Measuring steady-state absolute photoluminescence (APL) in perovskite solar cells (PSCs) is an important characterisation method, being able to provide information about defects and recombination pathways of various types of layer-stacks, due to its contactless nature. APL can also be measured on full devices with contacts, enabling the measurement of photoluminescence under voltage bias, or PL(V). This changes the nature of the measurement: one now measures both the extracted current and the photoluminescence simultaneously as a function of the applied voltage, in a similar fashion to the standard I-V sweep. Charge  extraction now also plays a role in the resulting photoluminescence, in addition to recombination pathways. The information gained by this technique, while valuable, is difficult to interpret, as both the current and photoluminescence are not independent of each other and dependent on measurement conditions.

This poster thus aims to present some results of PL(V) measurements done on PSC devices with various modifications to the layer stack, alongside the measurement setup and protocol that have been developed to best ensure an objective comparison between these devices.

The measurement setup consists of a bare fiber placed very close to the transparent electrode of a PSC, which is illuminated by a 455 nm LED to provide the excitation. The non-contacted parts of the cell are masked, to ensure no extraneous PL signal is measured. The electrical contact is with 2 pins on each metal contact, to have a 4-wire connection to the SMU. To convert the measured signal to absolute irradiance, a reference spectrum is measured using an integrated sphere in combination with a calibrated spectrometer.

The measurement protocol was developed to ensure of reproducibility and measuring in steady-state. The measurement itself consists of a reverse scan without illumination, illuminated reverse scan and illuminated forward scan, in that order. The amount of time and the state of illumination between measurements, as well as the scan speed, are determined in the software to ensure similar conditions for different devices. Additionally, multiple PL and current measurements are taken per bias point for checking if steady state is reached.

For this work, inverted PSCs with a bandgap of 1.63 eV were tested using this protocol, with modifications being made to either the perovskite-ETL interface via an interlayer or to the ETL itself, either by changing the thickness or ETL material. The goal of these changes was to see if changing the charge extraction properties would affect the PL(V) measurements.
Using 5 nm LiF and 5 nm BCP interlayers was the initial modification: these turned out to be unsuitable due to the IV properties of the cells drastically changing during the PL(V) measurement.
The next modification was to change the layer thickness of the ETL, C60, from the standard 20 nm to 150 nm. The higher thickness sample showed reduced charge extraction, as seen in its lower fill factor. This is also reflected in the PL(V) data, which shows a higher radiative current measured around the maximum power point. There is, however, little difference at short circuit, showing that charge extraction is not hindered under this condition.
Transient photovoltage data shows that, at high light intensities, the rise time between the two devices is similar, while it becomes longer for the 150 nm C60 device at low light intensities; decay time is similar between the devices.
In the end, while the device modifications do influence the extraction and thus the measured PL(V), a direct relation between extraction and PL(V) is still obscure, and warrants further research.