Measurements of reactive ion etching process effect using ...

Measurements of reactive ion etching process effect using long-period fiber gratings

Mateusz Śmietana,1,* Marcin Koba,2,3 Predrag Mikulic,2 and Wojtek J. Bock2 1Institute of Microelectronics and Optoelectronics, Warsaw University of Technology, Koszykowa 75, 00-662

Warszawa, Poland 2Centre de recherche en photonique, Université du Québec en Outaouais, 101 rue Saint-Jean-Bosco, Gatineau, QC

J8X 3X7, Canada 3National Institute of Telecommunications,Szachowa 1, 04-894 Warszawa, Poland

*[email protected]

Abstract: The paper presents for the first time a study of long-period fiber gratings (LPFGs) applied for the measurements of reactive ion etching (RIE) process effect in various places of a plasma reactor. For the purposes of the experiment a number of highly sensitive LPFGs working at the dispersion turning point was fabricated using electric arc discharges. We show that the LPFGs allow for monitoring of the phenomena taking place in the reactor, especially those resulting in reduction of the LPFG diameter. Results of the measurements supported by simulations have shown that etching rate significantly decreases with elevation of the sample up to 3.6 mm over the electrode in the reactor, and stays constant above this height.

©2014 Optical Society of America

OCIS codes: (280.4788) Optical sensing and sensors; (060.2370) Fiber optics sensors; (050.2770) Gratings; (310.3840) Materials and process characterization; (240.6700) Surfaces.

References and links 1. L. Martinu and D. Poitras, “Plasma deposition of optical films and coatings: A review,” J. Vac. Sci. Technol. A

18(6), 2619–2645 (2000). 2. H. Anma, J. Toki, T. Ikeda, and Y. Hatanaka, “Uniform deposition of SiC thin films on plastics surfaces,”

Vacuum 59(2-3), 665–671 (2000). 3. M. Smietana, W. J. Bock, and P. Mikulic, “Temperature sensitivity of silicon nitride nanocoated long-period

gratings working in various surrounding media,” Meas. Sci. Technol. 22(11), 115203 (2011). 4. M. Śmietana, M. Dudek, M. Koba, and B. Michalak, “Influence of diamond-like carbon overlay properties on

refractive index sensitivity of nano-coated optical fibers,” Phys. Status Solidi A 210(10), 2100–2105 (2013). 5. Y. Tzeng and T. H. Lin, “Dry etching of silicon materials in SF6 based plasmas. Roles of N2O and O2 gas

additives,” J. Electrochem. Soc. 134, 2304–2309 (1987). 6. D. Jose, M. S. John, P. Radhakrishnan, V. P. N. Nampoori, and C. P. G. Vallabhan, “An optical fibre based

evanescent wave sensor to monitor the deposition rate of thin films,” Thin Solid Films 325(1–2), 264–267 (1998).

7. L.-Y. Shao, J. P. Coyle, S. T. Barry, and J. Albert, “Anomalous permittivity and plasmon resonances of copper nanoparticle conformal coatings on optical fibers,” Opt. Mater. Express 1(2), 128–137 (2011).

8. A. M. Vengsarkar, P. J. Lemaire, J. B. Judkins, V. Bhatia, T. Erdogan, and J. E. Sipe, “Long-period fiber gratings as band-rejection filters,” J. Lightwave Technol. 14(1), 58–65 (1996).

9. X. Shu, L. Zhang, and I. Bennion, “Sensitivity characteristics of long-period fiber gratings,” J. Lightwave Technol. 20(2), 255–266 (2002).

10. J. Kanka, “Design of turn-around-point long-period gratings in a photonic crystal fiber for refractometry of gases,” Sens. Actuators B Chem. 182, 16–24 (2013).

11. X. Chen, K. Zhou, L. Zhang, and I. Bennion, “Dual-peak long-period fiber gratings with enhanced refractive index sensitivity by finely tailored mode dispersion that uses the light cladding etching technique,” Appl. Opt. 46(4), 451–455 (2007).

12. P. Pilla, P. F. Manzillo, V. Malachovska, A. Buosciolo, S. Campopiano, A. Cutolo, L. Ambrosio, M. Giordano, and A. Cusano, “Long period grating working in transition mode as promising technological platform for label-free biosensing,” Opt. Express 17(22), 20039–20050 (2009).

13. M. Smietana, W. J. Bock, P. Mikulic, and J. Chen, “Tuned pressure sensitivity of dual resonant long-period gratings written in boron co-doped optical fiber,” J. Lightwave Technol. 30(8), 1080–1084 (2012).

#204098 - $15.00 USD Received 3 Jan 2014; revised 11 Feb 2014; accepted 13 Feb 2014; published 6 Mar 2014(C) 2014 OSA 10 March 2014 | Vol. 22, No. 5 | DOI:10.1364/OE.22.005986 | OPTICS EXPRESS 5986

14. M. Smietana, W. J. Bock, P. Mikulic, A. Ng, R. Chinnappan, and M. Zourob, “Detection of bacteria using bacteriophages as recognition elements immobilized on long-period fiber gratings,” Opt. Express 19(9), 7971–7978 (2011).

15. M. Śmietana, M. Koba, S. Tripathi, P. Mikulic, and W. J. Bock, “Improving sensing properties of the long-period gratings by reactive ion etching,” in Proceedings of Seventh International Conference on Sensing Technology (IEEE, 2013), pp. 669–672.

16. K. S. Chiang, Y. Liu, M. N. Ng, and X. Dong, “Analysis of etched long-period fibre grating and its response to external refractive index,” Electron. Lett. 36(11), 966–967 (2000).

17. M. Smietana, W. J. Bock, and P. Mikulic, “Comparative study of long-period gratings written in a boron co-doped fiber by an electric arc and UV irradiation,” Meas. Sci. Technol. 21(2), 025309 (2010).

18. M. Smietana, W. J. Bock, and P. Mikulic, “Effect of high-temperature plasma-deposited nano-overlays on the properties of long-period gratings written with UV and electric arc in non-hydrogenated fibers,” Meas. Sci. Technol. 24(9), 094016 (2013).

19. T. Enomoto, M. Denda, A. Yasuoka, and H. Nakata, “Loading effect and temperature dependence of etch rate in CF4 plasma,” Jpn. J. Appl. Phys. 18(1), 155–163 (1979).

20. M. Daimon and A. Masumura, “Measurement of the refractive index of distilled water from the near-infrared region to the ultraviolet region,” Appl. Opt. 46(18), 3811–3820 (2007).

21. M. Smietana, W. J. Bock, P. Mikulic, and J. Chen, “Increasing sensitivity of arc-induced long-period gratings – pushing the fabrication technique towards its limits,” Meas. Sci. Technol. 22(1), 015201 (2011).

22. J. W. Coburn and H. F. Winters, “Plasma etching: a discussion of mechanisms,” J. Vac. Sci. Technol. 16(2), 391–403 (1979).

23. M. J. Goeckner, J. Goree, and T. E. Sheridan, “Measurements of ion velocity and density in the plasma sheath,” Phys. Fluids B 4(6), 1663–1670 (1992).

24. B. Jacobs, W. Gekelman, P. Pribyl, and M. Barnes, “Temporally resolved ion velocity distribution measurements in a radio-frequency plasma sheath,” Phys. Plasmas 18(5), 053503 (2011).

25. J. Hopwood, C. R. Guarnieri, S. J. Whitehair, and J. J. Cuomo, “Langmuir probe measurements of a radio frequency induction plasma,” J. Vac. Sci. Technol. A 11(1), 152–156 (1993).

1. Introduction

Plasma has been extensively used for deposition and processing of a number of materials [1]. Depending on the process parameters, especially on gases composition in plasma reactor, a deposition of a film, a material surface etching or a functionalization can be obtained. The plasma-based processes are widely applied for fabrication of advanced electronic devices. The processes are currently well controlled when standard surfaces, namely flat silicon wafers, are processed. However, for many applications anisotropic processing of complex shaped and structured materials is required [2–4]. In order to let plasma interact with all-around surface of the sample, it is required to hold it above the stage in the reactor. In such a case processing of the surface may take place differently than when the sample is placed directly on the stage. The effects of these processes are difficult to investigate since typically only a couple of nanometers are processed, and the measurements performed using, e.g., scanning electron microscope, often suffer from lack of precision. In case of microscopic analysis a whole fiber cross-section must be observed and its size is about three orders of magnitude bigger than the etched thickness.

The plasma etching process is based on an ion bombardment and chemical reactions on material’s surface. Reactive species generated in plasma, namely radicals and ions, are responsible for chemical reactions and surface bombardment, respectively. When chemical reactions occur on the material’s surface a volatile byproducts are formed and exhausted from the chamber. For low (10−3 – 10−4 Torr) and high (~1 Torr) pressure in plasma reactor physical (sputtering) or chemical etching dominates. Physical etching favors high anisotropy and low selectivity, where chemical etching results in higher selectivity and lower anisotropy. When the pressure is in range between 10 and 100 mTorr, both physical and chemical interactions occur and the process is often called as Reactive Ion Etching (RIE). For silicon-based materials RIE typically involves fluorine chemistries, i.e., CF4, CHF3 and SF6 [5]. Additional gases modify dynamics of the process, e.g., addition of oxygen increases fluorine concentration by combining with carbon or sulfur, which would otherwise bond to fluorine.

There are optical-fiber-based devices which are sensitive to processing of their surface. There has been presented that optical fiber sensors are able to monitor deposition rate of thin


films [6, 7]. In this work we employed long-period fiber gratings (LPFGs) for investigation of etching effects in different places in plasma reactor. The LPFGs have been known for over a decade and are a periodic modulation of the refractive index along the length of an optical fiber [8]. Under special phase-matching conditions, the grating couples the fundamental core mode (LP01) to m discrete cladding modes (LP0m) that are attenuated due to absorption and scattering. The coupling is wavelength-dependent, so one can obtain a spectrally selective loss for each of the cladding modes. There are many influences that can shift the resonance wavelengths m

resλ of a LPFG. According to (1), which describes a λ-dependent coupling from

the guided core mode to the mth cladding mode where 01effn is the effective refractive index of

the propagating core mode, 0meffn is the effective refractive index of the mth cladding mode and

Λ is the period of the LPFG, the shift can be induced by a variation of either the Λ of the grating or the effn of the modes.

( )01 0m mres eff effn nλ = − Λ (1)

Up to a certain point, an increase in LPFG sensitivity follows an increase in the order of the coupled cladding mode [9]. Shu et al. found that when the grating period of the LPFG is short (typically below 200 μm for standard fibers), it is possible to couple energy into the same cladding mode at two discrete wavelengths resulting in the appearance of dual resonant peaks of higher-order cladding modes [9]. The dual resonant wavelengths shift in opposite directions with the variation of a number of parameters. At this point, often called dispersion turning point (DTP) [9] or turn around point (TAP) [10], the gratings are extremely sensitive not just to external perturbations, including external refractive index ( extn ), but also to fiber

properties [11]. Several applications of LPFGs operating at the DTP for temperature, strain, pressure and refractive index sensing have been shown, including a number of chemosensors and biosensors [9, 12–14]. Due to some lack of repeatability in the fiber properties and LPFG fabrication processes, various methods have been applied for precise tuning of the grating leading up to the DTP. The most commonly a method based on controlled etching of the fiber cladding is applied [11, 15]. Reduction in fiber cladding decreases 0m

effn and induces shift of

the resonance wavelength towards the DTP. Analytical expression for shift of the mresλ

induced by variation in fiber cladding radius ρ is given in (2), where α is a core radius, u∞

is the thm root of the Bessel function 0J , ρ is a new value of cladding radius, cln is a

cladding refractive index, and ( )2 2 2/ 2co cl con n nΔ = − , where con is a core refractive index [16].

( )22

2 2 2 2 20 1

2 2 2 20 1 0 1

2 2

1 1

8 ( ) ( )

m

m clres

c

e

l

r s

n

u a u a u a u aJ J J

u n

J u J u

aJ

a

λδλ

π ρ ρ

ρ ρ ρ ρ ρ ρ

∞

∞ ∞

∞ ∞ ∞ ∞ + +

Λ ΔΛ

− − +

× −

(2)

In this work we focused on application of the LPFGs for measuring plasma-based etching rate for different sample placement in the reactor. Experimental results are correlated with results of simulations in order to estimate the etching rate. According to our best knowledge, the application of LPFGs for investigation of plasma-based etching process effectiveness and its distribution in plasma reactor is shown here for a first time.

2. Experimental details

A set of the LPFGs was fabricated using computer-assisted electric-arc-based setup described elsewhere [17]. A 10 cm-long piece of Fibrecore PS1250/1500 photosensitive fiber has been spliced between two input and output Corning SMF28 fibers. The LPFGs were written in PS


fiber only with period of Λ = 190-193 µm and typical number of periods required for appearing of a grating effect was 200 to 270. The optical transmission of the fiber was monitored during the LPFG fabrication process in order to obtain the desired spectral attenuation notch. The arc-induced gratings have been applied here due to their high resistance to temperature that may be significantly increased during the plasma-based process [18].

The etching process has been performed using Oxford PlasmaPro NGP80 system. We used etching scheme in SF6 and O2 plasma (O2 flow 100 sccm, SF6 flow 30 sccm, pressure 100 mTorr, RF power in range from 50 to 250 W). Temperature of the stage electrode was set to 20 °C and each etching was 5 minutes long. During the etching processes the LPFG samples and reference oxidized Si wafers were held from 1.2 to 12 mm above the surface and on the surface of the electrode, respectively (Fig. 1). Except the LPFG position, conditions inside the chamber during etching were constant for each of the processes. Since in case of plasma-based etching so called “loading effect” takes place, where etching rate depends on area of material to be etched [19], the amount of material each time in the chamber was constant.

Fig. 1. Schematic representation of the LPFG and silicon wafer placement in the process chamber. The LPFG’s position h has been changed in range 1.2 to 12 mm by placing the fiber between the glass plates of 1.2 mm in thickness. The elements are not in scale.

The etching rate on the electrode has been determined using the reference oxidized Si wafers. The thickness of the oxide (d), assumed to be a good reference for surface of the fused silica fibers, has been measured using Horiba Jobin-Yvon UVISEL spectroscopic ellipsometer according to a two-layers SiO2/Si model. The results obtained with etched samples were referred to d of the oxide before the etching (d = 115 nm). Since optical properties of the oxide and fused silica are very similar (nD = 1.458), we assumed similar etching rate of these materials.

The optical transmission of the LPFGs in the range of λ = 1100-1700 nm was monitored using a NKT Photonics SuperK COMPACT supercontinuum white light laser source and


Yokogawa AQ6370C optical spectrum analyzer. The ambient temperature (T) during the measurements was 22 °C and the tension of the LPFGs was held constant throughout the experiment. The etching effect on the LPFG has been investigated for the samples in air ( extn = 1) and for those immersed in deionised water ( extn = 1.318 at λ = 1550 nm) [20].

The optical simulations of the LPFGs were performed for the grating using Optigrating v4.2 software by Optiwave. We assumed fiber properties reported elsewhere [17]. Due to limited accuracy of fiber drawing process, especially in terms of its diameter, which is given by the manufacturer with accuracy of +/− 1 µm, the properties of the fiber were slightly tuned in order to obtain a good match of the simulated and measured spectra.

3. Results and discussion

For the needs of this experiment first we investigated the influence of RF generator power on etching rate of oxidized Si wafers placed directly on the electrode. The ellipsometric analysis have shown that etching rate on the electrode can be tuned from 6.5 to 103 nm/min by adjusting RF generator power from 50 to 250 W, respectively. In order to properly control the etching process on the fiber all the following experiments were performed with P = 100 W, for which etching rate on the reference Si sample reached 22 nm/min.

The measurement of the LPFG response to immersing the sample in water is shown and compared to the results of simulations in Fig. 2. The observed resonances are result of coupling core mode and 11th (LP011) cladding mode [21]. The unique feature of LPFGs that sets them apart from fiber Bragg gratings is their high sensitivity to changes in extn [8, 9]. An

increase of the spectral distance between the resonances is observed when extn is increased.

The effect is induced by the increase of 011effn with extn [21]. As a result of matching

experimental data and simulated spectra of the LPFGs, the initial fiber radius ( clr ) was set to

63 µm, where cladding thickness was 58.9 µm. The results of simulation and measurements for these properties of the grating agree well when grating is in air or is surrounded by water. In case of simulations for LPFG immersed in water, it is possible to determine position of the resonance appearing in λ range above 1700 µm, where taking measurements was impossible with the applied equipment.

Fig. 2. Initial spectra of the investigated LPFG surrounded by air (n = 1) and water (n = 1.318 at λ = 1550 nm). Results of simulation are given for comparison.

The spectra obtained after each etching process for the LPFG in air and immersed in water are shown in Figs. 3(a) and 3(b), respectively. With the increase in extn and decrease in clr , the


peaks increase and decrease their spectral distance, respectively. Each subsequent etching process brings the resonances closer to DTP, which is reached after 4th and 10th etching process for LPFG in air and water, respectively. Fiber and grating properties allow for obtaining DTP at about λ = 1560 nm. When DTP is reached, the distinguishing of the resonances is no longer possible, but just before its occurrence the sensitivity to both

extn and clr is highest. That is why the spectral distance between the resonances should be

minimized, but on the other hand the peaks must have enough spectral separation to precisely define the resonance wavelengths. The increase in extn is followed by variation in 011

effn and

results in tuning the LPFG away from the DTP. Thanks to higher separation between the resonances for the sample surrounded by higher extn it is possible to trace the resonance

wavelength position almost for all etching processes, but when water is applied the sensitivity to the reduction in fiber diameter is lower. In case of both media surrounding the fiber, the highest shift can be observed as a result of 1st to 3rd etching taking place up to 3.6 mm above the electrode and further when the LPFG works close to the DTP. The difference between resonance wavelength shift induced by subsequent etching can be well observed in the whole process range for the LPFG immersed in water (Fig. 3(b)). The irregularities observed in the spectra about λ = 1550 nm are induced by measurement setup.

a)

b) Fig. 3. Evolution of the LPFG spectrum induced by ten subsequent etching processes at different sample height (1.2 to 12 mm) when the LPFG was surrounded by (a) air and (b) water.


According to the results of simulations shown in Fig. 4(b), away from DTP the dependence between the resonance wavelengths and the changes in fiber radius is almost linear and reaches 0.22 and −0.29 nm/nm for resonance at lower and higher wavelength, respectively. It can be also seen that just before reaching the DTP, the sensitivity of the grating to the reduction in radius of the fiber is the highest and becomes nonlinear. When results of simulations are compared with experimental data shown in Fig. 4(a), the linear shift of the resonance wavelength with the number of processes for up to 3rd etching process taking place 3.6 mm over the electrode is significantly disturbed. In order to estimate the etching rate at each height taking into account the nonlinear character of the dependence between 11

resλ and clrΔ , we correlated the results of measurements and simulations. The effect

of matching the results and calculation of the etching rate for each of the processes is shown in Fig. 5. The results for both measurements in water and air agree well and show higher etching rate for processes taking place up to 3.6 mm over the electrode than when the sample was placed higher than 3.6 mm. When the results of etching rate obtained for oxidized Si wafer placed on the electrode are concerned, a decrease of etching rate with distance from the electrode reaching 4.8 nm/min/mm can be seen. For the samples placed higher than 3.6 mm over the electrode, the influence of height is insignificant and the etching rate stays at the level of 5 nm/min.

a)

b) Fig. 4. Resonance wavelength shifts at various stages of the etching experiment where (a) shows results for each etching step and (b) shows results of simulation when cladding radius is reduced.


Fig. 5. Etching rate at various height in plasma reactor calculated by matching results of λres for experimental results and simulations. Matching has been performed for both results obtained for the LPFGs surrounded with air and water. Etching rate obtained for oxidized Si wafer placed on the electrode and measured with spectroscopic ellipsometry is given for comparison.

In order to analyze the obtained results, a variety of physical and chemical processes taking place in the plasma reactor must be discussed. The main steps in the etching process are: formation of the reactive particles (e.g., F-), arrival of the particles to the etched surface, their adsorption followed by chemisorptions, formation of the product molecules (e.g., SiF up to SiF4), and finally their desorption (SiF4) and removal from the reactor. The process steps are highly influenced by ion bombardment. The measured effect of decrease in etching rate with distance from the electrode is related to the formation of the ion sheath at the surface of the electrode [22]. Since the electrode is isolated with a capacitor, the circuit is shortened and open for RF and DC field, respectively. Free electrons generated in plasma are able to follow the variations of the RF field and they can travel and gain high energy, where in turn the heavier ions are significantly less influenced by the field. The electrons moving with higher speed than ions collide with the electrodes and are removed from the plasma in first few generator cycles. As a result of this effect, the electrode becomes negatively charged with respect to plasma. The electrode repeals electrons towards plasma, while ions arriving at the interface between plasma and the sheath are attracted by the electrode and hit its surface. This effect has been investigated in this experiment using the LPFGs. The ions arriving to the self-biased electrode have higher kinetic energy than those hitting the fiber placed above it, where they are not yet highly accelerated by the electric field. The ion bombardment enhances adsorption of the reactive particles at the surface, their chemisorptions followed by formation of the product molecules, and their desorption. The impinging ions break some bonds between the atoms of the surface material, and thus form active sites, which easily react with fluorine. The ions also enhance chemisorption by delivering energy to form a covalent bond of the fluorine to the atom of the surface material. The formed molecules need to be removed from the surface, and the process is also enhanced by the incoming ions. For the applied process parameters at the height of above 3.6 mm over the electrode the fiber is immersed in plasma, where the ions energy is low. This fact significantly slows down the etching process. The shape of the dependence shown in Fig. 5 and obtained in our experiment, agrees well with the


results reported by other authors, when other optical [23, 24] and electrical [25] methods were applied for the measurements of ions velocity in the sheath.

4. Conclusions

This study shows for the first time an application of an LPFG for the measurements of RIE process kinetics. Thanks to high sensitivity of the LPFG working at a dispersion turning point to diameter of the fiber, it is possible to monitor a reduction in fiber cladding by tracing a resonance wavelength of the LPFG. We have found that for the investigated LPFGs, an etching of 1 nm in the fiber radius induces at least a 0.22 nm of the shift in resonance wavelength. The sensitivity of the LPFGs to reduction in the fiber diameter can be tuned by refractive index of the medium surrounding the LPFG during the experiments. Moreover, we’ve found that the effectiveness of the etching process highly depends on a distance between the sample and the stage electrode. The etching rate strongly decreases with the distance dropping to about 25% of the value at the electrode when the sample is placed only 3.6 mm over the electrode. The effect has been attributed to the presence of ion sheath at the electrode surface where ions gain their energies and enhance the etching process.

The presented method for investigations of effects of plasma process can be also applied when monitoring of other processing, i.e., deposition of a thin film, is required. In comparison to other methods of plasma diagnostics, which include application of Langmuir probes or measurements of laser-induced fluorescence, the application of the LPFGs allows for determination of the effects of the process, not only the ions velocity. When the LPFG is coated with thin high-refractive-index material, the method allows for monitoring of the thin film processing, which is not restricted only to plasma-based methods.

Acknowledgments

The authors gratefully acknowledge support for this work from the Ministry of Science and Higher Education of Poland within IUVENTUS Plus Program, the National Centre for Research and Development of Poland within the LIDER program, the Natural Sciences and Engineering Research Council of Canada, and the Canada Research Chairs Program.


Measurements of reactive ion etching process effect using ...

Documents

Transcript of Measurements of reactive ion etching process effect using ...