ISSN 0020-4412, Instruments and Experimental Techniques, 2018, Vol. 61, No. 1, pp. 94–98. © Pleiades Publishing, Ltd., 2018. Original Russian Text © G.M. Borisov, V.G. Gol’dort, A.A. Kovalyov, D.V. Ledovskikh, N.N. Rubtsova, 2018, published in Pribory i Tekhnika Eksperimenta, 2018, No. 1, pp. 87–91.

GENERAL EXPERIMENTAL TECHNIQUES

A Technique for Detecting Subpicosecond Reflection or Transmission Kinetics G. M. Borisova, b, V. G. Gol’dorta, A. A. Kovalyova, D. V. Ledovskikha, and N. N. Rubtsovaa, * a Rzhanov

Institute of Semiconductor Physics, Siberian Branch, Russian Academy of Sciences, Novosibirsk, 630090 Russia b Novosibirsk State University, Novosibirsk, 630090 Russia *e-mail: [email protected] Received March 10, 2017

Abstract—A double-modulation technique was developed and tested experimentally for the detection of the kinetics of reflection of subpicosecond pulses of probe near-IR radiation from a sample (or transmission through it) after the action of exciting radiation pulses. A probe signal is detected at a frequency that is equal to the sum of stabilized unequal frequencies of interrupting exciting and probe radiations, thus permitting the elimination of the contribution of scattered exciting radiation from the valid signal. When detecting the semiconductor-sample reflectance kinetics with a time resolution of 0.13 ps, the reflection sensitivity ΔR/R = 5 × 10–6 was reached. DOI: 10.1134/S0020441218010025

INTRODUCTION The kinetics of probe-radiation reflection and transmission makes it possible to evaluate the relaxation processes that occur in a sample after the action of a more powerful ultrashort exciting-radiation pulse. The reflection kinetics is the most important characteristic in the development of semiconductor mirrors with saturable absorption, which provide a passive laser-mode-locking regime. The kinetics is often investigated using the detection of probe pulses at the frequency of interruption of exciting radiation in the form of femtosecond pulses. However, this approach does not allow the scattered exciting radiation to be separated from a valid signal. Scattering of exciting pulses may produce a dc signal that does not depend on the delay between the exciting and probe pulses. The presence of this “pedestal” may prevent the correct interpretation of results, since a constant “pedestal” due to the absorption is also observed when studying single-crystal samples with lifetimes that exceed the femtosecond-radiation pulse repetition period. In principle, the above difficulty can be overcome by using multifrequency modulation. The two-frequency modulation technique was accomplished in [1–3] using a pair of acousto-optical or electro-optical modulators, which operated at similar but unequal frequencies. The use of such technique is complicated by “spreading” of radiation pulses due to the disper-

sion of the modulator material. In this study, the objective was to develop a high-sensitivity, simple, and inexpensive method for recording the subpicosecond probe-pulse reflection and transmission kinetics after the action of exciting-radiation pulses on a sample. SUBSTANTIATION OF THE METHOD Ultrashort pulses are measured using slow radiation detectors during recording of a nonlinear process that is formed by two continuous trains of ultrashort pulses with an adjustable delay between them [4, 5]. In this case, we deal with recording the amplitude of a subpicosecond probe-radiation pulse that passed through a studied sample (or was reflected from it) after the action of a more powerful saturating pulse, depending on the delay of the probing pulse relative to the saturating pulse. In this case, the sample-absorption saturation that causes a change in the reflection of the probe pulse from the sample surface (or in the pulse transmission through the sample) is the aforementioned nonlinear process. If the saturation is weak, the absorption coefficient α can then be expanded in terms of the powers of the saturating-radiation intensity Is as α0 ≈ α0(1 – βIs), where α0 is the unsaturated absorption coefficient and β is the dimensional factor. For probe radiation Ipr that was transmitted through the sample of low optical 94

A TECHNIQUE FOR DETECTING SUBPICOSECOND REFLECTION

0

2.5

5.0

7.5

10.0 12.5 Frequency, kHz

Fig. 1. The Fourier transform of radiation from the FL1000 continuous femtosecond laser. A decrease in noise at higher frequencies and a peak at low frequencies are observed. The vertical scale is 10 dB/division.

density (α0l  1, l is the sample length) at the frequency of the probe radiation, we can write Ipr ≈ I0exp(–αl) ≈ I0(1 – α0l + α0βlIs). The term that is proportional to the product of the intensities of the probe I0 and saturating Is radiations, which are incident on the sample, is the result of the action of the saturating radiation on the sample that is detected by the probe radiation. The modulation of the average probe-radiation power at the frequency F1 and the saturating-radiation power at the frequency F2 leads to the possibility of detecting the nonlinear term α0lβI0Is at the difference F2 – F1 or sum F2 + F1 frequency. The particular choice of the chopping frequencies of saturating and probe radiations and the choice of detection at the sum or difference frequency was determined in this study by the noise characteristics of the used subpicosecond-pulse source. In this study, we used radiation from a FL-1000 continuous femtosecond laser (Solar, Belarus) on the basis of a diodepumped Yb3+ : KY(WO4)2 crystal with a pulse repetition rate of 70 MHz, a pulse duration of 130 fs, an average power of 0.9 W, and a center radiation wavelength of 1035 nm. The Fourier transform of output radiation from the FL-1000 femtosecond laser, which was detected by an FD-7G photodiode, is shown in Fig. 1. Under steadystate laser thermal conditions, a characteristic peak in the low-frequency region is clearly observed; its presence was noted in other studies that used an analogous laser [6]. Because measures are taken to stabilize the pumping-laser power and the temperature conditions INSTRUMENTS AND EXPERIMENTAL TECHNIQUES

95

of both the active element and the mirror with the saturable absorption in the FL-1000 continuous femtosecond laser, the occurrence of the peak can be attributed to low-frequency mechanical vibrations of the elements of the laser optical resonator; a residual contribution of noise from pumping radiation is also possible. Figure 1 shows a decay of the optical-noise level with an increase in the frequency. The chopping frequencies F1 = 3.5 kHz and F2 = 4.2 kHz were provided by mechanical choppers based on DPM-25-N1-07Т motors. The valid signal was selected at a sum frequency of 7.7 kHz; this frequency is not a multiple of any of the mechanical beam-chopping frequencies and corresponds to a reduced level of optical noises of laser radiation. The maximum chopping frequency (4.2 kHz) in our case was determined by the diameter of the laser beam and the maximum rotation frequency of the motor at a reasonable chopper size (a disk with holes). The detection technique at the sum frequency makes it possible to reliably avoid the contribution of scattered radiation of exciting pulses to the valid signal. RESULTS A schematic optical diagram of the experiment is shown in Fig. 2. Beams of the probe (thin lines in Fig. 2) and exciting (bold lines) radiations were formed from radiation of the femtosecond laser (1) using a beamsplitting plate (2). Both beams passed through optical delay lines (3 and 4, respectively) and through mechanical choppers (5 and 6). The delay line for the saturating radiation was adjusted with a microscrew, while the delay line for the probe radiation (Thorlabs) operated under the control of a PC. Light-emitting diode–photodiode pairs (7 and 8) that generated signals at the frequencies F1 and F2 were placed on both sides of the chopper disks. If necessary, each beam could be attenuated using neutral filters (9 and 10). A short-focus lens (12) (the focal length is 15 mm) was used to focus both beams to the sample (11). The sample was fastened to a microscopic translation stage along three coordinates, thus allowing a sample to be adjusted in the focal region according to the maximum of a recorded signal. The coincidence of the focused beams on the sample was visualized with a webcam (13) by radiation that was scattered by the sample surface. Flat dielectric mirrors that are used in the system provide the maximum reflection at a specified angle of radiation incidence. The saturating radiation that was transmitted through the sample and reflected from it was blocked by absorbers (14 and 15), while the probe radiation was detected by photodiodes (16 or 17). The component at the sum frequency was selected from the probe signal in the electronic unit (18); this signal component was then detected with a synchronous detector (19). The zero delay between the satuVol. 61

No. 1

2018

96

BORISOV et al.

rating and probe radiation pulses was determined from the maximum of the second-harmonic generation signal in a noncollinear mode; for this purpose, a lithium niobate sample was fixed in a common plane with an investigated semiconductor sample. The electronic unit (18) performs three functions. These are the stabilization of the chopping frequencies of the probe (F1 = 3.5 kHz) and saturating (F2 = 4.2 kHz) radiation beams via control of the supply voltage; the formation of a reference signal of the synchronous detector at the frequency F1 + F2 = 7.7 kHz; and the amplification and filtering of the probe-radiation signal component at the sum frequency F1 + F2. The functional diagram of the electronic unit (18) is shown in Fig. 3. Figure 3а demonstrates the system for stabilizing the chopping frequencies F1 и F2 with a feedback with respect to the supply voltage of dc motors. Signals from the photodetectors at the frequencies F1 and F2 arrive at a frequency detector that consists of a generator of standard pulse signals, a lowpass filter, and an operational amplifier (OA). The resistors, which are denoted in Fig. 3 as F1 and F2, allow displacements of the “zero” frequency of the frequency detector. A signal from the OA output controls the supply voltage of the motors. The amplification and response speed of the stabilization loop makes it possible to maintain the frequencies F1 = 3.5 kHz and F2 = 4.2 kHz with an accuracy of 2–3 Hz. Using a switch, signals at the frequencies F1 and F2 can be fed to a frequency meter and a digital voltmeter to monitor the voltages of the motor power-supply unit and signal voltages at the frequencies F1 and F2. A block diagram of the formation of a reference signal at the sum frequency for synchronous detection is shown in Fig. 3b. Signals at the frequencies F1 and F2 arrive at a balanced mixer; subsequently, using bandstop and high-quality-factor band-pass filters, a signal of the sum F1 + F2 frequency is selected. The measuring signal arrives at the adjustable amplifier and then, using a high-quality-factor band-pass filter at the frequency F1 = 3.5 kHz and a high-quality-factor fourthorder band-pass filter at the sum frequency 7.7 kHz, is fed to the measuring input of the synchronous detector. Figure 3c shows a device in the form of a separate unit for selecting the nonlinear part of the probe-radiation signal within a band of 7.7 ± 0.01 kHz, which includes an adjustable amplifier, the band-stop filter for a frequency of 3.5 kHz, and the narrow-band filter at a frequency of 7.7 kHz. To obtain an idea of the sensitivity of the considered method, the reflection kinetics was recorded for attenuated intensities of the saturating and probing radiations. A semiconductor mirror with saturable absorption on the basis of quantum wells of InGaAs with nanostructured GaAs barriers, which is analogous to that in [7], was taken as the sample. The sam-

3

7 F2 4

2 1 F1

6

8 5

17

11

9

12

16

15

10 F1 19

18 13

14

F2

Fig. 2. A schematic optical diagram of the experiment: (1) laser; (2) beamsplitter; (3, 4) optical delay lines; (5, 6) choppers; (7, 8) LED–photodiode pairs; (9, 10) neutral light filters; (11) sample; (12) short-focus lens; (14, 15) blocking of exciting radiation; (16, 17) photodiodes for detecting probe radiation; (18) the unit for stabilizing the rotational frequency, forming the reference signal, and extracting a valid signal at the sum frequency; and (19) synchronous detector.

ple reflection kinetics was recorded using the abovedescribed method under the following experimental conditions: the pumping-radiation pulse energy was ~30 × 10–12 J at a peak intensity of 17 MW/cm2; the energy of the probe-radiation pulse was 0.8 × 10–12 J, and the peak intensity was 0.43 MW/cm2. The delay of the probe radiation pulse changed from –1 to 20 ps. The curve in Fig. 4 shows the relative change in the reflection coefficient of the investigated mirror ΔR/R as a function of the probe-pulse delay. The noise level in Fig. 4 does not exceed 5 × 10–6, which determines the limiting sensitivity of the method in the given realization: ΔR/R = 5 × 10–6. The short segment of the signal rise in Fig. 4 (on the order of the exciting-pulse duration) corresponds to the sample-reflection increase upon saturation. The descending segment of the curve is approximately modeled by the weighted sum of two exponents in the form S(t) ≈ s1exp(–t/T1) + s2exp(–t/T2). The relaxation times that were determined for this sample are T1 ≈ 0.4 ps and T2 ≈ 20 ps: the relaxation time T corresponds to the ionization of excitons that were formed under the action of saturating radiation, while the longer relaxation time T2 corresponds to the electron– hole recombination in the sample. The obtained value of T2 exceeds the mirror recovery time that was used in

INSTRUMENTS AND EXPERIMENTAL TECHNIQUES

Vol. 61

No. 1

2018

A TECHNIQUE FOR DETECTING SUBPICOSECOND REFLECTION

(а)

Frequency meter

97

Voltmeter

F1, F2  Setup and stabilization system +F1

Standardpulse generator

Operational amplifier

Lowpass filter

Motor power supply

Motor 1

Motor power supply

Motor 2

+F2

Standardpulse generator

(b)

Operational amplifier

Lowpass filter

Circuit of the formation of a reference signal of the synchronous detector F1 Balanced mixer

Band-stop filter, 700 Hz

Band-pass filter, 7700 Hz

Reference signal

F2

(c)

Signal selection circuit Input

Adjustable amplifier

Bandstop filter, 3500 Hz

Band-pass filter, 7700 Hz

High-pass filter

Output (synchronous detector input)

Fig. 3. (а) The functional diagram of the apparatus and stabilization of the frequencies of the mechanical choppers, (b) the formation of the reference signal at the sum frequency for the synchronous detector, and (c) the selection of a signal at the sum frequency.

[7]. This can be explained by a spatial inhomogeneity of the sample, which is known to users of semiconductor mirrors with saturable absorption. DISCUSSION The presented technique was developed for recording the subnanosecond kinetics of the reflection or transmission of semiconductor structures. However, the field of its application is much wider. This is primarily the detection of saturated absorption in the presence of a high-level scattered signal from exciting radiation: a pulsed or ultrafast process is not obligatory. Because the method is highly sensitive, its application in the cases where the studied object does not permit the action of high-intensity saturating radiation is obviously promising. INSTRUMENTS AND EXPERIMENTAL TECHNIQUES

In our case, the ultimate sensitivity of the recording system is limited by several factors. First, the main factor is the radiation-noise level of the continuous femtosecond laser. Measures that are taken to provide more careful temperature stabilization and stabilization of the pumping-diode laser current, as well as the use of stiffer fixtures in the optical resonator of the femtosecond laser, may obviously increase the limiting sensitivity. The quality of the recorded kinetic curve also depends on the stability of the movement of the linear-translation stage (delay line). Because the distances between optical elements are rather long and the transverse radii of the beams at the lens focus on the sample surface are 30 μm, the optical system is sensitive to microvibrations of its optical elements, thus also reducing the quality of the registered signal. The above drawbacks can be somewhat reduced via multiple scanning of the linear-translation carriage Vol. 61

No. 1

2018

98

BORISOV et al.

in the reflection coefficient. This system makes it possible to carry out investigations in a wide range of saturating-radiation intensities. This technique is of interest for a wide range of experiments on the study of the saturated-absorption kinetics in sample reflection or transmission, including medico-biological objects of investigation.

ΔR/R, 105 15

10

REFERENCES 5

0

0

5

15

10

20 Delay, ps

Fig. 4. The reflection kinetics of the semiconductor mirror with the saturable absorption. The dots are the experiment and the solid line is the approximation of the segment of the reconstruction of the linear reflection by the twoexponential decaying curve.

within a specified range of delays accompanied by storing and averaging of the results. CONCLUSIONS The system for recording the probe-radiation reflection or transmission after the action of a subpicosecond saturating-radiation pulse on a sample proved to be highly sensitive when recording changes

1. Savikhin, S., Rev. Sci. Instrum., 1995, vol. 66, no. 9, p. 4470. http://dx.doi.org/ doi 10.1063/1.1145344 2. Frolov, S.V. and Vardeny, Z.V., Rev. Sci. Instrum., 1998, vol. 69, no. 3, p. 1257. http://dx.doi.org/ doi 10.1063/ 1.1148792 3. Nechay, B.A., Siegner, U., Achermann, M., Bielefeldt, H., and Keller, U., Rev. Sci. Instrum., 1999, vol. 70, no. 6, p. 2758. https://doi.org/10.1063/1.1149841 4. Sverkhkorotkie svetovye impul’sy (Ultrashort Light Pulses), S. Shapiro, Ed., Moscow: Mir, 1981. 5. Kryukov, P.G., Femtosekundnye impul’sy. Vvedenie v novuyu oblast’ lazernoi tekhniki (Femtosecond Pulses. Introduction into a New Field of Laser Technique), Moscow: Fizmatlit, 2008. 6. Zhang, Zh., Sun, J., Gardiner, T., and Reid, D.T., Opt. Express, 2011, vol. 19, no. 18, p. 17127. doi 10.1364/ OE.19.017127 7. Kisel, V.E., Rudenkov, A.S., Pavlyuk, A.A., Kovalyov, A.A., Preobrazhenskii, V.V., Putyato, M.A., Rubtsova, N.N., Semyagin, B.R., and Kuleshov, N.V., Opt. Lett., 2015, vol. 40, no. 12, p. 2707. https://doi.org /10.1364/OL.40.002707

INSTRUMENTS AND EXPERIMENTAL TECHNIQUES

Translated by A. Seferov

Vol. 61

No. 1

2018