Flow cytometry teardown

Anatomy of a flow cytometer

A cytometer is a microscope that refuses to look at a picture. It runs cells past a laser focus one at a time and counts photons per event, so every choice in the instrument is really a choice about how many photons one cell can produce before it leaves the beam. Pick a part from the bill of materials and drop it on the cutaway; every parameter comes with the catalogue part numbers that carry it.

How to use this page

1Pick a part from the bill of materials.
2Drop it on the module, or just click the card.
3Read its job, its parameters and the part numbers behind them.

Bill of materials

Every part, one optical budget. Drag a card onto the photo, or click it - the panel beside it fills in with that part's parameters.

A teaching cutaway of a flow cytometer: three fibre-coupled lasers on the left, the sheath and sample fluidics above them, the flow cell at the centre where the beam, the cell stream and the collection optics meet, the dichroic and filter ladder with its detectors on the right, and the acquisition board along the bottom

Drag a part here or click a card in the bill of materials

A teaching cutaway, not an assembly drawing. Lasers on the left, fluidics above them, the flow cell and its interrogation point at the centre, the filter and detector ladder to the right, the acquisition board below.

Component positions are a teaching diagram, not an assembly drawing: the placement and the exploded view are indicative. Parameter values are those of the linked specification.

Every part, parameter by parameter

The same numbers the interactive view shows, in tables you can read, search and quote.

405 nm violet laser Excitation source

Sets how many photons a violet-labelled cell can emit. What matters is not the headline power but the power that survives the fibre, whether it holds steady over a run, and where the line sits relative to the filters: a 2 nm wide line at 405 nm drifts 0.3 nm per kelvin, so case temperature moves the excitation out of the dye's absorption band long before anything in the optics moves.

405 nm violet laser - Single-mode fibre-coupled | CW
Center wavelength	405 nm
Output power, CW	70 mW
Power offered on this line	70 mW \| 80 mW
Spectral width (FWHM)	2 nm typ
Wavelength temp. coefficient	0.3 nm/K
Fiber core diameter	3 um
Numerical aperture	0.13
Connector	FC/PC, SMA905, FC/APC
Dyes this line is chosen for	BV421, Pacific Blue, DAPI (examples)
Usual emission windows	450/50 nm, 525/50 nm (examples)

488 nm blue laser Excitation source

Sets the photon budget for every blue-excited channel at once. Because three windows share one line, the power is split between them and the dimmest dye decides the specification: the catalogue's 25 mW line is enough for a well-labelled panel, and the limit on the channel is usually the filter and the detector rather than the laser.

488 nm blue laser - Single-mode fibre-coupled | CW
Center wavelength	488 nm
Output power, CW	25 mW
Spectral width (FWHM)	2 nm typ
Wavelength temp. coefficient	0.3 nm/K
Fiber core diameter	3 um
Numerical aperture	0.12
Connector	FC/PC, SMA905, FC/APC
Dyes this line is chosen for	FITC, Alexa Fluor 488, PE, PerCP (examples)
Usual emission windows	530/30 nm, 585/42 nm, 695/40 nm (examples)

638 nm red laser Excitation source

Sets how far into the far red the panel can reach. Specifications usually quote a 640 nm line; 638 nm is the line the catalogue carries and it sits inside the same absorption band, so a filter stack built for 640 nm works unchanged. What the red line does not tolerate is a dark enough detector: the 660/20 and 780/60 windows are wide, and the limit on the channel is the detector's dark current, not the excitation.

638 nm red laser - Single-mode fibre-coupled | CW
Center wavelength	638 nm
Output power, CW	70 mW
Power offered on this line	70 mW \| 100 mW
Spectral width (FWHM)	2 nm typ
Wavelength temp. coefficient	0.3 nm/K
Fiber core diameter	4 um
Numerical aperture	0.13
Connector	FC/PC, SMA905, FC/APC
Dyes this line is chosen for	APC, Alexa Fluor 647, APC-Cy7 (examples)
Usual emission windows	660/20 nm, 780/60 nm (examples)

Fibre delivery Optical path

Sets the spot at the interrogation point and whether the three lines are co-linear there. A single-mode pigtail's output is already defined by the fibre: core diameter and numerical aperture fix the divergence, and the launch objective turns that into a waist of the size the flow cell needs. If the three lines are not co-linear to within a few micrometres, a cell labelled for one dye sees a different excitation intensity from a cell labelled for another.

Fibre delivery - Single-mode pigtail | combiner and launch optics
Lines delivered	405 / 488 / 638 nm
Fiber type	single-mode mode
Fiber core diameter	3 um
Numerical aperture	0.12
Fiber length	100 cm (110 cm max)
Connector	FC/PC, SMA905, FC/APC
Min. bend radius	>= 300 x fiber diameter
Launch	one objective shared by the three lines
Co-linearity at the interrogation point	set by the launch assembly you build Set by your design

Sheath and sample fluidics Fluidics

Sets the dwell time, and the dwell time sets the photon budget. A cell crossing a 10 um spot at 1 m/s is in the beam for 10 us; double the flow rate and every event loses half its photons, which only the laser can give back. This is the coupling that decides whether a 25 mW line is enough or whether the design needs the top of the range.

Sheath and sample fluidics - Hydrodynamic focusing | sheath, sample, waste
Dwell time in the beam	10 us across a 10 um spot at 1 m/s
Core diameter at the interrogation point	set by the nozzle and the sheath-to-sample ratio Set by your design
Flow rate	set by the event rate you need Set by your design
Sample and sheath pressure	set by the fluidics you build Set by your design
Waste handling	set by the assay and the site Set by your design

Flow cell Interrogation

Sets the geometry the rest of the optics is built to. The cuvette's wall thickness and surface quality decide how much excitation light is scattered into the collection path; its position fixes where the waist and the objective have to land; and it is the part that is hardest to change late, which is why the flow cell is usually chosen before the optics around it.

Flow cell - Quartz cuvette | one laser focus
Excitation lines focused into it	405 / 488 / 638 nm
Interrogation point	one point, all three lines co-linear
Waist diameter	set by the launch objective Set by your design
Cuvette geometry and wall thickness	set by the flow cell you source Set by your design
Stray excitation light reaching the collection path	set by the cuvette and the beam dump Set by your design

Collection optics Optical path

Decides which dye is seen by which detector. A bandpass filter is written as centre over width - 530/30 passes 515 to 545 nm - and the window has to sit on the dye's emission peak, never on the laser line. The dichroic that separates 405 nm from violet-triggered fluorescence is a different part from the one that separates 638 nm, which is why the excitation lines are chosen before the filter stack is laid out.

Collection optics - Objective, dichroics, bandpass filters
Collection angle	90 deg to the beam
Typical bandpass windows	450/50, 530/30, 585/42, 695/40, 780/60 nm
Excitation rejection	set by the filter stack you build Set by your design
Dichroic split wavelengths	set by the dye panel Set by your design
Long-pass cut for the tandem channels	set by the tandem dyes you run Set by your design

Detectors Detection

Gain, dark current and excess noise decide the dimmest dye the panel can carry. The catalogue's silicon APDs respond from 400 nm to 1100 nm, which spans the whole visible fluorescence band; whether one of them takes the place of a photomultiplier in a given channel is a question of photons per event, not of whether the technology fits the wavelength.

Detectors - Photomultiplier or avalanche photodiode
Response spectrum	400 - 1100 nm
Active diameter	230 um
Responsivity	80 A/W @ M=100
Dark current	0.4 nA @ M=100
Optimum gain	100
Response time	0.5 ns
Reverse breakdown voltage	100 - 200 V
Voltage temp. coefficient	1.1 V/C
Channel count	set by the dye panel Set by your design

Signal acquisition Electronics

Sets the noise floor and the dynamic range, which together decide whether the dimmest channel in the panel is real signal or the amplifier talking. It also has to hold each detector's bias: the APD channels run at 100 - 200 V with a 1.1 V/K breakdown coefficient, so a bias loop without temperature compensation drifts the gain out from under a calibration that was set at 25 C.

Signal acquisition - Analogue front end, digitiser, detector bias
Dwell time the front end has to cover	10 us across a 10 um spot at 1 m/s
APD bias	100 - 200 V
APD bias compensation	1.1 V/C
Channel count	set by the dye panel Set by your design
Sample rate	set by the dwell time you design for Set by your design
Dynamic range	set by the brightest and dimmest dye in the panel Set by your design

Housing Mechanical

Thermal design, mostly. Three numbers decide it: the excitation lines drift 0.3 nm per kelvin, a bandpass window is tens of nanometres wide, and the APD's breakdown voltage moves 1.1 V per kelvin. Ten kelvin of case rise spends 3 nm of a 30 nm window and 11 V of a 100 - 200 V bias; neither is fatal, but both are only recoverable if the chassis makes them predictable.

Housing - Chassis, optical bench, thermal path
Wavelength drift of the sources	0.3 nm/K on all three lines
Window consumed by a 10 K case rise	3 nm of a 30 nm bandpass
APD bias drift	1.1 V/K on the APD channels
Bench stiffness	set by the alignment you have to hold Set by your design
Shielding and EMC	set by the front-end noise floor Set by your design

Every parameter value on this page is pulled from the catalogue record that renders the linked product page, so the two cannot disagree.

Parts behind these parameters

Every figure below is read from the same catalogue record that renders the product page. Nothing on this page is retyped.

The three lines this page is built on

405, 488 and 638 nm, all single-mode pigtailed and all in the same coaxial can. The dye panel decides which of them you need; the dwell time and the detector decide how much power the dimmest channel has to see.

All fibre-coupled modules

LZFL405-70mW-3/125-FC/PC 405 nm The violet line feeds the shortest-wavelength windows and the dyes that put out the fewest photons, which is where its 0.3 nm/K drift matters most. Wavelength405 nm CW output70 mW Core3 um LZFL488-25mW-3.5/125-FC/PC 488 nm The blue line is the one most panels cannot do without: FITC, PE and the PerCP tandem channels all hang off this single excitation. Wavelength488 nm CW output25 mW Core3 um LZFL638-70mW-4/125-FC/PC 638 nm The far-red line closes the panel: APC and Alexa Fluor 647 need it, and their channels are usually the dimmest the design has to resolve. Wavelength638 nm CW output70 mW Core4 um

The lines either side

A panel rarely stops at three colours. 450 nm is the shorter neighbour of the violet line, 520 nm is the nearest green to the 561 nm line most panels specify, and 660 nm reaches further into the far red than 638 nm does.

All fibre-coupled modules

LZFL450-30mW-3.5/125-FC/PC 450 nm A violet-blue line, for panels that want the violet dyes excited a little further from their absorption edge. Wavelength450 nm CW output30 mW Core3 um LZFL520-50mW-3.5/125-FC/PC 520 nm The nearest green we carry to a 561 nm line: PE and its tandems are excited less efficiently here, and the trade is a cheaper, longer-lived source. Wavelength520 nm CW output50 mW Core3 um LZFL660-30mW-4/125-FC/PC 660 nm Twenty-two nanometres further into the far red than the 638 nm line, and inside the same APC band. Wavelength660 nm CW output30 mW Core4 um LZFL638-100mW-4/125 More power The red line at 100 mW, for the channel that will still be the dimmest one after every optical trick has been used. Wavelength638 nm CW output100 mW Core4 um

The detection side

Silicon APDs with a 400 - 1100 nm response cover the whole visible fluorescence band. Whether one of them replaces a photomultiplier in a given channel is a question of photons per event, gain and dark current - not of whether the technology fits the wavelength.

All APD detectors

LCC2-650APD-230-0.5 650 nm The visible-band part: a 650 nm window on a 230 um active area, and a 400 - 1100 nm spectrum that reaches every dye on this page. Wavelength650 nm Spectrum400 - 1100 nm Responsivity80 A/W @ M=100 To46-905APD-230 TO-46 The same die in TO-46, catalogued with a ranging filter and orderable without it - the cheapest way onto a solderable can. Wavelength905 nm Spectrum400 - 1100 nm Responsivity80 A/W @ M=100 16-ch-APD-Array Array Sixteen elements sharing one bias, for a parallel spectral or spatial channel that has to fit behind the same filter stack. Wavelength905 nm Spectrum400 - 1100 nm Responsivity80 A/W @ M=100

Building the filter stack

Aligning a four-colour collection path is easier with a beam you can see. These are the visible heads we keep for the bench, and the wavelengths they stand in for.

Collimated outputs

LZRGB-638-Collimated 638 nm A collimated red head at the same wavelength as the far-red line: the usual alignment beam for the red end of the stack. Wavelength range632 - 644 nm Rated outputCW 500 mW - 1 W ModeCW LZRGB-450-Collimated 450 nm The blue head, for the violet side - visible light where the dichroics have the least room to work in. Wavelength range430 - 463 nm Rated outputCW 5 - 5.5 W per emitter ModeCW LZRGB-520-Collimated 520 nm Green reads best under bright light, and it is the one that shows up cleanly on a card behind a 585/42 window. Wavelength range510 - 532 nm Rated outputCW 1 - 1.5 W per emitter / 13.5 W bank ModeCW

Three fibres do most of the work

Filter the catalogue by the fibre-coupled category and the three lines this page is built around come back as single-mode pigtailed modules: 405 nm for the violet dyes, 488 nm for FITC and PE, 638 nm for APC and the far-red tandems. The selector has no cytometry application yet - send us the panel of dyes you are building for and we will work back from the emission windows to the lines and the powers.

Browse the fibre-coupled lines Open the selector Fibre coupling and beam shaping

Send us the dye panel, not a part number

Every line on this page was chosen against an emission window. Tell us which dyes your panel has to resolve and we will come back with the excitation wavelengths, the powers they need and the filters they imply.

Request a quote Chat on WhatsApp APD vs PIN receivers