We report an experimental design for two-dimensional electronic spectroscopy (2D-ES) that avoids the need to measure notoriously weak pump-probe spectra. Retaining a fully non-collinear folded boxcar geometry, the described layout replaces pump-probe with heterodyned transient grating (het-TG). The absorptive component of the het-TG signal is measured directly, following a straightforward optimization routine. The use of het-TG achieves an improvement in signal to noise ratio by almost two orders of magnitude. As a result, 2D-ES-signals down to 0.5% can be clearly resolved.
©2013 Optical Society of America
Two-dimensional electronic spectroscopy (2D-ES)  is a modern, multifunctional technique in the ultrafast spectroscopy toolbox and is employed by a growing number of research groups. 2D-ES measures correlation between excitation- and emission-frequencies (ω1 and ω3, respectively), to provide further insights when compared to ω1-integrated techniques such as pump-probe (PP) or transient grating (TG). This is achieved by resolving so-called cross peaks (ω1 ≠ ω3). Cross peak dynamics describe the properties of various processes, e. g. energy transfer pathways in photosynthesis , excitonic dynamics in semiconductors  and molecular J-aggregates , or vibronic and vibrational coherences in molecules .
In 2D-ES, the excitation frequency axis ω1 is retrieved by Fourier-transformation (FT) from the time domain along t1, where t1 is the delay between the first two pulses in a four-wave-mixing sequence. The third pulse arrives after a so-called population time t2 and sets the system back into a state of coherence, which radiates for a time t3. FT along t3 is achieved in the experiment by frequency-resolved detection of the emerging signal. FT along t1 is more challenging, as the pulses separated by t1 have to be stable in phase for a meaningful FT. There are several reported methods to achieve such phase stability with atto-second accuracy. The first approach used diffractive optical elements (DOEs)  to retrieve two identical pulse replicas and common optics after the DOE to ensure phase stability. Even though DOEs inevitably induce angular dispersion on the excitation pulses, it has been shown  that this effect is reversed if the sample is placed accurately in the focal plane of the collimating mirror or lens. There are several experimental designs for 2D-ES, avoiding the use of DOEs, such as active phase stabilization , pairwise beam manipulation [6, 9], or pulse-shaper assisted layouts, either in a fully linear design , in the PP-geometry [11, 12], or based on diffractive pulse shaping  in folded boxcars geometry. Pulse-shaping approaches to 2D-ES bear many advantages, such as fast data acquisition based on rotating frame detection or direct measurement of the desired real (absorptive) component of the 2D-signal.
One limitation imposed by the PP-phase-matching geometry is the exclusion of double-quantum 2D-signals (2Q-2D). 2Q-2D-signals are useful in determining many-body interactions in semiconductors  or for describing the energy levels of molecules  and aggregates . 2Q-2D spectra require a FT along t2, which is not possible in standard realizations of 2D-ES in PP-phase-matching geometry  as phase locking is not maintained alongt2.Nevertheless, the ability to retrieve absorptive spectra directly is a key advantage for PP-geometry 2D-experiments, as a process called phasing is avoided: In non-collinear phase-matching geometries, the 2D-signal is only measured as its amplitude. To separate real (absorptive) and imaginary components, the signal’s phase has to be determined by heterodyne detection with a local oscillator pulse (LO) . This procedure called phasing relies on the so-called projection slice theorem ,stating that for a common value of t2, the purely absorptive component of a 2D-spectrum, integrated along ω1, equals the absorptive component of ω1-integrated techniques such as PP or TG. PP is typically used for phasing purposes as it inherently measures the absorptive component when ω3 is resolved. Thus, when using a non-collinear phase-matching geometry, it is necessary to record a PP-signal for all measured t2 times. In many cases the notoriously weak PP-signal creates a bottleneck for all 2D-experiments not using phase cycling schemes. Alternatively, the absorptive component of a TG-signal can be employed for phasing with the advantage that the TG-signal can be amplified by an independently tuned LO as for 2D-signals. In this article, we show how a directly measured absorptive component of a heterodyned TG-signal (het-TG) is an excellent alternative to PP for phasing purposes. The increased signal to noise ratio allows for the detection of signals down to the 0.5% level in a molecular J-aggregate (PVA/C8O3) . Furthermore, the described experimental design retains a fully non-collinear folded boxcar phase-matching geometry, in which both single- and double-quantum 2D signals can be measured.
2. Results and discussion
Figure 1 shows the schematic experimental design. The incoming 570 nm/8 fs pulse is generated by a home-built non-collinear parametric amplifier (NOPA)  and is split by a 50:50 beamsplitter (BS). The transmitted pulse is translated by a computer controlled linear motorized stage to introduce delay t2. Both pulses from the BS are then focused by separate spherical mirrors (SM1, radius of curvature = 750 mm) onto separate DOEs.
Two two-inch spherical mirrors (SM2, radius of curvature = 600 mm) are used to collimate the diffracted beams. Section A-A (Fig. 1) shows the employed phase-matching geometry. Pulses along k3and LO + , LO- are 0th and 2nd order diffractions from the variable DOE, while pulses along k1 and k2 are 1st order diffractions from the static DOE (Holoeye). A pair of 2° fused silica wedges (WP2), one of which is mounted on a linear motorized stage, is used to introduce delay t1 between pulses along k1 and k2. All other beams pass identical static wedge pairs (WP1 and WP3) to balance dispersion. The intensity of all five beams can be independently adjusted by variable neutral density filters (not shown). All beams derived from a common DOE, i.e.,k1, and k2 and k3, LO + and LO-,respectively, are steered by the same optics to maintain phase locking between the pulses. Over a 15 min. period, the phase stability was determined to be λ/60. This value compares well to a more compact experimental design from our group  with λ/90 for the same time interval.
In order to suppress unwanted low-frequency contributions, it is advantageous to record 2D-spectra at a sufficiently large signal-LO-delay, such as 540 fs in the present case. If the LO is to be used as a probe in a PP (het-TG)-experiment however, the delay should be set to zero. We achieve this by a custom made substrate (Hellma Optics), placed in the path of the excitation beams as shown in section B-B of Fig. 1. For het-TG and PP-measurements, the plate is positioned to have the same amount of glass in k1, k2, and k3 and the LOs, and therefore 0 fs delay between all pulses. For 2D scans, the plate is moved to a position with 400 µm less glass in k1-3, which lets the LOs precede the excitation pulses by 540 fs.
The three excitation beams are arranged in a triangular geometry to generate 3rd order signals in the direction of the LOs. For pulses separated and ordered in time, the rephasing signal (kI = -k1 + k2 + k3) co-propagates with LO + , while the non-rephasing signal (kII = k1-k2 + k3) emerges along LO-. This allows for the full 2D-spectrum (defined as kI + kII) to be measured using one of two methods: (i) delaying only t1 to negative values (via WP2 in Fig. 1) and collecting the signals along LO + (kI) and LO- (kII); or, similar to previous setups , measuring along a single phase-matching direction and scanning t1 to negative values to obtain kI and t1 and t2 to positive values for kII. We note that the same results cannot be achieved in a square phase-matching geometry as it requires the LOs be of different diffractive orders of the DOEs and brings upon asymmetric phase-matching between kI, LO + and kII, LO-. In addition to the single-quantum signals kI and kII, the double-quantum signal kIII = k1 + k2-k3 can be measured along LO- for positive delays of the pulse along k1.
The desired absorptive component of a 2D-spectrum is retrieved by employing the above mentioned projection slice theorem . In its two-dimensional form, the projection slice theorem states that the projection of a function in frequency space under an angle ϕ and a slice at the same angle ϕ in the time-domain are Fourier-transform pairs. Within the framework of 2D-ES, the 2D-function in frequency-domain is the 2D-signalat a given value of t2, while the time-domain function is the 3rd-ordersignal, a complex quantity, which are related by
The projection for ϕ = 0 leads to t1 = 0. Equation (1) becomes
The left hand side of Eq. (2) is the projection of onto the ω3-axis while the signal for t1 = 0 on the right hand side is proportional to transient grating signalresolved in detection frequency ω3. is complex valued an cannot be measured directly in experiment. To retrieve the desiredit is common practice to replacewith the spectrally resolved PP-signal whereis the probe field in frequency domain. relates to Eq. (2) by
is readily measured in the experiment depicted in Fig. 1 by blocking the pulses along k2 and k3 and using the pulses along k1 and LO + as a pump and probe respectively. The pulse along LO- can serve as a reference pulse if delayed by an additional glass plate (not shown). However, employing for phasing as shown in Eq. (3) brings upon the above mentioned difficulties of low signal strength and can be avoided in the presented setup by measuring the absorptive component ofThis is achieved by heterodyned detection of via a local oscillator field
This shows that a heterodyne detected transient grating signal is proportional to the spectrally resolved pump-probe signal (Eq. (4)) and can be used for purposes of phasing. For Eq. (5) to be useful, the absorptive component of the het-TG signal (right hand side of Eq. (5)) has to be determined experimentally. This procedure, described in the following, was first presented by Donaldson et al . The authors measured absorptive het-TG spectra with UV excitation and IR-probing, whereas in this work we employ all-visible excitation pulses and focus on the advantages of absorptive het-TG signals with respect to phasing of electronic 2D-spectra.
In changing from pump-probe to het-TG, we are replacing an auto-heterodyned signal with a heterodyned one. In other words, the relative phase between pump and probe is irrelevant towhile for het-TG, the phase difference betweenandaffects their heterodyne signalConsidering all experimentally relevant phase terms, we obtain for the measured intensities along LO + and LO- for t1 = 0 fs [22, 23]
In Eq. (6), the local oscillator and signal fields along LO + and LO- are assumed to be equal. ϕDOE is changed by moving the variable DOE into and out of the drawing plane of Fig. 1. ϕWin stems from an adjustable glass window in the path of the pulse along LO-. Rotation of this window will insert more or less glass. At t1 = 0 fs, pulses along k1 and k2 become interchangeable, leading to equal signal fields along LO + and LO-. Splitting into absorptive and dispersive components gives . After trigonometric manipulation, the difference between ILO+ and ILO- then becomes
In Eq. (7)it can be seen if ϕWin is chosen to be zero, the dispersive signal components vanish. The desired EAbs can be further enhanced by optimizing sin(ϕDOE-ϕTG). This means that the difference signal in Eq. (7) yields the absorptive components of a het-TG signal in the same fully non-collinear phase-matching geometry used for electronic 2D-measurements. On a daily basis, the absorptive het-TG signal is optimized via the following steps:
- 1. t1 = 0 fs and calibration of WP2 are determined by interference on a pinhole. This procedure was described in detail previously  and is crucial to the present optimization, as the signals along LO + and LO- are only equal for t1 = 0 fs, turning into rephasing (kI) or non-rephasing signals (kII) for t1 ≠ 0 fs.
- 3. The measured LO-intensities are set equal to by variable neutral density filters to ensure maximal fringe visibility in heterodyned detection.
- 4. For t1 << 0 fs, possible differences between the LO-intensities along LO + and LO- are subtracted as a background measurement.
- 5. The global maximum of the difference signal ILO+-ILO- is found iteratively by systematically varying both the position of the variable DOE (ϕDOE) and the rotation angle of the glass window (ϕWin). We note, for the desired zero phase differences, the signals along LO- and LO + are mirror images of each other and therefore out of phase. For strongly modulated signals, it is often easier to optimize this property rather than the overall amplitude of signal difference, though experimentally challenging.
An example of the data quality achieved using this procedure is given in Fig. 2.
Figure 2(a) shows a comparison between a PP-signal of PVA/C8O3 at t2 = 100 fs and an absorptive het-TG signal recorded at the same excitation pulse energy of 0.7 nJ per pulse (1.4 × 1013 Photons/cm2).The PP-signal was still discernible from noise at a maximal signal of 6 ΔmOD. At the same energy level, the absorptive het-TG trace shows the same spectral features, but an improvement in signal to noise ratio by almost two orders of magnitude (86). Accordingly, the energy per pulse can be reduced to 0.2 nJ (3.9 × 1012 Photons/cm2), while maintaining a relatively noise free absorptive het-TG signal. Comparison of the 2D-signal’s projected absorptive component and the corresponding absorptive het-TG-signal shows that the projection slice theorem holds (Fig. 2(b) and Eq. (5)). Phasing against absorptive het-TG instead of PP enables the phase of the 2D-signal to be determined with significantly higher accuracy. As an example, we plot the absorptive component of the 2D-signal from PVA/C8O3 at t2 = 100 fs in Fig. 3.
The strongest features in the 2D-spectrum in Fig. 3 are the positive ground state bleach and stimulated emission peaks near the diagonal line shown in black. Details of the underlying dynamics can be found elsewhere . The ability to accurately determine the 2D-signal’s phase makes it possible to analyze low-intensity-features. The cross peak between the first (16750 cm−1)and the third peak (17400 cm−1) in the absorption spectrum can be resolved clearly down to the 0.5% signal level. At such low signal levels, pulse propagation and beam geometry effects start to take effect and can be investigated quantitatively in the presented experiment .
Het-TG is an excellent alternative to notoriously weak PP-signals with respect to phasing. The experimental layout presented offers reduced data acquisition time by the simultaneous detection of kI and kII, along with the possibility to measure kIII-signals. The last point is an advantage over 2D-experiments in PP-geometry. The low excitation pulse energies possible in the presented setup are especially beneficial for molecular aggregates, as high-intensity effects such as exciton-exciton annihilation are avoided.
The authors acknowledge helpful discussion with Paul M. Donaldson and funding by the Austrian Science Fund (FWF): START project Y 631-N27.
References and links
2. T. Brixner, J. Stenger, H. M. Vaswani, M. Cho, R. E. Blankenship, and G. R. Fleming, “Two-dimensional spectroscopy of electronic couplings in photosynthesis,” Nature 434(7033), 625–628 (2005). [CrossRef] [PubMed]
3. K. W. Stone, K. Gundogdu, D. B. Turner, X. Q. Li, S. T. Cundiff, and K. A. Nelson, “Two-quantum 2D FT electronic spectroscopy of biexcitons in GaAs quantum wells,” Science 324(5931), 1169–1173 (2009). [CrossRef] [PubMed]
4. J. Sperling, A. Nemeth, J. Hauer, D. Abramavicius, S. Mukamel, H. F. Kauffmann, and F. Milota, “Excitons and disorder in molecular nanotubes: A 2D electronic spectroscopy study and first comparison to a microscopic model,” J. Phys. Chem. A 114(32), 8179–8189 (2010). [CrossRef] [PubMed]
5. T. Mančal, N. Christensson, V. Lukes, F. Milota, O. Bixner, H. F. Kauffmann, and J. Hauer, “System-dependent sgnatures of electronic and vibrational coherences in electronic two-dimensional spectra,” J. Phys. Chem. Lett. 3(11), 1497–1502 (2012). [CrossRef]
6. M. L. Cowan, J. P. Ogilvie, and R. J. D. Miller, “Two-dimensional spectroscopy using diffractive optics based phased-locked photon echoes,” Chem. Phys. Lett. 386(1-3), 184–189 (2004). [CrossRef]
8. T. H. Zhang, C. N. Borca, X. Q. Li, and S. T. Cundiff, “Optical two-dimensional Fourier transform spectroscopy with active interferometric stabilization,” Opt. Express 13(19), 7432–7441 (2005). [CrossRef] [PubMed]
9. U. Selig, F. Langhojer, F. Dimler, T. Löhrig, C. Schwarz, B. Gieseking, and T. Brixner, “Inherently phase-stable coherent two-dimensional spectroscopy using only conventional optics,” Opt. Lett. 33(23), 2851–2853 (2008). [CrossRef] [PubMed]
11. S. H. Shim and M. T. Zanni, “How to turn your pump-probe instrument into a multidimensional spectrometer: 2D IR and Vis spectroscopies via pulse shaping,” Phys. Chem. Chem. Phys. 11(5), 748–761 (2009). [CrossRef] [PubMed]
12. J. A. Myers, K. L. M. Lewis, P. F. Tekavec, and J. P. Ogilvie, “Two-color two-dimensional Fourier transform electronic spectroscopy with a pulse-shaper,” Opt. Express 16(22), 17420–17428 (2008). [CrossRef] [PubMed]
13. D. B. Turner, K. W. Stone, K. Gundogdu, and K. A. Nelson, “The coherent optical laser beam recombination technique (COLBERT) spectrometer: Coherent multidimensional spectroscopy made easier,” (Invited) Rev. Sci. Instrum. 82(8), 081301 (2011). [CrossRef] [PubMed]
14. D. Karaiskaj, A. D. Bristow, L. J. Yang, X. C. Dai, R. P. Mirin, S. Mukamel, and S. T. Cundiff, “Two-quantum many-body coherences in two-dimensional Fourier-transform spectra of exciton resonances in semiconductor quantum wells,” Phys. Rev. Lett. 104(11), 117401 (2010). [CrossRef] [PubMed]
15. N. Christensson, F. Milota, A. Nemeth, I. Pugliesi, E. Riedle, J. Sperling, T. Pullerits, H. Kauffmann, and J. Hauer, “Electronic Double-quantum coherences and their impact on ultrafast spectroscopy: The example of beta-carotene,” J. Phys. Chem. Lett. 1(23), 3366–3370 (2010). [CrossRef]
16. D. Abramavicius, A. Nemeth, F. Milota, J. Sperling, S. Mukamel, and H. F. Kauffmann, “Weak exciton scattering in molecular nanotubes revealed by double-quantum two-dimensional electronic spectroscopy,” Phys. Rev. Lett. 108(6), 067401 (2012). [CrossRef] [PubMed]
17. M. Khalil, N. Demirdöven, and A. Tokmakoff, “Obtaining absorptive line shapes in two-dimensional infrared vibrational correlation spectra,” Phys. Rev. Lett. 90(4), 047401 (2003). [CrossRef] [PubMed]
19. J. Piel, E. Riedle, L. Gundlach, R. Ernstorfer, and R. Eichberger, “Sub-20 fs visible pulses with 750 nJ energy from a 100 kHz noncollinear optical parametric amplifier,” Opt. Lett. 31(9), 1289–1291 (2006). [CrossRef] [PubMed]
20. A. Nemeth, J. Sperling, J. Hauer, H. F. Kauffmann, and F. Milota, “Compact phase-stable design for single- and double-quantum two-dimensional electronic spectroscopy,” Opt. Lett. 34(21), 3301–3303 (2009). [CrossRef] [PubMed]
22. P. M. Donaldson, H. Strzalka, and P. Hamm, “High sensitivity transient infrared spectroscopy: a UV/Visible transient grating spectrometer with a heterodyne detected infrared probe,” Opt. Express 20(12), 12761–12770 (2012). [CrossRef] [PubMed]
23. J. P. Ogilvie, M. Plazanet, G. Dadusc, and R. J. D. Miller, “Dynamics of ligand escape in myoglobin: Q- band transient absorption and four-wave mixing studies,” J. Phys. Chem. B 106(40), 10460–10467 (2002). [CrossRef]
24. F. Milota, V. I. Prokhorenko, T. Mancal, H. von Berlepsch, O. Bixner, H. F. Kauffmann, and J. Hauer, “Vibronic and vibrational coherences in two-dimensional electronic spectra of supramolecular J-aggregates,” J. Phys. Chem. A 130318064008000 (2013), doi:. [CrossRef] [PubMed]
25. B. Cho, M. K. Yetzbacher, K. A. Kitney, E. R. Smith, and D. M. Jonas, “Propagation and beam geometry effects on two-dimensional Fourier transform spectra of multilevel systems,” J. Phys. Chem. A 113(47), 13287–13299 (2009). [CrossRef] [PubMed]