DIRECT MEASUREMENT OF FEMTOSECOND PULSE FRONT DISTORTION DUE TO by asafwewe

VIEWS: 8 PAGES: 2

More Info
									DIRECT MEASUREMENT OF FEMTOSECOND PULSE FRONT DISTORTION DUE TO LENSES AND
ITS ELIMINATION USING AN ACHROMATIC LENS
S Ameer-Beg2 , A J Langley1,1 N Ross1 ,W Shaikh1 and P F Taday1
1 Central Laser Facility, Rutherford Appleton Laboratory, Chilton, Didcot, Oxon., OX11 OQX, U.K.
department of Physics, University of Essex, Colchester, Essex, C043SQ, U.K.
INTRODUCTION
Temporal deformation of ultra-dioxt optical pulses in laises can
significantly reduce (he focused power and time resolution available
from lasers. Aside from conventional aberrations, two types of
temporal pulse distortion can occur in laises 1-4 namely: pulse front
distortion (PFD) and group velocity disposion (GVD) . PFD arises
because of the difference betweoi the phase and group velocities in
T
50.4 -
50.2
§
^ 50.0
■g) 49.8
« 49.6
17^
Achromat
-C
the leas as it varies in thickness across the beam. This results in a
radially depmdmt delay betweoi the phase and pulse fronts. GVD on
the other hand results in pulse stretdi. Since the amount of GVD
depends on the thickness of dispersive material, axial beams are
affected most in positive looses.
- / Fused silica lens
P 49.4
o
u_
49.2
J.
600
700
BOO
900
It is important to find optical designs which avoid pulse distortion.
One solution to the problem is to use paraboloidal rdlective optics
since these do not distort pulse fronts. Such optics may not always
be convenimt, however, where on-axis focusing in the forward
direction is required, or where engineering constraints rule out off-
axis paraboloidal focusing. Sudi problems were encountered in the
design of an / /2 optic for focusing intense, 100 fs pulses in
multiphotan ianisation (~MPI) experiments. An alternative solution is
to use achromatic lmses for focusing since, as theoretical studies1,6
have pointed out, these lenses should produce a constant delay across
the beam between the phase and pulse fronts. This being an
acceptable solution for our application, an adrromat was designed
and tested to meet our focusing requirements.
300 F
J2
250
_______ Fused silica lens
C 200
O
150
E
o
100
u>
50
U
o
Achromat
o
-50
4> -100 -
(O
-150 fc
O.
600
700
800
900
Wavelength / nm
DESCRIPTION OF EXPERIMENT AND RESULTS
Figure 1: Focal lmgth (upper graph) and pulse front
distortion (lower graph) as a function of wavelatgjh for the
achromat described in the text and a fused-silica lens of SO
mm focal lmgth.
Two types of 50 mm focal length laises were compared in this
investigation. The firrf of these comprised a fused-silica 34 mm
diameter plano-convex doublet. Each component was 5.5 mm thick
cxi axis and was separated by 2.5 mm. One companmt was
aspherised to correct for spherical aberration. The doublet was used
for many years for picosecond MPI experiments and provided good
quality focal spots for up to //2 focusing. Two idmtical doublets
wavelcngjh range 650 - 775 tim Also shown for comparison are the
focal lengths and pulse fremt distortion calculated for a fused silica
lms of focal length 50 mm (at 740 nm).
were manufactured
Two idmtical achromats were manufactured to the design discussed
above by Optical Surfaces Ltd who also made mterferometric
measurements demonstrating a maximum optical path difference
across the lms aperture of < A/5 at 633 nm Measurements of the
focal f ot obtained with a 740 nm, 25 mm diameter beam of uniform
intmsity indicated that near to diffraction limited focusing was
achieved with these lmses. Each achromat was assembled in a
stainless steel mount designed for operation in an ultrahigh vacuum
apparatus. All optical surfaces were coated for minimum rdlectian
losses at 740 nm.
The other leas was an achromat. This was designed using Kidger
Optics ray-tracing software to provide 50 mm focal lmgth f /2
focusing with minimum pulse front distortion (chromatic abaration)
and spherical abaration at 740 nm Minimum aberrations wae
achieved with a three cexnponmt optic.
Pulse stretch due to GVD in the achromat was calculated to he 22 fs.
Thus a 100 fs duration sech2 pulse passing througi the lms would be
stretched to 102 fs. In principle this stretch could he ranoved by the
introduction of an appropriate amount of negative GVD.
Pulse front distortion measuranmts wae carried out using the
experimmtal arrangement riiown schematically in Figure 2. A
Spectra-Hiysics argon-ion pumped titanium-sapphire laser provided
near to transform-limited sech2 pulse of 100 fs duration at 740 nm
and at a repetition rate of 82 MHz.
The pulse front distorticxi, AtppQ is a function of the rate of change of
focal lmgth with wavelength df/dX 1 and is givm by
Xr2 df
AtpFD =
(1)
2cf2
The lasa beam was incident upon an optical beam-flitting
arrangement resembling a Michelsan interferometer. The idmtical
lens pair under investigation was placed in one arm and formed a
telescope of unit magnification. The position of the mirror in the
other aim was set to give a suitable delay between the plane and
curved pulse fronts. After recombination at the beam flitter, differmt
regions of the pulse fronts could be selected with a 1mm diameter
aperture.
where r is the beam radius at the lens of focal lmgth, f, c is the feed
of light and A is the wavelength of the incident light.
The ray tracing software was used to calculate the effective focal
lmgth of the achromat as a function of wavelmgth and this data was
used in equatiem 1 to calculate the pulse fremt distortion. Both these
data sets are diown in Figure 1. The lower graph illustrates that
pulse front distortion of less than 25 fs is expected ova the
176
The fused silica doublet introduced a significant curvature to the
pulse front. It is noted that the measured curvature was less than that
calculated (dotted curve) for a 50 mm focal lmgth fused silica singlet
lms. This is due to the reduced beam radius at the second
compcnmt of the doublet and aspherisation. Howeve1, the measured
delay remains quadratically dependent on radius, as indicated by the
closefit of a curve oftheform AtpFD = 0.82^ (solid line). Fromthis
the radius at the second component was calculated to be 8.5 mm
which is close to the value obtained by ray tracing.
-f
m
BS, „Lenaes„
Pulse from laser
1:
z
I
Also diown in Figure 3 is the measured pulse front delay for the
achromatic lens. There is evidence of a snail tilt betweai the
achromat and rrfs*ence pulse fronts of about 14 fs across the 25 mm
beams due to either a wedge on the Ions or a slight misalignment of
the instrument However, it is clear that the quadratic chromatic pulse
front delay has been eliminated.
Distorted pulse front
Plane pulse front
w
Aperture
4	►
BS,
f! number
CCD
2 25 3.3 5 10
i—1—i—1—i—«—i—'—r
CaaJiedaxve
10 5 3.3 25 2
i'ii—t—i—i—i—i
-175 -
-160 -\
-125
J2 -ioo
2 -75
Fised
-50
aica
dalM
-25
Adrcirrt
0
A utoco relator
jlJLa.
-125 -10.0 -7.5 -5.0 -25 0.0 25 5.0 7.5 10.0
Radius / mm
125
Figure 3: Measured pulse fronts ( 740 nm) after
focusing with a fused-silica doublet and an achromatic lens
Oscilloscope
Figure2: Arrangement used to measure the pulse front
delay introduced by lenses. Key: BS beam splittCT, m
CONCLUSION
mirror
An achromatic lens designed for high quality f /2 focusing has been
demonstrated to introduce no significant pulse front distortion as
predicted by theory.
As an alignment aid the beams were sampled by a 15% flitter placed
after the aperture and focused with a 100 mm focal length lens onto a
CCD detector, the output of which was displayed on a television
By insuring the images of the two focal spots were
superimposed, the angle between the beams was maintained below
100 prad commanding to a temporal delay of ®8 fs across the 25
Pulse front distortion caused by lenses has been measured directly
using a scheme comprising equipmoit likely to be found in any
femtosecond laser laboratory. The technique can be applied to any
cptical component and does not require apparatus designed for
Unlike methods which seek to
measure whole beam cross-correlation betweai plane and distorted
pulse fronts, the technique described here measures relative temporal
delay across a beam and is therefore not unduly affected by beam
intensity profile, temporal pulse shape or GVD.
screen.
mm beam.
interferometric measurements.
The relative delay between the two pulses emerging from the aperture
was measured by directing them to a Femtochrome rapid-scanning
autocorrelator. The autocorrelation measurement of the two pulses
consistsed of three peaks. The central one corresponded to the
autocorrelation of both pulses whilst the outer peaks was a cross-
correlation of one pulse with the other. The temporal separation
between the central and outer peaks give a measure of the delay
between the two pulses.
REFERENCES
1.	Z. Bor, J. Mod Optics,35 ,1907, (1988)
2.	M Kempe, U. Stamm and B. Wilhelnii, Optics Comm.,
89,119,(1992).
3.	Z. Bor and Z.L. Horvath, Optics Comm., 94 249, (1992).
4.	M Kempe and W. Rudolph, Fhys. Rev. A48 ,4721, (1993).
5.	R.L. Fork, C.H. Cruz, P.C. Beck a" and C. V. Shank, Optics
Letts.,12,483, (1987).
6.	S. Szatmari and G. Kuhnle, Optics Coinin.. 69 ,60, (1988).
Scanning the aperture across the beams allowed the relative delay
between the curved and reference pulse fronts to be measured Figure
3 shows such measurements obtained with the fused silica doublets
and achromatic lenses. The double pass through the ltns pairs has
been taken into account and the delay plotted on the vertical axis of
the graph is 0.25 times that measured
177

								
To top