Junction temperature, short fortransistor junction temperature,[1] is the temperature of the actualsemiconductor in an electronic device. In operation, it is higher than case temperature and the temperature of the part's exterior. The difference is equal to the amount of heat transferred from the junction to case multiplied by the junction-to-casethermal resistance.
Various physical properties of semiconductor materials are temperature dependent. These include the diffusion rate ofdopant elements,carrier mobilities and the thermal production ofcharge carriers. At the low end, sensor diode noise can be reduced bycryogenic cooling. On the high end, the resulting increase in local power dissipation can lead tothermal runaway that may cause transient or permanent device failure.
Maximum junction temperature (sometimes abbreviatedTJMax) is specified in a part's datasheet and is used when calculating the necessary case-to-ambient thermal resistance for a given power dissipation. This in turn is used to select an appropriateheat sink if applicable. Other cooling methods includethermoelectric cooling andcoolants.
In modern processors from manufacturer such asIntel,AMD,Qualcomm, the core temperature is measured by a network of sensors. Every time the temperature sensing network determines that a rise above the specified junction temperature (), is imminent, measures such asclock gating, clock stretching, clock speed reduction and others (commonly referred to as thermal throttling) are applied to prevent the temperature to raise further. If the applied mechanisms are not compensating enough for the processor to stay below the junction temperature, the device may shut down to prevent permanent damage.[2]
An estimation of the chip-junction temperature can be obtained from the following equation:[3]
where:=ambient temperature for the package [°C]
= junction to ambient thermal resistance [°C / W]
= power dissipation in package [W]
Many semiconductors and their surrounding optics are small, making it difficult to measure junction temperature with direct methods such asthermocouples andinfrared cameras.
Junction temperature may be measured indirectly using the device's inherent voltage/temperature dependency characteristic. Combined with aJoint Electron Device Engineering Council (JEDEC) technique such as JESD 51-1 and JESD 51-51, this method will produce accurate measurements. However, this measurement technique is difficult to implement in multi-LEDseries circuits due to high common mode voltages and the need for fast, highduty cycle current pulses. This difficulty can be overcome by combining high-speed sampling digital multimeters and fast high-compliance pulsedcurrent sources.[4]
Once junction temperature is known, another important parameter,thermal resistance (Rθ), may be calculated using the following equation:
AnLED orlaser diode’s junction temperature (Tj) is a primary determinate for long-term reliability; it also is a key factor forphotometry. For example, a typical white LED output declines 20% for a 50 °C rise in junction temperature. Because of this temperature sensitivity, LED measurement standards, likeIESNA’sLM-85Archived 2017-10-18 at theWayback Machine, require that the junction temperature is determined when making photometric measurements.[5]
Junction heating can be minimized in these devices by using the Continuous Pulse Test Method specified in LM-85. An L-I sweep conducted with anOsram Yellow LED shows that Single Pulse Test Method measurements yield a 25% drop inluminous flux output andDC Test Method measurements yield a 70% drop.[6]