Maskless laser processing of graphene
Fujio Wakaya, Tadashi Kurihara, Nariaki Yurugi, Satoshi Abo, Masayuki Abe, Mikio Takai
MMaskless laser processing of graphene
Fujio Wakaya, a) Tadashi Kurihara, Nariaki Yurugi, Satoshi Abo, Masayuki Abe, and Mikio Takai
Center for Science and Technology at Extreme Conditions, Graduate School of Engineering Science,Osaka University 1-3 Machikaneyama, Toyonaka, Osaka 560-8531, Japan (Dated: Oct. 24, 2014 submitted to Microelectronic Eng.; revised on Mar.10, 2015; accepted on Mar. 25, 2015)
Graphene on a SiO /Si substrate was removed by ultraviolet pulsed laser irradiation. Threshold laser powerdensity to remove graphene depended on the graphene thickness. The mechanism is discussed using kineticenergy of thermal expansion of the substrate surface. Utilizing the thickness dependence, thickness (orlayer-number) selective process for graphene is demonstrated. Maskless patterning of graphene using laserirradiation in the air is also demonstrated. Keywords: graphene; maskless laser process; ultraviolet laserJournal reference: Microelectronic Eng. , 203-206 (2015) (published online Apr. 2, 2015)DOI: 10.1016/j.mee.2015.03.049
I. INTRODUCTION
Graphene has been drawing attention due to theoutstanding electrical and optical characteristics sincethe so-called scotch-tape method was used to transfergraphene from graphite. Although extremely high room-temperature carrier mobility in graphene may be uti-lized for a high-speed transistor, realizing such transis-tors is still a challenging issue due to such problems asband-gap control and mobility suppression by extrin-sic scatterings.
Another most realistic application ofgraphene is a transparent electrode utilizing high trans-mittance and high electrical conductance of graphene.Graphene transparent electrodes have been reported insolar cells, touch screen panels, and flat panel displays, which means that mass-production technology is one ofthe most important issues in this field presently.Indium tin oxide (ITO) is widely used for the mate-rial of the transparent electrode. For patterning of ITO,maskless laser process using infrared or ultraviolet (UV)lasers was reported as a simple and fast process compar-ing to the conventional lithography process with a resistand masks.
The mechanism for the ITO etching bylaser irradiation was reported to be high-temperature ef-fects, i.e. melting, evaporation, and ablation.
In the previous study, we reported that UV pulsedlaser irradiation with a wavelength of 248 nm removedgraphene from a SiO /Si surface if the laser power den-sity is over a threshold value, and that no thicknesschanges and no defect generations were observed belowthe threshold power density. Mechanism for the observedremoval was supposed to be mechanical ejection from thesubstrate surface. Moreover, a possibility of selective re-moval of thick graphene was pointed out.In this paper, graphene thickness dependence of thethreshold laser power density is discussed, and thethickness-selective process using UV pulsed laser irradia-tion is demonstrated. Such a selective process should be a) Electronic mail: [email protected] quite useful because a specific-layer-number graphene ispreferable for many applications. Furthermore, the mask-less patterning of the graphene is also demonstrated.
II. EXPERIMENTAL
For the experiments of determining the threshold laserpower density to remove graphene from the substrate sur-face, a Si substrate covered with 100-nm-thick SiO layerwas used. Graphene pieces were transferred onto the sub-strate surface from natural graphite using the scotch-tapeprocess. The samples after the scotch-tape process werefirst observed by optical microscopy. The graphene thick-ness was determined by atomic-force-microscopy (AFM)observation and Raman spectroscopy with an incidentlaser wavelength of 488 nm. The samples were irradiatedby KrF excimer laser with a wavelength and a pulse widthof 248 nm and 20 ns, respectively. The laser beam washomogenized and shaped to a stripe (0.4 ×
60 mm ). Therepetition frequency of the pulsed laser was 30 Hz. Thesample stage was moved at 0.6 mm/s to the direction ofthe 0.4-mm width of the stripe, resulting in 95% overlapbetween the successive two pulses and 20 pulse irradia-tions at every point on the sample. This overlap and themany irradiations are to avoid the influence of ununiformintensity in the stripe of the laser light. Accumulationof the laser irradiation effects cannot be expected be-cause the laser pulse width of 20 ns is much shorter thanthe repetition cycle of 33 ms. After laser irradiation, thesamples were again observed by optical microscopy. Thisprocedure was repeated with higher laser power densityuntil the graphene was removed. The laser power densityat which the graphene was removed was defined as thethreshold laser power density. The threshold laser powerdensities for many graphene pieces with various thick-nesses were determined to obtain thickness dependenceof the threshold power density.For the demonstration of maskless patterning usinglaser irradiation, almost all setups were the same as de-scribed above except for the starting material and the a r X i v : . [ phy s i c s . a pp - ph ] D ec T h r e s ho l d l a s e r po w e r den s i t y ( M W / c m ) Graphene thickness (nm)
FIG. 1. Graphene thickness dependence of threshold laserpower density to remove graphene from SiO /Si substrate.Thinner graphene needs more laser power. -1 ) I n t en s i t y ( a . u . ) FIG. 2. Raman spectrum of graphene. The graphene is iden-tified as single layer because the ratio of the integrated inten-sities of the G and D (cid:48) peaks is 0.21. sample stage control. Commercially available graphene-on-SiO (100nm)/Si samples were used for this purpose.Almost all regions of the sample surface were coveredby single-layer graphene, which was grown by the chem-ical vapor deposition and transferred onto the SiO /Sisubstrate. The sample stage was not moved during thedemonstration. A 0.4-mm-wide stripe pattern was, there-fore, expected to appear by the demonstration. III. RESULTS AND DISCUSSION
Fig. 1 shows experimentally obtained graphene thick-ness dependence of the threshold laser power densityfor removing graphene from the substrate surface. It isshown that the thinner graphene needs the higher laserpower. While the AFM measurement was used to deter-mine the graphene thickness, Raman spectroscopy wasalso used for very thin graphene to be regarded as single-layer graphene. Fig. 2 shows the Raman spectrum forthe graphene identified as single-layer. As discussed in the dry laser cleaning process, during the laser irradi-ation the substrate material is thermally expanded witha surface velocity, v , of v = 1 + σ − σ ) βIcρ (1)Here, σ is the Poisson ratio, β is the thermal expansioncoefficient, c is the specific heat, ρ is the density, and I is the absorbed laser power density. Ignoring reflec-tion of the laser light and absorption by the graphene,we can estimate the velocity v to be 22 cm/s for 8.0MW/cm laser irradiation. For the estimation, σ = 0 . β = 7 . × − K − , c = 0 .
72 J/gK, and ρ = 2 . are used for Si at 300 K although these parametersshould depend on the temperature, and SiO is ignoredbecause it is transparent for the 248-nm-wavelength light.The kinetic energy per unit area which the graphenecould receive from the surface expansion can be estimatedas (1 / ρ s v , where ρ s is the sheet density of graphene.This energy is 1 . × − J/m for single-layer graphenefor 8.0 MW/cm laser irradiation. The energy (1 / ρ s v is proportional to the sheet density of graphene, whichmeans that the thicker graphene can receive more energyfrom the substrate expansion. We believe that receivingthis kinetic energy is the principal mechanism for the ob-served thickness dependence of the threshold laser powerdensity.The adhesion potential between graphene and a SiO surface was reported as ∼ − J/m , which is mea-sured at room temperature and seven orders of magni-tude larger than the kinetic energy discussed above. Thelarge discrepancy may be due to the temperature riseduring the laser irradiation. When a laser light with apower density of I is absorbed by a thick Si substrate,the temperature T at the Si surface after time t is T = 2 Ik (cid:114) Dtπ (2)where k and D are the thermal conductivity and thermaldiffusivity of Si, respectively. Although the layer struc-ture in the present work is different from the thick Sisubstrate without any other layers, the temperature atthe graphene and the SiO layer can be estimated usingEq. (2) because the absorption at the thin graphene andthe SiO layer can be ignored for a rough estimation. Thetemperature after 20 ns from 8.0 MW/cm irradiationsis actually estimated to be 790 o C. For the estimation, k = 168 W/mK and D = k/cρ are used. The ther-mal vibration energy per each atom can be estimatedas ∼ k B T , which is 1 . × − J for 790 o C. The atomdensity of graphene is 3 . × m − . The thermal vibra-tion energy per unit area is, therefore, ∼ . × − J/m ,which is on the same order as the room-temperature ad-hesion potential between graphene and SiO . Theestimated high temperature, therefore, might affect theadhesion potential effectively through the thermal vibra-tion of the atoms. The absorption coefficient of the 248-nm-wavelength light for the graphite along the c axis and
10 μm10 μm (a)(b)
FIG. 3. Optical microscope images of sample surface (a) af-ter scotch-tape process, (b) after 3.0 MW/cm laser irradia-tion. The 4.1-nm-thick graphene indicated by the white circleslightly changes the position and apparently changes the ro-tation angle. the corresponding absorption length are 0.13 nm − and7.7 nm, respectively, which means that relatively largeabsorption occurs at the graphene layer especially if thegraphene is thick comparing to the absorption length. Anumerical simulation taking the reflection at the top sur-face and the interface, the absorption at the graphenelayer, the heat conduction at all layers and the thermalexpansion of SiO and Si layers into account is necessaryto discuss the thermal expansion and the temperature inmore detail.The authors in Ref. 20 also reported that the thresholdlaser energy depended on the graphene thickness, whichis quite similar to the result in the present work. Their in-terpretation is, however, different from that in the presentwork. They argued that the mechanism for the grapheneremoval was thermal ablation and attributed the ob-served graphene thickness dependence to that the specificheat depends on the graphene thickness. They, however,observed that a part of the graphene edge was folded ontothe graphene after laser irradiation. As a similar example,we observed that a graphene piece was found at a differ-ent position with a different rotation angle after laserirradiation as shown in Fig. 3. Such experimental obser-vation suggests that the mechanism for the graphene re-moval is mechanical ejection from the surface. The largediscrepancy of the estimated kinetic energy to the ad-hesion energy reported in the literature, however, may 𝜇m 𝜇m 𝜇m (a)(b)(c) FIG. 4. Optical microscope images of sample surface (a) afterscotch-tape process, (b) after 5.5 MW/cm laser irradiation,and (c) after 6.5 MW/cm laser irradiation. The graphenethicknesses are indicated by white letters in the images. Thethickness (or layer-number) selective process is demonstrated. suggest that both of the mechanical and thermal effectsshould be considered.The graphene thickness dependence of the thresholdpower density shown in Fig. 1 suggests a possibilityof thickness (or layer-number) selective process. Fig. 4shows a demonstration of such a process. There are threegraphene pieces with a thickness of 0.7 nm, 1.0 nm, and1.5 nm as shown in Fig. 4(a). The AFM cross-sectionalplots of the graphene pieces are shown in Fig. 5. Af-ter 5.5 MW/cm laser irradiation, only the 1.5-nm-thickgraphene was removed as shown in Fig. 4(b). The 1.0-nm-thick graphene was removed by the 6.5 MW/cm laserirradiation as shown in Fig. 4(c). As a result, only thethinnest graphene with a thickness of 0.7 nm remainson the substrate. This experiment successfully demon-strates the thickness selective process of the graphene byUV pulsed laser irradiation.A demonstration result of the maskless patterning ofgraphene in the air by laser irradiation is shown in Fig. 6.The central part of the single-layer graphene/SiO /Sisample surface was irradiated by stripe-shaped laser m) H e i gh t ( n m ) FIG. 5. AFM cross-sectional plots of graphenes shown inFig. 4.
FIG. 6. Optical microscope image of sample surface after 10MW/cm laser irradiation. Graphene only in the irradiatedregion, which is the central part of the image, is removed.Maskless patterning of graphene using laser irradiation in theair is demonstrated. light at 10 MW/cm . The 0.4-mm-wide stripe where nographene existed was clearly observed in the figure. Ifthe laser light is shaped as a spot and the sample stageis computer-controlled, an arbitrary pattern can be real-ized by laser irradiation in the air without masks, whichshould be useful in the mass-production process. Thegraphene material used for the demonstration was not ahuge single crystal but consisted of many pieces of grain,the size of which was typically several µ m. Possible mech-anism for the demonstration is, therefore, that each pieceof the graphene is ejected without breaking the covalentbond between the carbon atoms when the substrate isirradiated by the laser. IV. SUMMARY
Graphene pieces on a SiO /Si substrate were removedby UV pulsed laser irradiation. The threshold power den-sity to remove graphene depended on the graphene thick-ness. The mechanism was proposed using the substratethermal expansion as is well known in the dry laser clean-ing process.Utilizing the thickness dependence of the thresh- old laser power density, thickness selective process forgraphene was demonstrated. The thickness selective pro-cess, or layer-number selective process, is quite interest-ing because a specific layer-number graphene is preferablein many applications.Maskless patterning of graphene using laser irradiationin the air was demonstrated using a SiO /Si substratecovered with single-layer graphene. This process will con-tribute to the mass production of graphene devices. ACKNOWLEDGMENTS
The authors acknowledge financial support from theJapan Society for Promotion of Science with KAKENHIGrant Number 25420287.
REFERENCES K. S. Novoselov, A. K. Geim, S. V. Morozov, D. Jiang, Y. Zhang,S. V. Dubonos, I. V. Grigorieva, and A. A. Firsov, Science ,666 (2004). E. McCann, Phys. Rev. B , 161403 (2006). T. Ohta, A. Bostwick, T. Seyller, K. Horn, and E. Rotenberg,Science , 951 (2006). K. I. Bolotin, K. J. Sikes, J. Hone, H. L. Stormer, and P. Kim,Phys. Rev. Lett. , 096802 (2008). K. Bolotin, K. Sikes, Z. Jiang, M. Klima, G. Fudenberg, J. Hone,P. Kim, and H. Stormer, Solid State Commun. , 351 (2008). J. Wu, H. A. Becerril, Z. Bao, Z. Liu, Y. Chen, and P. Peumans,Appl. Phys. Lett. , 263302 (2008). K. S. Kim, Y. Zhao, H. Jang, S. Y. Lee, J. M. Kim, K. S. Kim,J.-H. Ahn, P. Kim, J.-Y. Choi, and B. H. Hong, Nature ,706 (2009). X. Li, Y. Zhu, W. Cai, M. Borysiak, B. Han, D. Chen, R. D.Piner, L. Colombo, and R. S. Ruoff, Nano Lett. , 4359 (2009). S. Bae, H. Kim, Y. Lee, X. Xu, J.-S. Park, Y. Zheng, J. Balakr-ishnan, T. Lei, H. R. Kim, Y. I. Song, Y.-J. Kim, K. S. Kim,B. Ozyilmaz, J.-H. Ahn, B. H. Hong, and S. Iijima, Nat. Nan-otechnol. , 574 (2010). M. Takai, D. Bollmann, and K. Haberger, Appl. Phys. Lett. ,2560 (1994). O. Yavas and M. Takai, Appl. Phys. Lett. , 2558 (1998). F. Wakaya, T. Kurihara, S. Abo, and M. Takai, Microelectron.Eng. , 358 (2013). D. Graf, F. Molitor, K. Ensslin, C. Stampfer, A. Jungen, C. Hi-erold, and L. Wirtz, Nano Lett. , 238 (2007). N. Arnold, in
Laser Cleaning , edited by B. Luk’yanchuk (WorldScientific, Singapore, 2002) p. 51. S. P. Koenig, N. G. Boddeti, M. L. Dunn, and J. S. Bunch, Nat.Nanotechnol. , 543 (2011). Z. Zong, C.-L. Chen, M. R. Dokmeci, and K.-T. Wan, J. Appl.Phys. , 026104 (2010). K. Toyoda,
Laser Processing ed. the Laser Society of Japan [inJapanese] (Nikkei Gijutsu Tosho, Tokyo, 1990) pp. 19–32. National Astronomical Observatory of Japan, ed.,
Rika Nenpyo(Chronological Scientific Tables) [in Japanese] (Maruzen, Tokyo,1990). A. Borghesi and G. Guizzetti, in
Handb. Opt. Constants of SolidsII , edited by E. D. Palic (Academic Press, San Diego, 1991) p.449. S. Dhar, A. R. Barman, G. X. Ni, X. Wang, X. F. Xu, Y. Zheng,S. Tripathy, A. Rusydi, K. P. Loh, M. Rubhausen, a. H. C. Neto,B. ˝Ozyilmaz, and T. Venkatesan, AIP Adv.1