Raman Spectroscopy Setup: Laser, Filters, Spectrometer and Calibration
How a Raman spectroscopy setup is built and calibrated: choosing the excitation wavelength (532, 785 or 1064 nm), laser and filter requirements, collection optics, the spectrometer's resolution and coverage with a worked example, detectors, wavenumber calibration with silicon and lamp lines, and common problems such as fluorescence and sample heating.
Scope
This article describes the parts of a dispersive Raman spectroscopy setup, how to choose each one, and how to calibrate the result. It covers benchtop and microscope instruments that use a laser, a filter to remove the elastically scattered light, a grating spectrometer and an array detector. The physics of the effect is in the Raman scattering entry, and the technique and its uses in Raman spectroscopy.
In short: 785 nm excitation is the usual default because it suppresses most sample fluorescence while still working with silicon detectors; 532 nm gives about five times the signal where fluorescence is not a problem; 1064 nm is for strongly fluorescent samples and needs an infrared detector. In every case the laser must be spectrally clean and stable, the filter must block the laser line by a very large factor, and the wavenumber axis must be calibrated against known lines.
The signal
A Raman spectrum is plotted against the Raman shift, the difference in wavenumber between the laser and the scattered light:
usually in cm⁻¹. The shift is a property of the sample's vibrations, independent of the laser wavelength, but the scattered wavelength is not:
| Excitation | 1000 cm⁻¹ shift appears at | 3000 cm⁻¹ shift appears at |
|---|---|---|
| 532 nm | 561.9 nm | 633.0 nm |
| 785 nm | 851.9 nm | 1026.8 nm |
| 1064 nm | 1190.7 nm | 1562.9 nm |
Spontaneous Raman scattering is weak: only a small fraction of the scattered light is shifted, and the elastically scattered (Rayleigh) light at the laser wavelength is many orders of magnitude stronger. The design of the whole instrument follows from these two facts.
Choosing the excitation wavelength
Raman intensity scales roughly as the fourth power of the scattered frequency, so for the same laser power and detector efficiency 532 nm gives about = 4.7 times the signal of 785 nm. The competing effect is fluorescence: many organic, biological and impure samples absorb in the visible and emit a broad background that can be far stronger than the Raman lines. Longer excitation wavelengths excite less fluorescence. At 1064 nm the weak signal is often recorded with a Fourier-transform instrument, built on the same interferometer as FTIR spectroscopy.
| Excitation | Strengths | Limitations | Detector |
|---|---|---|---|
| 532 nm | Strong signal, good for inorganic materials, carbon and semiconductors | Fluorescence from many organic samples | Silicon CCD |
| 633 nm | Intermediate | Intermediate | Silicon CCD |
| 785 nm | Much less fluorescence; the common default for pharmaceuticals, polymers and biology | Silicon detector efficiency falls toward 1000 nm, limiting the high-shift range | Deep-depletion silicon CCD |
| 1064 nm | Least fluorescence | Weakest signal; needs InGaAs arrays or Fourier-transform Raman | InGaAs array, or FT spectrometer |
The first table shows the 785 nm limitation: a 3000 cm⁻¹ C–H stretching band lands at 1027 nm, where silicon is nearly transparent.
Laser
The laser linewidth adds directly to the width of every Raman line, so it should be well below the spectrometer resolution, typically under a wavenumber; frequency-stabilized diode lasers (with a grating or volume Bragg grating) and single-frequency DPSS lasers are used. Diode lasers also emit a weak broad background of spontaneous emission around the line, which overlaps the low-shift Raman region; a laser-line bandpass filter after the laser removes it. The power at the sample ranges from under a milliwatt for sensitive samples under a microscope to hundreds of milliwatts for bulk liquids.
Filtering the laser line
After collection, the light passes through a long-pass edge filter or a notch filter that blocks the laser wavelength and transmits the shifted light. The blocking at the laser line is specified as an optical density, typically OD 6 or more, and the edge position sets the lowest shift that can be measured. Edge filters are the usual choice for Stokes-only measurements; notch filters also pass the anti-Stokes side. Measurements very close to the laser line, below about 100 cm⁻¹, need volume Bragg grating notch filters or a multi-stage spectrometer.
Collection optics
Most instruments collect in backscattering: the same lens or microscope objective focuses the laser and collects the scattered light, separated by a dichroic beamsplitter or by the edge filter used at an angle. Scattered light is emitted over a wide angle, so collection efficiency rises with numerical aperture; microscope objectives of NA 0.5 to 0.9 are common, and a pinhole or slit in a conjugate image plane gives confocal depth selection. Fiber probes put the filters in the probe head: a bandpass filter on the excitation side removes the Raman and fluorescence background that the laser generates in the delivery fiber, and an edge filter on the collection side blocks the laser before it can generate the same background in the collection fiber.
Spectrometer and detector
A grating spectrometer disperses the light onto a cooled array detector. The reciprocal linear dispersion of a spectrometer of focal length with a grating of groove density in first order is approximately , where is the diffraction angle (see diffraction grating).
Worked example. A 300 mm spectrometer with a 1200 grooves/mm grating at 785 nm, in a symmetric geometry, diffracts at = 28.1° and disperses 2.45 nm per millimeter. On a detector with 26 μm pixels each pixel covers 0.064 nm, or 0.88 cm⁻¹ at 852 nm (1000 cm⁻¹ shift), and 1024 pixels cover 65 nm. The spectral resolution is set by the larger of the slit image and about two pixels, so a few wavenumbers in practice. Covering the full 200 to 3200 cm⁻¹ range at once needs a coarser grating or a shorter focal length; high resolution over a wide range needs several grating positions stitched together.
Detectors are thermoelectrically cooled silicon CCDs for 532 to 785 nm excitation, deep-depletion and back-illuminated types extending the response toward 1000 nm, and InGaAs arrays for 1064 nm. Cooling reduces dark current, which matters in the long exposures (seconds to minutes) that weak samples need.
Calibration
Wavenumber axis. Record the lines of a neon or argon lamp, whose wavelengths are tabulated, and fit the pixel-to-wavelength relation. The laser wavelength itself must be known to convert wavelength into Raman shift; measure it, or use a shift standard. A silicon wafer, whose first-order phonon line is at about 520.7 cm⁻¹, is the everyday check; liquid standards such as cyclohexane give several lines across the range (ASTM E1840 lists shift standards).
Intensity axis. The relative response of filters, grating and detector varies across the spectrum, so band intensities are only comparable after correcting with a calibrated white-light source or a luminescent reference material.
Common problems
Fluorescence background. Change to a longer excitation wavelength, photobleach the sample by exposing it before measuring, or subtract a fitted baseline; baseline fits can distort broad bands and should be applied consistently.
Sample heating and damage. 10 mW focused to a 1 μm spot is an irradiance of 1.3 × 10⁶ W/cm², enough to heat, burn or change dark samples. Reduce the power, defocus, or spin the sample, and check that repeated spectra agree.
Cosmic rays. Sharp single-pixel spikes on long exposures; remove by taking two exposures and keeping the lower value at each pixel, or by a spike filter.
Laser drift. A shift of the laser wavelength moves every Raman line; recheck against silicon or a standard when precise band positions matter.
References: E. Smith and G. Dent, Modern Raman Spectroscopy: A Practical Approach (2nd ed., Wiley, 2019); R. L. McCreery, Raman Spectroscopy for Chemical Analysis (Wiley, 2000); ASTM E1840, Standard Guide for Raman Shift Standards for Spectrometer Calibration.