Displacement current trigger SCR

[0002] 1. Field of the Invention

[0003] The invention relates to the protection of integrated circuits from electrostatic discharge (ESD), and more particularly to a low voltage protection of CMOS circuits by a parasitic silicon controlled rectifier (SCR) which is triggered by a displacement current.

[0004] 2. Description of the Related Art

[0005] The protection of integrated circuits from electrostatic discharge (ESD) is a subject which has received a lot of attention from circuit designers because of the serious damage that ESD can wreak as oxide thicknesses are reduced together with device dimensions. The seriousness of the problem is reflected in the number of articles published and U.S. Patents issued. Workers in the field and inventors have proposed many solutions, many trying to solve the problem of protecting sub-micron devices while still allowing them to function unencumbered and without undue, or zero, increase of silicon real estate. The main thrust of ESD protection for CMOS devices is focused on the use of parasitic npn and pnp bipolar transistor, which together form an SCR. Unwanted as this SCR is normally, it can safely discharge dangerous ESD voltages as long as its trigger voltage is low enough to protect those CMOS devices of which it is a part. Past solutions have offered low voltage lateral SCRs (LVTSCR), modified lateral SCRs (MLSCR), PMOS-trigger lateral SCRs (PTLSCR), NMOS-trigger lateral SCRs (NTLSCR), and modified PTLSCRs and NTLSCRs.

[0006] FIG. 1 is a cross-sectional view of a typical prior art CMOS structure protected from ESD pulses by an SCR and NMOS transistor. Shown is a semiconductor wafer 10 with CMOS devices, with parasitic bipolar transistors forming an SCR, and an additional NMOS device for lowering the trigger voltage of the SCR. In a p-substrate 11 an n-well 12 is formed, and a p-channel transistor with a p⁺ source 14 and a p⁺ drain (not shown) is created. An n⁺ contact region 13 is formed in the n-well and together with p⁺ source 14 connected to a voltage supply pad 19. In p-substrate 11 an n-channel transistor with an n⁺ drain 15, an n⁺ source 16, and a gate 17 is created. The n⁺ drain 15 straddles p-substrate 11 and n-well 12. A p⁺ contact region 18, formed in p-substrate 11, is connected together with n⁺ source 16 to a reference voltage 20.

[0007] The steps that produce the above CMOS structure also create parasitic bipolar pnp transistor 21 between p⁺ source 14 (emitter), n-well 12 (base), and p-substrate 11 (collector), and parasitic bipolar npn transistor 22 between n⁺ source 16 (emitter), p-substrate 11 (base), and n-well 12 (collector). The base of transistor 21 is connected via n-well resistor 23 to n⁺ contact region 13, and the base of transistor 22 is connected via p-substrate resistor 24 to p⁺ contact region 18. The base of one transistor is connected to the collector of the other transistor. Resistors 23 and 24 are equivalent resistors for the intrinsic resistance of the n-well and p-substrate material. FIG. 2 is the equivalent circuit of FIG. 1 showing the interconnection of transistor 21 and 22 forming an SCR. NMOS transistor Q1 is shunted across npn transistor 22 providing the trigger for the SCR. ESD voltage pulses are shunted from pad 19 via transistors 21 and 22 to reference voltage (ground) 20.

[0008] U.S. Patents relating to ESD protection are:

[0009] U.S. Pat. No. 5,754,381 (Ker) provides a modified PTLSCR and NTLSCR, and bypass diodes for protection of the supply voltage and output pad of an output buffer. The trigger voltage is the low snap-back trigger voltage of a short-channel PMOS (NMOS) device.

[0010] U.S. Pat. No. 5,754,380 (Ker et al.) is similar to U.S. Pat. No. 5,754,381 above but without bypass diodes. The invention requires a smaller layout area than conventional CMOS output buffers with ESD protection.

[0011] U.S. Pat. No. 5,744,842 (Ker) teaches the use of an ESD protection circuit between V_ddand V_ss, comprising a bipolar transistor and an n-type field-oxide device.

[0012] U.S. Pat. No. 5,742,085 (Yu) discloses a low-voltage trigger protection, consisting of two SCRs and an NMOS transistor for triggering disposed between an IC bonding pad and V_ss.

[0013] U.S. Pat. No. 5,719,733 (Wei et al.) presents an ESD protection device using one SCR for deep submicron CMOS devices, where the structure comprises a p-well and an n-well (or p-substrate) that are separated. A ground electrode is connected to a p⁺ and an n⁺ contact region and through a polysilicon region to a gate oxide region in the n-well. The triggering voltage for snapback of the SCR is tunable between 5-11 Volts.

[0014] U.S. Pat. No. 5,576,557 (Ker et al.) provides ESD protection for sub-micron CMOS devices supplying discharge paths at V_ddand V_ssusing two LVTSCRs. In addition a PMOS device is used in conjunction with one LVTSCR and an NMOS device with the other LVTSCR. Inclusion of the PMOS and NMOS devices allows lowering of the trigger voltage to 11-13 Volt.

[0015] U.S. Pat. No. 5,572,394 (Ker et al.) discloses an ESD protection circuit formed by two PTLSCRs and two NTLSCRs for protection from the four modes of ESD. The use of short-channel thin-oxide PMOS (NMOS) devices reduces the turn-on voltage of the lateral SCR to below the gate-oxide breakdown voltage of CMOS devices in the input stage.

[0016] U.S. Pat. No. 5,541,801 (Lee et al.) uses three LVTSCRs which are connected between V_dd, the circuit to be protected, and V_ss. Each of the SCRs uses a PMOS/NMOS transistor to lower the trigger voltage. The gates of the PMOS/NMOS transistors are each in turn connected via linked terminals of trigger gates to the circuit to be protected.

[0017] U.S. Pat. No. 5,452,171 (Metz et al.) describes a protection circuit using an inverter trigger device, with a voltage divider on its output, to control the amount of voltage necessary to cause latchup of the parasitic SCR. An NMOS device is disposed between the output of the trigger device and the SCR.

[0018] U.S. Pat. No. 5,400,202 (Metz et al.) is similar to U.S. Pat. No. 5,452,171 above in the use of an inverter trigger device, an NMOS transistor and the parasitic SCR, but does not appear to use the voltage divider.

[0019] It should be noted that none of the above-cited examples of the related art have reduced the snapback voltage below 4 Volt. The snapback voltage of a LVTSCR with a short-channel NMOS (or PMOS) of 4 Volt is still too high to protect the internal circuit of 0.18 micron processes, because the oxide thicknesses are in the order of 32 Ångstrom. The related art does not address protection for oxides that thin. What is needed is an SCR with an trigger voltage in the range of 1.2 Volt. This voltage is low enough to prevent internal device damage by ESD pulses. Instead of the breakdown of an NMOS (or PMOS) device utilizing the breakdown of n-well to p-substrate (or p-well) will be proposed.

[0020] It is an object of the present invention to present circuits and methods which protect CMOS semiconductor devices, having oxides as thin as 32 Ångstrom, from electrostatic discharge (ESD).

[0021] Another object of the present invention is to utilize a parasitic silicon controlled rectifier (SCR), intrinsic to the semiconductor device, for the protection.

[0022] A further object of the present invention is to provide low voltage triggering of the parasitic SCR in the order of 1.2 Volt.

[0023] These objects have been achieved by providing a displacement current trigger which causes components of the SCR (parasitic npn and pnp bipolar transistors) to conduct, i.e., to trigger the SCR. The displacement current is realized by a junction capacitance, which is connected on one side to the pad to be protected and on the other side to terminals of the aforementioned parasitic bipolar transistors. Two ways of realizing the junction capacitance are disclosed.

[0024] FIG. 1 is a cross-sectional view of CMOS transistors of the prior art with parasitic bipolar transistors shown schematically.

[0025] FIG. 2 is an equivalent circuit diagram of the parasitic bipolar transistors and an NMOS transistor of FIG. 1.

[0026] FIG. 3 is an equivalent circuit diagram of the parasitic bipolar transistors similar to FIG. 2, and showing the capacitor of the present invention

[0027] FIG. 4 is a cross-sectional view of CMOS transistors utilizing an n-well of a first preferred embodiment of the present invention showing the components of FIG. 3 and equivalent junction diodes.

[0028] FIG. 5 is a cross-sectional view of CMOS transistors utilizing an n-well and a p-well of a second preferred embodiment of the present invention.

[0029] FIG. 6 is a cross-sectional view of CMOS transistors and using a polysilicon capacitor of a third preferred embodiment of the present invention.

[0030] FIG. 7 is a graph of a latch-up test for a typical SCR.

[0031] FIG. 8a-bare equivalent circuit diagrams of a Human Body Model Electrostatic Discharge Tester (HBM ESD Tester).

[0032] FIGS. 9a-care graphs of the HBM Real-time I-V Characteristic of an NMOS device for three different ESD voltages.

[0033] FIG. 10 is a cross-sectional view of the CMOS structures utilized to perform the HBM ESD Test.

[0034] FIG. 11 is an equivalent circuit diagram of the CMOS structures of FIG. 10.

[0035] To avoid the high NMOS snapback voltage of low voltage trigger silicon controlled rectifier electrostatic discharge (LVTSCR ESD) protection devices of the prior art, a different circuit and method are proposed. The NMOS snapback voltage, which is greater than 4 Volt, would damage internal MOS structures for a 0.18 micron fabrication process, having an oxide thickness as low as 32 Ångstrom. The invention introduces a junction capacitance to initiate the triggering of the silicon controlled rectifier (SCR).

[0036] Briefly, when an electrostatic discharge (ESD) occurs, the displacement current through that junction capacitance produces a current flow which turns on the two parasitic bipolar transistors inherent in a CMOS structure and which together form an SCR. By use of that junction capacitance the SCR typically triggers near 1.2 Volt, low enough to protect those internal MOS structures with an oxide thickness equal to 32 Å or more. The trigger voltage (or operating range) can vary between 1 and 12 Volt depending on variations in the device parameters and zapping voltage. The junction capacitance C_jis formed by an n⁺ contact region to a p-substrate (or p-well) and provides the displacement current to trigger the SCR and short dangerous ESD voltages to ground (or reference voltage). Because of the low trigger voltage internal device damage by an ESD pulse is prevented.

[0037] When the NMOS transistor Q1 of FIG. 2 is removed and when, as shown in FIG. 3, capacitor 26 (with capacitance C_j) is added between pad 19 and point A, the SCR comprised of pnp and npn parasitic transistors 21 and 22 is triggered by the ESD induced surge current and passes that current safely to ground. Note that like numerals in FIG. 2 and 3 designate the same component. The function and arrangement of components 19 to 24 is the same as described earlier for FIG. 2.

[0038] We describe now in FIG. 4 a preferred circuit and method for a structure which produces such a displacement current triggered SCR:

[0039] 1) Forming an n-well 12 in the p-substrate 11 of a semiconductor wafer 10 and creating a p-channel transistor with a p⁺ source 14 and a p⁺ drain (the latter not shown) in n-well 12.

[0040] 2) Creating an n-channel transistor with an n⁺ source 16 and an n⁺ drain (the latter not shown) in p-substrate 11.

[0041] 3) Creating a p⁺ contact region 18 in p-substrate 11 and connecting p⁺ contact region 18 and n⁺ source 16 to a reference voltage 20.

[0042] 4) Forming a first n⁺ contact region 13 in n-well 12 and a second n⁺ contact region 25 in p-substrate 11.

[0043] 5) Forming a junction capacitor 26 between n⁺ contact region 25 and p-substrate 11.

[0044] 6) Connecting p⁺ source 14 and n⁺ contact regions 13 and 25 to a voltage supply pad 19.

[0045] The steps that produce the above CMOS structure also create parasitic bipolar pnp transistor 21 between p⁺ source 14 (emitter), n-well 12 (base), and p-substrate 11 (collector), and parasitic bipolar npn transistor 22 between n⁺ source 16 (emitter), p-substrate 11 (base), and n-well 12 (collector). The base of transistor 21 is connected via n-well resistor 23 to n⁺ contact region 13, and the base of transistor 22 is connected via p-substrate resistor 24 to p⁺ contact region 18. The base of one transistor is connected to the collector of the other transistor. The junction between resistor 24 and the base of transistor 22 is labeled Point A. Resistors 23 and 24 are equivalent resistors for the intrinsic, and parasitic, resistance of the n-well and p-substrate (or p-well) material.

[0046] When AC current, caused by ESD and flowing through capacitor 26, produces a voltage drop across resistor 24 (Point A) equal to or larger than 0.7 Volt, then the base-emitter junction of npn transistor 22 becomes forward biased acting as an equivalent junction diode (diode 28 in FIG. 4), electrons flow into n-well 12 and are collected by n⁺ contact region 13. This causes npn transistor 22 to turn on. If the electron current is large enough, then the emitter-base junction of pnp transistor 21 becomes forward biased and acts as an equivalent junction diode (diode 29 in the n-well in FIG. 4). Current (holes) will flow through the n-well, into the p-substrate, to the collector of npn transistor 22 and collect at p⁺ contact region 18. Therefore the pnp transistor 21 turns on, i.e., the SCR fires and the surge current from the ESD is safely conducted from pad 19 to ground 20. The triggering of the SCR can be viewed as the breakdown between n-well 12 and p-substrate 11 (or p-well 30), analogous to the breakdown of the NMOS (or PMOS) device of the prior art.

[0047] A second preferred embodiment of the present invention is depicted in FIG. 5, where elements 16, 18, and 25 are placed in a p-well 30. The arrangement of the parasitic transistors, equivalent junction diodes, and n-well resistors stays the same, except that the p-substrate resistor is replaced by a p-well resistor of the same value. A third preferred embodiment is shown in FIG. 6 and replaces n⁺ contact region 25 with a polysilicon capacitor 27 having the same capacitance C_jas capacitor 26 of n⁺ contact region 25.

[0048] FIG. 7 is a graph of the latch-up test for a typical SCR (pnpn), where Curve 71 depicts the rise of the trigger voltage versus time. When the maximum voltage is reached at Point B (exceeding 20 Volt) and time 0 the SCR fires, the curve drops steeply and the SCR enters latch-up mode, holding at about 2 Volt. A trigger voltage in excess of 20 Volt is too high to protect the typical MOS structure.

[0049] In FIG. 8awe show the circuit equivalent of a human body model for an electrostatic (HBM ESD) tester. This circuit allows the modeling of electrostatic discharge energy as represented by capacitor 81 with a capacitance of C₀as it affects a device 85, the test subject. Part of the model is inductor 82 (with inductance L₀) in series with a resistor 83 (with resistance R₀) driven by a current I₁. Capacitor 84 (with test-board capacitance C_t) models the decay time of the voltage as seen across device 85, i.e., the model of the SCR.

[0050] FIG. 8bis the same circuit as that of FIG. 8a, but shows the invention's circuit equivalent before the SCR turns on. The impressed voltage V with current I₁causes the circuit to ring since it contains inductive and capacitive components and, thus, models an electrostatic discharge. Current I₁splits into two branches with current I₂flowing through capacitor 84 and current I_subflowing through capacitor 86 (with capacitance C_jmodeling the junction capacitance of capacitor 26) and resistor 87 (with resistance R_sub) modeling the p-substrate (or p-well) resistance of resistor 24.

[0051] The magnitude of the substrate current I_subis critical, because if it is too small the SCR will not turn on, but once the product of I_suband R_subis equal or larger than 0.7 Volt the SCR will turn on, i.e., fire. I_subis determined by the following equation: $\begin{array}{c}{I}_{\mathrm{sub}}=\frac{V_0{C}_j}{2\beta L_0\left(C_j+C_t\right)}\left[\exp \left(\left(-\alpha +\beta \right)t\right)-\exp \left(-\left(\alpha +\beta \right)t\right)\right]\\ \text{where:}\\ \alpha =\frac{R_0}{2L_0}\\ \beta =\frac{\sqrt{R_0^2-4L_0\left(\frac{1}{C_0}+\frac{1}{C_j}-\frac{C_t}{\left(C_j+C_t\right)}\right)}}{2L_0}\end{array}$

[0052] Next we show in FIGS. 9a, b, and c three graphs of HBM Real-time I-V characteristics of NMOS devices with tests performed at +50 Volt, +100 Volt, and +200 Volt of electrostatic discharge. Voltages are displayed on the left side, currents are displayed on the right side of each graph. The tests were performed using the structure of FIG. 10. Curve 91 displays the voltage of the snapback of the NMOS and Curve 92 displays the corresponding current. Arrow 93 identifies the displacement current I_sub. In this case we assume C_t≈0 (C_tis the test board capacitance). So that a constant displacement current $C\frac{dV}{dt}=I_{\mathrm{dis}}$

[0053] appears at the point where the voltage (Curve 91) rises. The current I₁is approximately equal to I_sub. At a high enough ESD voltage of 200 Volt (see FIG. 9c), I_subis proportional to the ESD voltage and follows closely that voltage as compared to the 50 Volt and 100 Volt tests (see FIG. 9aand 9b). Once the SCR/CMOS devices are built, the latch-up (trigger) voltage should be approximately 1.2 Volt, as mentioned earlier.

[0054] The test structure of FIG. 10 shows in cross section a semiconductor wafer 10 with a p-substrate 11 having a number of p⁺ regions (101 and 109), n⁺ regions (102, 104, 106, and 107), gates 103, and p-substrate resistors 111. The test voltage is applied to pad 19 and current I flows to n⁺ regions 104 and 107. These two regions 104 and 107 have a junction capacitor 105 and 108, respectively. A parasitic npn bipolar transistor 110 has its emitter connected to n⁺ region 106, its collector to n⁺ region 107, and its base connected between two p-substrate resistors 111 and the junction capacitor 108. Both n⁺ region 106 and p⁺ region 109 are connected to V_ssor ground 20.

[0055] The cross-sectional diagram of FIG. 10 translates into the equivalent structural diagram of FIG. 11. FIG. 11 shows the preferred embodiment of FIG. 5, but adds the parasitic p-well resistor 24 (R_pw) between junction capacitor 26 and p⁺ contact region 18, and the equivalent junction diode 28 between junction capacitor 26 and n⁺ source 16. Note that like numerals in FIG. 5 and 11 designate the same component. The electrostatic discharge is applied to pad 19 and current I_subflows to p⁺ contact region 18 via capacitor 26 and p-well resistor 24. The voltage drop across resistor 24 forward biases the equivalent junction diode 28 allowing the SCR (not shown) to fire. The circuit diagram for FIG. 11 is identical to the circuit diagram of FIG. 3.

[0056] The present invention has the advantage that:

[0057] a) the trigger voltage of the parasitic SCR is about 1.2 Volt and, therefore, low enough to prevent internal device damage to oxides that are as thin as 32 Ångstrom for devices built using a 0.18 micron process, and

[0058] b) implementation of the junction capacitor is simple and does not consume valuable silicon real estate.

[0059] While the invention has been particularly shown and described with reference to the preferred embodiments thereof, it will be understood by those skilled in the art that various changes in form and details may be made without departing from the spirit and scope of the invention.