Output stage ESD protection for an integrated circuit

The present invention relates generally to integrated circuits, and in particular to electrostatic discharge protection (ESD) of circuitry coupled to an output stage of an integrated circuit.

The protection of integrated circuits (ICS) from damage due to ESD has received increased design attention, particularly as circuit geometries migrate to smaller dimensions. ESD damage can occur as a result of a voltage ESD event or a current ESD event. The book “ESD in Silicon Integrated Circuits” by A. Amerasekera and C. Duvvury, copyrighted in 1995 by John Wiley & Sons, which is hereby incorporated by reference, discloses much about the topic of electrostatic discharge phenomena. The damage may occur during manufacture of the IC chip, or more commonly, after the chip is packaged such as during handling, shipping or use.

One ESD protection technique to protect the input of a packaged IC employs resistors to reduce ESD voltages transmitted to the IC through bond pads. ESD events are transmitted to chip bond pads of packaged chips by package leads. Another technique employs a transistor to clamp the operating voltage on an input bond pad to a safe level. Yet another technique employs a four-layer device, such as a thyristor, to introduce hysteresis into the protective circuitry. Yet another technique protects the input of an integrated circuit from an ESD event has been to provide two steering diodes, each having an area large enough to conduct the expected current

Another prior art technique, shown in FIG. 8, has a first steering diode 30 h between the output bond pad 14 h and the positive supply bond pad 28 h . The first steering diode has a cathode coupled to the positive supply bond pad and an anode coupled to the output bond pad. The first steering diode is reverse-biased and non-conducting under normal operating conditions. The first steering diode provides a low impedance path from the output bond pad to the positive supply bond pad 28 h when the output bond pad voltage is more than one diode voltage drop above the voltage at the positive supply bond pad. The second steering diode 32 h is coupled between the ground bond pad 18 h and the output bond pad 14 h . The second steering diode has a cathode coupled to the output bond pad and an anode coupled to the ground bond pad. The second steering diode under normal operating conditions is reverse-biased and non-conducting. The second steering diode provides a low impedance path from the output bond pad to the ground bond pad in the event the output bond pad voltage is more than one diode voltage drop below the voltage at the ground bond pad. The first and second steering diodes are sized to accommodate the largest current expected due to an ESD event.

In radio frequency circuits, the output bond pad is typically inductively coupled to the positive supply. The DC bias voltage on the collector is almost equal to the positive supply voltage, therefore any positive going signal will cause the output voltage at the output bond pad to exceed the DC supply voltage. When the normal output voltage swing exceeds one diode voltage drop, the first steering diode 30 h conducts and limits the output voltage, undesirably distorting the radio frequency signal.

While such techniques have offered some ESD protection, further improvement is considered necessary. The need for a new ESD protection technique for circuits coupled to an output stage is desirable.

In accordance with the present invention, an integrated circuit includes a circuit coupled between an output bond pad and a ground bond pad. The circuit includes a transistor having a first electrode coupled to the output bond pad and a second electrode coupled to the ground bond pad. A degeneration device is coupled between the second electrode and the ground bond pad. At least one diode is coupled between the second electrode and the ground bond pad with the anode of the at least one diode coupled to the second electrode and the cathode of the at least one diode coupled to the ground bond pad.

FIG. 1 is a schematic diagram of a portion of an integrated circuit incorporating electrostatic discharge protection of an output stage of a circuit, comprising a bipolar transistor and a degeneration device, in accordance with the present invention;

FIG. 2 is a schematic diagram of a portion of an integrated circuit incorporating electrostatic discharge protection of an output stage of a circuit, the output stage comprising a metal oxide semiconductor transistor and a degeneration device;

FIG. 3 is a schematic diagram of a portion of an integrated circuit incorporating an alternate embodiment electrostatic discharge protection circuit comprising a plurality of at least two series coupled diodes;

FIG. 4 is a schematic diagram of a portion of an integrated circuit, similar to FIG. 1, in which a bipolar transistor is configured as a diode;

FIG. 5 is a schematic diagram of a portion of an integrated circuit, similar to FIG. 4, in which a metal oxide semiconductor transistor is configured as a diode;

FIG. 6 is a schematic diagram of a portion of an integrated circuit incorporating electrostatic discharge protection of an output stage of a circuit comprising a PNP transistor;

FIG. 7 is a schematic diagram of a portion of an integrated circuit incorporating electrostatic discharge protection of an output stage of a circuit comprising a four layer transistor; and

FIG. 8 is a schematic diagram of an output stage portion of an integrated circuit illustrating a prior art electrostatic discharge protection circuit.

FIG. 1 illustrates a portion of an integrated circuit 10 a having elements to provide electrostatic discharge protection to circuit 12 a , comprising an output stage, coupled to output bond pad 14 a . While only one output stage is illustrated, it is understood that there may be more than one output stage. Circuit 12 a , which is a portion of integrated circuit 10 a , may be but is not limited to being an amplifier operable at radio frequencies, such as frequencies greater than 500 megahertz. The output stage of circuit 12 a includes an NPN transistor 16 a in a common-emitter configuration, although the invention is not limited thereto. The collector of transistor 16 a is coupled to output bond pad 14 a and the emitter of transistor 16 a is coupled to AC ground bond pad 18 a . There may be other circuit components (not shown) coupled to the base of transistor 16 a . Between the emitter and the ground bond pad 18 a there is a degeneration device 20 a such as but not limited to a resistor or inductor. The degeneration device 20 a linearizes operation of circuit 12 a under normal operating conditions. The voltage drop across the degeneration device 20 a is small during normal operation, typically less than one diode voltage drop. However, during an ESD event, the voltage developed across the degeneration device can be quite large and potentially cause damage to both transistor 16 a and degeneration device 20 a.

Diode 22 a is coupled across degeneration device 20 a with its anode coupled to the emitter of transistor 16 a and its cathode coupled to ground bond pad 18 a . Diode 22 a is forward-biased under normal operating conditions of circuit 12 a . Since the emitter of transistor 16 a under normal operating conditions of circuit 12 a is not more than one diode voltage drop above the potential of ground bond pad 18 a , diode 22 a does not conduct during normal operating conditions of circuit 12 a . Diode 22 a protects circuit 12 a , the output stage of the amplifier, including degeneration devices therein, from excessive voltage/current during an ESD event.

A diode manufactured as small as possible in a given processing/line width technology has a smaller cathode-to-anode capacitance than a larger diode manufactured in the same processing/line width technology. Diode 22 a is sized to accommodate the anticipated ESD current to protect regeneration device 20 a . The capacitive reactance of diode 22 a translates back to the base of transistor 16 a as a negative resistance. A smaller capacitance provides less of a capacitive load to be charged by an output signal developed across the degeneration device. However, a small diode has a larger impedance and therefore provides less protection during an ESD event. Thus, a tradeoff is made between the ESD protection provided and the magnitude of the parasitic capacitance introduced by diode 22 a.

A positive ESD event raises the voltage of output bond pad 14 a . When a positive ESD event occurs such that the voltage at the emitter of transistor 16 a attempts to exceed one diode voltage drop above the potential of ground bond pad 18 a , diode 22 a is sufficiently forward-biased to conduct. Diode 22 a conducting prevents the voltage at the emitter of transistor 16 a from exceeding about one diode voltage drop above the potential at ground bond pad 18 a . By switching to the conducting state, diode 22 a minimizes the voltage developed at the emitter-degeneration device junction of transistor 16 a , thereby protecting the base-emitter junction of transistor 16 a , degeneration device(s) 20 a and hence the output stage of circuit 12 a from an ESD event. Diode 22 a is sized to accommodate the current of any expected ESD event. Transistor 16 a and diode 22 a are sized to carry the anticipated ESD current without damage to the devices themselves or to circuit 12 a and degeneration device 20 a.

A negative ESD event reverse biases the voltage of output bond pad 14 a such that the voltage at output bond pad 14 a becomes more negative than the voltage of ground bond pad 18 a . When a negative ESD event occurs, current flows in the inherent diode ID in transistor 16 a , shown in phantom in FIG. 1 . Current flows between the collector of transistor 16 a and the substrate in which the transistor is formed preventing reverse bias of the base-emitter junction of transistor 16 a from increasing and eventually exceeding the power dissipation the base-emitter junction can withstand without damage. One skilled in the art would know there is either a metallic connection (as shown) or back-to-back diodes(not shown) between the substrate and the ground bond pad.

A common emitter configuration of the invention illustrated in the circuit of FIG. 1 couples the radio frequency input signal to the base of transistor 16 a . A common base configuration of the invention illustrated in the circuit of FIG. 1 couples the radio frequency input signal to the emitter of transistor 16 a . The output signal in both configurations of the invention would appear at the collector of transistor 16 a.

The transistor being protected is not limited to being a bipolar transistor. The transistor may, for example, be a metal oxide semiconductor (MOS) transistor. FIG. 2 illustrates an output stage circuit 12 b including a metal oxide semiconductor transistor 16 b . In circuit 12 b of integrated circuit 10 b , the gate of transistor 16 b is coupled to other components, not shown. The drain of transistor 16 a is coupled to output bond pad 14 b and the source is coupled to ground bond pad 18 b . The breakdown of the gate-to-source capacitance of transistor 16 b must be designed to be greater than one diode voltage drop. Transistor 16 b includes an inherent lateral NPN transistor 17 b , shown in phantom in FIG. 2 . During a positive ESD event, transistor 7 b carries the expected current thereby contributing to protection of transistor 16 b from oxide breakdown. Diode 22 b limits the voltage developed across degeneration device 20 b during an ESD event, thereby preventing damage to circuit 12 b . In an ESD event, the voltage on output bond pad 14 b does not exceed one voltage drop above ground pad 18 b due to the presence of diode 22 b , thereby protecting MOS transistor 16 b and preventing the gate-source capacitance from breaking down.

Transistor 16 a during a positive ESD event conducts from collector to emitter. The transistor operates in breakdown mode which will not damage the transistor if the transistor is sized large enough. If power dissipation due to the ESD event is excessive, the transistor will be damaged. With transistor 16 a operating in breakdown mode, the base-emitter junction voltage may become reverse biased and damage the base-emitter junction. It has been empirically determined that a 70×10E-6 meter by 70×10E-6 meter by 0.15×10E-6 meter transistor 16 a is able to withstand the power dissipation from an ESD event without damage. A transistor of this size is not necessarily the smallest transistor that could withstand power dissipation from an ESD event without damage.

There is an inherent diode, normally reverse biased, between the drain of transistor 16 b and the substrate in which transistor 16 b is formed. The substrate is tied to ground pad 18 b as mentioned above. During a negative ESD event, current flows between ground pad 18 b and output bond pad 14 b , thereby protecting transistor 6 b and circuitry 12 b of integrated circuit 10 b.

FIG. 3 illustrates an alternate embodiment integrated circuit 10 c in which diode 22 a of the FIG. 1 embodiment is replaced with two or more series coupled diodes D 1 , D 2 , . . . , DN coupled in series across degeneration device 20 c . Providing two or more series coupled diodes between the emitter of transistor 16 b and ground bond pad 18 b reduces the capacitive load on output bond pad 14 c , and thus the capacitive load on the emitter of transistor 16 b.

Each of diodes D 1 , D 2 , . . . , DN, is sized to handle currents in the range expected during an ESD event, such as one to two amperes. Since diodes D 1 , D 2 , . . . , DN, are coupled in series, the total capacitance on the emitter of transistor 16 c due to the presence of the diodes is less when compared to the capacitance of a single diode since the total inverse capacitance is the sum of the inverse of each of the individual capacitances. When the diodes are identical, and there are two diodes, the capacitance is reduced to one-half of the capacitance of a single one of the diodes. Reducing the capacitance on the emitter of transistor 16 c reduces the loss of signal provided by circuit 12 c that is developed across regeneration device 20 c.

Furthermore, since the voltage developed across degeneration device 20 c is less than (N−1) diode voltage drops, where N is the number of diodes, under normal operation of circuit 12 c , the diodes D 1 , D 2 , . . . , DN, although forward biased, are not sufficiently forward biased to be conductive. Concomitantly, the voltage at the emitter of transistor 16 b must rise to a voltage level corresponding to the number of diode voltage drops represented by the series coupled diodes to switch the forward-biased series coupled diodes to be conductive. Under conditions of an ESD event where output bond pad 14 c is at a higher potential than ground bond pad 18 c , diodes D 1 , D 2 , . . . , DN would be sufficiently forward biased to be conductive, thereby providing a path to ground for current produced by an ESD event and in turn the diodes protect transistor 16 c and circuit 12 c from damage.

The maximum number of diodes across degeneration device 20 b that may be used to provide ESD protection is limited by the breakdown voltage of the emitter-base junction of transistor 16 b . The breakdown of the emitter-base junction of transistor 16 b is undesirable. The number of series coupled diodes is selected such that the voltage at the emitter-base junction rising causes the diodes to become conductive at a voltage level before the voltage is reached that would cause breakdown of the emitter-base junction of transistor 16 b . More than a few diodes result in lessened ESD protection. Preferably, no more than three diodes are employed. One skilled in the art could design a metal oxide semiconductor circuit 12 c using multiple series coupled diodes D 1 , D 2 , . . . , DN.

FIG. 4 illustrates an alternate embodiment integrated circuit 10 d in which a transistor is configured as a diode TD 1 across degeneration device 20 d . The transistor configured as a diode in FIG. 4 is illustrated as a bipolar transistor but the invention is not limited thereto.

FIG. 5 illustrates an alternate embodiment integrated circuit 10 e in which a metal oxide semiconductor transistor is configured as a diode MD 1 across degeneration device 20 e.

FIG. 6 illustrates an alternate embodiment integrated circuit 10 f in which a PNP transistor 16 f is employed in circuit 12 f of the amplifier output stage. Transistor 16 f is configured in a common emitter configuration. One skilled in the art would recognize the need to reverse polarity of the power supply, as well as the location of regeneration device 20 f and the location and polarity of diode 22 f.

FIG. 7 illustrates an alternate embodiment integrated circuit 10 g in which a four-layer transistor, such as a silicon-controlled rectifier or triac, provides electrostatic discharge protection. The four layer transistor provides an alternative path to ground for current due to an ESD event. A silicon-controlled rectifier may be triggered by the voltage at the emitter of transistor 16 g . A silicon-controlled rectifier for such application should have a holding current that is greater than the emitter bias current of transistor 16 g under normal operating conditions.

FIG. 8 is a schematic diagram of a portion of an integrated circuit 10 h illustrating prior art electrostatic discharge protection of an output stage.

As is known in the art, the integrated circuits 10 a through 10 h are typically enclosed in a package. The integrated circuit is positioned on the paddle of a lead frame having a paddle and isolated leads. Wires are bonded between corresponding bond pads on the integrated circuit and leads of the lead frame. The integrated circuit chip and bond wires are enclosed in the package, such as in an overmold operation.

While reference has been made to an output bond pad and a ground bond pad, coupling to the electrical equivalent of these bond pads would suffice. Transistors illustrated in the various embodiments are shown as a particular type of transistors, although the invention is not limited thereto. Other types of transistors may be used with the invention. While the above embodiments give typical design values suitable for an illustrative embodiment implemented in a particular technology, a wide range of transistors, resistors, and diode sizes are possible depending upon the protection needs required and component values.