Hello,
I've been working on this issue for more than a month now....
I made a basic SONOS flash structure using TCAD Sentaurus sProcess.
(tunnel oxide/nitride/blocking oxide : 4/6/6 [nm])
Then, I programmed this structure and got the threshold voltage.
The problem comes in here.
When the structure was programmed, the threshold voltage only increased by 0.5V compared with the same non-programmed structure.
(Traps in the nitride are 5e19)
In reality, I believed that the threshold voltage should increase more than 1 V when programmed.
I've looked into the channel potential after programming to see how much the trapped charge decreases the channel potential and found out it's about 0.5V
what should I change to encounter the real-life effect on my simulation structure?
(I want the change in threshold voltage when programmed to be increased by more than 1V.
Thank you for reading
If anyone has any idea, please share with me
Juju
Dear Ju Lee ,
In order to increase the threshold voltage at the same charge you need to increase the oxide thickness. Then to increase Vth to 1 Volt you need to increase the oxide thickness to double its value.
You can also increase Vth by increasing the stored charge. You need to increase the programming pulse time.
Or you can study the effect of the device parameters on its performance and then you can purposely choose the proper construction and operational parameters of the device to achieve the targeted specifications.
Best wishes
Dear Ju Lee ,
In order to increase the threshold voltage at the same charge you need to increase the oxide thickness. Then to increase Vth to 1 Volt you need to increase the oxide thickness to double its value.
You can also increase Vth by increasing the stored charge. You need to increase the programming pulse time.
Or you can study the effect of the device parameters on its performance and then you can purposely choose the proper construction and operational parameters of the device to achieve the targeted specifications.
Best wishes
Dear Ju Lee,
Abdelhalim is right, here are some additional recommendations. As you said, you created the structure with sprocess. Have you calibrated the process simulations? Just a simple example, if the diffusion or segregation coefficients do not match you experiemental devices, it is obvious to have differences between simulation and expectation...
Best
Mike
- Badges
- Science topic
- Similar topics
- Computer Science
- Program