Specifying a Raman spectroscopy laser starts with the excitation wavelength, and the choice is a trade rather than a lookup. Pick too short and fluorescence can swamp the spectrum. Pick too long and you give away signal and detector sensitivity. The sections below give the reasoning and the usual starting points, not a table you can apply without testing.

Signal against fluorescence

	Shorter wavelength	Longer wavelength
Raman signal	stronger, non-resonant scattering scales roughly with ν⁴	weaker
Fluorescence background	often higher, but sample dependent	often lower, not guaranteed
Detector	silicon CCD, sensitive and quiet	InGaAs beyond roughly 1000 nm, noisier
Acquisition time	usually shorter	usually longer
Spatial resolution	finer focus	coarser focus

In the simple non-resonant case, Raman intensity scales roughly with the fourth power of the excitation frequency. Comparing 532 nm with 1064 nm gives an idealised factor of about sixteen. Treat that as a comparison rule, not a prediction of measurement time: fluorescence, optical throughput, laser power and detector response all enter the real result.

Where fluorescence does dominate, it raises the shot noise floor across the spectrum, and shot noise cannot be removed by baseline correction. Longer integration still improves signal to noise, roughly with the square root of acquisition time, but from a much worse starting point. In practice the useful limit is usually set by detector saturation, photobleaching, sample heating or drift rather than by the background itself.

Wavelength by wavelength

405 nm

Largest Raman cross section of the common lines, and the usual choice for resonance Raman when the sample has an electronic transition nearby. On many organic materials the fluorescent background is prohibitive, but this is not a rule: resonance effects mean some materials give a better result at 405 nm than at longer wavelengths. See our 405 nm laser module.

488 and 532 nm

532 nm is the most widely used line in materials work. Strong signal, well matched to silicon detectors, and a large body of published reference spectra to compare against. Common for minerals, carbon materials, catalysts and semiconductors, though some inorganic samples fluoresce strongly as well. 488 nm behaves similarly with slightly more signal. See our 532 nm laser module.

633 and 660 nm

A middle ground that often reduces fluorescence against 532 nm while keeping more signal than 785 nm. Used for pigments, dyes, some biological samples and polymer formulations. Whether it helps on your material is an empirical question. See our 633 nm laser and 660 nm laser module.

785 nm

The most frequently specified single line in commercial instruments for organic chemistry, pharmaceutical analysis, forensics and biological tissue. It is a common default when the full range of samples is not yet known, which is not the same as being the safest choice for every material. See our 785 nm laser module.

830 and 1064 nm

For samples where fluorescence has defeated the shorter options. 1064 nm, usually in an FT-Raman configuration, suppresses fluorescence strongly but does not eliminate it in every case, and it requires InGaAs detection.

One detail that is easy to miss at 830 nm: the excitation line still sits inside silicon sensitivity, but the Raman scattered light does not stay there. A shift of 3000 cm⁻¹ from 830 nm lands near 1090 nm, already past the useful silicon range. Check where your bands of interest fall, not just the laser line. See our 1064 nm laser.

Starting points by sample class

Rules of thumb for a first experiment, not a classification. Fluorescence behaviour is specific to the material, and exceptions in both directions are common.

Sample class	Usual first try	If that fails
Minerals, inorganics, oxides	532 nm	405 nm for more signal, 633 nm if it fluoresces
Carbon materials, graphene, CNT	532 nm	488 nm
Catalysts, semiconductors	532 nm	633 nm
Pigments, dyes, artwork	633 nm	785 nm
Polymers, plastics	785 nm	1064 nm if filled or coloured
Pharmaceuticals, APIs	785 nm	1064 nm
Biological tissue, cells	785 nm	830 nm
Dark or unknown organics	1064 nm	830 nm

Excitation sources Raman laser modules from RGB 405 to 1064 nm · free space or fiber coupled · stabilized versions available View range

What else the source has to deliver

Parameter	What it affects	How to decide
Linewidth	Can broaden measured bands if it is not clearly narrower than the feature of interest	Compare against your spectrometer resolution and your narrowest band, not against a fixed number
Wavelength stability	Shifts the position of the Raman axis over time and temperature	State the period and conditions. At 785 nm, 0.015 nm corresponds to about 0.24 cm⁻¹
Spectral purity	Side modes and ASE raise the baseline near the laser line	Ask for a side mode suppression figure over the range you measure
Power stability	Relative peak intensities between measurements	Specify the measurement period and where it is measured

Linewidth and wavelength stability are separate specifications and are often confused. The first describes the spectral width of the line, the second how far its centre drifts over a stated period and temperature range. A drift of 0.24 cm⁻¹ matters for high resolution work, but it is not automatically an uncorrectable error: calibration against a reference standard addresses a good part of it.

How much power

• There is no universal figure. What matters is power density at the sample, which depends on spot size as much as on output power. Bulk measurements on robust materials tolerate far more than a focused microscope spot on a biological sample.
• Start low and raise it stepwise. Stop when signal to noise stops improving.
• Compare the spectrum at several power levels. If peak positions or relative intensities change, you are seeing heating, photochemistry or damage rather than better statistics. That check costs two minutes and catches the most common Raman artefact.

Frequently asked questions

Which wavelength is best for Raman spectroscopy?

There is no single best wavelength. 532, 785 and 1064 nm are the three most common choices and cover most situations between them. Which one suits your work follows from the fluorescence behaviour of your samples, your detector, and the Raman shifts you need to reach.

Does longer integration help when fluorescence dominates?

Yes, but inefficiently. In a shot noise limited measurement, signal to noise still improves roughly with the square root of acquisition time even with a large background. What usually stops you is detector saturation, photobleaching, sample heating or instrument drift, not the background as such.

What linewidth does a Raman spectroscopy laser need?

Clearly narrower than the spectral feature you want to resolve. For ordinary dispersive systems almost any single mode diode already satisfies this, because the spectrometer is the limiting element. Tight linewidth specifications belong to high resolution instruments and to low wavenumber measurements near the laser line.

Can one laser cover several sample classes?

785 nm covers a wide range of real samples, which is why many general purpose instruments use it. Laboratories working across both inorganic and organic materials commonly run two lines, often 532 and 785 nm.

Free space or fiber coupled?

Free space gives direct optical access and flexibility during development. Fiber coupling takes alignment out of assembly and makes optical behaviour more reproducible across a production series, which is why OEM builds often choose it.