Measurement Techniques, Vol. 55, No. 10, January, 2013

THERMAL MEASUREMENTS THE DEVELOPMENT OF EMISSIVITY MEASUREMENTS UNDER VACUUM AT THE PTB

A. R. Adibekyan,1 C. Monte,1 M. Kehrt,1 S. P. Morozova,2 B. Gutschwager,1 and J. Hollandt1

The consistency of the values of the spectral radiance of two standard vacuum absolutely black bodies (low and medium temperature) in the 80–170°C range, forming part of the equipment for measuring emissivity under vacuum, is considered. It is shown that, using this equipment, one can measure spectral radiance at low radiation temperatures (down to –100°C) in the 4.5–20 µm wavelength band. The characteristics of the inner coating of the sample enclosure, required in order to make the measurements in the far infrared band and compatible with the vacuum conditions and low temperatures, are presented. Keywords: emissivity, vacuum, spectral radiance, absolutely black body.

Measurements of the dimensionless quantity, directional spectral emissivity, are being carried out at the PTB at the present time using equipment which operates in air at a temperature of 80–500 °C in the 4–40 μm spectral band, and are a routine metrological service. New apparatus for measuring emissivity under a vacuum, which is part of a low-background calibration facility (LBCF), considerably extends the possibilities of these measurements. This equipment enables samples to be investigated at temperatures of 0–600°C in an angular range of ±75° and in the spectral range from 0.4 μm to 1400 μm. Its use also reduces the uncertainty of emissivity measurements, since two main components of the uncertainty, which arise when finding the temperature of the sample being investigated, are removed. These losses are due to convection of the air at high temperatures (this value of the uncertainty is necessary for an accurate determination of the surface temperature of the sample in air), and as a result of atmospheric absorption when the partial pressures of H2O and CO2 change. The LBCF equipment, shown in Fig. 1, is described in detail in [1]. It contains several reference sources: two vacuum absolutely black bodies (ABB) with a regulated radiation temperature, situated in the radiation-source chamber – a low-temperature black body (LTBB) [2] operating in the –173 to +170°C temperature range, and a medium-temperature (MTBB) [3], for use in the 80–430°C temperature range. The third (cold) reference source – a liquid-nitrogen cooled ABB – is mounted in the upper part of the optical-mechanical unit. The chamber containing the radiation sources also has additional space to accommodate a test source for calibration, or a spherical enclosure with a sample for measurement of emissivity inside it (see Fig. 1b). The radiation emerges through a circular opening in the front part of the sphere. Two standard radiators – the MTBB and the LTBB – are mounted on the right on a mobile table. Two schemes for recording the radiation are currently being used: 1) a Fourier transform infrared spectrometer which records spectral radiance in the 0.4–1400 μm band. In this case, 1 2

Physikalisch-Technische Bundesanstalt (PTB), Berlin, Germany; e-mail: [email protected]. All-Russia Research Institute of Optophysical Measurements (VNIIOFI), Moscow, Russia.

Translated from Izmeritel’naya Tekhnika, No. 10, pp. 31–36, October, 2012. Original article submitted May 15, 2012.

0543-1972/13/5510-1163 ©2013 Springer Science+Business Media New York

1163

the radiation from the ABB or the sample is reflected from an off-axis ellipsoidal mirror mounted on the mobile table in the radiation-receiver chamber (see Fig. 1a); 2) a vacuum standard infrared radiation pyrometer [4] which records the radiation temperature in the –170 to +170°C range in the 8–14 μm spectral band. The vacuum Fourier transform spectrometer has several receivers, which can be used for measuring the emissivity: an InSb receiver (wavelength range 1–5 μm), a mercury cadmium telluride (MCT) receiver (covering the range 2.5–20.0 μm), a pyroelectric DLaTGS receiver for the mid infrared range (2.5–25 μm), a pyroelectric DTGS receiver for the far infrared range (15–200 μm), and a liquid-helium cooled Si-composite bolometer for the far infrared range (10–1400 μm). All parts of the optical path of the low-background calibration facility (LBCF) (see Fig. 1a) – the optical-mechanical unit, the optical radiation output channel, which passes between the source chambers and the radiation receivers, and also all the apertures, are liquid-nitrogen cooled. As a rule, individual parts are at a temperature below –150 °C. The sample holder for measuring the emissivity (see Fig. 1b) is temperature regulated over the range 0–600 °C. The heater with the sample can be rotated using a highly accurate electric motor, which enables measurements to be made in an angular range of ±75° about the normal to the surface. The temperature of the spherical enclosure can be controlled in the range from –80 to +80°C, and on its inner surface V-shaped grooves have been cut and given a special coating of Nextel Velvet Black 811-21. The purpose of this paper is to show the following: 1) the consistency of the two standard radiators – LTBB and MTBB – which is a necessary condition for correct operation of the emissivity measuring system; 2) the ability of the LBCF to measure radiances, corresponding to values of the radiation temperature, and to measure the direction spectral emissivity of samples in the temperature range from well below 0 up to 170°C; 3) the suitability of the inner coating of the spherical enclosure for measuring the spectral emissivity, to investigate the effect of the coating on the temperature stability, and also to answer the question of whether this structure and the type of coating is capable of considerably reducing the radiation incident from it on to the sample. The Spectral Radiance of the Reference Radiators of the LTBB and the MTBB. The calculation procedure for measurements of the emissivity in a vacuum, described in [1], is based on a comparison of the results of successive measurements of the spectral radiance of the sample being investigated, which is inside the temperature-stabilized spherical enclosure, and reference absolutely black bodies at different temperatures. The temperatures of one absolutely black body TABB2 and of the sample Ts are close, while the temperature of the other TABB1 differs considerably from Ts: L˜ (T ) − L˜ ABB1(TABB1) Q= ˜ s s . LABB2 (Ts ) − L˜ ABB1(TABB1)

(1)

Here and below 0 denotes the measured value of the spectral radiance L. The necessary condition for using this method of measurement is the consistency of the two reference absolutely black bodies. To check this condition, we measured and compared the spectral radiances of the LTBB and of the MTBB at 80, 110, 140, and 170°C. Measurements and estimation procedure. To eliminate drift and background radiation, we made sequential measurements each time, including a recording of the signal and background (from the liquid-nitrogen cooled ABB). In this experiment, the measured spectral radiance signal of the ABB is given by 0ABB(TABB) = s(LABB(TABB) + Lb – Lr),

(2)

where s is the spectral sensitivity of the spectrometer, LABB(TABB) = εABBLP(TABB) is the spectral radiance of the ABB, determined by the value of LP(TABB), found from Planck’s law at the temperature and effective directional spectral emissivity εABB of the ABB, and Lb and Lr are the spectral radiances of the thermal background of the LBCF and the receiver, respectively. The measured spectral radiance of the “cold” liquid-nitrogen cooled reference ABB is defined similarly (see (2)): L˜ ABB(LN ) (TABB(LN ) ) = s(ρm LABB(LN ) (TABB(LN ) ) + Lb + Lm (Tm ) − Lr ). 2

1164

2

2

2

(3)

Fig. 1. Outside view of the low-background calibration equipment (a) and the opened chamber containing the radiation sources (b).

Here ρm is the directional spectral reflection coefficient of the rotating disk (modulator), which reflects the radiation of the “cold” reference ABB with spectral radiance: L ABB(LN

2)

(TABB(LN

2)

) = ε ABB(LN ) L P (TABB(LN 2

2)

)

and directs it along the optic axis, and Lm(Tm) = εmLP(Tm) is the spectral radiance of the additional radiation of the rotating disk with high reflection coefficient, which adds only to the background measurements. The difference of equations (2) and (3) enables both the background radiation Lb and the natural radiation of the receiver Lr to be eliminated: 0 ABB (TABB ) − 0 ABB(LN

2)

(TABB(LN

2)

) = s ( L ABB (TABB ) − ρ m L ABB(LN

2)

(TABB(LN

2)

) − L m (Tm )).

(4)

Successive averaging of these differences during the measurements also enables the effect of thermal drift of the apparatus to be eliminated. In the experiment, for each absolutely black body, according to (4), a sequence of radiance differences is measured. To compare two such bodies, it is assumed that the radiation of one absolutely black body (in this case the LTBB) is determined by its effective emissivity and Planck’s law at its working temperature. The deviation of the radiance temperature of the other ABB from its working temperature is a measure of the consistency of both bodies. The radiance of the MTBB, obtained from the ratio of two independent equations (4) for the LTBB and MTBB, can be calculated from the formula L MTBB (T ) =

0 MTBB (T ) − 0 ABB(LN

2)

(TABB(LN

2)

)( L LTBB (T ) − ρ m L ABB(LN

0 LTBB (T ) − 0 ABB(LN

2)

(TABB(LN

2)

) − L m (Tm ))

(T ) 2 ) ABB(LN 2 )

+ Lm (Tm ) + ρm LABB(LN ) (TABB(LN ) ). 2

+

(5)

2

The value of the spectral radiance of the rotating disk can be calculated from the results of an independent series of measurements of the LTBB at temperatures T1 and T2. It follows from (4) that L˜ ABB (T1,2 ) − L˜ ABB(LN ) (TABB(LN ) ) = s( LABB (T1,2 ) − ρm LABB(LN ) (TABB(LN ) ) − Lm (Tm )). 2

2

2

2

1165

Fig. 2. Spectral distribution of the radiation temperature of the MTBB, determined with respect to the LTBB, as given by (5), at nominal temperatures of 80°C (a) and 170°C (b); the gray area represents the range of total expanded uncertainty for the radiation temperature.

After reduction, we have Lm (Tm ) = LABB (T2 ) − ρm LABB(LN ) (TABB(LN ) ) − 2

−

2

( L˜ ABB (T2 ) − L˜ ABB(LN ) (TABB(LN ) ))( LABB (T1) − LABB (T2 )) 2 2 . L˜ (T ) − L˜ (T ) ABB

1

ABB

(6)

2

Comparison of the absolutely black bodies. The measurements were made using a vacuum Fourier transform spectrometer with the liquid-nitrogen cooled mercury cadmium telluride (MCT) receiver and a potassium-bromide broadband beam splitter in the 4.2–15.4 μm wavelength band. This receiver was corrected for nonlinearity [5]. The results of measurements for temperatures of 80 and 170°C are shown in the form of graphs of the radiation temperature of the absolutely black body against the wavelength (wave number), calculated using the inverse Planck function (Fig. 2). The gray area in the figure represents the range of total expanded uncertainty of the radiation temperature, equal to ±130 mK, calculated taking the uncertainty of the temperature sensors and the effective emissivities of both absolutely black bodies into account. The difference in the radiation temperatures of the standard LTBB and MTBB radiators lies in the limits of ±150 mK at 80°C and ±200 mK at 170°C. The difference of ±200 mK is explained by the additional uncertainty, introduced by the thermal drift of the Fourier spectrometer, which is ignored in the expanded overall uncertainty of the radiation temperature of both absolutely black bodies presented in the diagrams. Taking the value of 150 mK as the uncertainty of the measurement using the Fourier spectrometer, we can conclude that the LTBB and the MTBB are consistent with one another and can be used as standard radiation sources.

1166

Fig. 3. Energy radiance of the LTBB radiator, calculated from the results of measurements (the continuous curves), recorded by the MCT receiver, compared with the corresponding theoretical values (the points), calculated using Planck’s law (a), and the spectral distribution of the radiation temperature of the LTBB, calculated from the inverse Planck’s law (b).

Measurement of the Radiation Temperature of the LTBB in the –100 to +140°C Temperature Range and in the 4.5–20 µm Wavelength Band. The purpose of this experiment was to show that it is possible to make sequential measurements of low temperatures (down to –100°C) using a Fourier spectrometer in this wavelength band. We assumed that the Fourier spectrometer can be used as a stable instrument with constant spectral sensitivity over a period of five days. For this purpose, the spectrometer was calibrated at two temperatures of the LTBB: 0 and 170°C. The spectral sensitivity s and the modulator radiation Lm(Tm) were determined from (4) and (6), and then used to calculate the difference in the values of the radiance, using (3) for the LTBB: L LTBB (T ) = ( 0 LTBB (T ) − 0 ABB(LN

2)

(TABB(LN

2)

)) / s + L m (Tm ) + ρ m L ABB(LN

2)

(TABB(LN

2)

).

(7) 1167

Fig. 4. The energy radiance of the LTBB radiator, calculated from the results of measurements (the continuous curves), recorded with the DLaTGS receiver, compared with the corresponding theoretical values (the points), calculated using Planck’s law (a), and the spectral distribution of the radiation temperature of the radiator (b).

The measurements were made with different receivers. The MCT receiver. The first series of measurements were made using the vacuum Fourier transform spectrometer, the broadband KBr-beam splitter and the liquid-nitrogen cooled MCT receiver. The experimentally obtained spectral radiances of the LTBB and the theoretical values, calculated from Planck’s law at different temperatures from –100 to +80°C, are shown in Fig. 3a. We also show the noise level of these measurements, which was determined taking into account the root mean square deviation of a series of measurements at 0°C. To represent the results in the form of a graph of the radiation temperature against a wavelength, we used the inverse form of Planck’s law. The resultant spectral distribution of the radiation temperature for different values of the absolutely black body temperature is shown in Fig. 3b.

1168

Fig. 5. Directional spectral emissivity ε of the sample at a temperature of 120°C, observed at an angle of 5°: before liquid-nitrogen cooling (curve 1), after the first liquid-nitrogen cooling (curve 2) and the second liquid-nitrogen cooling (curve 3); the gray region shows the range of the standard uncertainty.

The deviation of the radiation temperature from its nominal value lies in the range ±0.5 K for measurements at –60, –80, and –100°C, and is less than this value for measurements at –20 and –40°C. The LTBB and MTBB radiators agreed within the limits ±0.2 K for a direct comparison (see Fig. 2), The large difference observed here is mainly due to the spectral sensitivity drift of the Fourier transform spectrometer over a period of several days. As mentioned above, the purpose of this series of measurements was to check the operation of the low-background calibration facility over wide temperature and spectral ranges, and not a high-precision comparison of ABB radiation with the theoretical radiation, described by Planck’s law. Hence, the radiation temperature down to –100°C and the emissivity can be measured stably with a vacuum Fourier transform spectrometer and a liquid-nitrogen cooled MCT receiver. The DLaTGS receiver. The second series of measurements was carried out using a vacuum Fourier transform spectrometer with a broadband KBr beam splitter and a DLaTGS pyroelectric receiver. The results of an experimental determination of the spectral radiance of an LTBB and the theoretical values, calculated from Planck’s law at temperatures of –100 to +140°C, are shown in Fig. 4a, as well as the calculated noise level. The spectral radiation temperature of the LTBB was also calculated in terms of the radiance using the inverse form of Planck’s law using (7) (Fig. 4b). In this case, the spectral radiance of a liquid-nitrogen cooled ABB can be neglected due to the very high level of the DLaTGS receiver noise. The deviation of the radiation temperature from Planck’s law, as also in the MCT receiver, lies in the range ±0.5 K, depending on the wavelength, and is due to drift of the spectral sensitivity of the Fourier transform spectrometer over a period of several days. Measurements using the DLaTGS receiver at lower temperatures exhibit much higher noise levels than when using the MCT receiver, but the first has a wider spectral range and also provides consistent and stable measurements of radiation temperatures. Hence, the radiation and emissivity in the –100 to +140°C temperature range can be stably measured using the vacuum Fourier transform spectrometer and the DLaTGS pyroelectric receiver. Characteristics and Results of a Test of the Coating of the Spherical Sample Enclosure. The directional spectral emissivity of the sample, investigated in the low-background calibration facility (LBCF), can be calculated from (1) if the values of the temperatures of both standard absolutely black bodies, the receiver, the sample surface and the radiation density incident on the sample from the hemispherical enclosure are known. This radiation density is determined by the temperature of the enclosure and its emissivity. Hence, the exact value of the directional spectral emissivity of the spherical enclosure is required to calculate the directional spectral emissivity of the sample. Moreover, this value must be the maximum possible, in order to reduce multiple reflections between the sample and the hemispherical enclosure for an exact calculation of the fraction of the radiation incident on the sample from the enclosure, and the possible temperature variation of the enclosure emissivity.

1169

Fig. 6. Directional (12°/12°) spectral reflection coefficient of the sample surface after cooling.

For this reason, the inner surface of the spherical enclosure is covered with circular grooves (60°) and, in addition, a black coating is deposited on it to provide an emissivity of greater than 0.98, for which Nextel Velvet Black 811-21 pigment was used. To check the characteristics of the coating, we constructed two similar samples with the same surface structure as the spherical enclosure, with grooves of identical geometry, also made from copper and chemically plated with nickel. The samples were then cleaned using a sand-blasting machine, after which they were coated with Nextel 811-21 pigment, one with a primer and the other without it. To test the coating, and its suitability for emissivity measurements in the middle and far infrared bands and the compatibility with the vacuum and low-temperature conditions, we made the following measurements. The directional spectral emissivity of the two samples were determined using the facility for measuring emissivity in air [6, 7] before and after a series of cooling cycles in liquid nitrogen. Moreover, the directional spectral emissivity in the far infrared band of these samples was also determined before and after a cooling test. The samples without the primer showed insufficient mechanical stability of the coating. Hence, only the results of a measurement on the sample coated with primer are presented below. The directional spectral emissivity ε of the spherical enclosure was determined at a temperature of 120°C. The resulting spectral emissivity of the sample and the range of standard uncertainty (P = 0.95), calculated as in [7], are shown in Fig. 5, from which it follows that ε decreases by approximately 0.01 after the first cooling, and then remains constant with respect to the standard uncertainty and at the desired level of 0.98. The directional (12°/12°) spectral reflection coefficient ρ of the spherical enclosure was measured using the vacuum Fourier transform spectrometer and the silicon-composite bolometer. The results after the liquid-nitrogen cooling tests for the 12.5–100 μm wavelength band are shown in Fig. 6, whence it follows that ρ did not change after cooling and remained lower than 0.1. Hence, the spectral reflection coefficient and the directional spectral emissivity of this coating satisfied the requirements as regards the inner surface of the spherical enclosure with respect to its suitability for emissivity measurements in the far infrared band and its compatibility with the vacuum and low-temperature conditions. Conclusion. The equipment for measuring emissivity under vacuum at the PTB as part of the low-background calibration facility [1] has completed its first tests. For measurements on the low-background calibration facility, as a rule, two standard radiation sources are employed at the present time – vacuum absolutely black bodies – and two measuring systems – for spectral radiance using a vacuum Fourier transform spectrometer in the 0.4–1400 μm wavelength band, and for radiation temperature using a vacuum standard infrared radiation pyrometer in the 8–14 μm spectral band. 1170

The vacuum Fourier transform spectrometer was used here in order to show that the two standard absolutely black bodies of variable temperature agreed within the limits of ±150 mK at 80°C and ±200 mK at 170°C. This investigation was carried out in a range which covered the temperature range of operation of two absolutely black bodies (80–170°C), which will serve as primary thermal radiation sources from –173 to +430°C. Moreover, it has been shown that the spectral radiances at low temperatures (down to –100°C) in the 4.5–20 μm wavelength band can be stably and consistently measured by this spectrometer. Measurements of the emissivity and reflection coefficient of samples, cleaned by sand blasting and coated with Nextel Velvet Black 811-21 pigment, showed that this coating is suitable for use as the inner surface of the spherical enclosure of the sample for emissivity measurements under low-temperature and vacuum conditions.

REFERENCES 1. 2. 3. 4. 5. 6. 7.

C. Monte et al., “Radiation thermometry and emissivity measurements under vacuum at the PTB,” Int. J. Thermophys., 30, 203–219 (2009). S. P. Morozova et al., “Vacuum variable temperature black body VLTBB100,” Int. J. Thermophys., 29, 341–351 (2008). S. P. Morozova et al., “Vacuum variable medium temperature black body VLTBB100,” Int. J. Thermophys., 31, 1809–1820 (2010). B. Gutschwager, R. Gärtner, and J. Hollandt, “An infrared precision radiation thermometer for the calibration of remote sensing instrumentations,” SPIE Vacuum Image and Signal Processing for Remote Sensing XV (2009), p. 7477. Pat. 4927269 1-10 USA (1990). C. Monte and J. Hollandt, “The measurement of directional spectral emissivity in the temperature range from 80°C to 400°C at the Physikalisch-Technische Bundesanstalt,” High Temp. – High Press., 39, 151–164 (2010). C. Monte and J. Hollandt, “The determination of the uncertainties of spectral emissivity measurements in air at the PTB,” Metrologia, 47, S172–S181 (2010).

1171