We demonstrate for the first time two-photon induced holographic recording at an arbitrary point in three dimensional photopolymeric cube by overlapping two coherent pulses from a 200 femtosecond Ti: Sapphire tunable laser operating at 710 nm. Spatial overlap is achieved by a novel pupil division method. The polymer material is made of epoxy host, which is fully polymerized and filled with liquid photopolymerisable formulation comprising acrylate type monomer and two-photon photoinitiator. Measured diffraction efficiency is measured to be 3.5%.
© Optical Society of America
ErrataCornelius Diamond, Yuri Boiko, and Sadik Esener, "Two-photon holography in 3-D photopolymer host-guest matrix: errata," Opt. Express 6, 109-110 (2000)
Two-photon holographic recording possesses significant advantages over its single-photon counterpart for use in optically-recorded micro-optical systems (O-MOS). Primary among these are low background recording and non-destructive readout because of a nonlinear recording threshold. This allows the creation of more multiple data or optical component layers than what would ordinarily be possible [1–2], with virtually no degradation with usage. The first demonstration of two-photon holography was by Linde et al.  in LiNbO3, employing the first and second harmonic of a Nd:glass laser. Subsequentially, two-photon gated holography was demonstrated in a polymer host matrix by G.C. Bjorklund et al.  who used continuous-wave laser sources in a variety of four-level spectroscopic systems to demonstrate recording efficiencies up to 70% . However, such two-state processes suffer from undesirable single-photon side-band reactions, which limit the diffraction efficiency.
In this communication we describe to our knowledge the first implementation of holographic gratings in a polymer gel cube by resonant two-photon photopolymerization. The recording medium utilized, a Benzil-Di-Methyl Ketal derivative in a host monomer, is a member of a new class of ultra-low two-photon threshold photopolymers we have developed which are found to exhibit performance exceeding those of the recent work in bis-donor derivatives by Cumpston et al. . Additionally, the monomers used have index-of-refraction changes between polymerized and unpolymerized regions of up to Δn=.0394, making them particularly suited for the creation of additional optically coupled diffractive or refractive components in the same microcube. This concept, called Optically-recorded Micro-Optical systems (O-MOS), is based upon one-step optical fabrication of components in a photopolymer cube. Such components are aligned to each other upon fabrication, eliminating the need for complex and expensive packaging. Previous to this communication, refractive components such as waveguides and lenses have been demonstrated. . We now demonstrate holographic elements for O-MOS, the fabrication mechanism in all cases being a non-reversible two-photon addition polymerization reaction, the reaction speed being proportional to the square of the photon density at each spatial point in the gel.
Photopolymerizable formulation employed as a recording material consisted of liquid acrylate monomer and two-photon absorbing photoinitiator. Two different cases have been explored: 1) recording in a high viscosity liquid photopolymerizable formulation; 2) recording in polymeric host filled with a liquid photopolymerizable formulation. In both cases Benzil-Di-Methyl-Ketal (BDMK) was used as two-photon photoinitiator, two-photon absorption of which is enhanced due to extended structural features of the molecule, in particular side-methyl groups located between two benzene rings. The employed wavelength of 710 nm is far away from the nearest absorption band, favoring no single photon excitation at 710 nm but rather a two-photon excitation transition at 355 nm. This photoinitiator has been measured to have a two-photon threshold in DPEPA (Di-Penta-Erythritol-Penta-Acrylate) as low as .38 GW/cm2, depending upon the concentration used.
High viscosity (13,600 cps at room temperature) and a high reactivity penta-functional acrylate monomer, namely DPEPA, was used in the liquid photopolymerizable formulation (concentration of initiator BDMK was 2 w.%). Initial refractive index of 1.4885 of the formulation was increasing when polymerization took place up to 1.5279, giving substantial dynamic range for holographic imaging of Δn=0.0394.
For the polymeric Host-Guest system the two-component epoxy material Epotek-301  was used as a Host matrix. Prior to curing of the epoxy, the liquid photopolymerizable formulation, comprising Ethoxylated Trimethylol Propane Triacrylate Ester (n=1.4695, viscosity 168 cps at room temperature) and 1–2 w. % of BDMK as two-photon photoinitiator was mixed in with components of the epoxy. 40 w.% of epoxy was used in this experimentation. DPEPA was not employed as it reacted with the epoxy hardener. The premixed formulation was poured into a 10mm3 cuvette and cured 1 hour at 65 °C.
Figure 1 shows the optical set-up that we used for creating holographic gratings with the above method. The light source we used was a mode-locked 76 MHz Ti:Sapphire laser operating at 710 nm, generating 200 fs pulses, pumped by an Ar-ion 7.5W laser. An optical isolator was used to eliminate back-reflection scattering off the sample that would otherwise cause the mode locking of the Ti:Sapphire laser to become unstable. After the isolator, each pulse proceeds through a series of collimating lenses, and then onto a Michaelson-type interferometric stage. One arm of the interferometre has an Au-coated ultra-fast mirror (Newport Corp.)which can move axially on a micrometre stage, while the pulse is displaced from its forward propagation direction by an amount Δx by the retroreflector (Coherent Corp.). The retroreflector can move perpendicularly to the beam to adjust Δx. Thus by using a retroreflector in the second arm of the interferometre, the second pulse is automatically spatially coherent with the first pulse, if the focusing objective is centred correctly. Temporal coherence is obtained by adjusting the first arm’s axial position. This is easier than the alternative of adjusting the angular and spatial position of two different arms simultaneously. The disadvantage is that one throws away half the power of the original pulse, though this is not a problem with the sensitivity of our photopolymers. The sample is placed upon a 3D scanning stage as to access any point in the cube. A low-NA objective (Seiwa, either 5X,
NA=.16 or 10X, NA=.23) as to produce a large area for mutual overlap. A telecentric imaging system is used to image the light diffracted by the grating. The thickness t of the diffraction grating depends upon the amount of overlap above threshold of the l=60 µm pulses. This is given by
the shift in the focal point of the objective when placed a distance d2 away from a sample of index of refraction n.
Figure 2 shows the fringes obtained by temporal and spatial overlap of the two pulses. Fringe spacing is approximately 2 µm, and can be modified by adjusting Δx. The zigzag pattern is because of the presence of vibrations due to the coupling of the cooling pump of the argon laser to the table. Ultimate spatial resolution of two-photon holographic gratings is limited by the amount of polymerization outside the illuminated regions, however, because of the nonlinear threshold of the two-photon process, fringe spacing can be smaller than in one-photon holographic gratings, effectively at the polymer chain diffusion limit. 
Diffraction efficiency (D.E.) of holograms made in liquid and in solid gel was measured. Holograms made in liquid had much higher diffraction efficiency, however, these decayed to neglible D.E. on the order of minutes under constant illumination, and suffered from severe scattering effects. Figure 3 shows a typical D.E. curve of a hologram in liquid under constant illumination of 16 mW average power. Writing intensity of the grating is 2.6 GW/cm2, adjusted for reflective losses off the side of the cube. There is a small period of induction while localized inhibitors such as oxygen are being scavenged, then a sharp rise in D.E. to 69%, before falling to less than 1% on the same timescale. This is because of the loss of Bragg matching as the created dense grating sinks in the lighter surrounding fluid. Optimally, one would want to lock in diffraction efficiency at its peak by polymerizing the entire surrounding area with UV light, though final D.E. is severely reduced as the background index of refraction rises to the level of the grating. Thus for better stability and beam profile, three dimensional volume holographic writing in better performed in gel, whose non-reacting background component will provide the necessary refractive index break.
The solid gel hologram had a maximum efficiency of 3.6%, which decayed to 2.7% over the course of 72 hours. The threshold intensity for the above gel sample was four times that of the liquid sample, owing to use of a less reactive monomer. Thus the holographic recording in these two cases differ significantly — near threshold condition for gel-type sample and four times above the threshold for the liquid sample. For that reason a straightforward conclusion about stronger holographic performance of the liquid sample over the gel one would be incorrect — peak recording efficiency for the gel sample is expected to be higher when recorded sufficiently above threshold. This is because higher intensities will lead to higher densification in fringe regions due to increased diffusion rates. Increasing material sensitivity will constitute the subject of our future research efforts in the field.
Based upon a holographic element with a 40 µm thickness and a 5 µm radius, it would be possible to create a 3-layer, (collimating lens layer, wavelength splitting series layer, focusing lens layer) 20-channel DWDM multiplexer/demultiplexer for an array of 20 fibres in a 1mm3 O-MOS cube, with a 40 µm output spacing. Similar calculations show sub-terabit WORM holographic memories for a 1cm3 O-MOS cube, assuming 100 bits written in 20 µm2 micro-holograms with an angular multiplexing factor of 10.
In conclusion, we have demonstrated one-step in-situ two-photon holography at an arbitrary point within a gel cube containing photopolymers. This method involves a photopolymerization reaction initiated by a highly efficient two-photon fluorophore encapsulated in a host epoxy. Alignment of recording and addressing beams is by a novel pupil division method. Diffraction efficiency is limited only by the shrinkage of the host matrix.
References and links
1. A. S. Dvornikov and P.M. Rentzepis, Opt. Commun.119, 341 (1995). [CrossRef]
2. A. Toriumi, S. Kawata, and M. Gu, “Reflection confocal microscope readout system for 3D photochromic optical data storage,” Opt. Lett. 23, pp 1924–6 (1998). [CrossRef]
3. D. von der Linde, A. M. Glass, and K. F. Rodgers, Appl. Phys. Lett.25, 155 (1974). [CrossRef]
4. G. C. Bjorklund, Chr. Brauchle, D. M. Burland, and D. C. Alvarez, “Two-photon holography with continuous-wave lasers,” Opt. Lett. 6, 159–61 (1980). [CrossRef]
5. Chr. Brauchle, Urs P. Wild, D. M. Burland, G. C. Bjorklund, and D. C. Alvarez, “Two-photon holographic recording with continuous-wave lasers in the 750-1100nm range,” Opt. Lett. 7, 177–9 (1982). [CrossRef]
6. B.H. Cumpston, J.W. Perry, and S. Marder et al, “New Photopolymers based on two-photon absorbing chromophores and application to three-dimensional microfabrication and optical storage,” Mat. Res. Soc. Symp. Proc. 488, 217–25 (1998). [CrossRef]
7. Cornelius Diamond, “O-MOS: Optically Written Micro-optical Systems,” Ph.D. Thesis, University of California, San Diego, exp. Pub. January 2000.
8. Epoxy Technology, Inc. Fully cured refractive index was measured to be 1.54.