Hygrometry

Chapter 10 Hygrometry Lars-Olof Nilsson, Kurt Kielsgaard Hansen and Miguel Azenha 10.1 General Principles The humidity of air can be measured with a number of various methods, i.e. psychrometers, dew-point sensors, mechanical hygrometers and electronic RH-probes (RH = Relative Humidity). In most cases when measuring RH in materials a number of electronic RH-probes are used. The sensors in these probes are, in most cases, capacitive or resistive where a small piece of a sensor material (typically a polymer) has two electrodes between which the electrical capacitance or resistance is measured. The capacitance or the resistance between the electrodes is a measure of the moisture content of the sensor material. The readings must be translated to RH with a calibration. It is noted however that capacitance-based sensors are more disseminated, Rittersma (2002), probably because of their improved performance in harsh environments as compared to resistive sensors. Some RH-probes have dew-point sensors where a small mirror is chilled until the temperature reaches the dew point of the surrounding air and condensation occurs at the mirror. The detection of condensation is done in different ways. L.-O. Nilsson (&) Lund University, Lund, Sweden e-mail: lars-olof.nilsson@byggtek.lth.se L.-O. Nilsson Moistenginst AB, Trelleborg, Sweden K. K. Hansen Technical University of Denmark, Kongens Lyngby, Denmark e-mail: kkh@byg.dtu.dk M. Azenha University of Minho, Braga, Portugal © RILEM 2018 L.-O. Nilsson (ed.), Methods of Measuring Moisture in Building Materials and Structures, RILEM State-of-the-Art Reports 26, https://doi.org/10.1007/978-3-319-74231-1_10 73 74 L.-O. Nilsson et al. The thermoelement psychrometer works by use of the thermoelectric effect, the Peltier effect, that cools down the welded junction below the dew point and water from the surrounding condenses on the welded junction. When stopping the cooling process the condensed water evaporates from the welded junction and creates a thermoelectric effect (i.e. voltage). The voltage is measured as function of time and the nick point when stopping the cooling process is determined from which the RH can be determined. This measurement principle was introduced by Spanner (1951) and further developed by Wiebe (1971) and is today commercialized, see Table 10.1. Some RH-indicators are using a hygroscopic salt that is coloured. When RH is reaching a certain threshold level, determined by the type of salt, the salt is accumulating water and is spread. By containing the salt in a cloth, the movement of the coloured salt can be observed. The movement itself says that RH has reached a level above the threshold. The magnitude of the movement can say something about for how long time RH has been above the threshold. Alfasensor AB has commercialized this technique, see Alfasensor (2016). The swelling and shrinkage due to moisture changes is used in some humidity meters like in traditional hair hygrometers. A bundle of hair whiskers is mechanically connected to a hand that moves when the bundle is shrinking and swelling. This phenomenon was used by Monfore (1963) who connected a thread of Dacron (a polymer) to a very thin nickel-copper wire. Humidity changes caused the Dacron thread to swell and shrink and that changed stretched the nickel-copper wire more or less. This change in elongation could be measured as electrical resistance changes. The arrangement of the polymer thread and the nickel-copper wire was done inside a long, perforated tube with a diameter of 2.5 mm that constituted an RH-probe that could be inserted into small, long holes in concrete specimens. A polymer optical fibre sensor is a single-mode optical fibre with a Fiber Bragg Grating (FBG) inscribed into the fibre close to the tip of the fibre. A FBG is a periodic series of perturbations in the refractive index of the optical fibre. The FBG reflects a specific wavelength portion of the light impinged on it; the wavelength of the reflected portion is changed when the FBG is exposed to changes in strain, temperature or humidity. The change in reflection is due to the expansion or contraction of the FBG, which changes the periodicity of the pertubations. 10.2 Principles for Measuring in Materials The principle of using hygrometry to measure the moisture condition of a material is to utilize an instrument that can measure the RH of air and place it in a closed “space” in contact with the material. The “space” can be 10 Table 10.1 Information on some frequently used RH-probes and their sensors Sensor material Principle Humi-guard Woven polymer + salt solution Birch “Some kind of polymer” Resistive Sahlén Sensirion SHTW2 SHTC1 SHT1x SHT2x SHT3x SHT7x Testo 605-H1 Vaisala Polymer Size probe/sensor (mm)/(mm3) Resistive Capacitive U12  45/ 1.3  0.7  0.5 2  2  0.8 7.5  4.9  2.6 3  3  1.1 2.5  2.5  0.9 19.5  5.1  3.1 Capacitive Capacitive U12  125/ 144 Electrodes dw/dRH (mg/% RH)a Two straight metallic wires Two parallel nails Confidential 0.02–0.06 Two metal sheets; one porous Ref (1) 64–98 0.07 0.07 (2) (3) (4) (5) 93– 99.5 (6) (7) (8) 75 Hygropin U5 Shute Optical fiber Thermo couple Thermo Wescor PCT/ PST-55 electric a www.rbk.nu 1. www.industrifysik.se and private communication 2. http://www.fuktbutiken.se/sv/rf-matare/sahlengivaren-inkl-05m-kabel.html 3. https://www.sensirion.com/products/digital-humidity-sensors-for-reliable-measurements/ and private communication 4. http://www.testolimited.com and private communication 5. www.vaisala.com and private communication 6. http://www.proceq.com 7. http://shute.dk/ and private communication 8. http://water.wescor.com/pct55.html Range (% RH) Hygrometry Sensor 76 – – – – – L.-O. Nilsson et al. a a a a a container with a sample of the material, “cup” at a surface of the material, space in a material or between two materials in a structure, drilled hole in a material, cast hole in a concrete specimen. A few examples of “spaces” where the RH of a material is measured are shown in Figs. 10.1, 10.2 and 10.3. The sample or the surface of the material gives off a small amount of moisture that is enough to humidify the enclosed air volume and the RH-probe, with its sensor and its filter. The RH-probe must have a low moisture capacity so this gain of moisture does not cause too large an error. This is further analysed in Sect. 10.6.2. A certain time is needed to reach equilibrium conditions in the measuring situation. The time required depends on the geometry, the properties of the material and the properties of the RH-probe. The traditional way to handle this is to simply wait until a steady reading is obtained. Fig. 10.1 Examples of “spaces” where the RH of a material can be measured. From Nilsson (1980) Fig. 10.2 Examples of “spaces” where the RH at the surface of a material can be measured. From Nilsson (1979) 10 Hygrometry 77 Fig. 10.3 Example of “space” where RH is measured by a wooden sensor, inside an embedded, porous polyethylene cylinder, Sahlén (2016) 10.3 Sensors for Hygrometric Measurements with RH-Probes Today, most measurements of humidity in materials are done with a small electronic probe that reacts to the humidity in the surroundings of the probe. The decisive part of such probes is the sensor and the sensor material. As said in Sect. 10.1, the principles of sensors used for measuring RH in a material are a. small size to be able to measure in a defined, local point, b. equilibrium between the sensor material and its immediate surroundings, c. a measure of the moisture content in the sensor itself is obtained with one of several techniques: • electrical resistivity • electrical capacity • length change d. alternatively, the vapour content of the immediate surrounding air is determined by chilling the sensor and detect at what (dew-point) temperature condensation occurs; together with a temperature measurement the RH can be quantified. A large number of probes are available on the market. In many cases it is not always clear what the principles are behind each type of probe. The manufacturers are of course hiding information that is company secrets but some general information should be available to make it possible for the user to understand what is happening during a measurement. In Table 10.1 frequently used probes and sensors are described in more detail. The information is what could be obtained from open sources and directly 78 L.-O. Nilsson et al. from the manufacturers. Some information is missing because it is a company secret. A very important property of an RH-probe is the moisture capacity dw/dRH, i.e. the amount of moisture (from the surroundings) that is required for the RH-reading to change. It should be quantified as e.g. mg/%RH. All probes have a sensor that has a more or less significant moisture capacity. Additionally, a number of probes have a filter protecting the sensor and it is that filter that has the large moisture capacity. Some systematic errors are directly linked to the moisture capacity of the probe. Consequently, information about the moisture capacity of the sensor and the filter must be available. 10.4 Temperature Effects The temperature conditions are crucial when using a hygrometric method. A couple of different temperature effects are affecting the measurements. 10.4.1 Temperature Difference, Sensor—Material Any temperature difference between the surface of the material and the sensor of the RH-probe will cause a systematic error. A rough estimation of this error is, Nilsson (1988), DRHTdiff a  5
DT ¼ 5
ðTmtrl Tsensor Þð%Þ ð10:4:1Þ A temperature difference of 1 °C gives an error of about 5% RH. 10.4.2 Temperature Difference, Measurement—Sampling If the measurement is made at another temperature than where the sample was taken, the measured RH will be different from what the RH was where the sample was taken. The effect of this temperature difference can be calculated, with some uncertainty, if required. The “correction” of the measured RH is done by 10 Hygrometry 79 Fig. 10.4 Examples of the temperature dependency of the sorption isotherms for concretes at +5 to +20 °C (left), Sjöberg et al. (2002), and wood (right), Nilsson et al. (2006), expressed as dRH/dT DRHTdiff b  dRH=dT
DT ¼ dRH=dT
ðTin situ Tmeasure Þð%Þ ð10:4:2Þ where dRH/dT follows from the temperature dependency of the sorption isotherm. The correction factor dRH/dT depends on the material, the RH-level and the temperature level. Two examples are given in Fig. 10.4. Temperature effects when the temperature is varying in time in on-site measurements, creating temperature differences between the sensor and the material, are dealt with in Chap. 27. 10.5 Calibration The principle of calibrating a hygrometric probe or sensor is to create a known RH, at a certain temperature, in an air volume and read the signal from the sensor/probe after equilibrium is achieved. The ways to create a known RH are several and are described in Sect. 4.3. A calibration must include test for linearity, drift over time, hysteresis and repeatability, that all affect the result of the RH-reading. These four phenomena are described briefly in the following and selected results are shown. 10.5.1 Tests by Use of Saturated Salt Solutions The tests are all performed in a box to prevent disturbances from wind, etc. The calibration vessels in the box are placed in a piece of rigid insulation material to secure a constant temperature and to prevent direct exposure of light. The calibration vessel is shown in Fig. 10.5. 80 L.-O. Nilsson et al. Fig. 10.5 A calibration vessel with a saturated salt solution separated from the air around the RH-probe by a semi-permeable membrane, Hansen and Christensen (1998) 10.5.2 Linearity of the Calibration Curve Linearity of the calibration curve is important to ensure a precise determination of a RH-reading between two calibration points on the calibration curve. Increasing the number of salt solutions in the calibration process can minimize this factor. The linearity is evaluated according to calibration curves applied to the measured RH-readings found under desorption by a linear curve fit or a 2nd order polynomial curve fit, respectively. Figure 10.6 shows an example of non-linearity of the calibration curve for an RH-instrument. It is seen that at 87.8% RH the difference between the linear curve fit and the 2nd order polynomial fit is 1.2% RH. As the 2nd order polynomial curve fits the measuring points very well a polynomial curve should normally be used. In Fig. 10.6 (right) the maximum deviation from linearity for four types of RH-instruments is compared. 10.5.3 Drift Over Time Drift causes changes in the RH-reading of an instrument even at constant RH. Having drift in an instrument the RH-reading changes as time goes for the same RH. The drift might vary from time to time. In Fig. 10.7 an example of drift over time for a RH-instrument through a 55 days period is shown. It is seen that the 10 Hygrometry 81 Fig. 10.6 An example of non-linearity of the calibration curve for an RH-instrument (left). Maximum deviation from linearity for four types of RH-instruments (right). From Hansen and Christensen (1998) Fig. 10.7 An example of drift over time for a period of 55 days (left). Maximum drift over time for a period of 55 days for four types of RH-instruments (right). From Hansen and Christensen (1998) maximum drift is 1.6% RH at 85% RH. If a RH-instrument has high drift over time, the RH-instrument must be checked regularly by use of saturated salt solutions. In Fig. 10.7 maximum drift over time for four types of RH-instruments is compared. It is relevant to mention that capacitive sensors in particular are susceptible to deviate from initial calibration upon subjection to high humidity environments (e.g. >90%) for relatively small periods. It is therefore recommendable to perform calibration checks after exposure to high humidity, Granja et al. (2014). 82 L.-O. Nilsson et al. 10.5.4 Hysteresis A calibration curve performed under absorption (wetting) and a calibration curve performed under desorption (drying) might not be identical. The difference is called hysteresis. If the RH-measurement in the concrete is done over a short period the best result will be to correct the actual RH-reading with a calibration curve made under absorption. If the sensor is placed in the concrete for a longer period, where the sensor is drying together with the concrete, the best results will be to correct the actual RH-reading with the calibration curve made under desorption. In Fig. 10.8 an example of hysteresis for a RH-instrument is shown. It is seen that at 90% RH the difference between absorption and desorption values is 2.5% RH. In Fig. 10.8 the maximum hysteresis for four types of RH-instruments is compared. 10.5.5 Repeatability Repeatability for a RH-instrument is a characteristic of goodness to obtain the same RH-reading when measurements are repeated after a short while. The repeatability is determined as the difference between two desorption curves performed immediately after each other. In Fig. 10.9 an example of repeatability for a RH-instrument is shown. It is seen that at 75% RH the difference between two readings is 0.9% RH. In Fig. 10.9 the repeatability for four types of RH-instruments is compared. Fig. 10.8 An example of hysteresis and maximum hysteresis for four types of RH-instruments. From Hansen and Christensen (1998) 10 Hygrometry 83 Fig. 10.9 An example of repeatability and repeatability for four types of RH-instruments. From Hansen and Christensen (1998) 10.6 Errors and Uncertainties 10.6.1 General The errors can be divided in systematic errors and in accidental errors. According to Hedenblad (1995) the systematic errors are: a. b. c. d. e. f. Systematic error for the saturated salt solution used at the calibration. Missing linearity for the RH-instrument. The measuring temperature is not equal to calibration temperature. Drift for the RH-instrument. Temperature difference between the RH-sensor and the concrete. The calibration is performed at another temperature than the temperature at the measuring site. The accidental errors are: g. Hysteresis for the RH-instrument. h. Calibration error. i. Influence of temperature variation during measurement. Hedenblad (1995) gives an example on calculation of the total uncertainty from the errors listed above. 10.6.2 Systematic Error Due to Moisture Capacities The sample of a material, or the “contributing volume” of the material, gives off a small amount of moisture to humidify the enclosed air volume and the RH-probe. 84 L.-O. Nilsson et al. This amount of moisture causes a drop in RH that will be a systematic error. The error is, Nilsson and Åhs (2012), DRH ¼ mw ðRH0 ; RH Þ þ vs ðTÞ
ðRH RH0 Þ
Vair Vsample
KðRHÞ ð10:6:2Þ where mw(RH0, RH) is the moisture capacity (kg) of the RH-probe between the initial humidity RH0 and the measured RH; vs is the vapour content (kg/m3) at saturation at the current temperature T; Vair is the volume (m3) of the enclosed air; Vsample is the volume (m3) of the sample; K is the moisture capacity (kg/m3) of the sample material at the humidity level RH. This systematic error is shown in Fig. 10.10 for measuring on 20 g sample of a material with a density of 2400 kg/m3. The error is directly proportional to the inverse of the density. As seen in Fig. 10.10 the systematic error is small if the measurement is done on a material with a moisture capacity larger than some 50 kg/m3 but it is very dependent on the moisture capacity of the RH-probe. This systematic error has been verified in an extreme case when measuring RH on small samples of high-performance concrete that has a very low moisture capacity, cf. Fig. 10.11. Hedenblad (1999) made a series of ten identical measurements on the same sample in a test tube and got a repeated difference of −0.6% RH for each measurement. The relevance of the size of the measuring pocket and the moisture capacity of the sensor itself have been found to be small to negligible in the cases of embedded probes for humidity measurement in concrete, Granja et al. (2014). Indeed, this embedding technique balances the measuring pocket and sensor with an inner surface of the measured concrete, which has enough moisture capacity to feed the measurement pocket/sensor without overall disturbance. Fig. 10.10 Systematic error when measuring RH on a small sample (20 g) of material with a density of 2400 kg/m3 10 Hygrometry 85 Fig. 10.11 Systematic error of −0.6% RH per measurement in ten repeated RH-measurements on a small sample of high-performance concrete, Hedenblad (1999) In a recent study Johansson (2014) found large systematic errors, several % RH, when measuring RH on samples from low w/c concrete. The main reasons for these errors are the very low moisture capacity of the material and drying of the samples before being put into the measuring container. Parallel measurements in holes in the same concrete showed a very much-reduced systematic error mainly because no significant drying of the hole occurs. An example from Johansson (2014) is shown in Fig. 10.12. Fig. 10.12 Comparison between RH-measurements on samples and in drilled holes in a concrete slab with low w/c. Data from Johansson (2014) 86 L.-O. Nilsson et al. 10.6.3 Interface Between Measured Material and the Artificial Measurement Hole Grasley et al. (2014) have proposed the use of Gore-tex® membranes for protection of cast-in measurement sleeves in concrete. Gore-tex® is a densely knit fabric, known to be impermeable to liquid water, while allowing vapour transport. Tests performed in cement paste by Granja et al. (2014) have compared the measurement performance of cast-in holes with and without Gore-tex® applied between the end of the hole and the measured material. Such comparison led to the conclusion that this type of membrane does not seem to affect the accuracy of measurements. 10.6.4 Permanently Installed Probes or Intermittent Measurements? Most commercially available systems for humidity measurement in concrete through artificial holes (either drilled or cast-in) are based in the intermittent placement of humidity probe in the hole for taking measurements. In the periods when measurements are not being taken, the hole is sealed with a cap. In tests carried out to compare the performance of this strategy in regard to a potential permanent installation of the humidity probe (i.e. without disturbances associated to the opening/closing cycles of the cap), led to the observation of negligible differences in measurements along a two-month period, Granja et al. (2014).