Tip:
Highlight text to annotate it
X
Hello my name is Brad Macleod and work with David Ginger in the Department of Chemistry
at the University of Washington and today I’m going to show you how to perform the
external quantum efficiency measurement. External quantum efficiency or EQE is a measurement
of the number of electrons that you can extract from a photovoltaic device per incident number
of photons. First of all we have the lamp source, inside a lamp and some optics to focus
the light into the monochromator. And the monochromator allows us to select certain
wavelengths of light to perform our spectral studies. We focus this light with some optics
on to - in this case we are looking at photodiode that is calibrated so that we know how much
light is coming out of the source. We have a custom sample chamber that allows us to
test our samples under inert atmosphere or under vacuum. From that sample chamber we
have a switch box because the substrates actually have several devices on them, and a source
probe unit, the Keithley 2400 which an industry standard for measuring current when we apply
voltage. And in order to test our devices under vacuum we have a rotary vane vacuum
pump with a valve we can turn on and off to load samples. In addition to that we have
the power supply for the lamp and these different pieces of hardware are interfaced with our
computer via the LabVIEW software by National Instruments. In order to accurately determine
the external quantum efficiency we need to measure the photocurrent response of our calibrated
diode. And then we’ll need to measure the photocurrent response of our device under
test. So to do this we’ll take the calibrated photodiode and hook it directly into the source
probe unit and we actually have a mask for our photodiode. And what the mask allows us
to do is control the illuminated area on the photodiode and why we do this is we want to
illuminate the same area on the both the diode and the solar cell to get an accurate EQE
measurement because we need to insure that the flux of the incidental light is equivalent
for both measurements. So to measure the photocurrent response of
the diode we’ll make sure that the lamp, we might set the monochromator to some wavelength
that is easy to observe by eye. About 532 nanometers will give you a nice visible green
spot. And then we adjust back and forth and side to side using our micrometer stage to
center the beam on the masked off area of the photodiode. And we also have a measurement
to move this up and down to insure that it is directly in line with the light. And we’ll
use the source-probe unit to help us achieve the maximal photocurrent by adjusting these
positions and this insures us that we’ve aligned the masked area of the photodiode
with the peak intensity of the lamp. And that should be reproducible whether we are using
the masked photodiode or the masked test device. We can see that if I move side to side the
current goes down pretty rapidly. This is an easy way to align thing reproducibly by
just finding the maximum photocurrent at a given wavelength. So right now for the given
monochromator settings of our lamp and the given power settings from our lamp supply
this particular photodiode being masked off to a certain area is giving us 22 microamps.
And since we know the masked area of our mask we can determine a photocurrent density as
well. To control the experiments we use a graphical programming language known as LabVIEW
that is produced by National Instruments. This window that I have open is to measure
the photocurrent spectra of a device all we really need to do is set the start and stop
wavelengths for the scan. This will change the center wavelength of the monochromator
and allow us to step through different wavelengths and measure the current as a function of wavelength.
So we set the start wavelength, the finish wavelength the step size and any other settings
that we might want to change to the hardware or software to improve the sensitivity of
the measurement. So right now I have it set to start 400 nm, finish at 800 nm in steps
of 10 nanometers. When the code is executed from the computer it sends the signal over
standardized communication hardware by the acronym GPIB which is just the standard cable
communication system that National Instruments uses. And it will instruct the monochromator
to shift the grating internally to diffract a certain wavelength of light and additionally
we also have equipped on our monochromator order sorting filters so as we move to lower
energy wavelengths we filter out harmonic components from higher energy harmonics. So
I’ll just run this scan now. And give it a filename. This scan will probably take about
a minute or two, not very long. It could take longer if we want to increase the resolution
or sensitivity. Now we have a plot of the photocurrent spectrum for the calibrated silicon
photodiode. So next we want to actually test our device. We have a 15 mm square substrate
ITO that we have fabricated a device on. You’ll see 8 little fingers sticking off, 4 from
each side. These are the aluminum top electrodes. You’ll see 8 little blotches and a 9th larger
blotch that is silver paint that we use to soften the mechanical contact between our
test setup and the actual device. So normally when we have made a device and we want to
get it ready to load it for testing we do this from the inside the glove box so we have
designed this sample holder to be small enough to fit into the antechamber of the glove box
system. But for the purpose of demonstration we will just do this on the lab bench. This
has got kind of a tight fit from the seal there. So what you’ll notice is that we
have feed through for the wiring. There’s about 8 individual wires one for each pixel
on the substrate plus one for the common contact. There’s optical access from both sides.
And this particular device holder we have two thumb screws that you just have turn to
loosen and unsecure this plane here and the you just pull this out and in this particular
device holder we have push pins and the idea with this design is that it can be easily
manipulated inside the glove box when you have thick gloves on but also we want to minimize
the stress on wires during use so that it has a longer lifetime in operation. Then using
our tweezers we can just pick up the devices that was already in there and underneath the
device you can see this mask that was fabricated to mask off the illuminated area of our substrates.
You can see that there is 8 little slots corresponding with the 8 finger shaped pixels on our substrate.
So we replace the mask and make sure it’s faced properly and take our next device and
place the device face up so that the mask is touching the glass side of the substrate
and the aluminum contacts are facing upward because those parts need to come into contact
with our external circuitry. We want to try to align the pixels with the underlying mask
visually before we secure the device but we can sort of move it around after we have secured
the device with this back plane. We just want to make sure that the common electrode here
at the top is oriented such that it will hit the spot of silver paint that we have put
on the substrate. Gently put this in and give it a slight push down so we can secure the
thumbscrews. You might be able see that the mask wasn’t aligned very well because you
can look through it and see completely through to the other side of the device holder. The
easiest way to fix this is to hold the sample holder up in front of a light source and adjust
the mask until you can’t see light coming through it anymore.
So now we have the sample secured in the device holder and the mask is aligned with the 8
different pixels we’ll load it into the chamber. We want the device to be normal with
one of these four optical axis windows. In our particular sample chamber we have little
holes on the lid and a notch on the chamber to make that easy in the case we want to do
normal incidence experiments or even at an angle of 45 degrees. We just have to make
sure that this gets pushed in and the seal is set. And then we take our vacuum system
and latch it on. And open the vacuum to the chamber. At this
point our device is under vacuum. Using a rotary vane pump it should maintain pressures
in the millibar range. We do this to prevent effects from the presence of oxygen and moisture
to the degradation of polymers that we use in the devices. So now we’ll just place
this on our translational stage. So I’ll move the photodetector and secure the sample
chamber in this base and we just want to make sure by eye that the windows are kind of faced
off and normal to the incident light that we will be testing with. And at this point
you’ll notice that this assembly can be kind of top heavy so it might be a good idea
under normal testing conditions to secure this somehow and reduce vibrations. So then
we have a switch box so that we can test the 8 different pixels on our substrates and that
just plugs into the feed-through on the top of the chamber like so. Then we take the BNC
cable from the photodiode assembly and plug it on to the switch box. Now we are ready
to test the solar cell using the LabVIEW software. We can test one pixel at a time and in order
to test it under lighting conditions we’ll have to align it similarly to the way that
we aligned the photodetector. Using this technique where we look at the photocurrent and then
adjust the micrometer translational stage requires that the device is efficient enough
and produces enough photocurrent otherwise you may have to rely on some other kind of
signal or just visually aligning it by eye. So I’ll go ahead and attempt to align pixel
one. So I set the switch box to pixel one. I turn the source probe unit on to local and
output on and then we get a live display of the current coming out of that device. Now
the monochromator is still set to the last setting that we programmed when we ran the
photocurrent spectrum for the detector which ended at 800 nm so right now its shining light
that I can’t really see so I’ll have to go set it to something that I can actually
observe by eye. So I just pick the wavelength 523 nm in this case and run that program.
So we’ve set the monochromator to a wavelength that is easily observable by the human eye
in the green and now I’ll align pixel number one using the photocurrent and my visual observations
of the beam’s position to align the pixel. So right now it’s actually off to the side
so I’ll adjust this side to side micrometer. It’s getting close to the pixel so I’ll
look at the photocurrent for a change. So now that we’ve found the maximum photocurrent
for pixel one, we’ll go ahead and run the LabVIEW photocurrent spectral measurement
to measure the photocurrent as a function of incident wavelength. So now I’ll be measuring
the photocurrent spectra of our test device over the same range and same software settings
that we used to measure the calibrated photodetector. When this is completed we can perform the
EQE calculation using the photodetector calibration file and the two spectra we have collected.
So you will actually notice something different about this device’s photocurrent spectrum
and the silicon photodetector’s photocurrent spectrum. Thought the photocurrent spectrum
of the device actually peaks in negative current where the silicon photodetector peaked in
positive current. And that is just due to the polarity of the devices, we can arbitrarily
inverse that or switch the polarity of the test cell to accommodate for that. So now
I would take the two photocurrent spectra one of test device and one of the silicon
photodiode bring that into a spreadsheet with the calibration data for our silicon photodiode.
The calibration data known as the responsivity is just a measure of the amount of current
that the photodetector produces for a given incident amount of light power. And using
these three pieces of data we can calculate the number of electrons per unit time that
our test device is producing under illumination at a given wavelength for the number of incident
photons on the device per unit time and it’s the ratio of those two that is the EQE. This
is my way of doing this calculation I have a spreadsheet template that I just copy and
paste the two spectra that I have collected. The responsivity data for this particular
diode is already in this file so I don’t have to worry about that. So copy and paste
the photocurrent spectra of the detector to one column, and the photocurrent spectra from
the test cell to another column. What we would see from the external quantum efficiency is
that it has some spectral peak line shapes and those line shapes are correlated with
the absorbing polymer. So the peak efficiency is occurring somewhere around 500 nm incident
irradiation which corresponds with the peak absorption of polythiophene. We the photocurrent
spectrum of the calibrated silicon photodiode and to that we apply the calibration responsivity
which gives us the responsivity is in units of amps per watt, so for the amount of photocurrent
we measure it tells us how much power of light was incident on the detector at every wavelength
of the spectra so we are able to determine the power density spectrum is for the lamp
system and then we divide our photocurrent spectrum of the test cell by this power density
spectrum we are able to determine directly what the EQE is by knowing how many electrons
there is for the current that we measured. If you are not careful with the measurement
the units are going to be arbitrary. But you can still get spectral features. The actual
units would just be a percentage 0 to 1. If this were actual units this would be a very
poorly performing cell having .1 percent or .2 percent efficiency. We can still compare
the relative performance of devices with relative units assuming that all the different settings
in our setup were identical. The actual external quantum efficiency requires that we pay special
attention to things such as how we’ve defined the active area, how much of the active area
is illuminated compared to its total area, as well as the intensity of the light that
we use, the temperature it’s tested under, all these different parameters. We have to
consider comparing any two absolute values in units that are meaningful. For these particular
devices where we have ITO spin coated with PEDOT PSS, spin coated with a bulk heterojunction
blend of polyhexylthiophene and PCBM and some top contact like aluminum we should expect
anywhere from as low as 50% to as high as maybe 80% or more external quantum efficiency
at the peak wavelength which for these devices should be in 500 nm range.