• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Chipworks - TEM scans of 22nm Trigate chips

Neat link. On a side note, how do they prepare these samples? I was thinking that they used a FIB (focused ion beam machine) to mill down a vertical section, then open up a large horizontal chunk to look sideways through, and then take a SEM photo from the side with the beam at as close to a perpendicular angle as possible, but then I read this:
We have to digress here a little to explain what we’re looking at. A typical TEM sample is 80 – 100 nm thick, to be thin enough to be transparent to the electron beam and at the same time have enough physical rigidity so that it does not bend or fall apart.
Click to expand...
So, my assumption was wrong (no surprise... happens frequently). They aren't using a SEM but a TEM, which means that they are looking through a slice of chip (SEM uses electron backscatter, TEM uses electron transmission)... but how does one slice off a ~90nm thick slice of silicon?

It's a neat article but I am still struggling with the idea of using something like a wafer saw to slice a 90nm slice off of something. And if not with a wafer saw, then how?
 
Last edited:
Yeah, I too thought this was SEM rather than TEM. I've never used a TEM so I have no idea how they prepared these samples.
 
Yeah, Ben90, I agree. That one seemed especially cool. Although I like the one with all of the metal stacked up as well. But seeing the actual silicon lattice. Very cool. And, Smartazz, I'm glad to hear I'm not the only one who thought SEM first. 🙂
 
pm said:
Neat link. On a side note, how do they prepare these samples? I was thinking that they used a FIB (focused ion beam machine) to mill down a vertical section, then open up a large horizontal chunk to look sideways through, and then take a SEM photo from the side with the beam at as close to a perpendicular angle as possible, but then I read this:

So, my assumption was wrong (no surprise... happens frequently). They aren't using a SEM but a TEM, which means that they are looking through a slice of chip (SEM uses electron backscatter, TEM uses electron transmission)... but how does one slice off a ~90nm thick slice of silicon?

It's a neat article but I am still struggling with the idea of using something like a wafer saw to slice a 90nm slice off of something. And if not with a wafer saw, then how?
Click to expand...

lol look at TEM sample prep videos... I've seen a few on Youtube before. I've had classmates who've done it before. You use a FIB and you mill out a cross section. the whole thing can take a few hours.... yuck
 
i don't understand any of the EE stuff, so i'll comment from an MSE perspective...

The images didn't look like SEM to me, clearly TEM =) TEM is pretty cool, but samples are a bitch to prepare :\ (college, ugh lol)


What i'm curious is how, by looking at the photo, they know the sample is oriented in <110>? I'm drawing it out and can't quite say if I can literally eye ball the orientation and say "yeah, its like that". Maybe its a function of their TEM to help determine the lattice spacings?

Either way that is an awesome TEM, you can literally see the lattice 🙂
 
magomago said:
i don't understand any of the EE stuff, so i'll comment from an MSE perspective...

The images didn't look like SEM to me, clearly TEM =) TEM is pretty cool, but samples are a bitch to prepare :\ (college, ugh lol)


What i'm curious is how, by looking at the photo, they know the sample is oriented in <110>? I'm drawing it out and can't quite say if I can literally eye ball the orientation and say "yeah, its like that". Maybe its a function of their TEM to help determine the lattice spacings?

Either way that is an awesome TEM, you can literally see the lattice 🙂
Click to expand...

Awesome. Another MSE guy here. Yeah, don't want to be elitist but yeah most of us can look at an image and go "Yup that's TEM." I do work with a pretty awesome FE-SEM that has a resolution of 1.5nm technically, and I look at plenty of samples at like 250 kx no problem, but anything like this is definitely TEM, especially when you can see the lattice clearly.

I definitely do need to learn more about TEM. Perhaps I can pick some up as its right next door to the SEM I'm on all the time. But the problem is I don't think they have a FIB here, so it's just traditional TEM grid crap I think.
 
DLeRium said:
I definitely do need to learn more about TEM. Perhaps I can pick some up as its right next door to the SEM I'm on all the time. But the problem is I don't think they have a FIB here, so it's just traditional TEM grid crap I think.
Click to expand...

Haha, MSE here too. We're all coming out of the woodwork for this post 😛

Thank goodness I didn't have to use TEM...sample prep is a pain as magomago mentioned. AES, EPMA, and a bit of SEM is what I used for my thesis work.
 
Last edited:
thilanliyan said:
Haha, MSE here too. We're all coming out of the woodwork for this post 😛

Thank goodness I didn't have to use TEM...sample prep is a pain as magomago mentioned. AES, EPMA, and a bit of SEM is what I used for my thesis work.
Click to expand...

LOL, MSE here too 🙂

The sample prep for high-res xsem and tem is usually the same, fib the sample to retain the vertical structures you wish to then image.

TEM requires both front and back sides of the sample to be prepared and of course the thickness in the sample prep must be quite thin and carefully controlled (uniforum across the sample area), xsem only requires the frontside to be milled down by FIB and only requires you to be sure your sample surface is actually normal (perpendicular) to the features you want to cross section otherwise everything can appear to be larger than it really is.

The reason they absolutely must use TEM for this imaging is that most high resolution sems (the kind that can approach the resolution needed to resolve the features in these xtors) are entirely limited by surface charging of the sample. Techniques for minimizing surface charging help, but still at a detriment to final resolution of the SEM.

Second reason for going TEM instead of SEM is contrast. TEM gives you excellent z-constrast (heavier atoms, higher z-number, tend to be darker, as are more dense materials residing next to less dense materials, etc).

So you go TEM for high-res apps like xtor cross-sections. (there is a reason these tear-down reports usually go for $35-$40k)
 
pm said:
Yeah, Ben90, I agree. That one seemed especially cool. Although I like the one with all of the metal stacked up as well. But seeing the actual silicon lattice. Very cool. And, Smartazz, I'm glad to hear I'm not the only one who thought SEM first. 🙂
Click to expand...

Read the title...

(I thought that it was SEM at first too, hence sought to preempt this discussion... *sigh*)
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top