Simultaneous stimulated Raman scattering second harmonic generation in periodically poled lithium niobate
Gail McConnell
Centre for Biophotonics, Strathclyde Institute for Biomedical Sciences, University of Strathclyde, 27 Taylor Street,
Glasgow, G4 0NR, UK.
Allister I. Ferguson
Department of Physics, University of Strathclyde, 107 Rottenrow, Glasgow G4 0NG, UK.
1. Introduction
Extending the spectral coverage of stable laser sources is key to improving the range and
efficiency of applications as diverse as optical microscopy, spectroscopy and remote sensing.
For this reason, nonlinear optical frequency conversion methods are often used to create
coherent sources with emission wavelengths that are difficult or impossible to access with
standard laser gain media [1-3]. Stimulated Raman scattering (SRS) is one-such powerful yet
simple example of a nonlinear method that exploits the third-order material response to evoke
an intensity-dependent Raman shift of the input wavelength. This frequency shift to a longer
wavelength increases the spectral range of the existing platform source. We report the
application of periodically poled lithium niobate (PPLN) to perform simultaneous SRS and
second harmonic generation (SHG) to extend the wavelength range of a soliton self frequency
shifted (SSFS) Nd3+:YLF laser. Use of a PPLN crystal provides access to a very high
nonlinear gain coefficient for high efficiency of the SRS and SHG processes. In the simple
single-pass configuration adopted, conversion efficiencies of 19% and 12% were measured
for the SHG and SRS outputs respectively. Changing the interaction period and the
temperature of the PPLN crystal created a SHG signal that was tunable over λ=584 nm to
λ=679 nm. The concurrently generated SRS output spanned a full-width at half-maximum
(FWHM) of λ=85 nm, centred at λ=1583 nm.
2. Background
Periodically poled crystals are invaluable media for performing nonlinear optical frequency
conversion as they provide access to high nonlinear coefficients, possess high damage
thresholds and have simple phase-matching requirements [4]. The quasi-phase-matching
(QPM) procedure typically involves applying a high-magnitude electric field in a predetermined
pattern to a ferroelectric material such as lithium niobate, resulting in alternating
polarity domain periods of length Λ, which is twice the coherence length for this process.
After one coherence length, the wave-vector mismatch ∆k is reset to zero by reversing the
sign of the nonlinear coefficient, leading to a constructive conversion process. For the
specific case of periodically poled lithium niobate (PPLN), use of the d33 tensor component in
this arrangement allows second-order nonlinear coefficients of up to 16 pm/V to be accessed
[5]. This is a five-fold increase over the effective nonlinear gain coefficient of bulk lithium
niobate using birefringence phase matching. Furthermore, the acceptance bandwidth for
QPM structures is typically larger than those for conventionally phase-matched devices since
QPM allows the use of the same polarization for the input and generated frequencies [6].
Solid-state nonlinear materials such as PPLN are well-suited to efficiently generating
SRS radiation as the nonlinear gain coefficient is higher than in fluid media due to an increase
in the concentration of Raman scattering centres [7]. Although there are many reports of SRS
in bulk nonlinear materials, reports on SRS in periodically poled materials are currently
limited. Pasiskevicius et al [8] described SRS in an optical parametric oscillator (OPO) based
on periodically poled KTiOPO4 (PPKTP). PPLN has a substantially higher third-order
nonlinear gain coefficient than PPKTP [5, 9] and therefore potentially enables more efficient
SRS. The use of a material with a higher nonlinear gain coefficient also means that more
efficient single-pass SRS can be performed. Application of this more straightforward
experimental strategy circumvents possible stability problems caused by OPO cavity length
tolerances.
In a study by Sidorov et al [10], substantial Raman peaks in bulk lithium niobate were
observed in the X(ZZ)Y orientation. This is the standard orientation adopted in PPLN
crystals, as it provides access to the highest nonlinear coefficient. We therefore employed a
PPLN crystal cut in the X(YY)Z plane as a basis for our observations and study of
simultaneous SRS and SHG in PPLN.
The SRS process is acknowledged to be extremely sensitive to the wavelength of the
pump source [8]. We therefore exploit the SSFS effect in photonic crystal fibre (PCF) to
create a suitable high peak power source for investigating the SRS process in PPLN. The
SSFS occurs when Raman self-pumping of a soliton transfers energy from shorter to longer
wavelengths, as described recently by Reid et al [11] in PCF. The overall effect is a
nonlinearly dependent red-shift of the pulse centre wavelength, creating a soliton-pulsed laser
source emitting at longer wavelengths than the input source.
3. Experiment
The experimental arrangement for instigating SRS and SHG in PPLN involved a horizontally
polarised continuous-wave mode-locked Nd3+:YLF laser (Biolight, Coherent) and PCF
serving as the platform source, as shown in Fig. 1. The Nd3+:YLF source comprised a
picosecond-pulsed laser with an externalfiber-grating pulse compressor. The average output
power from the compressor was 680 mW. The emission wavelength of the source from the
compressor was λ=1047 nm with a 6 nm spectral FWHM, measured using a fibre-coupled
spectrometer (Ocean Optics). After compresson, pulses of 400 fs duration were measured
using a second harmonic autocorrelator, corresponding to a time-bandwidth product of 0.68.
Given the source repetition rate of 120 MHz, the peak power of the source was therefore
approximately 14 kW.

This pump light was propagated through a λ/2 plate that was anti-reflection coated for the
pump wavelength. Rotating the half-wave plate in the set-up and hence changing the state of
the pump light entering the fibre was used to determine the polarization dependence on the
efficiency of the SSFS process. The transmitted light was focused into an 84 cm long section
of photonic crystal fibre (PCF) using an aspheric lens of focal length f=+8 mm with a
numerical aperture of N.A.=0.5. The average power incident on the PCF was measured to be
480 mW. The PCF used in this investigation (Crystal Fibre A/S) had a hexagonal
arrangement of air holes surrounding a 3.5 µm diameter core. The separation between
neighbouring air holes was 1.8 ± 0.2 µm, with a pitch of approximately 0.35. This resulted in
a high non-linearity fibre with a zero-dispersion wavelength of around λ0=800 nm. At the
pump wavelength, the fibre exhibited a low and anomalous dispersion. The transmitted output
from the fibre was collimated using another aspheric lens of f=+4.5 mm focal length lens and
N.A.=0.55. Neither the focusing nor collimating aspheric lenses were anti-reflection coated at
the pump or SSFS wavelengths and therefore contributed to Fresnel losses. The collimated
PCF output was focused using a spherical lens with an anti-reflection coating at λ=1064 nm
of focal length f=+40 mm into a plane-faced PPLN crystal (Crystal Technologies). The spotsize
within the PPLN was approximately 31 µm, matching the theoretically optimum beam
size for second harmonic generation of the SSFS light calculated using the Boyd and
Kleinman method [12]. The uncoated PPLN crystal was 0.5 mm-thick and 6.5 mm-long, and
was cut along the X(ZZ)Y direction to provide access to the highest nonlinear gain
coefficient. The crystal comprised five regular grating periods of Λ = 10 µm to Λ = 12 µm in
0.5 µm increments. This period length was intentionally chosen to maximise SHG of the
SSFS output. The crystal was heated in a custom-built oven with 0.2 °C accuracy to 110 °C
to minimise photorefractive damage observed at lower temperatures [13]. Fast wavelength
changes were made possible by changing the period of the PPLN interacting with the input
beam by translating the crystal relative to the source.
4. Results
The total average power measured at the PCF output (transmission) was 211 mW, with the
reduction in power attributed to non-optimal and uncoated fibre input coupling optics and
PCF loss. The SSFS average power was measured using two long-wave pass interference
filters with a total transmission of 35% at λ>1135 nm and a calorimeter.

At wavelengths longer than λ=1135 nm, with optimum half-wave plate orientation, up to 56
mW of average power was measured, corresponding to a generated average power of 160
mW of frequency shifted light. A relative average power decrease of up to 24% was
measured when the half-wave plate was rotated through 90ş, as shown in Fig. 2. The
maximum conversion efficiency from pump to SSFS radiation was therefore 24%.

A maximum shift peak at λ=1258 nm with a spectral FWHM of λ=91 nm was measured using
a fibre-coupled spectrometer (Hewlett-Packard). Figure 3 shows an example of the recorded
trace for 160 mW of SSFS radiation. It was noted that when the pulse duration delivered by
the Nd3+:YLF laser was increased by varying the grating spacing in the pulse compressor, the
SSFS spectrum became increasingly narrow. This was not symmetric about the peak
wavelength. Instead, the longer wavelength components of the SSFS spectrum disappeared.
For the remainder of this report, the optimum pulse duration possible from the Nd3+:YLF for
maximum average power SSFS generation was employed.
In order to determine the pulse duration of the SSFS output from the PCF, a collinear
scanning autocorrelator based on two-photon absorption in a biased GaAsP photodiode
(Hamamatsu) was used. Pulses of 220 fs FWHM duration were measured, as shown in Fig. 4.
Assuming a sech2 pulse shape, this indicated pulses of approximately 140 fs duration. From
the spectral and power data, this measured pulse duration results in a time-bandwidth product
of 2.47 and a calculated peak power of around 9kW. The SSFS source served as the platform
for studying single-pass SRS and SHG in PPLN.

Collinear SHG of the SSFS source was observed in PPLN. Continuously tunable output from
λ=584 nm to λ=679 nm was measured using a fibre-coupled optical spectrometer with a
resolution of 1 nm (Ocean Optics). Temperature tuning the PPLN crystal from 80 şC to 170
şC and changing the PPLN grating period achieved this broad tuning range. Figure 5 shows a
typical tuning range measured by varying the PPLN period, at a fixed temperature of 110 şC.

By changing the PPLN period, the spectral FWHM varied from 2.8 nm to 4.2 nm. An
average power maximum at λ=628 ± 3.1 nm of 30 mW was measured using a calorimeter and
optical bandpass filter (Chroma), which is 19% efficient conversion from the SSFS source.
This correlates with the spectral range of the fundamental SSFS radiation as already described
in Fig. 3. Additionally, 39 mW of radiation at λ=1047 nm and 24 mW of radiation at
wavelengths longer than λ=1135 nm were measured in crystal transmission. For input pulses
of 140 fs duration, the effective crystal length should ideally be sub-mm for maximum second
harmonic conversion. However, the application of longer crystals is understood to reduce the
threshold and increase the efficiency of the SRS process [7]. For this reason and also for ease
of handling, the 6.5 mm long crystal was retained, although we acknowledge that this
ultimately reduced the SHG conversion efficiency.
Using the SSFS output with a spectral maximum at λ=1258 nm as previously described,
at a fixed temperature of 110 şC the SRS spectral maximum was centred at λ=1583 nm (632
cm-1) and had a FWHM of λ=85 nm. This is shown in Fig. 6. The SRS output was observed
to propagate noncollinearly with the SSFS input beam at an angle of 28° with a 4 mW
threshold, and was therefore easily distinguishable from the generated SHG. With full
incident power at the SSFS wavelengths focussed into the PPLN crystal, the power in the SRS
spectrum was measured to be 18 mW across the measurable spectral range. The conversion
efficiency from SSFS to SRS radiation was therefore calculated to be 12%. The second
harmonic of the SRS radiation was also emitted collinearly with the SRS output, at an average
power of approximately 1 mW. With the long-wave pass filters placed prior to the PPLN
crystal, the SRS signal was reduced in magnitude by almost a factor of two. This was
attributed to the 65% decrease in peak power delivered to the PPLN crystal.

5. Conclusion
In conclusion, we have demonstrated simultaneous SHG and noncollinear SRS in PPLN.
This system was based on SSFS in PCF providing the pump wavelengths necessary for the
concurrent SHG and SRS. A simple single-pass geometry provided conversion efficiencies of
the SSFS radiation to SHG and SRS of 19% and 12% respectively. The SHG process created
a tunable source from λ=584-679 nm and the SRS output was at a peak wavelength of
λ=1583, with a spectral FWHM of λ=85 nm.
With the application of a shorter crystal, the SHG average power would increase but the
SRS efficiency would be reduced. One approach to solving this problem would be to power
scale the initial laser platform, where the output powers of both the SHG and SRS processes
could feasibly reach >100 mW at the novel spectral regions described. Additionally, with
increased interest in understanding and exploiting the nonlinear processes in PCF, we
envisage the future development of a more efficient SSFS source to study simultaneous SHG
and SRS. Applying high-quality anti-reflection coatings to both the PPLN and the aspheric
lenses would also increase the efficiency of both the SRS and SHG processes. Future work
enabled by this novel wavelength source includes remote sensing of CO2 in the L-band, which
extends from 1565 nm to 1620 nm, and further spectroscopy applications.
Acknowledgments
This work was supported by the Royal Society of Edinburgh.
References
1. L.E.
Myers and W.R. Bosenberg, “Periodically poled lithium niobate and
quasi-phase-matched
optical parametric oscillators,” IEEE J. Quantum Electron.
33 1663-1672 (1997).
2. D.W.
Chen and K. Masters, “Continuous-wave 4.3-mu m intracavity difference
frequency
generation in an optical parametric oscillator,” Opt. Lett. 26 25-27 (2001).
3. W.R.
Bosenberg, J.I. Alexander, L.E. Myers and R.W. Wallace, “2.5-W, continuous-wave,
629-nm solidstate laser source,” Opt. Lett. 23 207-209 (1998).
4. M.M.
Fejer, G.A. Magel, D.H. Jundt and R.L. Byer, “Quasi-phase-matched 2nd
harmonic-generation
– Tuning and tolerances,” IEEE J. Quantum Electron. 28
2631-2654 (1992).
5. I.
Shoji, T. Kondo, A. Kitamoto, M. Shirane and R. Ito, “Absolute scale of
second-order nonlinear-optical
coefficients,” J. Opt. Soc. of Am. B. 14
2268-2294 (1997).
6. F. Laurell, “Periodically poled materials for miniature light sources,” Opt. Mat. 11 235-244 (1999).
7. P.G.
Zverev, T.T. Basiev and A.M. Prokhorov, “Stimulated Raman scattering of laser
radiation in Raman crystals,”
Opt. Mat. 11 335-352 (1999).
8. V.
Pasiskevicius, A. Fragemann, F. Laurell, R. Butkus, V. Smilgevicius and A.
Piskarskas, “Enhanced stimulated
Raman scattering in optical parametric
oscillators from periodically poled KTiOPO4,” App. Phys. Lett. 82 325-327
(2003).
9. I.
Jovanovic, J.R. Schmidt and C.A. Ebbers, “Optical parametric chirped-pulse
amplification in periodically poled
KTiOPO4 at 1053 nm,” App. Phys. Lett. 83
4125-412 (2003).
10. N.V.
Sidorov, M.N. Palatnikov, K. Bormanis and A. Sternberg, “Raman spectra and
structural defects of lithium
niobate crystals,” Ferroelectrics 285 685-694
(2003).
11. D.T.
Reid DT, I.G. Cormack, W.J. Wadsworth, J.C. Knight, PSJ Russell, “Soliton
self-frequency shift effects in
photonic crystal fibre,” J. Mod. Opt. 49 757-767
(2002).
12. G.D.
Boyd and D.A. Kleinman, “Parametric interaction of focused Gaussian light
beams,”
J. App. Phys. 39 3597-3639 (1968).
13. I.A. Ghambaryan, R. Guo, R.K. Hovsepyan, A.R. Poghosyan, E.S. Vardanyan and V.G. Lazaryan, “Periodically poled structures in lithium niobate crystals: Growth and photoelectric properties,” J. Optoelectron. Adv. Mat. 5 61-68 (2003).