Reflectance & Transmittance

The most important thing about choosing a light source for reflection is to find one with strong output over the wavelength range of interest. For color or when making a measurement to mimic the human eye, the light source needs to cover 380 to 780 nm. For chemical composition of organic material, near-infrared or infrared light will offer more information. Except in a very few specific cases, a narrow light source will not offer enough useful information, so lasers and most LED’s can be ruled out.

Visible light sources: The broad, smoothly varying output of a tungsten halogen light source is ideal for reflectance at visible wavelengths, and for sorting or color applications. The LS-1 is the most economical, and comes in long-lifetime and rack mounted versions. The HL-2000 has very similar spectral output, and has additional shuttered and high power versions. The bluLoop is an LED-based light source with four different bulbs to yield balanced spectral output over the visible range. It is ideal for color measurements due to its unique spectral shape.

UV-visible light sources: Both deuterium tungsten and xenon light sources can be used to make a UV-visible reflectance measurement, but each has its own advantages. A deuterium and tungsten based light source has a broad, smoothly varying spectrum and stable output from 190 to 2500 nm, and since it comes from two different bulbs, the UV and visible portions of the spectrum can be used separately. The DH-2000 comes in a shuttered version for light-sensitive samples, and in a balanced version with where the strong deuterium emission line at 655 nm is attenuated.

A xenon light source, in contrast, has a very jagged and pulsed spectrum, though it has good output into the UV. Averaging is absolutely necessary to get good quality measurements. It gets used primarily for UV reflectance measurements in the field, since it runs on 12 V DC power. Though higher in cost and power consumption, the DT-MINI-2-GS is just as portable, and features the ability to switch the visible portion off to focus on UV wavelengths.

NIR light sources: Both the LS-1 and HL-2000 have output into the NIR, out to about 2400 nm. Even though intensity decreases at the longer wavelengths, this effect is compensated by higher sensitivity of the detectors in our NIR spectrometers at those wavelengths. The Vivo NIR light source is also a tungsten halogen light source, and uses four spatially separated bulbs to avoid overheating the sample. The Cool Red is a silicon nitride light source, with output from 1 to 5 μm, and a shutter capable of rapid modulation.

A reflection probe is great for making quick measurements and for applications where a small spot size needs to be sampled. It can measure either specular or diffuse reflectance, and is compatible with a preconfigured UV-VIS or VIS-NIR spectrometer and any light source (provided the probe fiber matches the wavelength range of the light source). The downside is that it illuminates and detects from the same direction, so it only sees part of the reflected light. Measurements made with a reflection probe are relative measurements.

A reflection stage with reflection probe is convenient for granular samples, or when transmission also needs to be measured. The illuminated stage even has active cooling to reduce the risk of overheating samples placed directly on the sample stage, which can be important with biological and organic samples, or those with low melting points.

An integrating sphere is a good idea if the reflectivity of the sample seems to change at different viewing angles. This happens with rough surfaces like brushed metal, fish scales, and seeds. An integrating sphere has a 180° view of the reflected light, giving more accurate (and absolute) reflectance measurements. An integrating sphere can even be used for convex curved surfaces, or to measure the color of objects that are small enough to fit into the sphere. Ocean Optics integrating spheres view a 5 to 8 mm spot size of the sample.

An integrating sphere’s magic comes from the perfectly diffusing interior surface. Light enters through a circular input port and is scattered repeatedly by the sphere’s inner wall until the light inside the sphere is uniform, regardless of any spatial, angular, or polarization variations in the input. A fiber placed at 90° to the input port then samples a tiny fraction of the light within the sphere, sending it to the spectrometer. A baffle in front of the fiber port helps block any light rays making their first reflection from the sample port.

A variable-angle reflection sampling system is a much easier way to measure specular reflection from surfaces as a function of angle, as the angle of incidence can be varied continuously from 10° to 50°. With convenient fiber input and output connectors, it eliminates the need for realignment when the angle is varied. Note that this system is designed for specular reflectance only, and thus must be referenced against a reflectance standard each time the angle of incidence is changed.

Collimating lenses can be used at the ends of individual fibers to truly customize the angle of incidence and angle of collection when making reflectance measurements. Specular or diffuse reflectance can be measured this way inexpensively, but much more alignment is needed up front, as is extra fixturing. The collimating lenses need to be adjusted carefully to avoid beam divergence and get good signal, making this a more time-consuming method. We also find that color measurements taken using collimating lenses and fibers are not as accurate as those made using an integrating sphere.

A reflection probe collects light at the same angle as it illuminates, and can be used for either specular or diffuse reflection measurements. It’s made of 6 illumination fibers around a single read fiber, which results in a 25° full angle field of view. It seems intuitive to connect the 6 fiber leg to the spectrometer, but it is actually more efficient to use the single fiber leg for detection. That’s because each illumination fiber projects a cone of light from the source. All of them overlap at the sample in the center, exactly where the read fiber is “looking”.

To measure specular reflectance, set the probe at 90° to the sample. For diffuse reflectance, work at an angle. Remember that the rays from the illumination fibers need some space to overlap and create reflected light, so the reflection probe needs to be pulled back slightly from the surface of the sample. Be careful not to scratch or dent the surface of the sample with the probe tip. Using a reflection probe holder (RPH-1 or STAGE) keeps the working distance consistent from one sample to the next, and when taking a reference measurement. The matte black finish of the RPH-1 helps to reduce ambient light.

The most important factors in choosing the right reflection probe for your measurement are wavelength range and amount of light needed. System sensitivity should be optimized for the reference standard, as samples are almost always less reflective. Choose your fiber size based on the sensitivity needed, not the spot size. The spot size is determined primarily by the working distance of the probe from the sample, and is easily changed. Remember, all fibers have the same 25° full angle field of view, so a 600 μm fiber has a spot size only 400 μm (0.4mm) larger than a 200μm fiber at the same working distance.

Probes with reference legs can be used when the light source needs to be monitored continuously. Mixed probes have multiple light source input legs to allow illumination by different wavelengths and and/or multiple measurement legs for output to spectrometers with different configurations. If working with powders or dense solutions, an angled probe tipped with a 30° window makes it easy to immerse the probe directly into the sample and still achieve a consistent working distance.

The ISP-REF integrating sphere is very convenient for general-purpose measurements of reflectance in the lab and in the field thanks to the built-in tungsten light source. Not only is it compact, portable, low cost, but it works with a preconfigured UV-VIS or VIS-NIR spectrometer (adding an L2 or L4 collection lens boosts signal). It collects light that is reflected from the sample in all directions, giving it a full 180° field of view.

Integrating Sphere ISP-REF

So what is the switch on the side of the ISP-REF? It’s the gloss trap, a light trap inside the sphere designed to remove the specular portion of the reflection from the measurement. Flipping the switch opens and closes a shutter on the gloss trap, excluding or including the specular reflection component. The gloss factor of a surface can be calculated by comparing data taken with and without the specular reflection included.

Even with its many advantages, the ISP-REF does have some limitations, the biggest one being that you can only use the integrated tungsten light source for illumination. It also has the highest port fraction of the Ocean Optics integrating spheres, giving it somewhat higher error. Generally, port openings should be less than 5% of an integrating sphere’s internal surface area (it’s 2% for the ISP-REF).

Measurements with an ISP-R require an external light source, which is routed with a fiber to an input port at 8° to the sample port. These spheres are compatible with all UV, visible, and NIR light sources between 200 and 2500 nm. The read port is at 90°, with baffling to isolate it from the sample port. The gloss trap versions of the ISP-R spheres come with two plugs. One is coated in PTFE for including the specular component, and the other is a black light trap to exclude specular reflection.

ISP-R integrating spheres come in 30 to 80 mm diameter, with sample ports from 6 to 8 mm (and even up to 20 mm on a custom basis). Smaller spheres give the strongest signal, but larger spheres with smaller port fractions mix the light better due to the higher number of reflections. It’s best to figure out what sample port size is needed first, then pick a sphere diameter based on whether you need more mixing (larger sphere) or more signal (smaller sphere).

No matter which sphere is used, an L2 or L4 lens will probably be needed in the spectrometer configuration to get enough signal, and possibly a 50 μm slit. The increased sensitivity needed for integrating sphere measurements often results in a lower resolution spectrometer, so if sub-nanometer resolution is needed, a reflection probe may be a better choice.

Each sphere has a maximum illumination fiber size specified for input light. This helps match the illumination spot size to the sample port opening diameter accurately, reducing stray light inside the sphere.

This leads to another point. When coupling a light source to an ISP-R sphere, it is important to adjust the focus of the collimating lens used to deliver light to the sphere until the illumination spot size is small enough that it slightly under-fills the sample port. This can be checked visually with a piece of white paper over the sample port opening. Adjusting the collimating lens and using the proper diameter fiber is extremely important, otherwise stray light will be created when the light source reflects off the sphere around the edge of the sample port.

The SpectroClip probes are compact integrating spheres designed to be used for reflectance measurements in the field. The SpectroClip-R has an integrating sphere with an input port for coupling a light source to make reflection measurements. The SpectroClip-TR has a second, opposing integrating sphere, so it can be used for both reflection and transmission measurements.

The spectrometer in a reflection system needs to match the wavelength range of interest, and have the right sensitivity for the sampling optic being used. A preconfigured UV-VIS or VIS-NIR spectrometer is just fine for a measurement with a reflection probe, but a more sensitive spectrometer is better for integrating sphere measurements.

An enhanced sensitivity “-ES” spectrometer works well with an ISP-REF. ISP-R spheres require even more sensitivity, though, which can be achieved using a wider slit or a yet more sensitive spectrometer like a QE Pro or Maya Pro. The sensitivity needed will really depend on the size of the sphere and the light source being used, so contact one of our application engineers to find the right configuration.

If making color measurements with a tungsten halogen light source as the source, consider using a grating #2 in the USB2000+ or USB4000 spectrometer. It peaks at a shorter wavelength than the grating #3 used in preconfigured VIS-NIR units. When combined with the silicon CCD and a tungsten light source, the resulting spectral response varies less with wavelength, resulting in slightly better S:N over the important 380 to 780 nm range.

A reflectance system is not complete without a standard for reference. Reflectance measurements are a ratio of the reflected light spectrum to the incident light spectrum. Since there is no way to directly collect all of the light incident on a surface, reflectivity is usually measured relative to a reference standard. Except that multiple reflectance standards exist.

The standard chosen should be similar in reflectivity to the sample to keep signal levels about the same during measurement and thereby achieve the best S:N. The WS-1 diffuse reflectance standard is most popular, since it is matte white in color and is >98% reflective from 250 to 1500 nm (then >95% reflective up to 2200 nm). The WS-1-SL can be a good choice when working in the field or in dirty environments, since it can be smoothed, flattened and cleaned if it gets pitted or dirty.

The STAN-SSH high reflectivity specular reference standard is the best choice when measuring very shiny surfaces, but it varies in reflectivity from 85 to 98% over its range of 250 to 2500 nm. This can be corrected in OceanView; just upload the reflectance values and correct for the reflectivity of the standard. This data comes automatically with the STAN-SSH-NIST calibrated reference standard. If no correction is applied, OceanView will assume the standard is 100% reflective at all wavelengths, giving distorted data.

At the other extreme, the STAN-SSL low reflectivity specular reflectance standard is best for surfaces with low specular reflectance values like thin film coatings, anti-reflective coatings, blocking filters and substrates. It has just ~4.0% reflectance from 200 to 2500 nm. It is also possible to purchase calibrated “gray” standards (diffuse reflectance standards at a variety of intermediate values) from our sister company, Labsphere.

When taking a dark measurement with a reflection probe or integrating sphere, it is best to block the light at the light source if possible. Turning the light source off and then on again will throw the light source out of thermal equilibrium and require a new reference measurement. Another option is to point the lit probe or opening of the integrating sphere into a dark space to take the dark measurement so that no light scatters back in. A background measurement taken this way is more accurate because it includes any scattered reference light that will be present in the sample measurements, allowing it to be subtracted properly.

Resist the urge to point the sampling optic at something black (like a piece of paper or a cover cloth). Objects that appear to absorb all wavelengths usually reflect some colors better than others.

Perhaps the most confusing thing about the WS-1 diffuse reflectance standard is that it results in relative measurements. How can it be a standard if it doesn’t yield a conclusive % reflectivity measurement? The important thing to remember is that a standard is only absolute if it comes with calibrated values. A WS-1 is designed to be a diffuse reflector, like most surfaces encountered in daily life. Its absolute reflectivity, therefore, depends on what range of angles of reflected light are collected from it.

The only guarantee from a WS-1 is that it will be equally reflective at all wavelengths, regardless of the angle of collection. That property makes it a really convenient relative standard, useful for diffuse or low-level specular reflection, and for use with any sampling optic. When a WS-1 is used as the standard in a reflectance measurement, OceanView essentially defines its reflectivity as 100%. The relative reflectivity spectrum is accurate in shape, if not in amplitude, which is just fine for most purposes (including color).

In any relative measurement, the enemy is instability. Temperature changes or heating in the light source can cause the reference or dark measurement to change slightly, affecting all data. Here’s how to get the most accurate, repeatable measurements:

• Warm up the light source, since output will continue to change very slightly until it is in thermal equilibrium.
• Take frequent dark and reference measurements to re-establish an accurate baseline.
• Set the integration time so that your reference reaches 80% to 90% of the full scale of the y-axis. This allows you to take advantage of the full dynamic range of the spectrometer, improving signal to noise (S:N).
• Never turn off the light source to take a dark measurement or disconnect fibers. Instead, point the reflection probe or integrating sphere into a dark space that will dissipate the light source’s output. Any signal seen during this measurement is stray light that will also be present in reflected light from the sample, and needs to be subtracted.
• Use averaging to improve signal-to-noise. The S:N improves with the square root to the number of averages taken. For example, setting averages = 9 improves the S:N by a factor of 3.
• Increasing the boxcar value can also smooth out noise in the spectrum. It is a moving average with wavelength. If this value is set too high, it will begin to blur the spectral shape, so use this feature carefully. To smooth data without affecting resolution, set the boxcar value equal to the pixel resolution of your spectrometer.
• Use the nonlinearity and stray light correction options. The stray light correction factor can be determined using high pass filters inline between the sampling optic and the spectrometer. Stray light is often in the NIR, which means it could be reduced through use of a shortpass filter en route to the spectrometer.
• Use an inline filter like BG-34 between the sampling optic and spectrometer when making color measurements to improve the S:N in the blue portion of the spectrum.

Small objects like gemstones and seeds can be tricky to measure in reflection because of their size and rounded or irregular shape. So why not put them inside the integrating sphere? That allows them to be illuminated from all sides, and reflection from all sides is also collected. The WS-3-GEM reference standard is designed to do exactly that. It has a concave surface for holding the sample, and fits directly into the opening of an ISP-80-8-R integrating sphere with a custom-sized sample port.

When making measurements, take the reference with only the WS-3-GEM inserted in the sphere, then place your sample in the WS-3-GEM cavity and re-insert to measure the sample. Measuring the sample inside the sphere allows the spectrum to capture the color of the full exposed surface, as well as any light transmitted through the sample (as in the case of a gemstone). Keep in mind that a large integrating sphere will deliver less light to the spectrometer, so be sure to compensate with a larger slit and L2 or L4 collection lens, as well as a large read fiber. This technique works best with just one or a few small objects.

Think carefully before using the gloss trap to make color measurements, as it tends to decrease the value of L* considerably for dark colors. A glossy looks darker to the eye than a matte or textured sample, even if it has the same pigmentation. People instinctively position glossy objects like magazines so that the gloss component doesn’t enter the eye, and a gloss trap simulates that.

For medium to low gloss samples, however, using the gloss trap doesn’t completely exclude the specular component. Including the specular light for these samples will give better consistency in the values measured from one instrument to another. For textured samples with some gloss component, the best geometry is one that minimizes the surface effects.

Sorting of objects and quality assessment needs to be done quickly. An Ocean Optics spectrometer is an ideal tool for the job thanks to its compact size and rapid data acquisition, but it produces a lot of data. Analysis can appear overwhelming at first, but in fact a full spectral comparison isn’t always needed. A few key wavelengths, on their own or in ratios, can often provide enough information to differentiate samples or confirm quality.

First, characterize the range of objects to be sorted or quality tested by acquiring full spectra and overlapping the data. Similarities and differences will often appear to identify a few unique wavelengths or wavelength ranges that provide the most relevant information about each sample. Sorting then becomes a matter of comparing measured values at those wavelengths to the known range of values for a given sample type.

The same concept can be used for quality testing. By first measuring a large group of products with known properties, upper and lower boundaries for reflectance at each of the key wavelengths can be identified and used to determine if an unknown sample meets quality standards.

When working at 90°, the diameter of the sample area being measured will be equal to ~½ of the distance, d, between the end of the probe and the sample. At 45° angle of incidence, it becomes an oval that is 0.44d by 0.63d.

To see exactly where the reflection probe is reading light from the sample, attach the read fiber to the light source temporarily. The spot illuminated on the sample is exactly where that read fiber is “looking”. If using the RPH-1 probe holder, turn the holder over and place a piece of paper on the contact surface to see the spot size illuminated through the paper.

A regular reflection probe can be used to measure granular samples or dense liquids in a pinch, though an angled probe is much easier to use and gives more consistent results. Combining a standard reflection probe with a petri dish and illuminating from below creates a flat sample surface at a consistent working distance from the reflection probe. A flat sample surface will give more consistent results with similar baselines as compared to reflecting off the uneven upper surface of the sample.

Fill the petri dish with the sample powder or dense liquid, then point the reflection probe up at the petri dish from below. When no light can be seen coming through the sample from above, there is enough material present to ensure consistent measurements from one sample to another.

To take a reference measurement, just put the reference standard face down on the petri dish, being careful to keep the reference clean. This type of measurement is best done in a darkened lab to prevent variations in ambient light from influencing the data. Note – if working in the UV, make sure the petri dish is made of glass, as plastic will not transmit at those wavelengths.

There are times in the field when your reference standard gets lost or just so dirty that there is no hope of a salvageable reference measurement, and yet you desperately need to take data. When this happens, you may be tempted to use a white sheet of paper as your reference, but don’t. Paper is not as spectrally “white” as it appears (measure a piece sometime).

A piece of Styrofoam™ will work much better. It provides fairly diffuse reflection, and has relatively even reflectivity across the visible. Just be sure to take that piece of Styrofoam™ back to the lab with you so that you can measure it against a proper reference standard and correct all your spectra accordingly.

Think carefully about whether you want to turn on electrical dark correction. Dark current levels tend to vary slightly over time due to temperature variations, and the electrical dark correction option will keep the offset dark levels stable for a longer period of time. This correction does add noise, however, as it is a subtraction calculation. So consider your environment and the amount of temperature variation you typically see. If it’s minimal, the dark correction probably won’t add value, but it can be a benefit in unstable temperatures.

The non-linearity correction compensates for the non-linearity effects of the detector, which occur above 80% of the saturation level. Non-linearity correction improves the accuracy of the spectral shape, but also adds more noise to your original signal.