New traceability chain for spectral irradiance measurement at LNE-Cnam

At LNE-Cnam, the source radiometry team is in charge of the realization of national reference scales for spectral radiance and irradiance. Its objective is to set and maintain them at the highest level of accuracy. The team participates to international comparisons to ensure the consistency of primary achievements with other countries. It carries out calibrations for French laboratories and National Metrological Institutes (NMIs) in the framework of international cooperation.

Previously, French primary spectral irradiance and radiance scales were traceable to the International Temperature Scale, ITS-90 (figure 1). Measurement range was from 300 nm to 2500 nm [1]. It was done as follows:

Zoom In Zoom Out Reset image size

LNE-Cnam Temperature department maintains the fixed point of copper of the ITS-90 [2], that is at 1357 K. This fixed point was used to calibrate a pyrometric lamp with a radiance temperature of 1917 K at 650 nm. The pyrometric lamp was then used to calibrate a small blackbody at different temperature ranging from 1900 K to 2000 K. This blackbody was used as a spectral radiance standard. Radiance of transfer sources to be calibrated, i.e. tungsten ribbon lamps, was compared to the radiance of the blackbody using a spectroradiometer on the spectral range 300–2500 nm. The spectral irradiance transfer source was calibrated in two steps [1]: first in relative spectral value using a tungsten ribbon lamp on a specific bench, second in absolute value (at 492 nm and 575 nm) with a filter radiometer on another bench. This filter radiometer had been previously calibrated using different references: the cryogenic radiometer for spectral responsivity, the spectrophotometry bench for the filter and the diaphragm surface measurement bench.

In order to extend our measurement capacities in the UV range up to 200 nm and to reduce our uncertainties over the full range, a new spectral irradiance measurement facility has been developed and the traceability chain has been simplified (figure 2). It is now based on the use of a high temperature blackbody (HTBB), working at a temperature that is determined using the radiometric traceability chain [3]. HTBB are commonly used as reference by NMIs, as they are known to be stable, homogeneous and able to reach high temperatures. However, the overall method described in this article is a rather complex combination of procedures i.e. original to our traceability chain. In particular, using a filter radiometer rather than a radiation thermometer allows us to link our traceability chain to the cryogenic radiometer [4] and to reduce our uncertainties.

Zoom In Zoom Out Reset image size

With this new installation, which does not rely on the ITS-90 temperature reference anymore, we have eliminated the pyrometric lamp measurement step and removed the radiance transfer source measurement pass. We are not only extending our measuring range, but we are also simplifying the process by reducing the number of benches (three in one) and the number of operations. This allows us to reduce the measurement uncertainties by at least a factor of 2 over almost the entire spectral range.

The new approach for spectral irradiance measurement consists of directly comparing the flux from a standard lamp to be calibrated with the flux of a HTBB with a temperature of approximately 3000 K. The reference spectral irradiance is determined by measuring the temperature of the HTBB with a filter radiometer. The absolute responsivity of the filter radiometer is accessed using our traceability chain to the cryogenic radiometer, our primary spectrophotometer and the dimensional reference.

2.1. Filter radiometer

The filter radiometer consists of the following three components (figure 3):

• a 8 mm diameter diaphragm,
• a SFK9 bandpass filter (Schott) centered on 492 nm,
• a three silicon photodiodes trap detector.

Zoom In Zoom Out Reset image size

Diaphragm and filter are placed in a housing regulated at a constant temperature of 23 °C by water cooling.

Its absolute spectral responsivity is obtained in three steps:

• calibration of spectral responsivity of the trap detector at six wavelengths from 457 nm to 633 nm by comparison with our cryogenic radiometer;
• calibration of the spectral regular transmittance of its bandpass filter with our primary spectrophotometer from 400 nm to 1100 nm [5];
• calibration of surface of the diaphragm with our devoted setup [6, 7].

The reflectivity of the trap detector is assumed to be null. Therefore, the absolute responsivity of the filter radiometer is expressed in A W⁻¹m² given by the integral over the full spectrum of the product of the filter spectral transmittance by the detector spectral responsivity and the surface of the diaphragm.

The uncertainty on the filter radiometer responsivity comes from uncertainties associated to each filter radiometer component: filter spectral transmittance (0.29%, k = 2), trap detector spectral responsivity (0.08%, k = 2), and surface of the diaphragm (0.16%, k = 2). They are detailed in table 1.

Table 1. Type B of the filter radiometer calibration uncertainties (k = 2).

Parameters	Origin	Relative expanded uncertainty (k = 2)
Diaphragm surface	Aperture IV	0.16%
Calculation of the sensitivity integral	Filter, trap detector	0.30%
Filter heating	Filter	0.012%
Thermal correction	Filter	0.022%
Filter radiometer		0.34%

2.2. Irradiance reference

HTBB's temperature is determined by first measuring its irradiance with the calibrated filter radiometer (figure 4). By measuring the geometrical extend of the system, the spectral radiance of the HTBB at 492 nm is deduced. Using Planck's law, the temperature of the blackbody is calculated.

Zoom In Zoom Out Reset image size

The temperature of the blackbody is obtained thanks to the following model:

$$\begin{equation*}{T_{{\lambda _0}}} = \frac{{{C_2}}}{{n{\,}{\times}{\,}{\lambda _0}{\,}{\times}{\,}\ln \left( {\frac{{{C_1}}}{{\pi{\,}{\times}{n^2}{\,}{\times}{\,}{\lambda _0}^5{\,}{\times}{\,}\varepsilon{\,}{\times}{\,}{L_{{\lambda _0}}}}} + 1} \right)}}{\text{ }}\end{equation*}$$

$$\begin{equation}{\text{ With}}\,{L_{{\lambda _0}}} = \frac{{{\Phi }}}{G} = \frac{{{d^2}}}{{{S_e}}}{\,}{\times}{\,}{E_{{\lambda _0}}}\,{\text{and}}\,{\text{ }}{E_{{\lambda _0}}} = \frac{{{\Phi }}}{{{S_r}}};{\text{ }}\end{equation} \tag{ 1 }$$

where ${\text{ }}{T_{{\lambda _0}}}$: HTBB temperature measured at λ₀, C₁ and C₂: Planck constants, n: air refractive index at room temperature, normal pressure and at 492 nm, according to the Ciddor model [9] (n = 1.00027113 with a relative uncertainty of 2.4 × 10⁻⁸), λ₀: wavelength in vacuum used with filter radiometer (λ₀ = 492 nm), : emissivity of the HTBB ( = 0.9976 ± 0.0010), ${L_{{\lambda _0}}}$: HTBB radiance measured at λ₀ with filter radiometer, S_e: diaphragm area of the HTBB (S_e = 50.250 mm² ± 0.049 mm²), d: distance from the HTBB diaphragm to the filter radiometer diaphragm (d = 758.00 mm ± 0.46 mm), ${E_{{\lambda _0}}}$: irradiance of the HTBB measured at λ₀ with filter radiometer.

Once the HTBB's temperature is known, we can calculate the spectral radiance and irradiance of HTBB at all calibration wavelengths. The spectral irradiance emitted by the HTBB is given by:

$$\begin{equation}{E_{{\text{HTBB}}}}\left( {\lambda ,T} \right) = \frac{{{S_{\text{e}}}}}{{{d^{\,2}}}} \times \varepsilon \times \frac{{{C_1}}}{{\pi \times {n^2} \times {\lambda ^5}\left( {\exp \left( {\frac{{{C_2}}}{{n \times \lambda \times T}}} \right) - 1} \right)}}.\end{equation} \tag{ 2 }$$

2.3. Spectral irradiance of a lamp

The spectral irradiance at a wavelength λ of a lamp is calculated as follows:

$$\begin{equation}{E_{lamp}}\left( \lambda \right) = \frac{{Signal{\text{ }}lamp\left( \lambda \right)}}{{Signal{\text{ }}HTBB\left( \lambda \right)}}{\,}{\times}{\,}{E_{HTBB}}\left( {\lambda ,T} \right).\end{equation} \tag{ 3 }$$

Our new facility has three main parts (figures 5 and 6).

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The 1st part is the HTBB equipped with a 8 mm diameter diaphragm, heated up to 3000 K. It provides radiations that can be used in the 250 nm to 2500 nm range.

The 2nd part is a spectroradiometric system that compares the spectral irradiance of the HTBB and transfer lamps (lamps 1 and 2). It comprises:

• an integrating sphere equipped with a circular entrance port diaphragm to receive the flux from the sources and with a rectangular slit shape of 2 mm × 15 mm output port located at 90°;
• two Ø150 mm spherical mirrors that focus the image of the sphere exit slit on the entrance slit of a monochromator;
• a Jobin Yvon HRD monochromator (640 mm focal length) on which the manual wavelength selection mechanism has been replaced by a high precision motorized translation stage and equipped with three gratings to cover the spectral range. The monochromator can be configured in single path or double path depending on the wavelength range (table 2);
• a set of different detectors to be used depending on the wavelength range.

The gratings and detectors characteristics are described in table 2. In UV range, we use the full path of our double monochromator (double) combined with a photomultiplier detector. For VIS-IR range, we use only the first path of our double monochromator (single) combined with three different detectors chosen according to the wavelength.

The 3rd part is a calibrated filter radiometer with a responsivity centered at 492 nm.

The filter radiometer and the integrating sphere are placed on the same translation stage to perform in one measurement cycle the HTBB temperature measurement and the signals ratio of the HTBB and the lamps. The integrating sphere rotates by 180° around an axis exactly centered on its exit slit so that the entrance port faces one source (HTBB) or another (transfer lamp) to collect flux, but the exit slit remains steady, its low and high part being inverted. To our knowledge, using a rotating integrating sphere rather than a translation stage is original to our method. The proposed set-up yields a more compact measurement setup, and allows heavy elements (HTBB) and fragile elements (monochromator) to remain stationary. This significantly improves the stability of the measurements, particularly in the case of the monochromator, a delicate mechanical instrument, as it allows its settings and wavelength calibration to remain very stable.

The entire spectroradiometric system is placed in a box with two compartments, one for the sources light output collection and another one for the monochromator, in order to reduce stray light.

The bench is fully automated and controlled by a custom LabVIEW software. This facility allows two lamps to be measured at the same time.

The calibration of a lamp consists in comparing its spectral irradiance with that of the HTBB wavelength by wavelength. The diagram in figure 7 describes the different steps of the measurement sequence for the calibration of two lamps. The blue boxes represent measurements made at the output of the monochromator.

Zoom In Zoom Out Reset image size

The measurement model for the HTBB temperature is shown by equation (1). The uncertainties on the temperature measurement are detailed in table 3.

The uncertainty on HTBB spectral irradiance reference is depicted in table 4.

Finally, the uncertainty on the spectral irradiance of a lamp under calibration is given in table 5.

Our set of four reference transfer lamps have been measured and compared with their historical values on the range 300–2500 nm (table 6). There is at least a factor two between historical and new measurement uncertainties (k = 2) over the full range. At the range extremities, i.e. 250 nm and 2500 nm, the large uncertainties are due to the lack of flux.

In table 7, most of the mean deviations 2019/2016 observed are smaller than the expanded uncertainty of the comparison. At 300 nm, the significant deviation observed seems to correct the deviation that we had during the last CCPR K1-a comparison in 2005 [10]. It could be explained by the lack of flux and stray light of the former single monochromator.

From now on, LNE-Cnam spectral irradiance reference is based on the radiometric reference scale. Thanks to the new setup, it is possible to measure the spectral irradiance of transfer lamps with a direct traceability to our cryogenic radiometer. Our installation is simplified into a single measuring bench, which is also relatively compact and stable thanks to using a rotating integrating sphere. Our measurement capacity is extended to cover the spectral range from 250 nm to 2500 nm. Our measurement uncertainties are reduced by at least a factor two over almost the entire spectral range.

In the future, this measurement setup will be improved by using white or gold plane diffuser instead of integrating sphere to extend our measurement capabilities down to 200 nm and up to 3000 nm.

The authors would like to thank Dr Jeanne-Marie Coutin for the calibration of the trap detector and Mr Arnaud Richard for the calibration of the filter. These elements are part of the filter radiometer.

We would also like to thank Dr Lou Gevaux and Dr Jimmy Dubard for their valuable advice and proofreading of this article.

The data that support the findings of this study are available upon reasonable request from the authors.