Vrije Universiteit Brussel

Laser ablation- and plasma etching-based patterning of graphene on silicon-on-insulatorwaveguidesVan Erps, Jürgen Albert; Ciuk, Tymoteusz; Pasternak, Iwona; Krajewska, Aleksandra;Strupinski, Wlodek; Van Put, Steven; Van Steenberge, Geert; Baert, Kitty; Terryn, Herman;Thienpont, Hugo; Vermeulen, NathaliePublished in:Opt. Express

DOI:10.1364/OE.23.026639

Publication date:2015

Document Version:Final published version

Link to publication

Citation for published version (APA):Van Erps, J. A., Ciuk, T., Pasternak, I., Krajewska, A., Strupinski, W., Van Put, S., ... Vermeulen, N. (2015).Laser ablation- and plasma etching-based patterning of graphene on silicon-on-insulator waveguides. Opt.Express, 23(20), 26639-26650. https://doi.org/10.1364/OE.23.026639

General rightsCopyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright ownersand it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.

• Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying the publication in the public portalTake down policyIf you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediatelyand investigate your claim.

Download date: 11. Dec. 2020

Laser ablation- and plasmaetching-based patterning of graphene on

silicon-on-insulator waveguides

Jurgen Van Erps,1,∗ Tymoteusz Ciuk,2 Iwona Pasternak,2

Aleksandra Krajewska,2 Wlodek Strupinski,2 Steven Van Put,3

Geert Van Steenberge,3 Kitty Baert,4 Herman Terryn,4

Hugo Thienpont,1 and Nathalie Vermeulen1

1Vrije Universiteit Brussel, Brussels Photonics Team (B-PHOT), Department of AppliedPhysics and Photonics, Pleinlaan 2, B-1050 Brussel, Belgium

2Institute of Electronic Materials Technology, Wolczynska 133, 01-919 Warsaw, Poland3Centre for Microsystems Technology (CMST), imec and Ghent University, Technologiepark

914A, B-9052 Gent, Belgium4Vrije Universiteit Brussel, Research Group Electrochemical and Surface Engineering

(SURF), Pleinlaan 2, B-1050, Belgium∗jverps@vub.ac.be

Abstract: We present a new approach to remove monolayer graphenetransferred on top of a silicon-on-insulator (SOI) photonic integrated chip.Femtosecond laser ablation is used for the first time to remove graphenefrom SOI waveguides, whereas oxygen plasma etching through a metalmask is employed to peel off graphene from the grating couplers attachedto the waveguides. We show by means of Raman spectroscopy and atomicforce microscopy that the removal of graphene is successful with minimaldamage to the underlying SOI waveguides. Finally, we employ bothremoval techniques to measure the contribution of graphene to the lossof grating-coupled graphene-covered SOI waveguides using the cut-backmethod.

© 2015 Optical Society of America

OCIS codes: (140.7090) Ultrafast lasers; (140.3390) Laser materials processing; (130.3120)Integrated optics devices.

References and links1. J.L. Cheng, N. Vermeulen, and J.E. Sipe, “Third order optical nonlinearity of graphene,” New J. Phys. 16, 053014

(2014).2. J.L. Cheng, N. Vermeulen, and J.E. Sipe, “DC current induced second order optical nonlinearity in graphene,”

Opt. Express 22, 15868–15876 (2014).3. Z. Cheng, H. K. Tsang, X. Wang, K. Xu, and J.-B. Xu,“In-plane optical absorption and free carrier absorption in

graphene-on-silicon waveguides,” IEEE J. Sel. Top. Quant. Electron. 20, 4400106 (2014).4. Y. Ye, L. Gan, L. Dai, Y. Dai, X. Guo, H. Meng, B. Yu, Z. Shi, K. Shang, and G. Qin, “A simple and scal-

able graphene patterning method and its application in CdSe nanobelt/graphene Schottky junction solar cells,”Nanoscale 3, 1477–1481 (2011).

5. 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, “Large-scale pattern growth of graphene films for stretchable transparent electrodes,” Nature 457, 706–710 (2009).

6. A. Pirkle, J. Chan, A. Venugopal, D. Hinojos, C.W. Magnuson, S. McDonnell, L. Colombo, E.M. Vogel,R.S. Ruoff, and R.M. Wallace, “The effect of chemical residues on the physical and electrical properties ofchemical vapor deposited graphene transferred to SiO2,” Appl. Phys. Lett. 99, 122108 (2011).

7. S.-C. Jeon, Y.-S. Kim, and D.-K. Lee, “Fabrication of a graphene nanoribbon with electron beam lithographyusing a XR-1541/PMMA lift-off process,” Trans. Electr. Electron. Mat. 11, 190–193 (2010).

(C) 2015 OSA 5 Oct 2015 | Vol. 23, No. 20 | DOI:10.1364/OE.23.026639 | OPTICS EXPRESS 26639 #248327 Received 21 Aug 2015; revised 25 Sep 2015; accepted 27 Sep 2015; published 1 Oct 2015

8. R. Sahin, E. Simsek, and S. Akturk, “Nanoscale patterning of graphene through femtosecond laser ablation,”Appl. Phys. Lett. 104, 053118 (2014).

9. F. Wakaya, T. Kurihara, S. Abo, and M. Takai, “Ultra-violet laser processing of graphene on SiO2/Si,” Micro-electron. Eng. 110 358–360 (2013).

10. I. Bobrinetskiy, A.V. Emelianov, N. Otero, and P.M. Romero, “Patterned graphene ablation and two-photonfunctionalization by picosecond laser pulses in ambient conditions,” Appl. Phys. Lett. 107, 043104 (2015).

11. R. Sahin, S. Akturk, and E. Simsek, “Quantifying the quality of femtosecond laser ablation of graphene,”Appl. Phys. A 116 555–560 (2014).

12. T. Ciuk, I. Pasternak, A. Krajewska, J. Sobieski, P. Caban, J. Szmidt, and W. Strupinski, “Properties of chemi-cal vapor deposition graphene transferred by high-speed electrochemical delamination,” J. Phys. Chem. C 117,20833–20837 (2013).

13. A. Gupta, G. Chen, P. Joshi, S. Tadigadapa, and P.C. Eklund, “Raman scattering from high-frequency phononsin supported n-graphene layer films,” Nano Lett. 6, 2667–2673 (2006).

14. A.C. Ferrari, J.C. Meyer, V. Scardaci, C. Casiraghi, M. Lazzeri, F. Mauri, S. Piscanec, D. Jiang, K.S. Novoselov,S. Roth, and A.K. Geim, “Raman spectrum of graphene and graphene layers,” Phys. Rev. Lett. 97, 189401 (2006).

15. C. Casiraghi, S. Pisana, K.S. Novoselov, A.K. Geim, and A.C. Ferrari, “Raman fingerprint of charged impuritiesin graphene,” Appl. Phys. Lett. 91, 233108 (2007).

16. D.S. Lee, C. Riedl, B. Krauss, K. von Klitzing, U. Starke, J.H. Smet, “Raman spectra of epitaxial graphene onSiC and of epitaxial graphene transferred to SiO2,” Nano Lett. 9, 4320–4325 (2008).

17. H. Cao, Q. Yu, L.A. Jauregui, J. Tian, W. Wu, Z. Liu, R. Jalilian, D.K. Benjamin, Z. Jiang, J. Bao, S.S.S. Pei,Y.P. Chen, “Electronic transport in chemical vapor deposited graphene synthesized on Cu: quantum Hall effectand weak localization,” Appl. Phys. Lett. 96, 122106 (2010).

18. Y.Y. Wang, Z.H. Ni, T. Yu, Z.X. Shen, H.M. Wang, Y.H. Wu, W. Chen, and A.T.S. Wee, “Raman studies ofmonolayer graphene: the substrate effect,” J. Phys. Chem. C 112, 10637–10640 (2008).

19. D. Graf, F. Molitor, K. Ensslin, C. Stampfer, A. Jungen, C. Hierold, and L. Wirtz, “Spatially resolved Ramanspectroscopy of single- and few-layer graphene,” Nano Lett. 7, 238–242 (2007).

20. J.M. Liu, “Simple technique for measurements of pulsed Gaussian-beam spot sizes,” Opt. Lett. 7, 196–198(1982).

21. W. Zhang, L. Li, Z.B. Wang, A.A. Pena, D.J. Whitehead, M.L. Zhong, Z. Lin, and H.W. Zhu, “Ti:Sapphirefemtosecond laser direct micro-cutting and profiling of graphene,” Appl. Phys. A 109, 291–297 (2012).

22. J. Hoffman, J. Chrzanowska, S. Kucharski, T. Moscicki, I.N. Mihailescu, C. Ristoscu, and Z. Szymanski, “Theeffect of laser wavelength on the ablation rate of carbon,” Appl. Phys. A 117, 395–400 (2014).

23. H. Al-Mumen, F. Rao, W. Li, and L. Dong, “Singular sheet etching of graphene with oxygen plasma,” Nano-Micro Lett. 6, 116–124 (2014).

24. I. Childres, L.A. Jauregui, J. Tian, and Y.P. Chen, “Effect of oxygen plasma etching on graphene studied usingRaman spectroscopy and electronic transport measurements,” New J. Phys. 13, 025008 (2011).

25. Y.A. Vlasov and S.J. McNab, “Losses in single-mode silicon-on-insulator strip waveguides and bends,” Opt. Ex-press 12, 1622–1631 (2014).

26. H. Li, Y. Anugrah, S.J. Koester, and M. Li, “Optical absorption in graphene integrated on Silicon waveguides,”Appl. Phys. Lett. 101, 111110 (2012).

1. Introduction

In recent years graphene, a two-dimensional hexagonal lattice of carbon atoms, has been shownto exhibit unique optical properties that can strongly improve the functioning of photonic de-vices and even of entire photonic circuits [1, 2]. When depositing graphene on chip-scale pho-tonic circuits consisting of silicon waveguides, a high-resolution patterning method to locallyremove the graphene layer without affecting the underlying silicon waveguide is desired formost applications. Indeed, an accurate control of the length of the graphene strips on top of thewaveguides is of high importance, since the graphene induces a high loss, especially when thegraphene sheet covers the grating couplers that are used to couple light in and out of the on-chipwaveguides [3].

High-resolution patterning of graphene on silicon waveguides is typically established usingphotolithography or electron beam lithography [4,5]. Yet there are important limitations associ-ated with these techniques, such as minimal wafer size requirements which prohibit chip-scaleprocessing. Most importantly, however, the subsequent steps of lithography may seriously dam-age the underlying graphene sheet or at least significantly alter its electrical and optical proper-ties through chemicals-induced unintentional doping [6, 7]. A promising alternative technique

(C) 2015 OSA 5 Oct 2015 | Vol. 23, No. 20 | DOI:10.1364/OE.23.026639 | OPTICS EXPRESS 26640 #248327 Received 21 Aug 2015; revised 25 Sep 2015; accepted 27 Sep 2015; published 1 Oct 2015

Fig. 1. Raman spectrum measured after graphene transfer on the SOI waveguides, showingthe monolayer nature of the transferred graphene.

for realizing high-resolution graphene patterning without employing chemicals is laser abla-tion [8–10]. The viability of femtosecond laser ablation has already been demonstrated forgraphene deposited on SiO2, but ablation of graphene directly deposited on silicon has not yetbeen reported. Proper selection of the laser parameters is vital to avoid damage to the underlyingsubstrate [11]. As opposed to SiO2, silicon absorbs infrared light with wavelengths shorter than1100 nm and thus is more prone to laser-induced damage at these wavelengths. Furthermore,laser ablation of graphene on waveguide structures –regardless the material used– has to ourknowledge never been investigated, but has great potential for providing the required controlover the length of the graphene-covered waveguide sections in integrated optics applications.

In this paper we show that laser ablation can successfully remove graphene on a siliconwaveguide without deteriorating the waveguide shape. We also find that laser ablation is notsuitable for peeling off graphene from the grating couplers attached to the waveguide input andoutput to enable light incoupling and outcoupling. Instead, oxygen plasma etching through amechanical mask can be employed for this purpose. As a case study, we apply both grapheneremoval techniques to determine the linear loss of grating-coupled graphene-covered siliconwaveguides along the cut-back method.

2. Monolayer graphene transfer

Monolayer graphene was grown on high-purity polycrystalline copper foil in a commercialBlack Magic system by the chemical vapor deposition (CVD) method. The pre-growth treat-ment of copper included annealing at 1000◦C in Ar atmosphere followed by H2 etching at apressure of 20 mbar [12]. The purpose of this step is to improve the quality and enlarge thegrain size of copper as well as to ensure a reduction of oxides from the copper surface. Later,methane was introduced into the reactor with time and flow rate settings ensuring monolayergraphene growth. The as-grown graphene was covered with a 200 nm-thin layer of PMMAby spin-coating and transferred onto an SOI photonic chip by means of the electrochemicaldelamination method carried out in 1M aqueous solution of potassium chloride [12].

The monolayer character of the transferred graphene was confirmed by Raman spectroscopyperformed using a Renishaw inVia microscope powered by a 532 nm continuous-wave (CW)frequency-doubled Nd:YAG laser. A laser power of 5 mW and a spot size of 0.4 μm (obtainedby means of a 100× objective) were used. The shape, position and width of the G and 2D peaksin the Raman spectrum can serve to distinguish mono- and bilayer graphene [13–16]. A typical

(C) 2015 OSA 5 Oct 2015 | Vol. 23, No. 20 | DOI:10.1364/OE.23.026639 | OPTICS EXPRESS 26641 #248327 Received 21 Aug 2015; revised 25 Sep 2015; accepted 27 Sep 2015; published 1 Oct 2015

Fig. 2. Experimental measurement of the squared diameter of the ablated areas (D2) as afunction of the peak fluence to determine the ablation threshold, which is the intersectionof the fitted curve with the X-axis. The smallest measured diameter of the ablated spotwas 8 μm for graphene on SiO2, 5 μm for graphene on SOI, and 4 μm for bare SOI. ANikon Optiphot 200 high-resolution, high-magnification optical microscope was used todetermine D2.

Raman spectrum of transferred graphene with indicated D, G and 2D bands is shown in Fig. 1.For that spectrum the 2D band position is 2673 cm−1 and G band position is 1582 cm−1. Thisis consistent with other reports on monolayer CVD graphene [6, 17, 18]. The full width at halfmaximum (FWHM) of the 2D band (about 38 cm−1) as well as the ratio of the peak intensities(I2D/IG ≈ 2) are also typical for monolayer graphene [19].

3. Femtosecond laser ablation

Graphene ablation was investigated using an Yb-doped fiber laser (Satsuma, Amplitude Sys-tems, 5 W, 1030 nm, 400 fs). During the ablation experiments, the pulse repetition rate wasfixed to 10 kHz, whereas the laser pulse energy and the amount of shots (i.e. the laser scanningspeed) were varied. An objective lens with a focal length of 60 mm and a numerical aperture(NA) of 0.2 was used to focus the laser beam and all ablation experiments were carried out in airat room temperature. The beam waist of the laser beam was determined to be w0 = 9.9 μm [20].

We started with investigating laser ablation of a monolayer graphene sheet (with the PMMAlayer removed) on an SiO2 substrate. The graph in Fig. 2 shows the squared diameter of theablated areas (D2) as a function of the peak fluence for single shots. From this graph, theablation threshold was determined to be 0.076 J/cm2. This ablation threshold is higher than thevalue found by Zhang et al. [21], but we also worked at a longer wavelength, namely 1030 nmversus 800 nm. This seems to be in line with earlier wavelength dependence studies [22]. Thesmallest measured diameter of the ablated spot was 8 μm, which in combination with high-precision alignment stages enables patterning of graphene with micron-scale resolution.

Next, we investigated the ablation of graphene deposited on an SOI substrate with a220nm-thick top silicon layer, and we found that the ablation threshold for graphene removal,i.e. 0.147 J/cm2 (Fig. 2), was a factor 2 higher for graphene on silicon than for graphene on SiO2

(0.076 J/cm2), while the damage threshold for bare SOI was 0.546 J/cm2 (Fig. 2). The small-est measured diameter of the ablated spot was 5 μm for graphene on SOI and 4 μm for bareSOI (Fig. 2). Having determined the threshold for ablation of graphene on a blank SOI sample,we proceeded to the local removal of graphene on a waveguide-containing SOI photonic chipfeaturing a 220 nm-thick top silicon layer. Notice that since the absorption depth of siliconat 1030 nm is much greater than 10 μm, we expected to induce less damage to the silicon at1030 nm as compared to shorter wavelengths, especially since the silicon layer is only 220 nmthick. The chip contained 450 nm-wide waveguides fabricated in the 220 nm-thick silicon top

(C) 2015 OSA 5 Oct 2015 | Vol. 23, No. 20 | DOI:10.1364/OE.23.026639 | OPTICS EXPRESS 26642 #248327 Received 21 Aug 2015; revised 25 Sep 2015; accepted 27 Sep 2015; published 1 Oct 2015

Fig. 3. Raman spectra measured in the ablated and non-ablated regions of a graphene-covered SOI waveguide.

layer, which are the standard waveguide dimensions for operation at the telecom wavelengthof 1550 nm. The waveguides’ grating couplers were also designed for 1550 nm. We found afluence of 0.16 J/cm2 and a laser scanning speed of 70 mm/s to be the optimal parameters forcontinuous removal of graphene from the silicon waveguides. The threshold for graphene abla-tion on the SOI waveguides (i.e. 0.137 J/cm2) was found to be slightly lower than the thresholdfor graphene ablation on bulk SOI (i.e. 0.147 J/cm2). As a result, the diameter of the ablatedspot on the waveguides at our working point of 0.16 J/cm2 is slightly larger (i.e. 7 μm) than thecorresponding spot diameter found for ablation of graphene on bulk SOI (i.e. 5 μm). A scan-ning speed of 70 mm/s for a free-running laser with a repetition rate of 10 kHz corresponds toablation in the single pulse regime when taking into account this 7 μm ablation spot diameterfor graphene on SOI waveguides.

To show that the local removal of graphene was successful, we measured the Raman spec-trum at several positions along the waveguide that was ablated and at positions that were notablated. The Raman spectra were obtained using a Horiba Scientific LabRam HR with a 532 nmCW frequency-doubled Nd:YAG laser and 0.25 mW laser power (notice that the Raman spec-tra for the ablation experiments were obtained with a different Raman microscope than theone used after graphene transfer and the plasma etching experiments because we measured thespectra directly after the ablation process, which was carried out at a different location than theplasma etching). The spot size was 0.7 μm, obtained with a 100× objective (NA of 0.9). Figure3 shows a typical result of these measurements. Whereas in the blue curve the characteristicpeaks clearly show the presence of monolayer graphene, the red curve shows that there is nographene remaining on top of the waveguide after laser ablation. In addition, atomic force mi-croscopy (AFM) was used to check the geometrical shape and surface roughness of the SOIwaveguides after laser ablation, to identify potential damage to the waveguide during ablation.A high aspect ratio tip was used in tapping mode (256 × 256 data points, 1 Hz scan rate) toacquire the AFM images in Figs. 4 and 5, showing the surface profile and cross-section of anSOI waveguide before and after ablation respectively. The root mean square (RMS) surfaceroughness Rq of a graphene-covered SOI waveguide, measured in an area of 400 nm × 400 nm,was 0.57 nm (Fig. 4). After ablation, we measured an RMS surface roughness of Rq = 2.12 nm(Fig. 5), which means that the ablation-induced increase in surface roughness equals 1.55 nm.We found that the parameter window for graphene removal with such small damage to the un-derlying waveguide is relatively small: the lower limit is determined by the ablation threshold

(C) 2015 OSA 5 Oct 2015 | Vol. 23, No. 20 | DOI:10.1364/OE.23.026639 | OPTICS EXPRESS 26643 #248327 Received 21 Aug 2015; revised 25 Sep 2015; accepted 27 Sep 2015; published 1 Oct 2015

Fig. 4. AFM measurement showing the waveguide profile before laser ablation (left). Theright part shows a cross-section of the waveguide along the blue line in the left AFM image(bottom) and a zoom on the 400 nm × 400 nm area indicated by the dashed square in theleft AFM image (top). The measured surface roughness is 0.57 nm.

Fig. 5. AFM measurement showing the waveguide profile after laser ablation (left). Theentire area shown in this measurement has been ablated. The right part of the figure showsa cross-section of the ablated waveguide along the blue line in the left AFM image (bottom)and a zoom on the 400 nm × 400 nm area indicated by the dashed square in the left AFMimage (top). The measured surface roughness is 2.12 nm.

(C) 2015 OSA 5 Oct 2015 | Vol. 23, No. 20 | DOI:10.1364/OE.23.026639 | OPTICS EXPRESS 26644 #248327 Received 21 Aug 2015; revised 25 Sep 2015; accepted 27 Sep 2015; published 1 Oct 2015

1mm

50µm

Fig. 6. Microscope image of the photonic chip on which a PMMA/graphene layer has beentransferred, which is then selectively O2 plasma etched to create openings in the layeron top of the grating couplers and tapered sections of the SOI waveguides. The dashedrectangle refers to the area of the chip which was used for the cut-back measurements (seeSection 5). Left: zoom on three grating couplers in an etched opening, and the transitionto a graphene-covered region (i.e. the slightly darker stripe that runs vertically across thefigure).

for graphene on SOI waveguides (i.e. 0.137 J/cm2), whereas the upper limit is found to be0.16 J/cm2 since the ablation-induced surface roughness increases rapidly for ablation fluenceslarger than this value.

We also observed that whereas for the waveguides the removal of graphene with laser ab-lation was successful, it was not possible to achieve the same result on the grating couplers.We believe that resonant reflections of the laser ablation beam on the gratings results in dam-age to the gratings even at very low ablation powers. For the tapered SOI sections betweenthe waveguides and the grating couplers, we found that the level of adhesion of the graphenelayer depends on the local width of the tapers, so that removing the graphene along the ta-pers would require a gradual variation of the laser ablation settings along the taper. To avoidthis and to solve the issue of graphene removal from the grating couplers, we used O2 plasmaetching [23, 24] to peel off the graphene from the grating couplers and the tapers.

4. Oxygen plasma etching

As opposed to laser ablation, the plasma etching step can be performed before the PMMA layerused for the graphene transfer is washed off of the graphene sheet. The O2 plasma etching wascarried out in a 13.56 MHz chemically reactive plasma etcher. The oxygen flow was set to5 sccm and the pressure reached 0.3 mbar. At 50 W of applied RF power the etching continuedfor 30 min. A mechanical mask was designed and laser-cut in a 200 μm-thin steel plate to definethe openings to be etched. Taking into account that the minimal attainable size of the openings isabout 100 μm, this approach of O2 plasma etching with a mechanical mask is particularly suitedfor the removal of graphene on larger areas such as in this case grating couplers and waveguidetapers. The alignment of the mask with respect to the chip proved a challenging task and wascarried out with the help of custom-adapted pick-and-place equipment which allowed us toachieve an alignment accuracy of 10 μm. The patterning resolution for graphene removal with

(C) 2015 OSA 5 Oct 2015 | Vol. 23, No. 20 | DOI:10.1364/OE.23.026639 | OPTICS EXPRESS 26645 #248327 Received 21 Aug 2015; revised 25 Sep 2015; accepted 27 Sep 2015; published 1 Oct 2015

Fig. 7. Raman spectra measured in a graphene-covered region (blue) and in the O2 plasmaetched grating coupler area (red).

(a) (b)

Fig. 8. AFM measurement showing the grating coupler profile before (a) and after (b)plasma etching. The bottom part shows a cross-section of the grating along the line in thetop AFM image.

(C) 2015 OSA 5 Oct 2015 | Vol. 23, No. 20 | DOI:10.1364/OE.23.026639 | OPTICS EXPRESS 26646 #248327 Received 21 Aug 2015; revised 25 Sep 2015; accepted 27 Sep 2015; published 1 Oct 2015

(a) (b)

Fig. 9. AFM measurement showing the tapered waveguide section profile before (a) andafter (b) plasma etching. The bottom part shows a cross-section of the grating along theline in the top AFM image. The measured surface roughness over an area of 400 nm ×400 nm was measured to increase from 0.66 nm before plasma etching to 3.16 nm afterplasma etching.

plasma etching is thus much lower than with laser ablation. Figure 6 shows the SOI chip afterthe mask alignment has been carried out and openings have been etched in the graphene/PMMAlayer. The brown area indicates where the graphene/PMMA layer was still present and the darkblue area where it was etched away. The selectively etched graphene/PMMA layer is afterwardsrinsed in an acetone/ethanol bath to strip the PMMA layer and reveal the underlying graphene.Figure 6 also shows a zoom of three grating couplers situated within one of the etched openings.

Figure 7 shows a typical Raman spectrum of graphene outside the O2 plasma etching re-gion, showing that the remaining graphene monolayer is of good quality. In the same figure,a typical Raman spectrum taken within the plasma etched area is shown which exhibits nographene-related peaks and shows no traces of carbon residue. This clearly shows that the O2

plasma completely removes the graphene layer in the selectively etched areas. The Raman spec-tra were obtained with the same Renishaw inVia Raman microscope referred to earlier. AFMmeasurements of the gratings (Fig. 8) confirmed that while a higher surface roughness is in-troduced, the grating profile shape is not significantly degraded due to O2 plasma etching (asopposed to laser ablation). From Fig. 9 it is clear that the plasma etching process also induces anincrease in RMS surface roughness on the tapered sections of the waveguides to Rq =3.16 nmfor an area of 400 nm × 400 nm (i.e. an increase of 2.5 nm compared to the surface roughnessmeasured before plasma etching). This could be improved by removing the PMMA layer priorto the plasma etching to shorten the plasma etching duration.

(C) 2015 OSA 5 Oct 2015 | Vol. 23, No. 20 | DOI:10.1364/OE.23.026639 | OPTICS EXPRESS 26647 #248327 Received 21 Aug 2015; revised 25 Sep 2015; accepted 27 Sep 2015; published 1 Oct 2015

Fig. 10. Schematic lay-out of the laser ablation and O2 plasma etching regions on an arrayof identical graphene-covered waveguides (indicated by the dashed rectangle in the rightpart of Fig. 6), to be used for cut-back measurements.

5. Measurement of linear loss of graphene-covered waveguides using cut-back method

To showcase its practical use for optical experiments, we also employed our combined laserablation and plasma etching technique for determining the optical transmission properties ofgraphene-covered silicon waveguides. More specifically, we have plasma etched and laser ab-lated the gratings and waveguides according to the pattern shown in Fig. 10. This way, thelength of the graphene-covered waveguide is gradually increased from 0 μm (i.e. the grapheneis removed over the full waveguide length) to 300 μm such that we can determine the linearpropagation losses by means of the so-called cut-back method [25].

The experimental setup is schematically illustrated in Fig. 11. To couple light between single-mode fibers and the silicon waveguides, we positioned the fibers above the grating couplers atan angle of 9◦ with respect to the vertical axis for optimal coupling efficiency. To be ableto monitor how much power we couple into the chip, we split off 1% of the light comingfrom the laser source. A variable optical attenuator (VOA) is used to accurately control thepower to be coupled into the chip, and just prior to entering the chip, the light passes througha fiber polarization controller (PC) to align the polarization to the desired Transverse-Electricmode of the SOI waveguides. As a reference, we first measured the grating coupler efficiencybefore graphene transfer on the chip. The source employed is a continuous-wave wavelength-division multiplexing (WDM) telecom laser (emitting at the wavelength of 1550 nm for whichthe grating couplers were designed) connected to an Erbium-doped fiber amplifier (EDFA) setto 1mW output power. The total insertion loss, including the grating coupler loss at input andoutput as well as the waveguide propagation loss over 800 μm (the latter comprises a 400 μm-long waveguide section and two 200 μm-long tapered sections) was found to be 13.5 dB. Sincethe propagation loss of the waveguides is of the order of 3 dB/cm [25], the insertion loss pergrating is estimated to be 6.6 dB when using flat-cleaved fibers to couple light in and out of theSOI waveguides.

Then we performed the cut-back experiments to measure the contribution of graphene to theloss of grating-coupled graphene-covered SOI waveguides. Figure 12 shows the total insertion

(C) 2015 OSA 5 Oct 2015 | Vol. 23, No. 20 | DOI:10.1364/OE.23.026639 | OPTICS EXPRESS 26648 #248327 Received 21 Aug 2015; revised 25 Sep 2015; accepted 27 Sep 2015; published 1 Oct 2015

Out-of-planecoupling setup

WDMSource

Fiber1:99

1%

Fiber

VOA

PCFiber

Chip under test

Powermeter

Powermeter99%

EDFA

Fig. 11. Schematic of the experimental setup to measure the linear loss properties ofgraphene-covered waveguides (WDM: wavelength-division multiplexing; EDFA: Erbium-doped fiber amplifier; VOA: variable optical attenuator; PC: polarization controller).

y = -0.1309x - 18.503R² = 0.9923

-70.0

-60.0

-50.0

-40.0

-30.0

-20.0

-10.0

0.00 50 100 150 200 250 300 350

Inse

rtio

n lo

ss (d

B)

Length of graphene strip (μm)

Fig. 12. Results of the linear transmission cut-back experiments.

loss (i.e. the waveguide loss and the coupling loss at both grating couplers) versus the lengthof the graphene strip in-between the ablated zones. The leftmost data point in Fig. 12 shows atotal insertion loss of 18.3 dB for a waveguide which was ablated over its full length. This isabout 4.8 dB higher than the insertion loss of 13.5 dB measured before the graphene transferand patterning through laser ablation and plasma etching. This difference is due to the plasmaetching-induced increase in surface roughness of the grating couplers and the tapered waveg-uide sections (which can be observed in Figs. 8 and 9), and the ablation-induced increase insurface roughness of the silicon waveguide (which can be observed in Fig. 5).

When applying a linear regression on all datapoints in Fig. 12, we find that the contributionof graphene to the propagation loss for the graphene-covered waveguides is about 13.2 dB per100 μm of graphene, with a very good R2 value of 0.9925 for the fit. This loss of 0.132 dB/μmis in line with other experimental results reported by Li et al. [26]. The fact that the graphene-induced loss is so high clearly underlines the necessity of local patterning of the graphenetransferred on top of SOI waveguides when targeting on-chip photonic applications. Li et al.also showed that the linear loss caused by the graphene layer on top of a waveguide stronglydepends on the waveguide geometry [26]. This is because for different geometries the electricfield strength in the graphene cover layer, and as such the interaction between the graphene andthe waveguide mode, is different.

6. Conclusions

In conclusion, we have demonstrated the successful removal of a monolayer of graphene onSOI waveguides through femtosecond laser ablation at 1030 nm. Because laser ablation of thegrating couplers was not successful, we employed O2 plasma etching through a metal maskto remove graphene at these locations, as well as on the tapered waveguide sections. Atomic

(C) 2015 OSA 5 Oct 2015 | Vol. 23, No. 20 | DOI:10.1364/OE.23.026639 | OPTICS EXPRESS 26649 #248327 Received 21 Aug 2015; revised 25 Sep 2015; accepted 27 Sep 2015; published 1 Oct 2015

force microscopy revealed that laser ablation induced an increase in surface roughness of theunderlying waveguides of 1.55 nm, whereas O2 plasma etching induced an increase in surfaceroughness of the tapered waveguide sections of 2.5 nm. This resulted in a 4.8 dB insertion losspenalty for a 400 μm-long ablated waveguide section and O2 plasma etched gratings and ta-pers. Both the laser ablation and plasma etching techniques offer the important advantage thatthey do not involve additional chemical treatment of the graphene sheet. Laser ablation has thesupplementary advantage that it allows patterning of the graphene with a much higher reso-lution than plasma etching. Finally, we employed both techniques for determining the opticaltransmission of graphene-covered SOI waveguides along the cut-back method. For a standardwaveguide width of 450 nm and height of 220 nm we found the contribution of graphene tothe optical loss to be 0.132 dB/μm. This high loss value per unit distance clearly shows thatto apply graphene in photonic chip applications, accurate control of the overall loss throughlocalized patterning of the graphene is essential.

Acknowledgments

This work was supported by the EU-FET GRAPHENICS (grant agreement no. 618086),BELSPO-IAP, Methusalem, ERC-FP7/2007-2013 grant 336940 and by the EU-FP7 GrapheneFlagship (grant agreement no. 604391). The work of J. Van Erps was supported by the ResearchFoundation Flanders (FWO-Vlaanderen) under a postdoctoral research fellowship.

(C) 2015 OSA 5 Oct 2015 | Vol. 23, No. 20 | DOI:10.1364/OE.23.026639 | OPTICS EXPRESS 26650 #248327 Received 21 Aug 2015; revised 25 Sep 2015; accepted 27 Sep 2015; published 1 Oct 2015