The PMOD/WRC Precision Spectroradiometer

PSR

Julian Gröbner, Natalia Kouremeti, Ricco Soder, Diego Wasser, Manfred Gyo, Fabian Dührig,

Etienne De Coulon

Physikalisch-Meteorologisches Observatorium Davos, World Radiation Center, Davos

Switzerland

Some of the work was performed within the EMRP project ENV03.

The European Metrology Research Programme (EMRP) is jointly funded by

the EMRP participating countries within EURAMET and the European Union.

Thanks to Saulius Nevas and Peter Sperfeld, PTB

for characterising the PSR stray light and linearity

Potential Applications

• Spectral aerosol optical depth for the eventual replacement of filter sunphotometers

• Retrieval of atmospheric parameters (ozone, water vapour, aerosols)

• absolute spectral solar irradiance measurements for photovoltaic applications

• Climatological measurements of solar irradiance (direct & global)

Some history...

August 2008 Project submitted to the regional Stiftung for Innovation, Entwicklung and

Forschung GR

September 2009 Design Study

October 2010 Benchtop model

Jan 2011-March 2012 Prototype Development

March 2012 PSR V0 First light

Mar 2014 PSR V2 First light

Mar-June PSR_003 in Izaña

June 2014 First 2 PSRs delivered to

customer

July 2014 AOD Campaign Finokalia

PSR Specifications

• Holographic flat-field grating 200 lines/mm from Zeiss

• Hamamatsu NMOS 1024 linear image sensor

• 18 bit ADC ( 262 kcounts)

• 10 ms to 40 sec integration time

• Temperature stabilised sensor ±0.1K (invar)

• Temperature insensitive optical bench (carbon)

• Wavelength range 300 – 1020 nm

• Resolution 1.5 nm to 6 nm

• Wavelength step 0.7 nm

• 2 entrance optics , direct irradiance and generic SMA

• stray-light optimised optical design with zero-order light-trap

PSR Characterisation

• Slit function and bandpass

• Linearity

• Temperature coefficients

• Stray-Light

• Telescope Flat-field

• Direct irradiance calibration

Linearity of PSR_003

Integration time variation from 10 ms to 8000ms Irradiance flux variation

PSR_003 is linear to better than ±1% over the full range:

• IT from 10ms to 8000ms

• Intensity variation between 1000 and 245000 counts

courtesy P. Sperfeld, PTB 0 0.5 1 1.5 2 2.5

x 105

-3

-2

-1

0

1

2

3

Dark Corrected Signal

Non-L

inearity

(%

)100*[

1-S

cp

ms(I

T)/

Scp

ms(I

Tre

f)]

Linearity - PSR_006

300 400 500 600 700 800 900 1000 1100-0.016

-0.014

-0.012

-0.01

-0.008 PSR_003 Temperature Depedency of Wavelength 19-Mar-2014

Wavele

ng

th S

hif

t (n

m/o

C)

wavelength (nm)

300 350 400 450 500 550 600 650-0.01

-0.005

0

0.005

0.01

F

WH

M (

nm

/oC

)

wavelength (nm)

Temperature behavior of PSR

Wavelength shift versus temperature

Optical Stability is obtained by using a carbon bench with vanishing temperature expansion coefficient.

1) from solar measurements and matSHIC

Wavelength /nm

Temp. coeff nm/K

350 0.01 ±0.004

500 0.02 ±0.01

605 0.00 ±0.03

795 -0.005 ±0.02

Using data from 60 days

dWL/T~0.01 nm K-1

sensor holder

2) Laboratory measurements

dFWHM K-1 < 0.005 nm K-1

0.01 nm = 1/70 pixel

Temperature behavior of PSR The temperature dependence of the sensitivity was determined for 4 PSR in the laboratory.

PSR Temp. coeff

%/K

003 -0.08 %

004 < 0.04 %

005 -0.06 ± 0.02 %

006 < 0.05 %

Example for PSR 006

0 50 100 150 200 250 3000

5

10

15

Erlapsed time (min)

Ben

ch

Tem

pera

ture

(oC

)

0 50 100 150 200 250 300-0.2

0

0.2

0.4

0.6

0.8

Erlapsed time (min)

Tem

pera

ture

Gra

die

nt

over

2

.7 m

in (

oC

/min

)100 200 300 400 500 600 700 800 900 1000 1100

-0.1

-0.05

0

0.05

0.1

pixel

Irra

dia

nce T

em

pera

ture

Dep

en

den

ce F

acto

r (%

)

±0.05%

The change in sensitivity is less than 0.1% / K for all PSR

Line Spread functions of various PSRs using lasers and spectral emission lamps

By default, a single gaussian fit is used

to determine FWHM and wavelength

dispersion

The stray-light rejection is > 10-5 The LSF seems very similar between

different PSRs. Here, one can clearly

see the second order line emerging from

the background

For some lines, an improved fit to the

experimental data is obtained using a

double gaussian fitting

Bandpass using spectral emission lines

FWHM varies between 1.5 nm and 6 nm

The location of the minimum FWHM can be selected

during the alignment process

The smallest FWHM of 1.5 nm is where the image of

the entrance slit is optimally focussed on the detector

300 400 500 600 700 800 900 10001

2

3

4

5

6

wavelength (nm)

FW

HM

of

Slit

Fu

nc

tio

n (

nm

) (m

ult

i-g

au

ss

)

PSR_006

Alignment of PSR on: 09-May-2014

Ar

CdHg

Kr

Xe

Stray-light The slit functions were determined at the tuneable laser setup PLACOS at PTB in the range 300 to 1000 nm.

YIB=C Ymeas

The stray-light corrected signal YIB can

be retrieved from the measurement by

applying a stray-light correction matrix C

to the measured signal, followingf the

method of Zong, 2006.

Stray-light The slit functions were determined at the tuneable laser setup PLACOS at PTB in the range 300 to 1000 nm.

YIB=C Ymeas

The stray-light corrected signal YIB can

be retrieved from the measurement by

applying a stray-light correction matrix C

to the measured signal, followingf the

method of Zong, 2006.

Is that effort really necessary?

Why not use a single slit function to determine the stray-light matrix?

The Line-Spread functions of PSR 003 were measured at the tuneable laser facility PLACOS at PTB

The enhanced stray-light above 700 nm is specific to PSR 003 and is not consistent with

PSR 004, 005 and 006 even though the instruments were all built in house as the same batch ???

For PSR 004, 005 006 we need to use a IR corrected version (gray-line in the video).

Stray-light validation

1) Tungsten-halogen lamp and Band-pass filters

2) Solar measurements and band-pass filters

3) Solar measurements at Izaña

Stray-light validation 1) Lamp & band-pass filters

Raw Spectra from a FEL Lamp Spectrum with several band-pass and cut-on filters

Stray-light validation 1) Lamp & band-pass filters

Stray-light from the LSF of

the IR spectrum.

Stray-light validation 1) Lamp & band-pass filters

4 PSR, Stray-light from PSR 003 is significantly higher than from all others

Stray-light validation 1) Lamp & band-pass filters

PSR 003

PSR 003

The Stray-light correction using the

appropriate SR-Matrix is effective in

removing more than 95% of the SR-Signal

in PSR 003

Direct optic flat field

Following WMO specifications for sunphotometers measuring direct solar

irradiance:

• less than 2° field of view

• Slope angle of 1 °

Direct optic flat field – spectral "waves"

±1%

±2%

Standard Diffuser,

sandblasted on one

side

Standard Diffuser,

sandblasted on two

sides,

3 times more

transmission loss

Direct optic flat field – spectral "waves"

±1%

±2%

Standard Diffuser,

sandblasted on one

side

Standard Diffuser,

sandblasted on two

sides,

3 times more

transmission loss

Where do these spectral waves come from ?

AND

How to get rid of them?

Langley-Calibration campaign Izaña March-June 2014 2 May 2014

Langley-Calibration campaign Izaña March-June 2014

Langley Plot

Langley-Intercept and literature spectrum

27 May 2014

Stray-light effect on solar measurements

Ratio between Langley extrapolated

ET Spectrum and literature ET

Stability and comparison to literature ET

Variability ±0.8%

±5%

Spectral AOD compared to CIMEL and PFR

CIMEL PFR

O2-O2

Water vapour

Spectral AOD compared to CIMEL – 10 May 2014

08 10 12 15 17180.010.020.030.04

340

08 10 12 15 17180.010.020.030.04

380

08 10 12 15 17180

0.010.020.03

440

08 10 12 15 17180

0.010.020.03

500

08 10 12 15 17180

0.010.020.03

675

08 10 12 15 17180

0.010.020.03

870

08 10 12 15 17180

0.010.020.03

10-May-2014

PSR-CIMEL < ±0.01

PSR

CIMEL

Further steps...

• Investigate the strange line spread functions of PSR 003 above 750 nm, which seem to be due to enhanced light scattering (from black anodized metal?)

• Connect a telescope to the second SMA port for lunar irradiance and nighttime AOD retrieval

• Construct additional PSRs ...

Black anodized

Courtesy Andreas Höpe, PTB

First lunar light at Izaña, May 2014

The PSR Team

Dani

Diego

Natalia

Ricco

Manfred

Fabian

Claudio

Silvio

Peter

Julian

Etienne

2009: 2014

>6 (wo)man-years development time