ARTICLE

Received 22 Jan 2014 | Accepted 25 Apr 2014 | Published 3 Jun 2014

Magnetocaloric effects in a freestandingand flexible graphene-based superlatticesynthesized with a spatially confined reactionHaiou Zhu1, Chong Xiao1, Hao Cheng2, Fabian Grote3, Xiaodong Zhang1, Tao Yao2, Zhou Li1,

Chengming Wang1, Shiqiang Wei2, Yong Lei3 & Yi Xie1

Superlattices have attracted great interest because of their tailorable electronic properties at

the interface. However, the lack of an efficient and low-cost synthetic method represents a

huge challenge to implement superlattices into practical applications. Herein, we report a

space-confined nanoreactor strategy to synthesize flexible freestanding graphene-based

superlattice nanosheets, which consist of alternately intercalated monolayered metal-oxide

frameworks and graphene. Taking vanadium oxide as an example, clear-cut evidences in

extended X-ray absorption fine structure, high-resolution transmission electron microscopy

and infrared spectra have confirmed that the vanadium oxide frameworks in the superlattice

nanosheets show high symmetry derived from the space-confinement and electron-donor

effect of graphene layers, which enable the superlattice nanosheets to show emerging

magnetocaloric effect. Undoubtedly, this freestanding and flexible superlattice synthesized

from a low-cost and scalable method avoids complex transferring processes from

growth substrates for final applications and thus should be beneficial to a wide variety of

functionalized devices.

DOI: 10.1038/ncomms4960

1 Hefei National Laboratory for Physical Sciences at the Microscale, and Collaborative Innovation Center of Chemistry for Energy Materials, University ofScience & Technology of China, Hefei, Anhui 230026, China. 2 National Synchrotron Radiation Laboratory, University of Science & Technology of China, Hefei,Anhui 230029, China. 3 Institute of Physics and IMN MacroNanos (ZIK), Ilmenau University of Technology, 98693 Ilmenau, Germany. Correspondence andrequests for materials should be addressed to Y.X. (email: yxie@ustc.edu.cn) or to S.W. (email: sqwei@ustc.edu.cn).

NATURE COMMUNICATIONS | 5:3960 | DOI: 10.1038/ncomms4960 | www.nature.com/naturecommunications 1

& 2014 Macmillan Publishers Limited. All rights reserved.

Superlattices have been the subject of active research in thepast two decades because of their fascinating physicalphenomena arising from new electronic states at the

interface1, which opens exciting opportunities to design newmaterials with novel properties2,3. Since the first experimentalsuperlattice growth by molecular beam epitaxy4, a spectacularseries of advances has been accomplished in areas such asthe discovery of high-temperature superconductors5, integerand fractional quantum Hall effect6,7, giant and colossalmagnetoresistance8,9 and topological insulators10,11. However,the conventional superlattice-fabrication strategies are usuallybased on pulsed laser deposition or molecular beam epitaxy,which are complex and expensive as well as involve difficulty inthe transferring process from the growth substrate for finalapplication12. In fact, the lack of an efficient and low-costsynthetic method represents a huge challenge to implementsuperlattices in practical applications, which inspires us to furtherpursue the low-cost and scalable bottom-up techniques for thesynthesis of superlattices.

Graphene, consisting of stacked sheets (o10) of two-dimensional sp2-hybridized carbon atoms arranged in six-membered rings, has generated enormous interest in the pastyears because of its prominent intrinsic properties and greatpotential in various applications13,14. The unique electronicstructure of graphene endows its extraordinary ambipolarelectronic property, which implies that graphene can serve asan electron donor or acceptor to continuously tune the holes andelectrons when combined with other substances. In this sense,graphene is a good matrix for hybrid materials and in particulargraphene-based superlattices15–19, which are envisaged tosynergistically offer tremendous opportunities for improvingand expanding applications. On the other hand, as is well known,few-layered graphene is an intermediate state between graphiteand monolayered graphene, which shows superiority overgraphite and monolayered graphene for growing graphene-based superlattice, for the interlayer spacing of few-layeredgraphene is remarkably larger than that of graphite. At the sametime, the graphene nanosheets could be better stacked comparedwith monolayered graphene. Therefore, few-layered graphene is aperfect material for growing such graphene-based superlattice.

Owing to the discovery of many distinctive physical properties,such as superconduction20,21, superionic conduction22,23, giantmagnetocaloric effect24,25, optical storage26 and so on27, thesubject of phase transition in solid materials has growncontinuously and enormously for centuries, forming not onlythe basis of materials technology but also the central issue ofsolid-state chemistry. VO2 is a well-known prototype material tounderstand the electronic structure and related physicalproperties during phase transitions28. For example, in the rutileVO2(R), the isometric V–V bond (2.88 Å) results in itinerant3d electrons in the overlapping V 3d orbitals, leading to anelectrically conducting state. Meanwhile, in the monoclinicVO2(M), the V–V bond splits into alternately altered bondlengths of 2.62 and 3.16 Å; thus, the outer V 3d electrons arelocalized, which results in an electrically insulating state. Astemperature varies, a fully reversible structural phase transitionprocess associated with optical, electric and magnetic propertychanges takes place between these two phases of VO2. AlthoughVO2(M/R) has been regarded as a thermodynamically stablephase, it needs to overcome a high reaction barrier to form thisphase in a solution reaction system. As a result, a metastablephase with shear structure consisting of vertex and edge sharingof deformed VO6 octahedra, VO2(B), is regarded as the mostcommon phase obtained from solution reactions29, which hasbeen used as the exclusive precursor for obtaining VO2(M/R)through solid-state transformation, even though thistransformation usually requires the rigid synergic effects ofhigh-temperature post treatment, inert-gas atmosphere withprecisely controlled flow and long synthesis time30. Recently,our group reported that the injection of electrons into the V–Osystems by adventitious hydrogen can lead to an enhancement ofthe e–e correlation effects and strengthening of the electronicorbital overlap of the infinite V–V chains, which results in thestabilization of high-temperature-phase metallic VO2(R) at roomtemperature31. These facts suggest that the introduction ofstructural lattice perturbation by an additional substance wouldsimultaneously activate the spin, charge and/or orbital of amaterial, and would provide an effective way to modulate itselectronic structure, resulting in a controllable electronic structureand its related physical properties.

VO(acac)2

VO(acac)2 hydrolyse Polycondensation

Polycondensation

–H2ODisperse

RGO

HRTEM Superlattice

11.5 Å

+

Figure 1 | Schematic representation of the formation mechanism for the superlattice through space-confined nanoreactor strategy. VO(acac)2 is firstly

hydrolysed to yield [VO(H2O)5]2þ , then intercalated into the interlayers of RGO nanosheets. The intercalated [VO(H2O)5]2þ condenses to VO6

octahedra in the RGO interlayers formed space-confined nanoreactor. Finally the VO6 octahedra further condenses to yield vanadium oxide nanosheets in

the interlayers of RGO nanosheets, resulting in the formation of the superlattice.

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms4960

2 NATURE COMMUNICATIONS | 5:3960 | DOI: 10.1038/ncomms4960 | www.nature.com/naturecommunications

& 2014 Macmillan Publishers Limited. All rights reserved.

Here we report a space-confined method to synthesizegraphene-based superlattices, in which the graphene layersprovide not only a space-confined nanoreactor to direct thegrowth of monolayered metal oxide frameworks in a two-dimensional manner, but also serve as an electron donor toreduce the formation energy for highly symmetrical metal-oxideframeworks. The electron transfer and stress field at the interfaceof the superlattice lead to the structural rearrangementsof the metal-oxide frameworks, which endow the superlatticenanosheets more fascinating physicochemical properties. Takingvanadium oxide as an example, we experimentally obtain thefreestanding and flexible vanadium oxide/graphene superlatticenanosheets, which show significantly enhanced magnetocaloriceffect compared to that observed in the VO2(M)–VO2(R) phasetransition. Likewise, these superlattice nanosheets also showsmart switching behaviour in electrical conductivity and near-IRoptical as well as magnetic properties, which are very attractivefor intelligent devices such as temperature sensors.

ResultsStructural identification. These graphene-based superlatticeswere synthesized in a space-confined nanoreactor formed by theinterlayers of reduced graphene oxides (RGO) (as seen in Fig. 1).As is well known, RGO nanosheets usually possess some oxygen-containing groups, such as hydroxyl groups and carboxylic acidgroups, which make the RGO negatively charged after dispersingin water. Meanwhile, the VO(acac)2 hydrolyse to the positivelycharged species of [VO(H2O)5]2þ in water, which can intercalateinto the interlayers of negatively charged RGO nanosheets

through the electrostatic attraction effect. Under the followingsolvothermal reaction condition, the homogenously intercalated[VO(H2O)5]2þ condenses to VO6 octahedra in the RGO’sinterlayers formed space-confined nanoreactor, which yields thehomogenously distributed superlattice. During the condensationpolymerization process, the oxygen-containing groups on RGOdevelop to act as bridging oxygen between the graphene layer andthe V–O framework, through which the electron transfer processoccurs between the graphene layer and the V–O framework.As demonstrated in our previous report29, the VO(acac)2 firsthydrolyses to [VO(H2O)5]2þ and then condenses to VO6

octahedra, and the vertex and edge sharing of this VO6

octahedra usually form VO2(B) in an open solution reactionsystem, which is confirmed by a parallel experiment where otherexperimental conditions are kept the same except for absence ofRGO (see Supplementary Fig. 1). However, in this space-confinednanoreactor strategy, the graphene layers could provide a space-confined nanoreactor to avoid the fast aggregation of VO6

octahedra and direct the VO6 octahedra to connect with eachother in a two-dimensional manner, forming a monolayered V–Oframework similar to monolayered VO2(B), with some structuraldeformation, which is also observed in exfoliated VO2(B)32.Furthermore, the V–O framework could accept electronsfrom graphene to reduce the formation energy for obtaining ahigh symmetry. As expected, X-ray absorption fine structurespectroscopy, a powerful element-specific tool to study the localatomic arrangements and electronic structures of condensedmatters, provides evidence on the formation of highly symmetricV–O frameworks in the superlattice nanosheets. Extended

000 1 2 3 4 5 6

ab

ba

V2p2/3

O1s C-C

C-O

C-O-C

C=OV2p1/2In

tens

ity (

a.u.

)

530 520 510Binding energy (eV)

292 288 284

c

~2.6 Å

Bridgingoxygen O

1.96 Å 1.96 Å

2.77 Å

V1

2

3

4

5

–0.8

–0.6

–0.4

–0.2

0

0.2

0.4

0.6

0.81.0

a

e f g

b c dSuperlatticeVO2(B)

2 4 6k (Å–1)

k3 χ(k)

FT

(k3 χ(

k)

R (Å)8

G

Measured superlattice

Calculated superlattice

VO2(B)

JCPDS 81-2392

10 202� (°)

30 40 50

D

Superlattice

RGO

600 800 1,000Raman shift (cm–1)

1,200 1,400 1,600

Inte

nsity

(a.

u.)

Inte

nsity

(a.

u.)

10

SuperlatticeVO2(B)

Figure 2 | Atomic arrangements and structural characterization of the superlattice nanosheets. (a) V K-edge extended XAFS oscillation function

k3w(k) and (b) Fourier transforms of k3w(k) for the V–O framework in superlattice nanosheets and VO2(B), respectively. (c,d) Representative atomic

models of superlattice nanosheets extracted from the quantitative fitting results of the V–O and V–V bonds. (e) The Raman spectra for the obtained

superlattice and starting RGO, respectively. (f) Experimental XRD patterns of the superlattice nanosheets and VO2(B). The standard pattern of VO2(B)

in JCPDS card no. 81-2392 and the calculated XRD pattern of the constructed model structure of superlattice nanosheets are also shown. (g) The XPS

spectra of the obtained superlattice.

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms4960 ARTICLE

NATURE COMMUNICATIONS | 5:3960 | DOI: 10.1038/ncomms4960 | www.nature.com/naturecommunications 3

& 2014 Macmillan Publishers Limited. All rights reserved.

X-ray absorption fine structure (EXAFS) spectroscopy is themeasurement of the x-ray absorption coefficient of a material as afunction of energy, typically in a 500-1000 eV range beginningbefore the absorption edge of an element in the sample. It is auseful tool for the investigation of local bond lengths in theenvironment of a given element. The k3w(k) EXAFS oscillationcurves shown in Fig. 2a reveal remarkable differences between theV–O framework of the superlattice nanosheets and VO2(B) in thek-region, implying the different local atomic arrangement. Thesedifferences are further illustrated by the Fourier transformed (FT)k3w(k) functions in R-space shown in Fig. 2b. The intensities ofthe nearest V–O and V–V coordination shells of the superlatticenanosheets are reduced compared with VO2(B) because thedegree of order in the V–O framework in the superlatticenanosheets was decreased. The schematic structure model of theintercalated V–O framework is presented in Fig. 2c,d, which isobtained from the quantitative fitting results of the V–O bonds(see Supplementary Fig. 2). Compared with VO2(B), the V–Oframework in the superlattice nanosheets exhibits a moresymmetric structure. The highly symmetrical V–O frameworkin superlattice nanosheets shows a zigzag V–V chain with anisometric bond length of about 2.77 Å, which corresponds closelyto the value of linear V–V chains with an isometric bond lengthof 2.88 Å in VO2(R)33; thereby both of the V–V chains in thesuperlattice nanosheets and VO2(R) are derived from the edgesharing of the adjacent VO6 octahedron. Except for a moresymmetric structure, the V–O framework of the superlatticenanosheets possesses an expanded structure along the inter-calation direction. This means that the average length of the V–Obonds along the intercalation direction is elongated comparedwith VO2(B), which may be the result of the electron transferfrom the graphene layer to the V–O framework layer. Thiselectron transfer is confirmed by Raman spectra. Figure 2e showsthe typical Raman spectra recorded for RGO and the obtained

superlattice samples, which yield D bands at 1346 and1378 cm� 1, respectively. Compared to RGO, the D band of thesuperlattice shows a clear Raman peak shift towards highfrequency and intensity reduction, which demonstrates theelectron transfer from the graphene layer to the V–Oframework layer. (A detailed explanation is provided inSupplementary Fig. 3 and Supplementary Note 1.)

The superlattice structure has been verified by XRD as shownin Fig. 2. As seen from the XRD patterns, compared to that of thestarting graphene oxide (GO) and RGO (see SupplementaryFig. 4), a newly arising d-spacing at d¼ 11.5 Å appears in thesuperlattice nanosheets, implying a new interlayer distance. Themeasured XRD pattern of the superlattice nanosheets matcheswell with the calculated XRD pattern of the constructed modelstructure with alternately intercalated V–O frameworks andgraphene layers.The transmission electron microscope (TEM)image and the selected area electron diffraction (SAED) patternswere recorded to further confirm the superlattice structure. Asseen from Fig. 3a, the corresponding TEM image confirms thetwo-dimensional (2D) morphology of the as-obtained superlatticenanosheets. Meanwhile, the folded cross-section edge confirmsthe flexibility of our superlattice nanosheets, which is differentfrom that of the synthetic superlattice obtained by pulsed laserdeposition or molecular beam epitaxy, implying a more practicaluse in applications, especially in the flexible functionalizeddevices. The high-resolution TEM (HRTEM) image and SAEDpattern of the cross-section edge shown in Fig. 3b indicate theexistence of a distinctive superlattice structure that consists of11.5 Å periodic slabs along the [001] direction. The periodicityof 11.5 Å is approximately the sum of the layer spacings ofvanadium oxide and graphene, agreeing well with the XRD resultsand suggesting that the monolayered V–O frameworks alternatelyintercalate into the adjacent graphene layers. A HRTEM image ofthe basal plane is shown in Fig. 3c, in which the interplanar

a b

c d

1

2

11.5 Å

11.5 Å

1.96 Å

1.96 Å

1.39 Å

1.99 Å

1.40 Å

1.40 Å

1.99 Å

1.99 Å

1.99 Å

Figure 3 | TEM and HRTEM images of the superlattice nanosheets. (a) Low-magnification TEM image confirms the two-dimensional (2D) morphology

and flexibility of our as-obtained superlattice nanosheets, (b) HRTEM image of the cross-section edge of the superlattice nanosheets as highlighted by

dotted box 1 in (a). The inset of (b) shows the corresponding SAED pattern of the cross-section edge. (c) HRTEM image of the basal plane of the

superlattice nanosheets as highlighted by dotted box 2 in (a). The inset of (c) shows the interplanar spacing obtained from the structure models.

(d) Corresponding SAED pattern of the basal plane. Scale bar, 5 nm.

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms4960

4 NATURE COMMUNICATIONS | 5:3960 | DOI: 10.1038/ncomms4960 | www.nature.com/naturecommunications

& 2014 Macmillan Publishers Limited. All rights reserved.

spacings were measured to be 1.99 and 1.40 Å according to theperiodic pattern of the lattice fringes, matching up with the planespacings of the fitted V–O framework by EXAFS (inset in Fig. 3c),respectively. The corresponding SAED pattern of the basal plane(as seen in Fig. 3d) displays a strong spot pattern associated withelongated satellite spots, indicating that the graphene layers aredistorted in the V–O framework layers, which has been widelyrecognized in superlattice intercalate systems of a Bragg signatureof a modulated structure34,35. The uniformity of this obtainedsuperlattice is subjected to detailed discussion in SupplementaryFig. 4 and Supplementary Note 2. X-ray photoelectronspectroscopy (XPS) (Fig. 2g) was also done to ascertain thecomposition of the superlattice. The two peaks at 516.5 and523.7 eV can be attributed to V2p2/3 and V2p1/2 (ref. 32),respectively The deconvoluted C-1s XPS spectrum shows the fourpeaks at 284.8, 286.3, 287.2 and 288.3 eV, due to C–C, C–O,C–O–C and C¼O bonding, respectively36.

Emergence of the phase transition and its characterization. Asaforementioned, due to the space-confinement effect, the V–Oframeworks in the superlattice nanosheets show high symmetrywith an isometric V–V bond length of about 2.77 Å, similar to theVO2(R), which allows us to predict that it will show similarproperties to that of the VO2(R). The apparent endothermic andexothermal peaks in the differential scanning calorimetry (DSC)curves (as shown in Fig. 4a) undoubtedly indicate reversible first-order phase transition in our superlattice nanosheets. FT infrared(FTIR) spectroscopy was also used to further confirm the phasetransition in the superlattice nanosheets because the vibrationfrequencies of chemical bonds vary with the atoms’ geometric

arrangement, providing a ‘fingerprint’ information in a givenmolecular structure37. As shown in Fig. 4b,c, the high-temperature FTIR spectra show absorbing peaks at 560,926 and 1,003 cm� 1, which can be indexed to the edge-sharingV–O stretching vibration, V–O–V bending vibration and V¼Ostretching vibration in the V–O framework, respectively, whilethe peaks at 1,109, 1,400 and 1,645 cm� 1 can be indexed tovibrations of C-O, –OH, and –C¼O in the graphene layers.Evident difference in the absorbing peaks arises at about 560 and926 cm� 1 corresponding to a blue shift of D¼ 26 and 30 cm� 1,respectively, as the temperature decreased across the transitionpoint, implying the symmetry breakdown of the V–O frameworkbelow the transition temperature, which is analogous to thechange of IR spectra during the phase transition of VO2(M) andVO2(R) at around 340 K.

Next, we employed Raman spectroscopy to confirm the phasetransition of the superlattice, since rotational and vibrationalinformation specific to the chemical bonds and symmetryof molecules could be inferred from the shifted frequencies. Asseen in Fig. 4d, the Raman spectra of the superlattice showdistinct difference between the high-temperature phase andlow-temperature phase. Except for the peaks of graphene(1,300–1,600 cm� 1), most of the Raman peaks ascribed to thelow-temperature phase disappeared as temperature increased toroom temperature. The distinctly dwindled numbers of Ramanshift peaks in the high-temperature phase indicate that the V–Oframework in the superlattice at high temperature is of highersymmetry.

It is noteworthy that the IR transmittance is connected throughmetallic free electrons, the so-called free electrons in metallic

a b

c d

0.4

0.2

0

–0.2

–0.4

–0.6

Cooling

Heating

230 240 250 260 270

220 K

220 KSuperlattice 300 K

DG

Superlattice 220 K

300 K

300 K

Temperature (K)

DS

C (

mW

mg–1

)

80

60

40

20

Tra

nsm

ittan

ce (

%)

Inte

nsity

(a.

u.)

Wavenumber (cm–1)

Wavenumber (cm–1)

1,000 2,000 3,000 4,000

80

60

40

20

Tra

nsm

ittan

ce (

%)

400 400600 800 1,000 800 1,200 1,600

Raman shift (cm–1)

Figure 4 | Optical and electrical properties of the superlattice nanosheets. (a) DSC thermogram, (b) temperature-dependent FTIR spectra indicate

that the superlattice nanosheets experience a first-order phase transition as temperature varies. (c) The magnified FTIR spectra. The temperature-

dependent FTIR spectra also imply the optical switching of the superlattice nanosheets in middle-wavelength IR regions. (d) Temperature-dependent

Raman spectra of the obtained superlattice.

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms4960 ARTICLE

NATURE COMMUNICATIONS | 5:3960 | DOI: 10.1038/ncomms4960 | www.nature.com/naturecommunications 5

& 2014 Macmillan Publishers Limited. All rights reserved.

materials are responsible for electrical conductivity and theirexcitation energies match with the light energies of the IRwavelength; thus the IR light is reflected off or absorbed ratherthan transmitted in a metallic (or narrow gap semiconductor)material. On the other hand, in isolating materials the excitationenergies are much larger than the energies of the IR light and thusthe IR spectrum is transmitted. The temperature-dependentresistivity curve in Supplementary Fig. 5 leads us to theconclusion that the superlattice nanosheets undergo a first-ordertransition at about 252 K from an insulating to a narrow-gapsemiconductor phase, accompanied by remarkable IR opticalswitch as temperature increases. Besides, the magnetism of thespace-confined synthesized superlattice nanosheets also showsemergent behaviour: an abrupt change in the magnetic suscept-ibility (Fig. 5a) was observed as the temperature decreased, whichis similar to the effect in VO2(R) (ref. 38). Considering that thehighly symmetric V–O framework in the superlattice nanosheetshas zigzag V–V chains with an isometric bond length of about2.77 Å, analogous to the linear V–V chain in VO2(R) by edgesharing of the regular VO6 octahedron, a similar rearrangementof V–V chains may take place in the superlattice nanosheets, andshould be responsible for the observed change in magnetism. Forcomparison, the M–T curve of VO2(B) synthesized from the sameraw materials in the open solution reaction system was alsomeasured, and the result (Fig. 5a) clearly showed no abruptchange in the magnetic susceptibility of VO2(B), consistent withthe M–T curve reported previously39. This result indicates thatthe atomic arrangement of the V–O framework in the superlatticenanosheets is indeed different from that of VO2(B), presumablybecause of the space-confinement effect of graphene.

The magnetocaloric effect of the superlattice. Generally, amagnetic substance shows magnetocaloric effect in a reversiblemagnetic susceptibility change experience through adiabatic

magnetization or demagnetization, which has been used as anenvironmentally friendly and highly efficient cooling technologyto replace conventional gas compression refrigeration during thepast decades40–44. In our present work, the field dependence ofmagnetization for the as-obtained superlattice nanosheets wasmeasured up to 1.5 T in the temperature range from 220 to 300 K,and the representative M–H curves are presented in Fig. 5b andSupplementary Fig. 6. The magnetization increases progressivelywith increasing temperature, presenting a step-like behaviour inthe temperature range covering the phase transition. This step-like behaviour in the M–H curve provides direct information onthe magnetocaloric effect because the magnetocaloric effect isrelated to the rearrangement of the magnetic moment undera magnetic field caused by either pure magnetic or latticetransformation. The magnetic entropy change can be calculatedfrom the M–H data on the basis of the Maxwell relationequation37:

DSðT;HÞ ¼ m0

Z H

0

@M@T

� �H

dH ð1Þ

The temperature dependence of DS was calculated and plottedin Fig. 5c. The maximum value of DS is about 0.4 J kg� 1 K� 1,which is remarkably larger than that of the VO2(M)–VO2(R)phase transition-driven magnetocaloric effect38. As is known, therapid change of magnetization would cause a large magneticentropy change, resulting in larger magnetocaloric effect. Theorientation of sharing V d-orbit electrons in a V–V chain shouldbe responsible for the magnetic transformation from magneticmoment disordering to ordering, which means a single-domainstructure could be expected to benefit from the rapidmagnetization change. In our single-crystalline superlatticenanosheets, the V–O frameworks are confined to the interlayerof multilayered graphene, and all of the V–O frameworks areparallel, which should form the single-domain structure to realize

aVO2(B) FC

b

c d

0.0022

0.0020

0.0018

0.0016

0.0014

0.0012150 180 210 240 270 300

Temperature (K)

Temperature (K)

264 K0.106

0.104

0.102

0.100

0.098

0.096

0.094

M (

emu

g–1)

ΔS (

JKg–1

K–1

)

M (

emu

g–1)

Inte

nsity

(a.

u.)

VO2(B) ZFCSuperlattice FCSuperlattice ZFC

234 K

230 K

220 K

ΔT =

3 K

300 K

260 K

240 K

220 K

180 K

150 K

14,000 14,200 14,400 14,600 14,800 15,000

H (Oe)

0.4

0.3

0.2

0.1

0.0220 230 240 250 260 270 280 6 7 8 9 10 20 30 40 50

2� (°)

Figure 5 | Magnetic properties of the superlattice nanosheets. (a) Temperature-dependent ZFC and FC magnetization in an applied field of 100 Oe for

the superlattice nanosheets and VO2(B). (b) The field-dependent magnetization plots at indicated temperatures. (c) Magnetic entropy change DS of the

superlattice nanosheets. (d) Temperature-dependent XRD patterns for the superlattice nanosheets.

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms4960

6 NATURE COMMUNICATIONS | 5:3960 | DOI: 10.1038/ncomms4960 | www.nature.com/naturecommunications

& 2014 Macmillan Publishers Limited. All rights reserved.

the rapid magnetization change resulting in the largermagnetocaloric effect. On the other hand, the structuraltransition-driven magnetocaloric materials usually show strongcoupling between crystallographic structure and magnetism; thus,any induced changes in the lattice entropies during the phasetransition should also contribute to the ultimate performance ofthe magnetocaloric effect45. In fact, recent reports havediscovered that the strain-mediated feedback from substratesnear the first-order phase transition can strongly enhance thetotal magnetocaloric effect46,47. In our present work, asaforementioned, the V–O frameworks are confined to theinterlayer of the multilayered graphene; thus, it is reasonable tobelieve that extrinsic strain feedback from the graphene layers tothe intercalated V–O frameworks should exist, which has beenverified through the temperature-dependent XRD. As seen fromFig. 5d, with decreasing temperature, the shift of the low-anglediffraction peak towards higher 2y values indicates that thedistance between two graphene layers is gradually shortened. Thismeans that the graphene layers exert extrinsic strain on the V–Oframeworks, resulting in an enhanced DS value compared to thatof our previously reported VO2(M)–VO2(R) phase transition-driven magnetocaloric effect.

DiscussionIn summary, the space-confined strategy within the interlayers ofmultilayered graphene enables the synthesis of freestanding andflexible superlattice nanosheets with special atomic configurationof the component to realize emerging physicochemical properties.In this approach, taking VO2 as an example, the graphene layersnot only provide a space-confined nanoreactor to avoid the fastaggregation of VO6 octahedra and direct the VO6 octahedra toconnect each other in a two-dimensional manner, but also serveas an electron donor to reduce the formation energy to obtainhighly symmetrical V–O frameworks. These highly symmetricalV–O frameworks endow the obtained superlattice nanosheetsemerging magnetocaloric effect compared to the VO2(B) sampleobtained from the same raw materials in the open solution-reaction system. Owing to the two-dimensional single-domainstructure of the V–O frameworks and extrinsic strain exerted bygraphene layers, the DS value of the superlattice nanosheets wassignificantly enhanced compared to that observed in theVO2(M)–VO2(R) phase transition. These flexible freestandingsuperlattice nanosheets synthesized from a low-cost and scalablemethod avoid the complex transferring process from growthsubstrates, as required in conventional synthesizing strategies,and thus can be adapted to a large variety of applications infunctionalized devices.

MethodsSynthesis of grapheme. Firstly, GO was prepared via a modified Hummer’smethod. After removing the salts and acids, the GO was ultrasonically treated inwater for over 1 h until it became clear without visible particles to form GOcollosol. Then the collosol was transferred into a 250-ml round-bottom flask,followed by addition of 0.031 ml hydrazine hydrate (85%) and 0.784 ml ammoniasolution (25%) into the flask under stirring. Then the solution was heated in an oilbath to 100 �C using a water-cooled condenser for 24 h. Finally, the as-obtainedRGO black collosol was cooled to room temperature for further intercalation.

Synthesis of superlattice nanosheets. The RGO precursor was exfoliated underultrasonic treatment for several hours to obtain small and thin RGO pieces, and anRGO solution was obtained: 0.18 mmol VO(acac)2 and 0.8 ml H2O2 were dissolvedin 30 ml acetone, then 4 ml of the RGO solution was added, subjected to magneticstirring for several minutes and then ultrasonicated for about half an hour tointercalate the vanadium source homogenously into the interlayer of adjacent RGOnanosheets. Then the well-distributed suspension was transferred into a 40 mlstainless Teflon-lined autoclave followed by solvothermal process at 180 �C for24 h. Afterwards the system was cooled to room temperature in air. The productwas collected by centrifugation at 10,000 r.p.m. for 2 min, washed with distilled

water and ethanol, and dried in vacuum at 60 �C for 8 h. For comparison, a productwithout graphene was prepared according to the above-mentioned process.

Characterization. The phase structure of the product was characterized by X-raypowder diffraction (XRD) using a Philips X’Pert Pro Super diffractometer with aCu Ka radiation source (l¼ 1.54178 Å). Field emission scanning electron micro-scopy (FESEM) images were obtained using a JEOL JSM-6700F scanning electronmicroscope. The TEM, HRTEM images and SAED analyses were captured by aJEOL JEM-ARF200F atomic resolution analytical microscope. Raman spectrumwas performed on a SPEX 1403 spectrometer equipped with an Arþ laser at anexcitation wavelength of 514.5 nm. The XPS spectrum was recorded by a VGES-CALAB MK8X-ray photoelectron spectrometer with an Al Ka. exciting source. Themagnetization was characterized by a superconducting quantum interferencedevice (SQUID, quantum design MPMS XL-7) magnetometer. The XAFS spectraof the superlattice nanosheets were measured at the U7C beamline of the NationalSynchrotron Radiation Laboratory.

References1. Smith, D. L. & Mailhiot, C. Theory of semiconductor superlattice electronic

structure. Rev. Mod. Phys. 62, 173–234 (1990).2. Luckyanova, M. N. et al. Coherent phonon heat conduction in superlattices.

Science 338, 936–939 (2012).3. Driza, N. et al. Long-range transfer of electron–phonon coupling in oxide

superlattices. Nat. Mater. 11, 675–681 (2011).4. Esaki, L., Chang, L. L., Howard, W. E. & Rideout, V. L. in Proc. 11th Int. Conf.

Physics of Semiconductors, 431 (Warsaw, Poland, 1972).5. Schuller, I. K. New class of layered materials. Phys. Rev. Lett. 44, 1597–1600

(1980).6. Klitzing, K. V., Dorda, G. & Pepper, M. New method for high-accuracy

determination of the fine-structure constant based on quantized Hall resistance.Phys. Rev. Lett. 45, 494–497 (1980).

7. Tsui, D. C. & Gossard, A. C. Resistance stand using quantization of the Hallresistance of GaAs/AlxGa1-xAs heterostructures. Appl. Phys. Lett. 38, 550–552(1981).

8. Bailich, M. N. et al. Giant magnetoresistance of (001)Fe/(001)Cr magneticsuperlattices. Phys. Rev. Lett. 61, 2472–2475 (1988).

9. Jin, S. et al. Thousandfold change in resistivity in magnetoresistiveLa-Ca-Mn-O films. Science 264, 413–415 (1994).

10. Bernevig, B. A., Hughes, T. L. & Zhang, S. C. Quantum spin Hall effect andtopological phase transition in HgTe quantum wells. Science 314, 1757–1761(2006).

11. Konig, M. et al. Quantum spin Hall insulator state in HgTe quantum well.Science 318, 766–770 (2007).

12. Garcia-Martin, S., King, G., Urones-Garrote, E., Nenert, G. & Woodward, P. M.Spontaneous superlattice formation in the doubly ordered perovskiteKLaMnWO6. Chem. Mater. 23, 163–170 (2011).

13. Stankovich, S. et al. Graphene-based composite materials. Nature 442, 282–286(2006).

14. Novoselov, K. S. et al. A roadmap for graphene. Nature 490, 192–200 (2012).15. Yankowitz, M. et al. Emergence of superlattice Dirac points in graphene on

hexagonal boron nitride. Nat. Phys. 8, 382–386 (2012).16. Hunt, B. et al. Massive Dirac Fermions and Hofstadter Butterfly in a van der

Waals heterostructure. Science 340, 1427–1430 (2013).17. Dean, C. R. et al. Hofstadter’s butterfly and the fractal quantum Hall effect in

moire superlattices. Nature 497, 598–602 (2013).18. Ponomarenko, A. L. et al. Cloning of Dirac fermions in graphene superlattices.

Nature 497, 598 (2013).19. Haigh, S. J. et al. Cross-sectional imaging of individual layers and buried

interfaces of graphene-based heterostructures and superlattices. Nat. Mater.11, 764–767 (2012).

20. Kasahara, S. et al. Electronic nematicity above the structural andsuperconducting transition in BaFe2(As1� xPx)2. Nature 486, 382–385 (2012).

21. Malliakas, C. D., Chung, D. Y., Claus, H. & Kanatzidis, M. G. Superconductivityin the Narrow-Gap Semiconductor CsBi4Te6. J. Am. Chem. Soc. 135,14540–14543 (2013).

22. Makiura, R. et al. Size-controlled stabilization of the superionic phase to roomtemperature in polymer-coated AgI nanoparticles. Nat. Mater. 8, 476–480 (2009).

23. Xiao, C. et al. Superionic Phase Transition in Silver Chalcogenide NanocrystalsRealizing Optimized Thermoelectric Performance. J. Am. Chem. Soc. 134,4287–4293 (2012).

24. Liu, J., Gottschall, T., Skokov, K. P., Moore, J. D. & Gutfleisch, O. Giantmagnetocaloric effect driven by structural transition. Nat. Mater. 11, 620–626(2012).

25. Mozharivskyj, Y., Pecharsky, A. O., Pecharsky, V. K. & Miller, G. J. On thehigh-temperature phase transition of Gd5Si2Ge2. J. Am. Chem. Soc. 127,317–324 (2005).

26. Ohkoshi, S.-I. et al. Synthesis of a metal oxide with a room-temperaturephotoreversible phase transition. Nat. Chem. 2, 539–545 (2010).

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms4960 ARTICLE

NATURE COMMUNICATIONS | 5:3960 | DOI: 10.1038/ncomms4960 | www.nature.com/naturecommunications 7

& 2014 Macmillan Publishers Limited. All rights reserved.

27. Xiao, C. et al. High Thermoelectric and Reversible p-n-p Conduction TypeSwitching Integrated in Dimetal Chalcogenide. J. Am. Chem. Soc. 134,18460–18466 (2012).

28. Wu, C. Z., Feng, F. & Xie, Y. Design of vanadium oxide structures withcontrollable electrical properties for energy applications. Chem. Soc. Rev. 42,5157–5183 (2013).

29. Zhang, S. D., Li, Y. M., Wu, C. Z., Zheng, F. & Xie, Y. Novel flowerlikemetastable vanadium dioxide (B) micronanostructures: facile synthesis andapplication in aqueous lithium ion batteries. J. Phys. Chem. C 113,15058–15067 (2009).

30. Wu, C. Z. et al. Direct confined-space combustion forming monoclinicvanadium dioxides. Angew. Chem. Int. Ed. 49, 134–137 (2010).

31. Wu, C. Z. et al. Hydrogen-incorporation stabilization of metallic VO2(R) phaseto room temperature, displaying promising low-temperature thermoelectriceffect. J. Am. Chem. Soc. 133, 13798–13801 (2011).

32. Liu, L. et al. Room-temperature intercalation-deintercalation strategy towardsVO2(B) single layers with atomic thickness. Small 8, 3752–3756 (2012).

33. Grinolds, M. S., Lobastov, V. A., Weissenrieder, J. & Zewail, A. H.Four-dimensional ultrafast electron microscopy of phase transitions. Proc. NatlAcad. Sci. USA 103, 18427–18431 (2006).

34. Koski, K. J. et al. Chemical intercalation of zerovalent metals into 2D layeredBi2Se3 nanoribbons. J. Am. Chem. Soc. 134, 13773–13779 (2012).

35. Gu, J. et al. Liquid-phase syntheses and material properties of two-dimensionalnanocrystals of rare earth-selenium compound containing planar Se layers:RESe2 nanosheets and RE4O4Se3 nanoplates. J. Am. Chem. Soc. 135, 8363–8371(2013).

36. Huang, Y. L. et al. Effect of extended polymer chains on properties oftransparent graphene nanosheets conductive film. J. Mater. Chem. 21,18236–18241 (2011).

37. Yao, T. et al. Ultrathin nanosheets of half-metallic monoclinic vanadiumdioxide with a thermally induced phase transition. Angew. Chem. Int. Ed. 52,7554–7558 (2013).

38. Wu, C. Z. et al. Direct hydrothermal synthesis of monoclinic VO2(M)single-domain nanorods on large scale displaying magnetocaloric effect.J. Mater. Chem. 21, 4509–4517 (2011).

39. Liu, J. F. et al. Metastable vanadium dioxide nanobelts: hydrothermal synthesis,electrical transport, and magnetic properties. Angew. Chem. Int. Ed. 43,5048–5052 (2004).

40. De Campos, A. et al. Ambient pressure colossal magnetocaloric effect tuned bycomposition in Mn1� xFexAs. Nat. Mater. 5, 802–804 (2006).

41. Tan, X. Y., Chai, P., Thompson, C. M. & Shatruk, M. Magnetocaloric effect inAlFe2B2: toward magnetic refrigerants from earth-abundant elements. J. Am.Chem. Soc. 135, 9553–9557 (2013).

42. Svitlyk, V., Miller, G. J. & Mozharivskyj, Y. Gd5Si4� xPx: targeted structuralchanges through increase in valence electron count. J. Am. Chem. Soc. 131,2367–2374 (2009).

43. Manoli, M. et al. Mixed-valent Mn supertetrahedra and planar discs asenhanced magnetic coolers. J. Am. Chem. Soc. 130, 11129–11139 (2008).

44. Evans, M. J., Holland, G. P., Garcia-Garcia, F. J. & Haussermann, U.Polyanionic gallium hydrides from AlB2-type precursors AeGaE (Ae¼Ca,Sr, Ba; E¼ Si, Ge, Sn). J. Am. Chem. Soc. 130, 12139–12147 (2008).

45. Gschneidner, K. A., Pecharsky, V. K. & Tsokol, A. O. Recent developments inmagnetocaloric materials. Rep. Prog. Phys. 68, 1479–1539 (2005).

46. Moya, X. et al. Giant and reversible extrinsic magnetocaloric effects inLa0.7Ca0.3MnO3 films due to strain. Nat. Mater. 12, 52–58 (2013).

47. Mosca, D. H., Vidal, F. & Etgens, V. H. Strain engineering of themagnetocaloric effect in MnAs epilayers. Phys. Rev. Lett. 101, 125503 (2008).

AcknowledgementsThis work was financially supported by National Natural Science Foundation of China(11079004, 90922016, 11135008, 21331005, and 11321503), innovation project of theChinese Academy of Science (XDB01020300) and the Fundamental Research Funds forthe Central University (WK 2060190020).

Author contributionsH.Z., C.X., S.W. and Y.X. conceived the idea and co-wrote the paper. H.Z., C.X., X.Z.,Z.L., X.Z. and C.W. carried out the example synthesis, characterization and magneto-caloric effect measurements. H.C., T.Y. and S.W. carried out the XAFS measurementsand analysed the XAFS data. All the authors discussed the results, commented on andrevised the manuscript. H.Z. and C.X. contributed equally to this research.

Additional informationSupplementary Information accompanies this paper at http://www.nature.com/naturecommunications

Competing financial interests: The authors declare no competing financial interests.

Reprints and permission information is available online at http://npg.nature.com/reprintsandpermissions/

How to cite this article: Zhu, H. et al. Magnetocaloric effects in a freestandingand flexible graphene-based superlattice synthesized with a spatially confined reaction.Nat. Commun. 5:3960 doi: 10.1038/ncomms4960 (2014).

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms4960

8 NATURE COMMUNICATIONS | 5:3960 | DOI: 10.1038/ncomms4960 | www.nature.com/naturecommunications

& 2014 Macmillan Publishers Limited. All rights reserved.