Radiation-induced defect formation and reactivity of ... - Rutgers Physics

Radiation-induced defect formation and reactivity of model TiO2 capping layers with MMA: a comparison with Ru B. V. Yakshinskiy1, M. N. Hedhili1, S. Zalkind1*, M. Chandhok2 and Theodore E. Madey1** 1 Dept. of Physics and Astronomy, and Laboratory for Surface Modification, Rutgers, The State University of New Jersey, Piscataway NJ 08854 USA 2 Intel Corporation, Hillsboro OR 97124 USA ABSTRACT Our goal is to provide insights into surface processes that affect the reflectivity and lifetimes of TiO2 and Rucapped multilayer mirrors used in extreme ultraviolet (EUV) lithography. Several surface-sensitive ultrahigh-vacuum techniques are used to characterize thermally-induced and electron-induced surface reactions of methyl methacrylate (MMA), a model for hydrocarbons found in EUV lithography vacuum chambers; low-energy electron beams are used to mimic excitations initiated by EUV radiation. Carbon film growth is measured on TiO2 surfaces during electron bombardment (at 20 eV and 100 eV) in MMA vapor; C growth rates are compared on Ru surfaces. The initial rates on the clean surfaces are very different: a C film grows more rapidly on TiO2 than on Ru. However, the limiting growth rates are the same for C thicknesses greater than ~1 to 1.5 nm, when MMA interacts with a C film. Irradiation of the C films in O2 gas, or in MMA + O2, has a mitigating effect on TiO2 surfaces. Keywords: Extreme ultraviolet lithography (EUVL); EUV optics contamination; EUV optics lifetime; TiO2; ruthenium; methyl methacrylate (MMA); electron-induced reactions; carbon.

1. INTRODUCTION Irradiation of Ru-capped multilayer mirrors (MLMs) in EUV exposure tools can lead to accumulation of surface contamination, due to radiation-induced surface chemistry in the inevitable background pressure of hydrocarbons and water vapor [1-5]. Recently, to examine the effects of irradiation on the accumulation of C on initially-clean Ru surfaces, we have worked with model single-crystal surfaces; we have bombarded the surfaces with low-energy electrons (substitutes for EUV photons) while exposing them to methyl methacrylate (MMA, C5H8O2) vapor, a model hydrocarbon outgassing product, and characterized the C film formation [6-8]. Several surface-sensitive ultrahighvacuum techniques have been used to characterize the surface reaction processes, including x-ray photoelectron spectroscopy (XPS), temperature programmed desorption (TPD), and electron stimulated desorption (ESD). In the present paper, we focus on contamination and mitigation of another effective capping layer material for MLMs, TiO2 [9]. Bombardment of a crystalline TiO2(011) surface by electrons with energies >25 eV causes desorption of O and O+, creating a surface with oxygen vacancies (O-defects) that are quantified using XPS. Via a similar mechanism, EUV photons (13.5 nm) can induce O-vacancies by direct core level ionization of TiO2. MMA adsorbs and desorbs in molecular form on stoichiometric TiO2, but when TiO2 is bombarded by low energy (20 eV) electrons in the presence of gaseous MMA (~10-8 Torr), reaction occurs to form a surface carbide and to accumulate a thermally-stable carbon film. Such low energies are characteristic of the secondary electrons released by EUV photons on TiO2 cap layers. Thus, we find evidence for direct radiation damage that can be induced also by EUV photons (O-vacancy formation) and indirect damage by secondary electrons (carbon accumulation caused by low energy secondary electroninduced reactions). In comparing carbon film growth induced by ~100 eV electron bombardment on TiO2 with similar ________________________________________________________________________________________________ * Sabbatical leave 2007-2008. Permanent address N.R.C.N., POB 9001, Beer Sheva, Israel ** Author for correspondence. Email: [email protected]

Emerging Lithographic Technologies XII, edited by Frank M. Schellenberg Proc. of SPIE Vol. 6921, 692114, (2008) · 0277-786X/08/$18 · doi: 10.1117/12.772798

Proc. of SPIE Vol. 6921 692114-1

studies on Ru(10 1 0) we find that the initial rates on the clean surfaces are very different: a C film grows more rapidly on TiO2 than on Ru. However, the limiting growth rates are the same for C thicknesses greater than ~ 1 to 1.5 nm, when the MMA incident from the gas phase interacts with a C layer that completely covers the substrate. Finally, we have performed experiments related to possible mitigation processes. For TiO2, we see mitigation benefits from coexposure to MMA + O2 for increasing O2/MMA pressures ratios; measurements extended to a pressure ratio ~600. That is, the C accumulation rate (in ~10-8 Torr MMA) is greatly reduced for increasing oxygen coexposure, up to >10-6 Torr. Moreover, if we bombard the C-covered TiO2 and Ru surfaces by 100 eV electrons in ~3×10-7 Torr O2 in the absence of gaseous MMA, carbon is removed. 2. EXPERIMENTAL PROCEDURES The measurements are made in a stainless-steel ion-pumped and turbo-pumped ultrahigh vacuum (UHV) chamber at Rutgers (base P 1500 K. Cleaning of each surface is accomplished by a combination of Ar ion bombardment and annealing in oxygen. Temperature programmed desorption (TPD) is used to measure the desorption temperatures and binding energy of adsorbates as well as the surface coverage. To optimize thermal contact for TPD from TiO2, the TiO2 sample is mounted onto a thin Ti heating plate that insures uniform heating. As the surface is heated at a linear rate, the desorbing atoms are detected using a quadrupole mass spectrometer (QMS). The temperature at which the desorption rate is a maximum is related to the binding energy. The quadrupole mass spectrometer is also used for the electron stimulated desorption (ESD) of ions from the surfaces. To examine the effects of electron irradiation on the buildup of C on an initially-clean TiO2 (Ru) surface, we have bombarded the entire surface with low energy electrons from a defocused low energy source (~20 eV to ~100eV) while exposing it to MMA vapor from a molecular beam doser. C film thicknesses are determined using XPS, by measurements of the Ti 2p3/2 (Ru 3d5/2 ) attenuation as a function of electron bombardment dose. Pressures of MMA (and O2, for mitigation studies) are measured using an uncalibrated Bayard-Alpert ionization gauge, and the pressures reported here are local pressures at the sample surface, corrected only for the factor ~10× enhancement of the doser. The gauge sensitivity factor for MMA is ~5 times greater than for N2 [11]. Measurements are made for a range of MMA (O2) pressures, from 10-10 to 10-6 Torr.

3. RESULTS AND DISCUSSION 3.1. Oxygen vacancy formation on TiO2 by electron bombardment Photon-stimulated desorption (PSD) and electron-stimulated desorption (ESD) may cause loss of oxygen from the surface of oxides. ESD and PSD processes are a concern for EUV lithography: the main radiation damage mechanism in this case is the formation of O vacancies, i.e., creation of surface defects. Residual gas molecules in the background vacuum (hydrocarbons, water, etc.) are generally more reactive with defects than with fully oxidized (stoichiometric) surfaces, so the surface chemistry may be changed. ESD/PSD from maximal-valency oxides (e.g., TiO2, MoO3, CeO2,.) have high cross sections for radiationstimulated desorption of oxygen, in a process initiated by creation of holes in shallow core levels (e.g., Ti 3p at a binding energy ~35 eV for TiO2) [12, 13]. Moreover, PSD and ESD of O are observed even for non-maximal valency oxides (e.g., RuO2) and an Auger stimulated desorption mechanism is invoked; the key factor is the localization of the

Proc. of SPIE Vol. 6921 692114-2

repulsive electronic excitation. There are many subtleties in the details of excitation processes that cause desorption from oxides [14]. We verify that low energy electrons cause desorption of O+, and O from TiO2 (011). Figure 1 shows a series of spin-orbit-split Ti 2p XPS spectra that appear during electron bombardment of a clean TiO2 (011) surface in UHV. Data are taken for grazing emission angle, to enhance surface sensitivity. The dominant peak at 459 eV is Ti 2p3/2 associated with the Ti+4 valence state of TiO2. The curves correspond to different cumulative electron bombardment doses (I=100 µA, V=100 V) and are all normalized to the peak intensity for Ti 2p3/2 at 459 eV. The notable change with increasing electron dose is the appearance of a shoulder on the low binding energy side of each of the Ti 2p features. The inset contains a difference spectrum (difference between the 2.5×1018 el/cm2 bombarded surface and the defect-free surface); the features at 457 eV and 455.5 eV are associated with the appearance of Ti3+ and Ti2+ species on the surface, due to O vacancies created by electron-induced removal of O from the surface [15]. In a recent paper, we have shown that the O vacancies formed by electron bombardment of TiO2 (011) are not randomly distributed over the surface but are clustered in one-dimensional vacancy arrays; this unusual behavior is associated with a local site-dependent excitation mechanism [16].

Normalized intensity

1.0

Ti 2p Ee = 100 eV

Difference spectrum 0.4

Electron dose ×1018 el/cm2 2.5 0.8 1.3 0.9 0.6 0.6 0.3 defect free surface

Ti3+ 0.3

Ti2+ 0.2

0.1

0.0

460

459

458

457

456

455

454

453

0.4

0.2

0.0 470

465

460

455

450

445

Binding energy (eV)

Figure 1. XPS of Ti 2p showing increase of defect concentration on TiO2(011) vs electron bombardment dose at 100 eV.

Figure 2 shows two plots relating to O vacancy formation. Fig. 2 (left) shows the intensity of the Ti 2p3/2 defect signal (i.e, the intensity of difference spectra similar to inset, Fig. 1) as a function of electron dose, for different primary electron energies. The background vacuum during all ESD experiments is 1 nm, where the TiO2 (Ru) substrate is completely covered by C, the incoming MMA

Proc. of SPIE Vol. 6921 692114-6

molecules see only a C-surface; as a result, the C film thickness increases almost linearly with fluence at higher thicknesses independent of underlying TiO2 or Ru substrate. In the light of Fig. 4, one can understand qualitatively how the C accumulation of Fig. 5 occurs. The steady-state coverage of MMA is sufficiently high at 10-8 Torr that a substantial concentration of adsorbed MMA molecules can be dissociated by the incident electron beam. We attribute the differences in initial growth rates on clean TiO2 and Ru to different molecular lifetimes and surface chemistry on initially-clean Ru and TiO2, and to variations in electron-induced cross sections on the two surfaces, to be discussed in a forthcoming paper [20]. While we refer to the radiation-produced film as carbon, in fact, XPS indicates that there is a small contribution of O in the film, about 7 atomic %. Recall that the O/C ratio in MMA (C5H8O2) is 0.40, so the C concentration in a pure MMA film is 29 atomic %, relative to O. The inset to Fig. 5 shows a plot of C film thickness vs electron fluence for 20 eV electron excitation, for MMA/TiO2. For comparison, similar data are shown for benzene/TiO2. The notable observation is that the C film grows even for such low energy electron excitation, indicating that low energy secondaries (emitted by EUV photons) play an important roles in mirror contamination build-up. As described above for Fig. 2, 20 eV is below the threshold energy for O-vacancy production on TiO2 (011). In separate studies we find that C-accumulation induced by 20 eV electrons is accompanied by changes in Ti 2p photoemission indicative of surface carbide formation. It appears that electronic excitation of MMA by low energy secondaries is sufficient to induce dissociation, and initiate reaction with the substrate. Recent studies of H anion ESD from PMMA films reveals a low threshold energy, ~6 eV, for electroninduced C-H bond breaking in PMMA [22].

5

4

Growth of C film on TiO2 and Ru at 300K P(MMA, benzene)=10-8 Torr Ie=100µA, Ee=100eV MMA / Ru

3

2 1.0

C6H6 / TiO2

1

C film thickness, nm

C film thickness, nm

MMA / TiO2

1ML

0

Ee = 20 eV

0.8

MMA / TiO2

0.6 0.4 0.2

C6H6 / TiO2

0.0 0

2

4

6

8

10 12 14 16 18

Electron fluence (x1018 el/cm2)

0

20

40

60

80

100

120

140

Electron fluence (×1018 el/cm2) Figure 5. C film growth on TiO2 under electron bombardment in MMA; compare with growth on Ru. Parameters of the electron beam: 100 eV, 100 µA; Inset: Ee = 20 eV. Data are shown also for benzene vapor (C6H6) for comparison.

Proc. of SPIE Vol. 6921 692114-7

Figure 6 illustrates evidence for mitigation of C film growth during irradiation of the TiO2 surface in gaseous mixtures of MMA+O2. Each plot shows the initial growth (100× more O2 than MMA in the background gas inhibits C film growth substantially on TiO2. Moreover, exposing a C film to electron bombardment in O2 alone (in the absence of MMA) leads to C film removal. Other mitigation schemes to remove C involving VUV-activated oxygen, or photochemical reactions on TiO2 surfaces in oxidizing atmospheres [24], may prove to be useful in future studies.

5. ACKNOWLEDGEMENTS This work has been supported by Intel Corporation.

Proc. of SPIE Vol. 6921 692114-9

6. REFERENCES [1] S. Bajt, Z. R. Dai, E. J. Nelson, M. A. Wall, J. B. Alameda, N. Q. Nguyen, S. L. Baker, J. C. Robinson, J. S. Taylor, A. Aquila, N. V. Edwards, “Oxidation resistance and microstructure of ruthenium-capped extreme ultraviolet lithography multilayers”, J. Microlithogr. Microfabr. Microsyst. 5, 023004 (2006). [2] B. Mertens, M. Weiss, H. Meiling, R. Klein, E. Louis, R. Kurt, M. Wedowski, H. Trenkler, B. Wolschrijn, R. Jansen, A. Van de Runstraat, R. Moors, K. Spee, S. Ploeger, R. Van de Kruijs, “Progress in EUV optics lifetime expectations”, Microelectron. Eng. 7374, 16-22 (2004). [3] T. E. Madey, N. S. Faradzhev, B. V. Yakshinskiy, and N. V. Edwards, "Surface phenomena related to mirror degradation in extreme ultraviolet (EUV) lithography", Appl. Surf. Sci. 253, 1691-1708 (2006). [4] R. Wasielewski, B. V. Yakshinskiy, M. N. Hedhili, A. Ciszewski, and T. E. Madey, “Surface chemistry of Ru: relevance to optics lifetime in EUVL”, Proc. SPIE 6533, 653316 (2007). [5] H. Over, Y. B. He, A. Farkas, G. Mellau, C. Korte, M. Knapp, M. Chandhok, M. Fang, “Long-term stability of Ru-based protection layers in extreme ultraviolet lithography: A surface science approach”, J. Vac. Sci. Technol. B25, 1123-1138 (2007). [6] B. V. Yakshinskiy, R. Wasielewski, E. Loginova, and T. E. Madey, “Carbon accumulation and mitigation processes, and secondary electron yields of ruthenium surfaces”, Proc. SPIE 6517, 65172Z (2007). [7] B. V. Yakshinskiy, R. Wasielewski, E. Loginova, M. N. Hedhili, T. E. Madey, “DIET processes on ruthenium surfaces related to extreme ultraviolet lithography (EUVL)”, Surf. Sci. (2008) in press. [8] M. N. Hedhili, B. V. Yakshinskiy, R. Wasielewski, A. Ciszewski, T. E. Madey, J. Chem. Phys. (2008) in press. [9] S. Bajt, N. V. Edwards, T. E. Madey, “Properties of ultrathin films appropriate for optics capping layers exposed to high energy photon irradiation”, Surf. Sci. Rpts. (2008) in press. [10] B. V. Yakshinskiy, T. E. Madey, and V. N. Ageev, “Thermal desorption of sodium atoms from thin SiO2 films”, Surf. Rev. Lett. 7, 75-87 (2000). [11] K. Nakayama and H. Hojo, in Proceedings of the Sixth International Vacuum Congress, Japan J. Appl. Phys. Suppl. 2, Pt. 1, 113 (1974). [12] M. L. Knotek and P. J. Feibelman, “Stability of ionically bonded surfaces in ionizing environments”, Surf. Sci. 90, 78-90 (1979). [13] R. L. Kurtz, R. Stockbauer, and T. E. Madey, “Angular distributions of ions desorbing from TiO2” Nucl. Instr. Meth. Phys. Res. B13, 518-524 (1986). [14] S. Tanaka, K. Mase, S. Nagaoka, and M. Kamada, “Electron–ion coincidence study for the TiO2(110) surface”, Surf. Sci. 451, 182-187 (2000). [15] L.-Q. Wang, D. R. Baer, M. H. Engelhard, “Creation of variable concentrations of defects on TiO2(110) using low-density electron beams”, Surf. Sci. 320, 295-306 (1994). [16] O. Dulub, M. Batzill, S. Solovev, E. Loginova, A. Alchagirov, T. E. Madey, U. Diebold, “Electron-induced oxygen desorption from the TiO2(011)-2×1 surface leads to self-organized vacancies”, Science 317, 1052-1056 (2007). [17] M. N. Hedhili, P. Cloutier, A. D. Bass, T. E. Madey, L. Sanche, “Electron stimulated desorption of anionic fragments from films of pure and electron-irradiated thiophene”, J. Chem. Phys. 125, 094704 (2006). [18] J. Hollenshead and L. Klebanoff, “Modeling radiation-induced carbon contamination of extreme ultraviolet optics”, J. Vac. Sci. Technol. B24, 64-82 (2006). [19] For a discussion of adsorption isotherms, see R. I. Masel, “Principles of Adsorption and Reaction on Solid Surfaces”, (Wiley, New York, 1996) Chapter 4. [20] T. E. Madey, B. V. Yakshinskiy, S. Zalkind, M. N. Hedhili, in preparation. [21] S. Tanuma, C. J. Powell, and D. R. Penn, “Calculations of electron inelastic mean free paths for 31 materials”, Surf. Interf. Anal. 11, 577-589 (1988). [22] P. Cloutier, L. Sanche, T. E. Madey, unpublished data. [23] M. Weis, M. Wedowski, B. M. Mertens, B. T. Wolschrijn, B. van Mierlo, N. B. Koster, J. van Elp, A.E. Duiserwinkel, and A. van de Runstrat, US Patent 2007/0054497 A1 (March 8, 2007). [24] K. Hashimoto, H. Irie, A. Fujishima, “TiO2 Photocatalysis: A Historical Overview and Future Prospects”, Japan. J. Appl. Phys. 44, 8269-8285 (2005).

Proc. of SPIE Vol. 6921 692114-10