Note: The following material
`may be protected by copyright
`law (Title 17, U.S. Code)
`
`Impact of diffractive optics on the design of optical pick up
`
`THOMSON CSF Laboratoire central de recherches F91404 Orsay France
`
`J-C Lehureau
`
`1. INTRODUCTION
`
`With over 100 millions units per year, the optical pick up is probably the most widely produced electro
`
`optic assembly on these days. It is also the driving force of semiconductor laser production with development of
`
`higher power for high data rate recording and shorter wavelength for high density. The production of CD pick up
`
`is of course highly cost driven; over 20 years of optical storage history and ten years of CD, the reduction in size,
`
`power consumption and cost can be compared to the evolution of integrated circuit. Diffractive optics plays a
`
`significant role in this roadmap, although as will be seen not all functions can be expected to be played by
`
`diffractive optics. Refractive lens will remain the leading technology for all functions that involves strong
`
`bending of the rays.
`
`In the near future the production od CD and CD-ROM player is expected to reach a plateau at twice today’s
`
`production but MO drives and soon DVD are taking off.
`
`2. STRUCTURE AND FUNCTIONS OF OPTICAL PICK UP
`
`Whereas magnetic storage and audio disc rely on the proximity of media and sensing device to localize
`
`information, optical storage consists in focusing of a beam on the information elements. Although the density is
`
`strongly limited by the diffraction limit, optical storage is definitely advantageous in terms of spacing between
`
`media and head and therefore provides immunity to dust; it will remain the leading technology for removable
`
`media.
`
`The role of optical head is triple:
`
`-focus the beam on the spot to be read,
`
`-detect information as encoded by phase, amplitude or polarization,
`
`-sense the tracking errors and react toward optimal position.
`
`Simple considerations on data rate and sensor sensitivity show that single spot readout requires a high intensity
`
`light source that only coherent laser source can achieve; the semiconductor laser was a key element to the
`
`development of cheap and compact optical head. This component has a lot of wonderful peculiarities but two
`
`major drawbacks: LG Electronics, Inc. et al.
`EXHIBIT 1004
`IPR Petition for
`U.S. Patent No. RE43,106
`
`22 / SPIE Vol. 2783
`
`0-8194-2 169-3/96/$6.00
`
`Downloaded From: http://proceedings.spiedigitallibrary.org/on 06/26/2015 Terms of Use: http://spiedl.org/terms
`
`

`
`-its emission pattern is not axio-symetric: vertical diffraction angle is up to three times the
`
`horizontal one,
`
`-its wavelength is not well defined during the production process and may further vary with the
`
`temperature at a rate of .3nrn/°C.
`
`The design of optical pick up will have to cope with these problems.
`
`focusing lens
`
`disk
`
`beam
`
`tracking and focus
`actuators
`
`quarterwave
`collimab3r
`c~indrical lens
`
`4 quadrant detector
`
`Figure 1: slTucture of a pick up
`
`When a high numerical aperture beam is focused on the disk, its reflection is focused back right into the laser
`
`source; therefore a beam splitter or a circulator is needed in order to deflect the beam towards the detectors.
`
`Many types of sensors have been deviced in order to sense the focus and radial tracking errors.
`
`The most widely used focus sensors are:
`
`-astigmatic sensor: some cylindrical focusing effect is added on the return path in order to
`
`elongate the focused spot in one of the two directions 45° away from the track according to the focusing error. A
`
`four quadrant detector assembly senses the shape of the returning spot,
`
`-asymetric sensor: a part of the far field of the returning beam is blocked or deviated in order to
`
`generate a disymetry of the returning spot; this disymetry is cancelled when the beam is well focused.
`
`The most widely used tracking sensors are:
`
`-three spots: two sidebeams are focused slightly left and right to the track; the balance of
`
`reflected power ot these two beams provides the error signal,
`
`-push pull sensor: the tracking error generally introduces a disymetry of the far field pattern; the
`
`sign and amplitude of this disymetry is related to the pit depth and is specified in the disc standard.
`
`Downloaded From: http://proceedings.spiedigitallibrary.org/on 06/26/2015 Terms of Use: http://spiedl.org/terms
`
`SPIE Vol. 2783 1 23
`
`

`
`As can be seen, error sensing requires an action in the far field (splitting, masking, wavefront deformation) and
`
`detection in the near field. A diffractive element is useful in order to separate the reflected beam from the incident
`
`one while introducing such action in the far field (wave front shaping or beam division).
`
`3. ADVANTAGES OF DIFFRACTIVE COMPONENTS
`
`In a conventional CD pick up, the separation of the incident and the reflected beam is performed by a
`
`half reflective mirror and as the returning beam is exactly focused on the laser source, the detectors must be
`
`positioned precisely on the conjugate of the laser source through the half mirror. The stability of the return spot is
`
`so stringent that up to recently, the housing that holds together the laser, detectors and half mirror could not be
`
`made of injected polymer. A diffractive beam splitter, on the other hand is only slightly deflecting the beam and
`
`its positioning is far more tolerant. In other words, whereas a mirror must be adjusted within a fraction of half
`
`wavelength at the edge of the beam, the positioning of a grating is tolerant up to a fraction of its pitch. This is
`
`more than an order of magnitude more tolerant and may accomodate plastic housing.
`
`Another advantage of the diffractive beam splitter is that any type of optical function can be added at nearly no
`
`cost on the diffracted beam. This allows to implement the "three beam method" by acting on the incident beam or
`
`to introduce astigmatism on the returning spot. It is also possible to split the far field of the beam and focus it on
`
`different detectors as is necessary in the "asymetric focus sensor".
`
`All these functions can be implemented at very low cost; it is possible not only to replicate the diffracting
`
`function by photolithography but to realize very cheap elements by simple injection molding. The
`
`photopolymerization process ("2P") also allows the very precise replication of a nickel master in a UV-curable
`
`resist; since the substate may be glass, long term dimensional stability can be reached at low cost.
`
`4. PROBLEMS ASSOCIATED WITH DIFFRACTIVE OPTICS
`
`If so much versatility can be expected from diffractive optics, one may wonder why all the functions of
`
`optical pick up are not done that way. One of the strong limitations is the instability of the source wavelength.
`
`Thermal variations generate a shit~ that may reach 1% of nominal wavelength; even if refocusing corrects the
`
`first order of the effect, stigmatism is also a question since for the high numerical aperture needed to focus on a
`
`disk, the pitch does not vary linearly with the radius of the lens.
`
`Even in such a function as beam splitting, the chromatism of the grating affects the position of the returning spot
`
`and precludes sensors which need a 2-dirnensional positioning of the spot with respect to the detectors like
`
`astigmatism.
`
`Another problem encountered with diffractive optics is the existence of many orders of diffraction. This is taken
`
`as an advantage to generate the two lateral beams of the "three beam sensor" but it is disadvantage when using
`
`24 / SPIE Vol. 2783
`
`Downloaded From: http://proceedings.spiedigitallibrary.org/on 06/26/2015 Terms of Use: http://spiedl.org/terms
`
`

`
`grating as a beamsplitter since the -1 diffracted order of the incident beam may overlap with the +1 diffracted
`
`order of the returning beam; this is generally solved by using diffraction angle larger than the optics field, but at
`
`the cost of higher chromatism and more stringent positioning. We are developing beamsplitters with no -1
`
`diffracted order for use in a self adjusting pick up; some profiles are characterized by a perfect phase
`
`compensation of the different zones of the grating in one or another order: we choose a three level profile that
`
`provides respectively 0%, 25% and 41% on the -I, 0 and +1 orders.
`
`phase reference of order ÷1
`
`__h ’,.I
`
`/ / )’,/
`/ /. / /
`
`-~/2
`
`Oth order +1 order
`25% 41%
`Figure 2: holographic profile for oplJcal pick up
`
`-1 ordre
`0%
`
`Even if duplication of diffractive optics can be done at very low cost, the realization of the master is often a
`
`problem. As soon as more than two levels are required to define the profile of the grating simple photolithograhic
`
`process can not be used. Oblique ion etching has been used but it offers low versatility and requires long
`
`experimentation to get proper results. Fortunately, recent developments of multilevel binary optics allow to
`
`realize complex profiles provided the spatial frequency is not too high.
`
`5. SOME ACHIEVEMENTS OF PICK UP USING DIFFRACTIVE OPTICS
`
`The most important achievement today is the mass production of a laser assembly by Sharp. This
`
`assembly integrates all the function of a CD pick up except the moving lens; this includes laser, laser power
`
`sensor, RF and tracking sensor assembly, three beam grating and beamsplitter grating. The sequence of mounting
`
`and adjustements is of uttermost importance for the industrial process:
`
`-the three chips (laser,power sensor, detector assembly) are mounted and wired in the socket,
`
`Downloaded From: http://proceedings.spiedigitallibrary.org/on 06/26/2015 Terms of Use: http://spiedl.org/terms
`
`5PIE Vol. 2783 / 25
`
`

`
`-the double faced holographic plate is positioned in front of the case and rotated until the
`
`asymetric sensor is well balanced,
`
`-a reference index is made with respect to the three beam sensor for positioning the laser
`
`assembly into the final pick up.
`
`Such a laser assembly is also used for magnetooptic pick up for MD application; however, the magnetooptic
`
`readout function has not been integrated yet.
`
`¯ -:, laserchip
`
`dual hologram
`beam splitter
`
`power
`control
`
`:..
`
`RF and servo detectors
`
`Figure 3: holographic laser assembly
`
`6. FUTURE CHALLENGES FOR DIFFRACTING OPTICS
`
`6.1 Magnetooptic pick up
`
`Both the rapid growth of MO drives (3.5") and the emergence of cheap magnetooptic drive in the audio
`
`field (MD) creates the need for simpler magnetooptic pick up. Polarization effects appear when the diffraction
`
`angles are important; this raises once again the problem of wavelength dependance. An elegant solution was
`
`proposed by Maeda et al where two polarizing holograms are superimposed on the same plate and result in a
`
`small deflection of the beam.
`
`sliglnlJy differ, ent pitches
`
`incident
`polarizalJons
`
`li.
`
`\
`
`Figure 4: compensated polarizing gralJngs
`
`26/SPIE Vol. 2783
`
`Downloaded From: http:llproceedings.spiedigitallibrary.org/on 0612612015 Terms of Use: http:/Ispiedl.org/terms
`
`

`
`We also realized gratings at the interface between an isotropic substrate and a highly birefringent liquid crystal;
`for proper profile one polarisation can be fully deflected along one order of diffraction while the other remains
`undiffracted. The liquid crystal can also be used as a controllable birefringenee for optimum magnetooptic
`
`detection.
`
`P
`
`s\
`i \
`
`glass cover
`__ nernalic Ic
`m olded grating
`glass susb’ate
`
`,~"~
`
`Figure 5: Polarizing bearnsplitter
`
`6.2 Bi-standard lens
`
`The emergence of a new standard (DVD) in the audio-video market raises the question of compatibility
`
`with the former CD standard. Both cover thickness and numerical aperture have changed to the point where
`
`objective lenses are fully incompatible. Twin lens actuators are bulky, expensive and power consuming; on the
`
`other hand holographic function at the surface of the molded lens enables independant beam shaping for each
`
`type of disc at nearly no cost.
`
`6.3 Self adjusted pick up
`
`As mentioned earlier, a holographic pick up requires adjustment during manufacturing due to the large
`
`deflection angle which induces chromatism and sensitivity to positioning. If one can solve the -1 order problem
`
`and find the near field function directly on the laser chip, it is possible to rely on the precision of patterning and
`
`mechanical assembly to realize all functions of a CD without adjustment. We realized such a structure using the
`
`edge of the GaAs chip as a near field Foucault knife edge; however the edge is realized by sawing and it is not
`
`precise enough for self adjustment. Better results can be expected from using the GaAs component as a detector;
`
`it has been shown in the communication field that a laser stripe can be used as a detector when properly
`
`polarized. A further advantage of this solution would be to use a confocal microscope structure which is less
`
`sensitive to aberrations than the conventional central aperture detection.
`
`6.4 Planar pick up
`
`Slimmer structure would be appreciated, specially when multidisc player is needed. One elegant solution
`
`has been proposed which combines planar waveguide and out of plane beam forming hologram. Unfortunately,
`
`many problems are still to be solved:
`
`Downloaded From: http://proceedings.spiedigitallibrary.org/on 06/26/2015 Terms of Use: http://spiedl.org/terms
`
`SPIE Vol. 2783 / 27
`
`

`
`-the butt coupling of the laser with the waveguide is unefficient and requires precise positioning;
`
`-the chromatism of the diffracting structure is too high and introduces aberrations when
`
`wavelength changes;
`
`-for MO application, the phase of TE and TM waves are rapidly changing with wavelength and
`
`temperature and requires an additiormal adjustment.
`
`I~ol ographic lens
`
`beam splitter
`
`detectors
`
`,guiding layer
`g layer
`
`RF output
`
`focus error
`
`#’acking error
`
`Figure 6: Planar pick up
`
`7. CONCLUSION
`
`It is clear that diffracting optics offers advantages over conventional refractive optics in certain cases:
`-when beam splitting is needed, specially at low diffraction angle,
`
`-for low cost polarizing functions,
`- for complex patterning and wave front formation.
`However, diffractive optics also has strong limitations:
`
`-it is difficult to control the power splitting over different diffraction order;
`
`-it is highly chromatic;
`-some functions are difficult to master.
`
`28 15PIE Vol. 2783
`
`Downloaded From: http://proceedings.spiedigitallibrary.org/on 06/26/2015 Terms of Use: http://spiedl.org/terms
`
`

`
`On the cost issue, there is no signicant difference between duplicating planar diffracting structure or refracting
`surfaces. It must be a careful decision of optical engineering to choose where refractive or diffractive optics is
`needed. Most of the future progress should concentrate in mixing both solutions rather than concentrating
`exclusively on one or the other.
`
`8. BIBLIOGRAPHY
`
`[1]- Principles of optical disc systems G. Bouwhuis et al
`
`[2]- A multifunctional reflection type grating lens for the CD head
`
`[3]- High performance optical head using optimized HOE
`
`[4]- A high density dual type grating for MO disc head
`
`[5]- Polarizing grating beamsplitter
`
`[6]- Possibility of super resolution readout in integrated optic PU
`
`Adam Hilger Ltd
`K. Tatsumi et al
`Y. Kimura et al
`H. Maeda et al
`J-C Lehureau et al
`T. Suhara et al
`
`ISOM ’87
`ISOM ’87
`ISOM ’89
`ISOM ’89
`ISOM ’89
`
`Downloaded From: http://proceedings.spiedigitallibrary.org/on 06/26/2015 Terms of Use: http://spiedl.org/terms
`
`SPIE Vol. 2783 / 29