Gratings and etalons ease wavelength tests

22 Aug 2003

Knowing the precise wavelength or spectral content of your source can be crucial in many applications. Jacqueline Hewett investigates the technology behind wavelength meters and handheld spectrometers, two instruments that are designed for this task.

From Opto & Laser Europe September 2003

Whether you want to measure the wavelength of a tunable laser to five decimal places or find out the spectral output of an LED, a system is out there that can be tailored to suit your needs. This month, Opto & Laser Europe takes a look at two laboratory workhorses that are designed for such measurements: wavelength meters and spectrometers.

Although both set-ups can be used for the same purpose, they have quite different characteristics. A wavelength meter is a high-precision system for taking accurate measurements of the wavelength of a laser. By contrast, the miniature handheld spectrometers that can be a more convenient solution in many applications do not have the accuracy offered by wavelength meters, although they can gather spectral data. To help you choose, we examine both alternatives in depth below.

1. Wavelength meters If you need to take a very accurate measurement of the wavelength of your laser, you will need a wavelength meter. These high-precision devices are typically used in applications such as spectroscopy, in which it is important to know the exact wavelength at which a tunable laser is operating.

The devices currently on the market tend to have an accuracy of around one part per million or better. Wavelength meters achieve this level of accuracy by utilizing the coherent nature of laser light in an interferometer. Typically, one of two interferometer schemes is employed: either a scanning Michelson configuration, or a combined approach using multiple Fabry-Perot and Fizeau etalons.

Michelson interferometers contain a beamsplitter that divides an incoming beam into two perpendicular paths that form the arms of the interferometer. One of the arms has a fixed length, while the other's length can be altered, changing the path difference between the beams. Mirrors at the ends of both arms reflect the beams, which then recombine to form an interference pattern.

A photodiode tracks the intensity changes in the interference pattern. The unknown wavelength is determined by using a formula that relates the mirror displacement to the wavelength of a built-in reference source. A stabilized HeNe laser is generally used as a reference guide.

There are several points to note about this set-up. For one thing, although it can provide an extremely accurate measurement - down to 0.2 parts per million - it only suits continuous-wave (CW) sources. Also, the update rate from this system is slower than in an etalon-based system, due to the detection technology used to monitor the fringe pattern.

Etalon-based interferometers, in contrast with the Michelson-based product, have no moving parts and are able to measure the wavelength of either pulsed or CW sources. Fabry-Perot and Fizeau etalons are both used and are similar in design: the former use two plane parallel mirrors separated by a small air gap, while the latter use wedge-shaped reflecting surfaces.

A pulsed wavelength meter contains a Fabry-Perot and a Fizeau etalon on separate optical paths. An incident pulse is split and enters both etalons, creating two interference patterns characteristic of each set-up. CCD cameras capture these fringe patterns, which are compared to an internal reference providing continual calibration.

The typical update rate for an etalon-based system is 20 Hz, whereas the update rate for a Michelson system is 1 Hz. The update rates of the two technologies differ because the Michelson interferometer must be scanned, which takes about 1 s. By contrast, there is no scanning involved in the etalon-type systems. The drawback of the etalon approach is that it has a shorter operational wavelength range compared with a Michelson-based product.

Both Michelson and etalon-based systems use silicon-based detectors and can cope with wavelengths between 400 and 1100 nm. UV-enhanced silicon detectors can generally be added, allowing wavelengths as short as 250 nm to be detected.

When it comes to actually taking a measurement using either system, standard practice is to pick off a small portion of the laser beam that needs to be characterized using a beam splitter. The tapped beam is then directly fed through the entrance aperture of the wavelength meter, or coupled into a fibre that enters the meter. To avoid saturating the detector, a maximum of a few milliwatts must be used, so neutral-density filters may have to be inserted into the beam path before the light enters the wavelength meter.

It is also important to remember that wavelength meters cannot cope with multimode lasers. Passing a multimode laser beam through an interferometer will not result in a useable interference pattern - a single longitudinal mode is required. Similarly, these instruments cannot measure the wavelength of two CW lasers simultaneously. (Note that two pulsed lasers can be measured sequentially using a time-gating system.)

The readout rates of wavelength meters are limited, which restricts the rate at which the devices can evaluate successive laser pulses. For example, a typical wavelength meter cannot measure every pulse from a laser operating at a kHz repetition rate. Problems can also occur when measuring pulses with a short time duration, owing to spectral broadening. For instance, wavelength meters can evaluate picosecond pulses, but the accuracy of the measurement will suffer because the pulses have a broad wavelength spectrum.

2. Handheld spectrometers Spectroscopy - the study of how light interacts with matter - is naturally a broad topic and not surprisingly, a wide range of spectrometers is available. The list of potential applications for these devices includes fluorescence lifetime spectroscopy; Raman spectroscopy; and laser-induced breakdown spectroscopy. This section however will be dedicated to the miniature spectrometers that operate in the UV-visible-infrared region of the spectrum and are used to study the spectral characteristics of light sources.

The advent and growth of miniature fibre-optic spectrometers has been a significant development in recent years, allowing scientists to take the spectrometer to the sample for the first time. Such spectrometers are built from a range of mix-and-match components and can be specifically tailored to suit your application.

Unlike when a wavelength meter is used, the light source under scrutiny does not need to be coherent because the measurement does not rely on an interferometer. Most spectrometers use a grating or a prism to spatially disperse the wavelength components of the incident light. A series of lenses and mirrors then projects the dispersed wavelengths onto a linear detector array. Calibration is typically performed using a series of discharge lines, such as those from a mercury-arc lamp, rather than an internal HeNe source.

Three crucial components will define the performance of the device: the entrance slit, the grating and the detector. It's a good idea to discuss the options available with a sales engineer, as these components are generally fixed into the spectrometer at the time of manufacture. The optical resolution of the spectrometer will be determined by your choice of slit and grating. The typical optical resolution of a monochromatic source, measured at full-width half maximum, is about 1 nm.

The entrance slit sits flush against the end of the optical fibre and controls the amount of light entering the spectrometer. Slit widths typically range between 5 and 200 µm, while heights are 1-3 mm. After passing through the slit, the light is reflected from a collimating mirror onto a grating, which splits white light into its component wavelengths.

From the large selection of gratings available, make sure you choose one that covers the optimum wavelength range and gives you the resolution you require for your application. The amount of dispersion from the grating is determined by the density of its holographically-etched or ruled grooves. Measured in lines per millimetre, the greater the groove density, the higher the optical resolution of the spectrometer.

However, it is important to bear in mind that there is an inverse relationship between the groove density (and consequently the resolution) and the spectral range of the spectrometer. If the spectral range is small, say 100 nm, the spectrometer will be restricted and will only measure a specific 100 nm-wide waveband accurately.

Another key factor to consider is the blaze wavelength of the grating, which refers to the efficiency with which it disperses light. It is defined as the peak wavelength in the efficiency curve of a ruled grating, or the most efficient wavelength range for a holographic grating. Manufacturers may also quote a wavelength range over which the grating is more than 30% efficient.

For example, a UV grating featuring 600 lines/mm could have a blaze wavelength of 300 nm, a greater than 30% efficiency range of 200-575 nm and a spectral range of 650 nm. The spectral range of such a grating would be large and it would be most efficient over the 200-575 nm range. As a consequence, wavelengths of more than 575 nm will have a lower intensity at the detector, owing to the reduced efficiency of the grating.

The last of the three crucial components of a spectrometer is the detector. Your device will generally come equipped with (or you will be given the choice of) a linear CCD array or a linear photodiode array. In both cases these will consist of silicon detectors, which can be enhanced to detect UV wavelengths as short as 200 nm or infrared wavelengths of as much as 1100 nm. In general, the CCD camera has a faster readout rate than the photodiode, but the photodiode is inherently less noisy than the CCD.