Chemiluminescence is the generation of electromagnetic radiation as light by the release of energy from a chemical reaction. While the light can, in principle, be emitted in the ultraviolet, visible or infrared region, those emitting visible light are the most common. They are also the most interesting and useful. Chemiluminescent reactions can be grouped into three types:
- Chemical reactions using synthetic compounds and usually involving a highly oxidized species such as a peroxide are commonly termed chemiluminescent reactions.
- Light-emitting reactions arising from a living organism, such as the firefly or jellyfish, are commonly termed bioluminescent reactions.
- Light-emitting reactions which take place by the use of electrical current are designated electrochemiluminescent reactions.
Chemiluminescent and bioluminescent reactions usually involve the cleavage or fragmentation of the O-O bond an organic peroxide compound. Peroxides, especially cyclic peroxides, are prevalent in light emitting reactions because the relatively weak peroxide bond is easily cleaved and the resulting molecular reorganization liberates a large amount of energy.
In order to achieve the highest levels of sensitivity, a chemiluminescent reaction must be as efficient as possible in generating photons of light. Each chemiluminescent compound or group can produce no more than one photon of light. A perfectly efficient reaction would have a chemiluminescence quantum yield
ΦCL of one, i.e. one photon/molecule reacted according to the equation:
ΦCL = ΦCE × ΦF × ΦR
The chemiexcitation quantum yield
ΦCEis the probability of generating an electronic excited state in a reaction and has a value between 0 and 1, with 0 being a completely dark reaction and, when 1, all product molecules are generated in the excited state. The most useful chemiluminescent reactions will have a
ΦCEof about 10-3 or greater. The fluorescence quantum yield
ΦFis the probability of the excited state emitting a photon by fluorescence rather than decaying by other processes. This property, which can have values between 0 and 1 is frequently at least 0.1. The reaction quantum yield
ΦRis the fraction of starting molecules which undergo the luminescent reaction rather than a side reaction. This value is usually about 1.
It is possible to increase the yield of chemiluminescence when the emitter is poorly fluorescent low
ΦF. A highly fluorescent acceptor is used in these cases in order to transfer the excitation energy from the primary excited state compound to the fluorescent acceptor/emitter. The chemiluminescence quantum yield is then determined by the equation:
ΦCL = ΦCE × ΦR × ΦET × ΦF′
The energy transfer quantum yield
ΦETexpresses the efficiency of converting the primary excited state formed in the reaction into the excited state of the acceptor. This value is often near 1. The fluorescence quantum yield of the acceptor/emitter
ΦF′should be also be near 1.
An appreciation of some of these fundamental principles of chemiluminescent reactions will help in understanding how to design chemiluminescent assays. A discussion of chemiluminescent measurement basics is also available.
> Download this information (pdf)
The use of light-producing chemical reactions for quantitative detection in ligand-binder assays is a rapidly growing field. Automated immunoassay analyzers from several different manufacturers are in commercial use and still others are in development. The technical demands of these assay markets requires detection technologies which are highly sensitive, not overly prone to interferences, robust and simple to use. Chemiluminescent processes have sufficient sensitivity and so are well suited for these applications. Most assays are heterogeneous and require separation of bound from unbound label. These assays are simpler to develop but somewhat more complicated to perform. A homogeneous assay, in contrast, requires the light-producing reaction to be affected in some way by binding of the ligand to its receptor.
One reason accounting for the growing popularity of chemiluminescent assays is their exquisite detection sensitivity. Unlike absorbance (colorimetric) or fluorescent measurements, assay samples typically contribute little or no native background chemiluminescence. The most serious source of background signal comes from the label being nonspecifically associated with the analyte through nonspecific binding events. This problem occurs in all detection systems. Measurement of light intensity is relatively simple, requiring only a photomultiplier or photodiode and the associated electronics to convert and record signals. The lack of inherent background and the ability to easily measure very low and very high light intensities with simple instrumentation provide a large potential dynamic range of measurement. Linear measurement over a dynamic range of 106 or 107 using purified compounds and standards is routine.
Factors in Choosing a Detection Reaction:
Enzyme Label or Direct Label?
A common means for generating chemiluminescence is to use a label enzyme to catalyze a chemiluminescent reaction. Each label can initiate multiple chemiluminescent reactions, so it is usually only necessary to incorporate one or a few enzyme labels/analyte. The chemiluminescent compound is supplied as the enzyme substrate in excess to assure saturation kinetics.
Light intensity is a linear function of the amount of label enzyme. The intensity, which is the rate of emission of photons/sec, is the product of the catalytic turnover of substrate E + S → E + S' (steps 1-3 below) and the lifetime of the light-producing compound S' → P* (step 4). The latter process is usually first order with a rate constant k and can be characterized by its half life
t1/2 = (ln 2)/k.
The lifetime of the excited state product P* → P (step 5) is extremely short in comparison to the other steps and has no effect on the observed reaction kinetics.
The chemiluminescence intensity/time profile consists of an initial rise period up to a prolonged emission at a plateau or pseudo-plateau level. When the first order reaction of S' → P* is slow, the rise time to plateau will be extended as the steady state concentration of S' is reached. When the reaction of S' → P* is fast, there will be a rapid rise and prolonged emission. The absence of a steady plateau value indicates either substrate depletion or inactivation of the enzyme. Detection of enzyme-generated chemiluminescence provides great flexibility in the measurement process. Light intensity at any time point through the plateau can be related to the amount of enzyme. Measurement can be made during the rising portion if speed is an issue as either a single point or a multi-point slope type measurement. For maximal sensitivity, measurement can be performed on the plateau.
Another way to use a chemiluminescent reaction for detecting an analyte in an immunoassay is to covalently label one of the complementary binding partners directly with a chemiluminescent compound. Triggering the chemiluminescent label to undergo the light-emitting reaction produces a signal for detecting the analyte. This approach has the advantage of being more straightforward and avoids the need to use large, relatively "sticky" labels. However, not many suitable chemiluminescent compounds exist. A suitable candidate must first have some functionality to allow for attachment to other molecules. More importantly, the label, upon triggering, should emit all of the light in the briefest possible period of time. When the chemiluminescence is emitted gradually over a period of time, signal intensity (photons/sec) is diminished and, in the worst case, can be low enough to impair detection sensitivity.
The optimum number of chemiluminescent tags on the species to be detected is best determined empirically. Even though more labels should theoretically provide more photons, quenching effects can occur if labels are too closely spaced. The surface area and the number of derivatizable groups, typically amino or sulfhydryl groups, provide further limitations. In practice 10-20 labels, at most, can be incorporated onto a large molecule. Moreover, the more labels on the surface of a binding molecule, the more likely that the label will interfere with its binding properties.
Factors in Choosing Assay Format:
Homogeneous or Heterogeneous?
Homogeneous or nonseparation assays are desirable because of their simple mix-and-measure nature. This can be achieved when the key binding event (mix) modulates signal generation (measure) without the need for separating bound and unbound ligands. Signal modulation can, in principle, be any of several types. For example, the complement to the chemiluinescent labeled partner can be labeled with a quencher or a wavelength-shifting agent or a substance which affects the time course of emission. In practice, such chemiluminescent assays are very difficult to design. It is necessary that two critical components for the reaction be separately linked to the respective binding partners and that they interact to modulate the production of chemiluminescence only when the binding event occurs and not when the two labeled components are both physically present in the assay mixture in the unbound state. The discrimination between the two signal states must be at least 3-4 orders of magnitude.
It is far simpler to design heterogeneous assays in which labeled binding pair complexes are separated from unbound labeled reactants. Most chemiluminescent reactions can be adapted to this assay format, by labeling either with a chemiluminescent compound or with an enzyme and using a chemiluminescent substrate. Most commercially developed immunoassays are of this type.
The Chemiluminescent Reaction
There are many known chemiluminescent reactions to choose from in designing a detection scheme. However, only a fraction of the known reactions have found their way into high volume clinical diagnostic testing methods. Several factors influence whether a particular chemiluminescent technology is suitable for use in automated immunoassays.
- The reaction, of course, should demonstrate an adequate chemiluminescence quantum yield.
- All aspects of the detection chemistry, including labeling methods, labeling reagents, triggering agents and chemiluminescent reagents, must be robust and reproducibly prepared. Methods using enzyme labels face the additional requirement that the enzyme must be reasonably stable and easy to conjugate without diminishing its catalytic activity.
- Ideal chemiluminescent reagents possess an easily measured signal by virtue of an efficient chemiluminescent reaction with a predictable time course of light emission.;
- Analytical performance must be highly reproducible. Since inter- and intra-assay CVs must often be on the order of 3-5%, reagent performance must not contribute more than 1-2% to the CV. Of course, the reagent not be overly prone to interferences and must produce a stable and reproducible signal under laboratory use conditions. Constant reagent performance over a period of 1-2 weeks is important so that on-board calibration curves can be stored.
- Processing and shipping considerations mandate that the detection reagent be capable of bulk manufacture. The reagent must show extended storage stability at refrigeration temperatures or ideally at ambient temperature.
The attractiveness of chemiluminescence as an analytical tool is the simplicity of detection. The fact that a chemiluminescent process is, by definition, its own light source means that assay methods and the instruments used to perform them need only provide a way to detect light and record the result. Luminometers need consist of only a light-tight sample housing and some type of photodetector. Taken to the extremes of simplicity, photographic or x-ray film or even visual detection can be used.
The simple requirements of chemiluminescent methods make them robust and easy to use. But what about sensitivity?
Chemiluminescence has two built-in advantages here, too.
- Most samples have no 'background' signal, i.e. they do not themselves emit light. No interfering signal limits sensitivity.
- Measurement of chemiluminescence is not a ratio measurement in the way fluorescence and absorption or color are. In fluorescence this can lead to difficulties with fluorescers with a small Stokes shift. Fluorescence may not be easy to resolve from the exciting wavelength.
Another problem is associated with scattering of the incident light to the detector, especially when samples are somewhat turbid.
The fundamental factor limiting sensitivity in absorption measurement is the need to measure a small difference in two relatively large signals.
Care should be taken to match the spectral response of the detection device to the chemiluminescence spectrum as closely as possible to maximize sensitivity. The photomultiplier tubes commonly found in luminometers show maximum response to blue light and diminished sensitivity to the red end of the spectrum. Solid state detectors typically have better red response.
X-ray film - X-ray film is widely used to record images of chemiluminescent blotting assays performed on nylon, nitrocellulose or PVDF membranes. The user should bear in mind that x-ray film will only detect visible light in the ultraviolet to blue spectral region, although a specialty film with enhanced green sensitivity is available.
Detection in Solution
Some terms are used regularly in the literature and throughout this website in discussing the use of chemiluminescence in assays: sensitivity, linearity, and dynamic range. The meaning of each is described below.
- Sensitivity refers to the lowest level at which something can be reliably detected. That 'something' is typically an analyte to be detected in an assay. The analyte can be labeled with some detectable tag, such as a chemiluminescent compound or an enzyme. The analyte can also be detected by a specific binding reaction with an affinity binding partner having a label. The lowest 'reliable' level at which some signal is said to still be detected, over a blank test sample, is affected by the sample matrix, the nature of the signal compound and the ability of the detector to repeatably sense low levels of signal.
- Linearity describes the relationship of signal to amount of analyte over a range of concentration of analyte. Ideally the proportionality factor should be constant; a plot of signal vs. analyte would be a straight line. Calibration curves deviating from linearity, e.g. s-shaped or sigmoidal curves, can still be useful.
- Dynamic range is the span of concentration of analyte over which signal varies in a monotonic manner with concentration. This defines the working range of the assay - the wider the better in most cases.
What level of sensitivity can be expected from a chemiluminescent assay? The answer, unfortunately, is 'It depends'. Only in rare cases do limitations of detector sensitivity or the chemiluminescence output set a floor to detectability. Modern detectors can sense vanishingly small intensities. Most often other factors contrive to limit assay sensitivity well above levels imposed by the detector. The most common culprit is non-specific binding of biological components (antibodies, enzymes, etc.) to the surfaces of reaction containers and supports. Virtually all immunoassays, blotting assays, nucleic acid hybridization assays and other enzyme-linked binding assays are limited by this effect.
Tubes and Microwell Plates
Glass and transparent or translucent plastic tubes and cuvettes are suitable containers for light measurement. Ideally light should be measured through a flat surface in order to minimize edge effects. Curved surfaces can be used, e.g. through the bottom of a cylindrical test tube. Comparing results taken using different tubes should be made only after determining that the tubes are manufactured uniformly.
Microwell plates - Light emitted in chemiluminescent reactions is isotropic - it is emitted equally in all directions. If a chemiluminescent assay were conducted in the wells of a transparent microwell plate, light would radiate out not only vertically, in the direction of the detector, but also laterally in the direction of other wells. Light is easily transmitted through the inter-well gaps and through the plate material itself, a phenomenon termed light piping. Relatively bright wells will introduce significant interference in adjacent wells and beyond. For this reason, CHEMILUMINESCENCE SHOULD NEVER BE MEASURED IN CLEAR PLATES.
Opaque microwell plates and strips are commercially available from several suppliers. They come in two kinds - white and black. Users will notice a significant decrease in signal (approx. 10-fold) using the black plates due to light absorption. Since all wells are affected proportionately, regardless of intensity, quantitation is not compromised. The choice of plate should be based on the expected signal strength, white for dimmer reactions, black for brighter ones. Black plates can also be used to advantage to lessen a non-specific binding background problem.
In addition, all white plates are not equally opaque. White plates designed for fluorescence may have dramatically less white pigment than those designed for chemiluminescence. Fluorescence plates are usually only illuminated one well at a time which eliminates almost all inter-well crosstalk. The opacity of a white microwell plate is easily checked by holding a flashlight against the back of the plate and viewing the wells by eye. Plates exhibiting more than a dim transmission of light may be problematic with strongly chemiluminescent substrates.
An Applications Note illustrating the sensitivity of Lumigen APS-5 in the detection of alkaline phosphatase on a microplate luminometer is available from Turner BioSystems.
Detection on Solid Phase Supports
X-ray film - It has been recognized for several years that chemiluminescent detection of immobilized proteins in western blots and immobilized nucleic acids in Southern and northern blots is a powerful combination. Use of several of Lumigen's chemiluminescent reagents in blotting assays is described in our Product Applications section.
Sensitivity is generally more than sufficient for the task at hand, provided that adequate control of non-specific binding is exerted. If not, the power of luminescent reactions typically manifests itself in blackened (overexposed) x-ray films. The best strategy to avoid this problem is to reduce the quantities of binding reagents, i.e. antibodies and the like. Exposure times of 1-10 minutes are usually sufficient to image most blots. Longer times rarely improve signal/background.
On the opposite side of the sensitivity issue, failure to record signal may not necessarily indicate lack of analyte. Photographic film of any type has a threshold intensity below which photochemical conversion of the silver grains fails. This effect is called reciprocity failure. In practice the result is that chemiluminescent signals below a certain intensity simply fail to register. Longer exposures can not overcome the problem. Methods to combat reciprocity failure include gas-hypering, in which the film is soaked in a mixture of hydrogen and nitrogen gas at elevated temperatures for prolonged periods before exposure and pre-flashing, in which the film is flashed with a short duration low intensity light to raise the photon floor in the photographic grains before exposure to the sample. Neither method is convenient and pre-flashing is difficult to do reproducibly.
CCD camera imaging - Relatively recently CCD camera systems have become a competitive tool for obtaining and storing images of chemiluminescent blots. These systems can be superior in dynamic range of signal measurement. Response is linear over 3-4 orders of magnitude vs. about 1 order of magnitude for x-ray film. Imaging time is shorter and multiple images can be obtained and stored easily. Although costly up front, their use eliminates the ongoing cost of purchasing film and film processing, not to mention the cost of the processing unit.
Photomultiplier tubes (PMTs) have traditionally been the workhorse detector in luminometers. Their advantages include good sensitivity, a broad dynamic range and applicability over a reasonably broad spectral range. PMTs are known for their very low dark currents leading to excellent signal to noise for low intensity samples.
PMT based systems operate in two basic modes, single photon counting and current sensing. There are examples of hybrid systems which are single photon counting to a light level in the low millions of photons/second and then switch to current sensing above that level.
PMT single photon counting systems are capable of exquisite sensitivity. Use of this type of detector is the method of choice for low light detection and quantitation as in, for example, detecting the ultraweak luminescence associated with phagocytosis. The greater sensitivity comes at a cost however. Sample housings must be very light-tight. Moderate light levels saturate the detector; high levels can cause irreversible damage to the PMT.
PMT current sensing systems are also capable of excellent sensitivity and will often read higher light levels than single photon counting systems without damage.
There are differing opinions in the chemiluminescence instrumentation field regarding which system is "better", current sensing or single photon counting. In a modern luminometer, both systems achieve excellent sensitivity and are easy to use. A proper understanding of the characteristics of each system should allow the user to choose the one best suited to the application. We use both types to great advantage and our substrates work well with both.
Solid State Detection
Photodiodes are capable of recording higher light intensities than photomultiplier tube detectors. This facet makes them an excellent choice for applications where high light intensities are to be measured. However, the inherent dark current in solid state detectors is generally much higher than that of photomultiplier tubes. One method of mitigating this problem is to cool the solid state detector via a Peltier or other thermoelectric cooler. Dark currents in solid state detectors drop dramatically with temperatures in the 0 to -30 degree celsius range. Cooled detectors can then be used to integrate the light intensity for one to hundreds of seconds without the signal being overwhelmed by dark current.
CCD and other solid state detectors possess several inherent advantages:
- Solid state detectors typically offer a "flatter" optical response over the visible range. Luminescent reactions emitting red and even near infrared light can be detected with enhanced sensitivity.
- CCD camera systems allow imaging of a variety of objects. Virtually any kind of sample or container can be accommodated ranging from microwell plates and test tubes to bacterial or cell cultures in Petri dishes, electrophoresis gels and blotting membranes.
- Single PMT systems must have the sample position well defined before it can be read. A sample tube has to be brought to a reproducible position to be read repeatably. Microwell plate PMT readers rely upon the standard spacing of microwells and will generally move the plate around to a precalculated position so the wells can be read one by one. Camera systems have the advantage of being able to read a sample without knowing its position in advance, as in the example of a band on a blot. The camera imaging system gives positioning information along with sample intensity.
- CCD camera systems allow imaging of numerous objects simultaneously. In the present era of 96, 384 and higher number well plates, parallel data collection is no longer a luxury. Solid state camera imaging systems have the potential to permit imaging and quantitation of entire plates in one pass.
Imaging detectors have a potential serious disadvantage for the quantitation of microwell plates if the instrument designer or end-user is not aware of it. The vignetting problem arises from the 3-dimensional nature of the plate relative to a simple 2-dimensional target, such as a blot. Since light intensity drops off as the reciprocal of the square of the distance to the plate, it would be logical to assume that placing the camera as close to the plate as possible would yield the best result. However, since the well has depth, the entire bottom of the outermost wells might not be visible. This geometric cut-off of part of the emitting well will result in less light being detected coming from the well.
Raising the camera will mitigate the situation somewhat at the expense of a lower overall intensity at the detector. Geometric weighting patterns can be applied to correct for the edge intensity loss, however, the signal to noise ratio for the outer wells must be worse than for the central wells.
Most research luminometers and luminometric immunoassay analyzers incorporate a number of highly useful auxiliary components. Sample heaters/coolers and reagent injectors facilitate the chemical processes leading to light production. Components for optical filtering are often provided, while structural features for rejection of stray light such as light baffles and cut-off switches, are mandatory. The benefits of some of the more important accessories are:
- Temperature control - Comparing results within and between days is sometimes confounded by temperature variability in the samples. Particularly in the case of enzyme-catalyzed "glow" type reactions which require several minutes to reach a plateau intensity, temperature fluctuation can lead to changes in measured intensity, detracting from analytical precision. Thermostatted sample blocks and plate heaters can help provide uniform temperatures and permit running reactions at elevated temperatures. The variation of sample intensity can be caused by several effects including the temperature dependent kinetics of the enzymatic reaction or subsequent substrate chemical kinetics. Subtle effects, such as the pH shift of buffers with large temperature coefficients, can cause significant signal variations.
- Monochromators and optical filters allow the isolation of specific wavelengths or ranges. Generally their use is not needed except in specialized applications. Assays and protocols featuring two luminescent species emitting at different regions of the spectra are sometimes used to detect two different analytes. Protocols using a fluorescent acceptor energy transfer compound in conjunction with a chemiluminescent emitter are useful to probe binding processes. Wavelength selectivity comes at a cost though since any device that reduces the range of wavelengths of light reaching the detector inevitably decreases sensitivity by decreasing light throughput.
- Neutral density filters are useful optical elements for extending the range of light intensities measurable by a luminometer by 2-3 orders of magnitude. Consisting of a piece of special grayish glass, and fitting between the sample holder and the detector, these filters function like "sunglasses" to diminish essentially all wavelengths by a known factor. Measured light intensities are corrected to the "true" intensities by applying a correction factor.
- Injectors allow the introduction of substrates or triggers at precise times. This can be important when a kinetically controlled sample has a time varying light curve. Some substrates can be read on the grow-in portion of the curve and will yield accurate results only if all wells or tubes are read at exactly the same time after substrate or trigger introduction. A caveat to injector use arises from the extremely high sensitivity of chemiluminescent measurements. If the injector system becomes contaminated or if different substrate systems are to be used on the same instrument, the injector system can be very difficult to clean thoroughly. Some substrates can detect as little as a few hundred molecules of their trigger agent. Cleaning the pistons, syringe barrels, tubing and valve systems to this level is very difficult without complete disassembly and autoclavability.
General discussions of chemiluminescence measurement can be found in these articles.
- J.E. Wampler, Instrumentation: Seeing the Light and Measuring It, in Chemi- and Bioluminescence, J.G. Burr, ed., Marcel Dekker, New York, 1-44 (1985).
- A.K. Campbell, Detection and Quantification of Chemiluminescence, in Chemiluminescence Principles and Applications in Biology and Medicine, Ellis Horwood, Chichester, 68-126 (1988).
- F. Berthold, Instrumentation for Chemiluminescence Immunoassays, in Luminescence Immunoassay and Molecular Applications, K. Van Dyke and R. Van Dyke, eds., CRC Press, Boca Raton, 11-25 (1990).
- T. Nieman, Chemiluminescence: Theory and Instrumentation, Overview, in Encyclopedia of Analytical Science, Academic Press, Orlando, 608-613 (1995).