Page 6: Guidelines for Canadian Drinking Water Quality: Guideline Technical Document – Turbidity

Part II. Science and Technical Considerations (continued)

5.0 Analytical methods

The turbidity of filtered water is usually measured using the nephelometric method. Nephelometry determines turbidity using the intensity of scattered light measured by a detector that is at 90 degrees to the incident light source. Table 3 lists seven nephelometric methods for the measurement of turbidity in drinking water that have been developed by consensus standards organizations or are approved by recognized organizations. These methods have been developed to standardize instrument design and calibration in order to achieve consistency in turbidity measurements. Depending on the range of turbidity in source water, instruments that conform to these standards may not be appropriate for monitoring turbidity in source water.

The U.S. Environmental Protection Agency (EPA), the American Public Health Association (APHA) / American Water Works Association (AWWA) / Water Environment Federation (WEF), the International Organization for Standardization (ISO) and ASTM International (ASTM) have developed or approved these standardized methods. Utilities should use turbidimeters that conform to one of the methods discussed below when monitoring drinking water.

Table 3: Recognized analytical methods for measuring turbidity in drinking water
Method Reference Description

FAU: formazin attenuation unit; FNU: formazin nephelometric unit.

APHA/AWWA/ WEF Standard Method 2130B APHA et al. (2012) Tungsten lamp at 2200-3000 K and one or more perpendicular detectors (and filters) with spectral response peak of 400-600 nm; light path less than or equal to 10 cm. Applicable measurement range of 0 to greater than 1000 NTU.
U.S. EPA Method 180.1 Rev. 2.0 U.S. EPA (1993) Tungsten lamp at 2200-3000 K and one or more perpendicular detectors (and filters) with spectral response peak of 400-600 nm; light path less than or equal to 10 cm. Applicable measurement range of 0-40 NTU.
ISO 7027 ISO (1999) Tungsten lamp (and filters), diode or laser as radiation source at 860 nm (or 550 nm if sample is colourless) with a perpendicular detector and aperture angle of 20-30 degrees. Two applicable measurement ranges are available, depending on the method selected. The diffuse radiation method has a range of 0-40 FNU. The attenuation of radiant flux has a range of 40-4000 FAU.
GLI Method 2 GLI International Inc. (1992) Two perpendicular 860 nm light sources alternately pulse each 0.5 seconds, and two perpendicular detectors alternately measure "reference" and "active" signals. Applicable measurement range of 0-40 NTU. The method allows dilution for measurement of samples above 40 NTU.
Hach FilterTrak Method 10133 Rev. 2.0 Hach Company (2000) Laser diode at 660 nm at 90 degrees to detector/receiver (light path less than or equal to 10 cm), which may use photomultiplier tube and fibre optic cable. Applicable measurement range of 0-5000 mNTUs (0-5.0 NTU).
ASTM D6698-07 ASTM International (2007) This method is for the online measurement of turbidity below 5 NTU in water. A variety of instrument technologies may be used in this method, including the design features listed in the methods above. Applicable measurement range of less than or equal to 0.02-5.0 NTU.
ASTM D6855-10 ASTM International (2010) This method is for the static measurement of turbidity below 5 NTU in water. A variety of instrument technologies may be used in this method, including the design features listed in the methods above. Applicable measurement range of less than or equal to 0.02-5.0 NTU or FNU.

A variety of reporting units are available for turbidity, depending on the design of the turbidity instrument that is used. In general, devices that use a tungsten lamp with a colour temperature of 2200-3000 K and measure the scattered light at an angle of 90 degrees to the incident light beam use NTUs. Instruments that measure turbidity in formazin nephelometric units (FNUs) use a light-emitting diode with a wavelength of 860 ± 60 nm as a light source and a detector at 90 degrees to the incident light beam. Instruments that measure turbidity in formazin attenuation units (FAUs) use a light-emitting diode with a wavelength of 860 ± 60 nm and a detector at 180 degrees to the incident light beam. These units are equivalent when measuring a calibration solution; however, each different type of instrument may not produce directly comparable results when measuring the turbidity of a water sample (USGS, 2005).

The U.S. EPA recently reviewed the methods available for measuring turbidity in drinking water and has approved four versions of APHA/AWWA/WEF Standard Method 2130B, which were published in 1991, 1995, 1998 and 2005 (U.S. EPA, 2008). Of the methods listed in Table 3, the U.S. EPA has also approved U.S. EPA Method 180.1 Rev. 2.0, GLI Method 2 and Hach FilterTrak Method 10133 Rev. 2.0.

5.1 Instrumentation

Nephelometric turbidity instrumentation varies in design, range, accuracy and application. The design of nephelometric instruments should take into account the physics of scattered light. The size, shape and concentration of the particles affect the intensity pattern and distribution of the scattered light. Small particles less than one tenth of the light wavelength will scatter light uniformly in both forward and backward directions. As the particle size approaches and exceeds the wavelength of the incident light, more light is transmitted in the forward direction. Because of this intensity pattern, the angle at which the light is measured is a critical factor; the current international standards have determined the most appropriate angle to be 90 degrees for the measurement of low turbidities (generally below 40 NTU) (APHA et al., 2012). Nephelometric turbidimeters can also include ratio technologies which are based on the use of a 90 degree detector in combination with another detector set at a different angle that determine the turbidity of a sample. Ratio technologies can help to compensate for interferences due to colour and particulate absorbance that are common in turbidity measurements (ASTM International, 2011). As noted above, as the concentration of particles increases, more particles reflect the incident light, increasing the intensity of the scattered light. Once the concentration of particles in a sample exceeds a certain level, which is determined by the specific optical characteristics of the process, the particles themselves begin to block the transmission of the scattered light. The result is a decrease in the intensity of the scattered light, which establishes the upper limit of the measurable turbidity (Sadar, 1998). Nephelometers are most effective for measuring light scattered by particles in the 0.2-1 µm size range, with a peak scatter at approximately 0.2 µm. The intensity at which various wavelengths of light are reflected or absorbed is also determined by the colour of the liquid and the reflecting surface. Industry standards require nephelometers to operate in the visible or infrared ranges: 400-600 and 800-900 nm, respectively (ISO, 1999; APHA et al., 2012).

All of these factors, along with the optical geometry of a particular instrument, cause measured values between instruments to vary widely; thus, criteria for instrument design have been developed to minimize these variables. The manufacture of nephelometric turbidimeters is guided by the instrument design requirements that are specified in the standards listed in Table 3.

Measurement technologies other than the nephelometric technique discussed above are available and vary by the light source type, the number of detectors and the detection angles that are used to obtain a turbidity measurement. In particular, different technologies may be more suitable for measuring higher levels of turbidity (generally greater than 40 NTU) or for measuring turbidity in the presence of colour. These technologies include ratio, surface scatter, back scatter, forward scatter and multi-beam techniques. Recently, a consensus-based guide on the application of various technologies for turbidity measurement has been developed and may aid readers in selecting the most appropriate technology for their water type (ASTM International, 2011).

5.2 Instrument performance

Filtered water turbidity is typically well below 1.0 NTU and is often below 0.1 NTU. Certain filtration methods, such as reverse osmosis, can achieve turbidity values that approach those of pure water, in the range of 0.010-0.015 NTU. The sensitivity of turbidimeters and the precision and accuracy of the measurements at low turbidity levels are important aspects in the practical application of turbidity monitoring (Sadar, 1998).

5.2.1  Sensitivity

According to U.S. EPA Method 180.1, GLI Method 2 and APHA/AWWA/WEF Standard Method 2130B, nephelometers designed under these methods should be able to detect turbidity differences of 0.02 NTU or less in waters having a turbidity of less than 1.0 NTU. All three methods state that turbidity readings should be reported to the nearest 0.05 NTU when the turbidity range is 0-1.0 NTU. ISO 7027 (ISO, 1999) indicates that results should be reported to the nearest 0.01 FNU when turbidity is below 0.99 FNU. ASTM D6855-10 for the static measurement of turbidity states that the resolution of the instrument should permit detection of turbidity differences of 0.01 NTU or less in waters with a turbidity of less than 5.0 NTU. Results should be reported to the nearest 0.01 NTU for water with turbidity of less than 1.0 NTU and to the nearest 0.05 NTU for water with turbidity between 1.0 and 5.0 NTU (ASTM International, 2010). ASTM D6698-07 for online turbidity measurements states that turbidity differences of 0.01 NTU or less should be detected in water with a turbidity less than 1.0 NTU and that differences of 0.10 NTU or less should be detected in waters with turbidity between 1.0 and 5.0 NTU. Results should be reported to the nearest 0.01 NTU for water with turbidity less than 1.0 NTU and to the nearest 0.1 NTU for water with turbidity between 1.0 and 5.0 NTU (ASTM International, 2007).

Laser turbidimeters, although more costly, are another option for measuring turbidity and typically have a higher sensitivity than standard nephelometric meters. Hach FilterTrak Method 10133 (Determination of Turbidity by Laser Nephelometry) has an applicable range of 0-5000 mNTU (0-5.0 NTU) (Hach Company, 2000). This method states that the instrument has a sensitivity that should permit the detection of a turbidity difference of 1 mNTU (0.001 NTU) or less in waters having turbidities less than 5.0 NTU. It is suggested that laser turbidimeters are better suited for monitoring treated water from membrane filtration because of the extremely low levels of turbidity that can be achieved using this treatment method. Research has indicated that the increased sensitivity of laser turbidimeters may make them more effective than standard nephelometers at detecting membrane integrity breaches (Banerjee et al., 1999, 2001; U.S. EPA, 2005). More recent studies also suggest that laser nephelometers are capable of measuring early end-of-run filter breakthrough and other very small increases in turbidity that are useful for conventional filtration plant optimization (Sadar and Bill, 2001; Sadar et al., 2009). Sadar et al. (2009) also demonstrated that measurement of a submicrometre (<0.01 µm) particle breakthrough event was possible using a laser nephelometer and that the sensitivity of laser nephelometers was equivalent to that of particle counters.

5.2.2  Performance

Several studies have evaluated the performance of turbidimeters in measuring turbidity in the range of 0.1-0.3 NTU. The U.S. EPA conducted a study of the ability of different types of turbidimeters to measure low turbidity levels by distributing standard suspensions with a reported value of 0.150 NTU to a variety of laboratories. The results indicated that benchtop, portable and online turbidimeters all had a positive bias compared with the true value of the samples provided, with results between 0.176 and 0.228 NTU. This suggests that errors in turbidimeters may be conservative from a filtered water perspective; that is, plants may actually achieve slightly lower levels than those indicated on the meter. The standard deviations on the samples analyzed by each type of meter ranged from 0.0431 to 0.0773 NTU (U.S. EPA, 2003b). Similarly, ASTM conducted an interlaboratory study of static turbidimeters (benchtop or portable). A standard sample with a turbidity of 0.122 NTU was provided to seven laboratories, and the precision and bias of the laboratory measurements were calculated. This study found a laboratory standard deviation of 0.0190 NTU and a single analyst standard deviation of 0.0089 NTU (ASTM International, 2010). This indicates that there may be some variability between measurements obtained from different laboratories; however, when a single analyst is employed, the standard deviation can be quite low.

Letterman et al. (2002) conducted a detailed evaluation of the effect of turbidimeter type, design and calibration method on low-level turbidity measurements. The authors found that factors such as light source and calibration material did not have a significant effect on turbidity measurements using benchtop or portable instruments. The calibration procedure did, however, have a significant effect on the turbidity measurements and resulted in two categories of instruments. One group of instruments (Group A) used a calibration procedure to automatically set a low particle reading sample at either 0.00 or 0.02 NTU. This group of instruments had lower average readings than the second group (Group B) of instruments, which did not automatically assign a predetermined reading to a low particle sample. When the turbidity of a sample was less than 0.15 NTU in the Group A instruments, the Group B instruments measured between 0.00 and 0.02 NTU.

Letterman et al. (2002) also evaluated online turbidimeter performance. The study found poor agreement between different online instruments, with an average range in turbidity measurements of 0.5 NTU. The authors believed that bubble interference may have resulted in some of the discrepancies between instruments. In contrast, ASTM International (2007) conducted an independent intralaboratory study of online instruments and found that the relative standard deviation varied between 7.3% and 12% for different instruments measuring a turbidity standard of 0.1 NTU. Although some degree of interinstrument variability has been demonstrated, it is generally believed that low-level turbidity measurements can be used as a performance indicator for achieving very high quality filtered water (less than 0.1 NTU) and as an indicator of treatment plant optimization within one treatment plant (U.S. EPA, 2006b).

Overall, currently available instruments are capable of measuring turbidity reliably at levels below 0.1 NTU. However, analysts must be aware of the factors that can affect turbidity measurements and be careful to minimize potential sources of measurement error. In addition, low-level turbidity measurement must be accompanied by careful instrument calibration and verification as well as comprehensive standard operating procedures, including rigorous analyst training (U.S. EPA, 2003b).

5.3 Quality assurance/quality control

As discussed above, in order to be able to accurately measure turbidity below 0.1 NTU, rigorous standard operating procedures and a high level of quality assurance and quality control (QA/QC) are required. Utilities should ensure that the appropriate operation, maintenance and calibration programs are in place for all turbidimeters. For example, all utilities should have operating procedures for cleaning turbidimeters, creating or using standards, sampling and calibrating turbidimeters. It is recommended that utilities calibrate online turbidimeters at least quarterly, or more frequently if recommended by the manufacturer. The calibration of turbidimeters should then be verified weekly with the appropriate standard and re-calibrated if the turbidimeter has drifted more than 10% from the value assigned to the standard. Most of the analytical methods listed in Table 3 include detailed information on the preparation of appropriate standards for calibration and the calibration procedure for turbidimeters. Preventive maintenance should also be part of a routine turbidimeter QA/QC program. Weekly inspections and regular cleaning of lenses, light sources, sample reservoirs, air bubble traps and sample lines are important to ensure proper operation of the turbimeter. A detailed discussion of the development of QA/QC programs can be found in the literature (Burlingame et al., 1998; Sadar, 1998; U.S. EPA, 1999, 2004).

Other factors, such as air bubbles, stray light, coloured water and particle contamination, should also be considered in QA/QC programs, as these can cause false high or low readings for turbidity (Burlingame et al., 1998; Sadar, 1998; APHA et al., 2005). In some cases, the factors listed above can have a significant effect on turbidity measurements. A recent study of bubble interference in online turbidimeters found that bubbles can cause turbidity spikes as large as 2.0 NTU, depending on the type of instrument used and the level of gas supersaturation in the sample (Scardina et al., 2006).

Several of the methods listed above also provide guidance on sampling and sample handling. As the turbidity of a sample can change due to changes in temperature and particle flocculation and sedimentation, samples should be analyzed immediately (ISO, 1999; ASTM International, 2010; APHA et al., 2012). It is recommended that samples be analyzed using on-site turbidimeters in the treatment plant or portable turbidimeters when conducting sampling in the field.

5.4 Particle counting

Electronic particle counters are now available that are capable of accurately counting and recording the number of particles as a function of size (often in the 1-150 µm range). Although in some cases there may be a general relationship between particle counts and turbidity at levels below 1.0 NTU, a direct correlation does not exist (Bridgeman et al., 2002).

A simple conversion factor relating particle counts to turbidity is not possible, because the two techniques for their measurement differ fundamentally in terms of discernment. Particle counting measures two characteristics of particulates: particle number and particle size. Samples with identical clarity can be distinguished on the basis of these two features; one sample may contain many small particles, whereas another may contain a few large particles. Turbidity, on the other hand, cannot distinguish between two samples of identical clarity and different particulate composition. It is difficult to correlate turbidity with the particle concentration of suspended matter. As the size, shape and refractive index of particles affect the light-scattering properties of the suspension, they therefore, affect the turbidity (APHA et al., 2012). In addition, turbidimeters can detect particles smaller than 1 µm, whereas the lower size for detection by particle counters is in the range of 1-2.5 µm. As a result, data from the two instruments may not correlate.

Particle counters are an excellent tool for optimizing treatment processes and for detecting the onset of filter breakthrough. They are restricted to performance verification only, and no limit is set as a maximum acceptable concentration for the number of particles in the treated water.

Page details

Date modified: