- Email: [email protected]
Tailoring properties of lossy-mode resonance optical fiber sensors with atomic layer deposition technique
Optics and Laser Technology 102 (2018) 213–221
Contents lists available at ScienceDirect
Optics and Laser Technology journal homepage: www.elsevier.com/locate/optlastec
Full length article
Tailoring properties of lossy-mode resonance optical fiber sensors with atomic layer deposition technique q Kamil Kosiel a,⇑, Marcin Koba b,c, Marcin Masiewicz b, Mateusz S´mietana b a
Institute of Electron Technology, Al. Lotników 32/46, 02-668 Warsaw, Poland Institute of Microelectronics and Optoelectronics, Koszykowa 75, 00-662 Warsaw, Poland c National Institute of Telecommunications, Szachowa 1, 04-894 Warsaw, Poland b
a r t i c l e
i n f o
Article history: Received 14 October 2017 Received in revised form 29 December 2017 Accepted 2 January 2018
Keywords: Optical fiber sensors Lossy-mode resonance Thin films Atomic layer deposition Optical properties
a b s t r a c t The paper shows application of atomic layer deposition (ALD) technique as a tool for tailoring sensorial properties of lossy-mode-resonance (LMR)-based optical fiber sensors. Hafnium dioxide (HfO2), zirconium dioxide (ZrO2), and tantalum oxide (TaxOy), as high-refractive-index dielectrics that are particularly convenient for LMR-sensor fabrication, were deposited by low-temperature (100 °C) ALD ensuring safe conditions for thermally vulnerable fibers. Applicability of HfO2 and ZrO2 overlays, deposited with ALD-related atomic level thickness accuracy for fabrication of LMR-sensors with controlled sensorial properties was presented. Additionally, for the first time according to our best knowledge, the doublelayer overlay composed of two different materials - silicon nitride (SixNy) and TaxOy - is presented for the LMR fiber sensors. The thin films of such overlay were deposited by two different techniques – PECVD (the SixNy) and ALD (the TaxOy). Such approach ensures fast overlay fabrication and at the same time facility for resonant wavelength tuning, yielding devices with satisfactory sensorial properties. Ó 2018 Elsevier Ltd. All rights reserved.
1. Introduction Nowadays optical sensors already allow for measuring majority of physical and chemical measurands of interest [1]. These are known various temperature [2], [3], pressure [4,5], strain [3], displacement [6], electromagnetic field [7–9] and refractive index [10,11] sensors, as well as chemical [12] and biosensors targeted towards certain molecule or even whole biological cell [13–15] to name only a few. While the modern optical sensors technologies cover a wide range of device architectures, physical effects and sensing strategies, the sensors often require deposition of thin films (overlays) or surface structures. The properties of these films or structures can be various and in particular they can be metallic or dielectric. The mentioned overlays or structures are used to initiate or modify sensorial response of the sensors, allowing for precise adjustment of the sensors properties according to the needs. Therefore in most cases the thin overlay thickness and its optical properties (primarily the complex refractive index - n⁄) should be strictly controlled. Contemporary thin film fabrication techniques are expected to provide high deposition accuracy and process-to-process reproducibility, as well as coating uniformity on various substrates and q
Invited article by the Guest Editor: Ignacio Del Vilar.
⇑ Corresponding author.
E-mail address: [email protected] (K. Kosiel). https://doi.org/10.1016/j.optlastec.2018.01.002 0030-3992/Ó 2018 Elsevier Ltd. All rights reserved.
shapes. These circumstances impose stringent requirements on control of the overlay deposition rate and the overlay properties on different stages of deposition. The matter of technological requirements becomes even more critical when three-dimensional objects like optical-fiber-based sensors are considered. All these make the matter of choice of the thin film deposition technique crucial. Many nanofilm deposition techniques have already been used in the field of optical sensing, such as Langmuir-Blodgett method [16], spin coating [17], sol-gel process [18], layer-by-layer technique [19], chemical vapor deposition (CVD) [20], or physical vapor deposition (PVD) methods, which include evaporation [21] or sputtering [22]. They all have specific drawbacks, such as poor deposition process control, low coating robustness or lack of accuracy and coating uniformity. Atomic layer deposition (ALD), which is a specific variation of the CVD method, seems to be the first-choice technique enabling control of the sensors properties by means of precise adjustment of the overlay features. The ALD chemical precursors are delivered to the reaction zone separately in time, and hence the complementary and sequentially repeated chemisorption half-cycle reactions responsible for the layer growth undergo at a surface in a selflimiting manner [23]. Thanks to this very special paradigm an atomic control of the layer thickness is possible, with uniquely conformal, uniform and tight layers, even when they are deposited
214
K. Kosiel et al. / Optics and Laser Technology 102 (2018) 213–221
on complicated high-aspect-ratio surfaces. These and other advantages of ALD, such as capability for deposition of dense layers in a wide range of materials (oxides [24,25], nitrides [26–28], metals [29,30] and many others isolators, semiconductors and conductors), possible deposition at relatively low temperatures (even a way below 100 °C [31,25]) lead to a situation where the ALD technique is a must in nowadays microelectronics [32,33]. At the same time, it reaches more and more applications in other very divers branches [34,35], and has already begun to be used in optical sensors technology too [36–42]. The phenomenon of lossy-mode-resonance (LMR), being the physical basis of operation of a specific class of sensors, results from the fact that for certain thickness and optical properties of the dielectric overlay the light in the thin-film-coated fiber core experiences wavelength-dependent attenuation, i.e., in the optical spectrum characteristic transmission minima appear [43]. The effect is a consequence of coupling between core modes and specific lossy modes of the overlay. The condition of appearance of LMR is that the real part of the overlay permittivity is positive and larger in magnitude than both its own imaginary part and the real part of permittivity of the materials surrounding the overlay, i.e., of the fiber core and an external medium. The crucial issue for sensing applications is that spectral positions of the minima depend also on refractive index of the surrounding medium (external RI). This fact enables the use of LMR structures as refractometers. On top of it, the sensor surface functionalization allows also for biosensing applications of these sensors [44]. It must be noted that the spectral position of the resonant wavelengths and the device sensitivity can be fine-tuned just by adjustment of the thickness and optical properties of the overlay. The most often used for the purpose of overlay for LMR-based fiber sensors are oxides of aluminium (Al2O3) [45], titanium (TiO2) [46,47], tin (SnO2) [48], indium (In2O3) [49], or indium doped tin oxide (ITO) [50–55]. Also silicon nitride (SixNy) [56], diamond-like carbon [57] and polymeric overlays, like polyallylamine hydrochloride/polyacrylic acid (PAH/PAA) [58] have already been used. However, the overlays on the fiber sensors have been deposited in the form of single layers only, i.e., just one material at a time has been used for formation of an individual overlay. The possible reason for this can be a complicated numerical modelling of such a sensor, but also a number of technological difficulties. According to our best knowledge only once multilayer LMRbased structure was discussed theoretically, but the concept has never been practically implemented [59]. In this paper we present capability for tuning the properties of optical fiber LMR sensors by using the ALD-based dielectric overlays. We show the possibility of tuning its sensing properties by ALD-deposited single-layer overlays, and also we put into practice the concept of double-layer dielectric overlays for the first time. Such overlays comprise two layers, each deposited by using different technique, namely ALD and plasma-enhanced CVD (PECVD). The presented solution offers to the field of thin film optical fiber sensors technological flexibility available thanks to application of the ALD technique.
2. Experimental details 2.1. Numerical modelling The numerical model used for the simulations of the LMR phenomenon has already been described elsewhere [57]. The aim of the performed simulations is to show the effect of variation in the LMR response when properties of the overlays change and to determine the range of the properties where the effect can be seen. Wavelength-dependent response of the structures to variations in
external RI was normalized using spectrum obtained for a bare sample surrounded by air. 2.2. The fibers and their preparation before deposition of the overlays For preparation of the sensors polymer-clad silica (PCS) multimode optical fibers with 400 lm core diameter were used. While total length of the fiber samples was up to 150 mm the only 25 mm-long central section of each fiber was stripped mechanically and chemically from outer polymer coating and cladding [57]. Then the fiber samples were extensively cleaned with isopropanol. 2.3. Deposition of the overlays by ALD The metal oxides were deposited by ALD using thermal mode at the thermocouple-controlled reactor temperature of 100 ± 0.1 °C, in Beneq TFS-200–190 system. Such low temperature of the process was unavoidable because of the thermal vulnerability of a polymer coating that majority of optical fiber length is originally covered by. Additionally, low temperature enables deposition of amorphous films and hence smoother surfaces than for also available for ALD polycrystalline forms. Amorphous films are preferred due to more uniform film thickness distribution as well as for suppression of light scattering. The carrier/purging gas was 6N-purity argon (Ar) flowing through the reactor with constant rate of 1000 sccm that generated the pressure inside the reactor of approx. 2 mbar. As the oxygen chemical precursor was used deionized water (DI H2O) kept at 19 ± 0.1 °C in Beneq’s standard LS container. Tantalum pentachloride (TaCl5, kept at 85 ± 0.1 °C), tetrakis(ethylme thylamino)zirconium (TEMAZ, kept at 75 ± 0.1 °C), and tetrakis(et hylmethylamino)hafnium (TEMAH, kept at 78 ± 0.1 °C), all delivered by Volatec Oy and kept in Beneq’s standard HS 300 containers, were used as chemical precursors for tantalum, zirconium and hafnium, respectively. The precursor exposure times, that are responsible for substrate-surface-saturation conditions for all complementary half reactions of chemisorption involved in the employed ALD processes, were experimentally determined before the experiments described in this paper; the crucial role of exposure of complementary precursors and surface saturation in ALD reactions has already been explained, e.g., in Ref. [60,61]. They were studied using Si test substrates, and they are not described here. On the basis of these experiments the appropriate precursor/purge-cycling schemes were determined. For the optimized cycling conditions used in this work the growth per cycle (GPC) measured for Si-based samples was approx. 0.133, 0.150 and 0.145 nm/cycle for the TaxOy, ZrO2 and HfO2 films, respectively. The film growth rate was approx. 0.019, 0.022, and 0.021 nm/s for the TaxOy, ZrO2 and HfO2, respectively. It should be noticed here, that generally the composition of as-deposited TaxOy may vary and contain compounds of tantalum with various oxidation states between +2 and +5, i.e., TaO, Ta2O3, TaO2 and Ta2O5. The thickness range of the ALD-based layers investigated in this paper was 201–211 nm for the HfO2, 216–244 nm for ZrO2, and 50 nm for the TaxOy. Before entering the ALD reactor the fibers were rinsed extensively with acetone and isopropanol at the ambient temperature (approx. 20 °C). For each ALD process three fibers were introduced into the reactor and they were positioned in a run-to-runrepeatable arrangement. 2.4. Deposition of the overlays by PECVD The SixNy thin films were deposited on the fiber samples using Oxford Plasmalab 80 Plus system. The fiber samples were placed in the plasma reactor on U-type holder on height of approx. 5 mm over the electrode. The reference Si wafers were placed next to
K. Kosiel et al. / Optics and Laser Technology 102 (2018) 213–221
215
them on the electrode. The radio-frequency (RF) PECVD process was preceded by 1 min-long plasma cleaning in Ar. The deposition process was performed with SiH4 flow of 500 sccm, NH3 flow of 50 sccm, pressure of 1.2 Torr, power of 50 W. The electrode temperature was kept at 200 °C remaining a controlled parameter and deposition on Si wafers took place in such thermal conditions. The temperature of the fiber samples was not measured directly and taking into account their placement it was slightly lower than 200 °C. The resulting films with thicknesses in the range of 410– 510 nm were deposited with growth rate of approx. 1.2 nm/s. In general, PECVD-deposited SixNy films can have different compositions of y/x in the range of approx. 0.5–1.5 [62]. 2.5. The ex situ control of thicknesses and optical properties of the overlays In order to allow for ex situ control of thicknesses and n⁄ values of the overlays deposited on the sensors, also test layers were deposited within the ALD and PECVD processes on reference Si wafers. Only these were analyzed by spectroscopic ellipsometry (SE), since such analyses could not be performed directly on the fiber samples. The ALD-related reference Si wafers were positioned in a run-to-run-repeatable manner inside the reactor. As it was revealed above the PECVD-related reference Si wafers were placed next to fibers on the electrode. The SE measurements were performed using Horiba–Jobin–Yvon UVISEL spectroscopic ellipsometer in the wavelength range 350–900 nm. The properties of the films were calculated by fitting SE data to single layer Tauc–Lorenc dispersion model [57,63]. 2.6. Measurements of the optical fiber structures After the ALD and PECVD deposition processes both ends of each fiber were polished in order to remove the overlays covering their ends and to minimize the transmitted optical signal losses. The transmission spectral response of the optical fiber samples was measured in wavelength range 350–1050 nm using Ocean Optics HL-200-HP white light source and USB4000 spectrometer. The applied measurement setup was described in details earlier in [57]. For the spectral response measurements performed in various external RIs the coated segment of the fiber was immersed in a series of aqueous solutions of glycerin with various concentrations. The respective RIs of the obtained liquids were measured using Reichert AR200 Automatic Digital Refractometer with accuracy ±10 4 RI units (RIU). Wavelength-dependent response of the structures to variations in external RI was normalized using spectrum obtained for a bare sample surrounded by air.
Fig. 1. Dispersion of n and k for the ALD HfO2, ZrO2, TaxOy and the PECVD SixNy films.
resonant wavelengths for the LMR sensors. The examples of this will be presented in further subsections. Here, an assumption was made that the properties of the films deposited on test substrates could satisfactorily correspond with the properties of the overlays deposited on the fibers. It should be noted here that properties of ALD process, e.g., the GPC, can vary depending on the applied substrate [60]. Mainly surface density of the functional groups that are active in chemisorption reactions is determined by the substrate. However, we confirmed that the thickness and n⁄ of oxides deposited within the same process on Si wafers and glass plates were comparable. The glass plate surfaces can be assumed to be similar to the surfaces of silica fibers used for the sensors fabrication. These observations would justify the use of Si wafers as reference substrate for ellipsometric analysis. On the other hand routine usage of the glass-based reference samples, would be much more time-consuming, as they would need complex preparation aiming for reduction of multiple reflections in a glass plate. At the same time one should also be aware that the properties of PECVD SixNy on the fibers could very likely be different (in terms of thickness and n⁄) than these measured on the Si-based references. The discrepancies steam from the mentioned earlier different placement of the Si wafers and fibers in the reactor. Moreover the values of n and k can be influenced by strain and temperature, which can be different for the planar layers and cylindrical overlays on the fibers. However, the obtained SE results were critically compared with the results of transmission measurements performed for the fibers and with their simulations in order to draw final conclusions on the overlay properties.
3. Results and discussion 3.2. The fiber structures coated by ALD HfO2 or ZrO2 3.1. Properties of the films The dispersion relations of refractive index (n) and extinction coefficient (k) measured for the films deposited on the reference Si wafers are presented in Fig. 1. As it is seen within the tested group of ALD-based materials we are able to tune the n value in the range of approx. 2.05–2.1 RIU given here for the wavelength equal to 650 nm, which we can further expand using the PECVD SixNy. The k values were low for all materials in the tested wavelength range. Given additionally the wide range of film thickness that is accessible for these deposition techniques (for ALD typically up to 250 nm and for PECVD even by an order of magnitude larger) it should be noticed that this provides very wide range of possible tuning of sensor properties, particularly in terms of accessible
Assuming optical properties of the films shown in Fig. 1, the numerical simulations predict presence of a set of LMR-effects occurring in available range of transmission spectrum for above certain overlay thickness. Fig. 2a shows a collection of spectra simulated in a wide range of wavelength for the ZrO2-coated structures of various thicknesses and for a fixed value of external RI. It is seen, that each spectrum contains two prominent resonant minima – which we further refer to as the 1st order minimum (for shorter wavelength, between approx. 350–400 nm) and the 2nd order minimum (for longer wavelength, over 600 nm). For both minima a redshift is seen with increasing overlay thickness. For better clarity Fig. 2b presents dependence of resonant minima wavelengths of different orders vs. overlay thickness.
216
K. Kosiel et al. / Optics and Laser Technology 102 (2018) 213–221
Fig. 2. Simulated response for a set of ZrO2-coated structures with different overlay thicknesses d in external RI equal to 1.33, where (a) shows transmission spectra and (b) resonance wavelength. In (a) the arrows indicate shift of the resonance minima for the increasing overlay thickness.
Fig. 3. Simulated evolution of the transmission spectra with external RI for the ZrO2-coated structure with 234 nm overlay thickness. The arrow indicates shift of the resonance minimum for the increasing external RI. The external RI values (in RIU) are given in the legend of the figure.
Fig. 3 refers to the simulated evolution of spectra for the ZrO2coated structure with changes in external RI, where redshift of the resonance with external RI is clearly seen. Due to equipment limitations we are able to measure the transmission spectra only in a limited wavelength range of a few hundred of nanometers. In Fig. 4 experimentally observed evolution of transmission response spectra with external RI changes for the HfO2- and ZrO2-coated structures is presented. The overlays show different values of n (nZrO2 > nHfO2) and the thickness of ZrO2 overlay is noticeably larger than for the HfO2-coated sample. This is the reason of very different shapes of their transmission responses and more resonances can be seen for the ZrO2-based structure. Anyway, for both sensors the redshift of each resonant minimum with the external RI is observed. This stays in qualitative agreement with numerical simulations shown as an example in Fig. 3. Evolution of the resonance wavelengths (Z1 and Z2) induced by external RI for the fiber structures with different ZrO2 thicknesses is presented in Fig. 5. The wavelength experiences a redshift with increase in thicknesses of the overlays that is anticipated by the numerical simulations (see Fig. 2). In particular it is seen in Fig. 5 that less than 10 nm change of overlay thickness can induce a shift of resonant minima by tens of nanometers - this gives an idea of the level of overlay thickness control that is necessary, when it is designed to tune the LMR sensors response. Moreover, an increase of the sensor sensitivity with external RI and with the overlay thickness is observed. It should be explained here that due to asymmetrical shape of the resonance Z2, which shows split into two minima, only the left-hand-side minima of Z2 (at shorter wavelength) were taken into account and presented in Fig. 5. The wavelength shifts as well as sensitivities for both Z1 and Z2 resonant minima measured in two different ranges of external RI were given in Table 1 (the two external RI exemplary ranges were selected: 1.33–1.35 and 1.41–1.43 RIU). They are presented for three fiber structures with different ZrO2 overlay thickness. It is seen that sensitivities for the two resonances are considerably higher in higher external RI range. The sensitivity for the resonance Z2 is higher than for the Z1 resonance. Moreover, for most the cases the sensitivity increases with overlay thickness, although the increase is not monotonic. The dependence of the observed resonance wavelength on overlay thickness and/or on external RI is a result of interaction of the polychromatic electromagnetic wave with the particular spatial distribution of the n⁄ with cylindrical geometry. The explanation of exact physical phenomena behind these effects are not straightforward. Nevertheless - as it is presented in this section - the electromagnetic wave behavior can be predicted by numerical analysis of physical models. As it was mentioned above, the presented theoretical and the experimental data stay in a qualitative agreement. Possible reasons of the quantitative differences between the theoretical and the experimental spectra may be originated by both the simplified model i.e., assumptions on negligible imperfections of the real fiber structures and many effects associated with electromagnetic wave propagation in cylindrical medium, as well as accuracy of the input data. The numerical model assumes ideally smooth external surfaces of the fiber core and the overlay, thus sharp interfaces. What is more, the cylindrical fiber is assumed to be much bigger than the wavelength and thus can be treated together with the core and overlay as planar stack of layers. The fiber components and the external environment are also assumed to be ideally uniform and isotropic. Moreover, the model assumes Lambertian distribution of light source which does not ideally represent the propagation conditions for electromagnetic wave in the discussed structure. Regarding the input data it should be noted that, as it was explained in Section 2.5, the n and k values of the deposited materials are extracted from ellipsometric characterization performed
217
K. Kosiel et al. / Optics and Laser Technology 102 (2018) 213–221
Fig. 4. Evolution of the normalized measured transmission spectra with external RI for the 211 nm-thick-HfO2-coated (a) and for the 244 nm-thick-ZrO2-coated (b) fiber structures. The arrows indicate shift of the resonance minima for the increasing external RI. The external RI values (in RIU) are given in the legends of the figures.
Fig. 5. Evolution of the resonance wavelength induced by change of external RI for the fiber structures with different ZrO2 thickness. The overlay thicknesses were as follows: 244 nm, 224 nm and 216 nm for the samples ZrO2_1, ZrO2_2 and ZrO2_3, respectively. The Z1 and Z2 resonances correspond to spectra type shown in Fig. 4b.
using planar test structures. As it was already mentioned, some other differences are possible between optical properties of the planar layers and the cylindrical overlays. 3.3. The fiber structures coated by TaxOy/SixNy-double-layer deposited by PECVD and ALD Various overlay materials show generally different properties that can be useful for optical fiber sensor technology, e.g., they differ in terms of hardness, elasticity, resistivity to various chemicals, thermal resistance, facility for being functionalized at the surface and many others [42]. For this reason combining more than one
overlaying material as a component of the LMR-based sensor can be found useful, as it has a potential for introducing additional functionalities. The PECVD-deposited SixNy has an advantage of having the largest n value and the appropriate k profile that promotes induction of LMR effect, both features being the most promising within the group of dielectric materials tested here (see Section 3.1). As it was noticed earlier, ALD is generally the most appropriate technique for easy and precise tuning of n and thickness of overlays. Thanks to this, the optical properties of the sensor (e.g., its resonant wavelength) can be tuned according to the need – by careful adjustment of the optical thickness of the overlay. ALD also shows an advantage of low-temperature deposition availability of a wide range of materials, as shown in this work HfO2, ZrO2 or TaxOy. However, not every material can be low-temperature-ALDdeposited while retaining satisfactory properties that are necessary for given application. This is also a case of SixNy, which allows rather high deposition temperature (typically at least 200 °C) for satisfactorily high n value (over 2). However, the relatively low deposition rate of ALD generally results in very time consuming processes. A couple-of-hours-long high temperature ALD process for approx. 200 nm-thick SixNy deposition, would be detrimental for parts of thermally vulnerable optical fibers. Taking into account that one can expect higher sensitivity of the LMR-based device equipped with even thicker overlay, the ALD-only fabrication of such thicker SixNy overlay seems to be highly problematic. On the other hand the PECVD technique has an unquestionable advantage of being a relatively fast-deposition method for thin layers. Comparison of the thin layer deposition rates (nm/s) presented in this work shows that PECVD is a technique by two orders of magnitude faster than ALD. Additionally it is easy to deposit even a few micrometers thick layers by PECVD, while ALD deposition is rather limited to sub-micrometer thickness because of technical reasons and time consumption. However, the capability of PECVD
Table 1 Wavelength shifts and sensitivities of Z1 and Z2 resonant minima measured in different ranges of external RI, for the three fiber structures with different ZrO2 overlay thickness. The Z1 and Z2 resonances correspond to spectra shown in Fig. 4b. ZrO2 thickness (nm)
216 224 244
Z1 minimum wavelength shift (nm)/sensitivity (nm/RIU)
Z2 minimum wavelength shift (nm)/sensitivity (nm/RIU)
In the range of 1.33–1.35 RIU
In the range of 1.41–1.43 RIU
In the range of 1.33–1.35 RIU
In the range of 1.41–1.43 RIU
1.06/53 0.57/ 28.50 2.18/109.00
5.52/276 5.52/276 6.09/304.50
3.33/166.5 5.42/271 3.90/195.00
13.80/690 12.65/632.50 17.60/880.00
218
K. Kosiel et al. / Optics and Laser Technology 102 (2018) 213–221
Fig. 6. Evolution of the normalized measured transmission spectra with external RI for the 410 nm-thick-SixNy-coated fiber structure (a) and for the same structure additionally coated by 50 nm-thick TaxOy (b), for the 450 nm-thick-SixNy-coated fiber structure (c) and for the same structure additionally coated by 50 nm-thick TaxOy (d), for the 510 nm-thick-SixNy-coated fiber structure (e) and for the same structure additionally coated by 50 nm-thick TaxOy (f). The arrows indicate shift of the resonance minima for the increasing external RI. The external RI values (in RIU) are given in the figures legends.
relating to overlay uniformity or thickness accuracy is noticeably lower than for ALD. The amount of PECVD-deposited film on a sample depends mainly on effectiveness of plasma-enhanced decomposition of gases and not only on mass located in the reaction chamber, but also relatively strongly on sample shape, sample location in chamber (plasma), particle (ion) flux, material of the sample and parameters of the process, such as gas composition, pressure, power, and time [56]. Though, the uniformity and accu-
racy of the PECVD–deposited overlay can be increased by optimization of technological factors, for example by rotation of the fiber during the process [64], it will never reach the level of the ALD. As a way to solve the above described problems, here we propose an approach to the fiber sensor overlay deposition being a combination of these two methods, and giving an advantage of short-term deposition process of an overlay (thanks to using
K. Kosiel et al. / Optics and Laser Technology 102 (2018) 213–221
Fig. 7. Evolution of the resonance wavelength induced by change of external RI for the fiber structures coated by the 410 nm-thick SixNy (SixNy_1), 450 nm-thick SixNy (SixNy_2), 510 nm-thick SixNy (SixNy_3), and for the same structures additionally coated by 50 nm-thick TaxOy (SixNy_1 + TaxOy, SixNy_2 + TaxOy and SixNy_3 + TaxOy, respectively). The data were extracted from measurements presented in Fig. 6, for the S2 minima (for the SixNy-coated fibers) and ST2 minima (for the TaxOy/SixNycoated fibers) in each case.
Table 2 Wavelength shifts and sensitivities of S2 and ST2 resonant minima measured in different ranges of external RI, for the six fiber structures with different total thicknesses of the TaxOy/SixNy overlay, and different TaxOy thickness percentages in the bilayer overlay. The S2 and ST2 resonances correspond to spectra shown in Fig. 6. TaxOy/SixNy overlay thickness (nm)
410 450 460 500 510 560
TaxOy thickness %
0 0 10.9 10 0 8.9
Wavelength shift (nm)/sensitivity (nm/RIU) In the range of 1.33–1.35 RIU
In the range of 1.43–1.45 RIU
1.37/68.5 2.66/133 2.20/110 3.12/156 5.79/289.5 21.54/1077
8.08/404 10.96/548 10.12/506 9.67/483.5 11.87/593.5 9.66/483
PECVD) and possibility of fine-tuning the target overlay properties - mainly in terms of its thickness, but also in terms of n-profile engineering according to the need (thanks to using ALD). Fig. 6a–e present the collection of transmission spectra for three pairs of the fiber structures. There are the results for the fibers coated only by PECVD SixNy with various thickness (dSixNy was in the range of 410–510 nm) in Fig. 6a, c, e. There are also the results for the same structures coated additionally by 50 nm-thick ALD TaxOy in Fig. 6b, d, f; therefore, the total thickness of the tested overlays was in the range of 410–560 nm. The relatively hightemperature PECVD process of SixNy as a part of each doublelayer overlay took a few minutes. Then, the sensorial response of each SixNy-coated device was analyzed optically. Next the examined fibers were introduced to the ALD system in order to deposit the ALD-based part of the overlay. Then the low-temperature ALD processes of 50 nm-thick TaxOy took about 44 min. Such a technological scheme can be easily adopted for fabrication of overlays with various proportions of the PECVD-based and the ALD-based part of the overlay in order to optimize the technology, taking into account the potentially poorer accuracy of PECVD-based layer thickness and the much larger time-consumption of ALD used for tuning the sensor properties up to the designed level. Like in the case of overlays presented in Section 3.2 a redshift of resonant wavelengths is seen with the increasing (TaxOy)/SixNy overlay thickness. As in the case of earlier presented fiber
219
structures, the resonances in longer wavelength range (S2, ST2) are more sensitive than those in shorter wavelength range (S1, ST1). It is seen, that resonance wavelength shift induced by change of overlay properties in a collection of sensors related to Fig. 6 covers approx. 180 nm-broad range of wavelength - considering only the positions of the S2 minima (for the SixNy-coated structures) as well as corresponding ST2 minima (for the TaxOy/SixNy-coated structures). The S2 and ST2 positions are tuned from structure to structure by a few tens of nanometers. However, as it was noticed in Section 3.2 by the example of ZrO2 overlay (which thickness was altered by less than 10 nm), a way more precise tuning is possible by using ALD - it is known that ALD allows for thickness tuning even with atomic layer precision [42]. The evolution of the resonance wavelength induced by change of external RI for the fiber structures with the described above SixNy and TaxOy/SixNy overlays is presented in Fig. 7. There are wavelength shifts and sensitivities of S2 and ST2 resonant minima presented in Table 2. These are presented for different ranges of external RI (the two external RI exemplary ranges were selected: 1.33–1.35 and 1.43–1.45 RIU). They were calculated for the six fiber structures with different total thicknesses of the (TaxOy)/SixNy overlay, and different TaxOy thickness percentages in the bilayer overlay. It is seen that sensitivity (nm/RIU) of the SixNy-only-coated sensors increases with the overlay thickness and reaches considerably high values in higher external RI. This behavior is similar to that observed for discussed earlier ALD-coated fibers. Also for the fibers coated by TaxOy/SixNy double-layers the sensitivity in lower external RI range increases with the overlay thickness - it should be noted that in the series of fibers with increasing double-layer thickness the TaxOy thickness percentage slightly decreases. In higher external RI, the sensitivity even slightly decreases with TaxOy/SixNy double-layer thickness. The case of two structures is worth noticing: the fiber coated by 50 nm TaxOy/410 nm SixNy and the fiber coated by 450 nm SixNy. Although the total overlay thickness for the double-layer coated fiber is 10 nm larger, the structure shows lower sensitivity. According to these results it can be concluded that the impact of TaxOy component of the overlay is different in the two external RI ranges. In both ranges it does not improve the sensitivity of the structures but, as it was mentioned above, helps to adjust the wavelength range for device interrogation purposes. The increase of sensitivity is provided rather by increasing thickness of the SixNy component of the overlay. It should be noted, however, that the n value of TaxOy - which lies farther from the fiber core than the SixNy - is lower than the n of the latter (see Fig. 1). Therefore, n does not increase monotonically in radial direction. This can be the reason of the observed differences in the sensitivity. However, the theoretical results presented in Ref. [59], suggest that some double-layer combinations should provide enhancement of sensitivity in comparison to such overlay components used separately on the fiber structures.
4. Summary and conclusions Application of the ALD technique as a tool for tailoring sensorial properties of LMR-based optical fiber sensors was discussed. The ALD was used for deposition of high-refractive-index materials (at the wavelength of 650 nm the n value was in the range 2.05– 2.1 RIU) for overlays of devices fabricated from PCS multimode optical fibers with partly removed claddings. The ALD-deposited HfO2, ZrO2, and TaxOy were applied. Application of ALD for LMRbased sensors is superior to the other deposition techniques and gives potential for accurate adjustment of overlay thickness and its conformity, uniformity and tightness. At the same time it
220
K. Kosiel et al. / Optics and Laser Technology 102 (2018) 213–221
ensures wide variety of dielectric materials (and by this the flexibility of n⁄ tuning) and low process temperature (that ensures safe conditions for thermally vulnerable fibers). For this reason we found ALD a first-choice tool for careful modification of sensorial properties of LMR-based sensors in terms of adjustment of their resonant wavelength positions, by even sub-nanometer-level overlay thickness control. Consequently, we presented examples of LMR sensors which overlay fabrication was done exclusively by ALD. At the same time ALD can be used as an element in more complicated technological scheme comprising also other deposition methods for overlay formation. Namely, we presented LMR-based sensors which overlay fabrication combined application of PECVD and ALD techniques. Such a technological scheme allowed for taking the advantage of high deposition rate of PECVD for creation of dominant bottom part of the overlay (in terms of its thickness), and using the ALD technique for fine-tuning the sensorial properties by deposition relatively thin upper part of the overlay, i.e., by doing this with the ALD-related superior thickness control. At the same time this is the first-time demonstration of optical fiber sensor double-layer overlay, composed of two different dielectric materials. It should be noted here, that the LMR sensor technology seems to be cheaper than the SPR-based one, because of relatively high cost of overlay materials used in the latter - the noble metals, the most often gold. While the ALD technique for dielectrics used in LMR sensors fabrication can still be found relatively expensive due to costs of chemical precursors and time consumption, thanks to the proposed approach where the two techniques, namely PECVD and ALD were combined, the LMR sensors technology can be cost-effective. Double-layer overlays can also provide different functionalities, such as, e.g., application of chemically resistive films as the most external part of a fiber structure for ensuring the chemical stability of the overall sensor.
Acknowledgments The authors gratefully acknowledge support for this work from the National Science Centre of Poland under grant No. 2014/14/E/ ST7/00104. The work was also financially supported as statutory activities at the Instytut Technologii Elektronowej. Kamil Kosiel is grateful to Dr Anna Szerling for her help during the paper preparation.
References [1] N. Sabri, S.A. Aljunid, M.S. Salim, R.B. Ahmad, R. Kamaruddin, Toward optical sensors: review and applications, J. Phys.: Conf. Series, 423 (2013) 012064. [2]. [3] M. Ramakrishnan, G. Rajan, Y. Semenova, G. Farrell, Overview of fiber optic sensor technologies for strain/temperature sensing applications in composite materials, Sensors 16 (2016) 99. [4] É. Pinet, Pressure measurement with fiber-optic sensors: commercial technologies and applications, in: Wojtek J. Bock, Jacques Albert, Xiaoyi Bao (Eds.), 21st International Conference on Optical Fiber Sensors, Proc. of SPIE, vol. 7753, 775304, 2011 SPIE. [5] S. Poeggel, D. Tosi, D.B. Duraibabu, G. Leen, D. McGrath, E. Lewis, Optical fibre pressure sensors in medical applications, Sensors 15 (2015) 17115–17148. [6] S.W. Harun, M. Yasin, H.Z. Yang, H. Ahmad, Fiber optic displacement sensors and their applications, in: M. Yasin, S.W. Harun, H. Arof (Eds.), Fiber Optic Sensors, Publisher: InTech, 2012 (Chapter 15). [7] H. Togo, S. Mochizuki, N. Kukutsu, Optical fiber electric field sensor for antenna measurement, NTT Tech. Rev. 7 (2009) 3. [8] Y. Zheng, X. Dong, Chan Ch Ch, P.P. Shum, H. Su, Optical fiber magnetic field sensor based on magnetic fluid and micro fiber mode interferometer, Optics Commun. 336 (2015) 5–8. [9] R. Forber, W.C. Wang, D.-Y. Zang, S. Schultz, R. Selfridge, Dielectric EM Field Probes for HPM Test & Evaluation, Annual ITEATechnology Review, August 7– 10, Cambridge, MA, Non-Intrusive Instrumentation Technologies, 2006.
[10] Feng De-Jun, Liu Guan-Xiu, Liu Xi-Lu, Jiang Ming-Shun, Sui Qing-Mei, Refractive index sensor based on plastic optical fiber with tapered structure, Appl. Optics 53 (10) (2014) 2007–2011. [11] M. Smietana, D. Brabant, W.J. Bock, P. Mikulic, T. Eftimov, Refractive-index sensing with inline core-cladding intermodal interferometer based on silicon nitride nano-coated photonic crystal fiber, J. Lightwave Technol. 30 (8) (2012) 1185. [12] M.F. Jaddoa, A.A. Jasim, M.Z.A. Razak, S.W. Harun, H. Ahmad, Highly responsive NaCl detector based on inline microfiber Mach-Zehnder interferometer, Sens. Actuat. A 237 (2016) 56–61. [13] M. Smietana, W.J. Bock, P. Mikulic, A. Ng, R. Chinnappan, M. Zourob, Detection of bacteria using bacteriophages as recognition elements immobilized on long-period fiber gratings, Optics Express 19 (9) (2011) 7971. [14] A. Leung, P.M. Shankar, R. Mutharasan, A review of fiber-optic biosensors, Sens. Actuat. B 125 (2007) 688–703. [15] Ch. Caucheteur, V. Malachovska, C. Ribaut, R. Wattiez, Cell sensing with nearinfrared plasmonic optical fiber sensors, Optics Laser Technol. 78 (2016) 116– 121. [16] I.M. Ishaq, A. Quintela, S.W. James, G.J. Ashwell, J.M. Lopez-Higuera, R.P. Tatam, Modification of the refractive index response of long period gratings using thin film overlays, Sens. Actuat. B: Chem. 107 (2) (2005) 738–741. [17] D. Leamy, F. Regan, A robust spin-coated optical sensing device for rapid determination of predominant metal ions in wastewater streams, Int. J. Environ. Anal. Chem. 83 (10) (2003) 867–877. [18] Š.M. Korent, A. Lobnik, G.J. Mohr, Sol-gel-based optical sensor for the detection of aqueous amines, Anal. Bioanal. Chem. 387 (8) (2007) 2863–2870. [19] L.Y. Shao, M.J. Yin, H.Y. Tam, J. Albert, Fiber optic pH sensor with selfassembled polymer multilayer nanocoatings, Sensors 13 (2) (2013) 1425– 1434. [20] R. Bogdanowicz, M. S´mietana, M. Gnyba, Ł. Gołunski, J. Ryl, M. Garda, Optical and structural properties of polycrystalline CVD diamond films grown on fused silica optical fibres pre-treated by high-power sonication seeding, Appl. Phys. A 116 (4) (2014) 1927–1937. [21] W. Xie, M. Yang, Y. Cheng, D. Li, Y. Zhang, Z. Zhuang, Optical fiber relativehumidity sensor with evaporated dielectric coatings on fiber end-face, Optical Fiber Technol. 20 (4) (2014) 314–319. [22] K. Schmitt, K. Oehse, G. Sulz, C. Hoffmann, Evanescent field sensors based on tantalum pentoxide waveguides–a review, Sensors 8, 2008, 711–738. [23] R.W. Johnson, A. Hultqvist, S.F. Bent, A brief review of atomic layer deposition: from fundamentals to applications, Mater. Today 17 (5) (2014) 236–246. [24] Ch. Das, M. Kot, K. Henkel, D. Schmeisser, Engineering of sub-nanometer SiOx thickness in Si photocathodes for optimized open circuit potential, ChemSusChem 9 (2016) 2332–2336. [25] S.E. Potts, W. Keuning, E. Langereis, G. Dingemans, M.C.M. van de Sanden, W.M. M. Kessels, Low temperature plasma-enhanced atomic layer deposition of metal oxide thin films, J. Electrochem. Soc. 157 (7) (2010) 66–74. [26] M. Sowinska, K. Henkel, D. Schmeißer, I. Karkkanen, J. Schneidewind, F. Naumann, B. Gruska, H. Gargour, Analysis of nitrogen species in titanium oxynitride ALD films, J. Vacuum Sci. Technol. A: Vacuum Surf. Films 34 (1) (2016) 01A127. [27] M. Sowinska, S. Brizzi, Ch. Das, I. Kärkkänen, J. Schneidewind, F. Naumann, H. Gargouri, K. Henkel, D. Schmeißer, Analysis of nitrogen species in titanium oxynitride ALD films, Appl. Surf. Sci. 381 (2016) 42–47. [28] H. Kim, Atomic layer deposition of metal and nitride thin films: current research efforts and applications for semiconductor device processing, J. Vacuum Sci. Technol. B 21 (2003) 2231. [29] M.B.E. Griffiths, P.J. Pallister, D.J. Mandia, S.T. Barry, Atomic layer deposition of gold metal, Chem. Mater. 28 (1) (2016) 44–46. [30] Z. Guo, H. Li, Q. Chen, L. Sang, L. Yang, Z. Liu, X. Wang, Low-temperature atomic layer deposition of high purity, smooth, low resistivity copper films by using amidinate precursor and hydrogen plasma, Chem. Mater. 27 (2015) 5988– 5996. [31] M. Kot, C. Das, Z. Wang, K. Henkel, Z. Rouissi, K. Wojciechowski, H.J. Snaith, D. Schmeisser, ChemSusChem 9 (2016) 3401. [32] I.J. Raaijmakers, Current and future applications of ALD in micro-electronics, ECS Trans. 41 (2) (2011) 3–17. [33] H. Kim, Atomic layer deposition of metal and nitride thin films: current research efforts and applications for semiconductor device processing, J. Vacuum Sci. Technol. B: Microelectron. Nanometer Struct. Process. Measure. Phenomena 21 (6) (2003) 2231–2261. [34] H. Im, N.J. Wittenberg, N.C. Lindquist, S.H. Oh, Atomic layer deposition (ALD): a versatile technique for plasmonics and nanobiotechnology, J. Mater. Res. 27 (4) (2012) 663–671. [35] M. Godlewski, S. Gierałtowska, Ł. Wachnicki, R. Pietuszka, B.S. Witkowski, A. Słon´ska, Z. Gajewski, M.M. Godlewski, High-k oxides by atomic layer deposition-applications in biology and medicine, J. Vacuum Sci. Technol. A: Vacuum Surf. Films 35 (2) (2017) 021508. [36] M. Dominik, J. Niedziółka-Jönsson, E. Roz´niecka, Ł. Wachnicki, M. Godlewski, P. Mikulic, W.J. Bock, M. S´mietana, Regeneration of titanium oxide nano-coated long-period grating biosensor, in: Sixth European Workshop on Optical Fibre Sensors (Vol. 9916, p. 99160Z) International Society for Optics and Photonics, 2016. [37] M. Dominik, A. Les´niewski, M. Janczuk, J. Niedziółka-Jönsson, M. Hołdyn´ski, Ł. Wachnicki, M. Godlewski, W.J. Bock, M. S´mietana, Titanium oxide thin films obtained with physical and chemical vapour deposition methods for optical biosensing purposes, Biosens. Bioelectron. 93 (2017) 102–109.
K. Kosiel et al. / Optics and Laser Technology 102 (2018) 213–221 [38] M. S´mietana, M. Mys´liwiec, P. Mikulic, B.S. Witkowski, W.J. Bock, Capability for fine tuning of the refractive index sensing properties of long-period gratings by atomic layer deposited Al2O3 overlays, Sensors 13(12) (2013) 16372– 16383. [39] Y. Zhao, F. Pang, Y. Dong, J. Wen, Z. Chen, T. Wang, Refractive index sensitivity enhancement of optical fiber cladding mode by depositing nanofilm via ALD technology, Optics Express 21 (22) (2013) 26136–26143. [40] A. Purniawan, G. Pandraud, T.S.Y. Moh, A. Marthen, K.A. Vakalopoulos, P.J. French, P.M. Sarro, Fabrication and optical measurements of a TiO 2-ALD evanescent waveguide sensor, Sens. Actuat. A: Phys. 188 (2012) 127–132. [41] M. S´mietana, M. Koba, P. Mikulic, W.J. Bock, Towards refractive index sensitivity of long-period gratings at level of tens of mm per refractive index unit: fiber cladding etching and nano-coating deposition, Optics Express 24 (11) (2016) 11897–11904. [42] K. Kosiel, M. Dominik, I. S´cis´lewska, M. Kalisz, M. Guziewicz, K. Gołaszewska, J. Niedziółka-Jonsson, W.J. Bock, M. S´mietana, Alkali-resistant low-temperature atomic-layer-deposited oxides for optical fiber sensor overlays, Nanotechnology (2018) (submitted for publication). [43] I. Del Villar, F.J. Arregui, C.R. Zamarreño, J.M. Corres, C. Bariain, J. Goicoechea, C. Elosua, M. Hernaez, P.J. Rivero, A.B. Socorro, A. Urrutia, P. Sanchez, P. Zubiate, D. Lopez, N. De Acha, J. Ascorbe, I.R. Matias, Optical sensors based on lossy mode resonances, Sens. Actuat. B 240 (2017) 174–185. [44] P. Zubiate, C.R. Zamarreño, P. Sánchez, I.R. Matias, F.J. Arregui, High sensitive and selective C-reactive protein detection by means of lossy mode resonance based optical fiber devices, Biosens. Bioelectron. 93 (2017) 176–181. [45] S. Zhu, F. Pang, S. Huang, F. Zou, Y. Dong, T. Wang, High sensitivity refractive index sensor based on adiabatic tapered optical fiber deposited with nanofilm by ALD, Optics Express 23 (11) (2015) 13880–13888. [46] M. Hernáez, I. Del Villar, C.R. Zamarreño, F.J. Arregui, I.R. Matias, Optical fiber refractometers based on lossy mode resonances supported by TiO2 coatings, Appl. Optics 49 (20) (2010) 3980–3985. [47] S. Zhu, F. Pang, S. Huang, F. Zou, Q. Guo, J. Wen, T. Wang, High sensitivity refractometer based on TiO2-coated adiabatic tapered optical fiber via ALD technology, Sensors 16(8) (2016) 1295. [48] P. Sanchez, C.R. Zamarreno, C.R. Zamarreo, M. Hernaez, I.R. Matias, F.J. Arregui, Exhaled breath optical fiber sensor based on LMRs for respiration monitoring, in: SENSORS, 2014 IEEE (2014) 1142–1145. [49] I. Del Villar, C.R. Zamarreño, P. Sanchez, M. Hernaez, C.F. Valdivielso, F.J. Arregui, I.R. Matias, Generation of lossy mode resonances by deposition of high-refractive-index coatings on uncladded multimode optical fibers, J. Optics 12 (9) (2010) 095503. [50] I. Del Villar, C.R. Zamarreño, M. Hernaez, F.J. Arregui, I.R. Matias, Lossy mode resonance generation with indium-tin-oxide-coated optical fibers for sensing applications, J. Lightwave Technol. 28 (1) (2010) 111–117.
221
[51] C.R. Zamarreño, M. Hernaez, I. Del Villar, I.R. Matias, F.J. Arregui, Tunable humidity sensor based on ITO-coated optical fiber, Sens. Actuat. B: Chem. 146 (1) (2010) 414–417. [52] D. Kaur, V.K. Sharma, A. Kapoor, High sensitivity lossy mode resonance sensors, Sens. Actuat. B: Chem. 198 (2014) 366–376. [53] C.R. Zamarreño, S. Lopez, M. Hernaez, I. Del Villar, I.R. Matias, F.J. Arregui, Resonance-based refractometric response of cladding-removed optical fibers with sputtered indium tin oxide coatings, Sens. Actuat. B: Chem. 175 (2012) 106–110. [54] P. Zubiate, C.R. Zamarreño, I. Del Villar, I.R. Matias, F.J. Arregui, High sensitive refractometers based on lossy mode resonances (LMRs) supported by ITO coated D-shaped optical fibers, Optics Express 23 (6) (2015) 8045–8050. [55] Indium CorporationÒ Application Note, Indium Oxide and Indium-Tin Oxide (ITO) Coatings.. [56] M. S´mietana, M. Dominik, M. Mys´liwiec, N. Kwietniewski, P. Mikulic, B.S. Witkowski, W.J. Bock, Properties of silicon nitride thin overlays deposited on optical fibers—effect of fiber suspension in radio frequency plasma-enhanced chemical vapor deposition reactor, Thin Solid Films 603 (2016) 8–13. [57] M. Smietana, M. Dudek, M. Koba, B. Michalak, Influence of diamond-like carbon overlay properties on refractive index sensitivity of nano-coated optical fibres, Phys. Status Solidi A 210 (2013) 2100–2105. [58] A.B. Socorro, I. Del Villar, J.M. Corres, F.J. Arregui, I.R. Matias, Spectral width reduction in lossy mode resonance-based sensors by means of tapered optical fibre structures, Sens. Actuat. B: Chem. 200 (2014) 53–60. [59] N. Paliwal, J. John, Theoretical modeling of lossy mode resonance based refractive index sensors with ITO/TiO2 bilayers, Appl. Optics 53 (15) (2014) 3241–3246. [60] R.L. Puurunen, Surface chemistry of atomic layer deposition: a case study for the trimethylaluminum/water process, J. Appl. Phys. 97 (2005) 121301. [61] C.D. Travis, R.A. Adomaitis, Modeling alumina atomic layer deposition reaction kinetics during the trimethylaluminum exposure, Theor. Chem. Acc. 133 (2014) 1414. [62] M. Blech, A. Laades, C. Ronning, B. Schröter, C. Borschel, D. Rzesanke, A. Lawerenz, Detailed study of PECVD silicon nitride and correlation of various characterization techniques, in: Proceedings of 24th European Photovoltaic Solar Energy Conference 21-25 September (2009) Hamburg Germany. [63] M. Smietana, W.J. Bock, J. Szmidt, Evolution of optical properties with deposition time of silicon nitride and diamond-like carbon films deposited by radio-frequency plasma-enhanced chemical vapor deposition method, Thin Solid Films 519 (2011) 6339–6343. [64] B. Michalak, M. Koba, M. S´mietana, Silicon nitride overlays deposited on optical fibers with RF PECVD method for sensing applications: overlay uniformity aspects, Acta Phys. Polonica A 127 (6) (2015) 1587–1590.
Contents lists available at ScienceDirect
Optics and Laser Technology journal homepage: www.elsevier.com/locate/optlastec
Full length article
Tailoring properties of lossy-mode resonance optical fiber sensors with atomic layer deposition technique q Kamil Kosiel a,⇑, Marcin Koba b,c, Marcin Masiewicz b, Mateusz S´mietana b a
Institute of Electron Technology, Al. Lotników 32/46, 02-668 Warsaw, Poland Institute of Microelectronics and Optoelectronics, Koszykowa 75, 00-662 Warsaw, Poland c National Institute of Telecommunications, Szachowa 1, 04-894 Warsaw, Poland b
a r t i c l e
i n f o
Article history: Received 14 October 2017 Received in revised form 29 December 2017 Accepted 2 January 2018
Keywords: Optical fiber sensors Lossy-mode resonance Thin films Atomic layer deposition Optical properties
a b s t r a c t The paper shows application of atomic layer deposition (ALD) technique as a tool for tailoring sensorial properties of lossy-mode-resonance (LMR)-based optical fiber sensors. Hafnium dioxide (HfO2), zirconium dioxide (ZrO2), and tantalum oxide (TaxOy), as high-refractive-index dielectrics that are particularly convenient for LMR-sensor fabrication, were deposited by low-temperature (100 °C) ALD ensuring safe conditions for thermally vulnerable fibers. Applicability of HfO2 and ZrO2 overlays, deposited with ALD-related atomic level thickness accuracy for fabrication of LMR-sensors with controlled sensorial properties was presented. Additionally, for the first time according to our best knowledge, the doublelayer overlay composed of two different materials - silicon nitride (SixNy) and TaxOy - is presented for the LMR fiber sensors. The thin films of such overlay were deposited by two different techniques – PECVD (the SixNy) and ALD (the TaxOy). Such approach ensures fast overlay fabrication and at the same time facility for resonant wavelength tuning, yielding devices with satisfactory sensorial properties. Ó 2018 Elsevier Ltd. All rights reserved.
1. Introduction Nowadays optical sensors already allow for measuring majority of physical and chemical measurands of interest [1]. These are known various temperature [2], [3], pressure [4,5], strain [3], displacement [6], electromagnetic field [7–9] and refractive index [10,11] sensors, as well as chemical [12] and biosensors targeted towards certain molecule or even whole biological cell [13–15] to name only a few. While the modern optical sensors technologies cover a wide range of device architectures, physical effects and sensing strategies, the sensors often require deposition of thin films (overlays) or surface structures. The properties of these films or structures can be various and in particular they can be metallic or dielectric. The mentioned overlays or structures are used to initiate or modify sensorial response of the sensors, allowing for precise adjustment of the sensors properties according to the needs. Therefore in most cases the thin overlay thickness and its optical properties (primarily the complex refractive index - n⁄) should be strictly controlled. Contemporary thin film fabrication techniques are expected to provide high deposition accuracy and process-to-process reproducibility, as well as coating uniformity on various substrates and q
Invited article by the Guest Editor: Ignacio Del Vilar.
⇑ Corresponding author.
E-mail address: [email protected] (K. Kosiel). https://doi.org/10.1016/j.optlastec.2018.01.002 0030-3992/Ó 2018 Elsevier Ltd. All rights reserved.
shapes. These circumstances impose stringent requirements on control of the overlay deposition rate and the overlay properties on different stages of deposition. The matter of technological requirements becomes even more critical when three-dimensional objects like optical-fiber-based sensors are considered. All these make the matter of choice of the thin film deposition technique crucial. Many nanofilm deposition techniques have already been used in the field of optical sensing, such as Langmuir-Blodgett method [16], spin coating [17], sol-gel process [18], layer-by-layer technique [19], chemical vapor deposition (CVD) [20], or physical vapor deposition (PVD) methods, which include evaporation [21] or sputtering [22]. They all have specific drawbacks, such as poor deposition process control, low coating robustness or lack of accuracy and coating uniformity. Atomic layer deposition (ALD), which is a specific variation of the CVD method, seems to be the first-choice technique enabling control of the sensors properties by means of precise adjustment of the overlay features. The ALD chemical precursors are delivered to the reaction zone separately in time, and hence the complementary and sequentially repeated chemisorption half-cycle reactions responsible for the layer growth undergo at a surface in a selflimiting manner [23]. Thanks to this very special paradigm an atomic control of the layer thickness is possible, with uniquely conformal, uniform and tight layers, even when they are deposited
214
K. Kosiel et al. / Optics and Laser Technology 102 (2018) 213–221
on complicated high-aspect-ratio surfaces. These and other advantages of ALD, such as capability for deposition of dense layers in a wide range of materials (oxides [24,25], nitrides [26–28], metals [29,30] and many others isolators, semiconductors and conductors), possible deposition at relatively low temperatures (even a way below 100 °C [31,25]) lead to a situation where the ALD technique is a must in nowadays microelectronics [32,33]. At the same time, it reaches more and more applications in other very divers branches [34,35], and has already begun to be used in optical sensors technology too [36–42]. The phenomenon of lossy-mode-resonance (LMR), being the physical basis of operation of a specific class of sensors, results from the fact that for certain thickness and optical properties of the dielectric overlay the light in the thin-film-coated fiber core experiences wavelength-dependent attenuation, i.e., in the optical spectrum characteristic transmission minima appear [43]. The effect is a consequence of coupling between core modes and specific lossy modes of the overlay. The condition of appearance of LMR is that the real part of the overlay permittivity is positive and larger in magnitude than both its own imaginary part and the real part of permittivity of the materials surrounding the overlay, i.e., of the fiber core and an external medium. The crucial issue for sensing applications is that spectral positions of the minima depend also on refractive index of the surrounding medium (external RI). This fact enables the use of LMR structures as refractometers. On top of it, the sensor surface functionalization allows also for biosensing applications of these sensors [44]. It must be noted that the spectral position of the resonant wavelengths and the device sensitivity can be fine-tuned just by adjustment of the thickness and optical properties of the overlay. The most often used for the purpose of overlay for LMR-based fiber sensors are oxides of aluminium (Al2O3) [45], titanium (TiO2) [46,47], tin (SnO2) [48], indium (In2O3) [49], or indium doped tin oxide (ITO) [50–55]. Also silicon nitride (SixNy) [56], diamond-like carbon [57] and polymeric overlays, like polyallylamine hydrochloride/polyacrylic acid (PAH/PAA) [58] have already been used. However, the overlays on the fiber sensors have been deposited in the form of single layers only, i.e., just one material at a time has been used for formation of an individual overlay. The possible reason for this can be a complicated numerical modelling of such a sensor, but also a number of technological difficulties. According to our best knowledge only once multilayer LMRbased structure was discussed theoretically, but the concept has never been practically implemented [59]. In this paper we present capability for tuning the properties of optical fiber LMR sensors by using the ALD-based dielectric overlays. We show the possibility of tuning its sensing properties by ALD-deposited single-layer overlays, and also we put into practice the concept of double-layer dielectric overlays for the first time. Such overlays comprise two layers, each deposited by using different technique, namely ALD and plasma-enhanced CVD (PECVD). The presented solution offers to the field of thin film optical fiber sensors technological flexibility available thanks to application of the ALD technique.
2. Experimental details 2.1. Numerical modelling The numerical model used for the simulations of the LMR phenomenon has already been described elsewhere [57]. The aim of the performed simulations is to show the effect of variation in the LMR response when properties of the overlays change and to determine the range of the properties where the effect can be seen. Wavelength-dependent response of the structures to variations in
external RI was normalized using spectrum obtained for a bare sample surrounded by air. 2.2. The fibers and their preparation before deposition of the overlays For preparation of the sensors polymer-clad silica (PCS) multimode optical fibers with 400 lm core diameter were used. While total length of the fiber samples was up to 150 mm the only 25 mm-long central section of each fiber was stripped mechanically and chemically from outer polymer coating and cladding [57]. Then the fiber samples were extensively cleaned with isopropanol. 2.3. Deposition of the overlays by ALD The metal oxides were deposited by ALD using thermal mode at the thermocouple-controlled reactor temperature of 100 ± 0.1 °C, in Beneq TFS-200–190 system. Such low temperature of the process was unavoidable because of the thermal vulnerability of a polymer coating that majority of optical fiber length is originally covered by. Additionally, low temperature enables deposition of amorphous films and hence smoother surfaces than for also available for ALD polycrystalline forms. Amorphous films are preferred due to more uniform film thickness distribution as well as for suppression of light scattering. The carrier/purging gas was 6N-purity argon (Ar) flowing through the reactor with constant rate of 1000 sccm that generated the pressure inside the reactor of approx. 2 mbar. As the oxygen chemical precursor was used deionized water (DI H2O) kept at 19 ± 0.1 °C in Beneq’s standard LS container. Tantalum pentachloride (TaCl5, kept at 85 ± 0.1 °C), tetrakis(ethylme thylamino)zirconium (TEMAZ, kept at 75 ± 0.1 °C), and tetrakis(et hylmethylamino)hafnium (TEMAH, kept at 78 ± 0.1 °C), all delivered by Volatec Oy and kept in Beneq’s standard HS 300 containers, were used as chemical precursors for tantalum, zirconium and hafnium, respectively. The precursor exposure times, that are responsible for substrate-surface-saturation conditions for all complementary half reactions of chemisorption involved in the employed ALD processes, were experimentally determined before the experiments described in this paper; the crucial role of exposure of complementary precursors and surface saturation in ALD reactions has already been explained, e.g., in Ref. [60,61]. They were studied using Si test substrates, and they are not described here. On the basis of these experiments the appropriate precursor/purge-cycling schemes were determined. For the optimized cycling conditions used in this work the growth per cycle (GPC) measured for Si-based samples was approx. 0.133, 0.150 and 0.145 nm/cycle for the TaxOy, ZrO2 and HfO2 films, respectively. The film growth rate was approx. 0.019, 0.022, and 0.021 nm/s for the TaxOy, ZrO2 and HfO2, respectively. It should be noticed here, that generally the composition of as-deposited TaxOy may vary and contain compounds of tantalum with various oxidation states between +2 and +5, i.e., TaO, Ta2O3, TaO2 and Ta2O5. The thickness range of the ALD-based layers investigated in this paper was 201–211 nm for the HfO2, 216–244 nm for ZrO2, and 50 nm for the TaxOy. Before entering the ALD reactor the fibers were rinsed extensively with acetone and isopropanol at the ambient temperature (approx. 20 °C). For each ALD process three fibers were introduced into the reactor and they were positioned in a run-to-runrepeatable arrangement. 2.4. Deposition of the overlays by PECVD The SixNy thin films were deposited on the fiber samples using Oxford Plasmalab 80 Plus system. The fiber samples were placed in the plasma reactor on U-type holder on height of approx. 5 mm over the electrode. The reference Si wafers were placed next to
K. Kosiel et al. / Optics and Laser Technology 102 (2018) 213–221
215
them on the electrode. The radio-frequency (RF) PECVD process was preceded by 1 min-long plasma cleaning in Ar. The deposition process was performed with SiH4 flow of 500 sccm, NH3 flow of 50 sccm, pressure of 1.2 Torr, power of 50 W. The electrode temperature was kept at 200 °C remaining a controlled parameter and deposition on Si wafers took place in such thermal conditions. The temperature of the fiber samples was not measured directly and taking into account their placement it was slightly lower than 200 °C. The resulting films with thicknesses in the range of 410– 510 nm were deposited with growth rate of approx. 1.2 nm/s. In general, PECVD-deposited SixNy films can have different compositions of y/x in the range of approx. 0.5–1.5 [62]. 2.5. The ex situ control of thicknesses and optical properties of the overlays In order to allow for ex situ control of thicknesses and n⁄ values of the overlays deposited on the sensors, also test layers were deposited within the ALD and PECVD processes on reference Si wafers. Only these were analyzed by spectroscopic ellipsometry (SE), since such analyses could not be performed directly on the fiber samples. The ALD-related reference Si wafers were positioned in a run-to-run-repeatable manner inside the reactor. As it was revealed above the PECVD-related reference Si wafers were placed next to fibers on the electrode. The SE measurements were performed using Horiba–Jobin–Yvon UVISEL spectroscopic ellipsometer in the wavelength range 350–900 nm. The properties of the films were calculated by fitting SE data to single layer Tauc–Lorenc dispersion model [57,63]. 2.6. Measurements of the optical fiber structures After the ALD and PECVD deposition processes both ends of each fiber were polished in order to remove the overlays covering their ends and to minimize the transmitted optical signal losses. The transmission spectral response of the optical fiber samples was measured in wavelength range 350–1050 nm using Ocean Optics HL-200-HP white light source and USB4000 spectrometer. The applied measurement setup was described in details earlier in [57]. For the spectral response measurements performed in various external RIs the coated segment of the fiber was immersed in a series of aqueous solutions of glycerin with various concentrations. The respective RIs of the obtained liquids were measured using Reichert AR200 Automatic Digital Refractometer with accuracy ±10 4 RI units (RIU). Wavelength-dependent response of the structures to variations in external RI was normalized using spectrum obtained for a bare sample surrounded by air.
Fig. 1. Dispersion of n and k for the ALD HfO2, ZrO2, TaxOy and the PECVD SixNy films.
resonant wavelengths for the LMR sensors. The examples of this will be presented in further subsections. Here, an assumption was made that the properties of the films deposited on test substrates could satisfactorily correspond with the properties of the overlays deposited on the fibers. It should be noted here that properties of ALD process, e.g., the GPC, can vary depending on the applied substrate [60]. Mainly surface density of the functional groups that are active in chemisorption reactions is determined by the substrate. However, we confirmed that the thickness and n⁄ of oxides deposited within the same process on Si wafers and glass plates were comparable. The glass plate surfaces can be assumed to be similar to the surfaces of silica fibers used for the sensors fabrication. These observations would justify the use of Si wafers as reference substrate for ellipsometric analysis. On the other hand routine usage of the glass-based reference samples, would be much more time-consuming, as they would need complex preparation aiming for reduction of multiple reflections in a glass plate. At the same time one should also be aware that the properties of PECVD SixNy on the fibers could very likely be different (in terms of thickness and n⁄) than these measured on the Si-based references. The discrepancies steam from the mentioned earlier different placement of the Si wafers and fibers in the reactor. Moreover the values of n and k can be influenced by strain and temperature, which can be different for the planar layers and cylindrical overlays on the fibers. However, the obtained SE results were critically compared with the results of transmission measurements performed for the fibers and with their simulations in order to draw final conclusions on the overlay properties.
3. Results and discussion 3.2. The fiber structures coated by ALD HfO2 or ZrO2 3.1. Properties of the films The dispersion relations of refractive index (n) and extinction coefficient (k) measured for the films deposited on the reference Si wafers are presented in Fig. 1. As it is seen within the tested group of ALD-based materials we are able to tune the n value in the range of approx. 2.05–2.1 RIU given here for the wavelength equal to 650 nm, which we can further expand using the PECVD SixNy. The k values were low for all materials in the tested wavelength range. Given additionally the wide range of film thickness that is accessible for these deposition techniques (for ALD typically up to 250 nm and for PECVD even by an order of magnitude larger) it should be noticed that this provides very wide range of possible tuning of sensor properties, particularly in terms of accessible
Assuming optical properties of the films shown in Fig. 1, the numerical simulations predict presence of a set of LMR-effects occurring in available range of transmission spectrum for above certain overlay thickness. Fig. 2a shows a collection of spectra simulated in a wide range of wavelength for the ZrO2-coated structures of various thicknesses and for a fixed value of external RI. It is seen, that each spectrum contains two prominent resonant minima – which we further refer to as the 1st order minimum (for shorter wavelength, between approx. 350–400 nm) and the 2nd order minimum (for longer wavelength, over 600 nm). For both minima a redshift is seen with increasing overlay thickness. For better clarity Fig. 2b presents dependence of resonant minima wavelengths of different orders vs. overlay thickness.
216
K. Kosiel et al. / Optics and Laser Technology 102 (2018) 213–221
Fig. 2. Simulated response for a set of ZrO2-coated structures with different overlay thicknesses d in external RI equal to 1.33, where (a) shows transmission spectra and (b) resonance wavelength. In (a) the arrows indicate shift of the resonance minima for the increasing overlay thickness.
Fig. 3. Simulated evolution of the transmission spectra with external RI for the ZrO2-coated structure with 234 nm overlay thickness. The arrow indicates shift of the resonance minimum for the increasing external RI. The external RI values (in RIU) are given in the legend of the figure.
Fig. 3 refers to the simulated evolution of spectra for the ZrO2coated structure with changes in external RI, where redshift of the resonance with external RI is clearly seen. Due to equipment limitations we are able to measure the transmission spectra only in a limited wavelength range of a few hundred of nanometers. In Fig. 4 experimentally observed evolution of transmission response spectra with external RI changes for the HfO2- and ZrO2-coated structures is presented. The overlays show different values of n (nZrO2 > nHfO2) and the thickness of ZrO2 overlay is noticeably larger than for the HfO2-coated sample. This is the reason of very different shapes of their transmission responses and more resonances can be seen for the ZrO2-based structure. Anyway, for both sensors the redshift of each resonant minimum with the external RI is observed. This stays in qualitative agreement with numerical simulations shown as an example in Fig. 3. Evolution of the resonance wavelengths (Z1 and Z2) induced by external RI for the fiber structures with different ZrO2 thicknesses is presented in Fig. 5. The wavelength experiences a redshift with increase in thicknesses of the overlays that is anticipated by the numerical simulations (see Fig. 2). In particular it is seen in Fig. 5 that less than 10 nm change of overlay thickness can induce a shift of resonant minima by tens of nanometers - this gives an idea of the level of overlay thickness control that is necessary, when it is designed to tune the LMR sensors response. Moreover, an increase of the sensor sensitivity with external RI and with the overlay thickness is observed. It should be explained here that due to asymmetrical shape of the resonance Z2, which shows split into two minima, only the left-hand-side minima of Z2 (at shorter wavelength) were taken into account and presented in Fig. 5. The wavelength shifts as well as sensitivities for both Z1 and Z2 resonant minima measured in two different ranges of external RI were given in Table 1 (the two external RI exemplary ranges were selected: 1.33–1.35 and 1.41–1.43 RIU). They are presented for three fiber structures with different ZrO2 overlay thickness. It is seen that sensitivities for the two resonances are considerably higher in higher external RI range. The sensitivity for the resonance Z2 is higher than for the Z1 resonance. Moreover, for most the cases the sensitivity increases with overlay thickness, although the increase is not monotonic. The dependence of the observed resonance wavelength on overlay thickness and/or on external RI is a result of interaction of the polychromatic electromagnetic wave with the particular spatial distribution of the n⁄ with cylindrical geometry. The explanation of exact physical phenomena behind these effects are not straightforward. Nevertheless - as it is presented in this section - the electromagnetic wave behavior can be predicted by numerical analysis of physical models. As it was mentioned above, the presented theoretical and the experimental data stay in a qualitative agreement. Possible reasons of the quantitative differences between the theoretical and the experimental spectra may be originated by both the simplified model i.e., assumptions on negligible imperfections of the real fiber structures and many effects associated with electromagnetic wave propagation in cylindrical medium, as well as accuracy of the input data. The numerical model assumes ideally smooth external surfaces of the fiber core and the overlay, thus sharp interfaces. What is more, the cylindrical fiber is assumed to be much bigger than the wavelength and thus can be treated together with the core and overlay as planar stack of layers. The fiber components and the external environment are also assumed to be ideally uniform and isotropic. Moreover, the model assumes Lambertian distribution of light source which does not ideally represent the propagation conditions for electromagnetic wave in the discussed structure. Regarding the input data it should be noted that, as it was explained in Section 2.5, the n and k values of the deposited materials are extracted from ellipsometric characterization performed
217
K. Kosiel et al. / Optics and Laser Technology 102 (2018) 213–221
Fig. 4. Evolution of the normalized measured transmission spectra with external RI for the 211 nm-thick-HfO2-coated (a) and for the 244 nm-thick-ZrO2-coated (b) fiber structures. The arrows indicate shift of the resonance minima for the increasing external RI. The external RI values (in RIU) are given in the legends of the figures.
Fig. 5. Evolution of the resonance wavelength induced by change of external RI for the fiber structures with different ZrO2 thickness. The overlay thicknesses were as follows: 244 nm, 224 nm and 216 nm for the samples ZrO2_1, ZrO2_2 and ZrO2_3, respectively. The Z1 and Z2 resonances correspond to spectra type shown in Fig. 4b.
using planar test structures. As it was already mentioned, some other differences are possible between optical properties of the planar layers and the cylindrical overlays. 3.3. The fiber structures coated by TaxOy/SixNy-double-layer deposited by PECVD and ALD Various overlay materials show generally different properties that can be useful for optical fiber sensor technology, e.g., they differ in terms of hardness, elasticity, resistivity to various chemicals, thermal resistance, facility for being functionalized at the surface and many others [42]. For this reason combining more than one
overlaying material as a component of the LMR-based sensor can be found useful, as it has a potential for introducing additional functionalities. The PECVD-deposited SixNy has an advantage of having the largest n value and the appropriate k profile that promotes induction of LMR effect, both features being the most promising within the group of dielectric materials tested here (see Section 3.1). As it was noticed earlier, ALD is generally the most appropriate technique for easy and precise tuning of n and thickness of overlays. Thanks to this, the optical properties of the sensor (e.g., its resonant wavelength) can be tuned according to the need – by careful adjustment of the optical thickness of the overlay. ALD also shows an advantage of low-temperature deposition availability of a wide range of materials, as shown in this work HfO2, ZrO2 or TaxOy. However, not every material can be low-temperature-ALDdeposited while retaining satisfactory properties that are necessary for given application. This is also a case of SixNy, which allows rather high deposition temperature (typically at least 200 °C) for satisfactorily high n value (over 2). However, the relatively low deposition rate of ALD generally results in very time consuming processes. A couple-of-hours-long high temperature ALD process for approx. 200 nm-thick SixNy deposition, would be detrimental for parts of thermally vulnerable optical fibers. Taking into account that one can expect higher sensitivity of the LMR-based device equipped with even thicker overlay, the ALD-only fabrication of such thicker SixNy overlay seems to be highly problematic. On the other hand the PECVD technique has an unquestionable advantage of being a relatively fast-deposition method for thin layers. Comparison of the thin layer deposition rates (nm/s) presented in this work shows that PECVD is a technique by two orders of magnitude faster than ALD. Additionally it is easy to deposit even a few micrometers thick layers by PECVD, while ALD deposition is rather limited to sub-micrometer thickness because of technical reasons and time consumption. However, the capability of PECVD
Table 1 Wavelength shifts and sensitivities of Z1 and Z2 resonant minima measured in different ranges of external RI, for the three fiber structures with different ZrO2 overlay thickness. The Z1 and Z2 resonances correspond to spectra shown in Fig. 4b. ZrO2 thickness (nm)
216 224 244
Z1 minimum wavelength shift (nm)/sensitivity (nm/RIU)
Z2 minimum wavelength shift (nm)/sensitivity (nm/RIU)
In the range of 1.33–1.35 RIU
In the range of 1.41–1.43 RIU
In the range of 1.33–1.35 RIU
In the range of 1.41–1.43 RIU
1.06/53 0.57/ 28.50 2.18/109.00
5.52/276 5.52/276 6.09/304.50
3.33/166.5 5.42/271 3.90/195.00
13.80/690 12.65/632.50 17.60/880.00
218
K. Kosiel et al. / Optics and Laser Technology 102 (2018) 213–221
Fig. 6. Evolution of the normalized measured transmission spectra with external RI for the 410 nm-thick-SixNy-coated fiber structure (a) and for the same structure additionally coated by 50 nm-thick TaxOy (b), for the 450 nm-thick-SixNy-coated fiber structure (c) and for the same structure additionally coated by 50 nm-thick TaxOy (d), for the 510 nm-thick-SixNy-coated fiber structure (e) and for the same structure additionally coated by 50 nm-thick TaxOy (f). The arrows indicate shift of the resonance minima for the increasing external RI. The external RI values (in RIU) are given in the figures legends.
relating to overlay uniformity or thickness accuracy is noticeably lower than for ALD. The amount of PECVD-deposited film on a sample depends mainly on effectiveness of plasma-enhanced decomposition of gases and not only on mass located in the reaction chamber, but also relatively strongly on sample shape, sample location in chamber (plasma), particle (ion) flux, material of the sample and parameters of the process, such as gas composition, pressure, power, and time [56]. Though, the uniformity and accu-
racy of the PECVD–deposited overlay can be increased by optimization of technological factors, for example by rotation of the fiber during the process [64], it will never reach the level of the ALD. As a way to solve the above described problems, here we propose an approach to the fiber sensor overlay deposition being a combination of these two methods, and giving an advantage of short-term deposition process of an overlay (thanks to using
K. Kosiel et al. / Optics and Laser Technology 102 (2018) 213–221
Fig. 7. Evolution of the resonance wavelength induced by change of external RI for the fiber structures coated by the 410 nm-thick SixNy (SixNy_1), 450 nm-thick SixNy (SixNy_2), 510 nm-thick SixNy (SixNy_3), and for the same structures additionally coated by 50 nm-thick TaxOy (SixNy_1 + TaxOy, SixNy_2 + TaxOy and SixNy_3 + TaxOy, respectively). The data were extracted from measurements presented in Fig. 6, for the S2 minima (for the SixNy-coated fibers) and ST2 minima (for the TaxOy/SixNycoated fibers) in each case.
Table 2 Wavelength shifts and sensitivities of S2 and ST2 resonant minima measured in different ranges of external RI, for the six fiber structures with different total thicknesses of the TaxOy/SixNy overlay, and different TaxOy thickness percentages in the bilayer overlay. The S2 and ST2 resonances correspond to spectra shown in Fig. 6. TaxOy/SixNy overlay thickness (nm)
410 450 460 500 510 560
TaxOy thickness %
0 0 10.9 10 0 8.9
Wavelength shift (nm)/sensitivity (nm/RIU) In the range of 1.33–1.35 RIU
In the range of 1.43–1.45 RIU
1.37/68.5 2.66/133 2.20/110 3.12/156 5.79/289.5 21.54/1077
8.08/404 10.96/548 10.12/506 9.67/483.5 11.87/593.5 9.66/483
PECVD) and possibility of fine-tuning the target overlay properties - mainly in terms of its thickness, but also in terms of n-profile engineering according to the need (thanks to using ALD). Fig. 6a–e present the collection of transmission spectra for three pairs of the fiber structures. There are the results for the fibers coated only by PECVD SixNy with various thickness (dSixNy was in the range of 410–510 nm) in Fig. 6a, c, e. There are also the results for the same structures coated additionally by 50 nm-thick ALD TaxOy in Fig. 6b, d, f; therefore, the total thickness of the tested overlays was in the range of 410–560 nm. The relatively hightemperature PECVD process of SixNy as a part of each doublelayer overlay took a few minutes. Then, the sensorial response of each SixNy-coated device was analyzed optically. Next the examined fibers were introduced to the ALD system in order to deposit the ALD-based part of the overlay. Then the low-temperature ALD processes of 50 nm-thick TaxOy took about 44 min. Such a technological scheme can be easily adopted for fabrication of overlays with various proportions of the PECVD-based and the ALD-based part of the overlay in order to optimize the technology, taking into account the potentially poorer accuracy of PECVD-based layer thickness and the much larger time-consumption of ALD used for tuning the sensor properties up to the designed level. Like in the case of overlays presented in Section 3.2 a redshift of resonant wavelengths is seen with the increasing (TaxOy)/SixNy overlay thickness. As in the case of earlier presented fiber
219
structures, the resonances in longer wavelength range (S2, ST2) are more sensitive than those in shorter wavelength range (S1, ST1). It is seen, that resonance wavelength shift induced by change of overlay properties in a collection of sensors related to Fig. 6 covers approx. 180 nm-broad range of wavelength - considering only the positions of the S2 minima (for the SixNy-coated structures) as well as corresponding ST2 minima (for the TaxOy/SixNy-coated structures). The S2 and ST2 positions are tuned from structure to structure by a few tens of nanometers. However, as it was noticed in Section 3.2 by the example of ZrO2 overlay (which thickness was altered by less than 10 nm), a way more precise tuning is possible by using ALD - it is known that ALD allows for thickness tuning even with atomic layer precision [42]. The evolution of the resonance wavelength induced by change of external RI for the fiber structures with the described above SixNy and TaxOy/SixNy overlays is presented in Fig. 7. There are wavelength shifts and sensitivities of S2 and ST2 resonant minima presented in Table 2. These are presented for different ranges of external RI (the two external RI exemplary ranges were selected: 1.33–1.35 and 1.43–1.45 RIU). They were calculated for the six fiber structures with different total thicknesses of the (TaxOy)/SixNy overlay, and different TaxOy thickness percentages in the bilayer overlay. It is seen that sensitivity (nm/RIU) of the SixNy-only-coated sensors increases with the overlay thickness and reaches considerably high values in higher external RI. This behavior is similar to that observed for discussed earlier ALD-coated fibers. Also for the fibers coated by TaxOy/SixNy double-layers the sensitivity in lower external RI range increases with the overlay thickness - it should be noted that in the series of fibers with increasing double-layer thickness the TaxOy thickness percentage slightly decreases. In higher external RI, the sensitivity even slightly decreases with TaxOy/SixNy double-layer thickness. The case of two structures is worth noticing: the fiber coated by 50 nm TaxOy/410 nm SixNy and the fiber coated by 450 nm SixNy. Although the total overlay thickness for the double-layer coated fiber is 10 nm larger, the structure shows lower sensitivity. According to these results it can be concluded that the impact of TaxOy component of the overlay is different in the two external RI ranges. In both ranges it does not improve the sensitivity of the structures but, as it was mentioned above, helps to adjust the wavelength range for device interrogation purposes. The increase of sensitivity is provided rather by increasing thickness of the SixNy component of the overlay. It should be noted, however, that the n value of TaxOy - which lies farther from the fiber core than the SixNy - is lower than the n of the latter (see Fig. 1). Therefore, n does not increase monotonically in radial direction. This can be the reason of the observed differences in the sensitivity. However, the theoretical results presented in Ref. [59], suggest that some double-layer combinations should provide enhancement of sensitivity in comparison to such overlay components used separately on the fiber structures.
4. Summary and conclusions Application of the ALD technique as a tool for tailoring sensorial properties of LMR-based optical fiber sensors was discussed. The ALD was used for deposition of high-refractive-index materials (at the wavelength of 650 nm the n value was in the range 2.05– 2.1 RIU) for overlays of devices fabricated from PCS multimode optical fibers with partly removed claddings. The ALD-deposited HfO2, ZrO2, and TaxOy were applied. Application of ALD for LMRbased sensors is superior to the other deposition techniques and gives potential for accurate adjustment of overlay thickness and its conformity, uniformity and tightness. At the same time it
220
K. Kosiel et al. / Optics and Laser Technology 102 (2018) 213–221
ensures wide variety of dielectric materials (and by this the flexibility of n⁄ tuning) and low process temperature (that ensures safe conditions for thermally vulnerable fibers). For this reason we found ALD a first-choice tool for careful modification of sensorial properties of LMR-based sensors in terms of adjustment of their resonant wavelength positions, by even sub-nanometer-level overlay thickness control. Consequently, we presented examples of LMR sensors which overlay fabrication was done exclusively by ALD. At the same time ALD can be used as an element in more complicated technological scheme comprising also other deposition methods for overlay formation. Namely, we presented LMR-based sensors which overlay fabrication combined application of PECVD and ALD techniques. Such a technological scheme allowed for taking the advantage of high deposition rate of PECVD for creation of dominant bottom part of the overlay (in terms of its thickness), and using the ALD technique for fine-tuning the sensorial properties by deposition relatively thin upper part of the overlay, i.e., by doing this with the ALD-related superior thickness control. At the same time this is the first-time demonstration of optical fiber sensor double-layer overlay, composed of two different dielectric materials. It should be noted here, that the LMR sensor technology seems to be cheaper than the SPR-based one, because of relatively high cost of overlay materials used in the latter - the noble metals, the most often gold. While the ALD technique for dielectrics used in LMR sensors fabrication can still be found relatively expensive due to costs of chemical precursors and time consumption, thanks to the proposed approach where the two techniques, namely PECVD and ALD were combined, the LMR sensors technology can be cost-effective. Double-layer overlays can also provide different functionalities, such as, e.g., application of chemically resistive films as the most external part of a fiber structure for ensuring the chemical stability of the overall sensor.
Acknowledgments The authors gratefully acknowledge support for this work from the National Science Centre of Poland under grant No. 2014/14/E/ ST7/00104. The work was also financially supported as statutory activities at the Instytut Technologii Elektronowej. Kamil Kosiel is grateful to Dr Anna Szerling for her help during the paper preparation.
References [1] N. Sabri, S.A. Aljunid, M.S. Salim, R.B. Ahmad, R. Kamaruddin, Toward optical sensors: review and applications, J. Phys.: Conf. Series, 423 (2013) 012064. [2]
[10] Feng De-Jun, Liu Guan-Xiu, Liu Xi-Lu, Jiang Ming-Shun, Sui Qing-Mei, Refractive index sensor based on plastic optical fiber with tapered structure, Appl. Optics 53 (10) (2014) 2007–2011. [11] M. Smietana, D. Brabant, W.J. Bock, P. Mikulic, T. Eftimov, Refractive-index sensing with inline core-cladding intermodal interferometer based on silicon nitride nano-coated photonic crystal fiber, J. Lightwave Technol. 30 (8) (2012) 1185. [12] M.F. Jaddoa, A.A. Jasim, M.Z.A. Razak, S.W. Harun, H. Ahmad, Highly responsive NaCl detector based on inline microfiber Mach-Zehnder interferometer, Sens. Actuat. A 237 (2016) 56–61. [13] M. Smietana, W.J. Bock, P. Mikulic, A. Ng, R. Chinnappan, M. Zourob, Detection of bacteria using bacteriophages as recognition elements immobilized on long-period fiber gratings, Optics Express 19 (9) (2011) 7971. [14] A. Leung, P.M. Shankar, R. Mutharasan, A review of fiber-optic biosensors, Sens. Actuat. B 125 (2007) 688–703. [15] Ch. Caucheteur, V. Malachovska, C. Ribaut, R. Wattiez, Cell sensing with nearinfrared plasmonic optical fiber sensors, Optics Laser Technol. 78 (2016) 116– 121. [16] I.M. Ishaq, A. Quintela, S.W. James, G.J. Ashwell, J.M. Lopez-Higuera, R.P. Tatam, Modification of the refractive index response of long period gratings using thin film overlays, Sens. Actuat. B: Chem. 107 (2) (2005) 738–741. [17] D. Leamy, F. Regan, A robust spin-coated optical sensing device for rapid determination of predominant metal ions in wastewater streams, Int. J. Environ. Anal. Chem. 83 (10) (2003) 867–877. [18] Š.M. Korent, A. Lobnik, G.J. Mohr, Sol-gel-based optical sensor for the detection of aqueous amines, Anal. Bioanal. Chem. 387 (8) (2007) 2863–2870. [19] L.Y. Shao, M.J. Yin, H.Y. Tam, J. Albert, Fiber optic pH sensor with selfassembled polymer multilayer nanocoatings, Sensors 13 (2) (2013) 1425– 1434. [20] R. Bogdanowicz, M. S´mietana, M. Gnyba, Ł. Gołunski, J. Ryl, M. Garda, Optical and structural properties of polycrystalline CVD diamond films grown on fused silica optical fibres pre-treated by high-power sonication seeding, Appl. Phys. A 116 (4) (2014) 1927–1937. [21] W. Xie, M. Yang, Y. Cheng, D. Li, Y. Zhang, Z. Zhuang, Optical fiber relativehumidity sensor with evaporated dielectric coatings on fiber end-face, Optical Fiber Technol. 20 (4) (2014) 314–319. [22] K. Schmitt, K. Oehse, G. Sulz, C. Hoffmann, Evanescent field sensors based on tantalum pentoxide waveguides–a review, Sensors 8, 2008, 711–738. [23] R.W. Johnson, A. Hultqvist, S.F. Bent, A brief review of atomic layer deposition: from fundamentals to applications, Mater. Today 17 (5) (2014) 236–246. [24] Ch. Das, M. Kot, K. Henkel, D. Schmeisser, Engineering of sub-nanometer SiOx thickness in Si photocathodes for optimized open circuit potential, ChemSusChem 9 (2016) 2332–2336. [25] S.E. Potts, W. Keuning, E. Langereis, G. Dingemans, M.C.M. van de Sanden, W.M. M. Kessels, Low temperature plasma-enhanced atomic layer deposition of metal oxide thin films, J. Electrochem. Soc. 157 (7) (2010) 66–74. [26] M. Sowinska, K. Henkel, D. Schmeißer, I. Karkkanen, J. Schneidewind, F. Naumann, B. Gruska, H. Gargour, Analysis of nitrogen species in titanium oxynitride ALD films, J. Vacuum Sci. Technol. A: Vacuum Surf. Films 34 (1) (2016) 01A127. [27] M. Sowinska, S. Brizzi, Ch. Das, I. Kärkkänen, J. Schneidewind, F. Naumann, H. Gargouri, K. Henkel, D. Schmeißer, Analysis of nitrogen species in titanium oxynitride ALD films, Appl. Surf. Sci. 381 (2016) 42–47. [28] H. Kim, Atomic layer deposition of metal and nitride thin films: current research efforts and applications for semiconductor device processing, J. Vacuum Sci. Technol. B 21 (2003) 2231. [29] M.B.E. Griffiths, P.J. Pallister, D.J. Mandia, S.T. Barry, Atomic layer deposition of gold metal, Chem. Mater. 28 (1) (2016) 44–46. [30] Z. Guo, H. Li, Q. Chen, L. Sang, L. Yang, Z. Liu, X. Wang, Low-temperature atomic layer deposition of high purity, smooth, low resistivity copper films by using amidinate precursor and hydrogen plasma, Chem. Mater. 27 (2015) 5988– 5996. [31] M. Kot, C. Das, Z. Wang, K. Henkel, Z. Rouissi, K. Wojciechowski, H.J. Snaith, D. Schmeisser, ChemSusChem 9 (2016) 3401. [32] I.J. Raaijmakers, Current and future applications of ALD in micro-electronics, ECS Trans. 41 (2) (2011) 3–17. [33] H. Kim, Atomic layer deposition of metal and nitride thin films: current research efforts and applications for semiconductor device processing, J. Vacuum Sci. Technol. B: Microelectron. Nanometer Struct. Process. Measure. Phenomena 21 (6) (2003) 2231–2261. [34] H. Im, N.J. Wittenberg, N.C. Lindquist, S.H. Oh, Atomic layer deposition (ALD): a versatile technique for plasmonics and nanobiotechnology, J. Mater. Res. 27 (4) (2012) 663–671. [35] M. Godlewski, S. Gierałtowska, Ł. Wachnicki, R. Pietuszka, B.S. Witkowski, A. Słon´ska, Z. Gajewski, M.M. Godlewski, High-k oxides by atomic layer deposition-applications in biology and medicine, J. Vacuum Sci. Technol. A: Vacuum Surf. Films 35 (2) (2017) 021508. [36] M. Dominik, J. Niedziółka-Jönsson, E. Roz´niecka, Ł. Wachnicki, M. Godlewski, P. Mikulic, W.J. Bock, M. S´mietana, Regeneration of titanium oxide nano-coated long-period grating biosensor, in: Sixth European Workshop on Optical Fibre Sensors (Vol. 9916, p. 99160Z) International Society for Optics and Photonics, 2016. [37] M. Dominik, A. Les´niewski, M. Janczuk, J. Niedziółka-Jönsson, M. Hołdyn´ski, Ł. Wachnicki, M. Godlewski, W.J. Bock, M. S´mietana, Titanium oxide thin films obtained with physical and chemical vapour deposition methods for optical biosensing purposes, Biosens. Bioelectron. 93 (2017) 102–109.
K. Kosiel et al. / Optics and Laser Technology 102 (2018) 213–221 [38] M. S´mietana, M. Mys´liwiec, P. Mikulic, B.S. Witkowski, W.J. Bock, Capability for fine tuning of the refractive index sensing properties of long-period gratings by atomic layer deposited Al2O3 overlays, Sensors 13(12) (2013) 16372– 16383. [39] Y. Zhao, F. Pang, Y. Dong, J. Wen, Z. Chen, T. Wang, Refractive index sensitivity enhancement of optical fiber cladding mode by depositing nanofilm via ALD technology, Optics Express 21 (22) (2013) 26136–26143. [40] A. Purniawan, G. Pandraud, T.S.Y. Moh, A. Marthen, K.A. Vakalopoulos, P.J. French, P.M. Sarro, Fabrication and optical measurements of a TiO 2-ALD evanescent waveguide sensor, Sens. Actuat. A: Phys. 188 (2012) 127–132. [41] M. S´mietana, M. Koba, P. Mikulic, W.J. Bock, Towards refractive index sensitivity of long-period gratings at level of tens of mm per refractive index unit: fiber cladding etching and nano-coating deposition, Optics Express 24 (11) (2016) 11897–11904. [42] K. Kosiel, M. Dominik, I. S´cis´lewska, M. Kalisz, M. Guziewicz, K. Gołaszewska, J. Niedziółka-Jonsson, W.J. Bock, M. S´mietana, Alkali-resistant low-temperature atomic-layer-deposited oxides for optical fiber sensor overlays, Nanotechnology (2018) (submitted for publication). [43] I. Del Villar, F.J. Arregui, C.R. Zamarreño, J.M. Corres, C. Bariain, J. Goicoechea, C. Elosua, M. Hernaez, P.J. Rivero, A.B. Socorro, A. Urrutia, P. Sanchez, P. Zubiate, D. Lopez, N. De Acha, J. Ascorbe, I.R. Matias, Optical sensors based on lossy mode resonances, Sens. Actuat. B 240 (2017) 174–185. [44] P. Zubiate, C.R. Zamarreño, P. Sánchez, I.R. Matias, F.J. Arregui, High sensitive and selective C-reactive protein detection by means of lossy mode resonance based optical fiber devices, Biosens. Bioelectron. 93 (2017) 176–181. [45] S. Zhu, F. Pang, S. Huang, F. Zou, Y. Dong, T. Wang, High sensitivity refractive index sensor based on adiabatic tapered optical fiber deposited with nanofilm by ALD, Optics Express 23 (11) (2015) 13880–13888. [46] M. Hernáez, I. Del Villar, C.R. Zamarreño, F.J. Arregui, I.R. Matias, Optical fiber refractometers based on lossy mode resonances supported by TiO2 coatings, Appl. Optics 49 (20) (2010) 3980–3985. [47] S. Zhu, F. Pang, S. Huang, F. Zou, Q. Guo, J. Wen, T. Wang, High sensitivity refractometer based on TiO2-coated adiabatic tapered optical fiber via ALD technology, Sensors 16(8) (2016) 1295. [48] P. Sanchez, C.R. Zamarreno, C.R. Zamarreo, M. Hernaez, I.R. Matias, F.J. Arregui, Exhaled breath optical fiber sensor based on LMRs for respiration monitoring, in: SENSORS, 2014 IEEE (2014) 1142–1145. [49] I. Del Villar, C.R. Zamarreño, P. Sanchez, M. Hernaez, C.F. Valdivielso, F.J. Arregui, I.R. Matias, Generation of lossy mode resonances by deposition of high-refractive-index coatings on uncladded multimode optical fibers, J. Optics 12 (9) (2010) 095503. [50] I. Del Villar, C.R. Zamarreño, M. Hernaez, F.J. Arregui, I.R. Matias, Lossy mode resonance generation with indium-tin-oxide-coated optical fibers for sensing applications, J. Lightwave Technol. 28 (1) (2010) 111–117.
221
[51] C.R. Zamarreño, M. Hernaez, I. Del Villar, I.R. Matias, F.J. Arregui, Tunable humidity sensor based on ITO-coated optical fiber, Sens. Actuat. B: Chem. 146 (1) (2010) 414–417. [52] D. Kaur, V.K. Sharma, A. Kapoor, High sensitivity lossy mode resonance sensors, Sens. Actuat. B: Chem. 198 (2014) 366–376. [53] C.R. Zamarreño, S. Lopez, M. Hernaez, I. Del Villar, I.R. Matias, F.J. Arregui, Resonance-based refractometric response of cladding-removed optical fibers with sputtered indium tin oxide coatings, Sens. Actuat. B: Chem. 175 (2012) 106–110. [54] P. Zubiate, C.R. Zamarreño, I. Del Villar, I.R. Matias, F.J. Arregui, High sensitive refractometers based on lossy mode resonances (LMRs) supported by ITO coated D-shaped optical fibers, Optics Express 23 (6) (2015) 8045–8050. [55] Indium CorporationÒ Application Note, Indium Oxide and Indium-Tin Oxide (ITO) Coatings.