How to use and calibrate a saltwater refractometer

How a refractometer works

Light bends as it passes from the prism into the water sample, and the amount it bends depends on the sample's refractive index. Dissolved salt raises the refractive index: pure water is about 1.3330 and 35 ppt seawater about 1.3394.^{[1, 16]} The optics turn that bend into a light/dark boundary on a printed scale.

So a refractometer doesn't measure salt directly. It measures refractive index and converts it using seawater's known value, which is why the scale it was built with, and what you calibrate it against, matter so much.^[3]

How to use a refractometer, step by step

1. Let it reach room temperature. Calibrate and measure indoors, with the refractometer and the sample at the same temperature. Red Sea says to leave its unit at 22–25 °C for 30 minutes before calibrating.^{[1, 8]}
2. Clean and dry the prism and cover plate. Use a soft lens cloth; paper towels can scratch the prism. Any dried salt left from last time skews the reading.^[6]
3. Add the sample. Two or three drops, enough to cover the whole prism.^{[6, 10]}
4. Close the cover and check for bubbles. The water should spread into a thin, even film. If you see bubbles or dry spots, wipe and start again.^[10]
5. Wait. The manufacturers' waits differ: 15 seconds (Red Sea), at least 30 (VWR), 45 (Bulk Reef Supply). The point is to let the sample and prism reach the same temperature.^{[6, 8, 10]}
6. Aim at light and focus. Point the prism end at daylight or a bright white light and turn the eyepiece until the scale is sharp.^[8]
7. Read where the blue field meets the white. If the line is fuzzy, use a fresh sample.^{[1, 6]}
8. Rinse and dry. Rinse the prism and plate with RO/DI or distilled water and dry them. Never submerge the refractometer; water inside fogs it.^{[6, 8]}

How to read the scale

Most saltwater refractometers print two scales: salinity in parts per thousand (ppt) and specific gravity (SG). Bulk Reef Supply suggests reading ppt, because SG changes with temperature and ppt doesn't.^[6]

Natural seawater is 35 ppt, which is SG 1.0264 at 25 °C (1.0266 at 20 °C).^[2] Calibration standards are labeled the same way.^[7] The common "1.025–1.026" target spans about 33 to 34.5 ppt at 25 °C (worked out with our salinity converter), so 1.026 is the better match for a reef. Some printed scales pair 35 ppt with 1.025, which Randy Holmes-Farley points out is wrong.^[5] If yours does, trust the ppt side after calibrating with a 35 ppt standard.

How to calibrate a refractometer

Calibrating means putting a solution of known salinity on the prism and turning the small calibration screw until the boundary sits on that value. Most manuals tell you to zero it with distilled or RO/DI water.^{[8, 10]} For seawater that isn't enough, for two reasons:

• Brine versus seawater scales. Many hobby refractometers are built on a sodium chloride (brine) scale. Calibrated perfectly at zero, they read 35 ppt seawater about 1.7 ppt high, Holmes-Farley calculates; Red Sea puts it at about 1.5 ppt.^{[1, 9, 13]}
• Slope error. A zero corrects an error that is the same everywhere on the scale. It can't correct one that grows with salinity, which is common in inexpensive units, often around 1 ppt and sometimes much more.^[1]

Calibrating at the value you actually measure fixes both. Holmes-Farley's rule is that a 35 ppt seawater standard always gives the correct calibration for measuring seawater, whatever the refractometer's scale.^[1] Oceanographers do the same: the IODP shipboard lab calibrates its refractometer to 35 with standard seawater.^[14]

Use a standard made for refractometers. A 53 mS/cm conductivity standard made from potassium chloride has the refractive index of roughly 26 ppt seawater, so calibrating to it puts you about 9 ppt off.^[1] A homemade alternative is 36.5 ppt sodium chloride (3.65 % by weight), which has the same refractive index as 35 ppt seawater; it is valid only for refractometers, not for hydrometers or conductivity probes.^{[1, 3]}

How often: Bulk Reef Supply suggests monthly.^[6] More important is Holmes-Farley's rule: recalibrate before you act on any reading that looks unusually high or low.^[4] Keep the calibration fluid capped, since evaporation concentrates it.

Temperature and ATC

Refractive index changes with temperature, so most handheld refractometers have automatic temperature compensation (ATC), a bimetal strip that shifts the scale.^[1] It has limits:

• It works over a set range, typically 10–30 °C on analog units referenced to 20 °C.^[10]
• It assumes the sample and the prism are at the same temperature, which is why you wait before reading.^{[1, 8]}
• It does not adjust the reading to your tank's temperature. It corrects the instrument, not the water.^[9]

Digital refractometers

Digital seawater refractometers such as the Milwaukee MA887 and Hanna HI96822 shine an LED through a prism, find the boundary with a sensor and convert using seawater (not sodium chloride) relationships.^{[11, 12]} Both are rated at ±2 ppt with 1 ppt resolution, need about 100 µL of sample, and are zeroed with distilled or deionized water as their makers specify.^{[11, 12]} They aren't more precise on paper than a good analog unit, but they remove the guesswork of reading a fuzzy line. Give them about a minute if the sample is much warmer or cooler than the meter.^[11]

Against the alternatives: Holmes-Farley rates conductivity meters and glass hydrometers as suitable too, and plastic swing-arm hydrometers as more prone to error.^[1] Hydrometers can't be calibrated and depend on water temperature, which is why Bulk Reef Supply recommends a refractometer or salinity meter instead.^[15]

Common mistakes

• Zeroing a brine refractometer with RO/DI and trusting it. Reads seawater 1.5–1.7 ppt high, so you dilute a tank that was fine.^{[1, 9]}
• Calibrating with a conductivity standard. A potassium chloride 53 mS/cm fluid lands about 9 ppt off on a refractometer.^[1]
• Reading before the temperature settles. Red Sea reports errors around 1 ppt when the instrument and sample differ by more than a few degrees.^[9]
• Bubbles or dry spots under the cover. Make the boundary vague or wrong; reapply the sample.^[10]
• Mixing up the SG and ppt scales. 35 is ppt; 1.026 is SG. Pick one and stick with it.^[6]
• Leaving salt on the prism. Throws off the next reading and can scratch the glass if you wipe it dry.^[6]

For what to do with the number, see the reef salinity reference, and the salinity converter for switching between ppt, SG and conductivity.