Silicon die analysis: inside an op amp with interesting "butterfly" transistors

Some integrated circuits have very interesting dies under a microscope, like the chip below with designs that look kind of like butterflies. These patterns are special JFET input transistors that improved the chip's performance. This chip is a Texas Instruments TL084 quad op amp and the symmetry of the four op amps is visible in the photo. (You can also see four big irregular rectangular regions; these are capacitors to stabilize the op amps.) In this article, I describe these components and the other circuitry in the chip and explain how it works. This article also includes an interactive chip explorer that shows each schematic component on the die and explains what it does.

Die photo of the TL084 quad op amp. The bond wires got a bit bent while cleaning the chip.

Die photo of the TL084 quad op amp. The bond wires got a bit bent while cleaning the chip.

An integrated circuit consists of a tiny piece of silicon. To make an integrated circuit, regions are treated with various atoms to change the properties of the silicon, giving them different colors under a microscope. On top of the silicon, a thin layer of metal connects different parts of the chip. This metal is clearly visible in the photo as yellowish traces and regions. Under the metal, a thin, glassy silicon dioxide layer provides insulation between the metal and the silicon, except where contact holes in the silicon dioxide allow the metal to connect to the silicon. Around the edge of the chip, thin bond wires connect the metal pads to the chip's external pins.

NPN transistors inside the IC

Transistors are the key components in a chip. This op amp chip uses several types of transistors: NPN and PNP bipolar transistors as well as JFETs. (Many newer op amps use low-power CMOS transistors instead.) If you've studied electronics, you've probably seen a diagram of an NPN transistor like the one below, showing the collector (C), base (B), and emitter (E) of the transistor, The transistor is illustrated as a sandwich of P silicon in between two layers of N silicon; the N-P-N layers make an NPN transistor. It turns out that transistors on a chip look nothing like this, and the base often isn't even in the middle!

Symbol and simplified structure of an NPN transistor.

Symbol and simplified structure of an NPN transistor.

The photo below shows one of the transistors in the TL084 as it appears on the chip. The different brown, purple and green colors are regions of silicon that has been doped differently, forming areas called N and P regions (negative with an excess of electrons, and positive lacking electrons). The yellow areas are the metal layer of the chip on top of the silicon—these form the wires connected to the collector, emitter, and base. Underneath the photo is a cross-section drawing showing approximately how the transistor is constructed. There's a lot more going on than just the N-P-N sandwich you see in books, but if you look carefully at the vertical cross section below the 'E', you can find the N-P-N that forms the transistor. The emitter (E) wire is connected to N+ silicon. Below that is a P layer connected to the P+ base contact (B). And below that is an N layer connected (indirectly) to the collector (C).1

Structure of an NPN transistor in the TL084 op amp

Structure of an NPN transistor in the TL084 op amp

While most of the transistors follow the above pattern, some of the transistors in the TL084 chip are optimized in confusing ways, such as the part of the die below. In this circuit, two transistors share one collector (C), while a resistor (blue line) runs between them. (This took me a while to figure out, even with the schematic.)

A complex part of the TL084, where two transistors share a collector while a resistor runs through them.

A complex part of the TL084, where two transistors share a collector while a resistor runs through them.

The output transistors (below) in the TL084 are larger than the other transistors and have a different structure in order to produce the chip's high-current output. The output transistors must provide milliamps of current, compared to microamps for the internal transistors. Note the interlocking "fingers" of the emitter (E) and base (B), surrounded by the large collector (C). Although the NPN and PNP transistors look similar, the dark purple P silicon is visible on the base of the NPN transistor and the emitter and collector of the PNP transistor, showing their opposite construction.

High-current NPN and PNP transistors drive the output of the TL084 op amp

High-current NPN and PNP transistors drive the output of the TL084 op amp

How capacitors are implemented in silicon

The TL084 contains four capacitors to provide stability for the op amps. You can see the four capacitors in the die photo; they are the largest structures on the chip. A capacitor in the chip is essentially a large metal plate separated from the silicon by an insulating layer. The main drawback of capacitors on ICs is they are physically very large. The TL084's capacitors have a very small capacitance value (a few picofarads) but take up a large fraction of the chip's area.2

JFET transistors3

A special type of transistor called a JFET is the key to the high performance of the TL084 chip. The JFET transistor is related to the more common MOSFET transistor: they both controls current between the source and the drain, under control of the gate. But while the MOSFET has has an insulating oxide layer between the gate and the body of the device, the JFET lacks this layer and has a silicon P-N junction instead (and thus is called a Junction FET). The chip used P-channel JFETS, where current flows through a channel of P silicon; the schematic symbol and basic structure is shown below. Normally, current flows between the source (S) and drain (D) through the channel. As the voltage on the gate increases, it "pinches" the channel closed, reducing and then stopping the current flow. An important feature of a JFET is that very, very little current flows through the gate; the gate resistance is an amazingly large 1012Ω. (This is because the gate junction acts as a reverse-biased diode, blocking current flow.) This high input impedance is an important feature for an op amp.

Symbol and simplified structure of a JFET transistor (P-channel).

Symbol and simplified structure of a JFET transistor (P-channel).

On the chip, the JFETs are constructed like the diagram above but rotated horizontally. The diagram below shows a JFET as it appears on the die (left), along with a close-up slice. (The JFET channel is wide and snakes around in order to pass more current. It also has drains on both sides of the source.) The cross section below shows the internal structure of the JFET. The P region connects the source and the drain, and it is surrounded above and below by the gate's N region. (The connection to the lower N region is outside the region shown.) The JFETs in this chip are built with ion implantation, which shoots accelerated ions into the chip to produce the P and N regions. Ion implantation provides accurate control of the doping and dimensions of the P channel between the source and drain, allowing the input JFETs to be built for high performance.

Cross section of an input JFET transistor, showing the construction of the JFET.

Cross section of an input JFET transistor, showing the construction of the JFET.

Manufacturing JFET op amp ICs was difficult when they were first sold decades ago. Hybrid (two separate dies in one package) JFET op amps were introduced in 1970. These were followed shortly afterwards by monolithic (i.e. a single die) op amps, but difficulties in manufacturing consistent JFETs caused these op amps to have poor characteristics. In 1974, National Semiconductor engineers developed the ion implantation technique for fabricating consistent, high quality JFETs and used this "BIFET" technique to build better JFET op amps. Two years later Texas Instruments introduced their JFET op amps, including the TL084 which was the first four-in-one op amp using the BIFET process.4

You might have noticed that each op amp has four input JFETs on the die (forming the butterfly pattern below), even though the op amp only has two inputs. The explanation for this is that for good performance the input transistors in an op amps should have identical electrical characteristics. But unfortunately chips can have thermal gradients (i.e. hotter on one side than the other) that affect the transistor characteristics and unbalance the inputs. A standard solution used in the TL084 is that each input uses two cross-coupled transistors, diagonally opposite from each other. If one side of the chip is hotter than the other, both inputs will have an affected transistor, canceling out the effect of the temperature gradient.

To insure the input transistors are matched, each input transistor is actually two connected transistors, diagonally opposite. Wiring connects each transistor pair.

To insure the input transistors are matched, each input transistor is actually two connected transistors, diagonally opposite. Wiring connects each transistor pair.

IC component: The current mirror

There are some subcircuits that are very common in analog ICs, but may seem mysterious at first. Before explaining how the TL084 works, I'll first give a brief overview of the current mirror and differential pair circuits.

Schematic symbols for a current source.

Schematic symbols for a current source.

If you've looked at analog IC block diagrams, you may have seen the above symbols for a current source and wondered what a current source is and why you'd use one. The idea of a current source is you start with one known current and then you can "clone" multiple copies of the current with a simple transistor circuit.

The following circuit shows how a current mirror is implemented.5 A reference current passes through the transistor on the left. (In this case, the current is set by the resistor.) Since both transistors have the same emitter voltage and base voltage, they source the same current, so the current on the right matches the reference current on the left.

Current mirror circuit. The current on the right copies the current on the left.

Current mirror circuit. The current on the right copies the current on the left.

A common use of a current mirror is to replace pull-up resistors. Resistors inside ICs are both inconveniently large and inaccurate. It saves space to use a current mirror instead of a resistor whenever possible. Also, the currents produced by a current mirror are nearly identical, unlike the currents produced by two resistors.

IC component: The differential pair

The second important circuit to understand is the differential pair, the most common two-transistor subcircuit used in analog ICs.6 You may have wondered how the op amp subtracts two voltages since it's not obvious how to make a subtraction circuit. This is the job of the differential pair.

Schematic of a simple differential pair circuit. The current source sends a fixed current I through the differential pair. If the two inputs are equal, the current is split equally.

Schematic of a simple differential pair circuit. The current source sends a fixed current I through the differential pair. If the two inputs are equal, the current is split equally.

The schematic above shows a simple differential pair. The key is the current source at the top provides a fixed current I, which is split between the two input transistors. If the input voltages are equal, the current will be split equally into the two branches (I1 and I2). If one of the input voltages is a bit lower than the other, the corresponding transistor will conduct more current, so one branch gets more current and the other branch gets less. As one input continues to increase, more current gets pulled into that branch. Thus, the differential pair is a surprisingly simple circuit that routes current based on the difference in input voltages. The TL084 uses JFETS instead of bipolar transistors in the differential pair, but the principle is the same.

The internal blocks of the op amp

The internal circuitry of the TL084 op amp is similar to the 741 op amp, which has been explained in many places7, so I'll just give a brief description of the main blocks. The interactive chip viewer below provides more explanation.

The two input pins are connected to the differential amplifier, which is based on the differential pair described above. The output from the differential amplifier goes to the second (gain) stage, which provides additional amplification of the signal. Finally, the output stage has large transistors to generate the high-current output, which is fed to the output pin. The capacitor stabilizes the op amp to avoid oscillation. The current mirror at the top provides currents to other parts of the chip. The current mirror at the bottom functions as an active load increasing the gain of the differential pair.

Die for the TL084 op amp, showing the main functional units.

Die for the TL084 op amp, showing the main functional units.

Interactive chip viewer

The image and schematic8 below are an interactive exploration of the TL084. Click a component to see its location on the die and in the schematic highlighted. The box below will give an explanation of the component. The die image below shows one of the four op amps on the chip; the others are essentially identical.

Click components in the image below for more information.

How I photographed the die

Getting to the die of an integrated circuit can be difficult since integrated circuit usually come in a black epoxy package. I'd rather avoid using dangerous concentrated acid to dissolve the epoxy package and see the die. Fortunately some ICs, such as the TL084, are available in ceramic packages that can be easily opened with a chisel. The photo below shows the chip package after removing the lid. The four large capacitors are visible on the die even without a microscope.

The TL084 op amp. The ceramic package has been opened, exposing the die inside. A couple pins fell off when the package was opened.

The TL084 op amp. The ceramic package has been opened, exposing the die inside. A couple pins fell off when the package was opened.

To obtain the die photos, I used a metallurgical microscope, which shines light from above through the lens, unlike a normal microscope which shines light from below. A metallurgical microscope is the secret to getting clear photos at higher magnification, since the die is brightly illuminated. I used Hugin to stitch multiple images together into high-resolution pictures. Below is a second die photo; the bond wires are removed in this one.

Die photo of the TL084 op amp with the bond wires removed.

Die photo of the TL084 op amp with the bond wires removed.

Conclusions

Texas Instruments introduced the TL084 in 1976 as one of the first high-performance quad op amps. I was motivated to study this chip by the pretty butterfly-like patterns on the die, but found some interesting circuitry inside the chip. The butterfly-like structures turned out to be JFET transistors that improved the chip's performance by providing high impedance for the op amp inputs. If you enjoyed this look inside an analog silicon chip, you may also like my analysis of the 741 op amp and 555 timer. Follow me on Twitter at @kenshirriff for my latest blog posts, or use my RSS feed. The chip was provided by Eric Schlaepfer (@TubeTimeUS).

Notes and references

  1. You might have wondered why there is a distinction between the collector and emitter of a transistor, when the simple picture of a transistor is totally symmetrical. Both connect to an N layer, so why does it matter? As you can see from the die photo, the collector and emitter are very different in a real transistor. In addition to the very large size difference, the silicon doping is different. The result is a transistor will have poor gain if the collector and emitter are swapped. 

  2. The capacitor in the op amp is located at a special point in the circuit where the effect of the capacitance is amplified due to something called the Miller effect. This allows the capacitor to be much smaller than it would be otherwise. Given how much of the die is used for the capacitor already, taking advantage of the Miller effect is very important. 

  3. Yes, I realize that "JFET transistor" is a redundant acronym. Since some readers may not be familiar with JFETs, I want to remind them that JFETs are transistors. 

  4. For an extremely detailed history of op amps, including the development of JFET op amps in the 1970s, see Op Amp History by Walt Jung. My section on JFET op amp history is based on this source. 

  5. For more information about current mirrors, you can check Wikipedia or chapter 3 of Designing Analog Chips. If you're interested in how analog chips work, I strongly recommend you take a look at Designing Analog Chips. 

  6. Differential pairs are also called long-tailed pairs. According to Analysis and Design of Analog Integrated Circuits differential pairs are "perhaps the most widely used two-transistor subcircuits in monolithic analog circuits." (p214) For more information about differential pairs, see Wikipedia, any analog IC book, or chapter 4 of Designing Analog Chips

  7. The TL084 op amp's design is similar to the 741 op amp, which is described in Wikipedia, Operational Amplifiers, IC Op-Amps Through the Ages and UNCC class notes. See any of those sources for more details on how op amps are constructed. 

  8. The schematic is from the TL084 datasheet

15 comments:

Franc said...

Thanks! Nice article, as usual :), and really like the interactive die-diagram you put in there!

Jonathan Hendry said...

What's the difference between N and N+ or P and P+?

Sooner Boomer said...

The difference between P (or N) and P+ (or N+) is the level of doping. The + portion have higher levels of doping and act as ohmic contacts rather than semiconductor junctions.

Unknown said...

Great article. Thanks very much.

John Pi said...

Your differential pair is not shown correctly (or at least isn't useful as shown). The current source should drive the common emitters (long-tailed pair), and the resistors are in each collector. Now, the circuit will route the shared current source to the right or left, depending on the difference between the base voltages. The small-signal gain of this type of circuit is approximately IBIAS*RLOAD/0.026 -- this comes from the exponential relationship between transistor collector coronet and the base-emitter voltage.

Ken Shirriff said...

Thanks John for the info. I've updated the differential pair diagram.

Mark Jeronimus said...

The differential pair diagram doesn't show up at all. Maybe a private image which you see because you're logged in to Google?

Unknown said...

"Diode {next to C1} prevents stabilization capacitor from going negative."

My understanding: it clamps the gain-stage so it does not go all the way to saturation. If the gain-stage saturates it has slow recovery after overload.

Ken Shirriff said...

Thanks Unknown; that makes sense. I've updated the description.

Jonathan Hendry said...

Sooner Boomer: Thanks!

行天下 said...
This comment has been removed by the author.
Richard said...


My friend, awesome work!

I feel like I'm working as a palentologist, discovering the origin of what we have today.

Like the search for the missing link.

A real innovation of technology.

Congratulations one more time.

Anonymous said...

That was a very interesting read, thank you.
I played around with the interactive schematic and picture combination which was is really well done.

It does seem to me that the link from each of the 64Ω output resistors is swapped compared to the area that they highlight. That is, they don't match the output transistor that they are connected to.

Regards, PE1NUT

Richard said...


One of the things I'm most impressed with is how you can open a chip without damaging the view? and how can you shoot the DIAL so perfectly?

There is a video on youtube that shows a laser device that can open the chip without damaging it
https://www.youtube.com/watch?v=4aGrvZFdkSw - I do not know if it would suit the older chips.

Unknown said...

Very impressive. Just curious in case you might know, the 741 was made by many manufacturer's. Did they all use the same die pattern i.e. was it licensed so they would all be identical or did each manufacturer create their own?