Raman spectroscopy · 14 min
532, 638, or 785 nm is the first decision you make about a Raman system and the hardest one to reverse. Three physical rules decide it, and they pull in opposite directions.
The governing trade-off
Almost every disappointing Raman measurement traces back to one of three relationships. None of them is complicated on its own. The difficulty is that the wavelength which optimises one actively degrades another.
I_Raman ∝ ν⁴ ∝ 1 / λ⁴
Scattering strength rises with the fourth power of frequency. Moving from 785 to 532 nm multiplies the Raman signal by (785/532)⁴ ≈ 4.7 — the difference between a five-second and a twenty-five-second acquisition.
SNR ∝ I_Raman / √I_background
Fewer molecules have electronic transitions reachable at 785 nm than at 532 nm. Because the background contributes shot noise, a fluorescence hump ten times the Raman signal costs far more than the ν⁴ law gives back.
λ_Stokes = 1 / (1/λ_ex − ν̃)
Raman shifts land at longer wavelengths than the laser. From 785 nm, the C–H stretch at 2900 cm⁻¹ arrives near 1016 nm, where silicon quantum efficiency has already collapsed.
sampling depth in tissue: ~tens of µm at 532 nm → hundreds of µm to mm at 785 nm
Wavelength also decides how deep the measurement reaches. In anything that scatters or absorbs — tissue, tablets, powders, filled polymers — shorter wavelengths are attenuated within the top layers, while near-infrared light survives far deeper before it is lost. For a heterogeneous sample this changes what the number means: 532 nm reports on the surface grain the spot happens to land on, 785 nm averages a volume closer to the formulation you actually want to characterise. Transmission Raman pushes this to its conclusion, firing 785 or 830 nm through an entire tablet to assay its full contents.
The rule cuts both ways. For thin films, 2D materials, and strained semiconductor layers, the shallow reach of green excitation is precisely the point — the signal comes from the layer under study instead of the substrate beneath it. Unlike the other three rules this one never fights fluorescence: both push turbid organic samples toward the red.
Interactive
The simulation below applies the ν⁴ law, a fluorescence background, and a real silicon QE curve to the same set of Raman bands. Raise the fluorescence to a biological sample and watch 532 nm — the strongest scatterer — become the worst choice on the list.
Signal strength, fluorescence background, and detector reach all move at once. Change the wavelength and watch which of the three becomes your limiting factor.
Raman shift (cm⁻¹)
Raman signal
4.7×
ν⁴ law, vs 785 nm
Fluorescence
10.0×
background, vs 785 nm
Peak contrast
1.4×
signal ÷ √background
Lateral resolution
361 nm
0.61 λ / NA 0.90
Typical polymer or pharmaceutical
NA 0.90
Detector reach
Silicon still responds across the full fingerprint and C–H stretch region (to ~3400 cm⁻¹).
The ν⁴ scaling, Stokes-shift arithmetic, and silicon QE curve are quantitative. The fluorescence model is a schematic exponential — real fluorescence is wildly sample-specific and can swing by orders of magnitude between two samples of nominally the same material. Use this to understand the shape of the trade-off, never as a prediction for your sample.
At a glance
| Wavelength | Relative signal | Fluorescence risk | Lateral resolution | Typically chosen for |
|---|---|---|---|---|
| 405 nm | ~14× vs 785 nm | Severe on almost any organic | ~275 nm at NA 0.9 | Wide-bandgap semiconductors, resonance Raman, diamond |
| 532 nm | ~4.7× vs 785 nm | High — the usual reason it fails | ~360 nm at NA 0.9 | Carbon and 2D materials, minerals, inorganics, semiconductors |
| 638 nm | ~2.3× vs 785 nm | Moderate | ~432 nm at NA 0.9 | General-purpose work, mixed sample sets, some biological samples |
| 785 nm | 1× (reference) | Low — the standard escape route | ~532 nm at NA 0.9 | Polymers, pharmaceuticals, biological tissue, forensics |
| 1064 nm | ~0.24× vs 785 nm | Essentially none | ~721 nm at NA 0.9 | Strongly fluorescent samples where nothing else works |
Relative signal is the ν⁴ ratio against 785 nm before detector response is applied. Lateral resolution is 0.61 λ / NA at NA 0.9.
The physics above condenses into defaults that most labs converge on independently. If your sample is on this list, the first-choice column is the wavelength to price first — and the fallback is where you retreat when a particular sample proves the exception.
| Sample type | First choice | Fallback | Why |
|---|---|---|---|
| Carbon & 2D materials — graphene, CNTs, graphite, MoS₂ | 532 nm | 638 nm | No fluorescence to fear, so the ν⁴ advantage is free. The D, G, and 2D bands are strong and well characterised at green. |
| Semiconductors — Si, GaAs, strain and thin-film work | 532 nm | 405 nm | Strong signal, and the shallow optical penetration at green is a feature: the measurement stays in the film or strained layer you care about. |
| Diamond & wide-bandgap materials — GaN, SiC | 405 nm | 532 nm | Below-bandgap excitation avoids absorption while resonance and near-resonance enhancement boost the signal. |
| Minerals & geology | 532 nm | 785 nm | Most minerals are non-fluorescent. Rare-earth impurities are the exception — if a spectrum drowns in luminescence, move to red. |
| Polymers & plastics | 785 nm | 1064 nm | The polymer itself is often fine at 532 nm, but additives, dyes, and degradation products rarely are. Red is the safe default. |
| Pharmaceuticals — tablets, APIs, excipients | 785 nm | 1064 nm | Excipients fluoresce at green, and the deeper sampling volume at red averages the formulation instead of one grain of it. |
| Biological tissue & cells | 785 nm | 1064 nm | Autofluorescence at 532 nm buries almost everything, and the lower photon energy at red reduces photodamage during long maps. |
| Forensics & true unknowns | 785 nm | 638 nm | When you cannot predict the fluorescence, choose the wavelength that fails least often. A weak spectrum beats no spectrum. |
| Strongly coloured or dyed samples | 1064 nm | 785 nm | Chromophores that absorb visible light fluoresce or burn under it. Deep NIR sidesteps both — at the cost of an InGaAs detector. |
A defensible order
The order matters more than any single specification. Optimising signal first is the most common way to end up with an instrument that cannot measure the samples it was bought for.
Fluorescence is the only failure mode that returns no spectrum at all. A weak spectrum can be integrated for longer; a spectrum buried under a fluorescence hump cannot be recovered. Ask what your sample does under green light before optimising anything else.
Once fluorescence is survivable, move as far to the blue as you can tolerate. Every step shorter buys signal at the fourth power — the difference between 785 and 532 nm is roughly a factor of five in integration time for the same signal-to-noise.
Confirm the highest Raman shift you care about still lands where your detector responds. If you need the C–H stretch region near 2900 cm⁻¹ on a silicon array, 785 nm is already marginal and 1064 nm is out of the question.
Lateral resolution scales as 0.61 λ / NA. At NA 0.9, moving from 532 to 785 nm costs roughly 170 nm of resolution. That is irrelevant for bulk measurement and decisive for mapping small inclusions or single particles.
Excitation sources
Once the wavelength is settled, two laser specifications decide whether the instrument reaches its resolution — and neither is the one people ask about first.
Δλ = λ² · Δν̃ → 1 cm⁻¹ ≈ 0.028 nm at 532, 0.062 nm at 785
A 1 MHz linewidth is 3 × 10⁻⁵ cm⁻¹ — irrelevant. What is not irrelevant is a free-running multimode diode whose emission is spread over 0.1–0.5 nm, because at 785 nm that is 1.6 to 8 cm⁻¹ of instrumental broadening applied to every band you measure. Ask for the spectral width of the whole output, not the linewidth of one mode.
0.1 nm drift at 785 nm ≈ 1.6 cm⁻¹ shift on every peak
Raman shift is a difference, so if the laser moves, the entire spectrum moves with it. A diode that wanders with temperature and drive current will quietly recalibrate your instrument between the morning and the afternoon. This is why wavelength-locked 785 nm sources exist and why they are worth the premium.
The strongest scatterer on the list, and the wavelength most likely to be defeated by fluorescence. Buy it for inorganics, carbon materials, and semiconductors.
Single-frequency and specified for Raman, which is exactly what the name says. 500 mW leaves room for the notch filter, the objective, and a sample that cannot take full power.
Watt-class single-frequency fibre output for large-area mapping, where integration time per pixel is the whole experiment.
Compact single-frequency head when 150 mW at the aperture is genuinely enough — which for a confocal microscope on a strong scatterer it usually is.
The only 532 nm line here with a published power-stability grade and a built-in modulator — worth it if you shutter the beam between pixels or need quantitative intensities.
The sensible single choice for a mixed sample set: most of the ν⁴ advantage, much less of the fluorescence.
The stocked line nearest the 638 nm middle ground, and by far the most capable of them. 633 rather than 638 nm shifts your filter set slightly — specify them together.
Genuinely 638 nm and genuinely single-mode, but only 40 mW. Fine for a strong scatterer under a high-NA objective, thin for anything else.
The fluorescence escape route, and the wavelength where laser wavelength stability stops being a footnote and becomes the specification that matters most.
Wavelength-locked, which is the whole argument. A free-running 785 nm diode wanders with temperature and current, and every nanometre it moves drags your entire wavenumber axis with it.
Single-frequency fibre output. The narrowest and quietest 785 nm option, and the one to pick if you also care about low-wavenumber performance near the filter edge.
A mid-power single-frequency head for probe-based and benchtop instruments where 300 mW after the fibre is the realistic budget.
Both are specialist answers to a specific failure. Reach for them when the middle of the range has already been ruled out.
Violet excitation for wide-bandgap semiconductors, diamond, and resonance Raman. On almost any organic it produces a fluorescence wall instead of a spectrum.
The last resort for samples that fluoresce at 785 nm. Remember that the entire Stokes spectrum now lands beyond silicon — this needs an InGaAs array or an FT-Raman interferometer.
Watt-class 1064 nm when the ν⁴ penalty has to be bought back with raw power and the sample can survive it.
Wavelengths not listed here — 442, 457, 473, 660, 671, 808, 830 nm and more — are available across the same source families. If your filter set or an existing instrument fixes the wavelength for you, start from that constraint and we will work back to a source.
The rest of the instrument
The laser is one of four decisions, and not the one that most often limits a Raman system. Grouped below with the reason each belongs in the budget.
Excitation wavelength decides what reaches the slit. Focal length, grating, and slit width decide whether two neighbouring bands arrive as two peaks or one. Longer excitation spreads a given cm⁻¹ interval over more nanometres, so it is easier to resolve — but it needs more detector real estate to cover the same Stokes range.
The industry-standard focal length for Raman and photoluminescence, with a triple grating turret so you can cover UV to NIR without opening the box.
Throughput first. The faster aperture collects more of a weak signal at the cost of resolution — the right trade for routine work and for 1064 nm excitation.
When you need to separate bands a few wavenumbers apart and can afford the light lost to the slower aperture.
Flagship resolving power, to 0.028 nm with a PMT. For fine molecular structure and calibration work rather than everyday Raman.
Toroidal mirrors remove astigmatism across the full focal plane, which is what makes multi-track and multi-fibre spectroscopy work at all.
Work out which groove densities and blaze wavelengths cover your Stokes range at your excitation wavelength before you commit to a turret.
Everything in the wavelength lab above about silicon quantum efficiency lands here. Below 785 nm excitation a cooled silicon sensor is the obvious answer; at 1064 nm it is not an answer at all.
Deep-cooled mono sensors for the long integrations that weak Raman bands need. Cooling buys you dark current, which is what limits a 60-second exposure.
Large pixels and a cooled mono IMX432 sensor. Big pixels collect more of a spectral line per binned column, which matters more in spectroscopy than megapixel count ever does.
Filter by sensor, cooling, read noise, and interface against the spectrograph you have chosen rather than against a generic spec sheet.
The companion guide on quantum efficiency, read noise, and why the signal-to-noise calculation rarely favours the camera with the most pixels.
Where 785 and 1064 nm excitation push the Stokes bands: the InGaAs guide for everything past the silicon cutoff.
At 1064 nm excitation the whole Stokes spectrum sits past 1100 nm, where silicon quantum efficiency is zero. No amount of cooling or exposure recovers it — that configuration needs an InGaAs array or an FT-Raman interferometer, and it is worth confirming before the wavelength is chosen rather than after.
The Rayleigh line is six or more orders of magnitude brighter than the Raman bands sitting beside it. Everything in this group exists to remove it without removing the low-wavenumber signal you came for.
Laser clean-up before the sample and edge or notch rejection after it. Filter edge position is what sets how close to the laser line you can measure — often the real limit on low-wavenumber work.
Groove density trades dispersion against free spectral range; blaze wavelength decides where your efficiency peaks. Both need choosing against the excitation wavelength, not independently.
For probe-based and remote-head geometries. Note that fused silica fibre generates its own Raman background, which is why probes carry filters at the tip.
Sample burning at 532 nm is a common and often unrecognised failure. Attenuate rather than turning the laser down into its noisy, drifting low-current regime.
Mirrors, lenses, and windows specified for the excitation wavelength and the full Stokes range — not just for the laser line.
Wavelength-specific eyewear. A 785 nm confocal spot is invisible and focuses to the retina efficiently, which is the worst combination.
The Finder 930 is a confocal Raman microscope configurable across 532, 638, and 785 nm, with the spectrograph, detector, filter set, and mapping stage already matched to each other. It exists precisely because the honest answer to “which wavelength?” is often “measure one representative sample at two of them.”
Open source
Wavelength choice determines what reaches your detector. What happens next — calibration, baseline removal, peak identification — decides whether it becomes an answer. We develop and maintain Raman Studio, an open-source desktop application for acquiring, processing, calibrating, and identifying Raman spectra, built around OpenRAMAN spectrometers.
Live camera feed with hardware ROI management for FLIR/PointGrey sensors.
Baseline removal, Savitzky–Golay smoothing, median filtering, blank subtraction, and peak detection.
Wavelength calibration against known emission lines, with automatic peak-to-line assignment.
Reference-library scoring by cosine similarity (HQI) with spectral overlay, plus structure-based prediction from SMILES or compound names via Mol2Raman.
Common questions
785 nm if your samples are organic, biological, or unknown, because fluorescence is the failure mode that produces no usable spectrum at all. 532 nm if you work with inorganics, carbon materials, or semiconductors, where fluorescence is rarely a problem and the ν⁴ signal advantage is decisive. 638 nm is the sensible single choice for a mixed sample set.
Raman scattering intensity scales with the fourth power of the scattered frequency. Because 532 nm has a higher frequency than 785 nm, the ratio is (785/532)⁴ ≈ 4.7. All else being equal, 532 nm excitation returns roughly five times more Raman signal from the same sample.
A Raman shift of 2900 cm⁻¹ from 785 nm excitation lands at about 1016 nm, and 3200 cm⁻¹ lands beyond 1030 nm. Silicon detector quantum efficiency has collapsed to a few percent there and reaches zero near 1100 nm. The fingerprint region is unaffected, but high-wavenumber bands become difficult or impossible without an InGaAs detector.
Not inherently — the opposite, usually. For a fixed spectrograph, a given interval in cm⁻¹ spans more nanometres at longer excitation wavelengths, so each detector pixel covers fewer wavenumbers. The practical catch is that the full Stokes range then needs more detector real estate and runs into the silicon cutoff.
Baseline correction removes the shape of a fluorescence background, but not the shot noise it carries. If the background is much larger than the Raman bands, that noise buries weak peaks and no algorithm recovers them. Photobleaching, shifted-excitation difference methods, and time-gating all help, but choosing the right excitation wavelength is the only approach that prevents the problem instead of treating it.
In transparent samples, depth is set by the focusing optics, not the wavelength. In turbid samples — tissue, tablets, powders, filled polymers — scattering and absorption attenuate short wavelengths within the top layers, so 532 nm probes only tens of micrometres while 785 nm reaches hundreds of micrometres or more. For heterogeneous samples this determines whether you measure one surface grain or a representative volume, and it is a second reason, independent of fluorescence, why pharmaceutical and biological work standardised on near-infrared excitation.
It can. Shorter-wavelength photons carry more energy, and at the power densities of a focused confocal spot this matters for biological samples, pigments, and thermally sensitive materials. Sample burning during 532 nm measurements is a common and often unrecognised failure. Reduce power and lengthen integration before assuming a sample is simply a weak scatterer.
The honest answer usually requires measuring one representative sample at two wavelengths. Our confocal Raman system is configurable across 532, 638, and 785 nm for exactly this reason.