Saturday, May 15, 2010

Biological, inverted microscope layer image comparison

Biological, top metal (220X):
Biological interconnect doesn't image well as all black. Sample for comparison (not same area):
If you are stuck with a biological microscope, set the intensity as high as you can and you should be able to make things out with your eye, but cameras might have more difficulty.
Inverted metallurgical (inverted), top metal (440X):
Inverted metallurgical (inverted), interconnect (440X):
Although all three are from the same chip type, only the two inverted images are from the exact same chip. NOTE: I have done a vertical flip on the inverted images to correct them to what the actual object is like (as in the biological image).
For those wondering where the magnification levels come from, it is a combination of the eyepiece, objective, and camera magnification. For the biological microscope, the image was shot at 10X eyepiece * 10X objective * 2.2X camera = 220X. For the inverted metallurgical microscope, the images were shot at 10X eyepiece * 20X objective * 2.2X camera = 440X. Thus, I (poorly) manually stitched two images together from each to get closer to the size of the biological image.
Alas, despite my best efforts, I still can't see transistors. Today I tried some extended hot baths in HF acid on the chip above and even tried adding some 30% H202 which dramatically increases the action of the HF. I'm letting it sit overnight (or maybe till Wed after finals) and see if it ate through. Maybe these chips are just resistant? I'll have to go back to sanding though to get other sample since I haven't gotten another technique to work yet on epoxied ICs.

Thursday, May 13, 2010

Image stitching

One of the challenges of using computer analysis is that it is highly desirable to assemble a full image of the chip, but only a small section of the chip can be imaged at a time.
For the DRAM example chip that's been used throughout several of these posts (its because they are easy to decap and I have a lot of them), I started by manually pushing the chip around since I didn't have linear stages and taking pictures, trying to get 1/3 or so overlap on each side. However, these initial sets resulted in far inferior results compared to those made by using linear stages. That said, I need to make a comparison where I manually paste the images together. In any case, I first tried to use Hugin to manually stitch the images together. Unfortunately, while Hugin works very well for standard images, it seems to have issues with the vectorized patterns typical of ICs. I was only able to paste two well formed pictures together after setting 15 or so control points:
Thinking that I could do better, I looked into how to automate this process better. I cam across some information on the SIFT algorithm. Unfortunatly, while the SIFT algoirthm is patented by University of British Columbia, there are still some free implementations of it. The first I evaulated with autopano and later, after a reccomendation, was autostitch. Using the screwdriver pushed around die (as opposed to linear stage), with autopano, I go this:
However, I can't find a result using the linear stage version. I need ot rerun on that to confirm that autopano can't do better.
Next I tried autostich. After some help from my friend on the linear stage set, I go this:
My friend that reccomended it to me is a bit better at using it though, and was able to get a much nicer image seen here (warning: 16MB). Scaled down local version:
His version is slightly more linear. My original image was quite distorted before he gave me the setting advice. In order to get good images with autostich, adjust quality settings to 100% (otherwise it will shrink it), set to linear, and adjust theta value low.

Inverted metallurgical microscope

A few weeks ago, Tom Ditto was kind enough to lend me this:
A sample image:
It has a number of differences over a standard biological microscope.
First, its primary advantage is that it does not require light to shine through the sample. Instead, light is injected directly into the view path, allowing easy view of opaque objects. I get around this with the biological microscope by shining intense lights to the side of the sample. I'm currently using a small 650W cone shaped halogen on a rheostat. Before I was using a 500W with a roll cage. I don't push it anywhere near full power. I had to make a shield to only allow in the needed amount of light and not toast the microscope with the intense heat it puts off.
Second, it inverts the image. As an example, here is an image of of an Intel 80486 ("486") copyright on the die:
Note it is flipped. After cropping and inverting, a much nicer image is produced:
Those black spots are dirt on the lenses. I've cleaned a lot off, but some of the larger spots are still present. In order to use this analytically, I will obviously have to fully clean it so spots of dirt don't appear where some critical interconnect image should be.
Various filters are available. I haven't played much with these though. I think the options include regular, polarized, and green. Here is a regular image, showing interconnect layers on some DRAM:
With the green filter:
Finally, although many do, for some reason the biological microscope I have doesn't have a movable stage on it. Maybe you were expected to push a slide around with your fingers? On the other hand, this one has an XY knobs for smooth, precise movement of the stage.
For future work, first on the TODO list is a proper camera mount. I'm using a clip on type mount that Tom let me borrow, but, unfortunately, its not stable enough for good images. This should fix alignment issues resulting in fuzzy, darker colors at the bottom and will allow me to zoom in, removing the dark circle. Next, I'm working on fitting some timing pulley's to the XY knobs for CNC control. I have all of the parts assembled for that and could probably due the fit in several hours. However, I need a larger peice of particle board to mount it on. Unfortunatly, I probably won't get back to this until August when I'm done working for the summer. With luck, I'll find someplace to continue work with it at MIT over the summer. Extra lab space anyone?

Why I'm not to the transistor layer yet, initial NaOH etch results

Several factors behind this. First, I'm approaching finals and haven't had a lot of time to delve into the deeper layers. Second, I haven't had much success in the few tries I've done to get below the metal layer. Although I should try this more, the initial sanding I tried with a dremel was too course. It had a tendency to produce a relativly uneven surface and take too much off. The HF would take off the top metal layer, but nothing else. I talked to Travis Goodspeed last night and he pointed out I should be (near) boiling the HF. Yum. Tried today briefly and I think I have some progress. However, I really need to take some pictures as I go to get a better idea of exactly what is being corroded.
I had also tried with some NaOH on a P3 die since I heard that is also corrosive to glass. I ordered some KOH which I read somewhere is preferred more in the semi industry, but it got sent to my home in CA by accident, so maybe I'll play with that in a few weeks. First, this is your typical P3 slot 1 die after being tin snip cut from the BGA carrier:
This is actually a metalic blue, but apperas more black in the image. There is a gray epoxy holding it in place and a brown carrier around that. After inspecting the blue area under the microscope, I realized its just a coating to make heat dissipation better and not actually part of the silicon. In fact, you can see crystals poking through it (100X biological microscope):
Unfortunatly, it was somewhat difficult to get a good picture of them, but you get the idea. Thinking this, being a large chunk of silicon, was a good canidate for NaOH etching, I boiled it in some NaOH for a bit. Result (note this is not the exact same chip as above):
Hmm...what are those cracks? I think its stress from using the tin snips. If I was doing this for an imaging sample, I should probably dremel or carefully hacksaw cut them out. In any case, you can see the silicon crystalline surface much better (100X biological microscope):
in this picture, one of the cracks is also visible. I'm actually not sure how much the NaOH etched the surface since I didn't realize the blue wasn't silicon and wasn't really paying attention to see otherwise. I'll try the other side of thise or some other dies I have laying around in the near future.
Also on the not going so well side, this is how my first attempt at a rosin decap ended:
After ordering some borosilicate test tubes, I tried with a heat gun instead since I could control it with a variac and gave more regular results. After a few hours of boiling, a small nick came out of the underside of the chip along the die area, but not corroded as I was hoping. However, a large portion of this time was also devoted to boiling off the petrollium jelly after which the solution darkened and roes in temperature. The CCC IC RE Wiki page (written by Martin Schobert I think) shows the rosin doing more of a higher temp boil. I was at a lower boil. So, I might try to get a propane torch or a more proper alcohol lamp and try that. I'll have a few weeks soon to mess around at home and can probably figure something out. Also, this killed the TCO in the heat gun after I was done.

Tuesday, April 20, 2010

Paypal payload

There have been several advancements including work on finishing the automation of the biological microscope and automation of the inverted metelergical microscope, but school has kept me busy and I haven't had time to finish those. In the meantime, here is a small deviation on taking apart a security key to get to the IC. In such situations, there is a double decap. We not only have to decap the IC pacakge, but its a bit of an effort to even get to it.
At DEF CON 17, there was a side event of sorts called BSides/Neighborcon (thanks Travis!). This actually had my favorite talk of the entire trip by HD Moore on WarVOX. In any case, PayPal handed out stacks of the PayPal (or someone anyway, I think it was them) Security Key.

From some articles such as this it is based off of RSA’s Securid. I'm not a crypto guy, but I figure if I take some images of this and work out some of the logic, someone else more experienced in the field who can't do this type of hardware analysis might be able to build off of this work. I won't be be imaging the chip until I can get some more experience since future cards will cost me $5 a pop. Plus, I haven't made any agreements at this point not to tear it to shreds. I originally had a lot, but I gave them away to a number of people who thought they were cool.
One cool feature of these things its display Basically, it will retain the image on it even with power gone. It is the same (class?) of technology used in the more famous Amazon Kindle.
From what I hear, GM week at RPI use to be about getting wasted and they use to bring large amounts of beer for students to drink. But they don't do that anymore. I don't drink, but it would have been hillarious to watch. On the surface, its about elections...w/e. I still have my mug from last year which is better and I needed some glassware to dissolve the card in. To top it off, it had a Vegas theme, which seemed appropriete to make the card go full circle.
In any case, lets get started with the teardown. After a few minutes in acetone, the outter cover is starting to shed:
A side view showing the ridges a bit better:
I think peeled this off to speed things up and soaked it a bit more:
The other half is starting to break apart a bit:
A little dissolving later, I can peel off the outside plastic to reveal the circuit board:
Closeup of the label section:
There are very small surface mount components on the board. The label says "InCardIC006AV11". There's also a number 2, whatever that is for. My guess is that five dot gold pattern is for programming and/or testing. That black dot should be the IC, which is what I'm primarily after. Unfortunately, it has no external labeling of any kind. Finally, the last component is what appears to be a lithium polymer battery based on its shape. Voltage reading:
Amazingly, the card still works! (the battery was removed later, still had battery here)


The acetone was getting a bit dirty. Time to clean it up a little:
After soaking for the last time, I wasn't able to get much else to come off even after soaking for a while. I had been hoping the board was going to dissolve at least slightly and release the IC package. Final front board image:
The battery came off with minimal force. Final back image:
The black IC package was then forcably removed and stored into a vial for later analysis. As I get better suited to dissolve the resin, I'll dissolve it and take at least a top metal layer picture. In the meantime, I'll keep practicing on expendable chips so scarcer chips like this can be properly analyzed.
To top things off, what kind of person would I be if I let flammables go to waste?

Thursday, April 15, 2010

IC decapping round 4: burnt perfection

As you can hopefully see from some of the previous posts, people employ many techniques to remove the irritating resin casing. I've seen suggestions of simply burning away the casing, which I was somewhat skeptical of, thinking this would annihilate the chip. However, I am open to ideas, and thought I'd give it a spin.
First off, burning is probably a bad word. The image I had in my head of how this would work was the physically heat the chip until the resin was utterly destroyed, like burning away a piece of wood. It turns out its not hard to make the resin brittle through heating. Or in other words, all one has to do is apply a short burst of intense heat and you can crumble the casing away with minimal chip damage. A picture is worth a thousand words, lets see an example. This was the result of the first chip I tried:
It came out amazingly clean like this in under a minute. Microscope inspection seems to indicate the die is healthy. More on this later. Starting from the beginning, here is a virgin chip:
A CP82C59A interrupt controller. All torch images are of one of these, but not necesarily the exact same unit. A tube of them was being kicked around in the RPI Electronics Club junk drawer. Torch meat now:
You shouldn't heat the chip too strongly. If you get this, you've gone too far:
I actually would have never guessed you could get a red hot IC package. In any case, as will be seen in the video, the most important thing is even heating. I'm not sure what the shock temperature is, but there was a certain cutoff line where the chip was extremely brittle vs very hard. In the video I don't heat the lower half of the chip good enough and it only breaks moderately well.
After removing the torch, it will burn like a well down marshmallow for a few seconds:
It should go out by itself fairly quick, but I was waving or blowing it out. Result:
Breaking now:
This final image is the die from the video. The one from the sequence above is the one on top shown here (contrasted with the very first one I flame decapped):
The heat managed to separate the die and the carrier! So that's what happens when you apply too much heat. I wasn't paying attention at the time and I'm not sure what happened to the die. I figured it had been melted in some weird way where as it probably fell to the floor when I cracked the casing open. The first one I tried (bottom die above) couldn't have gone any better. The case split perfectly and no resin was left covering the die.
Seeing how well the die withstood the heat, I wondered how far can we go? So I torched a die red hot. Although it was not as clear under the microscope as before, it still was of decent quality. I'll try to add a pic in a bit comparing an area of the first die extracted with the one that I tortured.
Here is a short video showing an heat based extraction:

In this test setup, that blue thing is a filter system to help reduce fumes. I also have a gas mask handy for exactly this sort of work. It really does help a lot. The resin was really brittle and very little force was required to break it. However, I didn't heat it enough towards the bottom, so one side of the die was still relatively well set into the case. This can be easily fixed though and is partly due to me trying to get this on film and not doing for an ideal setup.
For some future work, I may have gotten away with this due to the relatively large traces the 2 um test die I used had. Thinner traces may be much more sensitive. Still, the results are much more promising than I was expecting. As I move down the semiconductor technology roadmap, I'll revisit this and see how well it does against finer masks.
Thanks to Will Carder for lending me the torch!

Sunday, April 11, 2010

PLACE SANDWICH HERE: the camera

The original camera setup involved zip-tying t-slot aluminum to the microscope neck and positioning the camera based on a series of adjustments to the slot angles. There were a few issues with this setup. First, I didn't have any t-slot nuts, so it was always a pain to tighten things. I at first didn't have any t-slot L brackets either, but I had some nearly equivalent aluminum brackets, so that wasn't as big of a deal. Next, the small distances between the neck and the eyepiece really limited the flexibility in positioning the camera. It was very hard to position it accurately. Finally, the zip ties were only moderately stable, so if you hit it too hard, it would shift around.
The second setup I tried was to use a neck lamp as a flexible mount. It was abandoned because the camera was so much heavier than the light bulb that it caused it to sag. I took an old style fuse and ran a 1/4" bolt through it so it could mount to a camera. The camera isn't mounted to it here, but here I am disassembling it since I'm not going to use it anymore and bad things would happen if someone plugged in the lamp by mistake:

As the current setup is being upgraded for CNC control, the camera positioning will still remain manual since it doesn't need to be moved as the die is being scanned. Here is an overview of it:
Basically, this has two easy to position axis, which really help. The microscope eyepeice is on an angular axis and there is a linear slide that the camera is mounted on. Usually only minimal adjustment is needed with the t-slots, mostly for height. The slide was spring loaded to keep accurate positioning. When viewing the image manually, the eyepeice is swung around to the front of hte microscope where the neck is rather than at a 90 degree offset to standard orientation. This makes switching between manual and camera based viewing convenient.
Interesting to see the focal length as it relates to magnification. Here is 10X objective:

And here is 40X objective:


It looks like its touching, but its not. However, its very, very close. You can still get it out of focus by moving it closer. If you move it too close, the spring loaded objective lenses will move up rather than break. Here is what you see on the camera at 400X (more like 800X actually since we are zoomed in on the camera itself):
There is also a 100X objective, but I haven't looked into what it would take to use it. I believe I need to do something with immersion oil.
The original setup used an incandescent light. However, it only provided moderately acceptable light levels with the 40X objective. I recently bought a 500W halogen light that has really helped. However, its not very directional, so I installed a sandwich wrapper as a deflector to increase directionality to the sample and not blind the operator. As my friend Alex finished his lunch, he probably never thought about his burger wrapper again until he reads this. Little did he know it would become a critical ingredient in the camera setup after being banished to the gray cylinder labeled "TRASH."
Although it is festive, I'll probably replace it with some aluminum foil when I get a chance to go to the store or something.
This setup is pretty stable and I'm a lot happier with it over the previous one. It seems like a flexible mount would have been nice, but with the stability I've gained from the T-slot, this setup is probably best.
Thanks to Dane Kouttron (http://transistor-man.com/Index.html) for the pictures! Also thanks to Magesh Alagiriraj for lending me an SD card since mine died :(