* * *

Field of the invention The present description relates to techniques for verifying the reliability of electronic devices.

Specifically, the description has been developed with particular attention paid to verification of the robustness of electronic devices in regard to electrical stresses induced by system transients that can arise during operation of the device.

Description of the prior art

Electrical OverStresses (EOSs) are a constant cause of failure for integrated circuits. According to certain reports, approximately 40% of the failures found in integrated circuits can be attributed to EOS phenomena or events.

In particular, for semiconductor devices, EOS events comprise a wide range of stresses of an electrical nature, due, for example, to Electro Magnetic Pulses (EMPs), system transients of various nature (the so-called "overvoltage spikes" on the supply lines and input/output lines) and Electrostatic Discharge (ESD) . In particular, events belonging to the latter type of EOS present with durations comprised between 1 ns and 1 μs and prevalently in the production and handling stages, i.e., when the device is not subject to electrical biasing. The failures linked to EOS events in semiconductor devices can be classified according to the respective primary failure mechanism: failures induced as a result of thermal or electrical effects (phenomena of thermo- migration or electro-migration of material, which involve the metallization) , phenomena of latch-up, breakdown of the gate oxide, and other failures that can be correlated to electrical fields.

The EOS events considered herein belong to the family of system transients (on the supply lines and input/output lines) .

For said events, the sensitivity as a function of temperature depends upon the failure mode considered. For example, it is minimal for (thermo-electrical) damage of the metallization and very serious for breakdown of the oxides.

These particular events can be defined as over- voltage or over-current phenomena, which, in temporal terms, have a duration of between 1 μs and 1 ms and which are present during applicational operation of the device.

Document "Automation of electrical overstress characterization for semiconductor devices", Diaz CH. , Hewlett-Packard Journal, Hewlett-Packard Co. Palo Alto CA, US, vol. 45 no. 5, pages 106-111, describes an automatic test system developed to characterize semiconductor devices and interconnect failures caused by electrical overstress. Electrical stress in the form of current pulses of increasing amplitude is applied to a device under test in the testing mode. The test system was developed for monitoring EOS robustness in advanced CMOS processors under test.

Document "EOS/EDS; ADI Reliability Handbook", 2000 Analog Devices Inc. Norwood, MA 02062, pages 1-22, describes Human Body Model and Machine Model test methods that simulates the application of EOS on ICs.

Document US 2007/0018670 Al describes a system and method for electrostatic discharge (ESD) testing. The system includes a circuit that has a switch coupled to an input/output (I/O) circuit of a device under test (DUT) , a charge source coupled to the switch, and a control circuit coupled to the switch, wherein the control circuit turns on the switch to discharge an ESD current from the charge source to the I/O circuit, and wherein the circuit is integrated into the DUT. The system provides on-chip ESD testing of a DUT without requiring expensive and specialized test equipment.

Document "Improvement of degradation detection in ESD test for semiconductor device", Satoh S ED - Institute of Electrical and Electronics Engineers, conference record of the 2002 IEEE Industry applications conference 37 ^th IAS annual meeting, Pittsburgh, PA, October 13-18, 2002 describes ESD test as evaluation tests for semiconductor products. In particular, the most popular and standardized test is the Human Body Model ESD test. In HBM ESD test some issues are experienced in the detection of device degradation as pass/fail judgment after ESD stress applied. Some kind of device degradation cannot be detected with the conventional test condition and criteria. Improved judgment test using effective combination of optimized DC leak test and device functional test was proposed.

Document US 2005/017745 Al describes an electrostatic withstand voltage test method that enables semiconductor integrated circuit testing to be performed with a high degree of precision and at low cost. In this method, with one of ground pins VSS and VSSI of a semiconductor integrated circuit grounded, static electricity is applied from a static electricity discharge apparatus to all pins of semiconductor integrated circuit, after which, with power supply apparatus connected to power supply pin VDD of semiconductor integrated circuit and the other grounded, a leakage current test apparatus is connected to all signal pins and pin leakage current is tested, and with ground pin VSSI of the internal circuitry of semiconductor integrated circuit grounded and leakage current test apparatus connected to power supply pin VDDI, a pattern generator that supplies a digital signal is connected to signal input pins (IN, I/O), and power supply leakage current is tested.

Document US 5,132,612 describes an apparatus for applying high current fast rise time pulses simulating electrostatic discharge (ESD) to various combinations of pins of a device under test (e.g., a microcircuit) . The apparatus also provides for testing of the DUT after the performance of ESD stress testing. The apparatus establishes electrical connections between the terminals of a high voltage pulse generator (HVPG) and several different combinations of the DUT pins in sequence in order to apply ESD stresses. The apparatus further provides functional parameter tests whether the connection to the DUT pins during ESD stressing has caused the DUT to fail. All these previous documents describe exposing to electrical overstress an electronic circuit being subjected to testing, i.e. when the electronic circuit is in the testing condition or testing modality.

Object and summary of the invention

On the basis of the state of the art outlined above, there emerges the need to have available solutions such as to enable evaluation of the robustness/sensitivity of the electronic devices, such as integrated circuits, in regard to system transients (EOS) in normal applicational conditions of operation.

This need is felt, in particular, for devices obtained using BCD (Bipolar CMOS DMOS) technology; these devices are more sensitive to said events as compared, for example, to devices of a system-on-chip (SOC) type in view of the different electro-applicative context of operation. Normally, in fact, BCD devices directly interface inductive loads, which notoriously generate spikes. The object of the present invention is to provide a solution capable of meeting the needs outlined previously.

According to the invention, the above object is achieved thanks to a method presenting the characteristics recalled specifically in the ensuing claims. The invention also relates to a corresponding device .

The claims form an integral part of the technical teaching provided herein in relation to the invention. An embodiment of the solution described herein envisages that a generator of stresses of a system transient type (EOS) is able to apply positive and negative EOS system transient stresses on all the input-output (I/O) terminals of the Device Under Test (DUT), including the supply lines both in static and in dynamic applicational conditions of operation.

An embodiment of the solution described herein is able to meet the need of verification outlined previously by taking into account the applicational conditions of the device (both static and dynamic) and to perform an action of verification in regard to system transients of an EOS type applied:

- in dynamic conditions, i.e., with devices under test (DUTs) fully operative, with the possibility of control in current of ohmic-inductive loads; in static conditions, i.e., with electrically biased devices in a wait state or in standby not traversed by considerable currents.

In one embodiment, the solution described herein is able to operate in dynamic conditions (i.e., with the device functioning) , for example in HTOL (High- Temperature Operating Life) conditions, with the alternative application of positive stresses to just the terminals of the supply lines.

Brief description of the annexed drawings

The invention will now be described, purely by way of non-limiting example, with reference to the annexed drawings, in which: - Figure 1, comprising two parts designated respectively by a) and b) , illustrates a first possible embodiment of the solution described herein;

- Figure 2 illustrates a second embodiment of the solution described herein; - Figure 3 provides further details on the embodiment of Figure 2; and

Figure 4 illustrates the possible plot of waveforms that can be applied to the devices under test using the solution described herein.

Detailed description of embodiments

In the ensuing description, various specific items are illustrated aimed at providing a deep understanding of the embodiments. The embodiments can be obtained without one or more of the specific items, or with other methods, components, materials, etc. In other cases, known structures, materials or operations are not illustrated or described in detail in order not to obscure the various aspects of the embodiments. The reference to "one embodiment" in the scope of this description is intended to indicate that a particular configuration, structure or characteristic described in relation to the embodiment is comprised in at least one embodiment. Hence, phrases such as "in one embodiment", possibly present in different points of this description, do not necessarily refer to the same embodiment. Furthermore, particular conformations, structures or characteristics can be combined in any adequate way in one or more embodiments. The references used herein are only for reasons of convenience and hence do not define the sphere of protection or the scope of the embodiments.

Figure 1, which comprises two parts designated, respectively, by a) and b) represents, respectively in plan view and in a vertical median section, a mother board 10 that can be used for conducting reliability tests of an HTOL type on one or more electronic devices under test (DUTs) .

The example of embodiment illustrated herein refers to a mother board 10 that can be used for testing two devices alternatively, designated by DUT A and DUT B.

Also visible in Figure 1 is the OMI (OLT Module Interface) that performs the function of dialogue with the testing apparatus of which the mother board 10 forms part.

Apparatuses of this type are in themselves known in the sector of product qualification, thus rendering superfluous any detailed description herein. In the specific case of Figure 1, the device for conducting EOS tests comprises a daughter board 12 that can be coupled to the mother board 10. Mounted on the daughter board 12 is a set of circuits 14, designed to generate pulses that simulate EOS system transient events according to the modalities described more fully in what follows.

The circuits 14 are configured typically as oscillator circuits that can generate pulses having the characteristics described more fully in what follows. As may be seen more clearly from part b) of Figure 1, in the example of embodiment illustrated herein the daughter board 12 is coupled to the mother board 10 on the face opposite to the face on which the devices under test DUT A and DUT B and the interface OMI are mounted.

This condition of assembly enables a dual advantage to be achieved. In the first place, no appreciable modification of the mother board 10 is necessary; moreover, the mother board 10 and the EOS daughter board 12 can be located directly in the testing apparatus.

Figure 2 illustrates schematically the possibility of configuring the EOS test board as a stand-alone board within a dedicated apparatus provided with a monitor 18 that can be managed by a personal computer integrated in the system.

Figure 2 highlights the presence of the supply sources 20, with the consequent possibility of applying EOS events also to said supply sources. Figure 3 highlights the possibility of using the solution described herein within a testing apparatus 100, constituted, for example, by the test vehicle device used.

Highlighted schematically in Figure 3 is the presence of the device under test, to which - in a normal test cycle - "stimuli" of various nature are applied necessary for the operation of the DUT.

The monitor 18 is used as terminal of the personal computer (programming of the stimulation pattern and data collection) .

In the embodiment illustrated herein, there is moreover envisaged the possibility of intervening on the supply lines 20 (for example, 12 V, 5 V, 3.3 V nominal) applying thereto - in dynamic conditions, i.e., whilst the DUT is being tested in operating conditions - system transients (EOSs) . Added to this is the possibility of observing/recording on the monitor 18 the outcomes produced by said events on the device.

The evaluation of the robustness of the DUT will be obtained following upon a definitive parametric control by means of ATE (Automatic Testing Equipment) at the end of the EOS stimulations.

The solution described herein hence enables the DUT (typically an integrated circuit) to be subjected to stresses of an EOS system transient type during normal operation (i.e., while the circuits are handling currents), i.e., while the mother board is operating in a normal HTOL configuration.

The EOS events can be applied in a way synchronous or asynchronous with respect to the pattern of stimulation of the DUT 22, with the added possibility of programming both the amplitude and the pulse widths.

For example, the diagram of Figure 4 shows the possible plot of a pulse train (bottom right part of Figure 4), the characteristics of which can be partially set, in particular as regards the following characteristics :

- absolute amplitude a - settable;

- rise time rt - in itself not settable, but a consequence of the amplitude a;

- pulse width pw - settable;

- fall time ft - in itself not settable, but a consequence of the amplitude a;

- overshoot voltage (negative) uv - in itself not settable, but a consequence of the amplitude a.

For example, the amplitude a can reach a value of 27 V with a rise time (rt) and fall time (ft) having a typical value of 5 μs .

Furthermore, for example, the pulse width pw is programmable from 10 μs to 470 μs, and the negative voltage uv can reach a value of 10% of the maximum value of the pulse set. The pulse-repetition time can be set with a minimum of approximately 1 s.

The aforesaid values are of course to be understood purely as an example and are not to be read in a sense that might in any way limit the scope of the present description.

Similar considerations apply to the values exemplified in the tables reproduced below.

The ensuing table makes reference to EOS events that can be applied on three DC voltages Vl, V2, and V3 - representing three supply voltages (for example, the voltages 20 of Figure 3 ) .

Voltage range Min. Max.

DC voltage Vl 5 V 20 V

(min . spike 7 V) (max . spike 7 V)

(max . spike 22 V) (min . spike 0 V)

DC voltage V2 3 V 8 V

(min . spike 3 V) (max . spike 2 V)

(max . spike 7 V) (min . spike 0 V)

DC voltage V3 3 V 8 V

(min . spike 1 V) (max . spike 2 V)

(max . spike 7 V) (min . spike 0 V)

The next table relates to programmable spikes that can be superimposed on the aforesaid voltages Vl, V2, V3.

During operation as generator of EOSs of a static type (i.e., for verifying the behaviour of a device with an integrated circuit when exposed to sudden spikes applied on the supply lines) , the solution described herein envisages that the circuit under test, or DUT, is biased without being traversed by the currents that characterize operation thereof with the application of spikes one at a time. During static operation, the device described herein enables production of EOS system-transient events in an asynchronous way, there not being any stimulation pattern.

The EOS events can be applied on as many as three supply lines and/or on all the inputs and outputs (both positive and negative) . Both the amplitude and the duration of the spikes are programmable (in a single or multiple way) .

In a possible experimental configuration, the embodiment to which Figure 3 refers has been used for applying EOS events to an integrated circuit with two supply lines Vl and V2 with nominal values equal to 13.2 and 5.5 V, respectively.

The solution described herein enables validation, from the point of view of the designer, of the circuit protection adopted in the integrated circuit DUT in regard to the EOS system transients, with the possibility of seeking the limits or margins of application with respect to said EOS events. The solution described herein can also be included in a program of control of production in real time (RTC - Real Time Control) with the possibility of intervening during verification on the production lots of the product itself. Embodiments of the solution described herein envisage the implementation of the stresses in static conditions, i.e., with the device biased, without the latter being traversed by the currents that characterize operation thereof, for example in HTRB (High-Temperature Reverse Bias) conditions, with application of positive and/or negative stresses on all the I/O terminals and on the terminals of the supply lines .

Embodiments of the solution described herein envisage implementation of the stress in dynamic conditions (i.e., with the device functioning), for example in HTOL conditions, with application of the positive and/or negative stresses both on the terminals of the supply lines and on the I/O terminals, alternatively.

Embodiments of the solution described herein can involve extension of the number of supplies of the DUT on which to apply the EOS system transients, one at a time, and/or increase in the maximum nominal value of the voltage of the DUT, for example, to +60 V.

For example, the maximum amplitude of the positive voltage applied to the DUT can be +100 V (maximum nominal value +60 V plus EOS spike) , with a maximum amplitude of the negative voltage applied to the DUT of -60 V (nominal value minus EOS spike) .

Other possible values are the following:

- max. under-voltage : 10%;

- typical rise and fall slew rate: 2 μs;

- time durations of the pulses that can be applied to the supply voltages: from a minimum of 5 μs to a maximum of 200 μs; and

- time durations of the pulses that can be applied to the I/O terminals: from a minimum of 5 μs to a maximum of 500 μs . In one embodiment, the electrical circuits for stimulation of the EOS system-transient events are integrated in a specific board and are independent of the type of the DUT.

In one embodiment, the specific configuration of HTOL mode of the DUT can be integrated in a mother board dedicated to the functionality of the DUT.

In one embodiment, the EOS system transients can be applied to the DUT, both to all the I/O terminals and to the supply lines in an alternative way in both static configuration and dynamic configuration, as described previously.

One embodiment uses a personal computer for acquisition and management of the data of the DUT, detecting operation thereof in real time. It follows that, without prejudice to the principle of the invention, the details of construction and the embodiments may vary, even significantly, with respect to what is described and illustrated herein purely by way of non-limiting example, without thereby departing from the scope of the invention, as defined by the annexed claims.