model, i.e., to transition from a search for correlation to
a search for causation. The previously introduced firstorder
reversal curve current-voltage (FORC-IV) SPM
technique [74] has been deployed in imaging and analysis
of spatially uniform Ca-substituted BiFeO3 and NiO systems
[74,75]. Those studies have shown that the locally
measured hysteresis in the FORC-IV curves is related to
changes in electronic conduction sensed by the tip in response
to a bias-induced electrochemical process, and the
area of the IV loop is overall indicative of local ionic activity.
FORC-IV spectroscopic imaging modes lack adequate
data analysis and interpretation pathways due to the flexible,
multidimensional nature of the data set and the volume
of the data collected. In this example, we combine
FORC-IV measurements with the multivariate statistical
methods based on signal de-mixing, in order to discriminate
between different conductivity behaviors based on the
shapes of the IV curves in the full spectroscopic data set.
A CoFe2O4-BiFeO3 nanocomposite thin film (CFOBFO,
Figure 8c) was grown by pulsed laser deposition
and is a self-assembled, tubular heterostructure that
forms spontaneously due to segregation of the perovskite
BFO matrix and the CFO spinel inclusions [58,76]. The
FORC-IV spectroscopy was performed at humidity values
ranging from 0% to 87%, with an intermediate 58% case
also shown. To gain insight into the fine structure of the
CFO pillars and the CFO-BFO tubular interface, we
first used conductive AFM (c-AFM) to image areas of
size 500 × 500 nm2 of the film and then collect FORCIV
data using a waveform with six triangular pulses and
a maximum DC peak bias of 3 V on a 50 × 50 pixel