D. Sofikitis, K. Stamataki, M. A. Everest, V. Papadakis, J. Stehle, B. Loppinet, and T. P. Rakitzis, Sensitivity enhancement for evanescent-wave sensing using cavity-ring-down ellipsometry, Opt. Lett. 38, 1224-1226 (2013)

Please download to get full document.

View again

of 3
0 views
PDF
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Document Description
We demonstrate a method to increase the sensitivity of the s-p phase shift under total internal reflection (TIR) for optical sensing. This is achieved by the introduction of two simple dielectric layers to the TIR surface of a fused silica prism. The
Document Share
Document Transcript
  Sensitivity enhancement for evanescent-wave sensingusing cavity-ring-down ellipsometry Dimitris Sofikitis, 1,2 Katerina Stamataki, 1,3 Michael A. Everest, 4 Vassilis Papadakis, 1 Jean-Louis Stehle, 5 Benoit Loppinet, 1,6 and T. Peter Rakitzis 1,2,7 1 Institute of Electronic Structure and Laser, Foundation for Research and Technology-Hellas, Heraklion-Crete 71110, Greece 2 Department of Physics, University of Crete, Heraklion-Crete 71003, Greece 3 Department of Chemistry, University of Crete, Heraklion-Crete 71003, Greece 4 Department of Chemistry, Westmont College, Santa Barbara, California 93108, USA 5 Sopralab, 7 rue du Moulin des Bruyeres, Courbevoie 92400, France 6  e-mail: benoit@iesl.forth.gr  7  e-mail: ptr@iesl.forth.gr  Received February 6, 2013; revised March 5, 2013; accepted March 7, 2013;posted March 8, 2013 (Doc. ID 185029); published April 1, 2013We demonstrate a method to increase the sensitivity of the  s -  p  phase shift under total internal reflection (TIR) for optical sensing. This is achieved by the introduction of two simple dielectric layers to the TIR surface of a fusedsilica prism. The enhanced sensitivity is demonstrated using evanescent-wave cavity-ring-down-ellipsometry bymeasuring the refractive index of liquid mixtures and by studying the adsorption of polymers to the TIR surfaceof the fused silica prism. © 2013 Optical Society of America OCIS codes:  120.2130, 280.4788, 240.2130, 260.6970. Total internal reflection offers a practical configuration for optical sensing applications, which have been exploredover the last 40 years for a variety of techniques, suchas surface plasmon resonances (SPRs), optical waveguidespectroscopy (OWS), ellipsometry, and others [1 – 5]. Theobservables in these experiments are modulations ineither the light intensity or in the phase of the reflectedlight, which is often measured through the polarizationdependence of the total internal reflection (TIR)-induced phase shift. Recently an approach for the measurement of the phase shift between the  s  and  p  polarization compo-nents,  Δ  δ   p  − δ  s , was demonstrated using evanescent-wave cavity-ring-down ellipsometry (EW-CRDE) [6 – 8].However, evanescent-wave ellipsometry suffers fromlimited sensitivity. The  s -  p  phase shift  Δ  upon TIR(  R >  0 . 995 ) in a simple Fresnel interface is not verysensitive to changes on the low refractive index side, sincethe  s  and  p  waves are affected in similar ways. Severalmethods have been proposed to enhance internal reflec-tion sensitivity [9,10]. Here we propose a way to improve itbyintroducingathinlayersystemattheTIRsurface.Theenhanced sensitivity is demonstrated by measurementsof the refractive indices of liquid mixtures and adsorptionof polymers at the prism TIR surface. A thin layer of a high refractive index material (TiO 2 )is added to the TIR surface. This layer acts as a partial polarization-dependent beam splitter, and results in a de-coupling of the  s  and  p  light. Finally, a thin SiO 2  layer isadded on top. This simple design increases the phase-shiftsensitivitywhilepreserving the propertiesofthe substrate.Figure 1(a) displays the phase shifts for   s  and  p  polarizedlight calculated using the transfer matrix [11] for the sim- ple fused-silica/water Fresnel interface and for the layeredsystem, as a function of the incidence angle  θ   under TIRconditions. For the calculations, the refractive index of the fused-silica substrate was set to 1.46, the TiO 2  high-refractive index layer had  n ≈ 2 . 5  with thickness of 50 nm,and the top layer of SiO 2  had  n  1 . 46  with thickness20 nm. The phase shifts of the  s  and  p  waves are very sim-ilar for the simple fused silica/water Fresnel interface.However, they are very different in the layered system,for which  δ  s  is almost independent of   θ  ,  δ   p  varies stronglywith  θ  , and thus  Δ  varies much more compared to thatof the simple Fresnel interface [see Fig. 1(a)]. This leadsto the increased sensitivity discussed bellow.The phase shift  Δ  was measured with the EW-CRDEtechnique. A laser pulse linearly polarized at 45° entersa cavity consisting of two highly reflecting mirrorsand a prism in which light undergoes TIR, shown in Fig. 1. (Color online) (a) Dependence of   δ   p  (red dashed lines), δ  s  (blue dotted lines), and Δ  δ   p  − δ  s  (black solid lines) on theangle of incidence for the case of the simple fused silica/water Fresnel interface (left) and for the layered system (right).(b) Experimental setup (see text for details). (c) EW-CRDEtraces acquired via the uncoated (top) and coated (bottom) partof the TIR prism surface, in contact with water.1224 OPTICS LETTERS / Vol. 38, No. 8 / April 15, 20130146-9592/13/081224-03$15.00/0 © 2013 Optical Society of America   Fig. 1(b). A linear polarizer oriented at 45° to the planeof incidence is placed at the output of the cavity. Thefinal output (ideally) has the simple form  I   t    Ae − t ∕ τ  Cos 2  ω t ∕ 2   [6], where  A  is a normalization constant, τ   is the photon lifetime in the cavity, and  ω  is the polarization beating frequency. The ring-down time  τ   isdetermined from an exponential fit, and  ω  from the fastFourier transform of the data. The beating frequency  ω  isrelated to  Δ , the single pass  s -  p  phase shift as  ω    c Δ ∕ d ,where  d  is the cavity length and  c  the speed of light. Weuse a microchip Nd:YAG laser (Horus HLX-V-F-100),which delivers 0.6 ns pulses of 0.6  μ  J at 532 nm at80 kHz repetition rate. The near concentric cavity,formed by two mirrors (  R    0 . 996 ) with radius of curvature 10 cm placed at  d    20  cm, is chosen to belonger than the laser pulse length. Fused silica prismswere custom made (Laser 2000) with only half of their TIR surface coated with the TiO 2  (50 nm nominalthickness) and SiO 2  (20 nm nominal thickness) layers.This allowed us to switch the TIR spot from the coatedto the uncoated region by translating the prism only a fewmm, which did not significantly degrade the cavity align-ment. The prism (with dimensions  7 . 5  ×  2 . 5  ×  2 . 5  cm) is placed at the center of the cavity, and the partialreflections from the input and output surfaces, whichare not antireflection coated, are kept in the cavity.The incidence angle of 70° satisfies the TIR conditionfor a range of sample refractive indices between 1 and1.365, which includes water (  n    1 . 33 ) and most proteinand polymers aqueous solutions.The prism has been characterized by spectroscopicellipsometry (GES5E Semilab) both in the internal andin the external reflection configuration, which revealedbirefringence as high as 20°, depending on the positionof the ray in the prism. In order to assess the role of  prism birefringence in our measurements, EW-CRDE wassimulated by the Jones matrix method. It showed that prism birefringence adds to the final result of   Δ , thus preventing the measurement of the absolute values of the TIR  s -  p  phase shift, but did not affect the dependenceof the measured  Δ  on external refractive index. Therefore,relative measurements are expected to be accurate. A glass flow cell with a gap of 0.1 mm (Hellma) was placed on the TIR surface and connected to syringes toallow introduction of liquids. Two characteristic CRDEtraces, acquired via the uncoated and the coated part of the prism, with pure water inserted in the flow cell areshown in Fig. 1(c). The polarization beating frequency ω ∕ 2 π   is  ∼ 50  MHz and  ∼ 300  MHz for the uncoated andcoated surface, respectively; however, there is no signifi-cant difference in the ring-down time.The measured phase shift  Δ  for various water/ethanoland water/methanol mixtures are shown in Fig. 2, asa function of the refractive index measured by an Abberefractometer, for both the uncoated (black circles) andthe coated (red squares) part of the prism. The depend-ence of   Δ  on refractive index is significantly increased bythe coating with a change of more than 35° over thecovered range of refractive index, compared to less than10° for the uncoated prism. The black line in Fig. 2 cor-responds to the  Δ  value expected for the simple fused-silica/water interface, with the addition of a constant value introduced to account for prism birefringence.The red dashed line is the expected dependence of   Δ on the refractive index when the nominal thicknessesand refractive indices are used. A better fit is obtainedwhen a smaller thickness of 30 nm of the TiO 2  layer isassumed, as shown by the solid red line.The difference in resolution between  Δ  measurementsusing the coated and uncoated prism is illustrated in theinset of Fig. 2, showing two typical measurements of   Δ over a period of 100 s for the uncoated (left) and coated(right) part of the prism in contact with water. A longtimefluctuation of characteristic time  ∼ 15  s was apparentin the phase measurements, independent on the natureof the TIR surface (coated/uncoated). The solid lines de-note the change of   Δ  that corresponds to a change of  2  ×  10 − 4 refractive index units (RIUs). This value appearstobewithinthedetectionlimitforthecaseofthedielectriclayers, while it is not resolvable in the absence of the di-electric layers. The predicted sensitivity of   Δ  ( d Δ ∕ dn ) for the coated region was calculated to be a factor of   ∼ 4  timeslarger than the uncoated region. However, the enhancedsensitivity resulted in an amplification of the longtimefluctuation, therefore reducing the net gain in resolutionand determining the detection limit  δ  Δ min  of the measure-ment. This fluctuation could relate to the sample condi-tions such as temperature fluctuations and mechanicalinstabilities.The resolution improvement is found to be signifi-cantly better when we consider interface phenomena sensing, like adsorption, rather than bulk refractometry( ∼ 10 − 4 RIU). The resolution  δ  Δ min , including the long-time fluctuation, was found in the absence of dielectriccoatings to be on the order of   δ  Δ min    0 . 034 °. Such a change in  Δ  would correspond to the formation of a 1.2 nm layer with refractive index  n    1 . 4  at the water/ fused silica interface, i.e., an adsorbed amount on theorder of   120  ng ∕ cm 2 for standard organic materials. Inthe presence of the dielectric coatings, the resolutionwas found to be  δ  Δ min    0 . 07 °, corresponding to a thick-ness of 0.21 nm, or   21  ng ∕ cm 2 . The total resolution toadsorption sensing was therefore found to be improved ∼ 5 . 5  times. Fig. 2. (Color online) Variation of phase shift  Δ  as a functionof refractive index of liquid solutions, for the uncoated(black points) and coated (red points) part of the prism surface.Black and red lines correspond to simulations for the twosituations, respectively. Statistical error bars are of order   10 − 2 (not shown here). April 15, 2013 / Vol. 38, No. 8 / OPTICS LETTERS 1225  The increased sensitivity allows the monitoring of the creation of multilayers with thicknesses of less thana nanometer. We demonstrated the sensitivity by meas-uring  in situ  the formation of polyelectrolyte multilayers produced by the standard layer-by-layer depositiontechnique [12], the successive introduction of dilute positively/negatively charged poly-electrolyte solutionsand water to the TIR face of the prism. Sodium polysty-rene sulfonate (PSS) and polydiallyldimethyl-ammoniumchloride (PDDA) solutions at 1% by weight in 1M NaClsolution, were used. Sequences of water/PDDA/water/ PSS were repeated where each liquid was left  ∼ 50  s incontact with the interface.  Δ  was measured in real timewith the EW-CRDE setup and results are displayed inFig. 3.  Δ  was found to decrease with increasing layer formation. Individual steps correspond to the change of refractive index between the different solutions. Thechange of   Δ  after each sequence was attributed to the variation of the thickness of the composite PDDA-PSSlayer. Thickness was best estimated when in contact withwater. The thickness of each layer could be estimatedusing a sensitivity of   0 . 32 ° ∕ nm calculated assuming the30 nm TiO 2  layer calibration obtained from the refrac-tometry study. The step in Δ between two brine solutionsis typically of the order of 0.11°, corresponding to a thick-ness of 0.34 nm (or   34  ng ∕ cm 2 ). We see that the formationof every individual layer could be resolved with the ex-ception of the first PDDA layer. Notably the step in  Δ is approximately the same for each layer, showing a linear dependence of   Δ  on layer number, as expectedfor the particular system [12].The inset of Fig. 3 shows the time-dependence of  Δ for one sequence. Given the subsecond time resolution of the technique, the observed changes of   Δ  on timescalesof few seconds were attributed to the actual evolution of the sample. Additionally, as shown in the lower part of the figure, the simultaneously recorded ring-down timewas found to be insensitive to this multilayer deposition process, as expected for transparent dielectric materials.In conclusion, we have demonstrated a significantenhancement in the sensitivity of   s -  p  phase shift uponTIR, accomplished with a simple dielectric coating. Our sensing resolution is increased by more than a factor of 5 to a level of approximately  20  ng ∕ cm 2 , on the secondtimescale, which is convenient for monitoring mostadsorption processes. Established techniques like SPRand OWS achieve detection limits lower than  1  ng ∕ cm 2 ,down to  0 . 1  ng ∕ cm 2 [13] or   10 − 6 RIU for SPR and silicon prism TIR measurements [9]. It is worth noting that useof resonant dielectric layer structures should lead tosensitivities up to  ∼ ng ∕ cm 2 , imposing, however, extra constraints on the probe wavelength and/or alignment.The enhancement achieved allows using the simple TIRgeometry to reach the area of ng ∕ cm 2 in a subsecondtime scale, which is attractive for sensing applications.In principle, the CRDE technique can detect phase shiftson the microsecond timescale. In our current setup theacquisition time is limited by the repetition rate of our data acquisition system (digital oscilloscope). However,in future implementations, a fast FPGA based detectionscheme will be limited only by the laser  ’ s 80 kHz repeti-tion rate, permitting either an increase in sensitivity byreducing the noise via accumulation, or operation withthe current sensitivity on sub millisecond timescales.We thank the EU for partial support through the Euro- pean Research Council grant TRICEPS (GA no. 207542),and the FP7 IAPP SOFORT (PIAP-GA-2009-251598) andESMI-Infrastructure CS&CSA-2010-262348. References 1. E. Kretschmann, Zeitschrift fur physik  241 , 313 (1971).2. M. Maisonneuve, I.-H. Song, S. Patskovsky, and M. Meunier,Opt. Express  19 , 7410 (2011).3. J. D. Swalen, J. Phys. Chem.  83 , 1438 (1979).4. W. Chen, L. J. Martinezmiranda, H. Hsiung, and Y. R. Shen,Phys. Rev. Lett.  62 , 1860 (1989).5. R. Konradi, M. Textor, and E. Reimhult, Biosensors  2 , 341(2012).6. A. Karaiskou, V. Papadakis, B. Loppinet, and T. P. Rakitzis, J. Chem. Phys.  131 , 121101 (2009).7. M. A. Everest, V. M. Papadakis, K. Stamataki, S. Tzortzakis,B. Loppinet, and T. P. Rakitzis, J. Phys. Chem. Lett.  2 , 1324(2011).8. K. Stamataki, V. Papadakis, M. A. Everest, S. Tzortzakis, B.Loppinet, and T. P. Rakitzis, Appl. Opt.  52 , 1086 (2013).9. S. Patskovsky, I. H. Song, M. Meunier, and A. V. Kabashin,Opt. Express  17 , 20847 (2009).10. S. Otsuki and M. Ishikawa, Opt. Lett.  35 , 24 (2010).11. R. M. A. Azzam and N. M. Bashara,  Ellipsometry and Polarized Light  (North-Holland, 1987).12. G. Decher and J. Schmitt,  Progress in Colloid Polymer  Science  (Steinkopff, 1992), Vol  89 , p. 160.13. P. Kozma, A. Hámori, S. Kurunczi, K. Cottier, and R.Horvath, Sens. Actuators B  155  446 (2011).Fig. 3. (Color online) Study of multilayer creation on thecoated prism. (Top) Phase shift  Δ  modification caused bysuccessively alternating the electrolyte and the polymer solu-tion in the flow cell. The red line is a simulation of the average phase shift modification due to the multilayer creation. (Bot-tom) Corresponding ring-down times have been simultaneouslyrecorded, showing no significant changes.1226 OPTICS LETTERS / Vol. 38, No. 8 / April 15, 2013
Search Related
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks