I have fabricated a Schottky junction (100nm Au on nGaAs substrate with Ohmic contact of AuGe(100nm)/Ni(30nm)/Au(50nm) on nGaAs) using ebeam evaporation...When I perform I-V measurements, there has been times where a current range of 500mA (or sometimes even lower) passes through the device and after that there is black hole (dot) on the surface of Schottky layer (where probe was touching). The device I-V characteristics becomes linear (previously it was a nice diode-like curve) and to me it seems like device is failing. What can be done to increase the tolerance of the device to withstand higher currents?
Hello Amir.
According to your description, after applying a rather large current, the Schottky junction at the Au/semiconductor interface is destroyed and the contact becomes an ohmic one.
Did you estimate the power dissipated in the junction (U times I) when the biasing current is reaching 500 mA? I don't know the lateral dimension of tour device, but the associated temperature increase is certainly rather high.
GaAs can easily react with gold even in the solid-phase: Ga atoms will diffuse into gold, destroying the stoechiometry of the GaAs material just below the interface. In the present case, the combination of heating and strong current degrades the interface and transforms the Schottky contact into an ohmic one. Sinc your gold layer is rather thin, the damage is localized in the tip area.
This is the reason why, in order to improve the robustness of Schottky diodes on GaAs, a first very thin (typ.50nm) titanium is first deposited (don't react with GaAs) followed by a very thin platinum layer (to avoid interdiffusion of Ti and Au), and then gold.
See e.g. Luu-Henderson et al., http://csmantech.org/OldSite/Digests/2006/2006%20Digests/8B.pdf
At what corresponding voltage the current has a value of approx. 500mA? The touching black dot might be an indication for some electro migration problem? On the other hand if you explain from a diode like behaviour until a certain point and then it becomes resistive seems that there are other effects. Do you have some IV curve?
However, is the contact at the interface to the semiconductor well defined? Just some questions which came into my mind with the limited information...
Hi Mike,
Thanks for your reply. After creation of the "black hole", the I-V become linear for the whole range of voltage.
I do not understand what you mean by well defining the contact at the interface to the semiconductor. Do you mean the probe contact to the Ohmic/Schottky contact? What do you mean by word "well"?
Also, what is your thought on improving the tolerance of the device for withstanding higher current? Increasing the thickness of Schottky/Ohmic contact?
Thanks,
Amir
Hi Amir,
thank you for your reply. So let me think about it. A few additional questions:
In the IV curve it continues without any change in slope, am I right?
What I mean with "well" the junction itself. If you check the crosssection by REM or TEM or whatever, that you have no voids, etc. at the interface between metal and semiconductor.
If it is some electromigration topic or a similar problem, this might help. Black dot seems to me, that there is melting something, caused by whatever. This in fact is a hint for high current densities... so this might be the way. But what I am woundering is, that your thicknesses compared to literature and papers do not deviate that much and are state-of-the-art, aren´t they?
Best
Mike
Hello Amir.
According to your description, after applying a rather large current, the Schottky junction at the Au/semiconductor interface is destroyed and the contact becomes an ohmic one.
Did you estimate the power dissipated in the junction (U times I) when the biasing current is reaching 500 mA? I don't know the lateral dimension of tour device, but the associated temperature increase is certainly rather high.
GaAs can easily react with gold even in the solid-phase: Ga atoms will diffuse into gold, destroying the stoechiometry of the GaAs material just below the interface. In the present case, the combination of heating and strong current degrades the interface and transforms the Schottky contact into an ohmic one. Sinc your gold layer is rather thin, the damage is localized in the tip area.
This is the reason why, in order to improve the robustness of Schottky diodes on GaAs, a first very thin (typ.50nm) titanium is first deposited (don't react with GaAs) followed by a very thin platinum layer (to avoid interdiffusion of Ti and Au), and then gold.
See e.g. Luu-Henderson et al., http://csmantech.org/OldSite/Digests/2006/2006%20Digests/8B.pdf
Hi Vincent,
that was my expectation in terms of a well defined interface between metal-semiconductor... even if I am not a process engineer...
Best
Mike
@Mike Schwarz
Hi Mark,
Yes, the slope remains same during all current values.
And the thickness of the Schottky barrier is pretty much close to what is in literature.
@Tillmann Krauss
Hi Tillmann,
I use a micropositioner to place the needle on the Schottky layer (see the attached picture). I usually bring down the needle until it starts to move horizontally (since I am watching the sample vertically using a microscope). There are scratches to the surface of the sample time to time, but the black holes I observe are created even without scratching the
surface.
Dear friends,
Attached, please find the figures of the devices with the black hole on the Schottky layer. Let me know what you think!
Also, I have another question: why the black hole is on Schottky layer, not on Ohmic layer....Does this relate to the structure of Ohmic layer (AuGe/Ni/Au) vs. Schottky layer (Au)? Does this mean that the structure in Ohmic layer can withstand more current?
Hi,
I guess you bias both in forward direction. For the ohmic contact the electrons from the semiconductor have no problem to move to the metal. In terms of the Schottky contact you have a barrier to surmount. If you bias as in the attached graph (you find under the following link: https://www.tf.uni-kiel.de/matwis/amat/mw2_ge/kap_6/backbone/r6_3_2.html) and your MS interface is not stable enough, the field may become too high for your MS contact and result in a junction breakdown. What do you think Till?
Best
Mike
Amir,
the power is dissipated in the Schottky junction, where the voltage drop is taking place.
Thank you Mark and Vincent! This was a fruitful conversation for me.
- Badges
- Science topic