Light emission from a material is a complex dynamic process in which several different mechanisms can play a vital role. Based on the exact processes and materials involved, the dynamics can encompass a wide range of time-scales from an order of seconds all the way into the attosecond (10-18 s) regime. By investigating these dynamics, you can gain an in-depth understanding of a broad range of material properties and physical processes.
Spontaneous Light Emission Process in Time
When electrons excite a material (or light as in photoluminescence), valence electrons in the material reach an excited state. The excited system can subsequently decay into the ground state with the emission of a photon in a spontaneous emission process as depicted in Figure 1 (a). The decay process is stochastic in nature, and for an ideal two-level system, it is expressed by a single exponential as illustrated in the decay trace in Figure 1 (b). Measuring when light is emitted after the material’s excitation helps in retrieving this exponential form, where the lifetime τe is the point in which the exponential reaches 1/e (or ~0.368) of the initial value. The average decay rate Γtot is the reciprocal of τe.
Figure 1. (a) Schematic of the CL emission process for a semiconductor system. (b) An example of a single-exponential decay where the time of excitation and the characteristic decay time τe are indicated. The rise time here is chosen infinitesimally small leading to a vertical ramp-up of the signal at the time of excitation.
As explained by Fermi’s golden rule, the magnitude of Γrad is dependent on the intrinsic material/emitter properties as well as on the local density of optical states (LDOS) which defines the number of optical states the emitter can decay into at a specified position in space. This LDOS can be modified if emitters are positioned in a different dielectric environment. The LDOS can be improved (or, in some cases, suppressed) using nanoantenna systems, substrates or mirrors, photonic crystals, metamaterials/surfaces, or (nano) cavities . These improvements can depend on momentum (angular direction) and energy (wavelength), and therefore, the light emission process can be customized for specific applications.
In real materials, all of the excitations are not converted into light and some of the energy is lost to, for example, phonons. As such, the total decay-rate is the sum of radiative (light emission) and non-radiative (no light emission) decay Γtot = Γrad + Γnrad. Moreover, the two-level approximation is usually not valid and multiple (interacting) transitions above two or more levels can be present in a specified wavelength window resulting in more complex multi-exponential decay.
Typically, the magnitude of non-radiative decay is directly related to the inherent material properties and quality. Generally, a particular rise time will be present, which results in a sloped ramp-up of the intensity after excitation (for instance, if there is carrier diffusion before emission). This provides an additional source of information.
Time-correlated single-photon counting (TCSPC) is the most common method to measure the lifetime of light emission. In this method, a fast photodetector, for example, a photomultiplier tube (PMT) or an avalanche photodiode (APD) can be used to detect single photons with timing-precision and high sensitivity. These detectors yield a single electrical TTL or NIM pulse for every detected photon which is subsequently counted and timed by time-correlator electronics. In order to measure a decay trace, the electron source must be pulsed at a well-defined time.
In that case, it is possible to determine the arrival times of the photons in relation to a well-defined input trigger coming from the electron pulse generation process, enabling proper reconstruction of the decay trace. In the SPARC, the cathodoluminescence (CL) is relayed to a time-resolved detector unit with the help of the fiber coupling module. This detector unit includes a single-photon detector (SPD) and neutral density filters/color filters for changing the signal’s intensity and wavelength. A schematic representation of this setup is shown in Figure 2. In a more sophisticated version of the system, a streak camera can be fitted directly on the spectrograph that enables hyperspectral decay-trace imaging with high time-resolution*.
*for more information on the available detection schemes contact [email protected]
Figure 2. Schematic overview of the decay trace acquisition scheme. CL is coupled into an optical fiber using the fiber coupler module (dashed red line) which sends the light to a single ultrafast detector (SPD in this case). The signal from the detector is read out by the time-correlator system.
Pulsed Electron Microscopy
A number of solutions are available for making a pulsed electron microscope. The simplest strategy is to use an electrostatic blanker comprising of two metallic capacitor plates that are charged by an electrical pulse and then the electron beam is deflected and blanked by an aperture somewhere lower in the column. The advantage of this approach is that it can already offer very short pulses (down to 30 ps) while being user-friendly, cost-effective, simple to integrate, and flexible with respect to duty cycle and frequency, and can be readily switched from pulsed to continuous beam conditions.
Based on the position of the blanker in the column, it may be required to adapt the condenser lenses to remain in conjugate blanking mode where the beam crossover is situated in the middle of the blanker, to prevent beam streaking on the sample that results in ghosting and loss in resolution [2, 3]. It is possible to enhance the time resolution that can be achieved by moving the blanker higher up in the column where the electron energy is lower and the necessary blanking voltage thus decreases congruently, enabling faster charging of the capacitor plates.
However, this approach is usually incompatible with FEG tips because of the ultra-high vacuum demands and restricted space. On the other hand, the angular distribution of the electron beam can be limited by introducing smaller apertures in the column. These apertures may restrict the attainable current in continuous mode but considerably enhance the time resolution (pulse widths down to 30 ps are possible).
Another approach to generate a pulsed electron beam is through laser triggering of the electron gun using an ultrafast laser pulse from a femtosecond laser. This method can generate even shorter electron pulses (< 1 ps) [4,5]. Furthermore, the laser can be used to pump the sample optically while interactions occur between the electron pulse and the sample which can be used for a range of pump-probe electron microscopy experiments such as pump-probe CL. This method also has some limitations.
The technique is considerably more intricate in comparison with the electrostatic approach. Additionally, the high power femtosecond laser and optics need much more lab space and renders the system considerably more expensive. Also, it is not possible to switch instantaneously from continuous to pulsed mode as done in the case of the electrostatic blanker. In both approaches, some spatial resolution is generally lost in pulsed mode than in continuous mode.
As a third approach, there has been a considerable development in resonant microwave cavities for fast beam blanking which is being commercialized . Moreover, there is also a photoexcited blanker which integrates an ultrafast laser with the electrostatic blanking concept . These may offer alternative methods for the generation of ultrafast electron sources. Listed in Table 1 are some typical numbers for electron pulse generation with ultrafast laser excitation and an electrostatic blanker.
The instrument response function (IRF) or final time-resolution of the whole system is established by the electron pulse duration and jitter in detectors/signal readout and triggering system. The quantified decay trace will be convolved with this IRF. The IRF can be (partially) corrected by deconvolution, or by fitting the quantified decay trace data with a theoretical decay trace convolved with the IRF, provided that IRF << τe. Measuring the time-response on a metallic interface (for example, aluminum or gold) is an excellent way to experimentally determine the IRF. In such a case, the transition radiation is emitted on the order of 1–10 fs after excitation which is generally infinitesimally small in comparison with the IRF. As such, the measured width basically corresponds to the system’s IRF .
Delmic does not provide electron pulse generation solutions; however, the company closely teams up with numerous OEM partners to provide complete advanced time-resolved electron microscopy and CL detection solutions, producing unparalleled results and flexibility
Table 1. Comparison between electrostatic and laser-driven blanking. Note that typical values and behavior is listed here and the precise response depends on the exact realization of system and conditions used (e.g. SEM type, blanker/laser type).
||Pulsed fs laser
||30 ps – 10 ns
||< 1 ps
||1 kHz – 80 MHz
||100 kHz – 80 MHz
||Fast (~ 1 min)
||Slow (> 30 min)
CL decay-trace measurements are important for an extensive range of applications. As discussed earlier, the extracted lifetime is dependent on both intrinsic material properties and the local optical environment (LDOS) and can be used to obtain information on both. The material’s lifetime is often directly related to the (local) quality, where shorter lifetimes can signify more non-radiative defects that are detrimental to the device/material. Some undesirable effects such as the Quantum Confined Stark effect result in longer lifetimes and can also be seen in CL lifetime mapping.
In certain instances, the carrier dynamics in the material can also manifest itself in the decay trace both in the rise-time and in the decay-curve. Since several carrier processes are extremely fast (< 1 ns), this generally needs a good time-resolution in the system. Such studies are very valuable for rare-earth doped materials in phosphor/scintillator materials or semiconductor materials used for optoelectronic devices (for example, photovoltaics and LEDs).
Illustrated in Figure 3 is an example of lifetime measurements conducted on InGaN LED stacks at different wavelength bands. Additionally, decay-trace mapping can be used for studying/mapping the LDOS in nanoscale geometries where (single) emitters such as NV centers or quantum dots are integrated with a nanophotonic structure (for example, nanoantenna(s), photonic crystal).
To conclude, the ability to monitor lifetime processes at the nanoscale, along with the full analytical options provided by the SEM, is a powerful innovative tool for the nanoscale characterization of materials and optical behaviors.
Figure 3. Decay-trace histograms acquired on an InGaN quantum well LED stack with (a) a λ < 500 nm short pass filter selecting the InGaN quantum well emission and (b) a λ = 510 – 590 nm band pass filter selecting the yellow band emission from the n-doped GaN layer. The yellow band has a longer lifetime and clearly shows multiexponential decay characteristics. Images courtesy of Dr. Sophie Meuret (AMOLF, Amsterdam) .
 L. Novotny and B. Hecht, Principles of Nano-optics, Cambridge University Press (2006)
 R. J. Moerland, I. G. C. Weppelman, M. W. H. Garming, P. Kruit, and J. P. Hoogenboom, Opt. Express 24, 24760 (2016)
 S. Meuret, M. Solà-Garcia, T. Coenen, E. Kieft, H. Zeijlemaker, M. Latzel, S. Christiansen, S. Y. Woo, Y-H. Ra, Z. Mi and A. Polman (submitted) (2018)
 J. Sun, A. Adhikari, B. S. Shaheen, H. Yang, and O. F. Mohammed, J. Phys. Chem. Lett. 7, 985 – 994, (2016)
 A. Feist et al., Ultramicroscopy 176, 63 – 73 (2016)
 W. Verhoeven, J. F. M. van Rens, E. R. Kieft, P. H. A Mutsaers, O. J. Luiten, Ultramicroscopy 188, 85 – 89 (2018)
 I. G. C. Weppelman, R. J. Moerland, J. P. Hoogenboom, and P. Kruit, Ultramicroscopy 184, 8 – 17 (2017)
 S. Meuret, T. Coenen, H. Zeijlemaker, M. Latzel, S. Christiansen, S. Conesa-Boj, and A. Polman, Phys. Rev. B 96, 035308 (2017)
This information has been sourced, reviewed and adapted from materials provided by Delmic B.V.
For more information on this source, please visit Delmic B.V.