Cost-effective time-stretch terahertz recorders using 1550 nm probes Journal: opticajournal Publication type: Research Article

Christelle Hanoun 1, Eléonore Roussel 1, Christophe Szwaj 1, Clément Evain 1, Marc Le Parquier 1,2, Jean-Blaise Brubach 3, Nicolas Hubert 3, Marie Labat 3, Pascale Roy 3, Marie-Agnès Tordeux, and Serge Bielawski 1 * · 2025

Read Cost-effective time-stretch terahertz recorders using 1550 nm probes Journal: opticajournal Publication type: Research Article by Christelle Hanoun 1, Eléonore Roussel 1, Christophe Szwaj 1, Clément Evain 1, Marc Le Parquier 1,2, Jean-Blaise Brubach 3, Nicolas Hubert 3, Marie Labat 3, Pascale Roy 3, Marie-Agnès Tordeux, and Serge Bielawski 1 * (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2501-03180.txt.

Introduction

Christelle Hanoun and a large team show how to build cheaper recorders for terahertz pulses in a 2025 paper. Their method stretches a pulse in time so it can be recorded in a single shot, at up to millions of shots per second. It serves accelerator physics and tabletop spectroscopy alike. The catch has been cost: the readout needed very fast, pricey electronics, and cost grew with each added sample.

The team show that probes at a common telecom wavelength change the math. Their design uses only standard parts and a standard commercial probe laser, with no extra pulse broadening. At a synchrotron facility it recorded coherent terahertz pulses over an unmatched number of samples. The readout electronics needed only modest bandwidth. The authors present the design as a route to wider use of the technique.

Catalog record

CREATORChristelle Hanoun 1, Eléonore Roussel 1, Christophe Szwaj 1, Clément Evain 1, Marc Le Parquier 1,2, Jean-Blaise Brubach 3, Nicolas Hubert 3, Marie Labat 3, Pascale Roy 3, Marie-Agnès Tordeux, and Serge Bielawski 1 *
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE27 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2501-03180

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Cost-effective time-stretch terahertz recorders using 1550 nm probes Journal: opticajournal Publication type: Research Article Christelle Hanoun 1, Eléonore Roussel 1, Christophe Szwaj 1, Clément Evain 1, Marc Le Parquier 1,2, Jean-Blaise Brubach 3, Nicolas Hubert 3, Marie Labat 3, Pascale Roy 3, Marie-Agnès Tordeux, and Serge Bielawski 1 * Abstract Time-stretch electro-optic detection allows THz waveforms to be recorded in single-shot, up to Megahertz acquisition rates. This capability is required in accelerator physics, and also opens new applications in table-top THz time-domain spectroscopy. However, the technique has also been notoriously known for the need of high speed – and high cost – ADCs or oscilloscopes for the readout. Furthermore, the resulting cost considerably increases with the number of samples that is needed per THz waveform, an issue that has severely limited the widespread of the technique so far. In this article, we show that particularly cost-effective designs can be obtained by using 1550 nm probes. We present the performances of an experimental design, that uses only standard components (including dispersive devices), and a standard commercial probe laser without additional pulse broadening. In these conditions, our time-stretch system could already record coherent THz pulses at the SOLEIL synchrotron radiation facility, over an unprecedented number of samples, using oscilloscopes and ADC boards with only 1-3 GHz bandwidth. 1 Introduction There is an increasing demand for methods allowing ultrafast optical signals to be recorded in single shot, and at multi-Megahertz acquisition rates, using femtosecond probes. Applications include the analysis of pulsed sources in the optical [ 1 , 2 ] , X-ray [ 3 ] and THz [ 4 ] domains. Such methods are also required in a range of cutting-edge spectroscopic applications, where samples under interest present fast and non-reproducible dynamics [ 5 , 6 , 7 ] , and in order to speed up analyses in two-dimensional spectroscopy [ 8 , 9 ] . Single-shot methods are also key in research and operation of accelerators, synchrotron light sources, and free-electron lasers that operate at multi-MHz repetition rates [ 10 , 11 , 12 , 13 , 14 ] . For achieving such single-shot recordings in the THz domain, at high repetition rates (i.e., millions of traces per second) an increasingly popular technique is photonic time-stretch electro-optic detection [ 4 , 5 , 6 , 7 ] , a technique that is directly inspired by the so-called photonic time-stretch Analog-to-Digital Converter [ 15 ] . The electric signal of interest modulates a chirped laser pulse, using the electro-optic effect in a Pockels crystal. Then, the modulated laser pulse is further stretched until it can be recorded by a conventional photodetector and a fast oscilloscope or ADC board. As a result, the recorded optical pulse is a replica of the electric field under investigation, that is temporally stretched by a factor M , known as the (temporal) stretch factor. Compared to other single-shot THz recording techniques [ 16 , 17 , 18 ] time-stretch THz digitizers are – by design – able to operate at high repetition rates. However, achieving THz photonic time-stretch typically requires high bandwidth – and therefore high cost – oscilloscopes for the readout. This high cost has therefore limited applications of the technique to research in large instruments [ 4 , 19 ] , and pioneer demonstrations in time-domain spectroscopy [ 5 , 6 ] . For reducing the cost of THz recording systems to reasonable values, a key point is to reduce further the bandwidth required for the readout. Several studies have been performed for overcoming this bottleneck, in particular by performing time-stretch using long chirped Bragg gratings [ 5 ] . Another solution consists in avoiding time-stretch by using spectrally-decoded electro-optic detection based on ultrafast linear detector arrays [ 20 , 11 , 13 ] . Although these methods have been demonstrated to be efficient at MHz repetition rates and above, the limited availability of key technological components (several meter-long Bragg gratings and multi-MHz rate cameras) are still limiting their use for the moment. In this Letter, we present a design that allows Terahertz photonic time-stretch electro-optic detection to be performed with much longer stretch durations, for a given laser bandwidth, by using standard commercial components. A key point of the strategy consists of avoiding the classical wavelengths for Terahertz Electro-Optic Detection – mainly 800 nm and 1030 nm – and operating at the 1550 nm wavelength. Before going further, it is important to note that, surprisingly, few studies of single-shot 1550 nm-based THz electro-optic detection have been performed so far [ 21 ] . This contrasts with the fact that the 1550 nm wavelength has been widely used for many other applications of the so-called time-stretch ADC since its introduction in 1998 [ 15 ] . We will show here that – in addition to be a viable option – the association of 1550 nm-based THz electro-optic detection [ 22 , 23 , 21 , 24 ] with time-stretch presents major advantages with respect to previous high repetition rate electro-optic detection systems. As a first consequence, much lower costs will immediately be obtained, by enabling the use of few GHz oscilloscopes for the readout. Furthermore, we will see that even 1 GHz bandwidth readout will provide larger effective number of samples to be recorded, compared to similar systems operating at the more popular 1030 nm wavelength. Tests will be performed by recording THz pulses emitted at the AILES THz beamline of the SOLEIL synchrotron radiation facility, at a repetition rate of 0.85 MHz. 2 Experimental setup 2.1 Time-stretch THz recorder design The experimental setup is displayed in Figure 1 . Chirped probe pulses are delivered by a femtosecond Erbium fiber laser (Menlo C-Fiber, with 100 mW output power, 68 nm bandwidth, and 88.05 MHz repetition rate). The laser pulses are first chirped using 40 m of normal dispersion fiber (-140 ps/nm/km dispersion-compensation fiber). Then, the pulse repetition rate is divided by 13 using an acousto-optic pulse picker (Gooch & Housego Fiber-Q). After amplification, the resulting probe pulses have 6.77 MHz repetition rate, and typically 100 mW average power and 30 nm FWHM bandwidth. The laser repetition rate is synchronized to the main RF clock of the SOLEIL facility, using an RRE synchronization module from Menlo, thus ensuring the synchronization of the probe pulses (at 6.77 MHz) with respect to the delivered THz pulses (at 0.85 MHz repetition rate). Note that, in this test, the acquisition rate of the detection system ( 6.77\times 10^{6} measurements per second) was overdimensioned with respect to the 0.85 MHz repetition rate of the THz source. Figure 1: Time-stretch setup and typical THz signals recorded in single-shot. THz pulses are focused on a Gallium Arsenide (GaAs) Pockels crystal, and interact with 1550 nm chirped probe laser pulses. After the polarizing beam-splitter (PBS), the two laser pulses are intensity-modulated. The second stretch is achieved by a normal dispersion fiber-based telecommunication Dispersion-Compensation Module (DCM, with a dispersion equivalent to 80 km of standard single-mode fiber). Green: single-mode fiber (standard telecommunication fiber, except the DCF and DCM). EDFA: SM Erbium-doped fiber amplifier, PC: polarization controller, FC: fiber collimator with 6.18 mm focal length, OAPM: 50 mm off-axis parabolic mirror, HWP and QWP: achromatic half-wave and

Read the complete text