Self-aligned setup for laser optical feedback imaging insensitive to parasitic optical feedback

Self-aligned setup for laser optical feedback imaging insensitive to parasitic optical feedback
of 5
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.
Related Documents
  Self-aligned setup for laser optical feedback imaginginsensitive to parasitic optical feedback  Olivier Jacquin,* Samuel Heidmann, Eric Lacot, and Olivier Hugon Laboratoire de Spectrométrie Physique, Université Joseph Fourier de Grenoble,UMR CNRS 5588, B.P. 87, 38402 Saint martin d ’ Hères Cedex, France*Corresponding author: ojacquin@ujf ‑ Received 15 July 2008; revised 12 November 2008; accepted 14 November 2008;posted 17 November 2008 (Doc. ID 98716); published 17 December 2008 We propose a new optical architecture for the laser optical feedback imaging (LOFI) technique whichmakes it possible to avoid the adverse effect of the optical parasitic backscattering introduced by allthe optical interfaces located between the laser source and the studied object. This proposed setup needsno specific or complex alignment, which is why we can consider the proposed setup to be self-aligned. Wedescribe the principle used to avoid the parasitic backscattering contributions that dramatically dete-riorate amplitude and phase information contained in the LOFI images. Finally, we give a successfuldemonstration of amplitude and phase images obtained with this self-aligned setup in the presenceof a parasitic reflection. © 2009 Optical Society of America OCIS codes:  110.3175, 110.2970, 110.3080, 110.4280, 280.3420. 1. Laser Optical Feedback Imaging Technique The laser optical feedback imaging (LOFI) techniqueis a sensitive imaging method combining optical het-erodyne interferometry with the dynamic propertiesof class B lasers [1]. In this method, the interferencetakes place within the laser, between the intracavitylight and the backscattered light by the studied tar-get. The backscattered light is frequency shifted tocreate an intracavity optical beating. The laser out-put power is then modulated at the shift frequency  Ω .If the shift frequency  Ω  is resonant with laser relaxa-tion frequency  Ω  R , a great optical amplification of theoptical beating contrast can be obtained. The detec-tion of this modulation with a lock-in amplifiermakes it possible to realize simultaneously ampli-tude (i.e., reflectivity) images and phase (i.e., profilo-metry) images of noncooperative targets [2]. Forexample, with a Nd:YAG microchip laser, the ampli-fication is of the order of   10 6 , which makes it possibleto easily measure reflectivity as low as  10 − 13 with alaser output power of a few milliwatts and with abandwidth detection of   1 kHz [3].IntheLOFI technique,thelaserandthe targetareconjugated via the optics of the system and the back-scattered photons come back into the laser cavity ac-cording to the reverse path principle. The system isthen self-aligned since the laser is used as bothsource and detector (Fig. 1). Consequently, the opti-cal system needs no complex alignment, which is an-other great advantage of the LOFI technique.Figure 1 shows a description of the LOFI experi-mental setup. The laser is a cw Nd 3 þ :  YAG micro-chip, lasing at wavelength  λ  ¼  1064 nm with anoutput power of   1 mW and a relaxation frequency of  Ω  R  ¼  700 kHz. A two-axis galvanometric mirrorscanner makes it possible to move the laser beamon the surface of the studied target to build an imagepoint by point. The frequency shift is obtained withtwo acousto-optic deflectors (AODs) operating, re-spectively, at  81 : 5 MHz (order  þ 1 ) and  81 : 5 MHz- Ω = 2 (order  − 1 ). The frequency shift is equal to  Ω = 2  whenthe light passes through the shifter. After a roundtrip, the total frequency shift of the reinjected lightinto the laser is thus equal to  Ω . 2. Parasitic Optical Feedback  The LOFI method is extremely sensitive to all-optical feedback. Consequently the method is 0003-6935/09/010064-05$15.00/0 © 2009 Optical Society of America 64 APPLIED OPTICS / Vol. 48, No. 1 / 1 January 2009  sensitive to optical parasitic backscattering, which isinherent in all-optical systems [4,5] and which de- pends on the quality of the optical elements. A signif-icant optical parasitic backscattering generated byan optical element located between the frequencyshifter and the studied target dramatically limitsthe LOFI performances. Indeed, we intuitively un-derstandthatitmaybedifficulttodetectareinjectedtarget signal lower than one reinjected by  “ the para-sitic object. ”  The sensitivity of the LOFI technique isthen strongly limited. For a  “ parasitic ”  diffusing ob- ject with an effective reflectivity  r  P  and for a targetwithaneffectivereflectivity r t  locatedatdistances d  P and  d t  from the laser (Fig. 1), the expressions of am-plitude and phase extracted by the lock-in amplifierare [6]  R  ¼  G LOFI  ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi r 2 t  þ r 2  P  þ  2 r t r  P cos ð ϕ t  − ϕ  p Þ q   P out ;  ð 1 Þ ϕ  ¼  a tan  r t sin ð ϕ t Þ þ r  P sin ð ϕ  p Þ r t cos ð ϕ t Þ þ r  P cos ð ϕ  p Þ  ;  ð 2 Þ with  ϕ  p  ¼  2 π  λ  2 d  P  and  ϕ t  ¼  2 π  λ  2 d t .  P out  is the output la-ser power and  G LOFI  is the optical amplification of the optical beating contrast [2].The previous equations show that, for a significantparasitic reflection ( r  P  ≈  r t ), the amplitude ( r t ) andphase ( ϕ t ) image of the target may be inaccessibleas we have shown in [7]. To illustrate this effect,in Fig. 2 we give the phase ( ϕ ) and amplitude (  R )of images obtained with a controlled parasitic reflec-tion in the experimental setup (Fig. 1). The parasiticoptical feedback is generated by a microscope slideplaced between the target and the frequency shifter.The orientation of the slide is accurately adjusted toobtain  r  P  ≈  r t . For the amplitude images, the studiedobjectisapieceofmetallicruler;forthephaseimage,it is the letter  “  e ”  etched on a silicon plate (the typicaldimension of the letter is about  100  μ m). The phaseimages given in Fig. 2 have been unwrapped using the maximum-likelihood binary-tree (MLBT) meth-od [8].Figures 2(a) and 2(b) show, respectively, amplitude and phase images obtained without parasitic back-scatter. We can see in Figs. 2(c) and 2(d) the effect of parasitic reflection on these images. The phaseand amplitude information is completely scrambled.Indeed,in the amplitude image [Fig.2(c)] we no long-er recognize the pattern of the metallic ruler, and inthe phase image [Fig. 2(d)] we no longer distinguishthe profile of the etched letter. The rectangular pat-terns in Fig. 2 are typical patterns of the MLTBmethod when the phase information phase cannotbe unwrapped.The use of antireflective (AR) coated optics [9] atthe working wavelength in the optical device makesit possible to minimize this critical effect introducedbytheparasiticbackscatter.However,problemsarisewhen we wish to satisfy the inequality  r  p  ≪  r t  for  r t reflectivity as low as  10 − 13 . For example, if the LOFItechnique is coupled with a microscope to realize bio-logical images [10], it is difficult to find microscopeobjectives with AR coating at the wavelength of  1 : 064  μ m, and very expensive or not reasonable touse AR coated microscope slides. The surface of the sample may also cause an important echo thatlimits the possibilities of investigations under thissurface. This example highlights the need to elimi-nate the adverse effects caused by the optical para-sitic backscattering. We have proposed in a previouspaper [7] an architecture that makes it possible to Fig. 1. Description of the classical LOFI experiment.Fig. 2. Images obtained with and without a significant parasiticreflection ( r  P  ≈  r t ) in the classical LOFI setup (All the images havebeen realized at frequency  Ω : (a) amplitude image without para-siticreflection;(b)phaseimagewithoutparasiticreflection;(c)am-plitude image with parasitic reflection; (d) phase image withoutparasitic reflection. 1 January 2009 / Vol. 48, No. 1 / APPLIED OPTICS 65  avoid these effects. However, this device is not self-aligned, which considerably complicates the imple-mentation of the LOFI optical setup. This motivatesus to develop a new architecture insensitive to para-sitic reflection that allows us to keep the self-alignedfeature and so the simplicity of the classicalLOFI setup. 3. Antireflection LOFI Device We propose a self-aligned LOFI device (Fig. 3) allow-ing the detection of backscattered light by the stu-died target without optical parasitic backscattering contribution. The laser beam is split into two parts,using double refraction phenomena in an anisotropicmaterial plate. This splitter is a beam displacer(Melles Griot reference: 03 PBD 001) that splits abeam of light into two mutually orthogonal, linearlypolarized beams that are parallel to one another andto the axis of the input beam. In our experiment, thelaserbeamdisplacementisaround 2 : 7 mm.Thelaserbeam polarization at the entrance of the splitter iscontrolled by means of a polarizer combined with ahalf-wavelength retardation plate. To control thereturn path of backscattered light from a studied tar-get or from a parasitic reflection source, a quarter-wavelength retardation plate is placed after thefrequency shifter. The control principle is explainedin the next paragraph. The slow and fast axes of thisplate are oriented at a  45 ° angle to polarizations of both orthogonally polarized laser beams. Both laserbeams are finally focused in a single point of the tar-get by a lens. The scanning is still performed by twogalvanometric mirrors. The LOFI experimental set-up presented in Fig. 3 needs no specific or/and com-plex alignment; it is self-aligned. Indeed, after thebeam splitter, both laser beams are parallel and  a priori  they overlap in the image plane of the focusing lens. The only critical alignment is the centering of the laser beams with the lens to limit astigmatismaberration, as in the classical LOFI setup. In the set-upofFig.3thisalignmentmaybemoredifficultthanin a classical LOFI setup because the beam has beensplit into two beams. However, it is even less difficultas the distance between both laser beams is small.InthisLOFIdevice,theopticalpath(upordowninFig.3)taken bythe backscattered lightmaydiffer forthe parasitic backscattered light and for the back-scattered light coming from the target. In the caseof parasitic reflection, which takes place outside of the overlapping point of both laser beams, the lightnecessarily goes back to the laser by the same path.Ontheotherhand,thebackscatteredlightbythetar-get (i.e., the signal) may come back toward the laserby a different path. The round-trip path of the back-scattered light is thus an optical loop. Depending onthe round-trip path (identical path for incident andbackscattered light or loop path), the polarizationand the frequency shift of the backscattered lightare different, which allows us to differentiate theparasitic backscattering from the target backscatter-ing. This filtering may be optical (via the polariza-tion) and/or electronic (via the frequency shift)depending on the position of the parasitic backscat-tering source.To optically isolate the parasitic reflection, wecontrolthepathofthebackscatteredlightand,there-fore, have different paths for the parasitic backscat-tering and the target backscattering. The selectedbackscattered light path is obtained using birefrin-gent properties of the beam displacer where the op-tical paths depend on the entering laser beampolarization. The principle used is illustrated inFig. 4. The beam displacer is a calcite crystal platewith an optical axis inclined toward the crystal plateedges. At the entrance of the beam displacer, we maydecompose the laser beam polarization into twoorthogonal linear polarized components, one polar-izedperpendicular totheoptical axis,calledordinary( o ) component (corresponding to ordinary ray), andone with its polarization in a plane that includesthe optical axis, called the extraordinary (  e ) compo-nent (corresponding to extraordinary ray). In Fig. 4,the ordinary ray is not displaced and the extraordin-ary ray exits at a distance away from the incidentbeam. In Fig. 4(a), we consider an incident light withordinary polarization and a parasitic reflection situ-ated beyond the quarter-wavelength retardationplate, which is oriented at a  45 ° angle to (  e ) and( o ) polarization directions. After reflection, the lightentering the beam splitter has an extraordinary po-larization because the incident light has twicecrossed the quarter-wavelength retardation platethat rotates initial polarization by  90 °. Consequently,the reflected laser beam is displaced relative to theincident beam and it cannot be reinjected into thelaser cavity. We can do a similar reasoning for anincident extraordinary ray [Fig. 4(b)]. For the back-scattered light by the target, the light has two possi-bilities to return toward the laser (identical path forincident and backscattered light or loop path). Forthe identical paths case, the light cannot be rein- jected into the laser, similarly to the cases illustratedin Figs. 4(a) and 4(b). For the loop path case, the backscattered light by the target may be reinjectedinto the laser due to the  90 ° polarization rotation.The ordinary incident ray comes back toward the Fig. 3. Quasi-self-aligned experimental setup insensitive to opti-cal parasitic reflection. 66 APPLIED OPTICS / Vol. 48, No. 1 / 1 January 2009  laser by the extraordinary polarization path and vice versa [Fig. 4(c)].Theelectronicisolationisobtainedbyselectingthereference frequency of the lock-in amplifier as in thesetup proposed in [7]. Indeed, for a parasitic reflec-tion, the laser/target and the target/laser paths arenecessarily the same, which means that the fre-quency shift is  Ω  if the optical path is the bottomarminFig.3orzeroiftheopticalpathisthetoparmsin Fig. 3. On the other hand, the frequency shift is Ω = 2  for backscattered light by the target because thelaser/target and the target/laser paths are necessa-rilydifferent.Adetectionatthe Ω = 2 frequencyallowsus to avoid the adverse effect of parasitic reflection.This double isolation ensures a phase end ampli-tude image without the parasitic reflection contribu-tion for any position of parasitic reflection in thesetup. Indeed, if the parasitic reflection source is si-tuated before the quarter-wavelength retardationplate then the isolation is electronic. If it is beyondthe quarter-wavelength retardation plate then theisolation is both electronic and optical. As a result,optics used in the setup presented in Fig. 3 do notneed an AR coating at the working wavelength, de-spite the extreme sensitivity of the LOFI technique.If the target completely depolarized the incidentlight then the quantity of reinjected light into thelaser is divided by two, but the parasitic reflectionis not affected if the parasitic reflection source doesnot depolarize the incident light. 4. Experimental Results To validate the principle of the optical device pre-sented in Fig. 3, we realized the same experimentas the one described in Section 2. The parasitic opti-cal feedback is generated by a microscope slideplaced on both arms beyond the quarter-wavelength retardation plate. We adjust the slideorientation to obtain  r  P  ≈  r t . We have realized phaseand amplitude images at frequency  Ω = 2 . The resultsobtained are given in Fig. 5. Despite a significantparasitic reflection, we obtain phase and amplitudeimages of comparable quality as the ones obtainedwithout parasitic reflection in Section 2 [Figs. 2(a) and 2(b)]. These results demonstrate that the ad- verse effects caused by parasitic reflections can beeliminated with the optical system proposed inFig. 3.In Fig. 5(a), as the image is slightly truncated, weassume that this effect can be generated by overlap-ping variations between both beams on the targetduring the scanning or by polarization variation of reinjected light through the polarizer during thescanning. During the scanning, the incident angleof the ordinary and extraordinary laser beamschanges, and this might introduce a different reflec-tivity for both polarizations. Consequently, the polar-ization of the reinjected light through the polarizermay differ from the polarization incident light. Thisdifference depends on the scanner angle, meaning thatthereinjectedlightintothelasermightdecreasefor high angles. The use of a beam displacing prismwith a smaller displacement should reduce this scan-ning effect. 5. Conclusion We have proposed a self-aligned setup for a LOFItechnique insensitive to parasitic optical feedback.Experimental results demonstrate that parasitic re-flection contribution in phase and amplitude imagecan be eliminated while keeping the self-alignedfeature of the LOFI setup. Moreover, the proposeddevice allows a double isolation of parasitic reflec-tions. The isolation may be electronic and/or opticaldepending on the location of the parasitic reflection.This double isolation allows using no AR-coated op-tics in the setup, despite the extreme sensitivity tooptical feedback of the LOFI technique. Successfulexperimental phase and amplitude images obtainedwith significant parasitic reflection validate the Fig. 4. Principle of the optical isolation: ( o ) ordinary and (  e ) ex-traordinary polarization.Fig. 5. Images obtained with parasitic reflection located beyondthe quarter-wavelength retardation plate in proposed quasi-self-aligned setup: (a) amplitude image realized at frequency  Ω = 2 and (b) phase image realized at frequency  Ω = 2 . 1 January 2009 / Vol. 48, No. 1 / APPLIED OPTICS 67  insensitivity to parasitic optical feedback. In futurework we plan to decrease the distance between bothlaser beams to limit the truncation effect and imple-ment the principle presented in this paper into theLOFI microscopy setup. References 1. E. Lacot, R. Day, and F. Stoeckel,  “ Laser optical feedback tomography, ”  Opt. Lett.  24 , 744 –  746 (1999).2. E. Lacot and O. Hugon,  “ Phase sensitive laser detection byfrequency-shifted optical feedback, ”  Phys. Rev. A   70 , 053824(2004).3. E. Lacot, R. Day, and F. Stoeckel,  “ Coherent laser detection byfrequency-shifted optical feedback, ”  Phys. Rev. A   64 , 043815(2001).4. M. E. Storm,  “ Controlled retroreflection: a technique for un-derstanding and eliminating parasitic lasing, ”  J. Opt. Soc. Am. B  9 , 1299 –  1304 (1992).5. P. Megret, L. Wuilmart, J. C. Froidure, and M. Blondel, “ Bit-error-rate in optical fiber links with optical reflections, ” in  Proceedings of IEEE Conference on Lasers-and-Electro-Optics-Society  (IEEE, 1997), Vol. 2, pp. 87 –  89.6. R. Day,  “ Une nouvelle technique d ’ imagerie laser basée sur lareinjection décalée en fréquence, laser optical feedback ima-ging, ”  Ph.D. thesis (University J. Fourier, 2000), pp. 51 –  55,http://www ‑ lsp.ujf  ‑ O. Jacquin, E. Lacot, C. Felix, and O. Hugon,  “ Laser opticalfeedback imaging insensitive to parasitic optical feedback, ”  Appl. Opt.  46 , 6779 –  6782 (2007).8. C. H. Russell M. I. Younus, and J. Blackshire,  “ Robust phase-unwrapping algorithm with a spatial binary-tree image de-composition, ”  Appl. Opt.  37 , 4468 –  4476 (1998).9. J. M. Mackowski,  “ Coatings principles, ”  in  Optics in Astrophy-sics: Proceedings of the NATO Advanced Study Institute onOptics in Astrophysics , R. Foy and F.-C. Foy, eds. (Springer,2005), Vol. 198, pp. 327 –  342.10. O. Hugon, I. A. Paun, C. Ricard, B. van der Sanden, E. Lacot,O. Jacquin, and A. Witomski,  “ Cell imaging by coherent back-scattering microscopy using frequency-shifted optical feed-back in a microchip laser, ”  Ultramicroscopy  108 , 523 –  528(2008). 68 APPLIED OPTICS / Vol. 48, No. 1 / 1 January 2009
Similar documents
View more...
Related Search
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