Deep-level Transient Spectroscopy of Schottky Diode

Case Study Deep-level Transient Spectroscopy on an aInGaZnO4 Schottky Diode

Yuxin Ji DATE: 2nd Dec, 2015

Chasin, A., Simoen, E., Bhoolokam, A., Nag, M., Genoe, J., Gielen, G. and Heremans, P. (2014). Deep-level transient spectroscopy on an amorphous InGaZnO4 Schottky diode. Appl. Phys. Lett., 104(8), p.082112.

Overview A detailed analysis of one single piece of literature concerning characteristics of Schottky diode: • • • •

What is deep-level transient spectroscopy (DLTS)? In what ways is DLTS better? How to extract subgap DOS of a Schottky diode correctly? Findings and conclusions

What is DLTS? • DLTS: – an experimental tool for studying electrically active defects in semiconductors

• How does it work? – Applying a voltage pulse •

Reduces electric field in the space charge region

•

Free carriers penetrate this region to recharge the defects

– After the pulse, when voltage returns to steady-state value •

Defects start to emit charges due to thermionic emission

•

This is when capacitance transient behaviour can be observed.

Why DLTS? • Conventional method – In C-V measurement, one adjusts the bias voltage, which simultaneously varies •

Width of the space-charge layer

•

Intersection of EF and DOS

– Results are convolution of energetic and spatial variations. – One assumes states are uniformly distributed in space.

• DLTS – Gives separate controls in energetic and spatial measurement •

Energy can be measured by observing thermionic emissions of charges

•

Spatial measurement can be done by changing the bias.

Subgap DOS • The main task is to derive concentration of deep levels (Nt) • How? – 𝑁𝑡 =

–

𝑁𝑑𝑒𝑝𝑙 ∆𝐶 1 2 𝐶 𝐾𝑇

•

where 𝑁𝑑𝑒𝑝𝑙 is the concentration in the depletion region

•

∆𝐶 is the magnitude of DLTS spectrum

𝐴2 𝐶2

2 )(𝑉𝑏𝑖 𝜀0 𝜀𝑠 𝑁𝑑𝑒𝑝𝑙

=(

−

𝐾𝑇 𝑞

− 𝑉)

•

Where 𝑉𝑏𝑖 is the built in potential

•

εS is the static dielectric constant of the semiconductor (13.9)

•

ε0 is the dielectric constant of vacuum (8.854× 10−12 F/m)

Prerequisites • Schottky diode cannot be frozen-in – Therefore a especially fabricated device is needed.

• Subgap traps should be completely filled – The pulse should be sufficiently long.

• T=300K, Vbi=0.45 V, tp=10 ms, and Up=0.5 V

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Findings

• DLTS signal for a-IGZO is featureless with no sharp peaks. • Two exponential regimes of traps. – First exponential distribution: EC to 0.2 eV below EC – Second exponential distribution: 0.3 eV below EC

• The curve is not accurate. – Spatially speaking, only traps in 𝑊𝑟 − 𝑊𝑝 region are filled. – Effective depletion width (ratio) is therefore – A correction factor of

𝑊𝑟 𝑊𝑟 −𝑊𝑝

𝑊𝑟 −𝑊𝑝 𝑊𝑟

.

should be multiplied to obtain holistic distribution.

Figure 7

Figure 8

Findings

• A lower bias yields higher spectrum magnitude. • A lower bias yields a lower subgap concentration, which indicates: – At region closer to Schottky junction •

subgap DOS is larger than depletion region DOS

•

Slope of the first exponential distribution is larger

– The second exponential distribution has a constant characteristic energy.

Figure 9

Findings

• Nt and Ndepl has curves of similar trends – Meaning they share the same physical origin.

• They are results from both effective shallow donor and subgap deep acceptor states.

Conclusion • DLTS gives separate measurement.

controls

in

energetic

and

spatial

• Some parameters must be determined beforehand. • The subgap states are exponentially distributed on two different energy levels. • Ndepl is a result from both effective shallow donor and subgap deep acceptor states.