As an analysis platform, flow cytometry relies on interrogation of individual cells by laser light and the collection of the resulting fluorescence and scatter. The optics system handles illumination and light collection within the instrument.
Note: If you are not familiar with basic fluorescence concepts or with filters and light sources in general we highly recommend that you review the Molecular Probes School of Fluorescence—Fluorescence Basics module. This will help you better understand the content discussed in this section.
The components of the optical system work in concert to shine different wavelengths of light onto the cell, collect the data (i.e. side and forward scatter as well as emission from the excited fluorophores) in the form of emitted photons and convert these photons to an electrical signal—a photocurrent—that goes into the electronics system. The components of the optical system include excitation light sources, lenses, and filters used to collect and move light around the instrument, and the detection system that generates the photocurrent.
The interaction of the cells with the laser occurs in the interrogation point. Here the cells pass in single file where the focused excitation light crosses through the flow stream.
The most common excitation light sources in flow cytometry instruments are lasers. Laser light is coherent (has a synchronized, identical wave frequency), monochromatic (has a single wavelength), and energetic—properties that ensure that the cells are illuminated with uniform light of a specific wavelength. Flow cytometers typically contain one or more laser lines, and Table 1 lists the most common excitation lines available on today's instruments. Figure 1 displays the available lasers across the spectrum.
|Laser line||Wavelength||Common fluorophores used with laser|
|Ultraviolet (UV)||355 nm||DAPI, Hoechst, LIVE/DEAD Blue, Brilliant Ultraviolet|
|Violet||405–407 nm||Pacific Blue, eFluor 450, Pacific Orange, eFluor 506, Super Bright 436, Super Bright 600, Brilliant Violet, LIVE/DEAD Yellow, LIVE/DEAD Aqua, LIVE/DEAD Violet, CFP|
|Blue||488 nm||FITC, Alexa Fluor 488, Dylight 488, PE, PE tandems, PerCP, PerCP tandems, PI, 7AAD, eGFP, YFP|
|Green||532 nm||PE, PE tandems, Alexa Fluor 532, PI, mCherry, dTomato, RFP|
|Yellow||561–568 nm||PE, PE tandems, PI, mCherry, dTomato, RFP|
|Red||633–647 nm||APC, Alexa Fluor 647, Alexa Fluor 700, APC tandems|
Figure 1. The UV-visible spectrum. Lasers with discrete wavelengths in the UV-visible spectrum are used to excite fluorophores.
Knowledge of excitation light sources is crucial for making decisions concerning the fluorophores you may use in your experiment. It’s also important to understand how these light sources are configured in your own flow cytometer. There are two main types of arrangements: parallel laser arrangements and co-linear laser arrangements (Figure 2). In a parallel arrangement, the lasers are spatially separated so that the cells are exposed to one excitation source at a time as they pass through the interrogation point (see Figure 3 for an example set-up). In a co-linear laser arrangement, the lasers share the same optical pathway, and the cells are excited by multiple lasers at the same time (see Figure 4 for an example set-up).
It is important to note that these arrangements are not mutually exclusive in that one instrument could have both parallel and co-linear laser arrangements.
Figure 2. Comparison of parallel and co-linear laser arrangements. (Left) In the co-linear arrangement, multiple lasers share the same optical path and the cell in the interrogation point will be excited by those lasers at the same time. (Right) In the parallel arrangement, the lasers do not share the same optical path and they will excite the cell at different times within the interrogation point.
Figure 3. Parallel laser arrangement example. The 488 nm (blue) and the 633 nm (red) laser light beams are placed at different locations in the interrogation point, and the cell will pass by each laser at different times. In this set-up, there are separate optical paths for each laser including distinct emission filters and detectors.
Figure 4. Co-linear laser arrangement example. In this co-linear layout both lasers are located at the same point within the flow cell and the light from both lasers hits the cell at the same time. The optical path for both lasers is shared in that both use the same emission filters and detectors.
Optimal fluorophore choices can differ for these two arrangements, so it’s really important that you understand your instrument’s setup before you plan your experiment. For example, in the parallel laser arrangement shown in Figure 3 where the lasers excite the same cell at different times and each laser has its own discreet detector, you might consider using two fluorophores that have the same emission but are excited at different wavelengths. One such combination would be phycoerythrin–Alexa Fluor 647 (PE-AF647) and allophycocyanin (APC) (Figure 5). PE-AF647 is excited at 488 nm and APC is excited at 633 nm. Both dyes emit in the 680 nm region. The PE-AF647 tandem dye would be excited only by the 488 nm laser and the emission light would be collected by the detector designated for that laser line. On the same cell, the APC dye would be excited only by the 633 nm laser after the first dye was excited and the emission would be collected by its designated detector. Because the lasers are spatially separate, the emission overlap is not an issue. If you were using the co-linear laser arrangement shown in Figure 4, however, both dyes would be excited at the same time and place by both 488 and 633 nm lasers and the emission by the different dyes at the same wavelengths would be indistinguishable.
Figure 5. Excitation and emission spectra for PE-AF647 and APC. (A) The excitation spectra for PE-AF647 (red dotted line) and APC (orange dotted line) are shown. The excitation lasers at 488 nm (blue solid vertical line) and 633 nm (orange solid vertical line) are also indicated. (B) Emission spectra overlap for PE-AF647 (orange curve) and APC (red curve) is shown.
In a parallel laser arrangement, there is a time delay inherent in the system. This delay represents the amount of time it takes a cell to move from one laser to the next laser in the relay (see Figure2,right image) and it allows the signals collected from that cell after excitation by all of the lasers in the interrogation point to be delivered to the electronics system all together.
Although time delay is typically set automatically by most instruments, it can be monitored and adjusted manually if needed. If this time delay is not set correctly, you may observe a loss of signal, or worse, a mix-up of signals from two different cells. The results of correct and incorrect time delay settings are shown below (Figure 6). The fluorescent particles used in this example have equal amounts of two different fluorophores, one that excites at 488 nm and emits green fluorescence and one that excites at 561 nm and emits red fluorescence. You would expect that the number of events would be equal to one another since they are on the same particle and that the pattern of their emission would be similar as well if the time delay is set appropriately. Figure 6A shows the data from such an experiment. In Figure 6B, the data starts out looking similar to Figure 6A but the operator adjusted the time delay setting around the 30 time point and again at the 50 time point. Notice how the fluorescence of the downstream laser (ie the second laser to interrogate the cell in the relay) is lower compared to the trigger laser (first laser in the relay) and the spread of the events has gotten broader. This is something you will want to remember if working on an instrument with a parallel laser arrangement.
Figure 6. Identifying time delay issues in parallel laser arrangements. In both examples, the particles are excited at both 488 nm and 561 nm in an instrument with a parallel laser arrangement. (A) With appropriate time delay settings the fluorescence intensity and the spread of the fluorescence events are very similar with both lasers. (B) In this experiment, the time delay was started at the correct setting but was adjusted during the run. As you can see, there is a shift in the fluorescence events intensity and spread on the downstream laser as compared to the trigger laser. This indicates the incorrect time delay setting.
There are a variety of options for emission filters and mirrors that guide the path of the light (photons). These are summarized below and in Figure 7. Each set of filters serves to direct the specific wavelengths of light to their matching detector (these are described in the next section). The mirrors help direct the light to the detector as well.
The filters, mirrors and detectors in your instrument will need to be manually assigned to the fluorophores you choose, so it is important to know what your set-up is when you design your experiment and when you analyze your results. And to make matters a little more confusing, by convention, flow cytometrists will use the term “detectors” when referring to all three of these components combined (emission filters, mirrors and the actual detector itself). They are all so closely tied together and have to be matched to one another in order for them to work properly.
Figure 7. Examples of emission filter types.(A) This is a longpass (LP) filter that allows all light above a specific wavelength to pass through. The colored lines indicate which wavelengths of light are deflected or allowed to pass through the filter. The graph below the filter is a plot of the light passing through the filter versus the wavelength of the light. (B) This is called a shortpass (SP) filter which means that it will only allow light below a specific wavelength to pass through the filter. (C) This is a bandpass filter that is able to limit wavelengths below a low end and above a high end of the indicated wavelengths.
Longpass (LP) filters allow all light above a specific wavelength to pass through. In the example in Figure 7A, it is an LP 500 filter, indicating all light above 500 nm will pass through the filter. Note that the UV and violet wavelengths are deflected while the blue, green and yellow light is permitted to pass through.
Shortpass (SP) filters allow all light below a specific wavelength to pass through. In the example in Figure 7B, it is an SP 500 filter which means it will only allow light below 500 nm to pass. The UV, violet, and blue light is allowed to pass through while the higher wavelength light, green and yellow, is deflected.
Bandpass (BP) filters can be thought of as a cross between LP and SP filters. It will allow only light within a specific range to pass through the filter. The example in Figure 7C is of a 520/20 BP filter; that means the “band” of light allowed through the filter is 10 nm below 520 nm and 10 nm above 520 nm or put another way only light that falls within the range of 510–530 nm can pass. All other light will be deflected.
Dichroic mirrors (also called dichroic beamsplitters) are forms of LP and SP filters that contain a mirror coating. In addition to passing light above a specific wavelength (LP) or below a specific wavelength (SP), dichroic beamsplitters reflect the light in a certain direction (Figure 8). For example, a 500LP dichroic mirror would transmit light above 500 nm and reflect the light below 500 nm in a different direction. A 525SP dichroic mirror would transmit all light below 525 nm and reflect all light above 525 nm in a different direction. These dichroic mirrors are critical in the directing and capturing of light by the detectors.
Figure 8. Dichroic mirrors. The dichroic mirrors (or beamsplitters) in this example are each associated with a specific detector. The first dichroic mirror (1) deflects the red light and redirects it to the red detector while allowing the light from all wavelengths below the red wavelength to pass through to the next dichroic mirror in the path. That second dichroic mirror (2) deflects the yellow light to the second detector and allows the blue and green light to pass through. The last dichroic mirror (3) deflects the green light to the third detector and allows the blue light to pass.
Detectors capture the photons that are emitted by the excited fluorophores and scattered laser light, and convert them into photocurrent (which is really just another way of saying electric current generated from photons) which is then passed to the electronics system. There are several detector types that can be used on a flow cytometer, the most common types being the photodiode (PD) and the photomultiplier tube (PMT). The avalanche photodiode (APD), traditionally used in fiber optic telecommunication, is starting to make an appearance on some flow cytometers and is especially good for the detection of long-wavelength emissions (>650 nm). There are some instruments that use a charged-coupled device (CCD) camera for detection, but it is not commonly used and we will not discuss these in this section.
When a photon hits a PD, it ionizes the atoms of the detector, creating an electron/hole pair within the depletion region of the PD (Figure 9A). The electrons in the depletion area are swept towards the positive potential on the cathode, and the holes are swept towards the negative potential on the anode, creating a photocurrent which is then transferred to the electronics system. PDs are inexpensive but have low sensitivity because they do not amplify the photocurrent as much as a PMT does. PDs are usually used in the channel responsible for handling the brightest amount of light or signal, like the forward scatter channel.
Figure9. Detectors used on flow cytometers. (A)Schematic of photodiode detector. The purple circles represent the holes created when the photon interacts with the p region of the PD and generates the electron (green circles). These electron/hole pairs are represented by the purple and green circles connected with the red line between them. The negatively charged electrons are attracted to the positively charged cathode while the holes are attracted to the anode. (B)Schematic of a photomultiplier tube. The photon is converted to an electron when it enters the PMT at the cathode. It then travels through the PMT, being amplified at the dynodes (electrodes) throughout and end up at the anode which is a collecting electrode.
A PMT is much more sensitive to lower levels of signal and can amplify the energy from a single photon many millions of times. They are typically used to measure fluorescence from excited fluorophores from the cells as well as side scatter signal. As photons enter the PMT they strike a photocathode thereby generating electrons (Figure 9B). These electrons move from dynode to dynode generating secondary electrons resulting in an amplification of the signal. Not all PMTs are created equal, nor are all photons. The sensitivity of the PMT to a given wavelength of light is highly dependent on the material that the PMT is made of. Also, photons in the far-red region typically have less energy, and when they enter the PMT, fewer of them generate a photocurrent output.
PMT sensitivity is also controlled by the amount of voltage applied to it, and this needs to be optimized for a given PMT on a given instrument with a given configuration. One of the most common methods for doing this is the ‘Peak 2’ method (Maecker and Trotter (2006)). With this method, a dim fluorescent particle is run using a series of different voltage settings (sometime called a voltration), and the spread of the signal (or the coefficient of variation, CV) is plotted over the voltage series. It is important to run this in all emission channels that you’ll be using, since the optimal voltage settings for each may vary.
An example voltration experiment is shown in Figure 10. We are only showing the data for a single fluorescence channel (in this case it is the green channel using fluorescein labeled particles) but as noted above, you would need to run this experiment for all of the channels you will be detecting in your experiment. Figure 10A shows the fluorescence events collected over time at increasing PMT voltage settings, ranging from 200 mV to 400 mV. You can see in the data that the spread of these events is very broad at the lowest voltage of 200 mV (which is another way of saying the variation of the events is much larger at the lower voltages than the variation that you see with the higher voltage runs). You will also notice that as the voltage setting is increased, the variation of the data gets progressively smaller and actually reaches a point where it doesn’t look different from the next voltage setting (350 mV is virtually indistinguishable from 400 mV). If you plot the percentage of the CV (%CV) of each data set against the voltage setting (Figure 10B), you can see this more clearly. The 200 mV setting gives a large %CV but by the time you get to the range of the 300–400 mV, %CV does not change very much at all. The optimal voltage setting is one in which the %CV or noise is at its lowest possible point and the lowest voltage setting that gives you the smallest CV. This is called the inflection point. In this example, you would probably set the voltage of your PMT for the green channel at 350 mV since going higher in voltage has no impact on the CV.
Figure 10. Optimzing PM sensitivity. Fluorescein labeled particles were run at increasing PMT voltage settings. (A) The fluorescence events in the green channel were detected at different PMT voltages (indicated in the plot for each data set) and plotted against time. (B) The percent of the Coefficient of Variation (%CV) for each data set in panel A was plotted against the PMT voltage setting. The point on the curve where the %CV begins to level out, demonstrating decreasing variation of the data at the higher voltages, is called the inflection point, indicated by the blue arrow. The red arrow indicates the optimal PMT voltage range for this fluorescence channel.
The optical elements of the flow cytometer direct the photons emitted by the fluorophores (bound to the cells) to the detection system. This is done in a highly controlled manner so that you know the specific wavelengths of light are going to each detector when you collect your data. The detectors convert the signal photons from scattered light or fluorescence to an electric current that is proportional to the number of photons hitting the detector. The resulting pulse is transferred to the electronic system, where it is digitized and recorded for analysis.
Not for resale. Super Bright Polymer Dyes are sold under license from Becton, Dickinson and Company.
For Research Use Only. Not for use in diagnostic procedures.