Nanoscale

PAPER

Cite this: Nanoscale, 2018, 10, 22429

Received 24th August 2018,Accepted 2nd November 2018

DOI: 10.1039/c8nr06854b

rsc.li/nanoscale

Low temperature solution synthesis of reducedtwo dimensional Ti3C2 MXenes with paramagneticbehaviour†

Yeoheung Yoon, ‡a,b Thi Anh Le,‡a,c Anand P. Tiwari,a,b Ikjoon Kim,c

Michel W. Barsoum*d and Hyoyoung Lee *a,c

MXenes – two dimensional, 2D, early transition metal, M, carbides and nitrides, X – are the latest addition

to the 2D materials’ world. Herein, we report on a facile low temperature solution chemical synthesis

method to reduce Ti3C2Tx multilayered, ML, MXenes. Using X-ray photoelectron spectroscopy, electron

spin resonance, magnetization measurements and other techniques, we concluded that immersing

Ti3C2Tx MLs in the reducing agent Li-ethylenediamine (Li-EDA) – held at temperatures varying from room

to 120 °C – reduces the 2D layers creating Ti3+ ions and oxygen vacancies. Above a temperature (T ) of

≈10 K, the magnetic susceptibilities, χ, are temperature independent, implying that the resulting powdersare Pauli paramagnetic. The loss of the magnetic signal upon intercalation of Li+ or EDA, together with a

Curie-like increase in χ at T < 10 K, is consistent with that of a disordered metal that is close to a metallic

to insulator transition and proves that the magnetism is associated with the 2D flakes. This result is the

first evidence of any magnetism of any MXene.

I. Introduction

Since the successful separation of graphene from graphite,1

interest in other two-dimensional (2D) materials such as hex-agonal boron nitride (h-BN),2 transition metal dichalcogenides(TMDs)3 and, more recently, transition metal carbides and car-bonitride labelled MXenes has increased significantly.4–6

MXenes are a new family of 2D materials that are synthesizedvia selective etching of the “A”-layers by HF, or other fluorinecontaining salts, from the MAX phases (Mn+1AXn, where n = 1,2, or 3). The latter are layered ternary carbides and nitrides,where M represents a transition metal, A is typically a IIIA orIVA element (e.g. Al, Si, etc.); and X is C and/or N.7 When theA-layers are selectively etched, they are replaced with surface

terminations. Thus the correct formula for MXenes isMn+1XnTx, where Tx represents various surface terminationfunctional groups, including O, OH, and F.8,9

MXenes exhibit unusual characteristics and properties. Forexample, they combine hydrophilicity and metallic-likeconductivity.4–6 This unique combination can be further tunedby manipulating their composition and/or surface chemistries,which is one reason they have attracted wide interest.6,10,11

Another important reason is that they are seriously being con-sidered in a large and diverse host of applications. Thenumber of papers published and number of citations referringto these materials are both growing exponentially. The same istrue for theoretical studies on their properties and potentialapplications since their discovery in 2011.11–14 The gist ofmany of these theoretical papers is that it is possible to dra-matically alter the properties. For example, it has been pre-dicted that some non-terminated MXenes would bemagnetic.11 Concomitantly, all studies predict MXenes, withsurface terminations, to be non-magnetic.11 However, since itis almost impossible to maintain un-terminated nanosheets –especially under ambient atmospheres – due to their reactivity,other approaches must be found to change/control their ter-minations and structures.

One obvious approach is to reduce the MXene flakes. Inone of the earliest attempts, Mashtalir et al. exposed Ti3C2Txmultilayers, MLs, to nitrogen-containing organic molecules

†Electronic supplementary information (ESI) available. See DOI: 10.1039/c8nr06854b‡These authors contributed equally to this work.

aCenter for Integrated Nanostructure Physics, Institute for Basic Science (IBS),

Sungkyunkwan University, Suwon 440-746, South Korea.

E-mail: hyoyoung@skku.edubDepartment of Material Science & Engineering, Korea Advanced Institute of Science

and Technology (KAIST), Daejeon 34141, South KoreacDepartment of Chemistry and Department of Energy Science,

Sungkyunkwan University, Suwon 440-746, South KoreadDepartment of Materials Science and Engineering, Drexel University, Philadelphia,

Pennsylvania 19104, USA. E-mail: barsoumw@drexel.edu

This journal is © The Royal Society of Chemistry 2018 Nanoscale, 2018, 10, 22429–22438 | 22429

Publ

ishe

d on

07

Nov

embe

r 20

18. D

ownl

oade

d by

Sun

gkyu

nkw

an U

nive

rsity

on

12/1

8/20

18 1

2:25

:52

PM.

View Article OnlineView Journal | View Issue

www.rsc.li/nanoscalehttp://orcid.org/0000-0001-6213-2680http://orcid.org/0000-0002-8031-0791http://crossmark.crossref.org/dialog/?doi=10.1039/c8nr06854b&domain=pdf&date_stamp=2018-11-29http://dx.doi.org/10.1039/c8nr06854bhttps://pubs.rsc.org/en/journals/journal/NRhttps://pubs.rsc.org/en/journals/journal/NR?issueid=NR010047

including hydrazine (N2H4), urea, and dimethylsulfoxide(DMSO).15 The original purpose of that effort was to reducethe surfaces as is done for reduced graphene oxides (rGOs).16

Instead of reducing the surfaces, however, the hydrazine, andother small organic molecules are simply intercalated betweenthe layers. And while this approach allows for the delamina-tion of the MXene layers, whether it also reduced the Ti–Csheets was not investigated.15

Herein we reduced Ti3C2 and to understand what occursupon reduction it is instructive to review what is known aboutreduced titania, r-TiO2. By now it is reasonably well establishedthat reducing TiO2 results in the formation of oxygenvacancies, VO.

17,18 For example, heating TiO2 in the presenceof urea reduced some of the Ti4+ ions to Ti3+ and createdoxygen vacancies.19,20 Over the years there have been multipleapproaches to reduce TiO2, which ranged from using metalssuch as Al, Zn, Fe, etc., to heating under reducing atmos-pheres, or exposure to hydrazine hydrate, urea, etc.18 Twodecades ago, Diebold and co-workers used atomic-resolutionscanning tunnelling microscopy (STM) to directly evidenceoxygen vacancies, VO, generated under vacuum on a TiO2 (110)surface.21,22 Early on, DFT calculations have shown that theremoval of neutral, bridging oxygen atoms from rutile TiO2(110) surfaces generated two unpaired electrons that were loca-lized at the 3d orbitals of neighbouring 5-coordinated Tiatoms.23 Note that in this work we adopt the Kroger–Vinknotation in which oxygen vacancies with two, one and no loca-lized electrons are denoted as Vo

x, V•O, and V••O, respectively.

Furthermore, if the electrons are localized on a Ti4+ site, thelatter becomes Ti3+, that – in the Kroger–Vink notation – isdescribed as Ti′Ti There is a quite recent comprehensive reviewon reduced TiO2,

18 in which the fundamental role of VO inreduced TiO2 is outlined and documented in much moredetail. For our purposes, what is important is that reducingTiO2 results in oxygen vacancies.

Herein, we present a new synthetic route to reduce 2DTi3C2Tx MXenes – henceforth referred to as r-Ti3C2Tx – via afacile, low temperature solution process. The resultingmaterial is Pauli paramagnetic, which constitutes in turn, thefirst evidence of magnetism of any kind in a MXene. Forthe reduction, we chose the “super-base”, lithium-ethylenediamine (Li-EDA). This solution-based process is facile andamenable to mass production. In addition, since the parentMAX phases are Pauli paramagnets,7 we expect that the chemi-cally reduced Ti3C2Tx would provide a strong paramagneticsignal. In addition, we show that by controlling the reductiontemperature, we can “dial in” the density of states, DOS, at theFermi level, EF.

This paper is structured as follows: First, we describe theprocess make and reduce Ti3C2Tx flakes in the presence of Li-EDA. We then characterize the resulting material after washingwith water and ethanol using a number of techniques includ-ing X-ray diffraction, XRD, Raman spectroscopy, electron spinresonance, ESR, magnetization measurements as a function oftemperature, scanning and transmission electron microscopyand X-ray photoemission spectroscopy (XPS). XPS spectra

provide strong evidence for the formation of Ti3+ ions and, byextension, oxygen vacancies. The ESR and magnetizationmeasurements show the r-Ti3C2Tx to be Pauli paramagnetic.Finally, we show that our results are consistent with whatwould be expected for a disordered metal, close to a metallicto insulator, M–I, transition, where EF is near the mobilityedge.

II. ExperimentalA. Sample preparation

To prepare Ti3AlC2, Ti2AlC (Kanthal) and TiC (>99%, Alfa Aesar)were mixed using a ball mill for 24 h at a molar ratio of1 : 1 (after adjusting ∼34 wt% Ti3AlC2 present in the commer-cial powder).4 The mixed powders were then heated to 1350 °Cfor 2 h under argon and were held at temperature for 0.3 hresulting in lightly sintered brick. Finally, the Ti3AlC2 wascrushed into powder by ball milling for further reactionsand characterization. To synthesize the Ti3C2Tx, we carefullyfollowed previous reaction conditions described elsewhere.4 Inshort, the Ti3AlC2 powders are immersed in a 50% concen-trated HF solution at room temperature, RT, for 2 h. Theresulting suspension was then washed more than 5 times withDI water, and centrifuged to separate the supernatant andobtain Ti3C2Tx powders.

Synthesis of r-Ti3C2Tx from Ti3C2Tx by lithium ethylenedia-mine. Because Li and EDA are quite sensitive to moisture,every piece of glassware used was fully dried in a drying ovento remove residual water molecules. Then, 100 mg of Ti3C2Txand 70 mg of Li metal (Sigma Aldrich) were carefully placed ina flask under a N2 atmosphere in a glovebox. Then, 100 mg ofEDA (Sigma Aldrich) was injected slowly over 20–30 min.During this stage, the mixture was sonicated to produce ahomogeneous solution and to ensure that the Li metaldissolved in the EDA solution. Once the Li metal was welldissolved, the reaction solution turned dark blue. This changein colour was taken as evidence for the formation of electridesalts. Subsequently, the reaction solution was stirred for 24 hunder an inert N2 atmosphere at 25, 40, 80 or 120 °C. Afterthe reaction was complete, deionized, DI, water and ethanolwere carefully added to the reacted solution to neutralize itand remove the Li ions intercalated between the Ti3C2Txnanosheets. This process was repeated several times. Theresulting ML samples were dried and stored under vacuum inan oven at 60 °C for further characterization and use.

Delamination of r-Ti3C2Tx by lithium ethylenediamine. Thisprocedure is identical to the synthesis described above exceptthat the r-Ti3C2Tx MLs were delaminated after the reaction wascomplete. In this case, only DI water was added to the solu-tion. The solution was sonicated for 6 h and centrifuged, andthe supernatant was decanted. The product was washed withDI water several times to remove any Li or EDA residues.Finally, the supernatant, containing mostly delaminatedMXene flakes, was filtered through a 200 nm anodisk toprepare free-standing r-Ti3C2Tx films.

Paper Nanoscale

22430 | Nanoscale, 2018, 10, 22429–22438 This journal is © The Royal Society of Chemistry 2018

Publ

ishe

d on

07

Nov

embe

r 20

18. D

ownl

oade

d by

Sun

gkyu

nkw

an U

nive

rsity

on

12/1

8/20

18 1

2:25

:52

PM.

View Article Online

http://dx.doi.org/10.1039/c8nr06854b

B. Chemical and physical characterization

The samples’ microstructures were observed using a fieldemission scanning electron microscope (FE-SEM, JSM-6701F/INCA Energy, JEOL) and a transmission electron microscope(JEOL JEM 3010). Atomic force microscopy, AFM, was per-formed using a SPA400 instrument with an SPI-3800 controller(Seiko Instrument Industry Co.) at RT. XRD patterns wereobtained using a Rigaku Ultima IV X-ray diffractometer withCu Kα radiation at a scanning rate of 5° min−1. All XPSmeasurements were performed on a Thermo VG MicrotechESCA 2000 with a monochromatic Al KR X-ray source at 100W. Raman spectroscopy measurements were obtained using amicro-Raman system (Renishaw, RM1000-In Via) with an exci-tation energy of 2.41 eV (514 nm). The thermal properties ofthe materials were characterized by thermogravimetric ana-lysis, TGA (Polymer laboratory, TGA 1000 plus).

The magnetic properties were measured using: (i) ESR spec-troscopy (CW-EPR, Bruker EMX plus 6/1 spectrometer) at RTwith a microwave power of 0.94 mW, a microwave frequency of9.64 GHz, a modulation frequency of 100 kHz, and a modu-lation amplitude of 10 G and (ii) a superconducting quantuminterference device (SQUID) and a vibrating sample magnet-ometer (VSM). The magnetization, M, vs. magnetic field, H,measurement was carried out at RT under 30 Oe.

III. Results and discussion

The details of the operative reaction mechanisms can befound in the ESI.† Fig. 1 schematically summarizes the stepsto produce r-Ti3C2Tx using Li-EDA that is a well-known interca-lant as a reducing agent and an exfoliator for 2Dmaterials.15,24–26 Besides forming free standing films, thistreatment also creates defects (oxygen and/or carbonvacancies) in the MXene layers, as well as, increases the DOS atthe Fermi level (see below).

The SEM images of as-etched Ti3C2Tx, r-Ti3C2Tx andr-Ti3C2Tx after delamination are presented in Fig. 2A, B, and C,respectively. These micrographs are typical of MXenes after

etching of Al from the parent MAX phase using HF. Note thatthe layers are retained after Li-EDA treatment (Fig. 2B), whichimplies that the reduction process has no significant effect onthe overall MXene morphology. This conclusion is further con-firmed by SEM images of freestanding filtered films (Fig. 2Dand E) that are typical of such MXene films. The filmsproduced were ≈2 µm thick measured by SEM and with a flexi-bility (Fig. 2F) comparable to that of graphene oxide16,26 andother MXene filtered films.25,27,34

The 2D nature of the flakes and their purity were furtherelucidated by TEM. As shown in Fig. 2G and Fig. S1A and S1B†the as-reduced and the delaminated MXene nanosheets existas single or few stacked layers. The high-resolution (HRTEM)analysis shows clear lattice fringes of ca. 0.26 nm that areidentified for the (1̄010) (Fig. 2H) and the correspondingselected area electron diffraction, SAED, (inset in Fig. 2H) con-firms that the hexagonal structure of the parent MAX phase ismaintained after reduction. In particular, the space para-meters of r-Ti3C2Tx attributed to (2̄110), (1̄100) and (1̄010)planes are displayed clearly in the SAED.

AFM of flakes deposited on SiO2 substrates (Fig. 2I) showsthat, at least, some of the ML flakes have lateral dimension ofthe order of a few micrometres, while being a few nm thick.

Recent XPS results28 have shown that the Ti oxidation statesin multilayered Ti3C2Tx are far from simple. Peaks in the rangeof 454.9 eV to 458.6 eV were identified and assigned Ti, Ti2+,Ti3+ and Ti4+. The Ti-peak – located at the same position as theTi peak in Ti3AlC2 – viz. 455 eV – was assigned to Ti bonded toC atoms.28 The Ti3+ and Ti4+ peaks are located at 457.9 eV and458.9 eV, respectively.28,29 Herein we unambiguously show thatFig. 1 Scheme of synthesis of r-Ti3C2Tx by exposure to Li-EDA.

Fig. 2 Characterization of r-Ti3C2Tx, (A, B, and C) SEM images ofTi3C2Tx after HF treatment, r-Ti3C2Tx after Li-EDA reduction (before thedelamination process) and r-Ti3C2Tx after the delamination process,respectively. (D and E) Cross-sectional SEM images of delaminatedr-Ti3C2Tx films produced by filtration. (F) Picture of delaminatedr-Ti3C2Tx film and (G and H) TEM images of delaminated r-Ti3C2Tx flakes(inset shows the selected area electron diffraction (SAED) image shownin H). (I) Typical AFM image of delaminated r-Ti3C2Tx flakes.

Nanoscale Paper

This journal is © The Royal Society of Chemistry 2018 Nanoscale, 2018, 10, 22429–22438 | 22431

Publ

ishe

d on

07

Nov

embe

r 20

18. D

ownl

oade

d by

Sun

gkyu

nkw

an U

nive

rsity

on

12/1

8/20

18 1

2:25

:52

PM.

View Article Online

http://dx.doi.org/10.1039/c8nr06854b

the Ti3+ peak in r-Ti3C2Tx occurs at 458.0 ± 0.2 eV and the Ti4+

peak at 459 ± 0.2 eV. It should be noted that the slight differ-ences of the peak positions of Ti3+ and Ti4+ in Ref. 28, 29 andour work are most probably due to the sputtering. Halim et al.showed that sputtering can slightly shift the peak positions.The overall XPS spectra, over the full range of energy, areshown in Fig. S2A.† The spectra of the Ti 2p-region of Ti3AlC2,Ti3C2Tx and r-Ti3C2Tx, reduced at 25 °C, 40 °C, 80 °C and120 °C, are shown in Fig. 3A. The intensities are normalized tothe Ti–C peak, labeled on the left. A perusal of these resultsclearly shows that: (1) there are three main peaks labelled fromleft to right as Ti–C, Ti3+ and Ti4+. A shoulder between the Ti–Cpeak which is almost at the same energy as that of Ti3AlC2

28,30,31

and the Ti3+ peak presumably due to Ti2+ is clearly observed insome of the spectra. Note that the etching broadens this peak,presumably due to the introduction of defects. (2) The Ti4+ peakfor Ti3AlC2 is due to a surface oxide layer. It is also present forTi3C2Tx but conspicuously absent from samples reduced at 25,40 and 80 °C. That peak, however, is prominent for the samplereduced at 120 °C. (3) The Ti3+ peak, on the other hand, increasesmonotonically with increasing reduction temperature. (4) Peaksat 461 eV and higher – denoted by arrows – belong to 2p1/2. Itfollows that all major peaks and prominent shoulders can beassigned to Ti–C, Ti3+ and/or Ti4+.

Most of our peak assignments agree with previous work onTi3C2Tx.

28,30,32 In previous work, however, a Ti3+ peak was notdirectly observable; it was deduced via deconvolution of amuch wider peak and thus subject to some subjectiveinterpretation. It follows that an important result obtained hereis the clear and unambiguous assignment of Ti3+ in Ti3C2Tx to458.0 ± 0.2 eV. As discussed below, the correlation between

the intensity of this peak – that increases monotonicallywith increasing reduction temperature – and the paramagnet-ism is good. More specifically, the ratio of the intensities ofthe Ti3+ and Ti–C peaks (Ti3+/Ti–C) after reduction at 25 °C,40 °C and 80 °C was 0.5 ± 0.2, 0.6 ± 0.2 and 0.9 ± 0.2(see Fig. S3A, B and C†). The results after reduction at 120 °Cwere more scattered (Fig. S3D†). The exact reason for thescatter is unclear at this time, but as discussed below, thesamples reduced at 120 °C were more prone to oxidation andindeed the increase in the peak intensities of Ti4+ at 459 eV isevidence for oxidation. Furthermore, it is obviously seen inFig. S4B† that the peak intensities Ti2+ between Ti–C and Ti3+

gradually decreased as the reduction temperature increased.More specifically, the Ti2+ concentration percentage forr-Ti3C2Tx reduced at RT, 40, 80 and 120 °C samples is calcu-lated to be 17.7 at%, 11.8 at%, 10.0 at% and 8.5 at%, respect-ively. Note that the oxidation does not occur during thereduction process, but later when the samples are exposed toair. This comment notwithstanding, in all cases, the intensi-ties of the peaks or shoulder at 458 eV, corresponding to Ti3+,are significantly higher than the Ti–C peaks or the shoulder at≈455 eV. Note that the ESR signals of the three samplesmeasured in XPS were quite comparable and strong (Fig. S3E†)and thus the oxidation observed most probably only affectedthe outermost layers probed by the XPS.

To sum up the Ti XPS results: a strong correlation wasfound between the reduction temperature and the intensitiesof the Ti3+ peaks. The simplest interpretation is that the Li-EDA treatment reduces some Ti4+ into Ti3+. It is important tonote in this context that for there to be something to reduce,Ti4+ must have been incorporated into the etched MXenes tobegin with and/or C-vacancies are formed (see below). Whenthe O 1s spectra for the MXenes tested herein are compared(Fig. 3B) it is obvious that increasing the reduction tempera-ture shifts the center of gravity of the spectra to higher energy,confirming the fact that Li-EDA treatment indeed reducesTi3C2Tx. However, the shoulder for the samples reduced at120 °C is consistent with our conjecture that this sample waspartially oxidized. Fig. S2B† plots the C 1s XPS spectra as afunction of etching and reduction temperature, from which itis clear that the peak intensity ratio of the C atoms associatedwith Ti3C2Tx (viz. 282 eV) and all other C components such asC–C (248.8 ± 0.2 eV), CHx (285.5 ± 0.2 eV), C–O (286.8 ± 0.2 eV)and COOH (289.0 ± 0.2 eV) decreases – from about 1.8 to 0.9 –with increasing reduction temperatures (see Table S1†). Alsonoteworthy is the gradual increase in the intensity of the gra-phitic C–C peak. The intensity of the CHx peak first increasesup to a reduction temperature of 80 °C before ultimatelydecreasing after reduction at 120 °C (Fig. S2B†). In contrast,the intensities of the C–O (286.8 ± 0.2 eV) and COOH (289.0 ±0.2 eV) peaks remain more or less constant. In summary, it isreasonable to assume that during the reduction process,C atoms – that subsequently react with oxygen to form hydro-carbons, or react with other C atoms to form graphitic sheets –are produced. It is currently well established, and confirmedherein, that etching of Ti3AlC2 in HF results in a mixture of O,

Fig. 3 Effects of Li-EDA treatments, at various temperatures, oncharacteristics of Ti3AlC2, Ti3C2Tx and r-Ti3C2Tx: (A) XPS spectra of the Ti2p region of Ti3AlC2, Ti3C2Tx, and r-Ti3C2Tx samples after Li-EDAreduction at temperatures indicated. (B) Same as A, with focus on the O1s region. (C) Raman spectra of Ti3AlC2, Ti3C2Tx, and r-Ti3C2Tx after Li-EDA treatment at 80 °C. (D) XRD patterns of Ti3C2Tx and ML-r-Ti3C2Txafter Li-EDA treatment, and de-r-Ti3C2Tx after delamination.

Paper Nanoscale

22432 | Nanoscale, 2018, 10, 22429–22438 This journal is © The Royal Society of Chemistry 2018

Publ

ishe

d on

07

Nov

embe

r 20

18. D

ownl

oade

d by

Sun

gkyu

nkw

an U

nive

rsity

on

12/1

8/20

18 1

2:25

:52

PM.

View Article Online

http://dx.doi.org/10.1039/c8nr06854b

OH and F-terminations as shown in Fig. S2A.†8,9,15,28

Increasing the reduction temperature, however, replacesF-terminations with O or OH terminations. This is best seen inFig. S2C† that shows that after reduction at 120 °C, the F 1ssignal almost disappears. This result is not too surprisingsince it is established that the F-terminations are less stablethan the O and/or the OH terminations.8 A small N 1s peakalso appeared around 400 eV due to EDA molecules inter-calated after Li-EDA treatment as shown in Fig. S2A.†15,25

Raman spectra of Ti3AlC2, Ti3C2Tx, and r-Ti3C2Tx are com-pared in Fig. 3C. The three main Raman peaks located at ca.260, 400, and 605 cm−1 can be assigned to the Eg(1); A1g andB1g(2); and Eg(2) vibration modes of in-plane vibration ofcarbon and oxygen bonds and Ti–C vibrations, respectively.4,45

The Raman spectra of Ti3AlC2 and Ti3C2Tx are comparable,showing almost no changes in the peak position. But, the peakintensities of Ti3C2Tx (as compared with Ti3AlC2 at 400 and605 cm−1) were weaker and noisier as a result of etching.4

More interestingly, compared to Ti3AlC2, Ti3C2Tx the Ramanpeak of r-Ti3C2Tx at around 145 cm

−1 is new, and can possiblybe attributed to the vibration of anatase, indicating thatr-Ti3C2Tx are prone more to oxidation than Ti3AlC2 orTi3C2Tx.

45,46 Furthermore, the peak at around 400 cm−1 wasshifted to ∼430 cm−1 which can be attributed to the thinnessof nanosheets during reduction.33,47 The two broad peaks at∼1390 cm−1 and ∼1590 cm−1 are characteristic of the D- andG-modes of graphitic carbon. The ratio of the D- and G-bandintensities, ID/IG, was ∼1.12. The broadening of G- andD-bands and the fact that the ID/IG ratio >1 (Fig. S2D†) areevidence for highly disordered or oxidized C, similar to whathas been reported for graphene oxide.16

Fig. S4A† compares the weight losses when the variouspowders were heated to 800 °C under a N2 atmosphere. TheHF-etched Ti3C2Tx exhibited a small (≈3%) weight loss, attrib-uted to the elimination of interlamellar water and possibly theloss of some F associated with the terminations. The weightlosses of the r-Ti3C2Tx films gradually increased as thereduction temperature increased up to ≈600 °C, before increas-ing again. The initial weight losses in the r-Ti3C2Tx samplesare higher than those in the unreduced material suggestingthat the inter-lamellar water is less tightly bound in the formerthan the latter. The weight gain is most probably associatedwith the reaction of the powders with residual O2 or N2, and isconsistent with the notion that the reduced powders are morereactive than their un-reduced counterparts.35,44

We now shift our attention to what occurs to the MXenestructure as a result of Li-EDA treatment followed by washingwith water. The effect of the latter is best seen in Fig. 3D,where the XRD patterns of Ti3C2Tx, r-Ti3C2Tx ML and filtereddelaminated films are compared. After HF etching, the typicalXRD peaks ((002), (006), (008), (010), (012) and (110)) of theTi3C2Tx pattern are obtained, wherein the diffraction peaks inthe 35–45° range broaden significantly. More importantly, the0002 peak shifts downward to a 2θ = 9.02°, and broadens.4,5,15

The latter signifies a c-lattice parameter, c-LP, of about 19.5 ±0.1 Å which is typical of HF-etched Ti3AlC2.

4,15 After the

Li-EDA treatment (middle pattern in Fig. 3D), the (0002) peakshifts down further to 2θ = 6.5° that corresponds to a c-LP of27.1 ± 0.1 Å, more typical of Ti3AlC2 etched in the presence ofLi.32 In the XRD patterns of the films, composed of mostlydelaminated flakes, the (0002) peak downshifts further to 2θ =6.28°, corresponding to a c-LP of 28.1 ± 0.3 Å. It follows thatexposing HF-etched Ti3AlC2 to Li-EDA followed by washing inwater results in the intercalation of, at least Li+ ions and theirwater of hydration, between the layers. Note the disappearanceof the peak at 61°2θ, in the free-standing films. Previously, itwas assumed that the loss of this peak signified complete dela-mination and the formation of a turbostratic MXene.15 Morerecent work36 has shown that not to be the case; the loss ofthis peak simply reflects a better alignment of the basal planeson the substrate as a result of the filtration process. XPS in theLi 1s region confirmed that Li was intercalated between thelayers (bottom spectrum in Fig. S2E†). Two peaks, one at54.2 ± 0.2 eV (red curve) and the other at 55.9 ± 0.2 eV (bluecurve), were fitted. The first we assign to Li cations betweenthe layers, and the second is due to LiF salts that are left overfrom the etching process.32 Note that when these films arewashed, the Li peaks disappeared (top spectrum in Fig. S2E†).It follows that simply washing the MLs, in excess water,replaces Li by, presumably, hydronium ions.

Finally we turn to the most important result obtainedherein, which is the appearance of a paramagnetic signal as aresult of the reduction. EPR is a highly sensitive techniquethat allows for the investigation of paramagnetic specieshaving one, or more, unpaired electrons either in bulk or solu-tions.37 Fig. 4A compares the RT EPR spectra – observed at amagnetic field of 3447 ± 1 G – of Ti3AlC2, HF-etched Ti3C2Txand r-Ti3C2Tx films as a function of reduction temperature.Why the Ti3AlC2 sample did not show an EPR signal isunclear, since it is well established that most MAX phases arePauli-paramagnets.7 For example, the magnetic susceptibility,χ, of Ti3SiC2 was found to be ≈4 × 106 and temperature inde-pendent. In contrast, the EPR signal of the HF-etched Ti3C2Txwas weak, but measurable (inset in Fig. 4A). In addition, as thereduction temperature increased from 25 to 120 °C, the EPRpeaks became stronger, sharper and more symmetric (Fig. 4A).Sharp EPR spectra of all r-Ti3C2Tx were observed at a reso-nance field of Bres = 3447 ± 1 G, with a peak-to-peak line widthof ΔBpp = 10 G. The EPR spectra of r-Ti3C2Tx reduced at 40, 80and 120 °C showed a signal at g = 1.9998, 1.9993 and 1.9981,respectively. The fact that g is essentially 2 is important in thiscontext because it is the value expected for free electrons. Saidotherwise, as discussed in more detail below, the magnetismdoes not originate from the Ti3+ ions – for which g rangesfrom 1.99 to 1.98817 – but rather from delocalized electrons.

Before proceeding further, it is important to describe whatstructural changes occur when r-Ti3C2Tx are exposed todifferent solutions and drying protocols. Fig. S5† shows theXRD results of various treatments carried out on the r-Ti3C2Txfilms. As shown in Fig. 3D and Fig. S5A,† after HF etching, the(0002) peak is found at around 2θ ≈ 9°, (peak labeled i inFig. S5A†) that corresponds to a c-LP ≈ 19.5 Å, typical of HF

Nanoscale Paper

This journal is © The Royal Society of Chemistry 2018 Nanoscale, 2018, 10, 22429–22438 | 22433

Publ

ishe

d on

07

Nov

embe

r 20

18. D

ownl

oade

d by

Sun

gkyu

nkw

an U

nive

rsity

on

12/1

8/20

18 1

2:25

:52

PM.

View Article Online

http://dx.doi.org/10.1039/c8nr06854b

etched Ti3C2Tx.4 After treatment in EDA, 2θ decreases to ≈6.5°,

or a c-LP ≈ 27.1 Å (curve labeled ii in Fig. S5A†). When thelatter is dried at 25 °C for 24 h, 2θ increases slightly to ≈6.6°,or a c-LP ≈ 26.7 Å (curve labeled iii in Fig. S5A†). Furtherdrying at 80° for 48 h shifts the peak to ≈7.3°, c-LP ≈ 24.2 Å(curve labeled iv in Fig. S5A†). Finally, drying at 150 °C for72 h returns the peak position to almost the initial value of≈9°, with a c-LP ≈ 20.0 Å (curve labeled v in Fig. S5A†).Fig. S5B† shows that after Li intercalation, but prior towashing, the (0002) peak is located at 2θ ≈ 6.5°, or a c-LP ≈27.1 Å (curve labeled ii in Fig. S5B†). However, after washingthe MLs in ethanol and a 5 wt% HCl solution, during whichthe Li+ ions are presumably exchanged by protons, the (0002)peak location increases to a 2θ ≈ 8.5°, or a c-LP ≈ 20.8 Å (curvelabeled iii in Fig. S5B†). In other words, by drying or simplywashing the intercalated MLs, it is possible to de-intercalateeither the EDA or Li ions. The effect of these treatments on theESR signal is significant. When Li or EDA intercalates betweenthe layers, the paramagnetism vanishes (Fig. 4B); when theyare de-intercalated, by washing for e.g., the ESR signal is recov-ered (Fig. 4B).

To assess the general stability of our r-Ti3C2Tx we storedpowders made from the material reduced at 40 °C, either inwater or in air. Like most aqueous Ti3C2Tx colloidal suspen-sions, they started off black (Fig. S6A†). After 2–3 days in water,

however, the color of the solution changed to sky blue(Fig. S6B†) indicating that r-Ti3C2Tx is not stable in aeratedwater. Storing the dried powder in air for 30 days, however, didnot alter its color (inset in Fig. S6C†). Most importantly, theESR signal after storing the samples in air for 30 d was almostidentical to that of fresh powders (compare Fig. S5C† andFig. 4A). It follows that the paramagnetism observed inr-Ti3C2Tx films is robust vis-a-vis atmospheric oxygen ifsamples are stored in air. These observations are in generalagreement with previous work on the oxidation of Ti3C2Tx thathas shown that indeed Ti3C2Tx powders are prone to oxidationin oxygen containing water, but much more stable if stored asdried powders.8,15,38

To complement the ESR measurements we carried out mag-netization, M, measurements as a function of temperature. Asshown in Fig. 4C, the M at 300 K increased with increasingreduction temperatures. The total magnetic susceptibility, χT –given by the slope of the M–H plots near the origin – alsoincreased with increasing reduction temperatures. When thesemeasurements were repeated as a function of temperature, χTwas found to be more or less constant over the 300 to 20 Ktemperature range (Fig. 4D). At T < 20 K, χT increased.

We can now suggest the following scenario – shown sche-matically in Fig. 1 – for what occurs when Ti3AlC2 is firstetched in HF, reduced with Li-EDA and finally washed. Any

Fig. 4 Paramagnetism in Ti3AlC2, Ti3C2Tx and r-Ti3C2Tx after Li-EDA treatment. (A) EPR results of Ti3AlC2, Ti3C2Tx, and r-Ti3C2Tx after reduction inLi-EDA at temperatures indicated. The EPR signal becomes stronger with increasing reduction temperature. The inset shows results after reductionat 25 °C; (B) effect of intercalation of either Li or EDA on the ESR signal. (C) Field-dependent magnetization curves (M–H) recorded at 300 K for theTi3C2Tx, and r-Ti3C2Tx after reduction at temperatures indicated. (D) Temperature dependence of χT of r-Ti3C2Tx after reduction at temperatures indi-cated. (E) Plot of 1/(χT − χp) vs. T as a function of reduction temperatures indicated. (F) DOS of Ti3C2O2 reproduced with permission from ref. 40.Copyright 2013, American Physical Society.

Paper Nanoscale

22434 | Nanoscale, 2018, 10, 22429–22438 This journal is © The Royal Society of Chemistry 2018

Publ

ishe

d on

07

Nov

embe

r 20

18. D

ownl

oade

d by

Sun

gkyu

nkw

an U

nive

rsity

on

12/1

8/20

18 1

2:25

:52

PM.

View Article Online

http://dx.doi.org/10.1039/c8nr06854b

model put forward has to explain a number of importantobservations, chief amongst them: (i) upon reduction, the Ti3+

concentration increases at the expense of the Ti4+, (ii) some ofthe Ti atoms in the Ti3C2Tx are in a +4 oxidation state and (iii)the average oxidation states of the Ti and C in Ti3C2Tx are,respectively, +2.4 and −2.6. For the sake of simplicity, they willhenceforth both be assumed to be 2.5;8 (iv) C atoms arereleased during the etching step.

We propose the following simplified scheme to model oursystem. In previous work,8 the Ti4+ signal was ascribed toTiO2 particles generated during the etching process. This workshows that it may not be the case. We will assume that afteretching the Ti3C2Tx chemistry is Ti3C2O(OH). If we now replaceone of the C atoms with O, the final chemistry is Ti3CO2(OH).This replacement changes the atomic arrangement from O–Ti–C–Ti–C–Ti–(OH) to O2−–Ti4+–O2−–Ti1.75+–C2.5−–Ti1.75+–(OH)1−.The oxidation states chosen are consistent with the fact thatthe average oxidation states of the Ti and C of +2.5 and −2.5,respectively, and as importantly account for clear evidence forpeaks belonging to “TiO2” in Ti3C2Tx (Fig. 3B). Said otherwise,as shown in Fig. 1B, the replacement of a C atom by an O atomcreates regions where the Ti ion carries an oxidation state of+4 because it is now surrounded solely by O2− ions.

When the MLs are exposed to Li-EDA, it is reasonable toassume that oxygen vacancies are created according to thefollowing reaction:

TiTix þ OCx ¼ Ti′ Ti þ V•O þ12O2 ð1Þ

Said otherwise, as shown in Fig. 1c, exposure to Li-EDAremoves a near surface O and one of the electrons releasedconverts one Ti4+ ion to Ti3+. The second electron most likelyis localized on the vacant oxygen sites, forming V•O. The latterare most probably the localized states near EF that are respon-sible for the Curie-type component of the Pauli paramagneticbehaviour observed at T < 10 K discussed below.

The density of electronic carriers in the MAX phase andtheir MXene is quite large (of the order of 1027 m−3).39

Consequently, as noted above, most MAX phases are Pauliparamagnets,7 wherein the magnetic susceptibility, χP, is inde-pendent of temperature and given by:

χP ¼ μ0μB2NðEFÞ ð2Þ

where µ0 is the permeability of free space, µB is the Bohr mag-neton and N(EF) is the DOS at EF, per energy per unit volume.

Knowing χP, N(EF) can be calculated from eqn (2). Theseresults are listed in Table 1.

Fig. 4F reproduces the DOS for Ti3C2O2 calculated by Yuand Kent.40 Here the DOS at EF is ≈0.5 eV per formula unit.Assuming that the thickness of a Ti3C2O2 flake is 1 nm – avalue slightly higher than 12 the c-LP of Ti3AlC2 and the valuemeasured by AFM on single flakes41 – the theoretical DOS atEF is 3.9 × 10

46 J−3 m−3. Coincidentally or not, this value is inexcellent agreement with the value, 5.3 × 1046 J−3 m−3,obtained here for Ti3C2Tx (Table 1). This agreement is evenmore impressive when it is appreciated that the chemistry ofour MXene is not Ti3C2O2 and is defective. Note that thevalues for the MXene samples reduced at 80 °C and 120 °C aresignificantly higher than the theoretical values, which can beattributed at least partially to defects and the fact that ourchemistry is relatively far from Ti3C2O2.

The results shown in Fig. 4D suggest that at T < 10 K, atemperature dependent Curie-like term, χC, given by:

42

χC ¼μ0μB

2NSkBT

¼ CT

ð3Þ

comes into play. Here, T is the absolute temperature, kB isBoltzmann’s constant and Ns is the number of singly occupiedstates, with spin 12, per unit volume. The total susceptibility, χT,is thus given by:

χT ¼ χC þ χP ð4ÞTo check the validity of eqn (3) and (4), 1/(χT − χP) vs. T was

plotted (Fig. 4E) and indeed straight lines – mostly passingthrough the origin, with R2 values >0.99 – were obtained.

To make sense of these results, we assume our system isthat of a disordered metal, close to a metallic to insulator,M–I, transition, and where EF is near the mobility edge. In thatcase, the susceptibility is given by:42

χðTÞ ¼ 2μ0μB2

kBT

X

i

12þ eβðμ�εi�UÞ þ e�βðμ�εiÞ ð5Þ

where β = 1/kBT, µ is the chemical potential, εi are the energiesof the localized states i and U is the average intrasite electron–electron Coulomb interaction energy. At high temperatures,eqn (2) is recovered.

However, at temperatures sufficiently low that kBT < U,states near EF become singly occupied and C is given byeqn (3). Kamimura et al.43 called the transition between atemperature independent Pauli type behavior and a Curie-lawtype a cross-over temperature, TC, for which kBTC ≈ U. It

Table 1 Magnetic parameters obtained herein and calculated density of states and concentration of defects

Samp. (°C) χ300 N(EF), J−1 m−3 C, K−1 TC, K Ns, m

−3 NP, m−3

MXene 5.7 × 10−6 5.3 × 1046

25 8.6 × 10−6 5.9 × 1046 0.00015 10 1.9 × 1025 1.1 × 1025

40 8.2 × 10−6 7.6 × 1046 2.0 × 10−5 5 0.26 × 1025 0.5 × 1025

80 18.6 × 10−6 17 × 1046 3.2 × 10−5 6 0.41 × 1025 1.4 × 1025

120 52.0 × 10−6 48 × 1046 0.00012 7 1.5 × 1025 4.6 × 1025

Nanoscale Paper

This journal is © The Royal Society of Chemistry 2018 Nanoscale, 2018, 10, 22429–22438 | 22435

Publ

ishe

d on

07

Nov

embe

r 20

18. D

ownl

oade

d by

Sun

gkyu

nkw

an U

nive

rsity

on

12/1

8/20

18 1

2:25

:52

PM.

View Article Online

http://dx.doi.org/10.1039/c8nr06854b

follows that the number of sites that are singly occupied, Np,per unit volume is given by:42

Np � UNðEfÞ � kBTCNðEfÞ ð6Þ

where TC is the temperature at which the electron–electroninteractions are comparable to the thermal energy in thisdisordered metal near the boundary of the M–I transition.From Fig. S7, TC can be estimated as the temperature atwhich the intersection of the T dependent and temperatureindependent portions of χT intersect. Based on this, admit-tedly, crude analysis, TC ranges from 5 to 10 K (see column5 in Table 1). It follows that at ≈0.4 to 0.7 meV, the U valuesare thus quite small presumably because of themetallic screening of the high concentration of mobileelectrons. For comparison, Sariciftci et al.42 have shown thatin some conductive polymers, TC is of the order of 50 K andU ≈ 4–5 meV.

The value of Ns obtained from Curie’s law should be equalto Np calculated from eqn (6). A comparison of columns 6 and7 shows that the maximum discrepancy is a factor of ≈4. Thisagreement – from totally independent measurements – has tobe considered good considering the many simplifying assump-tions made in deriving eqn (2) and (3), the most egregious ofwhich is the assumption of a spherical Fermi surface, totallyfree electrons, etc.

The Curie-like contributions to the susceptibility arise fromsingle occupancy of localized states near EF.

42 Based on thetotality of our results, it is reasonable to assume that theselocalized states are locations where C atoms were first replacedby O2− ions during the etching process. The reduction processremoves these ions and replaces them with oxygen/carbonvacancies. Again assuming the thickness of a singleTi3C2O2 flake to be 1 nm, then the total number of C-sites is≈2.5 × 1028 m−3. The fraction of vacant sites, fs, is thus of theorder of ≈1025/2.5 × 1028, or ≈0.0004. This value is quite lowand suggests that there is quite a bit of room to furtherenhance χ with say more aggressive or longer reduction proto-cols. The corresponding value reported by Sariciftci et al.42 forconductive polymers is 0.001.

The last piece of the puzzle that needs to be addressed iswhy when Li or EDA intercalates between the layers, the para-magnetism vanishes (Fig. 4B). If indeed the system is poisedbetween a M–I transition, then it would be reasonable toassume that upon intercalation and the transfer of electronsfrom the MXene sheets to the intercalant, the system ispushed into an insulating state. This important conclusion isbacked by DFT calculations on Ti3C2O2 reproduced in Fig. 4F,from which it is obvious that EF – given by the solid verticalline – is clearly at the upper edge of a pseudogap. WhenTi3C2O2 is reduced, EF moves presumably to the right and theDOS at EF increases accordingly. When, on the other hand,those same electrons are transferred to the intercalant, EF caneasily shift into the pseudogap, where the DOS is significantlyreduced. This interpretation would thus explain the lossof the ESR signal upon intercalation. In other words, the

results and conclusions are totally consistent with the assump-tion that the system is close to an M–I transition. In such asystem an increase in electrical conductivity upon reductionshould be observed and indeed one was found. The averageresistivity of 40 µm thick filtered films before reduction wasabout 40 ± 3 Ω cm−1; after reduction, at 28 ± 2.5 Ω cm−1, i.e.the average resistivities were reduced by a third. We note inpassing that when the EDA or Li+ ions are de-intercalated theyare replaced by protons. Why the latter do not reduce the EPRsignal is an open question at this time and beyond the scopeof this paper.

IV. Conclusions

We developed a new, facile, solution-based synthetic methodto reduce 2D Ti3C2Tx multilayers by immersing them in thestrong reducing agent, Li-EDA. XPS spectra provided strong evi-dence for the formation of Ti3+ ions and by extension oxygenvacancies. The chemically reduced Ti3C2Tx films are Pauliparamagnetic in that the magnetization is not a function oftemperature in the 20 to 300 K temperature range. The mag-netism, however, does not stem from the defects, but fromchanges in the DOS at Ef. This allowed us to experimentallydetermine, for the first time, the DOS at Ef. The latter are inreasonable agreement with theoretical predictions, lendingcredence to both. The fact that the paramagnetism disappearswhen the layers are intercalated is strong evidence that theparamagnetism is related to the MXene layers and is not, forexample, emanating from reduced TiO2 nanoparticles. In thereduced TiO2 literature, Ti

3+ is sometimes found to be theorigin of the ESR signal. The fact that herein g ≈ 2 eliminatesTi3+ as the magnetism’s source.

Reducing Ti3C2Tx flakes in the super base Li-EDA – at temp-eratures between 25 °C and 120 °C – resulted in a significantincrease in their Pauli paramagnetism and concomitantly theconcentration of Ti3+ ions. At temperatures

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

H. L. and T. A. L. acknowledge support from the Institutefor Basic Science (IBS-R011-D1). Y. Y. acknowledges supportfrom the National Research Foundation of Korea (NRF) undergrant no. NRF-2017R1A6A3A11033347. MWB acknowledges thesupport of the Swedish Research Council [Grant # 621-2014-4890].

Notes and references

1 K. S. Novoselov, A. K. Geim, S. V. Morozov, D. Jiang,Y. Zhang, S. V. Dubonos, I. V. Grigorieva and A. A. Firsov,Science, 2004, 306, 666.

2 K. K. Kim, A. Hsu, X. Jia, S. M. Kim, Y. Shi, M. Hofmann,D. Nezich, J. F. Rodriguez-Nieva, M. Dresselhaus,T. Palacios and J. Kong, Nano Lett., 2011, 12, 161.

3 Z. Yin, H. Li, H. Li, L. Jiang, Y. Shi, Y. Sun, G. Lu, Q. Zhang,X. Chen and H. Zhang, ACS Nano, 2011, 6, 74.

4 M. Naguib, M. Kurtoglu, V. Presser, J. Lu, J. Niu, M. Heon,L. Hultman, Y. Gogotsi and M. W. Barsoum, Adv. Mater.,2011, 23, 4248.

5 M. Naguib, O. Mashtalir, J. Carle, V. Presser, J. Lu,L. Hultman, Y. Gogotsi and M. W. Barsoum, ACS Nano,2012, 6, 1322.

6 M. Naguib, V. N. Mochalin, M. W. Barsoum and Y. Gogotsi,Adv. Mater., 2014, 26, 992.

7 M. W. Barsoum, MAX phases: properties of machinableternary carbides and nitrides, John Wiley & Sons, 2013.

8 J. Halim, K. M. Cook, M. Naguib, P. Eklund, Y. Gogotsi,J. Rosen and M. W. Barsoum, Appl. Surf. Sci., 2016, 362,406.

9 M. R. Lukatskaya, O. Mashtalir, Y. Dall’ Agnese, P. Rozier,P. L. Tabema, M. Naguib, P. Simon, M. W. Barsoum andY. Gogotsi, Science, 2013, 341, 1502.

10 B. Anasori, M. R. Lukatskaya and Y. Gogotsi, Nat. Rev.Mater., 2017, 2, 16098.

11 M. Khazaei, A. Ranjbar, M. Arai, T. Sasaki and S. Yunoki,J. Mater. Chem. C, 2017, 5, 2488.

12 T. Hu, J. Wang, H. Zhang, Z. Li, M. Hu and X. Wang, Phys.Chem. Chem. Phys., 2015, 17, 9997.

13 M. Khazaei, M. Arai, T. Sasaki, C. Y. Chung,N. S. Venkataramanan, M. Estili, Y. Sakka and Y. Kawazoe,Adv. Funct. Mater., 2013, 23, 2185.

14 S. Zhao, W. Kang and J. Xue, Appl. Phys. Lett., 2014, 104,133106.

15 O. Mashtalir, M. Naguib, V. N. Mochalin, Y. Dall’Agnese,M. Heon, M. W. Barsoum and Y. Gogotsi, Nat. Commun.,2013, 4, 1716.

16 D. R. Dreyer, S. Park, C. W. Bielawski and R. S. Ruoff,Chem. Soc. Rev., 2010, 39, 228.

17 R. F. Howe and M. Gratzel, J. Phys. Chem., 1985, 89,4495.

18 W. Fang, M. Xing and J. Zhang, J. Photochem. Photobiol., C,2017, 32, 21.

19 J. Su, X.-X. Zou, Y.-C. Zou, G.-D. Li, P.-P. Wang andJ.-S. Chen, Inorg. Chem., 2013, 52, 5924.

20 C. Zhang, Y. Xie, J. Ma, J. Hu and C. Zhang, Chem.Commun., 2015, 51, 17459.

21 W. S. Epling, C. H. Peden, M. A. Henderson andU. Diebold, Surf. Sci., 1998, 412, 333.

22 U. Diebold, J. Lehman, T. Mahmoud, M. Kuhn,G. Leonardelli, W. Hebenstreit, M. Schmid and P. Varga,Surf. Sci., 1998, 411, 137.

23 T. Bredow and G. Pacchioni, Chem. Phys. Lett., 2002, 355,417.

24 L. Reggel, R. Friedel and I. Wender, J. Org. Chem., 1957, 22,891.

25 O. Mashtalir, M. R. Lukatskaya, M. Q. Zhao,M. W. Barsoum and Y. Gogotsi, Adv. Mater., 2015, 27, 3501.

26 Y. Yoon, K. Lee, S. Kwon, S. Seo, H. Yoo, S. Kim, Y. Shin,Y. Park, D. Kim, J. Choi and H. Lee, ACS Nano, 2014, 8,4580.

27 Z. Ling, C. E. Ren, M.-Q. Zhao, J. Yang, J. M. Giammarco,J. Qiu, M. W. Barsoum and Y. Gogotsi, Proc. Natl. Acad.Sci. U. S. A., 2014, 111, 16676.

28 M. Y. Xing, J. L. Zhang, B. C. Qiu, B. Z. Tian, M. Anpo andM. A. Che, Small, 2015, 11, 1920–1929.

29 Y. Zhang, Z. Xing, X. Liu, Z. Li, X. Wu, J. Jiang, M. Li,Q. Zhu and W. Zhou, ACS Appl. Mater. Interfaces, 2016, 8,26851.

30 J. Halim, et al., Chem. Mater., 2014, 26, 2374.31 S. Myhra, J. Crossley and M. Barsoum, J. Phys. Chem. Solids,

2001, 62, 811.32 M. Ghidiu, J. Halim, S. Kota, D. Bish, Y. Gogotsi and

M. W. Barsoum, Chem. Mater., 2016, 28, 3507.33 L. Zhang and R. V. Koka, Mater. Chem. Phys., 1998, 57, 23.34 K. Hantanasirisakul, M. Q. Zhao, P. Urbankowski, J. Halim,

B. Anasori, S. Kota, C. E. Ren, M. W. Barsoum andY. Gogotsi, Adv. Electron. Mater., 2016, 2, 1600050.

35 Z. Li, L. Wang, D. Sun, Y. Zhang, B. Liu, Q. Hu andA. Zhou, Mater. Sci. Eng., B, 2015, 191, 33.

36 M. Ghidiu and M. W. Barsoum, J. Am. Ceram. Soc., 2017,100, 5395.

37 T. Bodziony, N. Guskos, A. Biedunkiewicz, J. Typek,R. Wrobel and M. Maryniak, Mater. Sci., 2005, 23,0137.

38 N. Guskos, J. Typek, T. Bodziony, G. Zolnierkiewicz,M. Maryniak and A. Biedunkiewicz, J. Alloys Compd., 2009,470, 51.

39 P. Finkel, M. Barsoum, J. Hettinger, S. Lofland and H. Yoo,Phys. Rev. B: Condens. Matter Mater. Phys., 2003, 67,235108.

40 Y. Xie and P. Kent, Phys. Rev. B: Condens. Matter Mater.Phys., 2013, 87, 235441.

Nanoscale Paper

This journal is © The Royal Society of Chemistry 2018 Nanoscale, 2018, 10, 22429–22438 | 22437

Publ

ishe

d on

07

Nov

embe

r 20

18. D

ownl

oade

d by

Sun

gkyu

nkw

an U

nive

rsity

on

12/1

8/20

18 1

2:25

:52

PM.

View Article Online

http://dx.doi.org/10.1039/c8nr06854b

41 A. Miranda, J. Halim, M. Barsoum and A. Lorke, Appl. Phys.Lett., 2016, 108, 033102.

42 N. Sariciftci, A. Heeger and Y. Cao, Phys. Rev. B: Condens.Matter Mater. Phys., 1994, 49, 5988.

43 H. Kamimura, Philos. Mag. B, 1980, 42, 763.44 R. Liu and W. Li, ACS Omega, 2018, 3, 2609–

2617.

45 K. J. Cai, Y. Zheng, P. Shen and S. Y. Chen, CrystEngComm,2014, 16, 5466.

46 M. Naguib, O. Mashtalir, M. R. Lukatskaya, B. Dyatkin,C. Zhang, V. Presser, Y. Gogotsi and M. W. Barsoum, Chem.Commun., 2014, 50, 7420.

47 R. B. Rakhi, B. Ahmed, M. N. Hedhili, D. H. Anjum andH. N. Alshareef, Chem. Mater., 2015, 27, 5314.

Paper Nanoscale

22438 | Nanoscale, 2018, 10, 22429–22438 This journal is © The Royal Society of Chemistry 2018

Publ

ishe

d on

07

Nov

embe

r 20

18. D

ownl

oade

d by

Sun

gkyu

nkw

an U

nive

rsity

on

12/1

8/20

18 1

2:25

:52

PM.

View Article Online

http://dx.doi.org/10.1039/c8nr06854b

Button 1: