WO2008083445A1

WO2008083445A1 - Optical analysis system and method

Info

Publication number: WO2008083445A1
Application number: PCT/AU2008/000029
Authority: WO
Inventors: Robert Fischer; Dragomir Neshev; Solomon Mois Saltiel; Wieslaw Zbigniew Krolikowski; Yuri Kivshar
Original assignee: The Australian National University
Priority date: 2007-01-12
Filing date: 2008-01-11
Publication date: 2008-07-17
Also published as: WO2008083445A8

Abstract

An optical system comprising: a nonlinear material having a ferroelectric domain structure, the nonlinear material capable of converting first and second optical signals respectively to first and second frequency-converted optical signals; and alignment means for respectively aligning the first and second optical signals such that they propagate collinearly, but in opposite directions, through the nonlinear medium to obtain a overlap region in the nonlinear material where the first and second optical signals overlap, wherein the nonlinear material being capable of converting the first and second optical signals to a third frequency converted optical signal in the overlap region; wherein the third optical frequency generated by the nonlinear material propagates in a direction that is either oblique or transverse to the propagation direction of both the first and second optical signals.

Description

OPTICAL ANALYSIS SYSTEM AND METHOD

TECHNICAL FIELD

[ 0001 ] The present invention relates to optical analysis systems and specifically to systems and methods for detection and analysis of an optical signal using second harmonic light s generated by the optical signal of interest.

[ 0002 ] The invention has been developed primarily as a system and method for detection and analysis of a pulsed optical signal using planar non-collinear sum-frequency mixing or second harmonic generation and will be described hereinafter with reference to this application. However it will be appreciated that the invention is not limited to this particular field of use. i o BACKGROUND OF THE INVENTION

[ 0003 ] Any discussion of the background art throughout the specification should in no way be considered as an admission that such background art is prior art nor that it is widely known or forms part of the common general knowledge in the field.

[ 0004 ] The nonlinear optical process of second harmonic generation (hereafter SHG) is is arguably one of the most important nonlinear processes employed in optical applications and is commonly used for generation of visible or near-ultraviolet coherent optical radiation. A drawback of most SHG processes, however is that the second harmonic output cannot in general be tuned over a broad range of output frequencies due to the phase matching conditions imposed by the nonlinear medium used. Birefringent crystals may be oriented for phase matching0 between the fundamental and second harmonic frequencies and are most commonly used for efficient second harmonic generation, although the nonlinear conversion at the particular orientation is limited to the designed fundamental wavelength. The birefringence method of nonlinear conversion can not be used in crystals which have either small or no birefringence.

[ 0005 ] Alternatively, quasi phase-matching (QPM) for second harmonic generation can be₅ achieved by periodic spatial modulation of the second order nonlinearity of the material (known as periodic poling) . This technique can be used to allow conversion of a wider range of fundamental frequencies in a particular type of nonlinear material. Each fundamental frequency, however, requires a unique period of the nonlinearity modulation to fulfil the phase matching condition and hence for each wavelength of interest, a sample with grating which has been0 written specifically for that wavelength is required. [ 0006 ] Angle phase matching can be used over wide range of wavelengths since the birefringence changes significantly with angle. However, the birefringence causes the energy flow (Poynting vector) of the ordinary and extraordinary beams to propagate in different directions, and this effect is known as walkoff. This reduces the overall efficiency, because the beam energies must remain overlapped for the nonlinear interaction to occur, and the walkoff does not allow perfect overlapping. Walkoff does not occur if all the beams propagate at 90 degrees to the optic axis of the crystal, but the birefringence must then be adjusted using its temperature dependence, and this is called noncritical phase matching or temperature phase matching. Noncritical phase matching has advantages over angle phase matching, because it is more tolerant of misalignment, which therefore allows stronger focusing of the beams to obtain higher efficiency. The disadvantage is that the temperature dependence of birefringence is rather small, so the range of wavelengths that can be noncritically phase matched is much smaller than for angle phase matching.

[ 0007 ] Furthermore, in each of these cases, the second harmonic frequency, in general, propagates either collinearly with, or with only a small angular walk-off with respect to the propagation direction of the fundamental frequency, thereby making the separation of the fundamental and second harmonic frequencies required, adding further complexity to the optical system.

[ 0008 ] Efficient second harmonic generation (SHG) has become a common tool for a wide range of applications. However, most of the employed techniques to achieve the necessary phase matching (PM) condition - such as angular, temperature or non-critical phase matching or quasi phase matching - are highly limited in the bandwidth over which the second harmonic of an input wavelength can be achieved.

[ 0009 ] Detection and analysis of an optical signal through its second harmonic has a number of benefits, not the least of which including detection and/or analysis of light having a fundamental wavelengths that are higher than that able to be detected by inexpensive detectors.

For example, the optical bandgap of silicon detectors is approximately 1100 nm, hence they cannot be employed for detection or analysis of optical signals in the 1450 nm to 1600 nm range, and particularly at wavelengths around 1550 nm, corresponding to important optical telecommunications frequencies. There is therefore a need for systems and methods to enable the detection and analysis of long wavelength optical radiation using inexpensive devices.

[ 0010 ] However, since the output frequency of the nonlinear process is limited by the phase matching conditions for current systems, second harmonic generation has therefore been generally limited to the realm of generation of a new optical frequency for performance of a particular task rather than as a detection and analysis tool for an optical signal of interest at a frequency which is more difficult to detect using less complex schemes and inexpensive equipment.

5 [ 0011 ] Therefore, a need also exists for an improved detection/analysis scheme utilising nonlinear optical conversion of an optical signal of interest.

[ 0012 ] Ultra-short pulses have become a common tool for a rapidly increasing number of applications in physics, chemistry and biology, including both research and industry and precise knowledge of the temporal characteristics of the ultrashort pulses is increasingly important. Q While in most cases a precise knowledge about the pulse shape and duration is essential for the success of the particular applications, characterization of pico- and femto-second pulses requires elaborated, alignment-critical and expensive tools with usually large footprint (i.e. the physical area on a benchtop for the characterization equipment).

[ 0013 ] Currently, several techniques are available to measure ultrashort optical pulses, fors example:

(i) Intensity autocorrelation, which is probably the most commonly applied technique providing information about the pulse duration when a particular pulse shape is assumed [see for example E. P. Ippen and C. V. Shank, in Topics in Applied Physics, edited by S. L. Shapiro (Springer-Verlag, New York, 1977), Vol. 18, p. 83]; Q (ii) Spectral interferometry, including SPIDER [see for example A. Walmsley and V.

Wong, Journal of the Optical Society of America B 13 (11), pp. 2453-2463 (1996)] wherein a pulse is interfered with a frequency-shifted or spectrally sheared copy of itself;

(iii) Frequency-Resolved Optical Gating (FROG), a spectrally resolved autocorrelation technique which provides more detailed temporal and spectral pulse shape [see for example D. J.5 Kane and R. Trebino R, IEEE Journal of Quantum Electronics 29, pp. 571-579 (1993)], available from Swamp Optics, LLC of Atlanta, GA 30339-2919, USA; and

(iv) GRating-Eliminated No-nonsense Observation of Ultrafast Incident Laser Light E-fields (GRENOUILLE), [see for example P. O'Shea, M. Kimmel, X. Gu₅ and R. Trebino, Optics Letters 26 (12), pp. 932-934 (2001)] which is a simplified version of FROG also available0 from Swamp Optics, LLC. [ 0014 ] Pulse-characterization techniques such as FROG or SPIDER are based on the analysis of the second harmonic (SH) signal generated by two pulses copropagating (i.e. in the same direction) collinearly or under a small angle through a nonlinear material. They rely on series of interferometric measurements with an accurately varied time delay between the two pulses, thus requiring accurate and time consuming alignment. Improved techniques such as GRENOUILLE overcome most of these problems but they assume the optical signal to consist of only a single repeated pulse, while information on a time scale larger than the pulse length, e.g., that coming from a delayed pulse echo, is lost.

[ 0015 ] In practice, using the above methods can be a complex and alignment-critical process and significant care in the set up and alignment of the system is required to obtain a meaningful measurement. The alignment process may be relaxed by the use of techniques similar to that of two-photon fluorescence (TPF) in dye solutions with counterpropagating pulses (i.e. propagating in opposite directions) and transverse detection [see for example J. A. Giordmaine, P. M.

Rentzepis, S. L. Shapiro, and K. W. Wecht, "Two-photon excitation of fluorescence by picosecond light pulses", Applied Physics Letters 11 216 (1967) and the review articles by D. J.

Bradley and G. H. C. New; "Ultrashort pulse measurements"; Proc.of the IEEE, vol. 62 (3) p.

313 (1974) or by J. H. Bechtel and W. L. Smith, J., Journal of Applied Physics 46 (11), pp.

5055-5056 (1975)]. TPF methods enable a single-shot detection where no scanning across the pulse or series of measurements is required for a single pulse characterization, and can measure much larger time windows (longer pulses) without the need for variable delay lines. However, this method has many drawbacks including limited bandwidth, low reliability of the dyes, high peak powers required, and dye degradation which prevents wide application of this method.

[ 0016 ] The extension of the TPF method to parametric frequency conversion processes in solid state materials is also not a trivial exercise since transverse phase matching of the these process by birefringence is physically impossible. An option is to employ a planar waveguide geometry leading to the so-called surface emitting SHG from counterpropagating beams. However, the efficiency of this process in the non-phase-matching regime is low. The efficiency can be enhanced by the use of transverse phase matching. In order to phase match a process where one photon from each counterpropagating pulse contributes to the generation of a single photon at the sum frequency, a periodic photonic structure is required that can compensate for the phase-mismatch. In the case of transverse phase-matching, this imposes a fundamental constraint on practical realization of devices as the required periodicity for first-order phase- matching is of a subwavelength scale. Attempts to realize such nanoscale photonic structures 8 000029

- 5 - have been made [see for example G. D. Landry and T. A. Maldonado, J. Opt. Soc. Am. B 21, p. 1509 (2004)], however, the device performance has been limited to a narrow frequency range, which, together with difficulties in fabrication, limits its practical applicability.

[ 0017 ] Therefore, there is a demand for a less complex, compact and inexpensive alternative to perform autocorrelation and crosscorrelation measurements of optical pulses, whether they are a part of a pulse train or an isolated pulse in a single-shot operating regime. There is also a demand for an autocorrelation method which is less critically alignment sensitive to allow for increased ease of operation and greater confidence in the results of the autocorrelation measurements. This demand is felt, for example, with respect to both optical communication systems and short or ultra-short optical pulse generation systems.

SUMMARY OF THE INVENTION

[ 0018 ] Disclosed herein are optical systems and apparatus configured for optical pulse analysis using second harmonic generated light. The optical pulse to be detected may be part of a pulse train and each pulse in the train may be compared to other pulses or to a reference pulse, or alternatively, the pulse under analysis may be a single pulse. The single optical pulse or pulse train may be from a pulsed laser source and the optical system or apparatus may perform an autocorrelation of the pulse(s) for analysis of the temporal optical properties of the pulse(s).

[ 0019 ] Also disclosed are methods and apparatus for the measurement and analysis of a first optical signal at a first fundamental frequency, Q₁. The first optical signal may be a pulsed optical signal. A first frequency converted signal, Q₁*, that is derived from the first optical signal may be generated by an optical non-linear conversion process, and may be at the frequency of the second harmonic of the first fundamental frequency, i.e. Q₁* = 2Q₁. The first frequency converted signal may propagate in an oblique direction to the propagation direction of the first optical signal, and the propagation direction of the first frequency converted signal may also be substantially perpendicular or transverse to the propagation direction of the first optical signal or may also include the propagation direction of the first optical signal. In some arrangements the nonlinear conversion process may be configured such that there is negligible or minimal disruption and/or distortion of the first optical signal during the measurement process, however, in other arrangements, the nonlinear conversion process may be configured to convert a substantial portion of the first optical signal to the first frequency converted optical signal to provide an increased signal to noise ratio if required. The portion of the first optical signal which is converted to the first frequency converted signal may be in the range of 0.001% to U2008/000029

- 6 - several percent. The converted portion may be 0.001% to 50% of the first optical signal, or alternatively, the converted portion maybe between 0.001% to 1%, 0.005% to 1%, 0.01% to 1%, 0.01% to 0.5%, 0.01% to 0.4%, 0.01% to 0.3%, 0.01% to 0.2%, 0.01% to 0.1%, 0.02% to 2%, 0.02% to 1%, 0.02% to 0.5%, 0.02% to 0.4%, 0.02% to 0.3%, 0.02% to 0.2%, 0.02% to 0.1%, or approximately 0.5% to 50%, 0.5% to 45%, 0.5% to 40%, 0.5% to 35%, 0.5% to 30%, 0.5% to 25%, 0.5% to 20%, 0.5% to 15%, 0.5% to 10%, 0.5% to 5%, 0.5% to 4%, 0.5% to 3%, 0.5% to 2%, 0.5% to 1%, 10% to 50%, 10% to 45%, 10% to 40%, 10% to 35%, 10% to 30%, 10% to 25%, 10% to 20%, or approximately 10% to 15% and the portion of the first optical signal at the first optical signal that is converted may be approximately 0.001%, 0.005%, 0.01%, 0.02%, 0.03%, 0.04%, 0.05%, 0.1%, 0.2%, 0.3%, 0.4%, 0.5%, 1%, 2%, 3%, 4%, 5%, 6%, 7%, 8%, 9%, 10%, 15%, 20%, 30%, 40%, 50%.

[ 0020 ] Further disclosed are methods and apparatus for the detection of a first optical signal at a first fundamental frequency by overlapping the first optical frequency with a second optical signal having a second fundamental frequency ω₂ in a nonlinear medium. The second optical signal may be a pulsed optical signal. The first and second optical signals may be configured to be propagating collinearly, but in opposite directions, through the nonlinear medium such that a region exists in the medium where the first and second signals overlap spatially. A portion of both the first and the second optical signals, with frequencies G)₁ and CO₂, are converted respectively to first and second frequency converted signals respectively having frequencies coj* and Co₂*. The first and second frequency converted signals may radiate in a direction that is at least oblique to the direction of propagation of the first and second signals respectively and may be perpendicular or transverse to the direction of propagation of the first and second optical signals. The first and second frequency converted signals may radiate in a direction that is at least oblique to the direction of propagation of the first and second signals respectively and may be perpendicular or transverse to the direction of propagation of the first and second optical signals. The first and second frequency converted signals may be the second harmonic frequency of the first fundamental frequency such that Co₁* = 2 CO₁ and co₂* = 2 CO₂.

[ 0021 ] In the region in the nonlinear medium where the first and second optical signals spatially overlap, a portion of the first and second signals may be converted by the nonlinear medium to a third frequency converted optical signal having a frequency co₁₂*. The third frequency converted optical signal may be generated in the nonlinear medium via sum frequency mixing of the first and second optical signals whereby ω₁₂* = Co₁ + ω₂ when the first and second 2008/000029

- 7 - optical signals have different frequencies, or alternatively via second harmonic generation when the first and second optical signals are of the same frequency i.e. ω₁₂* = 2 Go₁ or ω₁₂* = 2 ω₂.

[ 0022 ] The third frequency converted signal may propagate radially in all directions in a plane that includes the propagation direction of the optical input signal or an alternate arrangements the third optical signal may propagate in a direction that is oblique to the propagation direction of the first optical signal or in further arrangements the third frequency converted signal may propagate in a direction that is perpendicular or transverse to the propagation direction of the first optical signal. The propagation direction of the third frequency converted signal may substantially coincide with the propagation direction of the first and second frequency converted signals.

[ 0023 ] The nonlinear medium may be a solid state nonlinear medium and may be a solid state nonlinear material. The nonlinear medium may be a ferroelectric nonlinear medium and may have a plurality of ferroelectric domains. The ferroelectric domains may be available in the unpoled crystal structure and may be random or disordered or they may be artificially produced by electric poling techniques such that they are regularly arrayed. The ferroelectric domains may be disordered antiparallel domains. The random ferroelectric domains may be randomly sized and/or randomly oriented and/or randomly situated. The nonlinear medium may be capable of quasi-phased matching (QPM) nonlinear frequency conversion processes of optical signals. The nonlinear medium may phase match or QPM any second order process in the range of 0.2 to 10 μm, or alternately in the range, 0.2 to 9 μm, 0.2 to 8 μm, 0.2 to 7 μm, 0.2 to 6 μm, 0.2 to 5 μm, 0.3 to 10 μm, 0.3 to 9 μm, 0.3 to 8 μm, 0.3 to 7 μm, 0.3 to 6 μm, 0.3 to 5 μm, 0.4 to 10 μm, 0.4 to 9 μm, 0.4 to 8 μm, 0.4 to 7 μm, 0.4 to 6 μm, 0.4 to 5 μm, 0.4 to 4 μm, 0.5 to 10 μm, 0.5 to 9 μm, 0.5 to 8 μm, 0.5 to 7 μm, 0.5 to 6 μm, 0.5 to 5 μm, 0.5 to 4 μm, 0.5 to 3 μm, 0.5 to 2 μm, 1 to 10 μm, 1 to 9 μm, 1 to 8 μm, 1 to 7 μm, 1 to 6 μm, 1 to 5 μm, 1 to 4 μm, 1 to 3 μm, or 1 to 2 μm, 1.4 to 1.7 μm, 1.45 to 1.65 μm, 1.5 to 1.6 μm, 1.53 to 1.57 μm, or 1.54 to 1.56 μm. The second order process may be second harmonic generation, sum frequency mixing or difference frequency mixing. The nonlinear medium may be phase-matched or QPM to provided broadband operation for frequencies in the range of 0.2 to 10 μm, or alternately in the range, 0.2 to 9 μm, 0.2 to 8 μm, 0.2 to 7 μm, 0.2 to 6 μm, 0.2 to 5 μm, 0.3 to 10 μm, 0.3 to 9 μm, 0.3 to 8 μm, 0.3 to 7 μm, 0.3 to 6 μm, 0.3 to 5 μm, 0.4 to 10 μm, 0.4 to 9 μm, 0.4 to 8 μm, 0.4 to 7 μm, 0.4 to 6 μm, 0.4 to 5 μm, 0.4 to 4 μm, 0.5 to 10 μm, 0.5 to 9 μm, 0.5 to 8 μm, 0.5 to 7 μm, 0.5 to 6 μm, 0.5 to 5 μm, 0.5 to 4 μm, 0.5 to 3 μm, 0.5 to 2 μm, 1 to 10 μm, 1 to 9 μm, 1 to 8 μm, 1 to 7 μm, 1 to 6 μm, 1 to 5 μm, 1 to 4 μm, 1 to 3 μm, or 1 to 2 μm, 1.4 to 1.7 μm, 1.45 to 1.65 μm, 1.5 to 1.6 μm, 1.53 to 1.57 μm, or 1.54 to 1.56 μm.

[ 0024 ] In alternate arrangements, the ferroelectric domains of the nonlinear medium may be naturally ordered, or the nonlinear medium may be periodically poled such that the domains are regularly arrayed or regularly spaced with a period corresponding to a desired wavelength of operation to provide QPM for conversion of an optical signal at a desired wavelength or wavelength range of operation to a nonlinear converted signal. The regularly arrayed domains may alternatively be annularly arrayed. The ordered domain structure may provide increased nonlinear conversion efficiency for the optical signal at the desired wavelength or wavelength range of operation. The domains may be ordered such that the nonlinear converted signal propagates in a direction that is oblique to the propagation direction of the optical signal or in further arrangements the nonlinear converted signal may propagate in a direction that is substantially perpendicular or transverse to the propagation direction of the optical signal. The propagation direction of the optical signal may be along the poling direction, that is perpendicular to the plane of modulation / plane in which the reciprocal grating vectors, g, lie. The solid state nonlinear material may be for example lithium niobate (LiNbO₃) and lithium tantalate (LiTaO₃), potassium tantalate niobate (KTN) or strontium barium niobate (SBN), KTP (KTiOPO₄), KTiOAsO₄ (KTA), RbTiOAsO₄ (RTA), RTP crystal (rubidium titanyl phosphate or RTiOPO₄), CsTiOAsO₄ (*CTA*), PbTiO₃, BaTiO₃, Pb₅Ge₃O₁₁, or, in alternate arrangements, the nonlinear medium may be another suitable solid state nonlinear medium capable of QPM of an optical signal in a direction that is at least oblique to the propagation direction of the optical signal, or substantially perpendicularly or transversely to the propagation direction of the optical signal

[ 0025 ] An optical detector may be located on the plane of the third frequency converted optical signal for detection thereof. In some arrangements, the second optical signal may be an optical signal with known characteristics for comparison with the first optical signal. In other arrangements, the second optical signal may have unknown optical characteristics. In further arrangements still, the second optical signal may be derived from the first optical signal, or alternately may be a portion of the first optical signal. The apparatus may be configured such that the first optical signal is split into two portions and generates both the first and second frequency converted optical signals. In the region in the nonlinear medium where the first and second optical signals spatially overlap, i.e. in this situation the two portions of the first optical signal, the first optical signal interacts with itself within the nonlinear medium to generate the third frequency converted signal which is detected by the detector.

[ 0026 ] In further arrangements, the apparatus may include a filter intermediate the nonlinear medium and the detector. As described above, where the first and second optical signals have different frequencies, the third frequency converted signal is generated via sum frequency generation. Thus, the third frequency converted signal may be at a different frequency to that of both the first and second optical signals, and also the third frequency converted signal is at a different frequency to that of both the first and second frequency converted signals. The filter may be configured to pass only the third frequency converted signal to the detector, i.e. by selectively blocking at least the first and second frequency converted signals and maybe also blocking any scattered, stray or residual light at the frequency of the first and/or the second optical signals. Thereby, the detector may detect only the third frequency converted signal free from any background radiation deriving from either the first or second optical signal, or the first or second frequency converted optical signals. [ 0027 ] Using the methods disclosed herein, spatial and/or temporal parameters of the first optical signal may be measured and analysed. These parameters may include at least the temporal profile of the first signal and/or the spatial distribution of the first optical signal. The spatial distribution of the third frequency converted optical signal may be analysed to obtain temporal information about the pulse(s) of the first optical signal. [ 0028 ] It will be appreciated by the skilled addressee that, by utilisation of the methods and apparatus disclosed herein, an improved technique for obtaining an autocorrelation measurement of a pulsed optical signal of interest, and/or alternatively a cross-correlation measurement of a pulsed optical signal is provided. In particular, compared to TPF for example, the correlation measurement with transverse phase-matched nonlinear optical techniques disclosed herein may require one, two, three, four, five or more orders of magnitude less power and the spectral bandwidth of the technique may be limited only by the transparency window of the nonlinear material. It will further be appreciated that the disclosed methods and apparatus provide for a cross correlation technique which is not limited to use of a second pulsed optical signal which has the same frequency as that of the pulsed optical signal to be measured, i.e. the present methods and apparatus provide a cross-correlation technique capable of using two different frequencies, which in turn is capable of providing a background free cross-correlation technique.

[ 0029 ] According to a first aspect there is provided an optical system for detection and analysis of a first optical signal pulse. The system may comprise a nonlinear material having a ferroelectric domain structure, the nonlinear material capable of converting first and second optical signals respectively to first and second frequency-converted optical signals. The system may further comprise alignment means for respectively aligning the first and second optical signals such that they propagate collinearly, but in opposite directions, through the nonlinear medium to obtain a overlap region in the nonlinear material where the first and second optical signals overlap. The nonlinear material may be capable of converting the first and second optical signals to a third frequency converted optical signal in the overlap region. The third optical frequency generated by the nonlinear material may propagate in a direction that is either oblique or transverse to the propagation direction of both the first and second optical signals. The optical system may also include an optical detector for detection of the third frequency converted optical signals. The detector may also detect the first and second frequency converted optical signals.

[ 0030 ] The optical system may comprise a nonlinear material having a ferroelectric domain structure, the nonlinear material capable of converting first and second optical signals respectively to first and second frequency-converted optical signals; and alignment means for respectively aligning the first and second optical signals such that they propagate collinearly, but in opposite directions, through the nonlinear medium to obtain a overlap region in the nonlinear material where the first and second optical signals overlap, wherein the nonlinear material being capable of converting the first and second optical signals to a third frequency converted optical signal in the overlap region; wherein the third optical frequency generated by the nonlinear material propagates in a direction that is either oblique or transverse to the propagation direction of both the first and second optical signals. In other arrangements, the optical system may comprise a nonlinear material having a ferroelectric domain structure, the nonlinear material capable of converting first and second optical signals respectively to first and second frequency- converted optical signals, and the nonlinear material capable of converting the first and second optical signals to a third frequency converted optical signal in a region in the nonlinear material where the first and second optical signals overlap, wherein the third optical frequency generated by the nonlinear material propagates in a direction that is either oblique or transverse to the propagation direction of both the first and second optical signals. [ 0031 ] The system may be configurable such that the first and second frequency converted optical signals propagate in all directions in a plane that includes the propagation direction of the optical signal pulse. The third frequency converted optical signal may be a correlation signal between the first and the second optical signal pulses. [ 0032 ] The optical characteristics of the second optical signal pulse may be known, wherein the system may generate a cross-correlation signal which may be detected by the detector for analysis of the optical signal pulse. The known optical characteristics of the second optical signal pulse may include the temporal characteristics such as the pulse width, the pulse shape, or in the case of a pulse train, the time between successive pulses. The known optical characteristics may also include the spectral characteristics such as the frequency/wavelength of the pulse or the spectral profile of the pulse.

[ 0033 ] The optical characteristics of the second optical signal pulse, for example the pulse width, pulse shape , or the time between successive pulses in a pulse train, may be unknown. The second optical signal pulse may be derived from the first optical signal pulse. The first optical signal pulse may be divided into first and second spatially separated optical signal pulse portions each pulse portion having characteristics corresponding to the characteristics of the first optical signal pulse, the first spatially separated optical signal pulse portion forming the second optical signal pulse for analysing the second spatially separated optical signal pulse portion, from which an analysis of the first optical signal pulse is obtained.

[ 0034 ] The nonlinear material may be oriented such that the first and the second optical signal pulses each propagate through the nonlinear medium substantially perpendicular to the optic axis of the nonlinear material. The first and second optical signal pulses may each be linearly polarised substantially parallel to the optic axis of the nonlinear material. In another arrangement, the first and the second optical signal pulses each propagate through the nonlinear medium substantially along the optic axis of the nonlinear material.

[ 0035 ] The first and the second frequency converted optical signals may respectively be the second harmonic of the first and second optical signal pulses. The third frequency converted optical signal may be the second harmonic of the first and the second optical signal pulses (where the frequency of the optical signal pulse and the second optical signal pulse are equal) or the third frequency converted optical signal may be the sum frequency of the first and second optical signal pulses (where the optical signal pulse and the second optical signal pulse have unequal frequencies).

[ 0036 ] The nonlinear medium may be a solid state nonlinear crystal and may be a ferroelectric nonlinear crystal. The nonlinear medium may have randomly distributed and sized ferroelectric domain structures. The nonlinear medium may be lithium niobate (LiNbO₃) and lithium tantalate (LiTaO₃) , potassium tantalate niobate (KTN), PbTiO₃, BaTiO₃, Pb₅Ge₃O₁₁, KTP (KTiOPO₄ ), KTiOAsO₄ (KTA), RbTiOAsO₄ (RTA) , RTP crystal (rubidium titanyl phosphate or RTiOPO₄), CsTiOAsO₄ (*CTA*), or strontium barium niobate (SBN).

[ 0037 ] According to a second aspect, there is provided a method of analysing a first optical signal pulse comprising the steps of: a) directing the first optical signal pulse to a nonlinear material having a ferroelectric domain structure with a desired propagation direction with respect to the nonlinear material;

b) directing a second optical signal pulse to the nonlinear material, the propagation direction of the second optical signal pulse being approximately collinear with and opposite to the propagation direction of the first optical signal pulse, such that a spatial location in the nonlinear material exists where the second optical signal pulse overlaps with the first optical signal pulse;

c) in the nonlinear material at the spatial overlap location, converting a portion of the first optical signal pulse and a portion of the second optical signal pulse to a frequency converted optical signal;

d) detecting the frequency converted optical signal with an optical detector.

[ 0038 ] The method may further include analysing the detected signal and associating the detected signal with the first optical signal for analysis thereof.

[ 0039 ] The frequency converted optical signal may correspond to a correlation between the first optical signal pulse and the second optical signal pulse. The correlation may be a cross correlation of the first optical signal pulse with the second optical signal pulse of known characteristics, or the correlation may be an autocorrelation of the first optical signal pulse wherein the second optical signal pulse is derived from the first optical signal pulse.

[ 0040 ] The second optical signal pulse may be derived from the first optical signal pulse and the correlation may be an autocorrelation of the first optical signal pulse. [ 0041 ] According to a third aspect, there is provided a method of analysing a first optical signal pulse using second harmonic or sum frequency mixed light generated by the first optical signal pulse and a second optical signal pulse.

[ 0042 ] According to a fourth aspect, there is provided a crystal having a random ferroelectric domain structure when used for detection of an optical signal pulse using second harmonic light. [ 0043 ] According to a fifth aspect, there is provided a crosscorrelator comprising a nonlinear material having a ferroelectric domain structure. The ferroelectric domain structure may be random.

[ 0044 ] According to a sixth aspect, there is provided an autocorrelator comprising a nonlinear material having a ferroelectric domain structure. The ferroelectric domain structure may be random.

[ 0045 ] According to a seventh aspect, there is provided a method of obtaining an optical cross-correlation or an optical autocorrelation using a nonlinear material having a ferroelectric domain structure. The ferroelectric domain structure may be random. [ 0046 ] According to an eighth aspect there is provided an optical system for analysing an optical signal pulse comprising: a nonlinear material having a ferroelectric domain structure, the optical signal pulse being alignable to be incident thereon, the nonlinear material capable of frequency converting the optical signal pulse to a first frequency-converted optical signal; generation means for generating a optical analyser pulse from the optical signal pulse; alignment means for aligning the optical analyser pulse to be incident on the nonlinear material such that it propagates through the nonlinear material approximately collinearly to, although in the opposite direction of, the propagation direction of the first optical signal pulse, such that the optical signal pulse and the optical analyser pulse overlap in an overlap region within the nonlinear material; wherein, in the overlap region, the nonlinear material is capable of converting the signal and analyser optical signals to a second frequency converted optical signal, such that the second frequency converted optical signal propagates in an oblique or substantially transverse direction to the propagation direction of the signal and analyser optical signals.

[ 0047 ] According to a ninth aspect there is provided an optical system for detection and analysis of a first optical signal pulse comprising: a first optical signal pulse source for generation of a first optical signal pulse; a nonlinear material having a ferroelectric domain structure, the first optical signal pulse being alignable to be incident thereon, the nonlinear material capable of converting the optical signal pulse to a first frequency-converted optical signal; a second optical signal pulse source for generation of a second optical signal pulse to analyse the first optical signal pulse, the second optical signal pulse being alignable such that it propagates through the nonlinear material approximately collinearly to, although in the opposite direction of, the propagation direction of the first optical signal pulse, and the nonlinear material capable of converting the second optical signal pulse to a second frequency converted optical signal, and at a location within the nonlinear material where the first and second optical signals overlap, the nonlinear material being capable of converting the first and second optical signals to a third frequency converted optical signal; and an optical detector for detection of the third frequency converted optical signal.

BRIEF DESCRIPTION OF THE DRAWINGS [ 0048 ] A preferred embodiment of the present invention will now be described, by way of an example only, with reference to the accompanying drawings wherein:

[ 0049 ] Figure 1 depicts nonlinear signal generation for second harmonic generation with a plane shaped emission from a nonlinear ferroelectric crystal with randomly sized and spatially distributed ferroelectric domains when the fundamental incident radiation is aligned perpendicular or near-perpendicular to the c-axis of the crystal;

[ 0050 ] Figure 2 is a schematic representation of the transverse SH emission of counterpropagating pulses: the SH of both UDFD and CPFD are emitted in the X-Y plane (arrows indicate the transverse direction of observation)

[ 0051 ] Figure 3 shows an experimentally obtained image of the transverse SH emission of Figure 2 and also the corresponding quasi-phase-matching conditions in each region of the nonlinear material (i.e. unidirectional frequency doubling (UDFD) AA-S and BB-S and counterpropagating frequency doubling (CPFD) AB-S);

[ 0052 ] Figures 4A and 4B are schematic depictions of an arrangement for detection and analysis of two approximately collinearly aligned and oppositely propagating pulses in a nonlinear medium with randomly sized and spatially distributed ferroelectric domains;

[ 0053 ] Figure 5 is a schematic depiction of the interaction of a first optical signal in the form of a signal pulse train with a second optical signal pulse in a nonlinear medium with ferroelectric domains;

[ 0054 ] Figure 6 is an example of timing diagram for the signal pulse train of Figure 3, showing possible pulse distortions or timing discrepancies from an optimal signal pulse train;

[ 0055 ] Figures 7A to 7C are pulse timing diagrams of the signal pulse train of Figure 3 and a train of second optical signal pulses showing the analysis of the pulses in the signal pulse train;

[ 0056 ] Figure 8 is an alternative arrangement of a crosscorrelation setup utilising fibre optic components; [ 0057 ] Figures 9A and 9B show example arrangements of an autocorrelation setup using a nonlinear medium capable of transverse frequency conversion (eg. transverse second harmonic generation);

[ 0058 ] Figure 10 is an experimentally obtained image of a transverse frequency converted signal generated in a nonlinear medium arising from two approximately collinearly aligned and oppositely propagating pulses;

[ 0059 ] Figures HA and HB respectively show a camera image of an autocorrelation signal as per Figure 10; and a comparison between the measured autocorrelation signal (dots) and an autocorrelation signal of the same pulse obtained using the GRENOUILLE method (solid line); [ 0060 ] Figures 12A and 12B respectively show a further image of an autocorrelation signal as per Figure 1;, and a comparison between the measured autocorrelation signal and an autocorrelation signal of the same pulse obtained using the GRENOUILLE method;

[ 0061 ] Figures 13A and 13B respectively show an autocorrelation image; and phase and transverse profile of the pulse used in Figures 12 A and 12B; [ 0062 ] Figures 14 A and 14B respectively show the setup used to impart a pulse front tilt on an optical pulse; and an autocorrelation image of the tilted pulse obtained with the system of Figure 9B;

[ 0063 ] Figures 15A and 15B respectively show the setup used to generate a double pulse comprising two identical pulses separated by 989 fs from a single pulse; and an autocorrelation image of the double pulse obtained with the system of Figure 9B;

[ 0064 ] Figure 16 is an alternate arrangement of the autocorrelation setup of Figures 9A and 9B utilising fibre optic components;

[ 0065 ] Figures 17A to 17C respectively are schematic representations of cone SH generation in a nonlinear medium for a pulse propagating along the optic axis of the medium; [ 0066 ] Figure 18 is a schematic representation of toroidal-wave SH generation in a nonlinear medium for a pulse propagating along the optic axis of the medium;

[ 0067 ] Figure 19 is a further schematic of generation of transverse SH signal with two counter-propagating beams and corresponding experimentally observed SH signal (emitted from a nonlinear medium with random domain structure such as SBN) as seen on the rectangular screen around the crystal (the two outer traces are the conical waves and the weaker central line represents the transverse SH emission); [ 0068 ] Figure 20 is experimentally observed transverse SH emission (central trace) and conical waves generated in a nonlinear material with annular domains;

[ 0069 ] Figures 21A to 21C are plots of the polarization characteristics of SHG in SBN (Figure 21A) and annularly poled SLT (Figures 21B and 21C); and [ 0070 ] Figure 22 is an example image of a background-free autocorrelation trace from counterpropagating pulses along the Z-axis of an SBN crystal.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[ 0071 ] Materials having ferroelectric domain structures have emerged as a new class of nonlinear optical crystals which enable quasi-phase-matching (QPM) of any parametric process, e.g., SHG or sum-frequency mixing (SFM) in an ultrabroad frequency range. One example of such a medium is the unpoled grown strontium barium niobate (Sr_xBa₁-_XNb₂O₆ , referred to hereafter as SBN) crystal although others exist and are equally suitable. SBN has domain structures which are random in both size and orientation thus, without the necessity of poling, SBN contains a distribution of domains sized in the range of 1 - 8 μm, providing an infinite set of grating vectors g for QPM of parametric processes. These domains are anti-parallel orientated. Thus the limitations of the bandwidth are mainly given by the transparency window of the crystal in the range of 0.4 — 6 μm. This is of special importance for ultrashort pulses, for which the full spectrum is converted with almost constant efficiency.

[ 0072 ] Figure 1 shows a schematic depiction of a nonlinear process in a nonlinear ferroelectric crystal with randomly sized and spatially distributed ferroelectric domains. An optical signal 2, having a frequency ω_1} is incident on the nonlinear medium 4 such that the propagation direction of the optical signal 2 is aligned perpendicularly or near-perpendicularly to the c-axis (also known as the optic axis) of the medium. The nonlinear medium 4 converts a small portion of the optical signal 2 via second harmonic generation to a frequency converted optical signal 6 with frequency GO₁ ^* = 2CO₁. The frequency converted signal 6 radiates in the plane perpendicular the optic axis of the nonlinear medium 4. An optical detector 8 may then be used to detect the second optical signal 6 without the need to isolate it from the first optical signal 2.

[ 0073 ] The same effect may be achievable in periodically poled nonlinear media, where a periodic modulation of the nonlinearity (which has a quadratic dependence on the irradiance of the optical signal(s)) is designed to enhance the efficiency of parametric processes such as SHG or SFM by QPM for a specific frequency. Typically this is done such that the frequency converted signal propagates in the same direction as the input beam, however the periodic domains can be configured to achieve oblique or even transverse propagation of the frequency converted beam such as, for example, using an annularly poled nonlinear material as described below. The drawback of using a periodically poled nonlinear material is that the poling is specifically designed for a particular wavelength (or possibly even a small wavelength range of the order of a few nanometres), although the nonlinear conversion efficiency can be increased to enhance the signal-to-noise ratio of the detected signal which may be advantageous for the particular wavelength or range of wavelengths of interest.

[ 0074 ] Disordered structures of the type exemplified in Figure 1 on the other hand do not only allow for phase-matching a far broader frequency range but also enable to observe SHG and/or SFM in a direction that is not coincident with the input beam(s). In particular arrangements, the SHG/SFM may be observed in a direction which is perpendicular to the fundamental input beam(s). Typical values of the nonlinear conversion efficiency for nonlinear materials with disordered domains is of the order of a fraction of a percent i.e. approximately 0.02% to 0.2%. Without any additional poling of the ferroelectric domains of the nonlinear medium, the limitations of the bandwidth are mainly given by the transparency window of the crystal, which ranges between about 0.4 to 6 μm for SBN. Since the second harmonic signal must also lie within the transparency window of the crystal, this implies that SHG can be achieved in SBN for fundamental wavelengths in the range of approximately 0.8 to 6 μm. The random structure of the domains provides a broad and smooth spectrum of reciprocal gratings in the Fourier space that can provide phase matching of nonlinear processes as second harmonic generation and other second order processes for an equally broad range of fundamental wavelengths. Also, since the phase matching condition is practically equal everywhere in the crystal, the system is highly insensitive against misalignment. [ 0075 ] The optical detection system disclosed herein is configured for analysis of at least one optical pulse, a first optical signal pulse having a first fundamental frequency Q₁ , by correlating that first optical signal with a second optical signal pulse having a second fundamental frequency CO₂. The first and the second optical signal pulses may be derived from a single pulse and thus includes the situation where CO₁ = CO₂. The first optical signal pulse under analysis and the second optical signal pulse are each incident on the nonlinear material and respective first and second frequency converted signals are generated. The nonlinear material is oriented such that the first and second frequency converted signals propagate radially in a plane that includes the propagation direction of the first and second optical signals respectively, for example the nonlinear material may be oriented such that the first and second optical pulses propagate substantially perpendicular to the optic axis (interchangeably referred to as either the Z-axis, the c-axis, or the optical axis) of the nonlinear material. In other arrangements as described below, the nonlinear material may be oriented such that the pulses propagate substantially parallel to the optical axis of the nonlinear material. The first and second optical signal pulses are incident on the nonlinear material such that they are propagating substantially collinearly, although in opposite directions through the material, such that there is a region in the material where the first and second optical signals are spatially overlapping. In this region of spatial overlap, the first and second signals cooperatively generate a third frequency converted optical signal via a nonlinear optical process. The system is configured such that at least a portion of the third frequency converted optical signal propagates in a direction that is not collinear with either the first or the second optical signals. The third frequency converted optical signal may propagate radially in a plane perpendicular (or normal) to the c-axis of the nonlinear medium. The second optical signal pulse may be either derived from a different optical pulse source to that of the first optical signal — in which case the correlation between the two signals is known as a crosscorrelation.

[ 0076 ] In other arrangements, the second optical signal may be derived from the first optical signal itself, and the first optical signal may be split into two optical signals by beam splitting methods commonly known in the art and the system configured such that the two portions of the first optical signal propagate collinearly in opposite directions through the nonlinear medium such that they interact to generate the third frequency converted optical signal. This situation is known in the art as an autocorrelation.

[ 0077 ] The phase-matching conditions for SHG differ significantly when the SH photon is generated by two photons from the same pulse as in unidirectional frequency doubling (UDFD) (k₂ = 2k] + g) and when the fundamental photons originate from oppositely propagating pulses as in counterpropagating frequency doubling (CPFD) (k₂ = ki - ki + g')- hi both cases, the QPM in the crystal allows for SHG for all spectral components of the pulse in all spatial directions within a plane perpendicular to the z-axis of the nonlinear material.

[ 0078 ] Figure 2 shows a schematic of the transverse SH generation of counterpropagating pulses in a nonlinear material 200 with randomly ordered domains. The nonlinear material 200 is a solid state nonlinear SBN crystal. The pulses 201 and 203 are collinear and counterpropagating inside the nonlinear crystal 200. In order to make use of the biggest component of the second order susceptibility tensor χⁱ²⁾ [d£} ], both pulses are preferably extraordinarily polarized and the crystal 200 is aligned according to the crystal axes 202 such that the pulsed propagate normal (perpendicular) to the z-axis of the nonlinear material. The pulses may not be extraordinarily polarized, however, the transverse SH generation will still occur, but with a lower efficiency. [ 0079 ] The nonlinear crystal 200 is placed such that the oppositely propagating pulses 201 and 203 meet in the centre of the crystal. Each incoming pulse generates a second-harmonic signal (via UDFD) (208 and 208) which radiates (206) transverse to both the propagation direction of the pulses and the crystal z-axis. In the region 205 where pulses overlap, the CPFD process generates a strong SH signal 210 which leads to a clear peak on the SH background also propagating transverse to both the propagation direction of the pulses and the crystal z-axis. The emitted SH is extraordinarily polarized.

[ 0080 ] For the practically important case of transverse SHG the relevant phase-matching conditions are depicted in Figure 3. hi the two regions [211 and 212] where the two pulses are propagating independently in opposite directions, the total SH signal is a sum of the signals generated by each UDFD process (AA-S and BB-S). This signal constitutes the background of the autocorrelation trace. In the region where the two counterpropagating pulses meet (205), the total SH signal is stronger because of the additional contribution from the CPFD process (AB-S). It can be clearly seen that for transverse SHG, different grating vectors g (and consequently different grating periods d) are required for UDFD (215 and 217) and CPFD (219) processes. [ 0081 ] While the grating period needed to phase match transverse CPFD is 2d_CPFD Υhis means

that the phasematching period for the UDFD process is V2/2 smaller than that of CPFD and, consequently, the former will involve V2/2 smaller grating periods. In practice, such sub-wavelength size domains are not available in the crystal, thus the frequency conversion processes are phase matched via higher-order QPM processes. As UDFD and CPFD require different grating periods, a different efficiency for each of those processes is expected depending on the availability of appropriate spectral components in the random domain structure. Defining the ratio of the efficiencies of both processes as R = V_UDFDI⁷I_CPFD then the signal to background ratio will be given by P=2/R + 1. For R = I, then the signal-to-background ratio is P = 3 which is exactly the maximum value for the TPF process. In the typical case of SBN, however, the average domain size is ~2.5 μm, and consequently R ~ 1. Thus an improved signal to background ratio of P ~ 3 is expected. [ 0082 ] In the particular configuration shown in Figure 2, when the two counterpropagating signals are identical, the generated normalized signal /_tot(?) is given as I_tot(τ) = l + (2/R)G^{2)(τ), where represents exactly the autocorrelation of the pulse. This

experimental geometry results in a mapping of time (τ) into a space (longitudinal) coordinate (s) via the simple relation τ = s2n_ωjc , where n_m is the SBN refractive index dependent on the optical frequency, and c is the speed of light in vacuum. Hence, the longitudinal distribution of the SH signal represents a time-dependent autocorrelation signal. It is noted that this method is a single- shot technique, and it does not require any variable differential delay during the correlation recording. [ 0083 ] Figures 4A and 4B show a schematic depiction of a generalised optical detection system of the present arrangements. In Figure 4A, two pulsed optical signals (or alternatively, each of the two optical signals being a single pulse) of respective fundamental frequencies Co₁ and CO₂ (i.e. a first optical signal 410 and a second optical signal 412 respectively) are configured such that they propagate collinearly in opposing directions and are directed into a nonlinear crystal 414 which has a ferroelectric domain structure such that QPM is achieved in a transverse direction to the respective signals. Both optical signals generate second-harmonic (SH) signals at frequencies coi* = 2 CO₁ and CO₂* = 2 ω₂ (signals 416 and 418 respectively), which each propagates at least approximately perpendicularly to the direction of propagation of the fundamental optical signals 410 and 412 (more correctly, in the plane including the normal to the z-axis or substantially perpendicular to the z-axis, which includes the propagation direction of the fundamental optical signals). An optical detector 420 is located transverse to the propagation direction of the fundamental signals 410 and 412 for detection of the frequency converted optical signals 416 and 418. Figure 4B shows the situation when the two fundamental optical signals 410 and 412 meet in the nonlinear medium 414 such that they spatially overlap. The nonlinear medium frequency converted the combination of the two signals of frequency coi and co₂ to generate a third frequency converted optical signal 422, propagating perpendicularly to the propagation direction of the fundamental beams and having an optical intensity (irradiance) greater that that of either second harmonic signals 416 or 418. The frequency of the combined frequency converted signal 422 is ω₁₂* = coi + co₂. Therefore, where the two fundamental optical signals are of the same frequency, coi = CO₂, the combined frequency converted signal 422 is the second harmonic (i.e. generated via SHG) of the fundamental frequency ω₁₂* = 2CQ₁. Alternatively, where the frequencies of the fundamental optical signals are not equal, coi ≠ ω₂, combined frequency converted signal 422 is the frequency mixed signal (i.e. generated via SFM) of the two fundamental beams.

[ 0084 ] The combined frequency converted response signal 422 represents a correlation of the two pulses from optical signals 410 and 412 - either an autocorrelation in the case of the two pulses originating from the same source or a cross-correlation for pulses from different sources. Cross-correlations using reference pulses with known properties are used in order to gain more information about the signal pulse. In cross-correlation measurements, the second optical signal is typically derived from a pulsed source that is controlled by the user. In the situation where the optical signal that is being measured (i.e. the first optical signal) does not have enough power to achieve an appreciably correlation signal from the apparatus, the power of the second optical signal may be increased by the user to increase the overall power in the correlation signal received by the detector.

[ 0085 ] Even the smallest de-synchronisation between the two oppositely propagating pulses may be measured from the movement of the peak of the correlation signal 422, enabling an easy quantification of drift and jitter of a signal against a reference source, for example an optical clock. In particular, a de-synchronization between the reference and signal source will lead to a shift of the correlation peak whereas a pulse jitter will result in broadening and flattening of the correlation trace.

[ 0086 ] The reference pulse does not need to have the same carrier frequency as the signal pulse, since the parametric process (in this case SFM) is always phase-matched. It will be appreciated that, where the two oppositely propagating pulses have different optical frequencies, the transversely detected correlation signal is at a different optical frequency to the fundamental frequencies of the two input pulses, and also to the second harmonic signals of each of the pulses generated in the nonlinear medium wherever the pulses do not interact. Therefore, the correlation signal is able to be spectrally filtered from the other optical signals in the apparatus, thus enabling background free detection of the correlation signal.

[ 0087 ] This approach disclosed in the above methods and apparatus makes the need for a variable delay to perform autocorrelation or crosscorrelation measurements of optical pulses obsolete, thus removing an alignment- and synchronisation-critical component of standard correlation methods and providing a technique capable of a single shot operation. The spectral bandwidth is practically only limited by the transparency window of the nonlinear material (for SBN the transparency window ranges between approximately 0.4—5.0 μm) and no angular alignment is required. The advantages of the presented technique show it to be a preferable alternative for characterising ultra-short pulses.

[ 0088 ] Examples of both crosscorrelation and autocorrelation applications of the present optical detection system are disclosed below. In the examples below, a femtosecond Ti:sapphire oscillator was used having pulse energies of 6 nJ, a repetition rate of 76 MHz, and tunability in the range of 700-900 nm. For the results presented in herein, the laser was set to a wavelength of 850 nm.

EXAMPLES

Example 1: Cross-correlation - Pulse Train Analysis

[ 0089 ] The device of the optical detector disclosed above may be used for analysis of an optical signal beam comprising a train of optical pulses, for example such as that found in optical communications networks and systems by combining the signal beam in a nonlinear medium with random ferroelectric domains with a counterpropagating pulsed reference beam with known characteristics, e.g., generated by an optical clock such that the SH signal represents a cross correlation of the reference and signal pulses. Parallel with the increasing density of transmitted data in optical networks, the tolerance to signal distortions decreases and quality monitoring is required to ensure the functionality of the network. In order to achieve ever increasingly higher bit rates, modern telecommunication systems optically interleave multiple signals in an optical time division multiplexing (OTDM) circuit. Due to the ability of the random ferroelectric domain structure to phase match parametric processes over an ultrabroad frequency range, the reference signal does not need to have the same carrier frequency as the measured signal. To the contrary, different carrier frequencies of both pulse trains allow for the two background SH signals to be filtered out from the sum frequency correlation signal, resulting in a background- free correlation measurement. [ 0090 ] As shown in Figure 5, the arrangement of Figure 4A can be modified to provide independent pulse quality measurements and analysis of the interleaved channels of an OTDM signal using less complex and cost-effective components. The fundamental beam 530 which is desired to be analysed is depicted as having a train of optical pulses. When the optical pulses propagate through a nonlinear medium such a SBN, a corresponding optical signal 532 at the second harmonic is generated. The SH generated signals are depicted only in the transverse direction to the propagation direction of the fundamental pulses, however it will be appreciated that the second harmonic signal from each respective pulse also propagates radially in the plane of the optic axis of the nonlinear material 534 (i.e. as depicted in Figure 1 above). The fundamental beam 530 is mixed in the nonlinear SBN crystal 534 (or other similar nonlinear material with random domain structures) with a counter-propagating second optical signal 536. The second optical signal 536 may be any form of optical pulse about which its optical parameters are known for comparison against the unknown quality of the pulses in the pulse train. An optical detector 538, depicted in the present arrangement as a CCD line array, is used to detect the optical power of the SH wavelength generated perpendicularly to the propagation direction of both the fundamental beam 530 and the clock signal 536. As has been described above, whenever a pulse from the second optical signal 536 interacts with a pulse from the pulse train 530, the generated nonlinear frequency signal 540 is a cross-correlation of the two interacting pulses which is detected on the detector 538. Also, where the clock pulse is of a different frequency to that of the pulse in the OTDM signal, the correlation signal can be detected with effectively no background interfering radiation. It will be appreciated that, where the OTDM signal contains optical pulses of difference optical frequencies, for example in the case of a wavelength division multiplexed (WDM) signal, the frequency of the correlation signal resulting from each pulse at each different wavelength will have a characteristic frequency which can be used to analyse or monitor a particular wavelength channel in the WDM signal. [ 0091 ] The present arrangement further enables the measurement of the de-synchronization of the combined signal as well as between the different channels and thus makes it possible to recover the initial signal by suitable adjustment of the clock signal input.

[ 0092 ] As shown in Figures 7A to 1C, the input signal 530 comprises a plurality of optical pulses 542 and the second optical signal 544 is derived from an optical clock which generates a plurality of optical pulses 546 with known characteristics and a stable spacing in the time domain. In the present example, the optical input signal 530 under analysis comprises a group of 5 channels 542a to 542d, each separated by 160 GHz, and the optical clock signal 544 is a train of optical pulses each separated by 40 GHz. In the present example for simplicity second optical signals 530 and 544 respectively are treated as having the same optical frequency CO₁. [ 0093 ] Where two counter propagating pulses overlap in the nonlinear medium 534 (SBN in the present example) i.e. pulses 542c and 546a in Figure 7B, a correlation signal 548 is generated the local intensity of the radiation is increased, which leads to the generation of a peak (548 of Figure 7C) in the SH radiation by the SBN crystal at the (spatial) point of the intersection within the crystal . Thus the position of the intersection appears as a peak in the SH signal against a constant floor (550 of Figure7C) generated by the two pulses travelling outside the overlap region. The height and shape of the peak is a direct measurement of the pulse quality and any distortion such as due to dispersion can easily be detected.

[ 0094 ] Furthermore, since the timing of the counter propagating pulse train is known (i.e. the optical clock signal 544), the position of the peak on the detector 538 additionally reveals the timing information of the incoming probe signal therefore translating the time information of the signal pulse train into spatial information that can be recorded by the detector above the crystal. This information can then be used to determine whether there is any distortion of any pulses in the fundamental pulse train. Figure 6 shows a comparison of the SH output recorded on the CCD line array detector between an optimal signal 552 and a distorted one 554 as they would be recorded spatially on the detector 538 (not shown). The broader and lower peak for channel "a" (556a) is indicative of performance losses due to dispersion, while the misplacement of channel "c" (565a) is indicative of a delay of the channel relative to the optimum signal 552. [ 0095 ] Taking the example of an OTDM systems with 4 interleaved channels of 40 GHz each, resulting in a total bandwidth of 160 GHz. The separation of two peaks in the SH signal on the detector can be calculated as s = C₀ / (2.n.f), where CQ is the speed of light in vacuum, n is the refractive index of the nonlinear material (2.3 for SBN) and / is the pulse frequency. For the system described above this would give a spatial separation on the detector 538 of 0.4 mm between adjacent pulses in the signal beam. Assuming the detector 538 is a CCD line array detector with a pixel pitch of 10 μm in the array, the separation between adjacent pulses would be approximately 40 pixels. For example, a CCD line array detectors having 512 pixels would easily allow measurement of each of the 4 interleaved channels twice with a single clock pulse, thus allowing for the reduction of noise in the detected signal. The device size is determined by the dimensions of the CCD detector. Since even fast CCD detectors usually have a frame rate of less than 100 kHz, a typical device would enable the integration of approximately half a million pulses per channel to further improve the signal to noise ratio.

[ 0096 ] Due to the feature of broadband second harmonic generation as disclosed above with nonlinear crystals such as SBN and the fact that the system is able to be configured such that there is negligible distortion of loss to the fundamental signal beam, this device is especially suitable for optical communications systems having large numbers of optical channels such as WDM and/or OTDM systems.

[ 0097 ] Figure 8 shows an example configuration of a cross-correlator according to the above using fibre optical components. The first optical signal pulse 660 to be analysed, propagating in optical fibre 661 propagates through optical multiplexer/demultiplexer 662 into optical fibre 663. The pulse then enters a correlation module 664, the correlation module having input ports 665 and 666, which in this arrangement are standard fibre coupling ports. A second optical signal pulse 667 form a pulse source (not shown) that is controllable by the user enters multiplexer/demultiplexer 668 from fibre 669 and is transmitted via fibre 670 to input port 666. [ 0098 ] Once the pulses 660 and 667 enter the correlation module 664 they are collimated by respective lenses 671 and 672 and directed to a nonlinear material 673 having the ability to broadband phase-match both the first and second optical signals 660 and 667 in the transverse direction, for example SBN. The nonlinear material 673 generates via SHG transversely propagating first and second frequency converted optical signals (not shown) in response to the first and second optical signals. The input ports 665 and 666 and collimating lenses 671 and 672 are configured such that the first and second input pulses 660 and 667 propagate substantially collinearly and in opposite directions through the nonlinear material 673, such that they overlap spatially in the nonlinear material 673-

[ 0099 ] Where the first and second signal pulses 660 and 667 overlap spatially in the nonlinear material a third frequency converted optical signal 674 is generated. The third frequency converted signal propagates transversely to the propagation directions of pulses 660 and 667 through the correlation module 673. The correlation module 664 also has an optically transparent window 675 for transmission of the third frequency converted signal 74 to a detector

676. The window 675 may also include an optical filter (not shown) to block any residual light at the wavelength of the first and second frequency converted optical signals, and the frequencies of the first and second signals 660 and 667 if required, to achieve background-free detection of the third frequency converted signal 674. The detector 676 may be a camera, CCD array or other suitable detector. Once pulse 660 (respectively for pulse 667) exits the nonlinear material 673, it is collected by optical fibre 670 (respectively fibre 663), and exits the system though multiplexer/demultiplexer 668 (respectively multiplexer/demultiplexer 662) onto optical fibre

677 (respectively fibre 678).

Example 2: Optical Autocorrelator

[ 00100 ] Figure 9 A shows a schematic arrangement of an example optical setup for analysis of an optical pulse using an autocorrelation. The pulse 780 (propagating in direction 781) to be measured (having an optical frequency Q₁) is directed to a beam splitter 782 which splits the pulse into two spatially separated pulses, a first optical signal pulse 783 and a second optical signal pulse 784, each propagating in directions 785 and 786 respectively. The first and second spatially separated pulses are directed using turning mirrors 787 and 788 to a nonlinear material 790 with the ability to broadband phase-match (i.e. via QPM) in the transverse direction, for example SBN. The propagation direction of the two spatially separated pulses is configured such that they are propagating approximately collinearly, although in opposite directions, through the nonlinear material 790. The pulses may be optionally focussed or collimated prior to the SBN crystal using lenses 791 and 792 if desired. Both the first and second optical pulses each cause the SBN crystal 790 to generate respective first and second frequency converted optical signals 793 and 794, each with frequency ω₂ = 2ω\ via SHG. The SH light is directed using a lens 796 to a detector 797 (eg. a camera, CCD array, or other spatially sensitive detector) for detection.

[ 00101 ] When the first and second optical pulses 783 and 784 are spatially overlapping in the SBN crystal 790, they interact and a correlation signal 795 is generated. The correlation signal is seen by the detector 797 as a clear peak in the frequency converted SH signal generated by the SBN. The measurement accuracy is primarily limited by the imaging of the correlation trace onto the camera. For recording the autocorrelation pulses in the present example, an optical microscope at 4.5 magnification was used. To avoid an image distortion caused by the limited depth of field, the beams were focused in the crystal by cylindrical lenses (f= 50 mm), resulting in a beam width of 33 μm along the observation direction.

[ 00102 ] Figure 9B shows an alternate arrangement of the autocorrelation setup of Figure 9 A in the form of a Mach-Zender interferometer. (Figure 9A is viewed from the top, such that the autocorrelation signal generated in the nonlinear material is shown as it would appear on the detector.)

[ 00103 ] Figure 10 shows an example of an experimentally detected SH signal generated in the SBN crystal as two optical pulses propagate through the crystal. In the present experiment, the pulse to be analysed had a wavelength of 850 ran and, after splitting of the pulse, each of the two spatially separated pulses 783 and 784, propagating approximately collinearly and in opposing directions, had an average power of 96 mW at a repetition rate of 76 MHz. Measurements with an uncooled CCD camera were performed at power levels down to 0.26 MW/cm², four orders of magnitude lower than those usually used with TPF. This significantly lower power requirement is primarily due to the confinement of the emitted SH in a plane compared to the emission in the full solid angle from the TPF process. [ 00104 ] The SH signal 793 and 794 from each of the pulses 783 and 784 respectively is seen in the image of Figure 10 where the two pulses meet in the SBN crystal 780, as is the clear peak 798 in the SH signal, being the correlation signal 795 of the pulse.. Analysis of the peak can then be performed to determine the duration of the original pulse 780. [ 00105 ] Figure HA shows an image of the autocorrelation signal 795 obtained in a nonlinear medium 790 as described above. Figure HB shows a graph of the detected points 798 taken from the measurement image of Figure HA representing the intensity of the SH autocorrelation signal in arbitrary units 0 to 1. The signal to background ratio of the measured autocorrelation signal 795 of about 3 to 1 (is as seen in the graph of Figure HB around +400 and -400 fs with an intensity of approximately 0.3) is an excellent result and is in line with the theoretical maximum signal to background ratio obtainable in the SH autocorrelation signal.

[ 00106 ] Figures 12 A and 12B show a further autocorrelation measurement of a femtosecond pulse. This pulse duration was measured with a GRENOUILLE (Figures 13A and 13B), obtaining a reference measurement 760 for comparison with the SH autocorrelation trace 762 obtained with the present system. In a series of measurements, the two methods were found to disagree by no more than 10%, including the dispersion of the 190 fs pulse in 2.5 mm of SBN, which accounted for approximately 4.5% of the error. The effect of dispersion can be easily reduced by using shorter crystals. It was noted that the signal-to-background ratio P in this measurement was approximately 4. [ 00107 ] Apart from the pulse duration measurements, the correlator in Figure 9 A or 9B can also be used to visualize unambiguously the tilt of the front of an optical pulse. Such a pulse front tilt (PFT) (which is typically caused by dispersive elements such as prisms, gratings, or wedges) leads to an effective longer pulse duration and hence a lower peak power in the focal plane of the beam. [ 00108 ] Referring to Figure 9B, for measurement of a pulse 800 having a tiled phase front, Figure 9B is viewed from the top, such that the autocorrelation signal generated in the nonlinear material is shown as it would appear on the detector.) The pulse 800 to be measured having a tilt in its pulse front, is split by beam splitter 702 into two first and second spatially separated pulses 804 and 806. The spatially separated pulses 704 and 706 are directed with turning mirrors 808, to 812 and lenses 814 and 816 such they are incident from opposite directions on nonlinear crystal 816. The turning mirrors are configured such that the path lengths of the pulses are equal i.e. there is no delay line required for operation of the system. When the pulses 804 and 806 spatially overlap in the nonlinear crystal 818, an autocorrelation signal 820 is observed by a detector (not shown). The tilt in the autocorrelation signal can then be analysed to obtain the pulse front tilt of the initial signal pulse 800. By placing a 60° (SFl 1 Schott glass) prism (822 of Figure 14A) into the path of the pulse 800, a PFT was introduced that can be clearly seen in the 4° tilt of the correlation trace 830 shown in Figure 14B. [ 00109 ] The large time window of the technique described herein (i.e. ~1 mm SBN is sufficient to monitor 7 ps) allows for monitoring more complex temporal structures consisting of multiple pulses. To demonstrate this feature, a pulse doublet is generated by passing the beam through a thin (3.2 mm) birefringent (lithium niobate) crystal 840 followed by a polarizer 842, as depicted in Figure 15A. The ordinarily polarized component of every pulse gets delayed by roughly 1 ps with respect to its extraordinarily polarized counterpart. The polarizer combines these two components to a pulse doublet. The autocorrelation trace shown Figure 15B clearly resolves the two components of the doublet and allows for the delay between them to be precisely measured. For comparison, in the analogous experiment, GRENOUILLE averages over the two pulses due to its narrow time window, hence the information about the double structure is lost.

[ 00110 ] It will be appreciated that the approach described above renders the need for a variable delay (which is normally required to be added to one of the spatially separated pulse portions) obsolete, thus removing an alignment critical part and all problems in the measurement of the autocorrelation signal associated with synchronization and timing issues between the two pulses, thus providing a significant advancement over the current methods.

[ 00111 ] As shown in arrangement 900 of Figure 16, the arrangement of either Figure 9 A or 9B can be readily transformed into a version using fibre optic elements. The pulse 901 to be analysed, propagating in optical fibre 902 propagates clockwise through optical circulator 903 an into optical fibre 904. The pulse 901 is then split by a fibre coupler 905 into to two pulses 906 and 907 respectively propagating in optical fibres 908 and 909. Fibres 908 and 909 direct the two pulses to two optical input ports 911 and 912 of a correlation module 910. The input ports 911 and 912 may be standard fibre coupling ports, or they may simply be windows which are optically transparent at the frequency of the pulses 906 and 907. Once the pulses 906 and 907 enter the correlation module 910 they are collimated by respective lenses 913 and 914 and directed to a nonlinear material 915 having the ability to broadband phase-match in the transverse direction, such as SBN.

[ 00112 ] The nonlinear material 915 generates a frequency converted optical signal 916 via SHG in response to pulses which is radiated perpendicularly to the propagation directions (917 and 918) of pulses 906 and 907 in the correlation module 910 (note that directions 917 and 918 are collinear, but in opposite directions). The correlation module 210 also has an optically transparent window 919 for transmission of the frequency converted signal 216 to a detector 920. The window 919 may also include an optical filter to block any residual light at the wavelength of the pulses 906 and 907 if required. The detector 920 may be a camera, CCD array or other suitable detector. Once the pulses 906 and 907 exit the nonlinear material 915, they are collected by optical fibres 909 and 908 respectively, recombined onto optical fibre 904 by coupler 905, and propagate again through circulator 903 to exit the system 900 on optical fibre 921, thus not being transmitted back along fibre 902 to the source of the original pulse 901. [ 00113 ] It will be appreciated that the correlation module 910 of system 900 described above can be quite compact, the main limiting factor governing the module size being the nonlinear material 915. In the case where the nonlinear material is an SBN crystal, the correlation module 910 may be as small as 1 x 1 x 5 cm. Compared with other optical autocorrelation methods and apparatus. Such as the FROG or GRENOUILLE, which typically have a footprint of approximately 50 x 50 cm or greater, the present arrangement provides a significant advancement.

[ 00114 ] The correlation module 910 can also be used as a free space module by removal of the fibre couplers at the input ports 911 and 912 (note that the input ports may simply be optically transparent windows and removal of optical fibre couplers would thus be unnecessary) and aligning free-space propagating pulses by normal methods (eg. as shown in Figure 9A using turning mirrors). In this mode of operation , it is also possible to use the correlation module 910 as a cross-correlation module for measurement of a desired optical pulse using a second optical pulse with known characteristics as previously described, simply by directed the two pulses into the module through input ports 911 and 912 respectively (note that synchronisation of the two pulse sources in this case would be required as per a normal cross-correlation measurement).

Example 3: Background-free Optical Autocorrelator

[ 00115 ] In a further arrangement, analysis of two counter-propagating beams may also be performed where the beams each propagate through the non-linear material along the Z-axis (also referred to as the e-axis or optic axis) of the material (as opposed to propagating perpendicular to the z-axis as in the previous examples). This arrangement provides for a background- free autocorrelation of the counter propagating beams. In this case only the process of CPFD gives rise to a plane emission but at a lower efficiency due to the lower value of the governing component d_zyy^ = 0.5d_zzJ-²\ [ 00116 ] As has been shown above (refer to Figure 1) with beams propagating along the X or Y crystalline axis (i.e. normal to the Z- or c-axis), randomly poled nonlinear crystals allow non- collinear SHG that is emitted in all directions (including the transverse direction) in a plane that contains the fundamental wave. However, if the single fundamental beam (pulse) propagates along the Z-axis, it is not possible to fulfil the transverse SHG (TSHG) phase-matching conditions. Volume pure transverse SHG in a plane perpendicular to the Z-directed fundamental beams can, however, be achieved by interaction of two counter-propagating pulses along the optical axis of a quadratic nonlinear crystal. In the arrangement described below, it is observed that the transverse SH emission of a beam propagating along the z-axis is generated in the form of an expanding toroidal wave emitted from the zone in the nonlinear material where the counter-propagating pulses overlap.

[ 00117 ] In this example, two different types of samples with χ^{2) spatial modulation are used: a 5 mm-tick Strontium Barium Niobate (SBN) crystal with naturally disordered domain structure and an annular periodically poled Stoichiometric Lithium Tantalate (SLT). The SLT sample used was 0.49 mm thick and had an annular periodically poled structure in the X-Y plane with a period of 7.5 μm. In contrast to the fixed-period SLT structure, random distribution of anti- parallel ferroelectric domains of the "as-grown" SBN leads to a natural disordered photonic structure as described above. This structural disorder provides almost a continuous set of grating vectors in the X-Y plane of the crystal allowing to phase-match parametric processes over a broad range of frequencies. As previously noted, no TSHG takes place for the fundamental beam propagating along the Z-axis. Instead, each counter-propagating pulse emits continuously a SH signal in the form of a cone, as shown schematically in Figures 17A to 17C. Referring to Figure 17A (and Figure 17C), a pulse 1010 travelling to the right enters a nonlinear medium 1050 with random domain structure. The nonlinear medium is oriented such that the pulse 1010 propagates along the Z-axis (c-axis or optic axis). A portion of the pulse is converted via QPM to give a SH signal which emits from the nonlinear medium in the form of cone 1015. Similarly, in Figure 17B, a pulse 1020 travelling to the left enters the nonlinear medium 1050 and a SH signal is generated which is emitted in the form of cone 1025. Analogous conical emission takes place in the annular periodically poled sample of SLT. The emission angle β of the cone of SH light is defined by relation k₂.cos β = 2kj.

[ 00118 ] For two counterpropagating pulses in the nonlinear medium 1050 as depicted in Figure 18, the SH cones 1015 and 1025 are formed independently. Only when the two pulses are exactly overlapped in the nonlinear material is the transverse SH (TSH) wave 1030 generated and emitted from the material. In such case, the momenta of the two counter-propagating photons cancel out, and the transverse phase-matching can be achieved due to the reciprocal grating vector provided by the nonlinearity modulation. The TSH wave can readlily be detected by a detector 1040 without detection of any of the SH light (1015 or 1025) generated independently by the counterpropagating pulses (1010 and 1020 respectively) thus giving a pure background free detection scheme. The solid lines 1071 represents the effective reciprocal lattice vectors g while the dotted (1072) and the dashed (1073) lines denote the fundamental k_ω and SH fø_® wave-vectors, respectively. Due to these phase-matching restrictions, the SH is emitted only from the region of pulse overlap and only for the duration of the pulse interaction, It is exactly this spatiotemporal overlap that allows for the emission of a spatiotemporal wave of a toroidal shape (1030 of Figure 18). The width and intensity profile of this wave along the Z- direction is determined solely by the temporal correlation of the fundamental pulses, while the width and intensity profile in the transverse (X-Y) direction (propagation direction of the transverse SH emission ) depends on pulse length and the fundamental beams spatial profiles. [ 00119 ] In the present example, pulses from a regenerative Ti:Sapphire amplifier operating at a wavelength of 830 nm are used. The Ti: Sapphire system delivers linearly polarized 165 fs long pulses of energy up to 3 μJ at a repetition rate of 250 kHz. The beam, which has a Gaussian spatial profile, is split in a polarizing beam splitter (not shown) and directed from both sides to a quadratic nonlinear medium 1050 such that the two pulses meet roughly in the center of the sample. A set of λ/2 waveplates (not shown) allows control of the relative powers of both beams and their polarizations. The average beam power is ~ 340 mW. The two beams are loosely focused in the nonlinear medium 1050 with a beam waist of about 160 μm. All facets of the nonlinear material are polished and the emitted SH signal is recorded by a CCD camera (for example detector 1040 of Figure 18). [ 00120 ] The first-order TSHG phase-matching requires very fine grating periods. For the SLT sample, the necessary period is 183 nm for a 830 nm fundamental wave. As the period of the annular grating in the sample used in this example is approximately 7.5 μm, the observed TSHG is thus due to a 41-st order phase-matching which, to the best of the inventors knowledge, is the highest quasi-phasematched (QPM) order in crystals reported so far. Naturally, such a high-order process results in a very low efficiency.

[ 00121 ] The power in the SH signal is quadratically dependent upon the power of the fundamental wave, and this is verified by measuring the SH intensity in a particular single direction. To obtain angularly symmetric TSHG, both counter-propagating beams are focussed exactly at the centre of the annular domain structure. In contrast, the TSHG in SBN does not depend critically on alignment since the phase-matching conditions are the same everywhere in the crystal. Furthermore, since the average domain size of the SBN sample is approximately 2.5 μm, the phase matching order is 14 resulting in higher generation efficiency.

[ 00122 ] As different nonzero values of the χ^{2) components are involved in TSHG in SBN and SLT, the emission diagrams for both structures are also different, hi the case of an OO-E interaction (ordinary polarized fundamental beams and extraordinary polarized SH), the relevant nonzero χ^{2) components in both crystals are d_zxx and d^y = d_zxx. The generated transverse SH emission is polarized along the Z-axis of the crystal and its intensity is constant for all emission directions in the X-Y plane. The SH intensity of the transverse SH emission does, however, depends critically on the polarization of the fundamental waves:

where γi and γ₂ denote the angles of input polarizations for both the beams measured counterclockwise with respect to the X-axis. In contrast to SBN, where the OO-O interaction is impossible, in SLT the relevant χ⁽²⁾ components d_m and d_yxx = -d_yyy allow for the generation of an ordinary SH wave polarized in the X-Y plane with its intensity varying with the emission angle and input polarization directions as

[ 00123 ] Figure 19 shows a three-dimensional schematic of the interaction between two counterpropagating pulses as in Figure 18, and an image of experimental SH generated light in an SBN sample (1050) for 0 < a < π. Also shown in Figure 19 is the phase matching scheme 1032 for the TSH wave and the overlap region 1032 in the nonlinear material 1050 where the TSH wave is generated. Similar experimental results of the SH emission from counterpropagating pulses in the SLT sample a = πis shown in Figure 20. [ 00124 ] Since the samples used in the present example are not cylindrical, it was not possible to measure accurately the angular variations of the intensity of the SH emission. Instead, the SH intensity and polarization properties of the emitted wave along the X- and Y-axis of the crystal vs. the polarization angles γi and γz was measured. The experimental results for both SBN and SLT crystals are shown in Figures 21A to 21C together with theoretical derived curves. The graphs in Figure 21A show the measured SH signal emitted in the SBN crystal (points) as a function of χi for γ₂ = 0 °, 45 ° and 90° (1110, 1112 and 1114 respectively) The agreement with the expression for /j_,e (solid line) of the OO-E interaction is excellent.

[ 00125 ] For the SLT crystal, the polarization dependencies are more complicated due to the simultaneous contribution of both OO-O and OO-E interactions (Figures 2 IB and 21C). Plots in Figure 21B show the dependence of the total intensity of the SH signal on the polarization of the fundamental beams. On the other hand, in the case displayed in Figure 21C, both fundamental beams were either parallel (X direction — circles 1016) or orthogonally (X and Y directions - squares 1018) polarized. The SH signal can then be measured as a function of the angular position (δ) of a polariser (not shown) mounted in front of the CCD camera. For parallel polarized fundamental beams, both OO-E and 00-0 processes contribute to the SH signal. Hence the recorded signal contains both ordinary and extraordinary components and never vanishes. For orthogonally polarized input beams, the SH wave is ordinary polarized (due to OE-O process) and the recorded SH signal vanishes at the angles δ- {-nil, π/2, 3 π/2).

[ 00126 ] A quantitative analysis of the experimental data indicates that the contribution of the OO-E process governed by the d_zyy component is stronger than that of the OO-O process governed by the d_m component since the OO-E process is closer to the exact phase matching condition than the 00-0 process. It is also noted that for the SBN sample, the SH generation process is practically independent of the fundamental wavelength within a broad frequency range. [ 00127 ] Due to the transverse geometry of the parametric interaction, the TSHG signal effectively translates the time coordinate into the space coordinate such that the width of the transverse SH emission in the direction of the Z-axis is exactly the autocorrelation function of the interacting pulses. From a calibrated experimental photo similar to one in Figure 20, the thickness of the toroid wave generated in the SLT crystal was measured to be to be 34 μm, corresponding to about 160 fs assuming secant hyperbolic temporal shape. Since the beam size (i.e. the beam waist) is much bigger than the spatial extent of the pulse, the thickness of the transverse SH emissions in the propagation direction outside the sample is determined by the beam size and in the present example is found to be about 370 μm.

[ 00128 ] In conclusion, in the present arrangement, toroidal second harmonic waves have been generated via interaction of counter-propagating femtosecond pulses in annularly poled SLT structures and SBN crystals with disordered domains. As the thickness of the transverse SH emission is determined by the correlation function of the fundamental pulses, it will be appreciated that this effect can be used as a pure background-free single short-pulse autocorrelator. An example of an pure background-free autocorrelation trace from this system for a femtosecond pulse is seen in Figure 22.

[ 00129 ] It will be appreciated that the methods and apparatus described/illustrated above at least substantially provide an optical detection method and apparatus for analysis of optical pulses, in particular a solid-state pulse correlation scheme based on transverse phase-matched parametric upconversion in disordered nonlinear optical media. The correlation technique can be used to monitor pulse properties such as duration and front tilt as well as to measure timing and synchronization in pulse trains. The corresponding phase-matching conditions for the counterpropagating beams in this random ferroelectric domain structure allows for an operation over an ultrabroad frequency range and an improved signal to background ratio of the correlation trace. Its simplicity and the possibility of a compact integration with standard fibre components make this technique attractive for a wide range of applications.

[ 00130 ] The preferred arrangements of an optical detection method and apparatus described herein, and/or shown in the drawings, are presented by way of example only and are not limiting as to the scope of the invention. Unless otherwise specifically stated, individual aspects and components of the optical detection method and apparatus may be modified, or may have been substituted therefore known equivalents, or as yet unknown substitutes such as may be developed in the future or such as may be found to be acceptable substitutes in the future. The optical detection method and apparatus may also be modified for a variety of applications while remaining within the scope and spirit of the claimed invention, since the range of potential applications is great, and since it is intended that the present optical detection method and apparatus be adaptable to many such variations.

Claims

Claims:

1. An optical system comprising:

a nonlinear material having a ferroelectric domain structure, the nonlinear material capable of converting first and second optical signals respectively to first and second frequency- converted optical signals; and

alignment means for respectively aligning the first and second optical signals such that they propagate collinearly, but in opposite directions, through the nonlinear medium to obtain a overlap region in the nonlinear material where the first and second optical signals overlap, wherein the nonlinear material being capable of converting the first and second optical signals to a third frequency converted optical signal in the overlap region;

wherein the third optical frequency generated by the nonlinear material propagates in a direction that is either oblique or transverse to the propagation direction of both the first and second optical signals.

2. A system as claimed in claim 1 further comprising an optical detector for detection of the third frequency converted optical signal.

3. A system as claimed in claim 1 wherein the first, second and third frequency converted optical signals also propagate transversely to the propagation direction of the first and second optical signal pulses.

4. A system as claimed in claim 1 wherein, where the detector also detects the first and second frequency converted optical signals.

5. A system as claimed in claim 1 the third frequency converted optical signal is a correlation signal between the first and the second optical signal pulses.

6. A system as claimed in claim 1 wherein the optical characteristics of the second optical signal pulse are known.

7. A system as claimed in claim 6 wherein the known optical characteristics of the second optical signal pulse comprises one or more of temporal characteristics such as the pulse width, the pulse shape, or in the case of a pulse train, the time between successive pulses or spectral characteristics such as the frequency/wavelength of the pulse or the spectral profile of the pulse.

8. A system as claimed in claim 1 wherein the optical characteristics of the second optical signal pulse are unknown.

9. A system as claimed in claim 8 wherein the second optical signal pulse is derived from the pulsed optical signal.

5 10. A system as claimed in claim 9 wherein the first optical signal pulse is divided into first and second spatially separated signal pulse portions, each pulse portion having characteristics corresponding to the characteristics of the first optical signal pulse, the first spatially separated signal pulse portion forming the second optical signal pulse for analysing the second optical signal pulse portion, from which an analysis of the first optical signal pulse is obtained. Q

11. A system as claimed in claim 1 wherein the nonlinear material is oriented such that the first and the second optical signal pulses each propagate through the nonlinear medium substantially perpendicular to the optic axis of the nonlinear material.

12. A system as claimed in claim 11 wherein the first optical signal pulse and the second optical signal pulse each are linearly polarised substantially parallel to the optic axis of thes nonlinear material.

13. A system as claimed in claim 1 wherein the nonlinear material is oriented such that the first and the second optical signal pulses each propagate through the nonlinear medium substantially parallel to the optic axis of the nonlinear material.

14. A system as claimed in any one of the preceding claims wherein the first and the secondQ frequency converted optical signals are respectively the second harmonic of the first and second optical signal pulses.

15. A system as claimed in any one of the preceding claims wherein the third frequency converted optical signal is the sum frequency of the first and second optical signals.

16. A system as claimed in claim 15 further comprising a filter intermediate the nonlinear₅ material and the detector, the filter being configured to block the first and second frequency converted optical signals.

17. A system as claimed is any one of the preceding claims wherein, when the first and second optical signals are of the same frequency, the frequency of the third frequency converted optical signal is the second harmonic of the first and second optical signals.

18. A system as claimed is any one of the preceding claims wherein the first and second optical signals propagate in free space prior to being incident on the nonlinear material.

19. A system as claimed is any one of claims 1 to 17 wherein, when the first and second optical signals propagate in optical fibres.

20. A system as claimed in any one of the preceding claims wherein the nonlinear material has randomly distributed and sized ferroelectric domain structures.

21. A system as claimed in any one of the preceding claims wherein the nonlinear material is Strontium Barium Niobate.

22. A system as claimed in any one of claims 1 to 19 wherein the nonlinear material is periodically poled such that the domains are regularly arrayed or regularly spaced, wherein the regularly arrayed domains may alternatively be annularly arrayed.

23. A method of analysing a first optical signal pulse comprising the steps of:

a) directing the first optical signal pulse to a nonlinear material having a ferroelectric domain structure with a desired propagation direction with respect to the nonlinear material;

c) the nonlinear material at the spatial overlap location, converting a portion of the first optical signal pulse and a portion of the second optical signal pulse to a frequency converted optical signal; and

d) detecting the frequency converted optical signal with an optical detector.

24. A method as claimed in claim 23 further comprising the step of (e) analysing the detected signal and associating the detected signal with the first optical signal for analysis thereof.

25. A method as claimed in claim 23 wherein the frequency converted optical signal corresponds to a correlation between the first optical signal pulse and the second optical signal pulse.

26. A method as claimed in claim 23 wherein the correlation is a cross correlation of the first optical signal pulse with the second optical signal pulse, where the second optical signal pulse is of partially known characteristics, or the correlation is an autocorrelation of the first optical signal pulse wherein the second optical signal pulse is derived from the first optical signal pulse.

27. A method as claimed in claim 23 wherein the second optical signal pulse is derived from the optical signal pulse and the correlation is an autocorrelation of the first optical signal pulse.

28. A method as claimed in any one of claims 23 to 28 wherein the frequency converted optical signal propagates in all directions in a plane that includes the propagation direction of the first optical signal pulse.

29. A method as claimed in any one of claims 23 to 28 wherein the frequency converted signal is detected substantially perpendicularly to the propagation direction of the first optical signal pulse.

30. A method of analysing a first optical signal pulse using second harmonic or sum frequency mixed light generated by the first optical signal pulse and a second optical signal pulse.

31. A crystal having a random ferroelectric domain structure when used for detection of an optical signal pulse using second harmonic light.

32. A crosscorrelator comprising a nonlinear material having a random ferroelectric domain structure.

33. A crosscorrelator as claimed in claim 32 wherein the crosscorrelator is a background-free crosscorrelator.

34. An autocorrelator comprising a nonlinear material having a random ferroelectric domain structure.

35. A autocorrelator as claimed in claim 34 wherein the autocorrelator is a background-free autocorrelator.

36. A method of obtaining an optical cross-correlation or an optical autocorrelation using a nonlinear material having a random ferroelectric domain structure.

37. An optical system comprising:

a first optical input port for receiving a first optical signal; a second optical input port for receiving a second optical signal;

a solid state nonlinear material for receiving the first and second optical signal, the nonlinear material capable of generating, in the transverse direction to the propagation direction of the first or second optical signal, a frequency converted optical signal derived from the first and second optical signals;

an exit port for outputting the frequency converted optical signal.

38. A system as claimed in claim 37 wherein the first and second optical input ports are each adaptable to receive a respective optical fibre.

39. An optical system for analysing a first optical signal pulse comprising:

a nonlinear material having a ferroelectric domain structure, the first optical signal pulse being alignable to be incident thereon, the nonlinear material capable of converting the optical signal pulse to a first frequency-converted optical signal; and

a second optical signal pulse source for generation of a second optical signal pulse to analyse the first optical signal pulse, the second optical signal pulse being alignable such that it propagates through the nonlinear material approximately collinearly to, although in the opposite direction of, the propagation direction of the first optical signal pulse, and the nonlinear material capable of converting the second optical signal pulse to a second frequency converted optical signal, and at a location within the nonlinear material where the first and second optical signals overlap, the nonlinear material is capable of converting the first and second optical signals to a third frequency converted optical signal.

40. A system as claimed in claim 39 further comprising an optical detector for detection of the third frequency converted optical signal.

41. An optical system for analysing an optical signal pulse comprising:

a nonlinear material having a ferroelectric domain structure, the optical signal pulse being alignable to be incident thereon, the nonlinear material capable of frequency converting the optical signal pulse to a first frequency-converted optical signal;

generation means for generating a optical analyser pulse from the optical signal pulse;

alignment means for aligning the optical analyser pulse to be incident on the nonlinear material such that it propagates through the nonlinear material approximately collinearly to, although in the opposite direction of, the propagation direction of the first optical signal pulse, such that the optical signal pulse and the optical analyser pulse overlap in an overlap region within the nonlinear material;

wherein, in the overlap region, the nonlinear material is capable of converting the signal and analyser optical signals to a second frequency converted optical signal, such that the second frequency converted optical signal propagates in an oblique or substantially transverse direction to the propagation direction of the signal and analyser optical signals.

42. An optical system comprising:

an first optical signal pulse source for generation of a first optical signal pulse;

a nonlinear material having a ferroelectric domain structure, the first optical signal pulse being alignable to be incident thereon, the nonlinear material capable of converting the optical signal pulse to a first frequency-converted optical signal;

a second optical signal pulse source for generation of a second optical signal pulse to analyse the first optical signal pulse, the second optical signal pulse being alignable such that it propagates through the nonlinear material approximately collinearly to, although in the opposite direction of, the propagation direction of the first optical signal pulse, and the nonlinear material capable of converting the second optical signal pulse to a second frequency converted optical signal, and at a location within the nonlinear material where the first and second optical signals overlap, the nonlinear material being capable of converting the first and second optical signals to a third frequency converted optical signal;

wherein the third optical frequency generated by the nonlinear material propagates in a direction that is either oblique or transverse to the propagation direction of both the first and second optical signal pulses.