Comparative study on the influence of Al component at GaAlAs layer for GaAs/AlGaAs photocathode

    Corresponding author: Benkang Chang, bkchang@mail.njust.edu.cn
  • 1. School of Electronic and Optical Engineering, Nanjing University of Science and Technology, Nanjing 210094, China
  • 2. School of Electronics and Electrical Engineering, Nanyang Institute of Technology, Nanyang 473004, China

Key words: AlxGa1-xAs layervariable Al componentGaAs/AlGaAs photocathodequantum efficiency

Abstract: We designed two transmission-mode GaAs/AlGaAs photocathodes with different AlxGa1-xAs layers, one has an AlxGa1-xAs layer with the Al component ranging from 0.9 to 0, and the other has a fixed Al component 0.7. Using the first-principle method, we calculated the electronic structure and absorption spectrum of AlxGa1-xAs at x=0, 0.25, 0.5, 0.75 and 1, calculation results suggest that with the increase of the Al component, the band gap of AlxGa1-xAs increases. Then we activated the two samples, and obtained the spectral response curves and quantum efficiency curves; it is found that sample 1 has a better shortwave response and higher quantum efficiency at short wavelengths. Combined with the band structure diagram of the transmission-mode GaAs/AlGaAs photocathode and the fitted performance parameters, we analyze the phenomenon. It is found that the transmission-mode GaAs/AlGaAs photocathode with variable Al component and various doping structure can form a two-stage built-in electric field, which improves the probability of shortwave response photoelectrons escaping to the vacuum. In conclusion, such a structure reduces the influence of back-interface recombination, improves the shortwave response of the transmission-mode photocathode.

    HTML

1.   Introduction
  • GaAs photocathodes have been researched for several years, and it has been found that negative electron affinity (NEA) GaAs photocathodes are widely used in the fields of weak light detection, high energy physics, polarized electron source and so on; all the applications can be linked to the properties of low dark current, low energy spread for emitted electrons, great wavelength response, high spin polarization and high quantum efficiency[1-6]. Quantum efficiency is a key factor to evaluate the performance of NEA GaAs photocathodes. In recent years, improving the quantum efficiency of GaAs photocathodes is becoming the research focus, which is because of the increased performance demand for modern low-light-level night vision devices. There are some researches that report that the Al$_{{x}}$Ga$_{{1-x}}$As layer can be conducive to the improvement of the quantum efficiency of GaAs photocathodes[7-10], this is because Al$_{{x}}$Ga$_{{1-x}}$As and GaAs materials are lattice-matched, and the generated electrons in the Al$_{{x}}$Ga$_{{1-x}}$As layer can make contributions to the photoemission.

    Usually, the Al constituent of the introduced Al$_{{x}}$Ga$_{{1-x}}$As layer is 0.63, and this constituent is the optimal; transmission-mode exponential-doping GaAs photocathode with Al component 0.63 can obtain surface escape probability 0.52[11, 12].There are some reports about the influence of the GaAs emission layer and doping structure on the quantum efficiency of GaAs photocathodes[11, 13-17], but there are few reports about the influence of the Al$_{{x}}$Ga$_{{1-x}}$As GaAlAs layer on transmission-mode GaAs photocathodes, especially on the Al component. In this paper, using the first-principle based on the density functional theory (DFT) method to calculate the electronic structure and adsorption spectral of Al$_{{x}}$Ga$_{{1-x}}$As at $x=0$, 0.25, 0.5, 0.75 and 1, the differences are analyzed. We designed two transmission-mode GaAs photocathode samples with different Al$_{{x}}$Ga$_{{1-x}}$As layer, one has an Al$_{{x}}$Ga$_{{1-x}}$As layer with $x$ varying from 0.9 to 0 and the other one has an Al$_{{x}}$Ga$_{{1-x}}$As layer with $x=0.7$. Then we carry on experiments on the two samples after Cs-O activation, obtain the spectral response curves and quantum efficiency curves, combined with the band structure diagram of the transmission-mode GaAs/AlGaAs photocathode and the fitted performance parameters, analyzing the influence of the Al component of Al$_{{x}}$Ga$_{{1-x}}$As on the quantum efficiency of the transmission-mode GaAs/AlGaAs photocathode.

2.   Calculations
  • By the first-principle calculation based on density functional theory (DFT) with an ultra-soft pseudopotential method[18, 19], which is performed with the quantum mechanics program Cambridge Serial Total Energy Package (CASTEP)[20, 21], the Al$_{{x}}$Ga$_{{1-x}}$As bulk models were built and the influence of the Al component on the electronic structured and adsorption spectral of Al$_{{x}}$Ga$_{{1-x}}$As at an Al constituent of $x=0$, 0.25, 0.5, 0.75 and 1 were calculated. The calculation models are shown in Fig. 1.

  • 2.1.   Formation energy

  • After being optimized by the Broyden-Fletcher-Goldfarb-Shanno (BFGS)[22] method, the Al$_{{x}}$Ga$_{{1-x}}$As structure is stable. The formation energy can reflect the stability of the structure, and the formation energy of one unit of the Al$_{{x}}$Ga$_{{1-x}}$As model is defined as:

    In Eq. (1), $E_{\rm {tot}}$ is the total energy of the model, $n$, $m$, and $l$ are the number of As, Ga, Al atoms, respectively, while $E_{\rm {As}}$, $E_{\rm {Ga}}$, $E_{\rm {Al}}$ are the energy of the As, Ga, Al atoms, respectively. The calculated formation energy of Al$_{{x}}$Ga$_{{1-x}}$As with different Al component is shown in Table 1.

    From Table 1, it can be seen that the formation energy of the Al$_{{x}}$Ga$_{{1-x}}$As model with different Al component is negative, which indicates all the structures are stable, and with the increase of Al component, the formation energy decreases, indicating the stability of Al$_{{x}}$Ga$_{{1-x}}$As becomes stronger with Al component increasing. This phenomenal can be explained that the polarity of Ga atom is weaker than the Al atom, thus Al and As combines more tightly.

  • 2.2.   Band gap

  • The band gap is an important characteristic parameter of semiconductors, which is decided by the band structure of the semiconductor. For GaAs, it is a direct band gap material, and AlAs is an indirect band gap material. So, with the Al component x increasing, Al$_{{x}}$Ga$_{{1-x}}$As varies from direct band gap to indirect band gap. Theoretical formula of the Al$_{{x}}$Ga$_{{1-x}}$As material energy gap is:

    The calculated band gap values of the Al$_{{x}}$Ga$_{{1-x}}$As model with different Al component using the first-principle method are shown in Table 2. From Table 2, it can be seen that as the Al component increases, the value increases, too. It shows when Al component $x$ is lower than 0.5, the Al$_{{x}}$Ga$_{{1-x}}$As is a direct band gap material, while when x is equal to or higher than 0.5, it is an indirect band gap material, which is in accordance with the theory and the conclusion obtained in Ref. [23]. In general, the calculated value is lower than the literature [23], this is because the band gap is excited, and in the DFT calculation process, the band gap is in ground state, consequently, the calculated value is small, which tends to be low 30%-50% or even more, but this is a common phenomenon which does not affect the theoretical analysis of the electronic structure.

  • 2.3.   Absorption spectra

  • The absorption coefficient is another parameter indicating the percentage of the light intensity decay during the spread through unit distance. Absorption spectrum of Al$_{{x}}$Ga$_{{1-x}}$As with different Al component is defined as follows:

    where $\alpha $ is the unit direction vector of the vector potential $A$, $c$ is the propagation velocity of light in a vacuum, and $\lambda_{\rm {0\thinspace }}$is the wavelength of light in a vacuum.

    The calculated absorption spectrum of Al$_{{x}}$Ga$_{{1-x}}$As by the first-principle method is shown in Fig. 2. From Fig. 2, it can be seen that with the increasing of the Al component, the absorption band edge moves towards the high energy side. This is because the attractive force between the atomic nucleus and the outer shell electron of atom is linked with the atomic radius, and the atomic radius of Al is stronger than Ga, then the attractive force between the atomic nucleus and the outer shell electron of Al is smaller than that of Ga. Thus for the Al atom, the energy needed to excite the outer shell electron is higher.

3.   Experimental results and discussions
  • We designed two transmission-mode GaAs photocathode samples which were grown on high quality n-type GaAs substrate, and the grown method is metal organic chemical vapor deposition (MOCVD). The two samples are all p-type Zn-doped, and one has an Al$_{{x}}$Ga$_{{1-x}}$As layer with $x$ varying from 0.9 to 0 and the other one has an Al$_{{x}}$Ga$_{{1-x}}$As layer with $x$ $=$ 0.7. The structure diagram of the transmission-mode GaAs photocathode is shown in Fig. 3. It consists of a GaAs cap-layer, an Al$_{{x}}$Ga$_{{1-x}}$As layer, a GaAs emission layer, an Al$_{{x}}$Ga$_{{1-x}}$As anti-reflection layer and a GaAs substrate from the top down. The emission layer adopts a quasi-exponentially doped structure, and the layer of 1.0 $\mu$m in total thickness is divided equally into four sections, and the Zn doping concentration varies from 1 $\times$ 10$^{\mathrm{19}}$ to 1 $\times$ 10$^{\mathrm{18}}$ cm$^{\mathrm{-3}}$. For the two samples, the Zn doping concentration and thickness of Al$_{{x}}$Ga$_{{1-x}}$As layer are all designed as 1 $\times$ 10$^{\mathrm{19\thinspace }}$cm$^{\mathrm{-3}}$ and 0.5 μm, while for sample 1, the Al component ranges from 0.9 to 0, and for sample 2, the Al component is designed as 0.7.

    Before the activation, there is a need to undertake heat cleaning and chemical cleaning in an ultra-high vacuum to obtain the clean atomic surface[24]; in the process, it mainly eliminates the surface contamination, such as oxides and carbon. Following the two kinds of clean methods, "high-low temperature" two-step Cs-O activation[25]was performed on the samples in the ultra-high vacuum chamber with pressure lower than 10$^{\mathrm{-9}}$ Pa. After the activation, the spectral response curves of the two samples were measured by the independent development multi information on-line monitoring system[26, 27] after being transferred into ambient air[28]. The kind of source light used to do the spectral response is a halogen tungsten lamp of 12 V/100 W. The experimental spectral response curves of two samples are shown in Fig. 4.

    From Fig. 4, it is easily found that sample 1 has higher integral sensitivity and long wavelength response, and the short wavelength response is also higher than that of sample 2; sample 1 obtains a great blue extending effect, which is suitable for marine detection devices.

    Considering the influence of the Al$_{{x}}$Ga$_{{1-x}}$As layer, the conventional quantum efficiency formula is revised as follows[29, 9]:

    where $P$ is the surface electrons escape probability, $L_{\rm {1\thinspace }}$and $L_{\rm {2}}$ are the diffusion length of minority carrier (electron) in the Al$_{{x}}$Ga$_{{1-x}}$As layer and GaAs emission layer, respectively, $\mu $ is electron mobility, $L_{\rm {E}}$ is the lifetime of minority carrier (electron) under the function of the build-in electric field, $n(d_{\rm {1}}^{{-}})$ represents the number of electrons generated in the Al$_{{x}}$Ga$_{{1-x}}$As layer at the antireflection film/Al$_{{x}}$Ga$_{{1-x}}$As layer interface, that is Ⅱ shown in Fig. 6, $\tau_{\rm {1\thinspace }}$and$_{\rm {\thinspace }}\tau_{\rm {2}}$ are the lifetime of minority carrier (electron) in the Al$_{{x}}$Ga$_{{1-x}}$As layer and GaAs emission layer, respectively, $d_{\rm {1\thinspace }}$and $d_{\rm {2\thinspace }}$are the thickness of Al$_{{x}}$Ga$_{{1-x}}$As layer and GaAs emission layer, respectively, R is the reflectivity on the surface of the photocathode, $I_{\rm {0}}$ is the intensity of incident light, $\alpha_{\rm {1\thinspace }}$and$_{\rm {\thinspace }}\alpha_{\rm {2\thinspace }}$are the absorption coefficient of the Al$_{{x}}$Ga$_{{1-x}}$As layer and the GaAs emission layer, respectively, and $S_{\rm {v\thinspace \thinspace }}$is the electron recombination velocity at interface Ⅱ.

    Through the formula $Y(h\upsilon )=1.24\frac{S_{ \lambda } }{\lambda }$, the quantum efficiency curve can be got by transforming the spectral response curves. Here, $Y(h\upsilon )$ represents the quantum efficiency and $S_{ {\lambda }}$ is the spectral response value at the corresponding wavelength $\lambda $. The experimental and theoretical quantum efficiency curves of the samples are shown in Fig. 5. We then fitted the quantum efficiency curves by the revised quantum efficiency formula, and the fitted quantum efficiency curves are also shown in Fig. 5.

    The relative fitted performance parameters about the samples by using Eq. (4) are listed in Table 3. The fitted electron escape probability $P$ all exceed 0.5, and the electron diffusion-drift length $L_{\rm {DE\thinspace }}$of the two samples are similar, which are smaller. This is attributed to that the emission layer is divided into 4 layers not more layers, so the built-in electric field can be distributed equally, then the influence of the built-in electric field on electrons becomes weaker. It is notable that the backface electron recombination velocity $S_{\rm {v}}$ of sample 1 is smaller than that of sample 2, which is only 10$^{\mathrm{4}}$ cm/s, this is because the Al$_{{x}}$Ga$_{{1-x}}$As layer of sample 1 adopts the variable Al component design, for the Al$_{{x}}$Ga$_{{1-x}}$As layer, the closer to the back interface, the Al composition x is smaller until it is zero, so the Al$_{{x}}$Ga$_{{1-x}}$As layer material and the emission layer material lattice match well, there is no interface between them.

    The differences can be explained by the built-in electric field in the transmission-mode GaAs photocathode with an Al$_{{x}}$Ga$_{{1-x}}$As layer and the electrons generated from the Al$_{{x}}$Ga$_{{1-x}}$As layer. The band structures of the two samples are shown in Fig. 6. As shown in Fig. 6, in the schematic of the band structure of transmission-mode GaAs/AlGaAs photocathode, $E_{\rm {c}}$ is the bottom of conduction bands, $E_{\rm {v}}$ is the top of valence bands, $E_{\rm {g1\thinspace }}$and $E_{\rm {g2}}$ are the band gap of the Al$_{{x}}$Ga$_{{1-x}}$As layer and GaAs emission layer, respectively, $E_{\rm {F}}$ represents the Fermi level, $E_{\rm {vac}}$ is the vacuum level, $\delta_{\rm {s}}$ and $d_{\rm {s}}$ are the height and width of the surface band-banding region (BBR), respectively.

    Although the variable Al component and various doping structure changes the back-interface potential barrier, reduces the reflection probability of photoelectrons at the back-interface, but the structure also greatly reduces the back-interface recombination velocity, and forms two-stage built-in electric fields in the photocathode, the first-stage one is formed by the variable Al component structure in the GaAlAs layer, which leads to the decrease of the band gap, the second-stage one is formed by the various doping structure in the GaAs emission layer, which results in the formation of the band bending. Since GaAlAs materials can act as photoemission material, then the GaAlAs layer with gradient energy band structure will absorb short wavelength light and generate photoelectrons, the generated photoelectrons transport to the GaAs emission layer under the function of a first-stage built-in electric field, and then transport to the surface of the photocathode under the action of the second-stage one. In addition, the photoelectrons generated in the emission layer absorbing short wavelength light also transport to the surface under the function of the built-in electric field, consequently, more shortwave response photoelectrons can escape to the vacuum. Such a variable Al component and various doping structure reduce the influence of back-interface recombination, improving the shortwave response of the transmission-mode photocathode.

4.   Conclusions
  • In the paper, we firstly use the first-principle method to calculate the influence of the Al component on the electronic structure and absorption spectra of Al$_{{x}}$Ga$_{{1-x}}$As, the calculation results indicate with the increasing of the Al component, the band gap of Al$_{{x}}$Ga$_{{1-x}}$As is becoming larger and the absorption band edge moves towards the high energy side. Then we designed two exponential-doped transmission-mode GaAs/AlGaAs photocathodes with different Al$_{{x}}$Ga$_{{1-x}}$As layer: sample 1 has an Al$_{{x}}$Ga$_{{1-x}}$As layer with the Al component ranging from 0.9 to 0, and sample 2 has an Al$_{{x}}$Ga$_{{1-x}}$As layer with fixed Al component 0.7. After the cleaning and activation, the spectral response curves were obtained by the multi information on-line monitoring system. The experimental spectral response curves and quantum efficiency curves indicate the transmission-mode GaAs/AlGaAs photocathode with variable Al component Al$_{{x}}$Ga$_{{1-x}}$As layer can have a good shortwave response. Through the revised quantum efficiency formula, we fit the quantum efficiency curve, the fitted performance parameters show this variable Al component structure has smaller back-interface electron recombination velocity $S_{\rm {v}}$, which can be neglected. Combined with the band structure diagram of the transmission-mode GaAs/AlGaAs photocathode and fitted performance parameters, it is easily understood that the transmission-mode GaAs/AlGaAs photocathode with variable Al component and various doping structure can form a two-stage built-in electric field, which improves the probability of the shortwave response photoelectrons escaping to the vacuum. In conclusion, such a structure reduces the influence of back-interface recombination, improving the shortwave response of the transmission-mode photocathode.

Figure (6)  Table (3) Reference (29) Relative (20)

Journal of Semiconductors © 2017 All Rights Reserved