1142 Vol. 41, No.

6 / March 15 2016 / Optics Letters Letter

High-efficiency optical parametric chirped-pulse

amplifier in BiB3O6 for generation of 3 mJ,
two-cycle, carrier-envelope-phase-stable
pulses at 1.7 μm
YANCHUN YIN,1,† JIE LI,1,† XIAOMING REN,1,† KUN ZHAO,1,2 YI WU,1 ERIC CUNNINGHAM,1 AND ZENGHU CHANG1,*
1
Institute for the Frontier of Attosecond Science and Technology, CREOL and Department of Physics, University of Central Florida, Orlando,
Florida 32816, USA
2
Beijing National Laboratory for Condensed Matter Physics, Institute of Physics, Chinese Academy of Sciences, Beijing 100190, China
*Corresponding author: Zenghu. Chang@ucf.edu

Received 15 December 2015; revised 29 January 2016; accepted 1 February 2016; posted 2 February 2016 (Doc. ID 255667);
published 8 March 2016

We produce a 3 mJ, two-cycle (11.4 fs), 1 kHz, carrier- lasers, which allows a trade-off between the gain bandwidth and
envelope phase (CEP)-stable laser source at 1.7 μm via a the damage threshold of the nonlinear crystals [7,8]. While a
three-stage Ti:sapphire-pumped optical parametric chirped- number of kHz-repetition rate, few-cycle, high-energy IR
pulse amplifier in BiB3 O6 . We achieve a pump-to-signal OPCPA sources have been reported [9–13], only two of these
conversion efficiency of 18% in the last stage, which is, to systems yielded mJ-level near-octave pulses (two cycles and be-
the best of our knowledge, the highest yet achieved for near- low): the 3 kHz, 1.2 mJ, 2.1 μm source pumped by a Yb:YAG
octave bandwidth amplification. A f-to-2f measurement laser in [10] and the 1 kHz, 1.5 mJ, 1.6 μm source pumped by
shows a CEP instability of 165 mrad over 1 h. This is an a Ti:sapphire laser in [13]. While commercially available Yb:
ideal light source for generating isolated attosecond pulses YAG pump lasers with hundreds of mJ of pump energy
in the soft x-ray region. © 2016 Optical Society of America may facilitate OPCPA output energies scaled to tens of mJ
in the future, the pump-to-signal conversion efficiency of the
OCIS codes: (320.7090) Ultrafast lasers; (190.4970) Parametric
Yb:YAG-pumped near-octave OPCPA in [10] was limited to
oscillators and amplifiers. around 7%. On the other hand, the previously reported highest
http://dx.doi.org/10.1364/OL.41.001142 efficiency in a Ti:sapphire-pumped near-octave OPCPA was
12%, achieved in a BiB3 O6 (BIBO)-based system [13]. Better
understanding of the Ti:sapphire pumping scheme is interest-
ing, since researchers can develop such an OPCPA either by
Attosecond science, emerging at the onset of the 21st century, making use of the existing Ti:sapphire lasers in their labs or
has opened a new era of understanding electron dynamics and by taking advantage of the mature high-energy Ti:sapphire
correlations which occur on attosecond timescales [1]. So far, products that have become available from several companies.
isolated attosecond pulses as short as 67 as have been generated In this Letter, we report a method for further increasing the
in an extreme ultraviolet (XUV) wavelength range by using pump-to-signal conversion efficiency of a Ti:sapphire-pumped
spectrally-broadened Ti:sapphire lasers operating in the visible near-octave, mJ-level OPCPA to 18% by designing the ideal
to near-infrared spectral region [2]. Rapid development of Ti:sapphire pumping bandwidth and center wavelength for
few-cycle, long-wavelength driving lasers has occurred in recent optimizing the complex phase matching (PM) between a
years to further reduce the pulse duration toward 24 as (one linearly stretched polychromatic pump pulse and a nonlinearly
atomic unit of time) and cover the water-window region of stretched seed pulse. The implementation of this technique in
the soft x-ray domain for ultrafast chemistry, condensed matter, a BIBO-based OPCPA yielded 3 mJ, two-cycle, 1 kHz, carrier-
and biologic applications [3]. Such light sources, however, are envelope phase (CEP)-stable pulses at 1.7 μm.
limited to low brightness because of the low pulse energy of In BIBO, it has been calculated previously that 780 nm light
available few-cycle IR driving lasers and the dramatically is optimal for pumping an octave-spanning spectrum centered
reduced single-atom response due to the quantum diffusion at 1.6 μm when using Type 1 PM [14]. Rather than tailoring
of electron wave packets [4–6]. Therefore, the development the pump bandwidth and central wavelength for optimizing
of high-energy few-cycle IR laser sources is in strong demand. PM conditions, however, many BIBO-based optical parametric
IR pulses with both high-pulse energies and few-cycle amplification (OPA) and OPCPA systems have been content
durations are primarily enabled by optical parametric chirped to operate with pump wavelengths centered instead around
pulse amplification (OPCPA) pumped by few-picosecond (ps) 800 nm, which corresponds conveniently to the peak gain

0146-9592/16/061142-04 Journal © 2016 Optical Society of America

 Letter Vol. 41, No. 6 / March 15 2016 / Optics Letters 1143

of room-temperature Ti:sapphire [11,14,15]. This is reasonable

for OPA systems, because the specifications for pump wave-
lengths are less stringent: since the near-transform-limited
pump and signal pulses allow all pump wavelengths to be
phase-matched with the signal wavelengths simultaneously,
PM in the temporal domain is essentially uniform and straight-
forward. In OPCPA systems, however, the pump spectrum
plays a more crucial role because the pump and signal pulses
are both chirped, with PM inextricably occurring only between
suitable spectral portions of the pump and the signal pulses that
also overlap in the temporal domain.
With this in mind, the optimal selection of the Ti:sapphire Fig. 1. The calculated PM efficiency as a function of pump and
signal wavelengths for (a) the collinear configuration (PM angle
pumping bandwidth and central wavelength in a BIBO-based
θ  10.9°, noncollinear angle α  0°); and (b) the noncollinear con-
OPCPA system involves two critical aspects:
figuration (PM angle θ  10.82°, noncollinear angle α  0.6°). The
1. Dispersion management. Gratings are not considered black line represents the chirped signal pulse with its corresponding
because of the relative high loss and the CEP variation that wavelengths marked by the left axis and its temporal chirp marked
requires active stabilization [16]. Chirped mirrors are only good by the top axis.
to compress a broadband pulse within a few bounces, because
too many bounces will decrease pulse contrast significantly and
also cause a high loss. Here the scheme of an acousto-optic pro- relatively narrow, as suggested by the unshaded area. To broaden
grammable dispersive filter (AOPDF) as the stretcher plus a the phase-matchable pump bandwidth, a noncollinear angle α
bulk material as a compressor is employed due to the adaptive can be introduced to the BIBO in order to adjust the PM rela-
phase control by AOPDF and high transmission and compact- tionship between the pump and signal wavelengths. Figure 1(b)
ness of the bulk material. Unfortunately, many bulk materials shows the noncollinear configuration with α  0.6°, with the
have a zero-dispersion point near the short-wavelength side of unshaded area indicating that the region of good PM (red con-
the desired spectrum, such as SF-57 at 1.8 μm and fused silica tours) is stretched over a broader range of pump wavelengths
at 1.3 μm, which results in large higher order dispersions. IR compared to the collinear case in Fig. 1(a). Therefore, this opti-
materials such as ZnSe do not have a zero-dispersion point over mized noncollinear configuration helps reduce the constraints
the desired spectrum, but they usually have very large nonlinear on the pulse stretcher in the pump-producing CPA system.
refractive index n2 , relatively low bandgap, and nonvanishing To further study how the pump bandwidth influences the
second-order nonlinearity χ
2. As fused silica (small n2 and large OPCPA gain and its bandwidth, a 1D three-wave mixing
bandgap) is chosen as the compressor, the OPCPA seed pulse numerical model has been developed by modifying the previ-
must be nonlinearly stretched to match the inherent material ously developed one in [17]: (1) the model includes the effects
dispersion, which complicates the PM in the temporal domain. of dispersion and nonlinear refractive index and excludes the
Thus, the nonlinear chirp must be accounted for when choosing effects of diffraction; (2) all three waves are assumed to be plane
the bandwidth and central wavelength of the pump laser. waves; and (3) the model only considers the OPCPA process
2. Design of pump bandwidth. While a narrow-band pump discussed herein. The value of the effective second-order non-
is more convenient for PM, stretching such a weakly polychro- linearity d eff is equal to 2.11 pm/V. The signal pulse (1.2–
matic pulse to several hundred ps–necessary to avoid damaging 2.2 μm) is stretched to 4.4 ps between 1.3 μm and 2.2 μm.
the gain medium in the pump-producing chirped-pulse ampli- The FWHM pump pulse is linearly chirped to 5 ps, under
fier (CPA)–requires a large stretcher-compressor design. Thus, which the damage threshold is around 20 GW∕cm2 deter-
an appropriately broad pump bandwidth can facilitate the com- mined by the BIBO antireflection coating. The peak intensities
pactness of the CPA without sacrificing the PM efficiency. of the signal and the pump are assumed to be 1.5 MW∕cm2
and 8 GW∕cm2 , respectively, in order to achieve a gain of
To address the first design concern: the PM efficiency, around 1000 in a 5 mm BIBO first-stage amplifier.
sinc2
ΔkL∕2, is plotted as a function of the pump (bottom Figure 2(a) shows the simulated results using three pump spec-
axis) and signal wavelengths (left axis) in Fig. 1(a), where Δk tra: (1) a 730–830 nm Gaussian pump spectrum typical of a
represents phase mismatch and L is the BIBO crystal length home-built Ti:sapphire CPA described hereafter (green), (2) a
equal to 5 mm. The color contours represent the PM efficiency 775–785 nm super-Gaussian spectrum for best PM (red), and
supported by a collinearly configured BIBO crystal with a (3) a 760-810 nm super-Gaussian spectrum that represents
phase-matching angle θ  10.9°, and the black line demon- a compromise between the other two (blue). The power spec-
strates the group delay for a stretched signal pulse overlapped trum shapes of the pump laser are similar to the corresponding
with a linearly stretched pump pulse for PM in the temporal pump pulse shapes in the temporal domain shown in Fig. 2(b).
domain (top axis). The most distinct feature of the stretched Compared to the ideal case of spectrum (2), the signal spec-
signal pulse is that the shorter wavelengths cannot be stretched trum pumped by spectrum (1) is gain-narrowed and the gain is
as much as the longer wavelengths, due to the zero-dispersion >25% smaller–both due to the poor PM at the wings of the
point of the fused silica bulk compressor near 1.3 μm. This sets spectrum and the nonflattop pump pulse shape, which is indi-
a reference for determining the proper pump wavelengths for cated in Fig. 2(b). On the other hand, the amplified signal spec-
good PM in the temporal domain. trum using spectrum (3) is similar to the one pumped by the
To address the second design concern: Fig. 1(a) also demon- ideal spectrum (2), with the gain only being <4% smaller–a mat-
strates that the pump bandwidth for achieving good PM is ter easily compensated by using either a slightly higher pump
 1144 Vol. 41, No. 6 / March 15 2016 / Optics Letters Letter

(DFG), sum-frequency generation, and parasitic second-

harmonic generation, are not included in the model.
To meet the requirements for such an OPCPA system, a
three-stage Ti:sapphire CPA system is employed, as represented
by the left side of Fig. 3. Nanojoule-level oscillator pulses (730–
830 nm) are stretched to 360 ps in an Offner-type stretcher
before the pulse energy is boosted to 4 mJ with a homemade
14-pass Ti:sapphire amplifier system. The central wavelength
does not shift, because the Ti:sapphire crystal is cooled to
173 K by the cold finger of a liquid immersion chiller. The
4 mJ from the 14-pass amplifier is split into two beams. First,
a 1.4 mJ portion is chopped to 760–810 nm with a flattop
shape by a transmission filter; an energy of 21 mJ is achieved
after two additional cryocooled amplifiers–a 6-pass and a 3-
pass–before a transmission grating pair compresses the pulse
to 5 ps (FWHM) with 18 mJ. This beam is split into three
portions for pumping the three OPCPA amplifiers described
Fig. 2. Simulation of a first-stage OPCPA in a 5 mm BIBO: (a) in- in the next paragraph. Second, the remaining 2.6 mJ portion
put signal spectrum (black) and output signal spectrum when pumped is compressed to 2.2 mJ, 30 fs by a transmission grating pair
by 775–785 nm (red), by 760–810 nm (blue), and by 730–830 nm and then sent to a hollow-core fiber (HCF) filled with 30 psi
(green); (b) temporal shape of the signal input pulse (black), the 775– neon for white light generation. The white light is compressed
785 nm and 760–810 nm pump pulse (purple), and the 730–830 nm
to ∼7 fs using chirped mirrors. A broadband IR (1.2–2.2 μm)
pump pulse (green); (c) and (d) simulation of a third-stage OPCPA in
seed with a pulse energy of around 1 μJ is generated via intra-
a 3 mm BIBO.
pulse DFG of the white light in a 1 mm type-II BIBO crystal
(PM angle θ  60°). The seed is then stretched to 4.4 ps using
an AOPDF with an output energy of ∼50 nJ.
intensity or a thicker crystal. Thus by using a bandpass filter to The OPCPA is composed of three stages, each one consist-
cut the 730–830 nm Gaussian spectrum to a 760–810 nm ing of a Type 1 BIBO crystal with a PM angle of θ  10.8° and
super-Gaussian spectrum, the home-built Ti:sapphire CPA a small noncollinear angle of α  0.6°. Both the first and sec-
can be outfitted for pumping a high-efficiency OPCPA without ond stages have 5 mm-thick BIBO crystals and are pumped by
gain narrowing. Meanwhile, with a 360 ps (730–830 nm) 360 μJ and 3.2 mJ, respectively. The output pulse energy of
stretcher-compressor design, a 180 ps pulse (760–810 nm) is these stages are 30 μJ and 400 μJ, correspondingly. The second-
safe enough to avoid damage in the high-energy CPA system. stage output is further boosted from 400 μJ to 3 mJ in a 3 mm-
To study the conversion efficiency that can be achieved in thick BIBO pumped by 14.4 mJ pulse energy in the third
such an OPCPA pumped by a 760–810 nm super-Gaussian la- OPCPA stage. The spectra at each stage are shown in Fig. 4.
ser, we simulated the third-stage amplifier, as shown in Figs. 2(c) In this last stage, the highest energy (3 mJ) and conversion
and 2(d). The peak intensities for the signal and pump are efficiency (18%) among any current near-octave OPCPA
0.8 GW∕cm2 and 7 GW∕cm2 , respectively. The BIBO thick- are achieved. The IR pulse is compressed in a 150 mm fused
ness is 3 mm. With a high pump depletion, a conversion efficiency silica bulk compressor to a FWHM pulse duration of 11.4 fs
of 28% is achieved at the expense of a slightly gain-narrowed spec- as retrieved by a home-built second-harmonic-generation
trum. The experimentally achievable conversion efficiency is ex- frequency-resolved optical gating (SHG FROG), whose trace
pected to be lower, since it is determined by spatially integrated is depicted in Fig. 5.
energy extraction from the pump to the signal and since other For CEP measurements, a single-shot f-to-2f interferometer
nonlinear processes, such as difference-frequency generation is utilized in which a sapphire plate generates a white light

2.2 mJ, 30 fs Hollow core fiber with 30 psi neon

Ti:Sa Osc.
80 MHz, 8fs
Chirped
Si BW BIBO DFG
mirrors
Ti:Sa
14-pass Amp BIBO 1 BIBO 2 BIBO 3
50 nJ 30 µJ 400 µJ
Filter
0.36 mJ 3.2 mJ 14.4 mJ 3 mJ, 11.4 fs
Ti:Sa
6-pass Amp
Delay 1 Delay 2 Delay 3
Fused silica
BS 2
Ti:Sa 18 mJ, 5 ps BS 1
3-pass Amp

Fig. 3. Schematic setup of the OPCPA system. BS1, 20% reflection beam splitter; BS2, 10% reflection beam splitter; Si BW, silicon window at
Brewster’s angle.
 Letter Vol. 41, No. 6 / March 15 2016 / Optics Letters 1145

To conclude, a high-efficiency, 3 mJ, two-cycle, 1 kHz, and

CEP-stable IR OPCPA light source was developed with a cen-
tral wavelength at 1.7 μm. Scaling of the performance of this
light source to over 10 mJ pulse energy is expected in the near
future by enhancing the Ti:sapphire output energy to 60 mJ.
Such a laser system paves the way for the generation of a bright
broadband soft x-ray continuum covering the water-window
region (280–530 eV), thus enabling attosecond pulses shorter
than the atomic unit of time (24 as). This attosecond source
itself poses a groundbreaking opportunity for ultratime-
resolved spectroscopy of biologically relevant samples and
strongly correlated solid-state materials.

Funding. Division of Physics (PHY) (1068604); Army

Research Office (ARO) (W911NF-14-1-0383); Air Force
Office of Scientific Research (AFOSR) (FA9550-15-1-0037);
Defense Advanced Research Projects Agency (DARPA)
(W31P4Q1310017).
Fig. 4. Spectra taken at successive stages of the laser system.
Acknowledgment. The authors thank Prof. Jiro Itatani,
Prof. Nobuhisa Ishii, and Dr. Yunpei Deng for sharing their
experience in few-cycle OPCPA.
1.0 1.0 1.0 †
Wavelength (µm)

Wavelength (µm)

1.0
(a) (b) These authors contributed equally to this work.
0.9 0.5 0.9 0.5

0.8 0 0.8 0
REFERENCES
0.7 0.7
-200 -100 0 100 200
1. Z. Chang and P. Corkum, J. Opt. Soc. Am. B 27, B9 (2010).
-200 -100 0 100 200
Delay (fs) Delay (fs) 2. K. Zhao, Q. Zhang, M. Chini, Y. Wu, X. Wang, and Z. Chang, Opt. Lett.
37, 3891 (2012).
(c) (d) 3. F. Silva, S. M. Teichmann, S. L. Cousin, M. Hemmer, and J. Biegert,
Nat. Commun. 6, 6611 (2015).
FWHM 4. B. Shan and Z. Chang, Phys. Rev. A: At., Mol., Opt. Phys. 65, 011804
11.4 fs (2001).
5. J. Tate, T. Auguste, H. G. Muller, P. Salières, P. Agostini, and L. F.
DiMauro, Phys. Rev. Lett. 98, 013901 (2007).
6. A. D. Shiner, C. Trallero-Herrero, N. Kajumba, H.-C. Bandulet, D.
Comtois, F. Légaré, M. Giguère, J.-C. Kieffer, P. B. Corkum, and
Fig. 5. (a) Experimental SHG FROG trace; (b) retrieved SHG D. M. Villeneuve, Phys. Rev. Lett. 103, 073902 (2009).
FROG trace; (c) independently measured spectrum (black), retrieved 7. H. Fattahi, H. G. Barros, M. Gorjan, T. Nubbemeyer, B. Alsaif, C. Y.
spectrum (red), and retrieved spectral phase (green); (d) retrieved pulse Teisset, M. Schultze, S. Prinz, M. Haefner, M. Ueffing, A. Alismail,
(red) and temporal phase (green). L. Vámos, A. Schwarz, O. Pronin, J. Brons, X. T. Geng, G. Arisholm,
M. Ciappina, V. S. Yakovlev, D.-E. Kim, A. M. Azzeer, N. Karpowicz,
D. Sutter, Z. Major, T. Metzger, and F. Krausz, Optica 1, 45 (2014).
8. A. Vaupel, N. Bodnar, B. Webb, L. Shah, and M. Richardson, Opt.
continuum, and a BBO crystal yields the frequency-doubled Eng. (Bellingham, Washington, U.S.) 53, 051507 (2014).
component. The CEP fluctuation is measured to be 165 mrad 9. O. D. Mücke, S. Ališauskas, A. J. Verhoef, A. Pugžlys, A. Baltuška,
(RMS) after 1 h of single-shot collection, as shown in Fig. 6. V. Smilgevičius, J. Pocius, L. Giniūnas, R. Danielius, and N. Forget,
Opt. Lett. 34, 2498 (2009).
10. Y. Deng, A. Schwarz, H. Fattahi, M. Ueffing, X. Gu, M. Ossiander, T.
Metzger, V. Pervak, H. Ishizuki, T. Taira, T. Kobayashi, G. Marcus, F.
Krausz, R. Kienberger, and N. Karpowicz, Opt. Lett. 37, 4973 (2012).
11. N. Ishii, K. Kaneshima, K. Kitano, T. Kanai, S. Watanabe, and
J. Itatani, Opt. Lett. 37, 4182 (2012).
12. K.-H. Hong, C.-J. Lai, J. P. Siqueira, P. Krogen, J. Moses, C.-L.
Chang, G. J. Stein, L. E. Zapata, and F. X. Kärtner, Opt. Lett. 39,
3145 (2014).
13. N. Ishii, K. Kaneshima, T. Kanai, S. Watanabe, and J. Itatani,
Conference on Lasers and Electro-Optics (CLEO: S&I) (Optical
Society of America, 2015), paper SF1M.3.
14. A. Gaydardzhiev, I. Nikolov, I. Buchvarov, V. Petrov, and F. Noack,
Opt. Express 16, 2363 (2008).
15. I. Nikolov, A. Gaydardzhiev, I. Buchvarov, P. Tzankov, F. Noack, and
V. Petrov, Opt. Lett. 32, 3342 (2007).
16. C. Li, E. Moon, and Z. Chang, Opt. Lett. 31, 3113 (2006).
Fig. 6. Top: f-to-2f interferograms collected for 1 h; bottom: CEP 17. Y. C. Yin, D. French, and I. Jovanovic, Opt. Express 18,
fluctuations (165 mrad RMS) in 1 h. 18471 (2010).