What can transient absorption spectroscopy reveal about the trap distribution in a semiconductor?
Graphical abstract
Introduction
Shallow traps directly affect the transport of charge carriers, and hence the performance, of many semiconductor based devices, see for instance the cases of TiO2-based photocatalysts [1], [2], [3], [4], [5], [6], [7] and dye sensitized solar cells [8], [9], [10], [11], [12]. It is essential to determine the spatial and energetic distribution of traps in order to access how a certain sample growth procedure or an after-growth treatment affect the device performance.
Transient absorption spectroscopy (TAS) [13] is a widely used method to investigate the trapping dynamics of electrons and holes [8], [14], [15], [16], [17], [18], [19], [20], [21] in semiconductors. The temporal resolution for this technique is of the order of 100 fs, providing information about the electronic relaxation in an ultrafast time scale.
In this experiment an initial pump pulse creates transient excited species which are analyzed employing a white light probe beam to study the absorption at different wavelengths (UV–vis–NIR) as a function of time. If at least one of the wavelengths can be associated with the conduction band (CB) absorption, the corresponding kinetic trace reveals the relaxation of the CB population. In the supplementary material we comment on the main processes involved following the pump pulse.
It is fairly common for the kinetic traces in TAS experiments to be fitted by sums of up to three exponentials [20], [22], [23], [24], [25], suggesting the presence of discrete trap levels without attempting to pinpoint their energy or addressing the role of the disturbed Fermi level. To properly interpret the TAS kinetic traces a thorough theoretical description of the electronic relaxation process in such experiment is needed, one that clearly establishes what exactly the CB kinetic trace reveal about the trap energetic distribution. Our work addresses this fundamental question.
We model the relaxation between the CB and the shallow traps with a master equation [26], assuming that the pump pulse is of such low intensity that: (i) we can neglect any second-order processes like direct electron-hole recombination and (ii) the relaxation can be treated in the near-equilibrium approximation. As argued in Refs. [18], [27], the weak-excitation regime is the relevant one when using TAS to study photocatalysis.
We solve the master equation and obtain an integral equation relating the CB kinetic trace to the trap distribution and to the equilibrium Fermi level. The analysis of this expression reveals that: (i) at room temperature, the traps that actually contribute to the kinetic trace have energies in a range of 260 meV centered around the equilibrium Fermi level, (ii) very different trap distributions lead to nearly identical kinetic traces, (iii) trap distributions entirely concentrated below the Fermi level give rise to kinetic traces that can be perfectly fitted by single-exponentials, whereas other types of distributions produce kinetic traces almost perfectly fitted by double-exponentials (and certainly not by single-exponentials), (iv) in the first case the decay rate reveals the Fermi level, not any feature of the trap distribution, and provides an upper limit for the average trap capture cross-section; in the second case the two decay rates inform nothing about the trap distribution but the Fermi level can still be obtained exploiting an approximate scaling between kinetic traces on nearby temperatures. In both cases it is essential to measure the kinetic trace at more than a single temperature.
Section snippets
The master equation
Consider a semiconductor with many shallow trap states below the conduction band (CB). The electrons in the CB and in the traps are initially in equilibrium when some extra electrons are added to the CB. In transient absorption spectroscopy (TAS) experiments these electrons are photo-excited by the pump pulse from the valence band, but their actual physical origin is not relevant in what follows.
Our first assumption is that the ensuing relaxation process is entirely restricted to trapping and
The kinetic trace of the CB
We want to model a situation in which the system is initially in equilibrium with a Fermi level μ0 when extra electrons are added to the CB at t = 0. Our results apply to both intrinsic and extrinsic semiconductors but, as we shall see, the actual value of μ0 have a direct impact on the relaxation time scale.
If the density of extra electrons is small enough that allows us to ignore direct recombination, we may accordingly assume that the initial condition represents a modest deviation from the
Conclusion
Using a master equation linearized close to equilibrium we were able to present an expression for the relaxation dynamics via trapping of CB electrons excited by a pump pulse, as in a TAS experiment, see Eqs. (12) and (13). The expression shows that only traps in a small energy window of width 10 kT around the disturbed Fermi level (μ) contribute to the relaxation. Traps with energy below μ contribute to the CB kinetic trace with basically the same rate, whereas traps with energy above μ
Supplementary material
See supplementary material for a comment on the main processes following the pump pulse in the TAS experiment, and for the solution of the linearized version of the master equation (6).
Author statement
The authors declare that they had equal participation in the manuscript.
Conflicts of interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgements
GAE thanks CAPES for financial support. JAF thanks INEO for financial support.
References (30)
- et al.
Catal. Today
(2019) - et al.
J. Photochem. Photobiol. A
(1997) - et al.
J. Photochem. Photobiol. C
(2019) J. Electroanal. Chem.
(2007)- et al.
J. Power Sources
(2014) - et al.
C. R. Chim.
(2006) - et al.
J. Photochem. Photobiol. C
(2019) - et al.
Catal. Today
(2007) - et al.
Chem. Rev.
(2014) - et al.
J. Am. Chem. Soc.
(2014)