ManipulateInterfaceThermalConductanceofMetal-Graphene-SubstrateUsingGrapheneThermalStability.doc

精品論文manipulate interface thermal conductance of metal-graphene-substrate using graphene thermal stability5zhao weiwei1, zhang chunwei1, yong guoqing1, huang peng1, bi kedong1, chen yunfei1, ni zhenhua2(1. school of mechanical engineering, southeast university, nanjing 211189;2. department of physics, southeast university, nanjing 211189)abstract: graphene defects have played an important role in graphenes electrical, optical and thermal10properties. in this work, we measured the interface thermal conductance of al/graphene with different defect situations/silica using transient thermoreflectance technique. we use the graphene thermal stability of being annealed in different temperatures to manipulate the graphene defects. the transient thermoreflectance technique measurement results suggest that as graphene defects increase, the interface thermal conductance increases as well. the cause of the interface thermal conductance15increase might be the graphene disintegration which starts from graphene imperfection rings and the oxygen doping which will enhance the metal-graphene bonding energy.keywords: thermal stability; graphene; interface thermal conductance; transient thermoreflectancetechnique200introductiongraphene, ever since its discovery, has drawn great attention due to its unique electrical, optical and thermal properties 1. with a sub-nanometer thickness, the amazing robust graphene could be widely applied to micro device fabrication. as we know, different electrical, mechanical and thermal properties are measured when graphene is suspended and supported on the substrate252, 3. nevertheless, regardless of being suspended or supported, interface conductance between graphene and other material is an inevasible topic when it comes to heat transfer. however, the interface thermal conductance (itc) between graphene and different substrates remains uncertain since computer simulations and experiments often present a huge difference 4, 5, 6. the value of itc is various with materials properties on each side of the interface and contact surface situation30changing 7, 8, 9.transient thermoreflectance technique (ttr) is widely used to measure the thermal diffusion, thermal conductance and itc since applied by paddock and eesley for the first time 10. more experiments about the itc between materials and graphene are performed since the discovery of graphene. fai et al measured the itc at the interface between graphene and silica using ttr11.35koh et al did the research on the thermal conductance across the au/ti/graphene/silica surface 12. these results are comparable to those measured by using other methods, such as raman spectrum, 3 method and null point scanning thermal microscopy 3, 5, 13. in all above cases, graphene is considered as a perfect two dimensional material. however, in many cases, defects are introduced to be a powerful method to modulate the properties of graphene. impurity doping,40annealing, plasma etching and e-beam exposure are all feasible methods to introduce defects into graphene 14-16. although hopkins et al investigated the effects of plasma etching on the thermal properties of graphene 6, the changing tendency of itc while graphene being annealed remains uncertain.in this paper, a fabrication method of defects in graphene using graphene thermal stability isfoundations: doctoral fund of ministry of education of china (20090092120004, 20100092110051) brief author introduction:zhao weiwei, (1986-), male, micro-nano scale heat transfer. correspondance author: bi kedong, (1979-), male, associate professor, micro-nano electro-mechanical systems. e-mail: - 6 -45depicted and the defect characterization is given. the itc at the interface between aluminum, graphene with different defect condition and silica is measured using ttr around the room temperature.1experimentultra-short femtosecond laser transient thermoreflectance technique (ttr) is set up based on50the theory of certain materials reflectivity being proportional to the surface temperature 17. a76mhz laser signal is split into two beams, the stronger one being used to heat the sample after being modulated while the weak one being used to monitor the sample surface temperature change after a time delay. optical signal of the weak ones reflectance is translated into voltage signal by the photodiode and collected by a lock-in amplifier. after being analyzed, the thermal properties55are captured.fig. 1. sample characterization. (a) schematic of samples for transient thermoreflectance technique. (b) sem image of transferred graphene. (c) optical image of transferred graphene. (d) raman spectrem of transferred graphene.60graphene films are fabricated using low pressure chemical vapor deposition (lpcvd)method 18. the growth pressure is around 120130 pa and proportion of methane plus hydrogen is 60 and 40 standard-state cubic centimeter per minute (sccm). hence, we can get monolayer graphene on both sides of copper. using pmma, graphene is transferred onto silica, and then is65confirmed to be one monolayer graphene film by raman spectrum. the 2d fwht of 38 cm-1,shown in fig. 2(d), indicates that the graphene is mostly monolayer with small part being bilayer, since the 2d mode helps identify the graphene thickness 19. because of the graphene growing principle on copper, there will be a few graphene islands giving the raman signal of higher g peak intensity, indicating the islands are thicker than other parts. the number of thicker islands70can be minimized by controlling the growth time and gas flow proportion. the sem image shows that the graphene surface is quite clean, with negligible pmma residuals after being washed by using anisole and ethanol.2results and discussionsdefects in graphene films are introduced by using annealing method in this experiment. after7580859095100transferring graphene using usual washing method, we expose the growing quartz tube in the air ambient, place graphene sample into the tube, and maintain it at a different temperature for 10 minutes. then there is a different d peak intensity and g peak intensity ratio (id/ig) at a different temperature. compared with the unheated graphene, there are a higher d peak, a lower 2d peak and a small preiection signal near the g-peak shown in raman spectrum of heated ones because of the heating process, which could be used to identify the defect type, as shown in fig. 2(a) 20. as shown from the heating result, with the temperature increasing, the d peak intensity gets higher and the g peak intensity gets lower. it is known that the intensity of the d peak suggests the filmdisorders. severe change happens at around 575 after the g peak intensity gets quite low at 575, indicating the graphene has been seriously damaged and most of the carbon atoms have gone. from fig. 2 (c) and (d), it is demonstrated that once the temperature reaches up to 500, even though being annealed for the same time, the id/ig of the samples gets higher and the ig getslower as the temperature goes higher. hence, it is possible to manipulate the defects in graphene by controlling the anneal temperature more meticulously.fig. 2. sample defection raman characterization. (a) a typical raman signal of graphene annealed at 550oc. (b) a typical raman signal of graphene annealed at 575oc. (c) the ratio of graphene d peak and g peak after annealed. (d) the intensity of graphene g peak after annealed.before performing the thermoreflectance measurement, a layer of metal is deposited on the surface of the graphene. aluminum of 70 nm thickness is thermally evaporated onto the graphene. to avoid the damage of graphene structure by the metal evaporation, gold of 5 nm in thickness is thermally evaporated onto a silica film with etched sink. raman spectrum of suspended and supported graphene is investigated. as shown in fig. 3, the raman signal confirms that the graphene is not damaged by the thermally evaporating process and no visible defects are induced. hence, thermal evaporation is an acceptable choice for metal deposition onto the graphene.105fig. 3. raman characterization of thermal evaporation onto the graphene. (a) a raman g peak mapping ofgraphene with 5 nm thick gold on top. red part suggests a higher g peak intensity, which is graphene supported onthe substrate. blue part is graphene suspended on the sink. (b) raman signal of graphene suspended and supported with gold on top.110115120125using the samples and the measuring technique introduced above, we get the voltage signals of graphene samples which include amplitude part and phase part. both of these two parts could be used to fit the theoretical model and should get the same result. the phase part of the signal is employed to do the data fitting since it has a very low sensitivity to the laser diffuse which might occur and cause a deviation. figure 4 shows the phase signal obtained from the lock-in amplifier as a function of time delay. the value of itc at the interface of non-annealed al/graphene/silicais around 10 mw/ (m2*k) from data fitting, which is obviously lower than that of the al/silica inour former experiments 21. as shown in fig. 1, d peak signal is quite weak and the ratio of the id/ig is similar to that of graphene obtained by mechanical exfoliation method, confirming the integrity of the six-member ring structure. consequently, according to the depositing principle of metal, a one-layer metal atoms form along the graphene lattice boundary, or in the shape of nanoclusters on the surface of graphene, without metal atoms deposited below the graphene 22. hence, there should be a well-arranged al-graphene-silica interface. the depressed thermal interface conductance value confirms that when it comes to the interface between metal/graphene and silica, the whole interface should be considered as two interfaces combined together since phonons could not traverse across the whole interface directly from metal side to oxide side 12.fig. 4. data fitting of the non-annealed films phase signal.the annealing progress described above introduces a series of defects in graphene. as theid/ig goes relatively higher, the interface thermal conductance increases. the decreasing raman130135140145150155160g peak shows that there is less graphene left but more defects, while the increasing d peak shows there is actually several oxygen atoms doped into the graphene boundary. here the interface, lower than al/silica, could still be considered as two separated interfaces combined together when the defect is not significant. the reason that annealing cause an increasing interface thermal conductance is complicated and the annealing defects in the graphene might be attributed to the graphene growing process. according to the graphene growing principle, graphene islands start to form on cooper surface, become bigger and connected with each other as a whole film finally. compared with the graphene islands, the grain boundary of graphene always has more imperfections and more five-member or seven-member rings. when graphene is annealed in oxygen environment, lots of carbon atoms are bonded with oxygen atoms. however, during the annealing process, the escaping of carbon atoms starts from where the imperfections are, since six-member ring is more stable than five-member or seven-member ring. this might be why graphene made by cvd method shows defects at a lower temperature than graphene made by mechanical exfoliation. therefore, after annealing process, multi-member rings appear, and the itc of the total interface can be seen as the al/graphene/substrate and al/substrate working together. also, when graphene is annealed in oxygen environment, not only the carbon atom escaping happens, but also oxygen doping happens. the bonding energy of al/graphene could be enhanced due to the oxygen atom. hence, the increasing itc could be attributed to these two factors working together. nerveless, by controlling the defects in graphene, it might be possible to manipulate the itc more precisely.fig. 5. experiment results of various samples annealed at different temperatures.3conclusionto summarize, we have fabricated graphene on cu using cvd method. after graphene is transferred onto silica wafers, annealing process is used to introduce some defects. the itc of annealed graphenes increases with the increasing defects. judging from the raman spectrum of graphene with defects, it is inferred that the escaping carbon atoms caused by the increasing temperature and oxygen doped into graphene might result in the increasing itc.references1 novoselov k s, geim a k, morozov s v, et al. electric field effect in atomically thin carbon filmsj. science, 2004, 306: 666-669.1651701751801851901952002052102152 seol j h, jo i, moore a l, et al. two-dimensional phonon transport in supported graphenej. science, 2010,328: 213-216.3 cai w w, moore a l, zhu y w, et al. thermal transport in suspended and supported monolayer graphenegrown by chemical vapor depositionj. nano letters, 2010, 10: 1645-1651.4 chang s w, nair a k, buehler m j. geometry and temperature effects of the interfacial thermal conductance in copper- and nickle-graphene nanocompositesj. journal of physics: condensed matter, 2012, 24: 245301-6.5 chung j, hwang g, kim h, et al. measurement of thermal contact resistance between cvd-grown graphene and si02 by null point scanning thermal microscopy z. birmingham, united kingdom: 12th ieee international conference on nanotechnology, august 20-23, 2012.6 hopkins p e, baraket m, barnat e v, et al. manipulating thermal conductance at metalgraphene contacts via chemical functionalization j. nano letters, 2012, 12: 590-595.7 swartz e t, pohl r o. thermal boundary resistancej. reviews of modern physics, 1989, 61: 605-668.8 stevens r j, smith a n, norris p m. measurement of thermal conductance of a series of metal-dielectricinterfaces by the transient thermoreflectance techniquej. journal of heat transfer, 2005, 127: 315-322.9 hopkins p e, norris p m. thermal conductance response to a changein cr/si interfacial propertiesj. appliedphysics letters, 2006, 89: 131909-3.10 paddock c a, eesley g. transient thermoreflectance from thin metal filmsj. journal of applied physics,1986, 60: 285-290.11 fai m k, hung l c, tony h. measurement of the thermal conductance of the graphene/silica interfacej. applied physics letters, 2011, 97: 221904-3.12 koh y k, bae m h, cahill d g, et al. heat conduction across monolayer and few-layer graphenesj. nanoletters, 2010, 10: 4363-4368.13 chen z, jang w, bao w, et al. thermal contact resistance between graphene and silicon dioxidej. appliedphysics letters, 2009, 95: 161910-3.14 wang y, shao y y, matson d w, et al. nitrogen-doped graphene and its application in electrochemicalbiosensingj. acs nano, 2010, 4: 1790-1798.15 lin y c, lu c c, yeh c h, et al. graphene annealing: how clean