GB2123552A

GB2123552A - Apparatus for determining the time-behaviour of ultra-short optical pulses

Info

Publication number: GB2123552A
Application number: GB08318271A
Authority: GB
Inventors: Peter Geittner
Original assignee: Philips Gloeilampenfabrieken NV
Current assignee: Koninklijke Philips NV
Priority date: 1982-07-09
Filing date: 1983-07-06
Publication date: 1984-02-01
Also published as: GB8318271D0; JPS5923225A; FR2530018A1; DE3225769A1

Abstract

Apparatus is provided for determining the time-behaviour of ultra-short optical pulses which avoids the disadvantages of fast detectors, streak cameras, or second harmonic generation apparatus. The pulse shape as a function of time and the pulse duration are determined by means of a linear auto-correlation function obtained by means of an optical interference process. A pulsed beam LS is divided into two subbeams a, b via a beam splitter ST. These subbeams are reflected by reflectors S1, S2 and are recombined by the beam splitter. The interference pattern formed in the recombined beam is measured D as a function of the shift (WZ2) of one of the reflectors S2, which is movable along the subbeam b axis. The intensity distribution in the interference pattern is analysed to produce information about the pulse shape and duration. <IMAGE>

Description

SPECIFICATION Apparatus for determining the time-behaviour of ultra-short optical pulses The invention relates to apparatus for determining the time behaviour of ultrashort optical pulses.

Various apparatuses of the type mentioned in the opening sentence are known. For example, a first known apparatus effects a direct measurement by means of fast optical detectors (generally avalanche photodiodes or MES field effect transistors) in conjunction with sampling oscilloscopes. The shape of the optical pulse is then measured directly. The time resolution is limited by the rise/fall time of the detector and the sampling system - or in other words: in general by the electrical response of the detection system.

Currently, the FWHM (Full Width Half Maximum) of such systems is approximately 50 to 100 ps. It is possible to determine real optical pulse widths smaller than 100 ps via square-law deconvolution of the system behaviour, but depending on the pulse width and shape systematic errors may occur. In general, optical pulses with small pulse energies (1 0-'4J) can be measured over a wide range of wavelengths (some 100 nm) depending on the spectral sensitivity curve of the detector.

In a further known apparatus the measurement is effected by means of a fast "streak camera".

Again the pulse shape is measured directly within the limits of a finite system behaviour, which in this system lies in the pico second range. In comparison with the above deconvolution method the time-resolution of the streak camera is about 2 orders of magnitude better, the spectral range is the same and the minimum pulse energies are also the same.

Finally, a non-linear auto-correlation measuring apparatus is known. This apparatus for pulse width measurement utilizes the physical principle of second harmonic generation (SHG) in a nonlinear crystal (for example KDP), which is also used for the frequency doubling of lasers in order to generate combined or very short wavelengths. By exposing the crystal to two frequencies z91 and w2 the combination frequencies , + )2 and w1 - W2 are generated via a non-linear effect. This is used for the pulse-width measurement by splitting the light from the optical emitter, which light comprises a sequence of recurrent optical pulses, into two subbeams in a beam splitter.Via a highly reflecting mirror these subbeams are reflected and recombined in the non-linear crystal. Starting from an initial position in which the distance between the mirrors and the beam splitter is the same, it is possible for example by shifting one mirror relative to the other, to obtain an optical pathlength difference for the two subbeams and thus a shift in time of the pulses relative to each other. The propagation-time difference T of the pulses is given by the relationship 2Az T= c0 where AZ the optical-pathlength difference of the subbeams and CO is the velocity of light in air. A shift of 1 mm of a mirror out of the initial position corresponds to a time delay of approximately 6.7 ps of the subpulses in the recombined beam.

The SHG signal in the KDP crystal appears with a signal frequency w112 and is proportional to the product of the overlapping subbeam intensities which recur with c9, (laser or emitter frequency).

When the subpulses coincide completely, i.e.

when A Z = 0, the SHG signal is a maximum. If the pulses do not overlap each other, i.e. A ZCd2 A TM where A Tm is the maximum width of the optical pulse the SHG signal disappears or, in accordance with the exact theory, decreases to one third of the maximum signal. For more details refer to Spectra Physics, Laser Technical Bulletin, February 1978, No. 8. Within limits the width of the SHG signal is therefore proportional to the pulse duration of the optical pulse. Except for specific pulse shapes, for example gaussian pulse shapes, the pulse width can be determined from the SHG signal. Reference is also made to the article "Autocorrelation function analysis" in the IEEE Journal, QE-1 6, 1980, No. 9, page 990.

The time resolution of the SHG measurement is very high (better than 0.1 ps), but since SHG is a non-linear effect - the required optical pulse energies are high. Moreover, the spectral range in which SHG is effected is greatly limited by the choice of the crystal and its adjustment. The socalled phase-matching condition must be met. As a result of this, the method is very expensive, requires complex adjustments and is limited to comparatively high pulse energies.

It is the object of the present invention to provide apparatus of the type mentioned in the opening sentence, which has a simple construction and operates in a simple manner and is particularly suitable for the measurement of the time-behaviour of optical pulses with low pulse energies and small absolute pulse widths.

Accordingly the invention provides apparatus for determining the time behaviour of ultra-short optical pulses, characterised by a beam splitter, two reflectors which are arranged in the paths of two subbeams formed by the beam splitter, at least one of said reflectors being movable along the corresponding subbeam axis, a light-sensitive detector, which is arranged in the path of the subbeams which are recombined to one beam by the beam splitter and which interfere with each other, and a system for processing the detector signal.

A first embodiment of the apparatus is characterised in that the reflectors comprise plane mirrors. After being reflected by the mirror and after recombination the two subbeams can interfere with each other uniformly over the entire cross-section of the beam which is formed by the recombined subbeams and which is referred to hereinafter as "recombined beam". This interference occurs only if the difference in pathlengths of the subbeams does not exceed the coherence length of the pulsed laser light. The resulting interference pattern exhibits a light distribution which is uniform across the crosssection of the recombined beam, the intensity of the uniform distribution depending on the relative phase of the subbeams. The intensity varies periodically with a continuous variation of the pathlength of one of the two subbeams as a result of a shift of one of the reflecting mirrors.One period then corresponds to a shift of the mirror by half the wavelength of the light used. The amplitude of the intensity variations moreover depends on the absolute path length difference of the two subbeams, i.e. the pathlength difference starting from a position in which the two mirrors occupy optimum positions, and on the duration of the optical emission process. The intensity variations are a maximum for equal absolute optical pathlengths of the subbeams and when one of the mirrors is shifted they decrease as the degree of overlapping of the subpulses in the recombined beam decreases. For pathlength differences which are so great that the subpulses no longer overlap i.e. the subpulses are shifted relative to each other by more than the pulse width, the interference effect no longer occurs.

The use of perpendicularly arranged mirrors is advantageous if the available intensities of the pulsed optical beam are low because, as a result of the uniform light-distribution across the entire cross-section of the recombined beam, the intensity modulation produced by a shift of one of the mirrors is measured by integration over the entire beam cross-section.

In a further advantageous embodiment of the invention one of the mirrors is tilted in such a way that the normal to the surface of the mirror is inclined relative to the corresponding subbeam axis. The resulting interference pattern is then a pattern of alternate bright and dark bands. The distance between adjacent parallel bands then depends on the tilt angle of the mirror and the wavelength of the pulsed light. The maximum local intensity differences between two adjacent bands depend on the pathlength difference of the subbeams and the optical-pulse duration. If one of the reflecting mirrors is shifted the interference pattern travels perpendicularly to the longitudinal direction of the bands. A shift by half the wavelength of the light used then corresponds to a shift of the interference pattern by a full band period.The local intensity modulation between two adjacent bands (band contrast) is a maximum in the case of equal optical path lengths of the subbeams, decreasing as one of the mirrors is shifted the pulses of the subbeams in the recombined beam overlap each other in time, to a smaller extent and disappears in the case of relative pathlength differences which correspond to a time shift by more than the pulse-width of the subpulses relative to each other.

This apparatus has the advantage that the movement of the interference pattern across the beam cross-section directly provides information about the direction and velocity of the shift of the two subpulses relative to each other. In addition, the occurrence of interference, in particular for small modulation amplitudes (small subpulse overlaps), can be determine more simply than in the case of perpendicularly arranged mirrors. The interference pattern can then be recorded pointby-point.

A further advantageous embodiment of the invention uses prisms, in particular retroflection prisms, as reflectors. At least one of the prisms may be then shifted perpendicularly to the beam axis. This results in the recombined subbeams being shifted parallel to each other without the occurrence of an additional pathlength difference between the subbeams. In this way the area in which the subbeams interfere with each other can be varied by shifting one of the prisms perpendicularly to the corresponding subbeam axis. As in this case different areas of the two subbeam cross-sections interfere with each other, the interference pattern thus formed contains information about the local time-behaviour of the optical pulse.This information is necessary for determining the pulse duration and pulse shape if the emission does not occur simultaneously in all the areas of the beam cross-section, or if the pulsed light passes through inhomogeneous media in which different beam portions are subjected to different delays. In both cases the overall pulse-width is greater than would be anticipated given the coherence length of the light.

One of the advantages of apparatus in accordance with the invention is that when a retroreflection prism which can be shifted perpendicularly to the beam axis is used it is also possible to determine the overall pulse duration from the variation of the interference pattern as a function of the perpendicular shift of one of the reflecting components.

In accordance with a further characteristic feature of the invention a diaphragm whose aperture diameter is variable and/or whose aperture can be moved in a direction perpendicular to the axis of the recombined beam is arranged between the beam splitter and the detector. The use of a diaphragm with an aperture of variable diameter may be advantageous if a mirror of prism whose normal to the base surface is inclined relative to the corresponding subbeam axis. The travelling pattern of interference bands which then occurs can be scanned pointwise by means of the diaphragm. The diameter of the diaphragm aperture is small in comparison with the band distance.

A diaphragm aperture whose position transverse to the axis of the recombined beam is adjustable may be used to advantage in an apparatus in which the subbeams can be shifted parallel to each other by shifting one of the prisms perpendicularly to the subbeam axis. By varying the position of the diaphragm perpendicularly to the axis of recombined beam and by varying the diameter of the aperture the interference pattern can be scanned and recorded at different locations in the beam cross-section for different overlapping areas of the subbeams. The resolution then depends on the selected aperture diameter.

Embodiments of the invention will now be described in more detail, by way of example, with reference to the accompanying drawings in which: Figure 1 shows an apparatus comprising two mirrors which are arranged perpendicularly to the subbeams, Figure 2 shows an apparatus comprising a tilted mirror, Figure 3 shows an apparatus comprising two prisms as reflecting components, and Figures 4 and 5 show measured results obtained by means of apparatus in accordance with the invention.

Figure 1 shows an apparatus comprising two mirrors arranged perpendicularly to the subbeams as reflectors. In this apparatus a laser beam LS, which has passed through a beam expander SA is split into two subbeams a, b in a beam splitter ST.

A pulse P is split into two subpulses P1 and P2 in the beam splitter ST. The subbeams a, b are reflected from two mirrors S1 and S2 which extend perpendicularly to the beam axes and are recombined as they pass through the beam splitter ST. In the present embodiment the mirror S2 can be shifted along subbeam b axis by means of a motor M. A variable-attenuation filter DF ensures that the subbeam intensities at the location of the cross-section B of the recombined beam formed by the subbeams are substantially equal. The subbeams a, b interfere with each other, the intensity in the present case being the same across the entire cross-section B. The intensity depends on the distance z2 of the mirror S2 from the beam splitter ST.A diaphragm L whose aperture has a diameter # is arranged in the radiation path of the recombined beam. In the present case the diaphragm L is centred relative to the recombined beam and the aperture corresponds to the beam cross-section B (0L = B).

The diaphragm consequently serves only for reducing undesired ambient light. Behind the diaphragm L a light-sensitive detector D is arranged, by means of which the intensity of the recombined beam is converted into an eiectrical signal. In the present embodiment this signal intensity is recorded on the vertical (Y) axis of a recorder R, the position Z2 along the axis of the mirror S2 driven by the motor M being shown on the horizontal (X) axis of the recorder. The intensity is thus recorded directly as a function of the distance z2.

When the mirrors are arranged symmetrically (Az = z, - z, = 0) the coincidence in time of subpulses P1 and P2 in the recombined beam is a maximum (åT = O). When the mirror S2 is shifted the intensity on the detector varies periodically, the period being given by Azp = A/2, A being the wavelength of the pulsed laser light A shift of the mirror S2 by a distance A z, corresponds to an optical pathlength difference of 2 A z between the two subbeams and thus to a shift in time of the subpulse P2 relative to the subpulse P 1.Thus, the shift of the mirror 52 over one period corresponds to a shift in time of the two subpulses of ATp = A/cO~6.67 10-3 pS um As the displacement A z of the mirror increases the subpulses P1 and P2 are increasingly shifted in time relative to each other, so that those pulse portions which can interfere with each other, or in other words which still overlap each other in time, decrease. As a result of this the modulation amplitude of the interference signal decreases when A z increases. For shifts A z and consequent delays A T = 2 A z/cO greater than A To (h To being the maximum pulse width of the optical pulse P) the modulation disappears completely and only a signal which is constant in time will be observed.The light intensity recorded by the recorder R as a function of the shift A z of the mirror S2 away from the symmetrical position therefore exhibits the characteristic variation represented in Figure 1, from which the pulse shape and duration of the laser pulse P can be derived by auto-correlation analysis. When the pulse duration and shape are unambiguously correlated with the coherence length of the light a quantitative analysis of the pulse is possible.

Figure 2 shows an apparatus comprising a mirror S2 which is tilted through an angle a relative to the subbeam b axis.

The elements of this apparatus have the same functions as the corresponding elements in Figure 1. The interfering subbeams a, b form a pattern of interference bands whose interference contrast, band distance and band position depend on the shift A z of the mirror S2.

In the radiation path of the recombined beam a diaphragm L is arranged whose aperture diameter 4i, is variable. The diaphragm L is centred on the axis of the recombined beam and the aperture diameter is small compared to half the distance d between the adjacent interference bands (0, d/2).

The interference pattern which travels transversely of the longitudinal direction of the bands as the mirror S2 moves along the axis of subbeam b can then be scanned pointwise by means of the light detector D. In the present embodiment the signal from this detector is recorded as the vertical (Y) deflection of the recorder R, the movement along the horizontal (X)-axis of the recorder being controlled via the motor M in conformity with the position of the mirror S2. The light intensity at the location of the diaphragm aperture is thus recorded directly as a function of the mirror position z2.

In the case of a symmetrical arrangement of the mirrors (bz = z2 - z1 = 0) the coincidence in time of the subpulses P1 and P2 in the recombined beam is a maximum (d T = O). The modulation amplitude of the pattern of interference bands is then also a maximum. The band distance d is the same for all bands and in the present case, because of the interference condition, it is given by t2 tan a. When the mirror S2 is moved out of the symmetrical position the interference pattern travels perpendicularly to the longitudinal direction of the bands, to the left or to the right depending on the direction in which the mirror S2 is shifted.A shift of the mirror S2 by åz2,p = A/2 corresponds to one period of the interference pattern, i.a. a shift of the pattern of bands by a full band distance d. This shift also corresponds to a time delay of the pulse P2 of ATp ;i/c0=6.67 1 10-3PS-( ps tim c0 being the velocity of the laser light.

As the shift åz of the mirror S2 increases the subpulses P1 and P2 are shifted in time relative to each other to an increasing extent, and the pulse portions which can interfere with each other become smaller. As a result of this, the modulation amplitude of the interference signal decreases as å z increases. In the case of shifts A z, which correspond to time shifts AT = 2 å z/c0 > ATo (å To being the pulse width of the pulse P), the interference contrast or pattern of bands disappears completely-and only a constant signal is detected.The light intensity recorded by the recorder R therefore exhibits the characteristic variation, which is qualitatively shown in Figure 2, as a function of the shift å z of the mirror S2, from which the pulse shape and pulse duration of the pulse P can be derived in accordance with an autocorrelation analysis process. A quantitative analysis is possible if the two magnitudes are correlated unambiguously with the coherence length of the pulse emission process.

Figure 3 shows an embodiment comprising two retroreflection prisms, arranged perpendicularly to the subbeam axis, as reflecting components.

The elements of this embodiment again have the same functions as the corresponding elements in Figures 1 and 2. The prism RP1 can be moved perpendicularly to the axis of the subbeam a over a distance A r, and the prism RP2 can be moved along the axis of the subbeam b by means of the motor M. From the known properties of retroreflection prisms, a shift of the prism RP 1 by a distance å r, perpendicularly to the axis of the subbeam a results in a parallel shift of this subbeam by 2 A r, at the location of the crosssection B of the recombined beam. However, this does not change the optical pathlength of the subbeam a.The subbeams a, b, can only interfere if they still overlap each other after a lateral shift by2Ar1. The interference pattern in the area of overlap depends not only on the distance z2 of the prism RP2 but also on the shift A r,, if the time behaviour of the laser pulse P differs at different locations across the beam. In general this is the case if the emission process is not homogeneous over the beam cross-section or if different beam portions are deiayed to a different extent, for example as they pass through inhomogeneous optical media.

Again, diaphragm L whose radial aperture position and aperture diameter sI, are variable, is arranged in the radiation path of the recombined beam. Behind the diaphragm L a light-sensitive detector D is arranged, by means of which the light intensity in the diaphragm aperture, which intensity is modulated when the prism RP2 is shifted, can be converted into an electrical signal.

This signal is recorded as the vertical (Y) deflection of the recorder R. The movement along the horizontal (X) axis of the recorder is controlled simultaneously via the motor M in conformity with the distance z2 between the prism RP2 and the beam splitter ST. The intensity at the location of the aperture can thus be measured directly as a function of the distance Za for different values of the radial displacement Ar, of the prism RP1.

In the case of a symmetrical arrangement of the retroreflection prisms (Az = z1 - z2 = 0, A r1= 0) both the geometrical overlap and the time overlap of the subpulses P 1 and P2 is a maximum. When the prism RP2 is shifted the intensity, which is uniform across the entire beam cross-section B, varies periodically, the period being given by åzp = A/2. This period åzp also corresponds to a delay of the subpulse P2 relative to the subpulse P1 of ATp = A/cO ~ 6.67 10-3pS tim where c0 is the velocity of the laser light. As the shift A z becomes greater the subpulses are shifted further in time relative to each other, so that the pulse portions which can interfere become smaller.Accordingly, the modulation amplitude decreases as å z increases. For shifts åz producing a time shift AT = 2åz/cO greater than åTO (åTO = maximum pulse width of the laser pulse) only light of a constant intensity is observed.

A similar qualitative behaviour is observed if the prism RP 1 is shifted perpendicularly to the axis of the subbeam a and starting from the symmetrical position of the prism RP2 (z2 - z1 = Az = 0), the modulation amplitude is measured as a function of the shift AZ of the prism RP2. However, in the present case non-corresponding portions of the subbeams a and b overlap each other, so that if the time behaviour of the laser pulse varies across the beam cross-section, the areas of the maximum modulation are shifted as a function of the displacement å r, of the prism RP1 and the displacement Ar, of the diaphragm L.Therefore, the intensity curves recorded by the recorder R have different shapes for different values of år" as can be seen in Figure 3. For the curve e år, is zero and for the source d år, has a non-zero value. A complete analysis of these modulation curves as regards the shape and position of the maxima as a function of år, and år, provides information about the coherence length of the pulse (duration of the emission process) and the temporal and geometrical structure of the pulse (variations across the beam cross-section).

Figures 4 and 5 show examples of interference patterns and variations of the auto-correlation functions measured by means of apparatus in accordance with the invention.

Figure 4 shows the intensity variation as a function of Az, in the recombined beam as measured with an apparatus as shown in Figure 1.

As a pulsed-light source a mode-locked synchronously pumped dye laser is used emitting light of a wavelength of 900 nm. One period then corresponds to a pathlength variation Az of 0.251us or a time delay of the subpulse P2 by 5.7 1 10-3 PS (ATp = 2 åzp/cO). Characteristic time variations of the auto-correlation function Imod(åT)/lmax (åT = 0) can be derived from the variation of the modulation amplitude as a function of å Z.

Figure 5 shows such variations for a modelocked synchronously pumped dye-laser pulse (curve k) and for a mode-locked krypton-ion laser pulse (curve L). The upper horizontal scale graduation m corresponds to curve k and the lower one n to the curve L.

In comparison with the prior-art apparatus such as a fast detector with sampling, or a streak camera, or SHG the apparatus in accordance with the invention has the following advantages.

The construction of all the embodiments is compact simple and comparatively cheap.

Adaptation to different light sources is possible without intricate adjustments. Depending on the quality of the optical elements the measurement is possible over a wide range of wavelengths without adaptation of these elements. Since interference is a linear effect the modulations can also be measured in the case of extremely small pulse energies. Therefore, the sensitivity is very high and, for example by means of a lock-in technique, it can be raised even further. For example, the minimum pulse energies that can be detected without lock-in technique are ~1 0-9Ws and those that can be detected with iock-in technique are 10-12Ws. From the autocorrelation function obtained by means of the interference process the explicit waveform of the optical pulse can be reconstructed more simply than in the case of the non-linear SHG measurement.Since the measured signal is directly proportional to the overlapping pulse area, the pulse shape as a function of time can be derived from the simple time derivative of the auto-correlation curve. The time resolution is very high, whilst at the same time calibration marks for the time axis can be derived from the periodicity of the modulation variation via å z (one period-6 67. 1 O- ps ( ), cf. of Figure 4) tim The time resolution depends on the accuracy of the mechanical shifting elements and is better than 10-2 Ps.

The measurement of the interference modulation curves requires the use of fast detectors. The speed of the motor can be very low and may be adapted to the detector sensitivity.

Fast recurrent pulse trains can be measured integrally as a constant light signal, without the information about the pulse structure as a function of time being lost. By means of apparatus as shown in Figure 3 time delays within a geometrically expanded optical pulse can be measured with a high time resolution for example better than 1 0-1 ps.

Claims

1. Apparatus for determining the time behaviour of ultra-short optical pulses, characterised by a beam splitter, two reflectors which are arranged in the paths of two subbeams formed by the beam splitter, at least one of said reflectors being movable along the corresponding subbeam axis, a light-sensitive detector, which is arranged in the path of the subbeams which are recombined to one beam by the beam splitter and which interfere with each other, and a system for processing the detector signal.

2. Apparatus as claimed in Claim 1, characterised in that an element for adjusting the intensity of a subbeam is arranged in the path of at least one of said subbeams.

3. Apparatus as claimed in Claim 1 or Claim 2 characterised in that a diaphragm of variable aperture is arranged between the beam splitter and the detector.

4. Apparatus as claimed in Claim 1, Claim 2 or Claim 3, characterised in that a diaphragm whose aperture can be moved in a direction perpendicular to the axis of the recombined beam is arranged between the beam splitter and the detector.

5. Apparatus as claimed in any one of Claims 1 to 4, characterised in that the reflectors are plane mirrors.

6. Apparatus as claimed in Claim 5, characterised in that one of the mirrors is tilted relative to the axis of the corresponding subbeam.

7. Apparatus as claimed in any one of Claims 1, 2, 3 or 4, characterised in that the reflectors are prisms.

8. Apparatus as claimed in Claim 7, characterised in that the prisms are retroreflection prisms.

9. Apparatus as claimed in Claim 7 or 8, characterised in that at least one of the prisms can be moved perpendicularly to the axis of the corresponding subbeam.

10. Apparatus as claimed in Claim 7, characterised in that the base surface of one of the prisms is inclined relative to the axis of the corresponding subbeam.

11. Apparatus for determining the time behaviour of ultra-short optical pulses substantially as described with reference to Figure 1 of the accompanying drawings.

12. Apparatus for determining the time behaviour of ultra-short optical pulses substantially as described with reference to Figure 1 and Figure 2.

13. Apparatus for determining the time behaviour of ultra-short optical pulses substantially as described with reference to Figure 1 or Figure 3.