Applied Physics Express     

LETTER

A novel dual–signal output screwing oscillator based on carbon@MoS2nanotubesTo cite this article: Yan-Wen Lin et al 2019 Appl. Phys. Express 12 065001

 

View the article online for updates and enhancements.

This content was downloaded from IP address 150.203.213.242 on 03/06/2019 at 09:07

A novel dual–signal output screwing oscillator based on carbon@MoS2 nanotubes

Yan-Wen Lin1, Wu-Gui Jiang1* , Qing-Hua Qin2, and Yu-Jiang Chen1

1School of Aeronautical Manufacturing Engineering, Nanchang Hangkong University, Nanchang 330063, People’s Republic of China2Research School of Engineering, Australian National University, Acton ACT 2601, Australia*E-mail: jiangwugui@nchu.edu.cn

Received March 21, 2019; revised April 16, 2019; accepted April 18, 2019; published online May 8, 2019

A novel dual–signal output screwing oscillator built on carbon@MoS2 heterogeneous nanotubes (CNT@MST) is proposed and its oscillationperformance is investigated using the molecular dynamics method. The MST acting as a stator is the outer tube of the oscillator, while the CNTacting as a rotor is the inner tube to which an initial rotational excitation and an axial translational excitation are simultaneously applied. Comparedwith the traditional double-walled carbon nanotube (CNT@CNT) screwing oscillators, CNT@MST screwing oscillators have the followingadvantages: better stability, lower energy dissipation rates, and wider adjustable intertube spacing. © 2019 The Japan Society of Applied Physics

Carbon nanotubes (CNTs),1) one type of nanoscalecarbon material with a special form of circular tubes,have wide applications in nano electromechanical

systems (NEMS) due to their excellent electrical, mechanical,chemical and optical properties.2,3) As a consequence,designing nano-oscillators based on double-walled CNTs(DWCNTs) or multi-walled CNTs (MWCNTs) has becomean important task in NEMS technology studies. During thepast decade, lots of efforts4–14) have been devoted to theinvestigation of the oscillation behaviors of single–signaloutput nano-oscillators by applying an initial axial transla-tional excitation or a rotational excitation to the inner tube,but no study on dual–signal output screwing oscillators hasbeen reported so far.In some transition metal sulfide materials having two-

dimensional molecular structures,15,16) such as molybdenumdisulfide (MoS2) and wolfram disulfide (WS2), tubularstructures similar to CNTs can also be formed.17)

Reference 18 found that a single-layer MoS2 sheet has ahigher quality factor than graphene. In addition, MoS2 can beused in NEMS for designing electronic devices such as fieldemission sensors,19) biosensors,20) and memory chips.21)

Reference 22 studied the nanomechanical properties ofMoS2 nanotubes (MSTs) via nanoindentation and compres-sion experiments. MSTs have good compatibility with CNTsdue to the excellent physical and chemical properties ofMSTs and their high similarity in structure to CNTs.23,24) Asa result, CNT@MSTs have great potential applications inmultifunctional engineering systems.25,26)

Recently, Ref. 27 established a CNT@MST oscillatormodel by means of the molecular dynamics (MD) methodto study the effects of intertube spacing and chirality on theoscillation behavior of the inner tube, indicating that theoscillation effect of a CNT@MST oscillator is superior tothat of a DWCNT oscillator. Reference 28 simulated rotatingself-excited DWCNT oscillators by the MD method, andpointed out that a rotating self-excited oscillator can oscillateat frequencies above gigahertz, but the amplitude of the self-excited axial oscillation is small (less than 0.5 nm) andunstable, which significantly limits its application prospects.It can be expected that two signals can be output bysimultaneously applying an axial translational excitationand a rotational excitation to the inner tube. However, noreport on the coupling effect of screwing motion on theoscillation behavior of CNT@MST screwing oscillators hasbeen made so far.

In this letter, a new CNT@MST screwing nano-oscillatoris presented to investigate the coupling effect of screwingmotion on inner-tube oscillation behavior via the classicalMD method. The proposed oscillator is also compared withthe traditional DWCNT screwing oscillators. In our simula-tions, the second-generation reactive empirical bond-order(REBO) potential29) is used to account for the intratubalatomic interactions of CNTs, while the atomic interactions ofMSTs are calculated using the Stillinger–Weber (SW)potential, which has been successfully used to study themechanical behavior of MoS2 by Refs. 18, 30, 31. In recentyears, there have been a few studies on the potential ofMoS2.

18,32–38) Based on the SW potential developed byRef. 18, the mechanical properties of MoS2/grapheneheterostructures,33) the buckling of single-layer MoS2 underuniaxial compression,34) and the phonon modes in MSTs35)

have been investigated. Based on the REBO potentialdeveloped by Ref. 30, Ref. 36 simulated the mechanicalbehavior of MSTs under compression, tension and torsion.Reference 37 reviewed most of the available force fields forMoS2, focusing on the calculation of the properties involvedin determining tribological behavior. From that work, re-commendations for a proper choice of empirical potentials touse in a computational study were provided. Reference 38evaluated the accuracy of the mechanical properties ofmonolayer MoS2 using the MD method with the CVFF1,CVFF2, SW and REBO potentials and found that in the caseof large deformation, the SW potential developed by Ref. 18is more accurate than the other potentials, although theCVFF2 and REBO potentials can predict well the elasticproperties of monolayer MoS2 in the case of small deforma-tion. Recently, Ref. 27 has used the SW potential topredict the oscillation behavior of CNT@MST oscillators.Besides these, there are two potential functions we need tobe concerned about between non-bonded atoms, i.e. vander Waals potential and electrostatic potential. In mostprevious studies on interaction in MWCNTs3,10,39) andMoS2/graphene heterostructures,30) only the Lennard-Jones(LJ) 12–6 potential function is considered because theelectrostatic interaction is very small compared to the vander Waals interaction.40) So, the interaction between the MSTand the CNT is here evaluated using the LJ 12–6 potential,which is described in the literature.27)

In all models, the outer tube is an MST while the innertube is a CNT, and both ends of the inner and outer tubes areopen, as shown in Fig. 1. In the simulations, the MST is fixed

© 2019 The Japan Society of Applied Physics065001-1

Applied Physics Express 12, 065001 (2019) LETTERhttps://doi.org/10.7567/1882-0786/ab1acf

and the CNT is applied with an initial separation distance of2 nm and an initial rotational frequency of 200 GHz. Thelength of the inner and outer tubes is around 6 nm and 4 nm,respectively.At the beginning, the two tubes in each mode have a

symmetrical layout along the axis, where the centers of boththe inner and outer tubes along the axial direction aremaintained at the same position. The whole system is putin a heat bath at a temperature around 0.001 K for 20 ps afterenergy minimization, which allows the system to fullyconverge at the specified temperature after the energyminimization, where the number of particles, volume andabsolute temperature of the system are constant (called thecanonical NVT ensemble, where N is the total number ofparticles, V is the volume of the system and T is the absolutetemperature). Then, pull the inner tube outward for 20 ps at avelocity of 0.1 nm ps−1 under constant temperature. After20 ps (the initial separation distance of the inner tube is2 nm), the inner tube is released from this velocity restraintand applied with a rotational excitation at a frequency of200 GHz for 20 ps. As our simulations show that theinfluence of the duration of the rotational excitation on theoscillation behavior can be ignored, here the duration of therotational excitation is set to be 20 ps in all simulations.Thereafter, we remove the rotational excitation of the innertube and the whole system varies under constant energyconditions, where the number of atoms, volume, and energyof the system remain constant (called the microcanonicalNVE ensemble, where NV is the same as given above and E isthe total energy in the system). LAMMPS41) (Large-scaleAtomic/Molecular Massively Parallel Simulator) is employedin the simulation with an analysis step size of 1 fs and anNVE total simulation time of 2 ns.We define the intertube spacing α as the distance between

the inner S atoms in the MST and the C atoms in the CNT, asshown in Fig. 1(b). According to Ref. 23, the equilibriumdistance between a CNT and an MST is approximately0.44 nm. Reference 6 reported that the equilibrium distancebetween inner and outer tubes in an MWCNT is around0.34 nm. Here we construct CNT(9,9)@MST(14,14), CNT(9,9)@CNT(14,14) and CNT(18,0)@CNT(27,0) screwingoscillators with the intertube spacing α close to the respective

equilibrium distance and compare their oscillation behaviors,in which the α of the CNT(9,9)@MST(14,14), CNT(9,9)@CNT(14,14) and CNT(18,0)@CNT(27,0) screwing oscil-lators are 0.447 nm, 0.329 nm and 0.356 nm, respectively.The evolution of the positions of the mass centers of the innertubes (MCITs) and the rotational frequencies of the innertubes (RFITs) for the CNT@MST and CNT@CNT screwingoscillators is plotted in Fig. 2. It can be seen from Fig. 2(a)that, for the CNT@MST oscillator, the axial oscillationamplitudes hardly decay, whereas for the CNT@CNToscillators, the oscillation amplitudes dissipate more severely,which indicates that the oscillation behavior of theCNT@MST screwing oscillator is more stable than that ofthe CNT@CNT screwing oscillators. Figure 2(b) plots theevolution of the RFITs for the CNT@MST and CNT@CNTscrewing oscillators with time, indicating that for theCNT@MST screwing oscillator, the dissipation of the rota-tional frequency is very small and the inner tube can performa stable rotation at the initial frequency of 200 GHz, whereasfor the CNT@CNT screwing oscillators, the rotationalfrequency decays sharply with time. The reason is thatcompared with the traditional CNT@CNT oscillators, theMST in the CNT@MST oscillator has a three-layer structureand a lower frictional coefficient.24) Figure 2 providesevidence that the traditional CNT@CNT screwing oscillatorshave significant dissipation in both oscillation amplitude androtational frequency during the screwing oscillation process,while the CNT@MST screwing oscillator can oscillateaxially with an almost equal oscillation amplitude and rotatecircumferentially with the initial rotational frequency, indi-cating that CNT@MST is more suitable as a potentialcandidate for a stable and low-loss screwing oscillator thanCNT@CNT.In order to investigate the effect of α on the oscillation

behavior of the CNT@MST screwing oscillator, fourCNT@MST screwing oscillator models with different α areconsidered: CNT(6,6)@MST(24,0), CNT(9,9)@MST(24,0),CNT(10,10)@MST(24,0) and CNT(12,12)@MST(24,0),whose α are 0.653 nm, 0.440 nm, 0.382 nm and 0.246 nm,respectively. The CNT(9,9)@MST(24,0) oscillator with anequilibrium distance of 0.440 nm exhibits an almost equalamplitude oscillation and equal frequency rotation during the

Fig. 1. (Color online) Schematic diagrams of the heterogeneous CNT@MST screwing oscillator model in the (a) axial direction and (b) radial direction.

© 2019 The Japan Society of Applied Physics065001-2

Appl. Phys. Express 12, 065001 (2019) Y.-W. Lin et al.

screwing oscillation process. As shown in Fig. 3, oscillatorswith larger α have more severe off-axial rocking motion thanthose with smaller α. For example, the CNT(6,6)@MST(24,0) oscillator has the largest α among the four oscillators,so its inner tube performs an off-axial rocking motion withthe largest displacement, which is about over 100 times thatof the other three oscillators. The frictional effect between theinner and outer tubes is closely related to the off-axialrocking motion. From Fig. 4 it can be seen that α

significantly affects the inner tube’s dissipation rate of bothoscillation amplitude and rotational frequency: as α is farfrom the equilibrium distance, the dissipation of bothoscillation amplitude and rotational frequency is more severe.This is mainly because when α is smaller than the equili-brium distance, a decrease in α may enhance intertubeinteraction, which causes the frictional effect between tubesto rise sharply and results in an increasingly pronounceddissipation behavior of the inner tube during the screwingoscillation. In contrast, when α is greater than the equilibrium

Fig. 2. (Color online) Evolution of (a) MCITs (inset: enlarged portion of the signals) and (b) RFITs of the CNT@MST and CNT@CNT screwing oscillatorswith time.

Fig. 3. (Color online) Off-axial rocking motions of inner tubes ofCNT@MST versus time for investigating the gap width effect; the insetshows the off-axial rocking motions of the (9,9)@(24,0), (10,10)@(24,0) and(12,12)@(24,0) heterogeneous nanotubes.

© 2019 The Japan Society of Applied Physics065001-3

Appl. Phys. Express 12, 065001 (2019) Y.-W. Lin et al.

distance, an increase in α may cause the off-axial rockingmotion distance to rise and then enhance the friction effect.To further investigate the effect of α on the dissipation rate

of the axial translation oscillation amplitude and the rota-tional frequency of the inner tubes for both CNT@MST andCNT@CNT during screwing motion, the offset of α isdefined as l a a= - ,0 in which a0 is the equilibriumdistance. It should be noted that for the CNT@MST models,a = 0.440 nm, while for the CNT@CNT models, a = 0.340

nm. Dimensionless dissipation rates = DA A

nA0and w = w

wDn 0

are defined to describe the dissipation rates of the axialoscillation amplitude and the circumferential rotational fre-quency, respectively. DA and wD are the dissipation of theoscillation amplitude and the rotational frequency in thewhole simulation respectively, n is the total number of cyclesin the simulation, the initial separation distance A0 = 2 nmand the initial rotational frequency w0 = 200 GHz.The variations of A and w for the CNT@MST and

CNT@CNT screwing oscillators with respect to λ are plottedin Figs. 4(a) and 4(b), respectively, from which it can be seenthat their values are very small when λ is near zero and theyare increased when λ is far from zero. We assume that thescrewing oscillation is unstable when A or w rises sharply asλ is far from zero, as shown in Fig. 4. The figure indicatesthat the CNT@MST screwing oscillators have a wideradjustable α than the CNT@CNT screwing oscillators. Itcan also be seen that the dissipation rates of both the axialoscillation amplitude and rotational frequency of theCNT@MST screwing oscillators are lower than those ofthe CNT@CNT screwing oscillators during screwing oscilla-tion. References 6,10,42 also reported that CNT@CNToscillators perform stable and sustained oscillation when α

is about 0.34 nm. We also find that a stable and low-lossscrewing oscillation can be obtained based on CNT@MSTheterogeneous nanotubes in a wider range of spacing valuesfrom 0.289 to 0.653 nm, indicating a stronger advantage overoscillators that are based on traditional CNT@CNT nano-tubes, which exhibit stable, sustained and low-loss screwingoscillation only when α is 0.340 to 0.376 nm.

In summary, we have presented a novel CNT@MSTscrewing oscillator that can provide outputs of both rotationaland oscillatory signals simultaneously. The effect of intertubespacing on oscillation behavior is especially investigated. TheMD method results show that, compared with the traditionaldouble-walled carbon nanotube (CNT@CNT) screwing os-cillators, CNT@MST screwing oscillators have three obviousadvantages: better stability, lower energy dissipation rates,and wider adjustable intertube spacing.

Acknowledgments This work was supported by the National NaturalScience Foundation of China (Grant No. 11772145).

ORCID iDs Wu-Gui Jiang https://orcid.org/0000-0002-5338-8323

1) S. Iijima, Nature 354, 56 (1991).2) J. Zou, B. H. Ji, X. Q. Feng, and H. J. Gao, Nano Lett. 6,, 430 (2006).3) Z. Qin, J. Zou, and X. Q. Feng, J. Comput. Theor. Nanosci. 5, 1403 (2008).4) Y. H. Zeng, W. G. Jiang, and Q. H. Qin, Nanotechnology 27, 095705

(2016).5) Q. H. Luo, W. G. Jiang, and Q. H. Qin, J. Nano Res. 48, 148 (2017).6) S. B. Legoas, V. R. Coluci, S. F. Braga, P. Z. Coura, S. O. Dantas, and

D. S. Galvao, Phys. Rev. Lett. 90, 055504 (2003).7) J. Cumings and A. Zettl, Science 289, 602 (2000).8) Q. S. Zheng and Q. Jiang, Phys. Rev. Lett. 88, 045503 (2002).9) W. L. Guo, Y. F. Guo, H. J. Gao, Q. S. Zheng, and W. Y. Zhong, Phys. Rev.

Lett. 91, 125501 (2003).10) Q. S. Zheng, J. Z. Liu, and Q. Jiang, Phys. Rev. B 65, 245409 (2002).11) A. Kis, K. Jensen, S. Aloni, W. Mickelson, and A. Zettl, Phys. Rev. Lett. 97,

02550 (2006).12) P. Liu, Y. W. Zhang, and C. Lu, J. Appl. Phys. 98, 56 (2005).13) P. Liu, Y. W. Zhang, and C. Lu, Carbon 44, 27 (2006).14) Y. Zhao, C. C. Ma, G. Chen, and Q. Jiang, Phys. Rev. Lett. 91, 175504

(2003).15) M. Remškar, M. Viršek, and A. Mrzel, Appl. Phys. Lett. 95, 133122

(2009).16) A. C. Ferrari, F. Bonaccorso, V. Fal’Ko, K. S. Novoselov, S. Roche,

P. Bøggild, S. Borini, F. H. L. Koppens, V. Palermo, and N. Pugno,Nanoscale 7, 4598 (2015).

17) R. Tenne, L. Margulis, M. Genut, and G. Hodes, Nature 360, 444 (1992).18) J. W. Jiang, Nanotechnology 26, 315706 (2015).19) M. Strojnik, A. Kovic, A. Mrzel, J. Buh, and D. Mihailovic, AIP Adv. 4, 387

(2014).20) D. Sarkar, W. Liu, X. Xie, and A. C. Anselmo, ACS Nano 8, 3992 (2014).21) S. Bertolazzi, D. Krasnozhon, and A. Kis, ACS Nano 7, 3246 (2013).22) D. Maharaj and B. Bhushan, Sci. Rep. 5, 08539 (2014).

Fig. 4. (Color online) Dimensionless dissipation rates (a) A and (b) w with respect to the offset λ of the intertube spacing.

© 2019 The Japan Society of Applied Physics065001-4

Appl. Phys. Express 12, 065001 (2019) Y.-W. Lin et al.

23) X. C. Song, Y. F. Zheng, Y. Zhao, and H. Y. Yin, Mater. Lett. 60, 2346(2006).

24) X. F. Zhang, B. Luster, A. Church, and C. Muratore, ACS Appl. Mater.Interfaces 1, 735 (2009).

25) V. O. Koroteev, L. G. Bulusheva, I. P. Asanov, and E. Shlyakhova, J. Phys.Chem. C 115, 21199 (2011).

26) H. Yoo, A. P. Tiwari, J. Lee, and K. Doyoung, Nanoscale 7, 3404 (2015).27) W. G. Jiang, Y. H. Zeng, Q. H. Qin, and Q. H. Luo, Nano Res. 9, 1775

(2016).28) K. Cai, H. Yin, Q. H. Qin, and Y. Li, Nano Lett. 14, 2558 (2014).29) D. W. Brenner, O. A. Shenderova, J. A. Harrison, and S. J. Stuart, J. Phys.:

Condens. Matter 14, 783 (2002).30) J. W. Jiang, H. S. Park, and T. Rabczuk, J. Appl. Phys. 114, 064307 (2013).31) J. W. Jiang, H. S. Park, and T. Rabczuk, Nanoscale 6, 3618 (2014).

32) T. Liang, S. R. Phillpot, and S. B. Sinnott, Phys. Rev. B 79, 245110(2009).

33) J. W. Jiang and H. S. Park, Appl. Phys. Lett. 105, 033108 (2014).34) J. W. Jiang, Nanotechnology 25, 355402 (2014).35) J. W. Jiang, B. S. Wang, and T. Rabczuk, Nanotechnology 25, 105706

(2014).36) E. W. Bucholz and S. B. Sinnott, J. Appl. Phys. 112, 123510 (2012).37) P. Nicolini and T. Polcar, Comput. Mater. Sci. 115, 158 (2016).38) S. Xiong and G. X. Cao, Nanotechnology 26, 185705 (2015).39) J. Zou, B. H. Ji, X. Q. Feng, and H. J. Gao, Small 2, 1348 (2006).40) Y. Ma, Y. Dai, M. Guo, and C. Niu, Nanoscale 3, 3883 (2011).41) S. Plimpton, J. Comput. Phys. 117, 1 (1995).42) J. L. Rivera, C. McCabe, and P. T. Cummings, Nanotechnology 16, 186

(2005).

© 2019 The Japan Society of Applied Physics065001-5

Appl. Phys. Express 12, 065001 (2019) Y.-W. Lin et al.