Non-collinear sum-frequency generation of femtosecond pulses in a micro-structured β-BaB_2O_4...

Non-collinear sum-frequency generation of femtosecond pulses in a micro-structured β-

BaB2O4 crystal Carolina Romero,* Javier R. Vázquez de Aldana, Cruz Méndez and Luis Roso

Servicio Láser, Universidad de Salamanca, Pl. La Merced SN, E37008 Salamanca, Spain *Corresponding author: [email protected]

Abstract: Sum-frequency generation of femtosecond pulses by noncollinear phase matching in a micro-structured β-BaB2O4 crystal has been investigated. A 1D relief grating was recorded in the surface of a β-BaB2O4 crystal by direct irradiation with femtosecond pulses. Groove dimensions of 1 μm width and 200 nm depth were obtained for the recording parameters. This micro-structured crystal was used for generating ultraviolet pulses (265 nm) by noncollinear sum-frequency generation of the fundamental and the second harmonic of a Ti:Sapphire femtosecond laser system. Effective group-velocity matching between the incident pulses can be achieved for certain noncollinear processes in a compact way. The application of such device to intensity cross-correlation measurements of the second-harmonic pulse was shown.

©2008 Optical Society of America

OCIS codes: (190.7220) Upconversion; (320.7110) Ultrafast nonlinear optics; (230.4320) Nonlinear optical devices; (130.4310) Nonlinear; (050.1950) Diffraction gratings.

References and links

1. See, for instance, R. W. Boyd, Nonlinear Optics (Academic Press, San Diego, 2003). 2. D. Eimerl, L. Davis, S. Velsko, E. K. Graham, and A. Zalkin, “Optical, mechanical and thermal properties

of barium borate,” J. Appl. Phys. 62, 1968 (1987). 3. M. Nisoli, S. De Silvestri, V. Magni, O. Svelto, R. Danielus, A. Piskarkas, G. Valiulis, and A.

Varanavicius, “Highly efficient parametric conversion of femtosecond Ti :sapphire laser pulses at 1kHz,” Opt. Lett. 19, 1973–1975 (1994).

4. S. R. Greenfield and M. R. Wasielewski, “Near-transform-limited visible and near-IR femtosecond pulses from optical parametric amplification using Type II β-barium borate,” Opt. Lett. 20, 1394-1396 (1995).

5. P. Banks, M. Feit, and M. Perry, “High-intensity third-harmonic generation in beta barium borate through second-order and third-order susceptibilities,” Opt. Lett. 24, 4–6 (1999).

6. V. Krylov, J. Gallus, U. Wild, A. Kalintsev, and A. Rebane, “Femtosecond noncollinear and collinear parametric generation and amplification in BBO crystal,l” Appl. Phys. B 70, 163–168 (2000).

7. H. Liu, J. Yao, and A. Puri, “Second and third harmonic generation in BBO by femtosecond Ti:sapphire laser pulses,” Opt. Commun. 109, 139–144 (1994).

8. T. Zhang, H. Choo, and M. Downer, “Phase and group velocity matching for second harmonic generation of femtosecond pulses,” Appl. Opt. 29, 3927–3933 (1990).

9. C. Radzewicz, Y. Band, G. Pearson, and J. Krasinski, “Short pulse nonlinear frequency conversion without group-velocity-mismatch broadening,” Opt. Commun. 117, 295–302 (1995).

10. A. Smith, “Group-velocity-matched three-wave mixing in birefringent crystals,” Opt. Lett. 26, 719-721 (2001).

11. G. Arisholm, J. Biegert, P. Schlup, C. P. Hauri, and U. Keller, “Ultra-broadband chirped-pulse optical parametric amplifier with angularly dispersed beams,” Opt. Express 12, 518-530 (2004).

12. L. Sudrie, M. Franco, B. Prade, and A. Mysyrowicz, “Writing of permanent birefringent microlayers in bulk fused silica with femtosecond laser pulses” Opt. Commun. 171, 279–284 (1999).

13. X. Wang, H. Guo, H. Yang, H. Jiang, and Q. Gong, “Fabrication of beam shapers in the bulk of fused silica by femtosecond laser pulses,” Appl. Opt. 43, 4571–4574 (2004).

14. N. Takeshima, Y. Narita, S. Tanaka, Y. Kuroiwa, and K. Hirao, “Fabrication of high-efficiency diffraction gratings in glass,” Opt. Lett. 30, 352–354 (2005).

#98393 - $15.00 USD Received 7 Jul 2008; revised 29 Jul 2008; accepted 1 Aug 2008; published 22 Oct 2008

(C) 2008 OSA 27 October 2008 / Vol. 16, No. 22 / OPTICS EXPRESS 18109

15. K. Kawamura, T. Ogawa, N. Sarukura, M. Hirano, and H. Hosono, “Fabrication of surface relief gratings on transparent dielectric materials by two-beam holographic method using infrared femtosecond laser pulses,” Appl. Phys. B 71, 119–121 (2000).

16. C. Méndez, J. R. Vázquez de Aldana, G. A. Torchia, and L. Roso, “Integrated-grating-induced control of second-harmonic beams in frequency-doubling crystals,” Opt. Lett. 30, 2763–2765 (2005).

17. G. A. Torchia, C. Méndez, I. Arias, L. Roso, A. Ródenas, and D. Jaque, “Laser gain in femtosecond microstructured Nd:MgO:LiNbO3,” Appl. Phys. B 83, 559–563 (2006).

18. Y. Li, P. Lu, N. Dai, X. Wang, Y. Wang, B. Yu, and H. Long, “Surface relief diffraction gratings written on β-BaB2O4 crystal by femtosecond pulses,” Appl. Phys. B 88, 227–230 (2007).

19. N. Barnes and V. Corcoran, “Parametric generation processes: spectral bandwidth and acceptance angles,” Appl. Opt. 15, 696-699 (1976).

20. O. E. Martínez, “Pulse distortions in tilted pulse schemes for ultrashort pulses,” Opt. Commun. 59, 229-232 (1986).

21. J. Zhang, J. Huang, H. Wang, K. Wong, and K. Wong, “Second-harmonic generation from regeneratively amplified femtosecond laser pulses in BBO and LBO crystals,” J. Opt. Soc. Am. B 15, 200–209 (1998).

1. Introduction

Second-order nonlinear processes in crystals [1], such as second-harmonic generation (SHG), sum-frequency generation (SFG) or optical parametric amplification, are known to be very powerful techniques to produce ultrashort laser pulses in different spectral ranges. β-BaB2O4

(BBO) is one of the most widely used nonlinear crystals due to its high optical nonlinearity, even in the ultraviolet (UV), wide range of transparency and physical robustness [2]. Efficient second and third-harmonic generation (THG) or parametric generation/amplification were demonstrated with femtosecond pulses in BBO [3-6]. In all these processes, in addition to achieve the phase-matching conditions, the group-velocity mismatch (GVM) between the interacting waves needs special care provided that it is a limiting factor of the interaction length and it elongates the generated pulse [7]. Noncollinear phase-matching is very often used in order to avoid, or at least to minimize, the detrimental effects of GVM. SHG [8,9] or THG [9] can be benefited from this technique. The efficiency of parametric processes can be even improved by combining noncollinear phase matching with pulse front tilting [8,10]. This technique allows one to increase also the bandwidth of parametric amplifiers.

On the other hand, femtosecond infrared laser pulses have been successfully used in high-precision micro-structuring of transparent crystals and glasses, by ultrafast laser ablation or optical breakdown. The application of this technique for the inscription of diffractive elements has been widely investigated. Both internal [12-14] as relief [15-18] gratings can be designed and fabricated by direct writing with femtosecond pulses with arbitrarily complex geometries. Beam shaping [13] or noncollinear laser action [17] are examples of practical uses of these elements. Concerning nonlinear optics, the integration of diffractive gratings in crystals has been recently used for tailoring SHG in KH2PO4 [16] or in BBO [18] by noncollinear phase matching among several diffraction orders. In this work we have recorded a 1D relief grating in the surface of a BBO crystal with a femtosecond laser system. With precise control of the pulse energy the dimensions of the groove were greatly reduced thus allowing us to set a grating spacing of 3.5 μm. After optical and atomic-force microscopy characterization of the grating, it was used for noncollinear SFG between fundamental and SH femtosecond pulses. The incident pulses are diffracted at the grating that is placed at the entrance surface of the crystal. By changing the tilt angle of the crystal and the relative delay between the incident pulses, SFG take place by noncollinear phase matching among some of the diffraction orders thus generating TH pulses. Moreover, the correct choice of the grating parameter allows the propagation angle between fundamental and SH pulses to be such that the GVM can be effectively minimized for certain noncollinear processes. The application to intensity cross-correlation measurements of SH pulses is also presented.



2. Micro-structuring BBO and grating characterization

The laser system we used for recording the grating was a Ti:Shappire regenerative amplifier (Spitfire, Spectra Physics) that operated at 1 kHz repetition rate. It delivered pulses of 120 fs duration, 795 nm central wavelength and maximum pulse energy of 1 mJ. The beam size was around 10 mm (at 1/e2) and the energy was finely controlled with a λ/2 plate and a linear polarizer followed by a neutral-density filter. A 50x microscope objective (Epi Plan L, Leica) was used for focusing the beam. An uncoated BBO crystal (14x14x1 mm3, θ=44.7º) was placed on a computer-controlled motorized three-axes stage. The beam was focused on the surface of the sample while scanning the piece with constant velocity. In the search for the minimum groove size, several writing parameters (pulse energy and scanning velocity) were tested before recording the grating, analyzing the grooves with optical and atomic-force microscopes. We finally selected 300 nJ/pulse (measured in front of the objective lens) while scanned at 57 μm/s as engagement between high quality of the gratings and reasonable processing time (~9 hours, short enough to ensure the stability of the beam). For these parameters, the depth of the grooves was 200-250 nm and the width was around 1 μm. The achieved dimensions are smaller than those reported by Li et al. [18] (2-3 μm groove width) due to the different focusing optics (20x microscope objective) and writing parameters (5μJ pulse energy, 50 μm/s scanning velocity) used in their procedure.

Working near the precision limit of our stage, we fixed the groove spacing to d=3.5 μm and we recorded a 2.5×2.5 mm2 grating. In Fig. 1(a) we show a partial picture of the grating taken with an optical microscope in reflection mode (bright field). The efficiency of the grating was measured by irradiating it with a He-Ne laser (600 μm beam size) and the normalized power of the main diffraction orders is shown in the bar chart (Fig. 2(b)). The total energy travelling in the transmitted diffraction orders amounts to 49% of the incident. The beam quality of the transmitted orders was inspected by recording them on a CCD: in spite of the apparent good quality of the grating some lack of periodicity or defects in the writing procedure lead to noncircular beams, in particular for the orders 3, 4 and 5.

Fig. 1. (a) Partial view of the grating taken with an optical microscope in reflection mode. (b) Bar chart showing the normalized energy of the diffraction orders when irradiating the grating with a He-Ne laser.

3. Non-collinear SFG: experimental setup

The micro-structured BBO crystal was used for SFG between the fundamental and the second harmonic of the femtosecond laser system described in the previous section, thus producing the TH of the Ti:Sapphire (setup depicted in Fig. 2). We reduced the pulse energy of the beam to 70 μJ with a neutral density filter before passing through a first BBO sample (0.5x14x14 mm3, θ=29º, uncoated) where the SH pulses were generated by collinear type-I phase matching. Under our experimental conditions the measured (external) efficiency of the crystal for SHG was around 2%. The central wavelength (397.5 nm) and the spectral width of the SH

(a) (b)



pulses (~2.4 nm) were measured with a spectrometer (HR4000, Ocean Optics). Fundamental and SH pulses were separated in a dichroic mirror (DM1). A delay line with a motorized stage was arranged for the blue beam. A short-pass filter (cut-off wavelength 750 nm, 0.38 transmittance at 397.5 nm) was used to eliminate any infrared light reflected by the dichroic mirror. Infrared pulses transmitted through the dichroic mirror were attenuated by a neutral density filter decreasing their energy to 5 μJ. A half-wave plate was introduced also in the optical path in order to rotate 90º the polarization. Both beams were joined again in dichroic mirror DM2 so that they propagated spatially overlapped from it. The beams were then focused with an f=40 cm achromatic doublet. The pulse energy measured after the lens was 4 μJ for the fundamental and 0.3 μJ for the SH beam.

The micro-structured BBO crystal was placed on a positioning mount (with rotation and tilt control) behind the focus for the fundamental. At the grating plane the diameter of the largest beam (the SH) was around 500 μm so that both beams fitted the grating size. The orientation of the optic axis of the crystal is set in such a way that both beams are ordinary (polarization perpendicular to the grating grooves).

Fig. 2. Experimental setup for SFG. DM: dichroic mirror, SP: short-pass filter, ND: neutral density filter, L: focusing lens.

4. Measurements and results

As the two co-propagating beams enter the micro-structured BBO crystal, the diffraction grating produces the off-axis propagation of multiple beams (diffraction orders). The angles of these beams with the incidence direction at the exit of the crystal were measured for the fundamental and the SH pulses by intercepting them on a screen. They are summarized in Table 1 where we have also included the theoretical values calculated from Eq. (1) (see the next section).

Table 1. Measured and calculated angles (external) for the first and second diffraction orders

Fundamental Second harmonic

Angle Order 1 11

~ωα Order 2 12

~ωα Order 1 21

~ωα Order 2 22

~ωα

Measured 14.9º 34.1º 7.4º 14.9º

Calculated 13.1º 27.0º 6.5º 13.1º

For normal incidence, once the temporal overlap between the fundamental and the SH pulses was achieved, TH radiation was generated along the incidence direction. It was detected with the spectrometer after separating the three wavelengths with an UV-grade fused silica prism. The spectrum, centered at 265 nm, is shown in the graph on the right of Fig. 3 (solid line in panel (b)). Such TH beam was generated by type-I collinear SFG between the order 0 of the two wavelength pulses: the cut of the crystal (θ=44.7º) was optimum for phase matching this process at normal incidence (o795+o397.5 →e265). Some other weak TH beams



appeared propagating noncollinearly between the orders 0 and ±1 for the SH (see picture at the top in Fig. 3(a)) forming external angles of around ±4.4º with the order 0. Afterwards, the tilt of the crystal (φ) was finely changed from the normal incidence position (φ0=0). At φ1=0.5º the intensities of the TH beams at ±4.4º external angles were maximized. The tilting was again increased: these two beams vanished, but other two appeared propagating at angles of ±9.2º. The maximum of these signals, much weaker than the others, was obtained at φ2=2º. We increased the tilt and at φ3=2.7º two new TH beams emerging again at ±4.4º were maximized, at first glance more intense than the previous ones. By careful adjusting the delay between the fundamental and the SH pulses, the pulse energy of these beams growth up to 35 nJ each one. A further increase in the crystal tilt to φ4=11.3º lead to the generation of other two TH beams propagating at angles of ±9.2º. These two beams were very sensitive to the delay between the incoming pulses and it was necessary to very finely adjust it in order to get the maximum pulse energy of 10 nJ. The spectrum of such beams was carefully analyzed and some spatial chirp was apparent, probably due to the angular dispersion introduced by the grating in the noncollinear beams. In Fig. 3(a) we show the pictures of the emerging beams for several tilt angles and the spectra of the central part of the TH beams are also shown in Fig. 3(b).

Fig. 3. (a) Pictures of the emerging beams projected on a paper screen for different values of the crystal tilt φ. Red arrows point the ±1 diffraction orders for the SH and yellow arrows point the TH beams generated by noncollinear phase matching whose intensity is maximized for the given tilt. The bright spot in the middle is the order 0 for both frequencies. The orders ±1 for the fundamental are out of the picture frame. (b) Spectra of the TH beams optimized at the tilt angles indicated in the legend.

5. Theoretical description and discussion

5.1 Diffraction orders

For normal incidence the grating produces the propagation of multiple beams in the horizontal plane (x-y, see Fig. 4 (a)) inside the crystal along the directions defined by

dn

cm

ioi

im ωω

ωπα 2

sin = (1)

with m being the diffraction order, 2,1, =iiω the frequencies of the incident waves ( 1ω

fundamental and 2ω second harmonic) and ionω the corresponding index of refraction

(ordinary). imωα is the angle between the order 0 and the wave-vector for order m (see Fig.

4(a)). Note that the angles computed with Eq. (1) are internal; the external angles imωα~ shown

in Table 1 were calculated from Eq. (1) by setting 1=ionω .

(a) (b)



When the crystal is tilted an angle φ (rotation around x axis) the incident beams also experience the refraction at the entrance of the crystal resulting in an inclination of the plane

that contains the diffraction orders (diffraction plane) an angle ( ) ( )ionωφφβ /sinarcsin= as

depicted in Fig. 4 (a). The different inclination of the diffraction planes for the two incident beams, which is less than 0.2º for a tilt φ=10º, will be neglected along this work by taking

7.121 ≅≅ ωωoo nn for BBO in ( )φβ .

5.2 Phase matching conditions

The SFG process is efficient only if the phase-matching conditions are approximately satisfied, that is

( )⎪⎩

⎪⎨⎧

=+

=+321,

321

ωωω

ωωω

nmnm kkk (2)

where 1ωmk and 2ω

nk are the wave-vectors (inside the crystal) of order m for the fundamental

pulses and order n for the SH respectively. ( )3,

ωnmk is the wave-vector for the generated TH

beam. In case of perfect phase matching for type-I SFG (oo→e) the above vector condition splits into the next two scalar equations:

( )( ) ( ) ( )( ) ( )⎪⎩

⎪⎨⎧

=−+−

−+−=

0sin2sin

cos2cos,3

2211

22113

χαχα

αχχαχφθωωωω

ωωωωω

nomo

nomoe

nn

nnn (3)

where we have assumed ωω 22 = and ωωωω 3213 =+= , with ω being the central frequency of the fundamental pulses. The definitions of the angles inside the crystal are shown in Fig. 4 (a). The c-axis of the crystal is contained in the yz-plane, x is the tilting axis and y is the propagation direction of the incident beams. ( )φγ is the angle forming the c-axis with the

diffraction plane and depends on the tilt angle though ( ) ( )φβγφγ += 0 , with γ0=44.7º being

the collinear phase matching angle for which the crystal was cut. θ(φ) (not depicted in the figure for clarity) is the angle between the generated TH beam and the c-axis. It depends both on χ as on ( )φγ through the equation ( ) ( ) χφγχφθ coscos,cos = .

Fig. 4. (a) Definitions of the axes and angles inside the crystal. Only one half of the diffraction plane (i.e. positive diffraction orders) is shown for clarity. (b) Diagrams showing the diffraction plane for the two noncollinear SFG processes that involve orders 0 and 1. They are phase matched at different crystal tilts (φ1 and φ3); thus the inclination (β(φ)) of the depicted planes is different in the two cases.

(a) (b)



The most intense noncollinear TH beams will be generated when, in the SFG process, the order 0 for the fundamental or for the SH are involved because of the large fraction of the energy that travels along this direction (see bar plot at Fig.1). From the phase-matching conditions (Eqs. (3)) we can compute the tilt angles of the crystal for which noncollinear phase matching is achieved between the order 0 of the fundamental/SH, and the orders 1 (see diagrams in Fig. 4(b)) or 2 for the SH/fundamental respectively. We have also computed the angles ( )m,nχ~ at which the so-generated TH beam emerges from the crystal. They are

summarized in Table 2 where the measured values have been also included thus allowing us to identify the origin of the TH beams previously described in Sect. 4. Note that the processes involving the order 0 for one wavelength and the order m for the other generate the SF pulse along nearly the same angle, but for different crystal tilts.

Table 2. Calculated (th.) and measured (exp.) crystal tilts (φi, with i=1,2,3,4 as defined in Sect. 4) and emergence

angles of the TH pulses ( ( )m,nχ~ ).

( )m,nχ~ th. ( )m,nχ~ exp. φ th. iφ exp.

2110ωω kk + 4.3º 4.4º 0.5º 0.5º

2120ωω kk + 8.7º 9.2º 1.9º 2º

1210ωω kk + 4.3º 4.4º 2.3º 2.7º

1220ωω kk + 8.8º 9.2º 9.4º 11.3º

Figure 5(a) shows the angles between the fundamental and the SH beam for which phase matching is achieved as a function of the crystal tilt. The blue curve corresponds to the

processes involving the order 0 for the fundamental ( )010 =ωα and the order n for the SH

( )2ωαn , and the red curve the opposite. Filled circles are the experimental measurements.

Fig. 5. (a) Noncollinear phase matching curves. See text for explanation. (b) Bandwidth of the TH pulses as a function of the angle between the fundamental and the SH beam.

5.3 Third-harmonic bandwidth

The spectral bandwidth of the generated TH pulses is strongly dependent on the wave-vector

mismatch ( )321,

ωωωnmnm kkkk −+=Δ . The efficiency of the SFG process decreases rapidly with

increasing kΔ and one criterion for the maximum allowable wave-vector mismatch is

L/π≤Δk with L being the crystal length [19]. In order to estimate the spectral bandwidths

(a) (b)



of the TH pulses and to compare them with experimental measurements (Fig. 2) we assume that the SH pulse is monochromatic with wavelength 5.3972 =λ nm. We have computed the dependence of the phase mismatch on the fundamental wavelength in terms of the angle

between the fundamental and the SH ( 21 ωω αα nm − ). The bandwidth of the generated TH pulses is plotted in Fig. 5(b) for a L=1 mm BBO crystal. The very small change in the bandwidth with the non-collinear angle is in agreement with the spectra shown in Fig. 2. Note that the slightly larger bandwidth observed in the measured spectra (~0.5 nm) is well justified by the true bandwidth of the SH pulses (neglected in the calculations).

5.4 Pulse tilting

Ultra-short laser pulses after propagating through an angularly dispersing element (like a diffraction grating) experience a pulse-front tilt that depends on the grating parameters, incidence conditions and distance to the grating. For normal incidence in the grating and just after this element the tilt angle approximately equals the diffraction angle [20]: it means that inside the micro-structured crystal the pulse-fronts of all diffraction orders will be nearly parallel to the pulse-fronts of the incident pulses (see diagram in Fig. 6).

6. Effective group-velocity matching. Cross-correlation of SH femtosecond pulses

Group-velocity mismatch (GVM) between the incident pulses limits the effective interaction length of the SFG process due to the loss of temporal overlap in the crystal. In a collinear geometry with type-I phase matching, assuming a pulse duration of τ=120 fs (for both the fundamental and the SH pulses) the interaction length in BBO is reduced to

( ) ( ) 36011

12 =⎥⎦⎤

⎢⎣⎡ −

−− ωωτ gg vv μm. The micro-structured crystal studied along this work offer the

possibility to avoid, or at least alleviate, the disturbing effect of GVM between the incident pulses in a compact way for certain processes. In noncollinear SFG where the order 0 of the SH and another order of the fundamental are

involved effective group-velocity matching is achieve if 112 cos ωωω αmgg vv = (see Fig. 6). In our

case, for the interacting wavelengths, this condition is fulfilled for º191 ≅ωαm . The order 2 of the recorded grating for the fundamental beam propagates inside the crystal with an angle

slightly larger than 19º (measured external angle º1.34~ 12 =ωα , see Table 1) and then for the

process ( )321

0,202ωωω kkk =+ the GVM is expected to be effectively reduced.

Fig. 6. (a) Diagram of the noncollinear SFG process induced by the grating showing pulse tilting and effective group velocity matching between the incident pulses. (b) Cross-correlation traces obtained with the micro-structured BBO crystal for collinear (red) and noncollinear (blue) SFG processes. In solid line we show a theoretical cross-correlation trace computed by assuming GVM=0 for comparison.

(a) (b)



An application of this property is the cross-correlation of different-wavelength pulses by SFG. For measuring the intensity cross-correlation trace one has to use a crystal as short as possible to ensure the product GVM·L being much smaller than the typical pulse durations; otherwise the trace would be artificially broadened (time smearing). We used these ideas for measuring the SH pulse duration with our setup. In Fig. 6 (b) we present two cross-correlation traces of the SH pulses generated in the 0.5 mm thick BBO, performed in the micro-structured (1 mm thick) BBO used along this work. A GaP photodiode was used for measuring the power of the generated TH signal. One trace (triangles) corresponds to that obtained in a “standard” collinear SFG configuration (tilt φ=0, normal incidence) where the GVM is not compensated. The trace is very broad due to the time smearing in the 1 mm BBO. The second trace (circles) was obtained in the effective group-velocity matching configuration described above (φ=11.3º). The width of this trace is ~180 fs and it should be compared with the theoretical curve (solid line) included in the figure. For obtaining this curve we have used a numerical model in one dimension (see for instance [21]) to compute the fundamental and SH pulses emerging from the first BBO crystal; afterwards, the cross-correlation trace between them was evaluated by assuming GVM=0. The agreement of this trace with that measured in our noncollinear setup is quite good thus demonstrating the effective reduction of the GVM. Unfortunately, as shown in the spectra of Fig. 3 and in the calculations (Fig. 5(b)), the phase-matching bandwidth is not increased when working in such noncollinear geometry.

7. Conclusions

High-quality diffractive elements were recorded by femtosecond direct irradiation in the surface of a BBO crystal. Typical dimensions of the grating grooves were 200-250 nm depth and 1 μm width, thus allowing a grating spacing of 3.5 μm. The micro-structured crystal was used for non-collinear SFG of the fundamental and the second harmonic of a femtosecond Ti:Sapphire laser system. By slightly changing the tilt of the crystal SFG among several diffraction orders was achieved thus generating UV (265 nm) pulses along certain directions. The possibility to get effective GVM between the incident pulses for certain diffraction orders was discussed. This property was applied to intensity cross-correlation measurements of SH pulses demonstrating a clear reduction of the well known time smearing effect. The ideas discussed along this work are also expected to be applied to difference-frequency generation (optical-parametric amplification) of femtosecond pulses. Other applications of the micro-structured crystal are compact beam splitting or signal multiplexing. Different techniques of femtosecond laser writing are now under investigation in order to reduce the grating spacing and to improve the conversion efficiency and beam quality.

Acknowledgments

We acknowledge partial support from the Spanish Ministry of Science and Education (MEC) through Consolider Programme SAAUL (CSD 2007-00013) and project No. FIS2006-04151, and from the Junta de Castilla y León through project SA145A08. C. Romero also acknowledges MEC for support (grant BES-2007-17415). The authors would like to thank J.A. Pérez for atomic-force microscope measurements.



Non-collinear sum-frequency generation of femtosecond pulses in a micro-structured β-BaB_2O_4...

Documents

Transcript of Non-collinear sum-frequency generation of femtosecond pulses in a micro-structured β-BaB_2O_4...