would hinder the bond angle change and thus the basal plane strain. Conversely, the response of C₄0-asym- unzip graphene oxide is not hindered in any such way (see Fig. 9(a)). Upon low concentration hole injection (0.025 e/C-atom), C₄0-asym-unzip graphene oxide is also observed to contract by a considerable amount (-1.3%), following which further hole injection eventually induces large expansions (up to 9.6%). The origin of the latter (high concentration hole induced expansion) can be explained by considering the molecular orbitals associated with the C-O-C bond. To this end, we have calculated the spd- and site-projected density of states (DOS), as well as the excess charge density distributions, of C/_tO-asym-unzip graphene oxide upon hole injection (see Fig. 10). The O atom in GjO-asym-unzip graphene oxide has two sp ike atomic orbitals in the C-O-C plane, which from σ bonds with the sp² atomic orbitals of the C atoms. The O atom has a further two lone electron pairs, one of which exists purely as a p- orbital with its electron density perpendicular to the C-O-C plane, while the other is close to an sp² orbital in the C-O-C plane. As is evident in Fig. 10(a), the O atom has non-zero s- and p-DOS near the Fermi level (Er=0 eV). This is further confirmed by the excess charge density distributions for low concentration hole injection (see Figs. 10(c) & (d)), where there exists a clear spMike orbital surrounding the O atom. The C- ; O σ bonds should correspond to energies of between -15 and -5 eV (for Ef =0), where the s- and p-DOS of both the C and O atoms overlap significantly. Thus, the two lone pair orbitals of the O atom should correspond to energy levels of -5 to 0 eV (see Fig. 10(a)).

With a moderate level of hole injection (>0.05 e C-atom), the electron density of the C atom's π orbital, and the O atom's sp² lone pair orbital, shrinks (see Fig. 10(a)). Such a reduction in the overlap of these two orbitals appears to reduce the repulsive interaction between them, thus leading to an increase in the C-O-C bond angle and an expansion of the unit cell along the basal plane. For high concentration hole injections (0.15 e/C-atom), significant changes in the DOS are observed. Near the Fermi level, the p-DOS of the O atom significantly increases with a reduction in the s-DOS. Meanwhile, between energy levels of -22.5 and -24 eV, we observe an increase in the s-DOS and a clear reduction in the p-DOS of the O atom. This suggests that a hole injection at this level will change the sp²-like nature of the lone pair to a more p-like orbital. This is supported by the excess charge density plot in Fig. 10(e). In this case, the O atom becomes more C-like in the graphene oxide lattice, having both σ and π bonds with its nearest neighbours. As a result, the graphene .oxide structure becomes flat. As for the low concentration hole (0.025 e/C-atom) induced contraction of C₄0- asym-unzip graphene oxide, the above molecular orbital discussion is considered, by itself insufficient to fully describe this phenomenon, which is similar to that seen in pristine graphene (albeit with a much greater magnitude). With reference to Fig. 11, a quasi-linear relationship exists between the charge injection and the C-0 interatomic bond length change, where hole injection always leads to a contraction of the C-O bond. This C-0 bond length contraction tends to enhance the overlap of C atom's π orbital with the O atom's sp² lone pair orbital. As such, it is postulated that for low concentration hole injection, this effect overwhelms the shrinkage of the two orbitals due to the removal of electrons, and thus results in an observable decrease of the C-O-C bond angle.

Another observation from Fig. 8 is that graphene oxide compounds with lower O concentrations tend to exhibit higher electromechanical strains. For example, consider the responses of CeO-sym-unzip and CgO-sym-unzip graphene oxide for both electron and hole injection. Regardless of the sign of the charge injection, the CgO compound generates significantly higher strains; sometimes up to four times higher. The same can be said of the comparison between CjO-sym-clamp and GjO-sym-clamp graphene oxide. For both electron and hole injection, the G»0 compound produces in excess of double the strain magnitude of the C2O compound. Whilst one could reasonably expect that as the O concentration diminishes the electromechanical response of graphene oxide should approach that of pristine graphene, this is evidently not the case. It is proposed that this anomalous effect is due to an increase in the degree of structural rippling at lower O concentrations, as the distance between adjacent O atoms increases. To qualify this statement, consider again the comparison between C40-sym-unzip and CgO-sym-unzip graphene oxide. The maximum electron- induced contractions for these graphene oxide configurations (at -0.15 e/C-atom) are - 0.23 and -0.87%, respectively. If we consider the effective layer thickness of each configuration in an attempt to quantify the degree of structural rippling, which is given by the c-axis distance between the lowest C atom and the highest O atom, we find that the thicknesses are 0.51 and 0.89 A, respectively. The larger effective layer thickness of the CgO compound derives from its greater level of basal plane rippling, and thus corifirms that the degree of rippling is correlated with the strain capacity of graphene oxide. In light of this, one should clearly seek to maximise the degree of structural , rippling in a graphene oxide based electromechanical actuator in order to optimise its performance.

An important consideration during the development and optimisation of any actuation material, in addition to the producible strains, is the stress generation capacity. As the herein graphene oxide compounds retain the inherent high modulus of the pristine graphene lattice, with values on the order of 0.6 TPa, our very high predicted strains are coupled with equally high stresses of 100 GPa. This unique high stress-strain coupling makes graphene oxide a particularly suitable building block for artificial muscles. To test this, consider the electromechanical strain (ε), strain-voltage coefficient (S_v) and volumetric work density (W_vo/) comparisons in Table 1. Next to mammalian skeletal muscle, C40-asym-unzip graphene oxide produces the highest electromechanical strains compared to pristine graphene, CNT-based actuators, ferroelectric materials and the other graphene oxide configurations investigated. This is especially significant when orie considers that muscle operates via contraction, rather than expansion, and this particular graphene oxide compound produces significantly larger contraction values than previously reported. This is an important aspect to consider when seeking a material for artificial muscle, since an elongate material is prone to buckling during expansion, but not during contraction. Whilst the present contraction of -4.8% is presently lower than mammalian muscle (-20%), further improvements are expected. For artificial muscle applications it is also important to consider the voltages required to achieve a given strain, which is best computed via S_v. In order to produce strains of >20% at safe voltages of <20V, S_v values of greater than 1% V are required. From Table 1, it is evident that several materials satisfy this requirement. However, these suitable S_v values need to be coupled with sufficiently high strains in order for a material to be a suitable candidate for artificial muscle. Presently, C40-asym-unzip graphene oxide appears to best satisfy these conditions. The final consideration, which accounts for the stress and strain generation capacities in a collective manner, is a given actuation material's W_vo/. Due to its very low stress capacity (max. 0.35 MPa), mammalian skeletal muscle has a relatively low maximum Wvo/ value of 0.04 J/cm³. There are a multitude of materials that are capable of equal or higher W_vo/ values than muscle, including even piezoelectric, magnetostrictive and electrostrictive materials. By way of comparison, C₄0-asym-unzip graphene oxide is capable of generating much higher W_vo/ values than the other listed materials, due to its unique high stress-strain coupling. Finally, in terms of the response speed, as the herein predicted actuation responses are QM in origin, we expect that the response times should be on the order of, or less than, 1 ns. Across the board, C40-asym-unzip graphene oxide fares very well against other available materials and skeletal muscle, highlighting its potential as a possible building block for an artificial muscle actuator. TABLE 1. Electromechanical strain ε (%), strain-voltage coefficient S_v (% V), and volumetric work density W_vo/ (J/cm³) comparison between graphene oxide and other materials.

Actuators formed using the graphene oxide described herein can take various forms. In one arrangement, the actuator can comprise a composite of two or more layers of graphene oxide, with at least one layer being the graphene oxide that elongates and/or contracts on charge injection as defined herein. All layers of the composite can be graphene oxide, with at least one layer being said graphene oxide that elongates and/or contracts on charge injection. In another embodiment, the composite can comprise layers of graphene oxide, with at least one layer being said graphene oxide that elongates and/or contracts on charge injection, and one or more other materials. In a further embodiment, all of the graphene oxide layers of the composite can have the same configuration. In a still further embodiment, some or all of the graphene oxide layers can have different configurations to that of other graphene oxide, layers.

In yet another embodiment, the actuator can comprise one or more narrow strips or nanoribbons of graphene oxide that elongate and/or contract on charge injection. In a still further embodiment, the actuator can comprise graphene oxide that elongates and or contracts on application of an electric field. Such an actuator can have one, some or all of the features of the actuator defined herein. In yet another embodiment, the actuator can comprise graphene oxide that elongates and/or contracts on application of a magnetic field. Such an actuator can have one, some or all of the features of the actuator defined herein. The magnetic field can be created using electric charge (an electromagnet). As described herein, the described forms of graphene oxide can also be useful for micro or nano-electro-mechanical systems (MEMS NEMS) actuators. In another embodiment, the actuator comprising single or multiple layers of graphene oxide as defined can be usable as artificial muscle. In another embodiment, the actuator as defined can comprise or be part of a switch, particularly switches used in relatively long-term, low-power switching applications or, for example, a switch used in fibre optic communications. Where the elongation is only at least partially reversible, the actuator can comprise part of a single-use switch. The actuator can be used as a resonator, including broadband resonators. The actuator can also be actuated by an alternating current electrical signal. In other embodiments, the actuator as defined can be used in optical zoom lenses and/or autofocus lenses, including lenses used in mobile telephones, including smartphones, and digital cameras). The actuator as defined can also comprise or be used as part of a robotic actuator including robotic actuators for use in minimally invasive surgery. In other embodiments, the actuator as defined can be used in electromechanical valves and controllable channels for use in micro and/or nanofluidic devices. In yet another embodiment, the actuator as defined can be used in devices such as electromechanical carbon nanotube extruders.

It will be appreciated by persons skilled in the art that numerous variations and/or modifications may be made to the above-described embodiments, without departing from the broad general scope of the present disclosure. The present embodiments are, therefore, to be considered in all respects as illustrative and not restrictive.

Claims

CLAIMS: ,

1. An actuator comprising graphene oxide that elongates and/or contracts on charge injection.

2. The actuator of claini 1 wherein the graphene oxide has a zig-zag configuration whereby the oxygen atoms of the graphene oxide are aligned along the zig-zag direction of the graphene lattice.

3. The actuator of claim 1 wherein the graphene oxide has an armchair configuration whereby the oxygen atoms of the graphene oxide are aligned along the armchair direction of the graphene lattice.

4. The actuator of any one of the preceding claims wherein the oxygen atoms are assembled on only one side of the graphene lattice in a symmetrical configuration.

5. The actuator of any one of claims 1-3 wherein the oxygen atoms are assembled on both sides of the graphene lattice in an asymmetrical configuration.

6. The actuator of claim 5 wherein one or more oxygen atoms are assembled in such positions that the graphene oxide has an interchanging periodicity.

7. The actuator of any one of the preceding claims wherein the actuator comprises unzipped graphene oxide.

8. The actuator of claim 7 wherein the actuator elongates upon electron and/or hole injection.

9. The actuator of claim 7 wherein the actuator contracts upon electron and/or hole injection.

10. The actuator of claim 8 or claim 9 wherein the elongation or contraction is in the plane of the graphene oxide.

11. The actuator of claim 8 or claim 9 wherein the elongation or contraction is at least partly or fully reversible.

12. The actuator of claim 8 or claim 9 wherein the elongation or contraction increases with increasing charge (electron or hole) injection.

13. The actuator of claim 8 wherein the elongation increases up to about or exactly 3.5% for a 0.15 e/C-atom charge injection.

14. The actuator of claim 9 wherein the contraction increases up to at least about or exactly 0.25% for a -0.15 e/C-atom charge injection.

15. The actuator of any one of claims 1 to 6 wherein the actuator comprises clamped graphene oxide.

16. The actuator of claim 15 wherein the actuator elongates upon electron and/or hole injection.

17. The actuator of claim 15 wherein the actuator contracts upon electro and/or hole injection. ⁷ .

18. The actuator of claim 16 or claim 17 wherein the elongation and/or contraction of the graphene oxide is at least partly or fully reversible and the graphene oxide contracts from its elongated condition, and/or elongates from its contracted condition, on removal of the charge (electron or hole) injection.

19. The actuator of claim 16 wherein the elongation is irreversible even on removal of the charge (electron or hole) injection.

20. The actuator of claim 16 or claim 18 wherein the elongation is only at least partly reversible once a certain percentage of elongation is exceeded.

21. The actuator of claim 16 or claim 17 wherein the elongation and/or contraction is in the plane of the graphene oxide.

22. The actuator of claim 17 wherein the contraction is at least partly or fully reversible.

23. The actuator of any one of claims 15-17 wherein the elongation and/or contraction increases with increasing charge (electron or hole) injection.

24. The actuator of claim 16 wherein the elongation increases up to about or exactly 28.2% for a 0.15 e/C-atom charge.

25. The actuator of claim 16 wherein if the elongation reaches about or exactly 28.2%, upon removal of the charge, the elongation will reverse to about or exactly 23.8% strain.

26. The actuator of claim 16 or claim 18 wherein the elongation is fully reversible so long as elongation is at or about 6.3% or less.

27. The actuator of claim 15 or claim 16 wherein the elongation is up to about 0.4% for about a -0.15 e C-atom charge.

28. The actuator of claim 1 wherein the actuator comprises a composite of two or more layers of graphene oxide, with at least one layer comprising said graphene oxide that elongates and/or contracts on charge injection.

29. The actuator of claim 28 wherein all layers of the composite comprise graphene oxide.

30. The actuator of claim 28 wherein the composite comprises layers of graphene oxide and one or more other materials.

31. The actuator of claim 29 or claim 30 wherein all graphene oxide layers of the composite have the same configuration.

32. The actuator of claim 29 or claim 30 wherein some or all graphene oxide layers comprise different configurations to those of other graphene oxide layers.

33. The actuator of claim 1 or claim 28 wherein the actuator comprises one or more narrow strips or nanoribbons of graphene oxide that elongate and/or contract on charge injection.

34. An actuator comprising graphene oxide that elongates and/or contracts on application of an electric field.

35. The actuator of claim 34 wherein the actuator has any one or more of the features of the actuator defined in any one of claims 1-33.

36. An actuator comprising graphene oxide that elongates and/or contracts on application of a magnetic field.

37. The actuator of claim 36 wherein the magnetic field is created using electric charge.

38. The actuator of claim 36 or 37 wherein the actuator has any one or more of the features of the actuator defined in any one of claims 1-33.

39. The actuator of any one of the preceding claims when used for micro or nano- electro-mechanical systems (MEMS/NEMS) applications.

40. The actuator of any one of the preceding claims when used as artificial muscle.

41. The actuator of any one of claims 1-39 when used as part of a switch.

42. The actuator of ^" claim 41 wherein the switch is a single-use switch or a switch used in fibre optic communications.

43. The actuator of any one of claims 1-39 when used as a resonator, including broadband resonators.

44. The actuator of any one of claims 1-39 when actuated by an alternating current electrical signal.

*

45. The actuator of any of claims 1-39 when used in optical zoom lenses and/or autofocus lenses.

46. The actuator of any of claims 1-39 when used as part of a robotic actuatoT, including robotic actuators for use in minimally invasive surgery.

47. The actuator of any of claims 1-39 when used in electromechanical valves and controllable channels for use in micro and/or nanofluidic devices.

48. The actuator of any of claims 1-39 when used as an electromechanical carbon nanotube extruder.

49. A method of delivering a force comprising actuating an actuator as defined by any one of the preceding claims.