By Steve Ching and Manor Narayanan, Isolink, a subsidiary of Skyworks Solutions, Inc.
One of the first optocouplers used for high reliability applications was the simple phototransistor optocoupler. The basic devices were the 4N22, 23, 24 and the 4N47, 48, 49 series of hermetic phototransistor optocouplers in TO‐five type transistor cans made for military applications developed in the early 1970s. However, these devices were never designed specifically for space radiation environments.
Isolink’s involvement in the use of optocouplers for space applications came out of the work that was done in the military and medical use of optoelectronic devices in the late 1980s and early 1990s. In the military applications, most of the information on the radiation performance of various optoelectronic devices was initially obtained in the development of radiation-tolerant optoelectronic devices for radar, missiles, B1 and B2 bombers, M1A2 tanks, etc. The main types of radiation requirements were gamma total dose, dose rate, and neutron radiation. In the medical area, optoelectronic devices were developed for the control of high energy X-ray linear accelerators for cancer treatment. Based on this initial information, several types of LEDs and detectors were identified to be more radiation tolerant than others. For example, the amphoterically Si-doped 940nm GaAs liquid phase epi (LPE) LED cannot withstand displacement damage due to neutron particles or high energy X-rays. Other LED compositions such as double heterojunction (DH) AlGaAs LPE LEDs in various wavelengths (660nm, 700nm, 830nm, 870nm) are much more displacement damage tolerant.
In late 1996, Isolink was contacted by Boeing concerning the availability of proton radiation-tolerant phototransistor optocouplers for use in the International Space Station. This was the result of the premature failure of 4N49 type optocouplers in the TOPEX/Poseidon satellite and subsequent confirmation of displacement damage of the optocouplers.1 Four types of phototransistor optocouplers were prepared using various DH AlGaAs LEDs and phototransistor detector combinations that met all the electrical specification of the 4N49. All devices were tested for proton displacement damage radiation (63MeV) by Boeing Radiation Effects Laboratory (BREL) at the Crocker Nuclear Laboratory of UC Davis. All four types of optocouplers showed much better degradation performance than the 4N49 that was designed in the Space Station.2 It also confirmed that LEDs that performed well in neutron radiation displacement damage also performed well in proton radiation. The OLX249 was standardized for space applications given the higher CTR margin, simple construction, and better control on the sources of the dice. Since then, over 100,000 units of OLX249 phototransistor optocouplers were provided and made specifically for space environments for customers all over the world.
This article discusses the follow-up development of a high-gain phototransistor optocoupler, OLX449, based on the OLX249, which provides even more design margins for space applications. The OLX449 uses the same parts as the OLX249. The only difference is the more efficient LED.
Phototransistor Optocoupler Transfer Function
The transfer function of the phototransistor optocoupler can be simplified into the following equation:
IC = Ie n hFE
IC is the collector output current of the optocoupler. For optocouplers, the term current transfer ratio (CTR) is often used. It is the ratio of output current to the LED input current (IC/IF). Ie is the radiant intensity of the LED. n is the coupling factor between the LED and the phototransistor. hFE is the phototransistor gain. For simplification, the terms Ie n can be lumped into a measureable parameter as ICB, the photocurrent generated by the phototransistor collector/base junction when the LED photons are coupled through the optical medium and picked up by the collector/base junction.
IC = ICB x hFE
In the case of a phototransistor, the collector/base junction is made deliberately large to be the light-sensitive detector. Therefore, by measuring the ICB, the LED and coupling performance can be easily quantified. From past radiation test data, no significant change was found in the transparency of the silicone gel coupling material due to radiation. Therefore, any change in ICB is largely due to the change in the LED power output.
Figure 1 shows the ICB (photocurrent) vs. LED current IF comparison between the OLX249 and OLX449. The OLX449 ICB is about three times higher across the entire LED current range, and is about 10 times higher than the 4N49 ICB.
Figure 1: ICB (photocurrent) vs. IF, OLX249 and OLX449
Phototransistor Current Gain vs. ICB vs. Radiation
The graph of the OLX249 phototransistor current gain (hFE) versus photocurrent (ICB) after several levels of total dose is shown in Figure 2.
The typical behavior of bipolar transistor gain after radiation exposure has been well documented in various publications. [3, 4] In high injection regions (high IC or IB), the ß degradation will decrease less after radiation exposure. In low injection regions (low IC or IB), the ß degradation will decrease more rapidly. This is part of the reason that phototransistor CTR degrades less with higher LED currents. At higher LED currents, the photocurrent is higher and the phototransistor is operating in the higher injection region. However, higher LED drive current is usually a luxury in space application. Therefore, by changing to more efficient LEDs, the photocurrent will increase at the same LED drive current and will result in less CTR degradation and more remaining CTR after radiation.
Figure 2: ß curve vs. photocurrent (IB) after gamma total dose radiation
For example, for a typical 4N49 the photocurrent (IB) at IF=1mA is about 6µA. Before the total dose radiation, the phototransistor ß is about 400, resulting in an IC of 2.4mA, or a CTR of 240 percent. As the optocoupler receives more radiation, the phototransistor ß decreases rapidly, from 400 to about 180 at 95K rads. Even though the LED is quite immune to gamma total dose radiation, the decrease in ß will result in a lower IC of 1.08mA, or a CTR of 108 percent after 95K rads. That is a decrease of 55 percent in IC.
For the OLX249 with the more efficient DH AlGaAs LED, the photocurrent (IB) is about 20µA at IF=1mA. The corresponding ß is about 390, resulting in an IC of 7.8mA or a CTR of 780 percent before radiation. After 95K rads of gamma radiation, the ß is decreased to 220 for a post radiation IC of 4.4mA, or a CTR of 440 percent. That is a decrease of about 44 percent in IC. This is less than the CTR degradation of the 4N49 and with a higher remaining CTR for more design margin.
For the OLX449, an even more efficient DH AlGaAs LED than that of the OLX249 is used. All other parts of the optocoupler remain the same. This more efficient LED generates about three times more photocurrent than the OLX249 LED. At 1mA IF, the photocurrent is about 65µA with a corresponding ß of 360 for an IC of 23.4mA, or a CTR of 2340 percent. After 95K rads of gamma radiation, the ß is decreased to 220 for a post radiation IC of 14.30mA. The IC decreased by about 39 percent. This degradation is less than that of the OLX249 and with a much higher remaining CTR for more design margin.
In the case of proton radiation, where the displacement damage will reduce both the LED light output and the ß of the phototransistor, the nature of the ß curve rolloff at low IB will further exacerbate the degradation of IC or CTR (see Figure 3). Since the LEDs for the OLX249 and OLX449 have very similar behavior in proton radiation and the phototransistor detector is the same, one can extrapolate the OLX449 proton radiation performance by using the data from OLX249 but with higher LED current. By using the more efficient LEDs, the degradation of CTR will be reduced and contain more CTR design margin after proton radiation exposure.
Figure 3: Proton radiation performance comparison between Isolink phototransistor optocouplers
Saturated CTR vs. Linear CTR
The phototransistor optocoupler is commonly used in two ways in a circuit, with an emitter follower resistor or with a pull-up resistor (see Figures 4a and 4b).
Figure 4: Common phototransistor optocoupler circuit
With the emitter follower resistor mode, the optocoupler is operating in the linear mode, with Vce in the range of 1V and higher. With the pull-up resistor, the optocoupler is operating in the saturated mode and the Vce (sat) is usually less than 0.4V for logic switching. Due to the normal bipolar transistor Vce curve characteristics, the CTRs at different Vce of the optocoupler are different. Figure 5 shows the differences in the CTRs of the OLX449 at Vce=5V and Vce=0.4V. The implication for space applications where the optocoupler is used for saturated switching is that CTR (sat) must be specified and radiation data collected at that condition. CTR (sat) is always lower than CTR measured at higher Vce. The degradation of CTR (sat) due to radiation is usually lower than CTR at higher Vce.
Figure 5: CTR and CTR (sat) vs. IF, OLX449
Advances in LED fabrication technology paves the way for a designer to advantage of higher power LEDs coupled with a phototransistor to deliver high CTR values. This would allow for more performance design margins for radiation degradation without any major changes in the device and circuit design. In other words, the designer can get the same radiation performance as the present radiation-tolerant optocoupler, but at a reduced LED drive current.
 G. Swift, B. Rax, C. Barnes, A Johnston, "TOPEX/Poseidon Radiation Issues: Displacement Damage in Octocouplers", JPL Publication, September 15, 1997.
 D. Oberg, D. Egelkrout, "Flight 2A Optoelectronic Device Proton/Neutron Degradation Test Report", Boeing Radiation Effects Laboratory, March 18, 1997.
 M. Dentan, "Introduction to Radiation Effects on Electronic Components and Circuits", EFDA JET CSU, March 23, 2006, p. 50.
 A. Johnston, B.G. Rax, "Proton Damage in Linear and Digital Optocouplers", IEEE Trans. Nucl. Sci., 47(3), 675(2000).