The term “diode” is applied to a large number of two-terminal devices with varied functions. By itself it usually references a device intended to serve as a rectifier with little further implied, though adding qualifiers can result in reference to some specific subset of rectifier diodes, or devices intended for entirely different functions, such as voltage clamping or regulation, temperature measurement, light detection, light emission, frequency control, noise generation, or other functions. Or specialized subsets of these other functions. It’s a lot to keep track of, and the purpose of this section is to offer an overview of the diode landscape and directions to resources providing information in greater detail.
Common diode characteristics and qualifications
A diode’s reverse voltage rating, also called off state voltage characterizes the maximum voltage that can be applied to a device in the reverse-biased direction without causing breakdown –a rapid increase in reverse current flow that results in significant power dissipation in the device, leading to rapid temperature rise and usually destruction of the diode if the reverse current is not externally limited. The degree to which current flow under reverse breakdown conditions are concentrated or distributed across the physical expanse of a device’s active region vary with device type and construction method, and consequently different devices display variable degrees of robustness to operation in breakdown mode; some diode types are designed specifically for use in this mode of operation.
The conditions and qualifications under which manufacturers’ ratings apply can vary significantly. One device’s figure may be intended to communicate a safe operating limit under anticipated operating conditions, e.g. at elevated device temperature, and allowing for a margin of safety to the expected point of failure, while another device’s ratings may reflect a maximum performance figure under ideal conditions, from which a user is expected to de-rate the quoted value as appropriate for their particular conditions.
Diodes’ current carrying capacities in a forward biased mode of operation are not without limit, and several ways of characterizing that limit exist. One of the more common is an average rectified current rating, which as the name suggests refers to the maximum permissible time-averaged current flow through a device. It’s a limitation that derives from thermal considerations, so quoted values will come with thermal-related conditions and qualifications that are no less significant that the rated values themselves, and in most cases such ratings describe a steady-state mode of operation in which thermal equilibrium has been achieved.
Another means of characterizing forward current limitations are peak values, which describe maximum permissible forward current values for short time periods. Qualifications attaching to characterizations of this sort commonly include wave shape and/or duration, and whether the rating applies to a repetitive condition (e.g. once each half-cycle of an AC waveform) or to a one-time event. This too is often a thermally-derived limitation, but one describing device behavior under rapidly-changing conditions where thermal transfer to the surroundings is negligible because of the short time frames involved–a scenario complementary to the thermal equilibrium conditions limits described by average rectified current ratings.
Diodes are not lossless devices; some amount of voltage drop appears across them when carrying current in the forward biased direction, and this is called the device’s forward voltage and abbreviated Vf. Observed values for a given device will vary significantly with device temperature and the amount of forward current (If) applied when making the measurement. To a first and fairly good approximation until a device’s rated maximums are approached, Vf is linearly related to the logarithm of If and varies with temperature according to some roughly-constant coefficient.
For a given device rating and forward current, different diode types will exhibit different forward voltages. From a conduction loss standpoint minimizing Vf is desirable, though in practical devices making improvements in Vf characteristics often comes at a cost of worsening leakage and other characteristics.
Current flow through a diode under some amount of reverse bias that’s not sufficient to cause reverse breakdown is called leakage . Leakage currents for a given device will typically increase with temperature and the amount of reverse bias voltage applied, and varies by orders of magnitude depending on device type, usually as an engineering trade-off with forward voltage.
The power dissipation in a device that results from leakage currents can be non-trivial, particularly for high-leakage types with relatively high reverse voltage ratings. Such devices are vulnerable to a thermal runaway situation, wherein leakage currents cause a device to heat up, causing further leakage, more heating, more leakage, etc. until something breaks. As a loose rule of thumb, leakage currents can be expected to double with every 10°C increase in temperature. If a leakage characterization is offered for a 25°C device temperature (as is common) one should be prepared for observed values to be roughly a thousand times higher as the device temperature approaches its rated maximum.
When reverse-biased, the regions of a diode on either side of the junction act like the electrodes of a capacitor. Since the thickness of the junction region is physically quite small the amount of resulting capacitance can be significant, and because its effective thickness changes in response to the amount of reverse voltage applied, the capacitance is also voltage-dependent. The phenomenon is commonly harnessed to make a voltage-controlled capacitance and often used in RF applications, but in other use cases it’s simply another hidden parasitic element that’s easy to overlook until it causes problems.
To varying extents based on their construction, diodes exhibit a substantial reverse-recovery phenomenon wherein the process of transitioning from a forward-biased to a reverse-biased state involves a brief period of current flow through the device in the reverse direction, at significantly higher levels than characterized by steady-state leakage. Affected devices are often classified or described based on the time duration of this recovery period, e.g. “fast recovery” or “ultra-fast recovery” and further by the characteristic shape of the current waveform through the device during the recovery process, e.g. “soft recovery”.
A diode’s recovery characteristics are significant for a variety of reasons. At the most basic level, it limits the maximum signal frequency that a given device can usefully rectify, and affects the efficiency of the rectification process. More subtly, a diode’s recovery characteristics can strongly affect the peak voltage stresses that are experienced in a circuit, and the amount of noise generated by the switching process.
These latter effects occur because reverse current flow through a diode causes energy to be stored in connected inductances, whether the ones put there for a reason or the parasitic ones that exist as the unavoidable consequences of physics. Because reverse recovery currents are flowing in a “wrong” direction, energy stored as a result is generally destined to be wasted on undesirable things, like increased device stresses or unwanted noise emissions. Longer recovery periods result in more of this unwanted energy storage, and recoveries that end abruptly result in greater component stress and noise generation than those that end more gradually.
The image at right shows voltage (yellow) and current (blue) waveforms of a +/- 2v pulse being applied to a 1N4002T diode in series with a 51 ohm resistor, and illustrates the recovery process. The reverse current flow during recovery lasts almost a third of a microsecond, and has a peak amplitude roughly twice that of the forward current that had been flowing.
For more information related to diode basics and recovery phenomena in particular, see these suggested resources
The classic and most common use of diodes is for rectification; the electrical equivalent of a one-way valve. Though the terms “diode” and “rectifier” are often used interchangeably, it is also common to reserve the term “rectifier” for devices designed to handle higher current flows, and the term “diode” for devices designed for low-current, small-signal applications.
Rectifier diodes based on a simple P-N junction are known as standard diodes. As minority-carrier devices they have a substantial reverse-recovery behavior, and their forward voltage characteristics are among the highest of the common rectifier types. On the the positive side, standard diodes tend to have significantly lower leakage current and junction capacitance than other rectifier types, and are available with reverse voltage ratings well into the kV range, whereas a number of other types are limited to a few hundred volts at most. This combination of attributes makes standard diodes more favorable as application voltages increase, diminishing the significance of a higher Vf and amplifying the benefits of lower leakage.
Many standard diodes are described as being “glass passivated.” This refers to a process of protecting the edges of the P-N junction with a hermetic glass seal, which helps improve the stability and consistency of device behaviors such as leakage, reverse recovery, and reverse voltage tolerance.
Shottky diodes are rectifier diodes that differ from standard types by virtue of being constructed using a metal-semiconductor junction rather than a junction of differently-doped semiconductor regions. The practical effect of this difference is that shottky diodes typically have a lower forward voltage characteristic and a near-zero recovery time
compared to standard diodes, and the accompanying disadvantages of higher leakage currents, a lower degree of electrical robustness and long-term reliability, and a lower upper limit on maximum operating voltage of available devices. A common technique for mitigating these vulnerabilities known as a “guard ring” creates a parasitic P-N junction diode in parallel with the shottky device. Should aggressive forward drive current conditions cause this junction to become forward biased, it will exhibit a reverse-recovery behavior as any other standard diode, and degrade the recovery performance of the overall device.
Typical silicon-based shottky diodes are predominantly used at application voltages of 100v or less at the time of writing. Higher-rated devices are available and the bar will likely move as technology develops, but the low-Vf /high leakage characteristic of shottky devices renders them less favorable as application voltages increase.
Finally, shottky diodes should not be confused with the phonetically-similar but now obsolescent shockley diode, an early device used as a voltage-controlled switch.
For more information on shottky diode characteristics and application considerations, see these suggested resources.
FERD® (Field Effect Rectifier Diode)
Not to be confused with a FRED (Fast Recovery Epitaxial Diode, a standard diode designed for short recovery times.) FERD® diodes are a proprietary offering from ST Microelectronics that are marketed as an improvement on similarly-rated shottky diodes, offering a lower forward voltage with comparable leakage and a similar zero-recovery turn-off behavior. Notably, the Vf vs. If characteristic for FERDs tends to rise more rapidly than is typical for shottky devices, with the result that while a FERD likely will exhibit a lower Vf than a comparable shottky at low to moderate forward currents, the situation may reverse as forward currents approach the devices’ rated maximums.
Another proprietary designation (this one from Diodes Inc.) super barrier rectifier (SBR®) diodes are also marketed as an improvement on shottky diode technology, with similar characteristics relative thereto as the FERD® diodes discussed above, and predating them by a number of years. A few persons of note appear to have been involved with the development of both during their tenures with the respective manufacturers, and the two can be considered functionally similar families designed to improve on the forward conduction loss characteristics of shottky diodes, with attendant drawbacks of high leakage currents and limited reverse voltage capability.
Silicon carbide (SiC) is a semiconductor base material analogous to silicon, and one of a number of emerging materials known as wide bandgap semiconductors which offer enhanced performance potential relative to comparable silicon-based devices, chiefly in terms of allowing higher operating voltages and temperatures in the case of SiC. Most SiC rectifier devices available at the time of writing are based on the shottky metal-semiconductor design concept, and carry reverse voltage ratings of 600V or more. A likely factor in the trend is the characteristically higher (by about 50~100%) forward voltage of SiC shottky devices relative to standard types, which becomes proportionally less significant as application voltages increase.
Above: A plot of rated forward current vs. reverse voltage for rectifier diodes available from Digi-Key at the time of writing.
Above: Instantaneous forward voltage vs. current plots for a sampling of different rectifier diode types with nominal 20A, 100V ratings. Leakage current characteristics for each are indicated in the chart legend.
Diodes designed to break down
Rectifier diodes are generally optimized for rectifying, and aren’t designed to tolerate reverse voltage stresses high enough to cause the diode junction to break down and begin conducting beyond normal leakage in the reverse direction, either on a transient or continuous basis. Rectification is but one of several tricks that diodes have been taught however, and a number of diode families exist that are expressly designed to be tolerant of operation in a reverse breakdown mode.
In Digi-Speak and several other dialects, the term “avalanche diode” (or “controlled avalanche diode”) refers to a standard P-N junction rectifier diode designed to tolerate being pushed into reverse breakdown on a limited basis, above and beyond what other rectifier devices are expected to tolerate. Reverse voltage ratings in devices sold as avalanche diodes are relatively high compared to those of devices sold as zener diodes, ranging from about 50V to several kV, whereas available devices sold as zener diodes reach to a few hundred volts at most.
Worthy of note is that the avalanche phenomenon can be quite noisy, to the extent that in other dialects, an “avalanche diode” is understood as a device used as a noise source, rather than for rectification. Indeed, devices exist that are optimized for either case. Insofar as those available from Digi-Key at the time of writing are those oriented toward rectifier use, they are listed among the other rectifier diode types and characterized as such. One noteworthy exception is that avalanche diodes are likely to carry an avalanche mode energy rating, which characterizes the amount of reverse-breakdown energy a device can safely absorb. For non-avalanche types, this limit is typically not characterized and is presumed to be zero.avalanche diode datasheet with maximum avalanche energy specification highlighted.
Zener diodes are a family of P-N junction devices that are capable of rectification, but which are designed to enter reverse breakdown at a relatively precise voltage and to tolerate such operation indefinitely so long as device temperatures are maintained within acceptable limits. This behavior leads to their frequent use in voltage reference and regulation applications.
It’s a bit confusing from a terminology standpoint, but the physical phenomena known as zener breakdown and avalanche breakdown that are responsible for reverse conduction in semiconductor diodes are both at work in devices sold as zener diodes, to varying degrees depending on the breakdown voltage of the device. Zener breakdown dominates in devices with zener voltage ratings below about 5 to 7 volts, while above this point avalanche breakdown is prominent, with some mix-and-match action occurring around the transition point. The distinction can be seen in the reverse current vs. voltage curves in zener datasheets that cover a range of reverse voltage ratings; the low-voltage devices have a rather gradual curvature and gentle slope, whereas the higher-voltage devices have a relatively sharp “knee” and more aggressive slope.
Attributes of Zener Diodes
The nominal zener voltage of a device characterizes the voltage that appears across a device when a test current of a specified value is passed through it in the reverse direction. It’s typically measured under pulsed conditions with the device under test at room temperature. Because it’s a temperature-sensitive characteristic and significant power is dissipated in the device as a result of the current passing through it, observed values in practice are likely to differ from nominal values. Variation of zener voltage may have a positive or negative correlation with temperature, depending on whether the zener or avalanche breakdown is dominant; devices with breakdown voltages in the transition region (around 5.1v) often exhibit the lowest temperature drift characteristics among similar devices in a given product series. Observed values will also vary with the actual amount of reverse current flow present, both as a consequence of thermal effects, and due to the slope of a given device’s reverse voltage/current characteristic. This variation tends to become more pronounced as nominal zener voltages fall below about 7 volts.
Above: Zener current v. voltage plots for the BZT52C series of zener diodes, showing characteristics for devices with different nominal zener voltage values. Note the change in curve shapes that occurs in the 4~7v region.
The tolerance figure commonly quoted in conjunction with zener diodes characterizes variability in zener voltage among devices due to variability in materials and manufacturing. It is measured and applicable under the specified test conditions only; different conditions will likely result in observed values outside the indicated range.
An example of how this is communicated by a device datasheet is shown at right; temperature conditions are indicted by a catch-all 25°C ambient temperature and a note indicating pulse-based testing implying a device die temperature near that 25°C also. Test current conditions are shown by the column heading for VZ values, which indicate that the figures shown are applicable at the adjacent IZT values. While this example does not reference a “tolerance” per se, the deviations of the min & max values from nominal may be converted to a percentage, and are commonly communicated as such for trade purposes.
The maximum power figures quoted for zener diodes characterize the maximum amount of power that can be continuously dissipated in a device under specified test conditions. It’s a thermally-derived rating, commonly reckoned as the amount of power dissipation that results in the device’s die temperature reaching the listed allowable maximum under specified mounting and ambient temperature conditions. Since maximum-temperature operation tends to be quite stressful on a device and actual application conditions are often less favorable than those used for deriving the listed power rating, maximum safe power dissipation levels in practice are likely to be significantly less than the listed values.
The figure at right shows an example of how this is communicated in a datasheet; two maximum power figures are actually indicated, one for a scenario where the device’s lead temperature is maintained at 75°C, and one for a scenario where the ambient temperature is maintained at 25°C and the device is mounted as specified in a separate note elsewhere in the document.
Zener Impedance figures describe the relationship between zener voltage and current when a device is operating in its reverse breakdown region; an increase in zener current produces an increase in zener voltage, and the coefficient of proportionality between the two is known as the zener impedance, in the pattern of Ohm’s law. It is common for manufacturers to offer characterizations of zener impedance at two test current conditions; one at the same nominal test current value at which the zener voltage is specified, and another at a lesser current, near the “knee” of the device’s operating curve.
An example of how this information is represented in a datasheet is show at right. The information is often presented both in tabular and graphical form, with zener impedance amounting to the slope of a device’s operating curve, at the point corresponding to the test current in question.
The concept of reverse leakage in zener diodes is not much different than in the context of diodes generally, though its relevance may not be immediately apparent given that zeners generally are used in a reverse breakdown mode where they’re expected to ‘leak’ copiously. In clamping or detection-type applications however, it’s usually desirable for them to conduct as little current as possible when the applied reverse voltage does not exceed the zener voltage.
Insofar as the measured leakage current will increase from zero (at zero reverse voltage) up to the test current value at the zener voltage, the reverse voltage condition that applies to listed leakage values is nearly as meaningful as the current measurement itself; a zener that leaks a microamp at 30% of its rated zener voltage is a lot more “leaky” than one that leaks the same microamp at 80% of Vz.
For further information on avalanche and zener diodes, see these suggested resources.
Transient Voltage Suppression (TVS) Diodes
Devices marketed as TVS diodes are designed for use in protecting against momentary over-voltage events, and as such are adapted for and characterized in terms of their operation during brief periods of relatively high reverse current flow. They come in a few distinct types:
- Zener/avalanche type devices (example) that function as a voltage clamp and which directly dissipate the energy contained in an over-voltage event.
- Rectifier-type or “steering” devices (example) used to divert currents induced by over-voltage events around a protected device or circuit, frequently into the power supply network associated with the protected circuit. Such devices are actually not intended to be operated in a reverse-breakdown mode, but since they are characterized for pulsed operation and used in pursuit of the same end goals as zener and compound TVS devices, they end up being grouped alongside them. TVS diode attributes that pertain to reverse-breakdown modes of operation are not directly applicable to this class of devices, however they may be characterized in similar ways based on forward-biased operation.
- Compound/mixed-type devices (example) often containing an array of steering type devices that allow a single integrated zener-type device to provide protection for several independent circuits, but also found with a variety of subtly different architectures that impart some benefit of convenience or efficacy in various situations.
Above (L-R): Datasheet representations (links at left) of a bi-directional zener-type TVS diode, a steering-type TVS diode array, and a multi-channel mixed-type array.
Attributes of TVS diodes
TVS diode devices can be characterized as unidirectional or bidirectional , depending on whether their breakdown voltages are asymmetric or symmetric respectively with regard to applied polarity.
Though simple enough in context of single-channel, standalone devices with only two terminals, the concept rapidly looses clarity in context of multi-channel, mixed-type devices or those with multiple configuration possibilities, as the point of reference for determining the polarity of an applied voltage opens up to interpretation in those cases.
Also, very few devices are characterized as both bidirectional and unidirectional. For this reason, it’s advisable avoid applying parametric filters for unidirectional and bidirectional channel count concurrently.
Above: a sine wave voltage (yellow) applied to a unidirectional TVS diode and resulting current flow (green). Note the asymmetry of the voltages at which positive and negative peaks of the applied waveform are clipped.
Above: the same sine wave used at left applied to a bidirectional version of the same TVS diode. Note the symmetry of clipping levels and current flows.
Reverse Standoff Voltage
Rarely, if ever is it desirable for the voltage applied to a TVS diode to result in current flow beyond the inevitable leakage during normal/continuous operation–they are after all, called transient voltage suppressors… A TVS diode’s reverse standoff voltage rating characterizes the maximum voltage that can be applied before entering that realm, and is also sometimes called a maximum working voltage or maximum reverse working voltage . This figure will be less by some margin than the device’s rated breakdown voltage.
Differences in the way the concept is expressed can lead to the same essential idea being expressed as either a maximum or minimum value; the phrase “reverse standoff voltage” implies a focus on the characteristics of the part itself, and such values are likely to be quoted as minimums. In contrast, an expression like “maximum working voltage” implies a focus on the stresses that the user is applying to the part, and these figures are often quoted as maximums. They’re two sides of the same coin; “the part’s minimum stress withstanding capability is x” is the mirror image of “the maximum stress that can be safely applied is x.”
As might be expected, a TVS diode’s rated breakdown voltage characterizes the amount of reverse bias at which it begins conducting more than some specified (small) amount of current. It’s commonly quoted in terms of some minimum value that’s greater than the maximum reverse standoff voltage specified, with the gap between these two figures providing the needed margin to avoid unwanted conduction.
The datasheet excerpt below lists characteristics for the sample TVS diodes used to gather the accompanying waveforms at right and above.
The clamping voltage given for a TVS diode characterizes the maximum voltage appearing across the device when a standard test current waveform with the specified peak amplitude is applied. Since those peak current values are generally chosen by the manufacturer to communicate the upper limit of a device’s capabilities, clamping voltages can be loosely understood as describing the maximum voltage expected to appear across the device so long as it’s still doing it’s job. Clamping voltages are not exact figures; they’re characterizations based on a standardized test waveform. There’s more than one of those to chose from, and while they’re designed to model likely scenarios, in practice nature is not always so kind as to conform to the molds one tries to make of its character.
The waveforms at right show voltage across (yellow) and current through (green) an SA5.0A TVS diode when a charged 47,000uF capacitor is connected through a relay. The resulting 15A surge through the TVS is clamped to less than 8V.
Above: a representation of the circuit used to generate the waveforms shown at right.
Peak Pulse Current
The peak pulse current quoted for a TVS diode typically has a dual meaning, in that it describes both a measurement condition at which the stated clamping voltage is characterized, and also an upper limit to the amplitude of the transients a device is capable of suppressing without expectation of damage. The figure necessarily comes with a qualification regarding waveform shape and duration, with the 10x1000us and 8x20us standard waveforms being commonly referenced.
Peak Pulse Power
The peak pulse power quoted for a TVS diode characterizes the maximum allowable power dissipation in the device during a transient event, typically based on the same test waveform as the peak pulse current.
Light Emitting Diodes (LEDs)
Light emitting diodes, not surprisingly, are diodes that emit light when forward-biased. While low-power devices suitable for use in milliwatt-scale indication applications have been commercially available since about the 1970s, advancements in technology have enabled the development of high-power devices suitable for illumination and general lighting applications, with devices operating at hundred-watt input power levels being readily available. Variations in the products themselves and the ways that they are characterized lead to segmentation along several common lines; broadband (white) vs. single-color vs. non-visible, indication-class vs. illumination-class devices, and by the level of integration at which a product is traded; packaged semiconductors are handled differently than when those same products are integrated into a higher-functioning assembly.
Above, L-R: Surface-mount and through-hole indicator-class LEDs, a lighting-class LED, and a chip-on-board array. (photos not to scale)
Attributes of LEDs
Wavelength, Dominant Wavelength, Peak wavelength
The various permutations of the wavelength attributes used to describe LEDs communicate information about the emitted light’s spectral content.
This is relevant because perceived color and “actual” color aren’t necessarily in complete agreement. The basic model for human color perception is trichromatic; most people have 3 different types of color-sensitive cells in their eyes, which are stimulated by incoming light to different degrees based on that light’s spectral content. Regardless of actual spectral content, any two stimuli that produce the same ratio of responses from the color receptors in the eye are perceived as having the same color. Consequently, light sources with significantly different spectral content can have the same apparent color.
Dominant wavelength is a perceptual measure, indicating the wavelength of a monochromatic light that would produce the same perceived color (based on a standardized human model) as the light emitted by the device being characterized. This is most meaningful for indicator-like applications where an LED is being viewed directly rather than being used to illuminate some other object.
Peak wavelength on the other hand, indicates the high point of a device’s emission spectrum, and is of greater interest in applications where actual spectral content is of greater importance than perceived color, such as horticultural lighting.
When dominant and peak wavelengths are both characterized they often fall quite close together, though they don’t overlap perfectly and the difference tends to increase near the edges of the visible range. This occurs as a result of asymmetries in human color response, and the manner in which the “skirts” of a light source’s spectral density and human sensitivity curves overlap.
Wavelength by itself doesn’t imply reference to either dominant or peak values, and neither case should be assumed; the term may have mixed significance even within a datasheet for a single LED series, with “peak” figures being quoted for extreme-wavelength devices and “dominant” figures quoted for those nearer the middle of the visible spectrum. This is perhaps not so disorganized a practice as it might seem; devices with peak wavelengths near the extreme ends of the visible spectrum are less likely to be used for illumination purposes, since their emission spectra lie in regions to which the human eye is not very sensitive. Because they are not especially effective at providing illumination as a result, they’re more likely to be used in applications where their actual (vs. perceived) spectral qualities are of interest.
Finally, wavelength figures quoted in nanometers are only applicable to colored light sources; “white” light sources are mixtures of light across a broad spectrum encompassing most of the visible range, and are specified in terms of correlated color temperature with values quoted in degrees kelvin (°K). Since it’s often expedient to group white and colored LEDs together for product curation purposes, it’s common (if technically incorrect) to see CCT values with units of temperature listed under a “wavelength” heading.
Above: Normalized power spectral density of a WP3A8HD LED, plotted with the CIE 1931 standard color matching functions. The peak wavelength in the LED’s output spectrum is outside the range of significant sensitivity, but the “skirt” of the LED’s emitted spectrum has substantial overlap (shaded), with that encompassed by the X curve being about twice that of the Y. Note that the dominant wavelength listed for the device (635nm) plotted as a vertical line intersects the X and Y curves at values having a similar ratio.
Test current, test temperature
Measurements of device performance and behavior are directly influenced by the amount of forward current used when making the measurements, and the temperature of the device when measurements are made is also quite significant, though to a lesser degree. Listed Test current and Test Temperature values for an LED reflect the values of these variables at which other listed performance characteristics apply. While test current values are almost universally listed, test temperature values tend to be parametrically exposed only for lighting-class LED products. The relatively high power dissipation typical of these devices makes operation at 75 to 100°C more of a norm than an exception, rendering measurements of behavior at the customary 25°C device temperature an inaccurate reflection of practical results.
LED performance tends to degrade in all respects as device temperature increases, so a device with a higher test temperature is to be preferred when devices with numerically similar performance characteristics measured at different test temperatures are being compared. For estimation purposes, an increase in test temperature from 25°C to 85°C can be expected to result in a 10% decrease in luminous flux and efficacy, though changes in the range of 5 to 20% are not abnormal.
Lumens/Watt (luminous efficacy)
Luminous efficacy is a perceptual measure of an LED’s light output per unit of electrical input energy, based on a standard model for the wavelength-dependent sensitivity of human vision. Stated differently, it’s a measure of how effective a device is at turning a unit of electrical input energy into light that humans find useful. Measurements are typically quoted based on a photopic (color perception at normal light levels) vision model, in which the theoretical maximum is 683 lumens per watt. That maximum applies at the peak of the human spectral sensitivity curve around 555nm (green) so theoretical maximums for light sources with broader spectral content (e.g. “white” light) will be lower.
Flux @ ______
Luminous flux , measured in units of lumens , is a measure of perceived optical power. Because human vision is not uniformly sensitive to all wavelengths, the usual all-purpose unit for measuring power (the watt ) doesn’t serve well in situations where providing illumination is the goal; one watt of red light does not provide the same illumination benefit as one watt of green light, for example. The luminous flux concept works around this limitation by weighting the spectral content of a light source according to a standard luminosity function, which describes the variation of human vision sensitivity as a function of wavelength.
The various flux attributes used to describe LEDs communicate the amount of optical power produced by a device. These figures are directly dependent on the forward current applied to the LED when the measurement was made and somewhat less directly (though strongly) on the temperature of the device. Accordingly, the listed flux values are applicable at the also-listed test current and temperature.
The photo at right illustrates the difference in illumination benefit derived from applying 100mW to an LED with an indicated luminous efficacy of 46lm/W (left) versus a similar device with an efficacy of 201 lm/W (right). Both are mounted to the opposite side of the proto board in the foreground, aiming at a sheet of white paper, with another sheet folded into a V-shape separating the areas illuminated by each.
An LED’s viewing angle summarizes spatial variations in the intensity of the light emitted by an LED. A relatively well-observed convention is to quote viewing angles based on the total angle in in a plane through which the emitted light intensity is at least half of its peak value. Such values are sometimes referred to as full-width, half-max or FWHM measurements for greater clarity. Since emission patterns are generally symmetric about a device’s optical axis, it’s also fairly common to specify the angle between the optical axis and the half-intensity point. Values specified in this manner are sometimes called half-angles, and are equal to one-half of a viewing angle specified in FWHM fashion.
The image below compares two indicator-class LEDs, one with a nominal viewing angle of 180° (left), and another with a nominal viewing angle of 24° (right). The device with the narrower viewing angle projects a distinct bright region on a sheet of white paper a short distance away, while the wide-angle device emits a much more diffuse beam.
CCT (Correlated Color Temperature)
Correlated color temperature , frequently abbreviated as CCT , is a concept used to measure and communicate the color characteristics of nominally “white” light sources. The CCT of a given light source is the temperature of a blackbody radiator with a color that most closely approximates that of the light source in question. While perhaps not widely recognized by name, blackbody radiation is commonly understood in terms of its effects when (for example) a metal object is heated to a high temperature and begins to glow; a dull red at temperatures around 800°K (about 1000°F) on through orange, white, and eventually blue at temperatures beyond the melting points of currently-known solids.
It should be noted that “most closely” and “closely” are not the same thing, in the sense that Adak is a municipality in Alaska situated “most closely” to the equator, but is nevertheless over 3000 miles (~5000km) distant from it… Stated differently, CCT describes the closest point of reference for a “white” light’s color characteristics, but indicates nothing about the actual distance to that point of reference; this latter piece of information is characterized by a separate metric, known as Color Rendering Index or CRI.
The figure at right shows normalized emitted intensity versus wavelength for a blackbody radiator at different temperatures–note that both scales are logarithmic. At 2700°K (roughly the operating temperature of a traditional incandescent light bulb) there’s roughly 100 times more light at the red end of the spectrum being emitted than at the blue end, resulting in a light with a yellow-ish hue. At 5600°K, the difference is drastically reduced to a factor of only 4, and because of relatively high human sensitivity to light at blue wavelengths, such a light source would appear blue-ish. While the relative balance of spectral content changes with the temperature, the majority of the energy radiated by a blackbody does not fall within the visible range, making the luminous efficacy of lighting agents that rely on blackbody-like radiation characteristically low.
Also worthy of note is that the the descriptive terms “warm white” and “cool white” often used to refer to CCT in a general qualitative sense do not refer to a blackbody temperature, but instead to a subjective/aesthetic/psychological human response evoked by a light source, which is backwards from the CCT figures; a light source with a high CCT is called “cool white” while those with a lower CCT are called “warm white.”
Above: A plot of the emission spectra for blackbody radiators at several different temperatures.
Below: (L-R) An incandescent lamp, and LED-based lamps with advertised CCT values of 2700°K, 3000°K, and 5000°K
CRI (Color Rendering Index)
Color rendering index or CRI is a means of characterizing the color reproduction properties of nominally “white” light sources, based on how closely the apparent color of objects illuminated by a light source matches their apparent color when illuminated by a blackbody-based reference standard of the same CCT. The theoretical maximum CRI value is 100, and the less accurately an illuminant reproduces colors relative to the standard, the lower its CRI. In effect, CRI figures are the “how far” supplement to the “nearest reference” information provided by CCT data.
In the interest of trying to get more lighting benefit from a unit of input energy, illuminants such as fluorescent lamps and LEDs have been developed that sidestep the built-in inefficacies of lighting technologies based on blackbody radiation principles. Their resulting output spectra tend to be rather peak-ish, rather than the smooth curve characteristic of a blackbody.
As a result, the apparent color of illuminated objects can vary significantly depending on the spectral content of the light source. The sodium-based lamps often used for outdoor area lighting (street lights, parking ramp lights, etc.) and which tend to make everything appear some shade of orange-yellow or black are perhaps the most familiar example of a light source with a very low CRI.
A millicandela rating is a measure of perceived light intensity emitted by a device in a specific direction, usually along the device’s optical axis where this intensity is highest. It’s most often used in context of indicators, segmented displays, and similar devices that are designed for direct viewing, because in order to be readily visible the intensity of the light emitted by a device in the direction of an observer needs to be distinguishable from that of any ambient light reflecting off the device’s surface. This concept stands in contrast to the units of luminous flux used to characterize devices designed for illumination, where the total amount of energy emitted in all directions is of greater interest than the intensity of emission in one specific direction.
More precisely, the candela is the metric base unit for luminous intensity; the amount of light energy per unit of solid angle emitted by a device in a specific direction, adjusted to account for the spectral dependence of human vision sensitivity. Specifically, it is defined as the luminous intensity of a monochromatic 555nm light source with a radiant intensity of 1/683 watt per steradian. And of course, a millicandela is 1/1000th of a candela.
And for those that are wondering, a steradian (think “stereo radian”) is the three-dimensional cousin of the radian angle of measure used to describe angles in a plane; there’s 2π radians in a circle, and there are 4π steradians in a sphere. A close link exists between a device’s millicandela rating and its viewing angle because of this geometry concept; an LED that focuses its light output into a narrower beam will appear brighter than one that spreads the exact same amount of light energy over a larger solid angle.
Further information regarding LEDs and related topics can be found among these suggested resources.
Above: Segmented displays that are both easily visible in darkness (bottom) may not both be as readable under ambient light (top).
Photodiodes are a family of devices closely related to rectifier diodes and based on similar basic concepts, but which are designed to exploit (rather than suppress) the photoelectric effect and use it for detecting or measuring light in applications as varied as changing the channel on the TV or measuring the concentration of oxygen in one’s bloodstream. Though capable of performing a rectification function photodiodes are relatively poor at it, just as rectifier diodes can function (poorly) as photodiodes.
The photos at right illustrate the difference in behavior between a common 1N4148 rectifier diode in a glass package and an SFH229 photodiode, when illuminated with a red LED from the random parts box driven by an AC current. The purpose-built photodiode generates a signal of roughly 7 v amplitude, compared to about 2 mV for the '4148.
Above: 1N4148 rectifier diode illuminated by an LED.
Above: Photoelectric signal (yellow) produced by rectifier diode illuminated by LED driven with an AC current (green)
Above:The circuit used to generate the waveforms in the images at right.
Above: SFH229 photodiode illuminated by an LED
Above: photoelectric signal (yellow) and LED drive current (green)
Photodiodes are generally operated in one of two fashions. In photovoltaic mode a device is operated with little to no applied voltage, and behaves much like a very small-scale version of the photovoltaic panels used for solar energy harvesting. In this mode of operation the diode is forward-biased; the anode becomes positive with respect to the cathode terminal. If left open or applied to a high-value resistor, the resulting voltage has a logarithmic relationship with incident light, much like a rectifier diode’s forward voltage is logarithmically related to forward current. An example of this is shown middle right, using the same photodiode and LED as in the above.
On the other hand, if the photodiode is short-circuited and the resulting current flow is monitored, the resulting signal is quite linear with respect to incident light, as in the image at far right. The voltage amplitude of the resulting signal in this case is also several orders of magnitude higher per unit of current applied to the light source.
Because a photodiode is essentially unbiased when operated in photovoltaic mode, junction capacitance is at a maximum and response times are slowed. Excess noise sources and power dissipation in the device are minimized however, making this mode useful for low-speed/high precision applications, which in many cases are applications where light itself is being measured.
Above: Input and output waveforms for an LED/photodiode pair, operating the LED in high-impedance photovoltaic mode.
Above: Input and output waveforms for an LED/photodiode pair, operating the LED in short-circuit photovoltaic mode.
In contrast, operation in photoconductive mode amounts to placing a photodiode under some amount of reverse bias and measuring the resulting leakage current; incoming light that is absorbed by the device creates new charge carriers, which are then available to carry current under the influence of the applied reverse bias voltage. Increasing the amount of bias applied reduces the diode’s junction capacitance, enabling a faster response time at a cost of increased noise in the resulting signal. Due to this effect, photoconductive mode tends to be preferred for applications where speed is a concern; these are often applications concerned more with detecting light to decode information that it’s carrying rather than measuring the light itself, with fiber optic communications receivers being an excellent example.
The wavelength attribute listed for photodiodes on the Digi-Key website indicates a device’s wavelength of peak sensitivity; the wavelength of light which produces a maximum electrical response per unit of incident light energy.
Photodiodes designed to allow responsivity to some portion of the light spectrum are often described as “enhanced” for that region. This should not be interpreted to mean that these devices are especially sensitive in that region; it may mean that some effort has been made to improve that region of the sensitivity curve relative to the rest, or simply that sensitivity in the referenced region has not been deliberately reduced through the use of optical filtering. It’s probably best understood as an imprecise quick-select mechanism for the more accurate (and unwieldy) information surfaced under the Spectral Range attribute.
By way of example, the PDB-C152SM is described as a “blue-enhanced” device by the manufacturer, and it’s spectral response curve is shown at right, highlighted to show the visible colors associated with different wavelength regions. It can be seen quite clearly that responsivity in the “blue” region is roughly a third to a quarter of the peak value, which lies in the infrared. Nonetheless, the device does have a potentially useful level of blue-spectrum response, and is thus said to be “enhanced” in that region.
Above: Responsivity curve for a PDB-C152SM photodiode with visible spectrum highlighted.
The spectral range attribute for a photodiode describes the range of light wavelengths over which a device’s responsivity is above some manufacturer-defined threshold, commonly in the range of 10~50% of the peak value. Conventions on the exact value chosen are not particularly strong, so listed values should be interpreted somewhat loosely and their exact meaning ascertained if it is of possible relevance to the application in question. As an example, the TEMD1000 datasheet excerpt below indicates a 50% threshold is used for the indicated range, whereas the SFH2701 datasheet (middle right) indicates use of a 10% threshold. If the same 50% level is used for both, the TEMD1000’s infrared response is found to extend roughly 70nm further.
Above: Excerpt from the SFH2701 datasheet highlighting the spectral range characterization. Note use of a 10% threshold level.
Above: Spectral sensitivity plot from the SFH2701 datasheet. If evaluated using the same 50% sensitivity level as used for the TEMD1000, its range is roughly 480 to 970nm.
A PIN photodiode is a variant of a standard P-N junction diode incorporating an un-doped ‘intrinsic’ layer between the P- and N-doped regions. This has the effect of making a bigger target for incident light and reducing junction capacitance, allowing higher sensitivity and faster response than obtainable with a comparable P-N junction.
Avalanche photodiodes (APDs) are operated in photoconductive mode at relatively high reverse bias voltages, and exhibit significantly higher sensitivities to incident light than PIN-type photodiodes by incorporating a form of internal amplification. Rather than the output signal being directly due to the charge carriers freed by absorption of incident light, APDs use the applied reverse voltage to accelerate those charge carriers to a point where they have enough energy to cause an avalanche effect in an adjacent reverse-biased P-N structure, resulting in an output signal tens or hundreds of times larger than what would have been achieved were the original photo-generated charge carriers solely responsible for the output signal. This higher sensitivity comes at a cost of increased noise levels, and the combination of high reverse voltages and potentially large photocurrent signals can result in significant power dissipation within the device.
Above: Representations of PIN diode structure compared to a standard PN junction.
Above: Illustration of an avalanche photodiode in operation. Incident light striking the diode’s absorption region creates new charge carriers, which are accelerated by the applied reverse bias voltage and result in the freeing of a larger number of charge carriers through the avalanche effect upon striking an adjacent P-N junction.
A photodiode’s responsivity characterizes its sensitivity to incident light, in terms of the amount of output current produced per unit of incident light energy. It varies with the spectral content of the light and amount of any applied bias voltage, and the figures quoted are applicable for a monochromatic test source of the indicated wavelength, which is often different from a device’s wavelength of peak sensitivity. Aside from figures not being listed for a substantial number of products, this variability of context under which the quoted values apply limits the degree to which the listed values can be used as a fair basis of comparison between different devices. Additionally, other factors such as a device’s active area will influence the response of a device to a light source of given intensity.
A photodiode’s response time characterizes the delay between application of an optical stimulus and the resulting output signal approaching it’s final resulting value. It’s strongly affected by junction capacitance, which is in turn influenced by user-defined factors such as the chosen mode of operation and amount of any reverse-bias voltage applied. Advertised response time values therefore reflect the measurement conditions chosen by the manufacturer almost as much as they reflect the behavior of the parts themselves, and this fact should be borne in mind when making device selections–a device advertised as having a 10-nanosecond response time is likely faster than one with a 10-microsecond response time, but it’s less certain in a case where the advertised values are, for example, 5 microseconds versus 10.
Voltage - DC Reverse Max
The maximum reverse voltage attribute for photodiodes communicates essentially the same thing as is does in context of rectifier diodes, indicating the maximum voltage that can be applied in the reverse direction before breakdown becomes imminent. This does not necessarily represent a maximum safe working voltage however, since current flow (and therefore power dissipation) through a photodiode is a function of the optical input signal, by design. This is particularly relevant in context of avalanche photodiodes, which are typically operated at high reverse voltages and develop relatively high-current outputs.
Photodiode dark current is analogous to leakage current in a rectifier diode application–it’s current that flows through a reverse-biased photodiode in the absence of any incident light. Variations in this current constitute a noise source, and observed values vary significantly with temperature and applied bias voltage. As a result, these factors are important to keep in mind when comparing device characteristics or making design calculations.
The portion of a photodiode that’s light-sensitive is called its active area , the physical size of which directly influences the magnitude of the output signal for incident light of a given intensity. All else being equal, the increased sensitivity that comes with a larger active area is usually accompanied by higher junction capacitance, slower response times, and higher dark currents.
In the context of photodiodes, viewing angle refers to the manner in which a device’s sensitivity changes as a function of the direction from which incident light arrives. As in context of LEDs, conventional practice is to measure viewing angle as the full angle of incidence over which a device’s responsivity is at least half of its peak value, with half-angles also being common.
Variable Capacitance Diodes
Variable capacitance diodes, also known as tuning diodes or varactor diodes are designed to be used in a reverse-biased mode of operation, and exhibit a predictable change in junction capacitance that varies with the amount of bias voltage applied; put differently, they’re intended to function as electrically-adjustable capacitors, rather than rectifiers. They’re commonly used as a tuning element in RF applications and while most diodes can function in this manner, those marketed as variable capacitance diodes are purpose-built and characterized for this manner of use.
Junction capacitance receives more careful characterization in context of varactor diodes, though the principle giving rise to it is little different than in the case of common rectifier diodes. It is strongly dependent on the amount of reverse bias voltage applied, making the voltage at which a device’s nominal capacitance is characterized nearly as meaningful as the capacitance figure itself. Junction capacitance varies inversely with reverse bias voltage, so a capacitance figure that is characterized at a near-maximum bias condition can be considered a “minimum” capacitance value. Conversely, capacitance figures quoted at low bias voltages can be thought of as “maximum” values. Examples of both practices as well as the continuum between them can be found in the wild, making filtering by device capacitance an imprecise and/or cumbersome affair since one cannot know whether the top, bottom, or middle of a device’s capacitance range is being referenced without fairly close inspection. Since the number of available devices is not exceptionally large, a suggested selection strategy is to initially choose all values that are even remotely close to the range of interest; subsequent application of other filter criteria will tend to reduce the size of the result set, and reduce the amount of time one need spend interpreting test conditions.
In context of variable capacitance diodes, capacitance ratio refers to the ratio of junction capacitances produced at two different levels of applied reverse voltage. The ratio values themselves aren’t of much use without mention of the two voltage values on which they’re based (called the capacitance ratio condition in Digi-speak), and multiple values will often be offered by the manufacturer for characterization purposes. In such cases, the ratio based on the widest voltage spread (usually corresponding to the device’s maximum usable range) offered by the manufacturer is typically referenced in suppliers’ parametric data.
Q @ VR, F
In the context of variable capacitance diodes, a Q value refers to the ratio of a device’s reactive impedance to its parasitic series resistance, much as it does in the context of other reactive devices. Since a varactor diode is designed to provide a variable reactance depending on the amount of reverse voltage applied, that voltage is an important test condition when characterizing a device’s Q factor. Also relevant is the test frequency at which the characterization is made; parasitic elements such as packaging-related inductance can cause a Q value measured at 50 MHz for example, to differ significantly from one measured at 1 MHz.
For further information on variable capacitance diodes, see these suggested resources.
Fundamentals of Rectifiers (Vishay, 2 pages)
Contains a brief enumeration of rectifier classifications by recovery characteristics and an outline of descriptive parameters with their typical abbreviations.
Rectifiers-Physical Explanation (Vishay, 3 pages)
Contains brief descriptions of numerous diode attributes, with occasional elaboration on the significance or measurement of such.
Understanding Diode Reverse Recovery and its Effect on Switching Losses (Fairchild Semi, 11 pages)
Discusses the influence of diode recovery properties in the context of generalized half-bridge switching applications.
SNVA744: Choosing Standard Recovery Diode or Ultra-Fast diode in Snubber (Texas Instruments, 8 pages)
Discusses design considerations for diode selection in snubber circuits, in the context of flyback-topology switching power supplies.
IXAN0044: Fast Recovery Epitaxial Diodes (FRED) (Ixys, 9 pages)
Discusses the influence of diode recovery characteristics on switching applications, and differences among different fast-recovery rectifier diode types.
LED Color Mixing: Basics and Background (Cree, 23 pages)
Discusses color perception, colorimetry, and related concepts in context of LED lighting applications.
LED Color Characteristics (U.S. Dept. of Energy, 4 pages)
Provides a brief but useful summary of color-related concepts in lighting applications.
Optical Measurement Guidelines (Lumileds, 12 pages)
Describes processes for making optical measurements in context of illumination applications.
Converting Radiant Intensity in Units of mW/cm2 to mW/sr (TT Electrionics, 3 pages)
Describes the concept of solid angle and conversion of radiant intensity measurements between different measurement units.
AN32: Electrical Drive Considerations for Bridgelux Vero Series LED Arrays (Bridgelux, 20 pages)
Discusses drive considerations for high-power chip-on-board (CoB) LED arrays. Bridgelux products are discussed specifically, but the concepts are generally applicable to similar products from other manufacturers.
Cree Xlamp LED Electrical Overstress (Cree, 9 pages)
Illustrates the effects of electrical overstress events on LEDs and techniques for the mitigation of such. While Cree products are mentioned specifically, the information can be generalized to other products of a similar type.
AN30: Thermal Management for Bridgelux Vero series LED Arrays (Bridgelux, 31 pages)
Discusses heat transfer theory and application in context of high-power chip-on-board (CoB) LED arrays. Bridgelux products are discussed specifically, but the concepts are generally applicable to similar products from other manufacturers.
Reliability and Lifetime of LEDs (Osram, 19 pages)
Discusses reliability metrics and concepts as applied to LED components and related lighting applications.
Evaluating the Lifetime Behavior of LED Systems (Lumileds, 16 pages)
Discusses concepts and considerations relating to the longevity of LED lighting systems.
Cree XLamp LEDs Chemical Compatibility (Cree, 23 pages)
Provides examples of chemical compatibility issues in LED applications, mitigation techniques, and an empirical test method for determining chemical compatibility.
Chemical Compatibility of LEDs (Osram, 24 pages)
Describes LED construction and adverse effects that can arise from chemical imcompatibility.
Preventing LED Failures Caused by Corrosive Materials (Osram, 9 pages)
Offers a comparably brief summary of chemical incompatibility issues in context of LED lighting applications.