Reference cells and calibration

Aus PVMET
Zur Navigation springen Zur Suche springen

Ingo Kroeger

Physikalisch-Technische Bundesanstalt (PTB), Bundesallee 100, 38116 Braunschweig, Germany

Summary of reference cell properties, importance, different types, what is their purpose and how are they used.

Figure 1: Layout and circuit diagram of a WPVS reference solar cell [3]


For calibration measurements of solar cells or PV modules the irradiance level can be determined with a reference device. IEC standard 60904-2 [1] defines these as follows:

“Reference solar devices are specially calibrated devices which are used to measure natural or simulated irradiance or to set simulator irradiance levels for measuring the performance of other solar devices having similar spectral response, optical characteristics, dimensions and electrical circuitry.”

This definition follows the idea that measurement uncertainties can be kept as low as possible, if the reference device and the device under test are as similar as possible so that most systematic measurement uncertainties related to the device properties cancel out. This covers exemplary:

  • Figure 1: Layout and circuit diagram of a WPVS reference solar cell (2)
         If the spectral irradiance responsivity of the DUT and the reference are identical, the spectral mismatch correction factor according to IEC 60904-7 [2] becomes unity, even if the solar simulator spectrum is poorly matched to the AM1.5g spectrum
  •      If the size and the non-uniformity of the DUT and the reference are identical the non-uniformity correction of the solar simulator becomes negligible, even if the non-uniformity of the solar simulator is poor.
  •      If the optical characteristics of the DUT and the reference are identical, systematic measurement deviation arising from interreflections can be avoided.

This principle is applied for industry type “golden” reference solar cells and modules that are used in industry for production lines and that are calibrated by testing and calibration laboratories. Hence these “golden” reference solar cells and modules are well characterized, pre-aged and long term stable devices out of the manufacturing batch. The disadvantage of “golden” reference solar cells and modules is the relatively high measurement uncertainties that can be assigned to these devices that are limited by the state of the art calibration methods.

However, there is a second class of reference solar cells: the world photovoltaic scale (WPVS) reference solar cells. The purpose, its requirements and the design is described by Osterwald et al.  [3].

The WPVS reference solar cells are 2x2 cm² solar cells made from mono-crystalline float zone silicon encapsulated in silicone, epoxy or EVA with a durable glass window. The solar cell is electrically connected to female LEMO connectors that are fixed on the side of the package housing using 4 wire technique. On the side of the housing there must be printed or engraved a unique and permanent ID. Usually a Pt100 thermal resistor is attached to the backside of the solar cell allowing a direct measurement of the solar cell temperature. The solar cell is electrically isolated but in good thermal contact to the metal housing in order to enable effective cooling of the solar cell via the backside of the package upon irradiation. The advantage of these reference solar cells are their robustness in handling (due to the housing), their long term stability (due to the encapsulation), their reproducibility (due to the standardized connectors) and the possibility of calibration with very low measurement uncertainty using primary calibration methods (see Module 1 – Lesson 1 “Introduction to PV module power measurements”). The disadvantage of the WPVS reference devices is that they generally are not perfectly identical to the devices under test (industrial type solar cells or modules), so that the accuracy of the calibration methods and proper correction for measurement deviations becomes very crucial.


Additionally, there are options of optically filtered reference solar cells. In these cases, the glass window in front of the solar cell is replaced by a transmission filter in order to modify the spectral responsivity of the silicon device to better match the spectral responsivity of the DUT, e.g. for thin film technologies.

Classification, suppliers, certificates, Standards’ require.

The general requirements for a reference device according to IEC 60904-2 are stable photovoltaic characteristics and linear variation of the output signal with irradiance.


Suppliers of reference solar cells are:

  •     Fraunhofer ISE:

https://www.ise.fraunhofer.de/en/rd-infrastructure/accredited-labs/callab/wpvs-reference-cells.html

  •   PVMeasurements:

http://www.pvmeasurements.com/I-V-Health/calibrated-photovoltaic-reference-cells.html

  •      Konica Minolta:

https://www.konicaminolta.eu/en/measuring-instruments/products/photovoltaic/reference-pv-cells/introduction.html

  •      PRC Krochmann:

http://www.prc-krochmann.com/index.php/reference-solar-cells

  •      Solarmer:

http://solarmer.com/solar-reference-cell/

  •      Abet Technologies

http://abet-technologies.com/pv-iv-measurement-systems/reference-cells/

  •      ReRa Solutions

https://www.rerasolutions.com/product-category/solar-reference-cells/

  •      Photoemission

http://www.photoemission.com/referencecells.html


…and probably many more.

The requirements of the WPVS reference solar cell are as aforementioned described by Osterwald et al. [3]


Step by Step calibration procedure of a reference cell 

In principal there are 4 different procedures to calibrate reference solar cells:

a)      The direct sunlight method

b)     The global sunlight method

c)      The Differential spectral responsivity method

d)     The solar simulator method

All 4 methods can be conducted as primary or as secondary calibration methods, dependent on the reference applied. A description of these 4 methods can be found in IEC standard 60904-4 [4]. In the following section these methods are briefly summarized in a simplified version.

a) The direct sunlight method (DSM)

For the DSM method the direct component of the sunlight is taken as the light source. Therefore, a two axis sun-tracker is needed to properly point DUT and reference directly to the sun. Both devices must be temperature controlled and kept at 25°C. In order to block all diffuse light component, the DUT must be mounted in a collimator tube with a viewing angle identical to the. For secondary calibration, the reference would be another calibrated reference solar cell with an equivalent collimator tube or a pyrheliometer. For primary calibration the reference is a cavity radiometer traceable to the world radiometric reference (WRR). Additionally, a calibrated spectroradiometer with a collimator with identical viewing angle is recommended to measure the relative spectral irradiance distribution for spectral mismatch correction. The calibration procedure would be as follows:

  1.    Measurement should only be performed at clear sky conditions. Discard data that was taken when clouds or haze has been observed within the viewing angle of the instruments.
  2.    Point the DUT, reference and the spectroradiometer directly to the sun. Determine solar elevation and azimuth angle. Using this data and the geographical position to determine the airmass value (AMx).
  3.    Wait until reference and DUT show a stable temperature reading 25°C ± 2°C.
  4.    Measure (in an optimum case simultaneously) the irradiance via the reference, the short circuit current of the DUT, the spectral irradiance distribution and the temperature readings for DUT and reference.
  5.    Apply the spectral mismatch correction factor on the measured current of the DUT, if the spectral responsivity of DUT and reference differ significantly or if the spectral irradiance distribution differs significantly from the AM1.5g spectrum.
  6. Apply a temperature correction on the measured current if the DUT, if the temperature coefficient of the DUT is known.
  7.   Scale the measured current at given irradiance to 1000W/m² and correct for non-linearity of DUT and reference if non-linearity is known.


Repeat this procedure several times across the day and repeat these measurements for at least 2 more days. Average calibration values and identify and remove outliers. More detailed information on this method can be found in Ref. [5-9].

b) The global sunlight method (GSM)

For the GSM method the global component of the sunlight is taken as the light source. Also for this method, a two axis sun-tracker is needed to properly point DUT and reference directly to the sun. Both devices must be temperature controlled and kept at 25°C. For secondary calibration, the reference would be another calibrated reference solar cell or a pyranometer. Additionally, a calibrated spectroradiometer with an entrance optics for global spectral irradiance is recommended to measure the relative spectral irradiance distribution for spectral mismatch correction. The calibration procedure would be as follows:

  1.   Measurement should only be performed at clear sky conditions. Discard data that was taken when clouds or haze has been observed. Care should be taken to avoid strong reflecting objects in the vicinity of the measurements.
  2.   Point the DUT, reference and the spectroradiometer directly to the sun. Determine solar elevation, azimuth angle and atmospheric pressure. Using this data and the geographical position to determine the airmass value (AMx).
  3.   Wait until reference and DUT show a stable temperature reading 25°C ± 2°C.
  4.   Measure (in an optimum case simultaneously) the irradiance level via the reference, the short circuit current of the DUT, the spectral irradiance distribution and the temperature readings for DUT and reference.
  5.   Apply a temperature correction on the measured current if the DUT, if the temperature coefficient of the DUT is known.
  6.   Scale the measured current at given irradiance to 1000W/m² and correct for non-linearity of DUT and reference if non-linearity is known.
  7.   Apply the spectral mismatch correction factor on the measured current of the DUT, if the spectral responsivity of DUT and reference differ significantly or if the spectral irradiance distribution differs significantly from the AM1.5g spectrum. Alternatively, you may perform a linear regression of the measured short circuit current over the air mass value AMx and determine the short circuit current value for AM1.5.


Repeat this procedure several times across the day and repeat these measurements for at least 2 more days. Average calibration values and identify and remove outliers. More detailed information on this method can be found in Ref. [10-20].

c) The differential spectral responsivity method (DSR)

The DSR-method is a spectral method using quasi monochromatic light as a light source. Hence, for this method a reference is needed with calibrated differential spectral irradiance responsivity. This could be a reference solar cell or a photodiode.  The optical setup of a DSR-facility is typically built up the following way: A white light source such as a Xe arc lamp or a halogen lamp is coupled into a monochromator system that selects the desired wavelength. Following the monochromator there is an optical chopper that modulated the monochromatic light with a given frequency f. Subsequently the monochromatic light passes an imaging optics that generates a uniform irradiance distribution in the measurement plane. Within this optics there is a beam splitter that couples a fraction of the monochromatic light onto a monitor photodiode. Finally, there are white light bias lamps that can generate a bias irradiance between 0 – 1100 W/m² in the measurement plane additionally to the monochromatic irradiance. The DUT and the reference must be temperature controlled and kept at 25°C. Both reference device and DUT shall be connected to a transimpedance amplifier, that coverts the currents into voltages and keeps the solar cell in short circuit state. The AC modulated voltage generated by the monochromatic spectral irradiance shall be then measured by a Lock-In amplifier, the DC voltage generated by the bias irradiance shall be measured by a calibrated multimeter. Also the Monitor current should be converted into a voltage using a shunt or a transimpedance amplifier and measured by a Lock-In-amplifier. The calibration procedure would be as follows:

  1. Place the reference device in the centre (i.e. location of the best uniformity) of the monochromatic light field, wait until the temperature reading is stable within 25°C ± 1°C and measure the monitor corrected voltage of the reference for all wavelengths of interest (i.e. from 280 nm – 1200 nm if the DUT is made from c-Si): [math]\displaystyle{ \frac{I_{Ref}\left(\lambda\right)}{I_{MD,Ref}\left(\lambda\right)} }[/math]
  2. Place the DUT at the identical position as i), set the bias irradiance to a fixed level, wait until the temperature reading is stable within 25°C ± 1°C and measure the monitor corrected voltage of the DUT for all wavelengths of interest. Repeat this measurement for at least 7 different bias irradiance levels between 0 and 1100 W/m² and measure the bias current [math]\displaystyle{ I_{Bias} }[/math] of the DUT: [math]\displaystyle{ \frac{I_{SZ}\left(\lambda,I_{Bias}\right)}{I_{MD,SZ}\left(\lambda\right)} }[/math]
  3. Calculate the absolute differential spectral irradiance responsivity [math]\displaystyle{ {\widetilde{s}}_{SZ}\left(\lambda,I_{Bias}\right) }[/math] using the calibration values [math]\displaystyle{ {\widetilde{s}}_{Ref}\left(\lambda\right) }[/math] of the reference: [math]\displaystyle{ {\widetilde{s}}_{SZ}\left(\lambda,I_{Bias}\right)=\frac{\frac{I_{SZ}\left(\lambda,I_{Bias}\right)}{I_{MD,SZ}\left(\lambda\right)}}{\frac{I_{Ref}\left(\lambda\right)}{I_{MD,Ref}\left(\lambda\right)}}\bullet{\widetilde{s}}_{Ref}\left(\lambda\right) }[/math]
  4. Calculate the AM1.5g weighted absolute differential spectral irradiance responsivity  using the tabulated AM1.5g spectral irradiance data from IEC 60904-3: [math]\displaystyle{ {\widetilde{s}}_{AM1.5g}\left(I_{SC}\left(E_b\right)\right)=\frac{\int_{0}^{\infty}{\widetilde{s}\left(\left.\ \lambda,\ I_{SC}\ (E_B\right)\right)\bullet E_{\lambda,\ AM1.5g}\left(\lambda\right)d\lambda}}{\int_{0}^{\infty}{E_{\lambda,\ AM1.5g}\left(\lambda\right)d\lambda}} }[/math]
  5. Calculate the short circuit current under standard test conditions by approximating the upper integration limit, such that the following equation is fulfilled. [math]\displaystyle{ 1000=\int_{0}^{I_{STC}}\frac{1}{{\widetilde{s}}_{AM1.5g}\left(I_{SC}\right)}dI_{SC} }[/math]
  6. Additionally, the absolute spectral irradiance responsivity of the DUT can be calculated by firstly calculating the irradiance level of each measurement [math]\displaystyle{ E_{b,\ AM1.5g}=\int_{0}^{I_{SC}\left(E_b\right)}\frac{1}{{\widetilde{s}}_{AMx}\left(I_{SC}\right)}dI_{SC} }[/math] and secondly solving the following equation: [math]\displaystyle{ s_{STC,\ AM1.5g}\left(\lambda\right)=\frac{1}{E_{STC}}\int_{0}^{E_{STC}}{\widetilde{s}\left(\lambda,E_{b,\ AM1.5}\right)\partial E_{b,\ AM1.5}} }[/math]

In addition to this calibration procedure the following sources of uncertainty should carefully be evaluated and eventually corrected/accounted for: non-uniformity of the spectral irradiance, temperature deviation from 25°C, bandwidth and wavelengths of the monochromatic spectral irradiance, non-linearity of the measurement electronics, positioning of DUT and reference in the same position and measurement plane and reproducibility. More detailed information on this method can be found in Ref. [21-23].

d) The solar simulator method (SolSim)

The solar simulator method uses artificial light sources to emulate the AM1.5g spectral irradiance. The light engine is often made from Xenon lamps (or Xe- and halogen multi lamp systems) combined with optical filters and homogenizing optics to generate a uniform irradiance in the measurement plane. For this method a reference is needed with calibrated short circuit current under STC and spectral irradiance responsivity under STC. differential spectral irradiance responsivity. In the measurement plane there should be a monitor solar cell that can be measured simultaneously with the DUT or the reference. The DUT and the reference must be temperature controlled and kept at 25°C. The calibration procedure would be as follows:

  1. Place the reference device in the centre (i.e. location of the best uniformity) of the light field, wait until the temperature reading is stable within 25°C ± 1°C and adjust the solar simulator lamp current or the distance between light source and measurement plane until the calibrated short circuit current at STC of the reference is measured. Also measure the current of the monitor.
  2. Place the DUT in the exact same position as the reference, wait until the temperature reading is stable within 25°C ± 1°C and measure its short circuit current. Also measure the current of the monitor.
  3. If there is a difference of the two monitor readings, the irradiance level has drifted. Correct the measured short circuit current of the DUT by the correction factor given by the ratio of [math]\displaystyle{ f=\frac{I_{Mon,\ DUT}}{I_{Mon,\ Ref}} }[/math]
  4. Apply the spectral mismatch correction factor on the measured current of the DUT, if the spectral responsivity of DUT and reference differ significantly. For this correction the relative spectral irradiance distribution must be measured in the exact position of the DUT with the identical settings of the solar simulator of the calibration using a calibrated spectroradiometer. Additionally, the relative spectral irradiance responsivity of the DUT must be known. This could be measured using a simplified DSR-method as described above.
  5. If the DUT and the reference are of different size, a correction accounting for the non-uniformity of the irradiance in the measurement plane has to be applied

More detailed information on this method can be found in Ref. [24-28].

Case study: example reference cell, characteristics, calibration (possibly a video)

Figure 1: Picture of a typical WPVS reference solar cell.

A good example for a reference solar cell is shown in Figure 1. In the centre there is a 2x2 cm² solar cell made from mono-crystalline float zone silicon encapsulated with a durable glass window. The solar cell is electrically connected to female LEMO connected to the bottom. On the left side of the housing there is the printed ID as well as the name of the manufacturer and are marking of the physical position of the cell within the housing.

Figure 1: Picture of a typical WPVS reference solar cell.

The coloured graphs in Figure 2 show the absolute differential spectral responsivity (DSR) of this reference solar cell. These curves have to be taken, if the reference device is used as a reference for a DSR facility measuring the DSR of a DUT.

The spectral irradiance responsivity (SR) is shown as dashed line. The SR drops to zero at 280 nm and at 1200 nm. The short circuit current of this cell is 122,14 mA. This SR curve has to be taken for the spectral mismatch correction if this reference device is used as a reference for a solar simulator measuring the short circuit current of a DUT. The irradiance level of the solar simulator has to be adjusted to the short circuit current of this reference. The non-linearity of the device is better than 0.1 % as shown in Figure 3

Figure 3: This graph shows the AM1.5g weighted DSR dependent on the bias current i.e. the bias irradiance level. This graph indicates that the non-linearity of the reference solar cell is less than 0.1%.

Figure 4 shows the spectral temperature coefficient of this silicon reference solar cell. This graph shows the percentage change of the spectral responsivity per Kelvin temperature difference from 25°C. This correction has to be applied to the spectral responsivity, if the reference cell is operated at different temperatures. A weighting of this function with the SR of the device and the AM1.5g spectrum leads to a temperature coefficient of the short circuit current of 680 ppm/K. This correction has to be applied to the short circuit current, if the reference cell is operated at different temperatures.

Figure 2: Differential spectral responsivity of the reference solar cell (coloured) as well as its spectral irradiance responsivity at STC (dashed).
Figure 4: Spectral temperature coefficient of the reference solar cell.


References

[1]        IEC 60904-2:2015, Photovoltaic devices - Part 2: Requirements for photovoltaic reference devices

[2]        IEC 60904-7:2008, Photovoltaic devices - Part 7: Computation of the spectral mismatch correction for measurements of photovoltaic devices

[3]        Osterwald, C. R., Anevsky, S. , Bücher, K. , Barua, A. K., Chaudhuri, P. , Dubard, J. , Emery, K. , Hansen, B. , King, D. , Metzdorf, J. , Nagamine, F. , Shimokawa, R. , Wang, Y. X., Wittchen, T. , Zaaiman, W. , Zastrow, A. and Zhang, J. (1999), The world photovoltaic scale: an international reference cell calibration program. Prog. Photovolt: Res. Appl., 7: 287-297.

[4]        IEC 60904-4:2009, Photovoltaic devices - Part 4: Reference solar devices - Procedures for establishing calibration traceability

[5]        C.R. Osterwald, K.A. Emery, D.R. Myers, R.E. Hart “Primary reference cell calibrations at SERI: History and methods” Proc. 21st IEEE PVSC Orlando, FL, May 21-25 1990, 1062-1067.

[6]        K.A. Emery, C.R. Osterwald, L.L. Kazmerski, R.E. .Hart “Calibration of primary terrestrial reference cells when compared with primary AM0 reference cells" Proc. 8th European PVSEC, Florence, Italy, May 9-12 1988 p. 64-68.

[7]        C. Osterwald, K. Emery "Spectroradiometric Sun Photometry" Journal of Atmospheric and Oceanic Technology, 17 (200) 1171-1188.

[8]        ASTM E 1125 “Standard test method for calibration of primary non-concentrator terrestrial photovoltaic reference cells using a tabular spectrum”.

[9]        A Fehlmann, G Kopp, W Schmutz, R Winkler, W Finsterle, N Fox, metrologia 49 (2012) S34

[10]      K.A. Emery, C.R. Osterwald, L.L. Kazmerski, and R.E. Hart, (1988c), Calibration of Primary Terrestrial Reference Cells When Compared With Primary AM0 Reference Cells, Proceedings of the 8th PV Solar Energy Conference, Florence, pp. 64-68.

[11]      K. A. Emery, C.R. Osterwald, S. Rummel, D.R. Myers, T.L. Stoffel, and D. Waddington, “A Comparison of Photovoltaic Calibration Methods,” Proc. 9th European Photovoltaic Solar Energy Conf., Freiburg, W. Germany, September 25-29, 1989, pp. 648-651.

[12]      K.A. Emery, D. Waddington, S. Rummel, D.R. Myers, T.L. Stoffel, and C.R. Osterwald, “SERI Results from the PEP 1987 Summit Round Robin and a Comparison of Photovoltaic Calibration Methods,” SERI tech. rep. TR-213-3472, March 1989.

[13]      Gomez, T, Garcia L, Martinez G, "Ground level sunlight calibration of space solar cells. Tenerife 99 campaign," Proc. 28th IEEE PVSC, 1332-1335, (2000).

[14]      J. Metzdorf, T. Wittchen, K. Heidler, K. Dehne, R. Shimokawa, F. Nagamine, H. Ossenbrink, L. Fornarini, C. Goodbody, M. Davies, K. Emery, and R. Deblasio, “The Results of the PEP '87 Round-Robin Calibration of Reference Cells and Modules,- Final Report” PTB technical report PTB-Opt-31, Braunschweig, Germany, November 1990, ISBN 3-89429-067-6.

[15]      H. Müllejans, A. Ioannides, R. Kenny, W. Zaaiman, H. A. Ossenbrink, E. D. Dunlop “Spectral mismatch in calibration of photovoltaic reference devices by global sunlight method” Measurement Science and Technology 16 (2005) 1250-1254.

[16]      H. Müllejans, W. Zaaiman, E. D. Dunlop, H. A. Ossenbrink “Calibration of photovoltaic reference cells by global sunlight method”, Metrologia 42 (2005) 360-367.

[17]      H. Müllejans, W. Zaaiman, F. Merli, E. D. Dunlop, H. A. Ossenbrink “Comparison of traceable calibration methods for primary photovoltaic reference cells” Progress in Photovoltaics 13 (2005) 661-671.

[18]      F.C. Treble and K.H. Krebs, “Comparison of European Reference Solar Cell Calibrations”, Proc. 15th IEEE PV Spec. Conf., 1981, pp. 205-210.

[19]      R. Whitaker, G. Zerlaut, and A. Purnell, “Experimental demonstration of the efficacy of global versus direct beam use in photovoltaic performance prediction of flat plate photovoltaic modules”, Proc 16th IEEE PVSC, pp. 469-474, 1982.

[20]      A Fehlmann, G Kopp, W Schmutz, R Winkler, W Finsterle, N Fox, metrologia 49 (2012) S34

[21]      J. Metzdorf “Calibration of solar cells. 1: The differential spectral responsivity method”, Appl. Optics 26 (9) (1987) 1701-1708.

[22]      J. Metzdorf, S. Winter, T. Wittchen “Radiometry in photovoltaics: calibration of reference solar cells and evaluation of reference values” metrologia 37 (2000) 573-578.

[23]      S. Winter, T. Wittchen, J. Metzdorf “Primary Reference Cell Calibration at the PTB Based on an Improved DSR Facility” in “Proc. 16th European Photovoltaic Solar Energy Conf.”, ed. by H. Scherr, B. Mc/Velis, E. Palz, H. A. Ossenbrink, E. Dunlop, P. Helm (Glasgow 2000) James & James (Science Publ., London), ISBN 1 902916 19 0.

[24]      R. Shimokawa, F. Nagamine, Y. Miyake, K. Fujisawa, Y. Hamakawa “Japanese indoor calibration method for the reference solar cell and comparison with outdoor calibration” Japanese J. Appl. Phys. 26(1) (1987) 86-91.

[25]      R. Shimokawa, H. Ikeda, Y. Miyake, S. Igari "Development of wide field-of-view cavity radiometer for solar simulator use and intercomparison between irradiance measurements based on the world radiometer reference and electrotechnical laboratory scales" Japanese J. Appl. Phys. 41 (2002) 5088-5093.

[26]      H. Müllejans, W. Zaaiman, F. Merli, E. D. Dunlop, H. A. Ossenbrink “Comparison of traceable calibration methods for primary photovoltaic reference cells” Progress in Photovoltaics 13 (2005) 661-671.

[27]      CIE 53-1982 “Methods of Characterizing the Performance of radiometers and Photometers”, ISBN 92 9034 053 3.

[28]      CIE 63-1984 “The Spectroradiometric Measurement of Light Sources”.




Abgerufen von „https://twiki.hypradata.de/index.php?title=Reference_cells_and_calibration&oldid=135“