Working as a Chip designer

[PauloConstantino]

Hi all,

I am doing a masters degree in microelectronics, and I want to be a chip designer. But I'm a bit confused as to exactly what** I should study that is most important.

Firstly what kind of jobs exist? What kind of people work in the actual foundries? Are they chemists and physicists? Who are the people who actually make the chip, rather than just design them? I think this is the role I'm mostly interested in. I want to be part of the team that actually creates the chip. I love design too, but I want something more fundamental.

Lastly, do the people who do the design, using software, do they have to learn semiconductor physics? Or are they mostly ignorant of it, since all they probably need to know is how to use the design software and follow the rules?

So then, who should learn semiconductor physics? The guys working in the foundries? And who are the guys who design the equipment that builds the chips in the foundries?

Please I appreciate your advice.

[Brumby]

To start off, I suggest you have a read of this thread....

https://www.eevblog.com/forum/jobs/hardware-design/msg1195852/#msg1195852

[hans]

Lastly, do the people who do the design, using software, do they have to learn semiconductor physics? Or are they mostly ignorant of it, since all they probably need to know is how to use the design software and follow the rules?

So then, who should learn semiconductor physics? The guys working in the foundries? And who are the guys who design the equipment that builds the chips in the foundries?

Yes, if you want to be an (analog) IC design engineer you do have to learn that stuff. If you're doing analog IC design on a masters university level, pay some notable attention to electric fields & magnetics courses. I finished an analog design masters course last semester, and needless to say without that course it was a struggle, knowing also that the lecturer will fail anyone on the oral exam who can't explain fundamentals (like drift vs diffusion current), even if you're a legend on the rest.

It's not like you're constantly working on the physics level on each design step, but for example in CMOS design it's important to understand what parameters are tweak-able (like width/length of CMOS devices) and what implications those adjustments will have. Also there are a lot of 2nd order and even higher order effects in CMOS design, so understanding of those is essential. This is necessary to get any chance of a working design, let alone good.

In general I think (analog) IC design can sit between borderline physics level up (model of 1 transistor) to systems design level (like some kind of SoC).
There are also groups that focus more on the physics level and actually don't care much at all how systems are designed with them. They just want to build better transistors: more bandwidth, better matching, lower leakage currents, etc.
I would also like to add that I've seen a lot of fellow students and PhDs from mixing majors. Some (applied) physics students doing analog IC design course, where maybe an EE takes extra courses in device physics.

I didn't mention digital IC design much at all: it's mostly synthesized by computer from HDL languages to pretty standard "logic cells" at an ASIC level. Yet some device physics can still be useful, given that clock distribution in high-speed (GHz) digital IC's is not that easy, especially at low-power.

I am not sure what education requirements are needed for the people working in the "factory". But my intuition tells me that it's not going to be master university level at all.

ASML is the current leader of chip lithography machines. I think TSMC, etc. all buy their machines. I've been told the competition is not really up there, but kept alive artificially to have a '2nd source'.

[PauloConstantino]

Hi friends. Than you for your response.

I think my masters course is pretty weak. We have only 2 electronics modules, Integrated circuit design and digital system design. The rest is stuff like matlab, project management and embedded programming. I find it very disturbing and I want to switch universities and go to another that focus only on electronics design.

What I really want to know is, who are the guys who actually design the chips including the masks? and what does it mean to design the "process"?

I want to be the guy who designs the chips, including the masks, but I also want to understand the quantum mechanics of semiconductors. I want to know if the guys working in the fabric actually know physics, what do they do ? Is it all software? I kinda go in circles trying to figure out who are the guys who work at thhe physics level but also the design level, all combined.

Who is the engineer who develops better transistors and puts them in analog chips?

I have designed a cpu from scratch using logic chips, and I find it interesting. But I find it extremely boring to write in a HDL. I want to design analog or digital chips from first principles, including the masks, and to develop the process of making the chips too.

[KJDS]

.... I want to design analog or digital chips from first principles, including the masks, and to develop the process of making the chips too.

No-one does it from first principals any more.

The physics guys design the process, which defines what the transistors can do.
The chip designer sorts out either VHDL or an analog equivalent to design the circuitry. Occaisionally a new function will be needed, so someone will sort out a cell just for that function so it can be used in a bigger IC.
The software mostly sorts the masks out.

The world has moved on from one person doing it all.

[PauloConstantino]

So it's the physics guys that design the process at the fab? Hmm...

So basically the task is totally fragmented between various people....

I want to be the semiconductor quantum mechanics guy. Are these the process engineers?

[coppice]

I want to be the guy who designs the chips, including the masks, but I also want to understand the quantum mechanics of semiconductors.

You need a time machine. Head back to the 1960s, and you should find what you are looking for.

I want to know if the guys working in the fabric actually know physics, what do they do ? Is it all software? I kinda go in circles trying to figure out who are the guys who work at thhe physics level but also the design level, all combined.

If by working in the fabric you mean working in places like TSMC, they have people working on the basic physics of next generation processes. They use a lot of software to model the physics, and they apply what they learn to actual physical systems. Hardly any science or engineering these days is done in isolation from software tools of some kind. There is a reason fab companies, like TSMC, have been able to separate themselves completely from chip design companies - there's very little overlap in what the two groups do any more.

Who is the engineer who develops better transistors and puts them in analog chips?

You've conflated two things there. Most of the "design better transistors" work occurs in fab companies, like TSMC. Most of the "putting them in analog chips" work occurs in chip design companies.

I have designed a cpu from scratch using logic chips, and I find it interesting. But I find it extremely boring to write in a HDL. I want to design analog or digital chips from first principles, including the masks, and to develop the process of making the chips too.

Sounds like you really need that time machine.

[rstofer]

The physics level design is done by the chip manufacturers and they provide a library of cells that the device designers can use.  The cells are known to be manufacturable and are well understood.  So, device design is simply pulling cells from libraries and interconnecting them.  Well, perhaps 'simply' isn't the right word...

I haven't been involved with wafer fab since about '90 so I have no idea where things have gone in the last 27 years but I did note that the Intel Kaby Lake based Xeon processors are using 14 nm technology.  A human hair is about 100,000 nm in diameter!

https://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#.22Dunnington.22_.2845_nm.29

Here's some information on lithography:

http://www.phys.sinica.edu.tw/TIGP-NANO/Course/2016_Spring/Notes/Lithography_20160331.pdf

And small feature size:

https://en.wikipedia.org/wiki/14_nanometer

Other than very top tier universities (Standford, MIT, maybe Berkeley), I wouldn't think courses dealing with device physics at this level would be offered or have very much credibility.  There are probably only a few dozen people in the world that understand silicon at this level.  Actually, the problem isn't with silicon, it is masking, etching, doping and metalizing at this small feature size is the trick.

Don't worry about it!  Everything you need to know you can learn at work.  Get an entry level job in the field that interests you and pay attention.

In the late '80s, we could build a wafer fab, including equipment for perhaps $100M.  Today the cost is probably closer to $3-4 Billion!  That's why there are only a few foundaries, small chip makers can't afford to own a fab.  So, they design the chip and send it out for manufacturing.

https://en.wikipedia.org/wiki/Semiconductor_fabrication_plant

[coppice]

Other than very top tier universities (Standford, MIT, maybe Berkeley), I wouldn't think courses dealing with device physics at this level would be offered or have very much credibility.  There are probably only a few dozen people in the world that understand silicon at this level.  Actually, the problem isn't with silicon, it is masking, etching, doping and metalizing at this small feature size is the trick.

There are considerably more than a few dozen, but all the ones I have met have Dr in from of their names.

[PauloConstantino]

Excellent replies everyone. Thank you.

I didn't know that today they split chop design from manufacturing. I know Intel doesn't however, and maybe some others. I knew that ARM for example just make the designs and sell them to others.

One of my modules are university is working with cadence, and designing chips, but I find it extremely boring and artificial. If drawing rectangles on a screen today is chip design... Of course before doing that one needs to actually design the logic, but drawing rectangles is just boring. I thought tools like verilog would synthesize the physical layers for you? Doesn't it ?

So what I'm looking for then are the guys who work at the fab, researching on better transistors and designing the processes that create these transistors? So I should try and work for a fab when I graduate?

Actually I want to do a phd after i finish my msc. Is it possible to do a phd in the area of transistor research and process design? Or is this something you only learn inside a fab ?

Thank you all

Paulo

[coppice]

Excellent replies everyone. Thank you.

I didn't know that today they split chop design from manufacturing. I know Intel doesn't however, and maybe some others. I knew that ARM for example just make the designs and sell them to others.

One of my modules are university is working with cadence, and designing chips, but I find it extremely boring and artificial. If drawing rectangles on a screen today is chip design... Of course before doing that one needs to actually design the logic, but drawing rectangles is just boring. I thought tools like verilog would synthesize the physical layers for you? Doesn't it ?

So what I'm looking for then are the guys who work at the fab, researching on better transistors and designing the processes that create these transistors? So I should try and work for a fab when I graduate?

Actually I want to do a phd after i finish my msc. Is it possible to do a phd in the area of transistor research and process design? Or is this something you only learn inside a fab ?

Thank you all

Paulo

You need to consider where you want to spend your working life. Your tag indicates that you are in the UK. If you want to stay there, fab/foundry work is probably not for you. The UK used to have quite a few fabs, but they've mostly gone. The UK still has a number of chip design businesses, but if your focus is entirely on process few of those chip design businesses will have even one appointment that matches your skills.

[ECEdesign]

As others have said, fabrication and design are very different worlds.  One does need to know a little about the other side, however.  When the chips were much larger they actually used tape to "tapeout" the design fo the mask.  Finishing the design is still called a tapeout for historical purposes.  Cadence is pretty much as close to designing the mask as you can get.  Be aware that when a tapeout deadline approaches it is usually around 100 hours a week and can be quite stressful.  Silicon CMOS technology has very good models (such as BSIM) which were developed over the past 30 years.  GaN is starting to be used for military applications and requires the designer to know a fair amount of knowledge about the physics because the models are not anywhere near where silicon models are today.  Working for the military or on military contracts is usually a good option to learn about this area.  I would much rather be a designer than in the fab.  There are a few reasons for this, including having to work in Asia, working with hazardous chemicals, and reduction in the need for humans in factories.  Back in the day, people filled the fabs but today most of the work is automated.

[Hugoneus]

I just glanced over all the replies. Lots of good advice.

I am as ASIC designer (mm-wave, mixed signal, analog, etc.) I have knowledge about the process as it is important to understand the devices, but I would not consider myself an expert in that field. I also have knowledge on system modelling. I am right in the middle of those two worlds with overlapping skills into each.

I would not want to be a process engineer, nor a modelling engineer. There are not that many good position in that field, I do work some people with those skills however.