Mapping fast evolution of transient surface photovoltage

Preparing to load PDF file. please wait...

0 of 0
100%
Mapping fast evolution of transient surface photovoltage

Transcript Of Mapping fast evolution of transient surface photovoltage

Notice: This manuscript has been authored by UT Battelle, LLC, under Contract No. DEAC0500OR22725 with the U.S. Department of Energy. The United States Government retains and the publisher, by accepting the article for publication, acknowledges that the United States Government retains a non-exclusive, paid-up, irrevocable, worldwide license to publish or reproduce the published form of this manuscript, or allow others to do so, for the United States Government purposes. The Department of Energy will provide public access to these results of federally sponsored research in accordance with the DOE Public Access Plan (http://energy.gov/downloads/doe-public-access-plan).
1

Submitted to ACS Nano
Mapping fast evolution of transient surface photovoltage dynamics using G-Mode Kelvin probe
force microscopy
Liam Collins,†,‡ Mahshid Ahmadi,ζ Jiajun Qin,ζ Olga S. Ovchinnikova,†,‡ Bin Hu,ζ Stephen Jesse†,‡ and Sergei V. Kalinin,†,‡
† Center for Nanophase Materials Sciences, Oak Ridge National Laboratory, Oak Ridge, Tennessee 37831, USA
‡ Institute for Functional Imaging of Materials, Oak Ridge National Laboratory, Oak Ridge, Tennessee 37831, USA
ζ Joint Institute for Advanced Materials, Department of Materials Science and Engineering, University of Tennessee, Knoxville, Tennessee 37996, USA
Keywords: Atomic Force Microscopy (AFM), Kelvin probe force microscopy (KPFM), Surface photovoltage, Organic-inorganic hybrid perovskite.
2

ABSTRACT. Optoelectronic phenomena in materials such as organic/inorganic hybrid perovskites depend on a complex interplay between light induced carrier generation and fast (electronic) and slower (ionic) processes, all of which are known to be strongly affected by structural inhomogeneities such as interfaces and grain boundaries. Here, we develop a time resolved Kelvin probe force microscopy (KPFM) approach, based on the G-Mode SPM platform, allowing quantification of surface photovoltage (SPV) with microsecond temporal and nanoscale spatial resolution. We demonstrate the approach on methylammonium lead bromide (MAPbBr3) thin films and further highlight the usefulness of unsupervised clustering methods to quickly discern spatial variability in the information rich SPV dataset. Using this technique, we observe concurrent spatial and ultra-fast temporal variations in the SPV generated across the thin film, indicating that structure is likely responsible for the heterogenous behavior.
3

INTRODUCTION
Solar energy is critical towards meeting the ever increasing global energy demands, necessitating the continuous search for novel photovoltaic materials and device structures. Organic/inorganic hybrid perovskites (OIHPs) solar cells have gained significant attention with an impressive power conversion efficiency of 22.1%.1 Beyond this, methylammonium lead bromide (MAPbBr3) has attracted wide interest as an active material in light emitting diodes (LEDs)2 and as an ideal candidate in tandem structures with a Si or a Cu(InGa)Se2 solar cell.3 This is due to its superior charge transport with high and nearly balanced carrier mobility for both electrons and holes, large light absorption coefficient, strong photoluminescent quantum efficiency, large and tunable band gap, low defect density as well as solution processability and low cost.4-6
Despite tremendous advancement in OIHP optoelectronic device performances in the last decade, a full understanding of several observed physical behaviors, including coupled fast and slow relaxation time scales,7-9 hysteretic transport behavior,10 and non-uniform optoelectronic characteristics11-14 remain largely elusive. Many material challenges remain unsolved, and it is widely accepted that optimization and improved longevity of OIHP devices requires precise knowledge of the charge distribution behavior at local inhomogeneity’s (e.g. grain to grain variability, grain boundaries, interfaces and, and defects). 15-16
While scanning probe microscopies (SPM) are ideally suited for spatial probing of materials structure and functional properties on these length scales, they are inherently slow, restricting applications to equilibrium or very slow processes (e.g. > seconds). Of particular relevance for photovoltaics, Kelvin probe force microscopy (KPFM)17, is a non-invasive mode of AFM which allows the simultaneous mapping of the topography and the local contact potential
4

difference (CPD) between tip and sample with nanometer resolution. The CPD value measured by KPFM is related to work function for metal samples, surface potential in dielectrics, band bending in semiconductors and surface photovoltage (SPV) in photovoltaics.18-19 The ability to correlate nanoscale structural and electronic/electrochemical characteristics has made KPFM a powerful method to study electronic devices including solar cells and LEDs.20-24 In particular, spatially resolved SPV measurements can be related to important characteristics including carrier diffusion length20-21 carrier lifetime and local recombination rates,25 characteristics which are fundamental to understanding charge generation process in photovoltaic materials.26 At the same time, capturing information on these processes requires KPFM techniques which are capable of probing both fast (ns-µs) and slow (ms-s) processes.
The factors limiting the measurement bandwidth in classical KPFM are the lock-in amplifier time constant and the bandwidth of the bias feedback loop, well below the mechanical bandwidth of the AFM cantilever itself. Practically, measuring SPV using KPFM involves capturing the CPD or surface potential (Vsp) under both illuminated (Vsp-light) and dark (Vsp-dark) conditions, where SPV=Vsp-light - Vsp-dark. However, a KPFM image under either condition can take several 10s of minutes to capture, and hence, the accuracy of the SPV measurement depends on the stability of the tip-surface potential in the time frame of the entire measurement. Furthermore, the long measurement time makes studies of SPV under different environmental or illumination conditions (wavelength, intensity, and light soaking/degradation effect) impractical.27 This is especially important in the study of materials including OIHPs which involve ion migration.28
As the result, in KPFM measurements all information on dynamic processes below the measurement timescale (e.g. ms) is lost. At the same time, a complete understanding of the role
5

of trapped charges and ion migration play in anomalous observations such as the hysteretic behavior and light soaking effect in OIHPs would benefit greatly from time-resolved (tr) SPV measurements using KPFM. Indeed, the KPFM community have tried to respond in recent years by developing tr-Electrostatic force microscopy30-34 and tr-KPFM27,35 approaches for investigation of time dependent optoelectronic properties. However, these approaches have some limitations including either; lacking quantitative information on local potentials, 30-33 sometimes operated in single point mode (i.e. no spatial contrast),27, 33 or where quantitative KPFM imaging has been realized the time resolution is still limited to ~16 seconds.35
Recently we developed the G-Mode acquisition approach for SPM measurements,36 and subsequently developed open loop G-Mode KPFM37-38 as a method of probing surface potentials on microsecond timescales, well below the mechanical bandwidth of the cantilever.39 Here we combine G-Mode KPFM with photoexcitation as a method to map SPV dynamics of photovoltaic samples with temporal properties. As a proof of principle, we choose a thin films of methylammonium lead bromide (MAPbBr3) on ITO/PEDOT:PSS substrate. We further highlight the usefulness of adopting unsupervised clustering algorithms for quickly and effectively discerning local deviations in optoelectronic properties from the high dimensional and information rich datasets afforded by G-Mode KPFM. RESULTS AND DISCUSSION Figure 1 describes the measurement setup. All KPFM measurements were operated in lift mode, or dual pass mode, in which the sample topography is recorded in the first pass, and the KPFM measurement is performed during a second pass at a predefined distance above the sample (100 nm unless otherwise stated). The G-Mode platform was used to capture the photodetector signal at high sampling rates (~ 2-4 MHz) as the tip was raster scanned (scan rate ~ 0.4 Hz) over the sample in lift mode. For G-Mode KPFM, a sinusoidal voltage is applied to the conductive
6

cantilever generating a dynamic electrostatic force between tip and sample. The dynamic

cantilever response due to the electrostatic force is encoded in the photodetector signal. De-

noising of the photodetector signal in G-Mode KFPM is realized in a post processing step using

Fourier filtering (Noise thresholding and low pass filter) on the entire data, parsed into individual

line segments. We note that the knowledge of the noise floor and the capability to inspect the

data allows for adaptable post-experiment filtering, not traditionally afforded in laboratory

settings. Further, G-Mode KPFM allows flexibility in exploration of frequency filter and noise

threshold settings without effecting the original raw dataset. 40

Once the data has been processed, recovery of the true electrostatic force is achieved through

deconvolution of the cantilever transfer function (calibrated at the beginning of the

measurement) using the Fast Free Force (F3R) reconstruction method outlined previously.39 In

the case of EFM or KPFM, the tip-sample electrostatic force (𝐹𝑒𝑠), established between the

grounded sample and conductive probe is written as:

𝐹 = 1 𝐶′(𝑉 − 𝑉 )2

𝑒𝑠 2

𝑡𝑖𝑝 𝑆𝑃

Eq. (1)

Where C' is the tip-sample capacitance gradient, Vsp is the surface potential or more precisely

the CPD between tip and sample. The tip voltage is Vtip=Vdc+Vaccos(ωt); however, application of

a Vdc bias offset is not a requirement in G-Mode KPFM and in principle any arbitrary waveform

can be adopted. As seen from Eq (1), the recovered electrostatic force is expected to have

parabolic voltage dependence. The quantitative values of Vsp can be determined by fitting the

functional form of the force vs voltage relation. Correspondingly, the readout rate of the Vsp is

governed by the time per period of oscillation of the tip voltage. After functional fitting of the

recovered force, the data matrix comprises a multidimensional dataset of Vsp (x, y, time).

Although not considered in this work, the functional form of the force vs bias relationship can be

7

further related to information on the capacitance (i.e. second order fit parameter), as well as charging events or polarization effects (i.e. deviation for purely parabolic response) which can be inferred from the functional form of the response.
For SPV measurements the excitation laser is periodically modulated on and off to induce a photo response in the sample. The laser excitation waveform can be configured in multiple ways; here we chose to include a single illumination event per pixel (4.096 ms) such that the light was modulated on (2.048 ms) and off (~2.048 ms) at each spatial location. The SPV was calculated by subtracting the Vsp recorded under illuminated conditions from the Vsp under dark conditions.
Figure 1. Schematic of the measurement set-up for surface photovoltage measurements by GMode KPFM.
Shown in Figure 2(a) is the topography height profile of a representative area on the MAPbBr3 thin film sample, having grains of ~1-2 µm in size. The KPFM surface potential maps show variations in Vsp at grain boundaries, within grain facets, and in defective regions, as shown
8

in Figure 2(b). The heterogeneous variation in surface potential at grain boundaries or within single facets could be a result of any number of effects including local doping, or chemical segregation upon crystallization of the film, presence of defect states or shallow trapping levels or even creation of small polaronic states due to localized lattice strain and molecular orientations.13-14, 41-42 However, the focus of this work is not to determine the precise origin of such variation, instead we attempt to capture of characteristic light induced photovoltage behavior by developing a time resolved approach.
KPFM measurements under dark and illuminated conditions were performed on a smaller area of the sample as shown in Figure 2(c). Using standard KPFM, it was found that upon illumination (see Figure 2(f)), a decrease in the measured Vsp was observed relative to the Vsp measured under dark conditions (see Figure 2(e)), indicating a reduction of the work function of the MAPbBr3. The calculated SPV is shown in Figure 2(d), demonstrating an average SPV value of -78 ± 12 mV. Local grain to grain variations in the SPV can be observed, as well as variations in SPV in areas correlating with grain boundaries. In addition, the negative SPV (Figure 2(d)) indicates a downward band bending due to accumulation of negative charges at the surface. Meanwhile, a gradual positive drift in SPV can be seen in the direction of the scan (bottom to top), this can be indicative of a slow (>>sec) relaxation process within the material likely a result of excess charge relaxation through the film thickness, ion migration or the relaxation of trapped charges.35 Notably, while KPFM captures the time averaged processes and is suited for probing such slow processes, these measurements provide little to no information on fast processes taking place below the KPFM measurement time (~11 min).
9

Figure 2. KPFM surface photovoltage measurement on MAPbBr3 thin film. (a,b) Topography and surface potential of a 25 µm2 region. (c,d) Topography and SPV in a smaller region indicated by a red box in (a). SPV was calculated from subtraction of the Vsp measured under (f) dark and (e) illuminated conditions.
To explore the fast (µs – ms) light-induced SPV dynamics in MAPbBr3 thin film, we utilize G-Mode KPFM. As a first representation, the time averaged SPV data is shown (Figure 3(b)), determined by subtracting the mean SPV value recorded during the illuminated (Figure 3(c)) and dark (Figure 3(d)) states for each pixel. In agreement with Figure 2, we see a decrease in the surface potential upon illumination. However, for G-Mode KPFM we observe a much larger SPV values than that observed in standard KPFM (-140 ± 28 mV vs -78 ± 18 mV). This is likely due to differences in measurement timescales between G-Mode KPFM and classical KPFM that allow the observations of light-induced fast processes prior the onset of ionic screening, see Figure 2(d). Furthermore, the SPV contrast at the grain boundaries was found to be larger and more pronounced in G-Mode KPFM than in standard KPFM. This result could also
10
SpvVspKpfmTimeProcesses