Quality and Reliability Approach
In the current semiconductors market, two key factors for a company's success are product quality and reliability. These are interrelated: reliability represents long-term quality performance over a product's expected lifetime. Manufacturers must ensure products meet or exceed baseline quality and reliability standards. onsemi supplies products for demanding applications intended for harsh environments and subject to high quality and reliability requirements.
Rather than relying solely on "test to ensure quality," the preferred approach is "design to ensure quality." onsemi applies a dual approach to achieve final quality and reliability levels. First, robust processes are developed and implemented. Second, process specifications are strictly followed at every step from start to finish. Inspections and procedures are established to detect potential hidden failure modes. This systematic focus on long-term reliability supports consistent product performance.
Four-step reliability framework
- Strict process control and inspections
- Thorough evaluation of design and materials
- Process average testing, including 100% QA redundant testing
- Ongoing reliability verification through audits and reliability studies
These procedures, together with incoming inspection and outgoing quality control, help ensure product quality throughout the flow from silicon to delivery.
Reliability Testing Overview
onsemi IGBT devices undergo a broad set of reliability tests to verify consistency. These tests accelerate failure mechanisms that may occur in actual applications so that satisfactory performance in real-world use can be confirmed.
Below are the routine reliability tests commonly applied to onsemi IGBT devices.
High Temperature Reverse Bias (HTRB)
The HTRB test checks stability of the main blocking junction under reverse-bias conditions at elevated temperature as a function of time. For a given junction temperature and applied voltage, stability and leakage current over time indicate junction surface stability and therefore serve as indicators of device quality and reliability.
For IGBT testing, voltage is applied between collector and emitter with the gate shorted to the emitter. DC parameters monitored include ICES, V(BR)CES, IGES, VGE(th), and VCE(on). Failure occurs when leakage current rises sufficiently to cause power dissipation that leads to thermal runaway. For stable devices, leakage current should remain relatively constant, with only slight increases during the test period.
Typical conditions: VCE = 80-100% of maximum rated value, VGE = 0 V (shorted), TA = 150 C or Tj(max). Duration: 1,000 hours to meet qualification requirements.
High Temperature Gate Bias (HTGB)
The HTGB test applies DC bias stress across the gate oxide at elevated temperature to detect drift caused by random oxide defects and ionic contamination. For IGBT testing, voltage is applied between gate and emitter with the collector shorted to the emitter. Monitored DC parameters include IGES, VGE(th), and VCE(on). Oxide defects can cause early device failure.
Typical conditions: VGE = ±20 V or 100% rated VGE, VCE = 0 V (shorted), TJ = 150 C or Tj(max). Duration: 1,000 hours to meet qualification requirements.
High Temperature Storage Life (HTSL)
HTSL assesses device stability, high-temperature endurance potential, and internal manufacturing integrity of the package. Although devices are not usually exposed to such extreme temperatures in the field, this test accelerates failure mechanisms that could occur during long-term storage at elevated temperature.
Test method: devices are placed in a mesh tray and exposed to a controlled elevated-temperature chamber as a function of time.
Typical conditions: TA = 150 C (temperature on plastic package). Duration: 1,000 hours to meet qualification requirements.
High Humidity High Temperature Reverse Bias (H3TRB)
H3TRB evaluates resistance of components and materials to combined degradation from long-term operation in high-temperature, high-humidity environments. This test is applicable to non-hermetic devices only.
Humidity has long been a factor for semiconductor degradation, especially for plastic-packaged devices. Moisture-related degradation is often caused by moisture penetration through passivation and surface corrosion. onsemi addresses this through junction passivation processes, die coatings, and appropriate package material selection.
Typical conditions: VCE = 80-100% of maximum rated value, VGE = 0 V (shorted), TA = 85 C, RH = 85%. Duration: 1,000 hours to meet qualification requirements.
Unbiased Highly Accelerated Stress Test (UHAST)
UHAST evaluates moisture resistance by exposing devices to high vapor pressure. This test is performed on plastic/epoxy packaged devices, not hermetic packages. Devices are mounted on a tray and positioned about two inches above the surface of deionized water in the test chamber to avoid liquid condensation on devices. After reaching the target temperature and pressure, conditions are held for at least 24 hours, then devices are removed and dried. Typical monitored parameters are leakage current and voltage.
Typical conditions: TA = 131 C, P = 14.7 psi, RH = 100%. Duration: 72 hours to meet qualification requirements.
Intermittent Operating Life (IOL)
IOL evaluates integrity of the die and package under cyclic on/off operation that simulates real-world duty cycles: the device heats during on-time due to power dissipation and cools during off-time when power is removed. A DC supply is applied until the desired junction temperature is reached, then power is removed and forced air cooling reduces the junction temperature to ambient.
(Formula 1)
(Formula 2) This is typically an accelerated condition.
(Formula 3) The sequence repeats for the specified number of cycles. Careful control of temperature offsets is necessary to ensure repeatable results. IOL assesses thermal fatigue of the die-attach interface between the die and the mounting surface and the die-to-leadbond interfaces.
For IGBT devices, monitored parameters include thermal resistance, threshold voltage, on-resistance, gate-emitter leakage, and collector-emitter leakage. Failure occurs when thermal fatigue causes thermal resistance or on-resistance to increase beyond the maximum specified in the data sheet.
Temperature Cycling (TC)
Temperature cycling determines resistance of the device to high and low temperature excursions in air and the effects of cycling between extremes. Tests are performed by alternately placing devices in separate hot and cold chambers. Air circulation keeps chamber temperatures uniform and chambers have sufficient thermal capacity to reach the specified ambient temperature before devices are transferred.
Each cycle includes at least 15 minutes exposure at one extreme temperature followed by immediate transfer to the other extreme for at least 15 minutes to complete one cycle. Immediate transfer imposes greater stress than non-immediate transfer.
Typical extreme conditions: -65 C / +150 C. The number of cycles is tied to the severity of the expected application environment. Industry practice often considers ten cycles sufficient to reveal gross quality issues. Temperature cycling can identify excessive strain between internal materials that have differing coefficients of thermal expansion.
Low Temperature Storage Life (LTSL)
LTSL assesses device stability, low-temperature endurance potential, and internal package integrity. Although devices are not normally exposed to such extreme low temperatures in the field, this test accelerates potential failure mechanisms that could occur during long-term storage at low temperature.
Typical conditions: TA = -65 C (temperature on plastic package). Duration: 1,000 hours to meet qualification requirements.
Test method: devices are placed in a mesh tray and exposed to a controlled low-temperature chamber as a function of time.
Steady-State Operating Life (SSOL)
SSOL evaluates die and package integrity under steady-state continuous operating conditions. For IGBT devices, monitored parameters include thermal resistance, threshold voltage, on-resistance, gate-emitter leakage, and collector-emitter leakage.
Typical conditions: VGE >= 10 V, Delta TJ = 100 C, TA = 25 C. Duration: 1,000 hours to meet qualification requirements.
Failure occurs when thermal fatigue causes thermal resistance or on-resistance to increase beyond the maximum specified in the data sheet.
Package-related Tests
- Physical dimensions: Verify conformance to package outline specifications.
- Visual and mechanical inspection: Confirm appearance and functional standards such as mark legibility and absence of contamination.
- Solvent resistance: Verify solderability of device terminals.
- Terminal strength: Lead bend tests to check lead integrity.
Each manufacturing process produces a distribution of quality and reliability. Controlling that distribution is essential to ensure a high mean, narrow spread, and consistent distribution shape. Appropriate design and process control reduce the need to rely on screening programs to eliminate the lower tail of the distribution.
Accelerated Stress Testing and Data Review
Some tests described here subject devices to conditions far beyond those encountered in normal operation. These accelerated conditions stress the underlying failure mechanisms and allow prediction of failure rates over shorter test durations. Temperature-related failure modes are characterized using the Arrhenius model.
(Formula 4) AF = acceleration factor. EA = activation energy (eV). K = Boltzmann constant (8.62×10^-5 eV/K). T2 = operating temperature, T1 = test temperature.
Equivalent device hours equal the acceleration factor (from the Arrhenius model) multiplied by the actual device hours.
Data review examples:
- HTRB is used to determine leakage current stability related to field distortion in the IGBT. HTRB enhances the failure mechanism through high-temperature reverse-bias stress, making it a good indicator of device quality and process control effectiveness.
- HTGB checks stability under accelerated high-temperature gate-bias forward conditions. This stress on the gate oxide reveals drift due to random oxide defects, which typically appear at very low defect rates in the early, random portion of the reliability "bath-tub" curve.
- IOL is an excellent accelerated stress test to determine integrity of die and package components under cyclic on/off power conditions. It is among the most important tests because it simulates typical real-world usage. IOL exercises die attach, wire bond, on-state and off-state device behavior, and validates compatibility of materials under thermal expansion. onsemi performs extensive IOL testing as part of ongoing process control monitoring and correlates detailed Delta TJ analysis. onsemi has determined that a Delta TJ of 100 C is effective for stressing devices for reliability modeling, which is well above many customer application requirements.
- Temperature cycling also imposes thermal stress from the external environment, complementing IOL which imposes internal electrical stress on the device system.
HTSL, H3TRB, thermal shock, and pressure-cooker style tests are routine. Reliability engineering practitioners consider HTRB, HTGB, IOL, and TC among the most important tests for IGBT devices.
