WO2020081418A1 - Methods and systems for high-throughput multi-modal optical characterization of samples - Google Patents

Abstract

Methods and systems for multi-modal optical characterization of sample regions are provided. In an embodiment, a method comprises performing an initial screening by illuminating a pattern of sample regions with a patterned light beam to induce a patterned optical signal and detecting the patterned optical signal to generate a first set of data comprising optical values and associated sample region values; performing a first subsequent screening by illuminating a subset of sample regions of the pattern of sample regions with a light beam under confocal point illumination to induce optical signals and detecting the optical signals to generate a second set of data comprising optical values and associated sample region values; performing a second subsequent screening by conducting a correlative characterization technique on the subset of sample regions to generate a third set of data comprising chemical or structural values and associated sample region values; and performing a statistical analysis on multi-modal data comprising at least the second and third sets of data to generate correlation data.

Description

METHODS AND SYSTEMS FOR HIGH-THROUGHPUT MUUTI-MODAU OPTICAU CHARACTERIZATION OF SAMPUES

CROSS-REFERENCE TO REUATED APPUICATIONS

[0001] The present application claims priority to U.S. provisional patent application number 62/745,636 that was filed October 15, 2018, the entire contents of which are incorporated herein by reference.

REFERENCE TO GOVERNMENT RIGHTS

[0002] This invention was made with government support under FA9550-14-1-0005 awarded by the Air Force Office of Scientific Research. The government has certain rights in the invention.

BACKGROUND

[0003] The discovery of new molecules and materials is slow, expensive, and laborious. The high costs necessitate either a slow, high-quality search or a fast, low-quality search through the enormous chemical and physical phase space where the most promising targets are often missed. Several long-standing problems in the field include (i) the need for multimodal observations that require multiple sophisticated techniques, each with their own specialized hardware and software, (ii) a lack of standardization across different laboratories, (iii) variations in datasets arising from heterogeneity in materials processing, sample preparation, and characterization techniques, and (iv) fundamental limits on the information available from current techniques. In view of these issues, major advancements have been largely serendipitous.

SUMMARY

[0004] Methods and systems for high-throughput multi-modal optical characterization of samples are provided. Principal features and advantages of the disclosure will become apparent to those skilled in the art upon review of the following drawings, the detailed description, and the appended claims. BRIEF DESCRIPTION OF THE DRAWINGS

[0005] Illustrative embodiments of the disclosure will hereafter be described with reference to the accompanying drawings.

[0006] FIG. 1 shows a block diagram of an illustrative system for multi-modal optical characterization of samples.

[0007] FIG. 2 shows the block diagram of FIG. 1, with additional detail showing an illustrative controller of the system.

[0008] FIG. 3 shows a flow diagram illustrating exemplary operations performed by an application of the controller of the system of FIG 2.

[0009] FIG. 4 illustrates the technique of multi-modal imaging using the system of FIG.

1. The left-most image shows an SEM image of a perovskite particle. Spectrally-resolved transient absorption (TA) at each point is measured as shown in the two middle panels. The SEM image identifies structural features of the perovskite particle. As shown in the six right most images, Fermi level, carrier relaxation, carrier density, and initial carrier temperature are optical properties which may be mapped by analyzing the TA signal from multiple locations in the particle.

[0010] FIG. 5 shows two-dimensional scatter plots and histograms of multi-modal data generated using the system of FIG. 1, revealing the correlation between amplitude of a sub- bandgap state (y-axis) and the initial carrier temperature (x-axis). Triangles show the positions of median values. The difference in the median values for locations at grain boundaries and within grain interiors is evident. Grain boundaries and interiors are two different morphologies found within single perovskite particles as shown in the correlated SEM images of FIG. 4.

[0011] FIGS. 6A-6G show spectrally-resolved TA spectra at grain interior (GI) and grain boundaries (GB) of a single particle of CH3NH3Pbl3. (FIG. 6A) SEM image of the perovskite particle. The solid curve traces the edge of the particle, and the dashed curves trace the grain boundaries between grains. The ones in the inner circle are not marked, as they are too dense. Scale bar: 1 pm. (FIG. 6B) TA surface and (FIG. 6C) TA spectra with fits (black solid curves) at different pump-probe delays at one position picked in grain interiors. (FIG. 6D) TA surface and (FIG. 6E) TA spectra with fits (black solid curves) at different pump-probe delays at one position picked at grain boundaries. (FIG. 6F) Sub-bandgap feature extracted from fit, with spectrally integrated (1.570 ± 0.015 eV) sub-bandgap intensity and single exponential fit (solid curve). The population lifetime is labeled. (FIG. 6G) Cuts of TA spectra at grain boundary at 1.0 ps and 1.8 ps, with fits (black solid curve) and residues (dashed curve, scaled). An arcsinh-scaled color map is applied on FIGS. 6B, 6D, and 6F.

[0012] FIGS. 7A-7D show the correlation between sub-bandgap states and carrier properties and the results of a statistical analysis on the multi-modal data. Two-dimensional scatter plots of sub-bandgap state intensity and (FIG. 7A) carrier temperature and (FIG. 7C) quasi-Fermi energy at different delay times. Two-dimensional scatter plots of carrier cooling rate and temporally averaged (FIG. 7B) sub-bandgap state intensity and (FIG. 7D) quasi- Fermi energy. Triangles show the positions of median values. Closed circles: GB;“X”: GI. The p-value for testing the null hypothesis of no correlation between the two properties shown along x and y axes is marked.

[0013] FIG. 8 shows a block diagram of the transient absorption microscopy (TAM) setup used in the Example. BBO: beta barium borate crystal. YAG: yttrium aluminum garnet crystal. DCM: double chirped mirror pair. Inset: initial carrier density map. Scale bar: 1 pm.

[0014] FIG. 9A shows two-dimensional scatter plots of sub-bandgap state intensity and carrier temperature at different delay times. Triangles show the positions of median values. FIG. 9B shows two-dimensional scatter plots of temporal averaged sub-bandgap state intensity and carrier cooling rate. Triangles show the positions of median values. The p-value for testing the null hypothesis of no correlation between averaged sub-bandgap state intensity and carrier cooling rate is marked on the left comer.

DETAILED DESCRIPTION

[0015] Methods and systems for high-throughput multi-modal optical characterization of samples are provided. The methods and systems may be used for the rapid discovery of new materials, including the discovery of new semiconducting materials, e.g., for use in photovoltaics and light-emitting diodes, as well as the discovery of room-temperature qubits based on optical excitation.

[0016] In an embodiment, a method for multi-modal optical characterization of a sample comprises performing an initial screening of a pattern of sample regions using a patterned light beam. The“pattern of sample regions” may refer to a plurality of distinct, individual samples arranged in a particular pattern. Alternatively, the“pattern of sample regions” may refer to a plurality of distinct, individual spatial locations within a larger single sample, the spatial locations arranged according to the particular pattern.

[0017] Individual sample regions may be defined and labelled using a substrate patterning procedure in which a polymer (e.g. SU8) is deposited on a substrate in a unique pattern (i.e., with unique labels based on a numbering, lettering, or other encoding scheme). Other technologies may be used for patterning such as e-beam lithography, which gives smaller and more well-defined patterns, or by using a reference TEM copper grid. In any case, the sample (or plurality of distinct, individual samples) is placed in contact with the patterned substrate. This patterning allows a set of coordinates to be associated with each sample region within the pattern, thereby identifying the physical location of each sample region within the pattern. These coordinates may be referred to as sample region values throughout the present disclosure. The patterning also allows the patterned substrate to be removed from one microscope and taken to other instruments, e.g., an electron microscope, where the exact same positions (i.e., sets of coordinates) may be found. This registration is important because each sample region being interrogated is only a small fraction of the total area of the patterned substrate (< 10 ⁶). Software may be used to align the fie ld-of- view of the sample region to be interrogated by an optical microscope with an image to be acquired by an electron microscope or other high-resolution instrument (e.g. an atomic force microscope).

[0018] The“patterned light beam” refers to a light beam which comprises a plurality of distinct, individual“beamlets” arranged in a pattern. Such a patterned light beam may be achieved, e.g., by using a diffractive optic(s) as described below. The particular patterns used for the sample regions and the light beam are not limited, but are generally the same or similar enough so that each sample region of the pattern of sample regions may be illuminated by an associated beamlet of the patterned light beam.

[0019] The patterned light beam interacts with the pattern of sample regions (each beamlet interacting with an associated sample region) to induce a patterned optical signal, which is detected using an appropriate detector, e.g., an array detector as described below. Thus, the initial screening step provides both parallel, simultaneous illumination of each of the plurality of individual sample regions as well as parallel, simultaneous detection of each of the induced optical signals. [0020] In the initial screening step, a single patterned light beam may be used in order to perform linear spectroscopy (e.g., absorption spectroscopy) or additional light beams may be used to illuminate and interact with the pattern of sample regions to perform non-linear spectroscopy (e.g., pump-probe transient absorption (TA) spectroscopy). Additional light beams may or may not be patterned; for instance, in a pump-probe experiment, the pump beam may be patterned, while the probe beam may be unpattemed such that a single light beam illuminates the plurality of individual sample regions simultaneously. The type of spectroscopic technique used is not particularly limited.

[0021] Regardless of the particular spectroscopic technique performed, the result of the initial screening step is a first set of data comprising optical values and associated sample region values. That is, each sample region has associated with it an optical value generated from the spectroscopic experiment. The optical value may be raw data output directly from the detector or processed data representing desired optical information calculated from the raw data. The sample region value is a set of coordinates identifying the physical location of the sample region (e.g., within a patterned substrate). By way of illustration, for an 8 x 8 array of sample regions and a pump-probe TA experiment, a first set of data may comprise 64 optical values, each optical value associated with one of the 64 individual sample regions.

The optical values may be processed data, e.g., various carrier properties calculated from raw data such as carrier temperature cooling rate. (See Example and below.) Depending upon the number of spectroscopic experiments and types of processing conducted, multiple first sets of data may be obtained from the initial screening step.

[0022] The initial screening step may be carried out on a wide-field microscope apparatus comprising any of the optical components, electrical components, electro-mechanical components, light sources, detector, controller, etc., typically used to carry out the desired spectroscopic technique(s). An illustrative wide-field microscope apparatus is described below. Illustrative data sets obtained using such an apparatus could include those comprising optical values (processed from raw detector data) such as absorption and emission spectra, bleach recovery time, hot-carrier relaxation rate, bandgap energy, Fermi level, exciton/free- carrier ratios, sub-bandgap energy, sub-bandgap rise and decay times, carrier

(electron/hole/exciton/polaron) temperature and density, and exciton binding energy. In the data sets, each of these optical values has associated with it the sample region value from which the optical value was generated. [0023] The method for multi-modal optical characterization of a sample further comprises performing a first subsequent screening of a subset of sample regions of the pattern of sample regions using confocal point illumination. The subset of sample regions may be those sample regions which exhibited desirable optical values from the initial screening step. This may be determined by comparing the obtained optical value(s) for a particular sample region to a predetermined optical value(s). By way of illustration, the subset of sample regions could include only those sample regions exhibiting an absorbance at a predetermined wavelength above a predetermined value. The desired/predetermined value depends upon the particular application for the sample regions being analyzed.

[0024] The first subsequent screening step may be carried out analogously to what has been described above with respect to the initial screening step. However, the resolution achieved in the first subsequent screening step is greater due to the use of confocal point illumination. Also, the confocal light point may be scanned to illuminate, sequentially, each sample region of the subset. In addition, the confocal light point may be scanned over each individual sample region, resulting in optical signals induced from individual spatial locations (which may be referred to as sample subregions) within the individual sample region. Briefly, light from the confocal point illumination interacts with each sample region of the subset to induce optical signals, which are detected using an appropriate detector, e.g., an array detector as described below. A single light beam may be used for the confocal point illumination in order to perform linear spectroscopy or additional light beams may be used to perform non-linear spectroscopy. Again, the type of spectroscopic technique used is not particularly limited. The result of the first subsequent screening step is a second set of data comprising optical values and associated sample region values. When a confocal light point is scanned over an individual sample region, the associated sample region values may be referred to as associated sample subregion values. The optical values may be raw data or processed data as described above with respect to the initial screening step. The sample region values are the set of coordinates that identify the physical location of the sample region as described above (or sample subregion values identifying the physical location of the subregions within the sample region).

[0025] The first subsequent screening step may be carried out on a confocal microscope apparatus comprising any of the optical components, electrical components, electro mechanical components, light sources, detector, controller, etc., typically used to carry out the desired spectroscopic technique. An illustrative confocal microscope apparatus is described below. Illustrative second sets of data obtained using such an apparatus could include those comprising any of the optical values described above with respect to the initial screening step. However, as noted above, the optical values from the first subsequent screening step will generally be of higher resolution and of higher signal-to-noise. In the second sets of data, each optical value has associated with it, the sample region value (or sample subregion value) from which the optical value was generated. Again, depending upon the number of spectroscopic experiments and types of processing conducted, multiple second sets of data may be obtained from the first subsequent screening step.

[0026] The method for multi-modal optical characterization of a sample further comprises performing a second subsequent screening of the subset of sample regions from the pattern of sample regions using a correlative characterization technique. A“correlative characterization technique” refers to a technique which is capable of generating chemical or structural information from a sample. The chemical or structural information obtained is then correlated to the optical information obtained as described above (e.g., from the initial screening step, the first screening step, or both). Correlative characterization techniques include Raman spectroscopy, for extracting information about the chemical entities within a sample, as well as atomic force microscopy (AFM), scanning electron microscopy (SEM), transmission electron microscopy (TEM), and near-field scanning optical microscopy (NSOM), each of which may be used to extract structural information from the sample. The result of the second subsequent screening step is a third set of data comprising

chemical/structural values and associated sample region values. The chemical/structural values may be raw data or processed data. The sample region values are the set of coordinates that identify the physical location of the sample region as described above. Depending upon the resolution, chemical/structural values from sample subregions may also be collected. Again, depending upon the number of correlative characterization experiments and types of processing conducted, multiple third sets of data may be obtained from the second subsequent screening step.

[0027] As noted above, the chemical/structural information in a third set of data may be correlated to optical information (e.g., from a second set of data obtained in the first screening step) in order to generate“maps” of the optical information. By way of illustration, SEM may be carried out on each sample region of the subset of sample regions to generate a structural image of each sample region in the subset. (See“High-Resolution Scanning

Probe” and“Morphology Analysis” panels in FIG. 4.) Since each sample region value of the SEM image also has associated with it an optical value obtained from a second set of data, those optical values can be correlated to the SEM image, thereby providing maps of each optical value. (See the“Fermi Level Map,”“Carrier Relaxation Rate,”“Carrier Density Map,” and“Carrier Temperature” panels in FIG. 4.)

[0028] In addition, the chemical/structural information in a third set of data may be used to group certain sample regions together. This is useful when implementing statistical analysis as further described below. By way of illustration, SEM images of sample regions may be used to identify all grain boundaries in the subset of sample regions and all grain interiors in the subset of sample regions.

[0029] The second subsequent screening step may be carried out on an appropriate microscope apparatus comprising any of the optical components, electrical components, electro-mechanical components, light sources, detector, controller, etc., typically used to carry out the desired technique, e.g., Raman, AFM, SEM, TEM, NSOM, etc.

[0030] The collection of data sets described above composed of

optical/chemical/structural values and associated sample region values may be referred to as “multi-modal data.” In a fourth step of the method for the multi-modal optical

characterization of a sample, statistical analysis may be carried out on the multi-modal data. The type of statistical analysis is not particularly limited and illustrative types are described in the Example and below.

[0031] As noted above, grouping of sample regions, as facilitated by the second subsequent screening step, is useful during statistical analysis. This is illustrated via FIG. 5, showing the correlation between the amplitude of a sub-bandgap state and initial carrier temperature (two optical values) for grain boundaries and grain interiors (two groupings of sample regions; in this case, the sample regions are distinct, individual spatial locations of a single perovskite particle) at two different times.

[0032] Also provided are systems for carrying out the present methods. A block diagram of an illustrative system 100 is shown in FIG. 1. The system comprises an initial screening microscope apparatus 102 (e.g., a wide-field microscope apparatus) configured to carry out the initial screening step as described above; a first screening microscope apparatus 104 (e.g., a confocal microscope apparatus) configured to carry out the first subsequent screening step as described above and operably coupled to apparatus 102; and at least one second screening microscope apparatus 106 configured to carry out a correlative characterization technique as described above and operably coupled to the apparatuses 102, 104. Each of the microscope apparatuses may be distinct apparatuses comprising distinct components from another microscope apparatus but operably coupled through one or more shared controllers 108. However, some microscopes may be shared, e.g., a microscope for carrying out Raman spectroscopy in the second subsequent screening step could be the same microscope apparatus used in the initial screening step or the first subsequent screening step. Patterned samples may be transferred between individual microscopes but the individual microscopes can illuminate the exact same regions due to the use of sample region registration via patterned substrates as described above.

[0033] Regarding controller(s) 108 of the system 100, such controllers 108 may be configured to control one or more components of the system 100. Such controllers 108 may also be configured to process data generated from one or more components of the system 100 (e.g., perform calculations on raw detector data in order to provide desired optical information; make comparisons between data to compare generated optical values to predetermined optical values; apply statistical analysis techniques to data; and combinations thereof). The controllers 108 may be integrated into the system 100 as part of a single device or their functionality may be distributed across one or more devices that are connected to other system components directly or through a network that may be wired or wireless. A database (not shown), a data repository for the system 100, may also be included and operably coupled to the controller(s) 108.

[0034] An illustrative controller 108 which may be included in the system 100 is shown in FIG. 2. The controller 108 may include an input interface 200, an output interface 202, a communication interface 204, a computer-readable medium 206, a processor 208, and an application 210. The controller 108 may be a computer of any form factor including an electrical circuit board.

[0035] The input interface 200 provides an interface for receiving information into the controller 108. Input interface 200 may interface with various input technologies including, e.g., a keyboard, a display, a mouse, a keypad, etc. to allow a user to enter information into the controller 108 or to make selections presented in a user interface displayed on the display. Input interface 200 further may provide the electrical connections that provide connectivity between the controller 108 and other components of the system 100. [0036] The output interface 202 provides an interface for outputting information from the controller 108. For example, output interface 202 may interface with various output technologies including, e.g., the display or a printer for outputting information for review by the user. Output interface 202 may further provide an interface for outputting information to other components of the system 100.

[0037] The communication interface 204 provides an interface for receiving and transmitting data between devices using various protocols, transmission technologies, and media. Communication interface 204 may support communication using various transmission media that may be wired or wireless. Data and messages may be transferred between the controller 108, a database, other components of the system 100 and/or other external devices using communication interface 204.

[0038] The computer-readable medium 206 is an electronic holding place or storage for information so that the information can be accessed by the processor 208 of the controller 108. Computer-readable medium 206 can include any type of random -access memory (RAM), any type of read only memory (ROM), any type of flash memory, etc. such as magnetic storage devices, optical disks, smart cards, flash memory devices, etc.

[0039] The processor 208 executes instructions. The instructions may be carried out by a special purpose computer, logic circuits, or hardware circuits. Thus, the processor 208 may be implemented in hardware, firmware, or any combination of these methods and/or in combination with software. The term "execution" is the process of running an application or the carrying out of the operation called for by an instruction. The instructions may be written using one or more programming language, scripting language, assembly language, etc.

Processor 208 executes an instruction, meaning that it performs/controls the operations called for by that instruction. Processor 208 operably couples with the input interface 200, with the output interface 202, with the computer-readable medium 206, and with the communication interface 204 to receive, to send, and to process information. Processor 208 may retrieve a set of instructions from a permanent memory device and copy the instructions in an executable form to a temporary memory device that is generally some form of RAM.

[0040] The application 210 performs operations associated with controlling other components of the system 100. Some of these operations may include receiving data (e.g., raw data from detectors(s)), and processing data as described above. The result of such operations may be outputting information to a user or outputting to components of the system 100. Some or all of the operations described in the present disclosure may be controlled by instructions embodied in the application. The operations may be implemented using hardware, firmware, software, or any combination of these methods. In embodiments, the application 210 is implemented in software (comprised of computer-readable and/or computer-executable instructions) stored in the computer-readable medium 206 and accessible by the processor 208 for execution of the instructions that embody the operations of application 210. The application 210 may be written using one or more programming languages, assembly languages, scripting languages, etc.

[0041] It is noted that devices including the processor 208, the computer-readable medium 206 operably coupled to the processor 208, the computer-readable medium 206 having computer-readable instructions stored thereon that, when executed by the processor 208, cause the device to perform any of the operations described above (or various combinations thereof) are encompassed by the disclosure. The computer-readable medium 206 is similarly encompassed.

[0042] With reference to FIG. 3, exemplary operations associated with the application 210 are described. Additional, fewer, or different operations may be performed depending upon the embodiment. The order of presentation of the operations of FIG. 3 is not intended to be limiting. Thus, although some of the operational flows are presented in sequence, the various operations may be performed in various repetitions, concurrently, and/or in other orders than those that are illustrated.

[0043] In an operation 300, multi-modal data is received from the at least the first screening microscope apparatus 104 and the second screening microscope apparatus 106. Thus, the multi-modal data comprises at least the second and third sets of data described above.

[0044] In an operation 302, a statistical analysis is applied to the multi-modal data. The statistical analysis may be conducted to determine whether a correlation exists between a pair of different optical values (e.g., amplitude of a sub-bandgap state and initial carrier temperature) from an associated sample region value (or grouping of associated sample region values, e.g., sample region values at grain boundaries or sample region values within grain interiors). The statistical analysis may also be conducted to determine whether the correlation is statistically significant. By way of illustration (and depending upon the statistical analysis applied), the determination may involve, e.g., calculating a correlation coefficient and an associated p-value for the data being analyzed. These determinations can be made at various different sample region values (or groupings thereof) and for various different pairs of optical values. The results of the statistical analysis, e.g., correlation coefficients, p-values, etc. may be referred to as correlation data. If sufficient data volume exists, the data may be modelled using deep learning approaches such as convolutional neural networks to make predictive models of the data. These models may then be used to predict the properties of materials that have yet to be measured or to suggest new materials using active learning methods.

[0045] In an operation 303, the correlation data may be output, e.g., to a display, or output for further processing.

[0046] Other operations (not shown in FIG. 3) could include a determination of the subset of sample regions to be interrogated in the first subsequent screening step. This may be accomplished by comparing optical values from the first set of data from the initial screening step to predetermined optical values as described above.

[0047] Still other operations (not shown in FIG. 3) could include initiating one or more of the screening steps, e.g., causing the initial screening microscope apparatus 102 to illuminate the pattern of sample regions with the patterned light beam; causing the first screening microscope apparatus 104 to illuminate the subset of sample regions with the light beam under confocal point illumination; and/or causing the second screening microscope apparatus 106 to conduct the correlative characterization technique on the subset of sample regions.

[0048] Details of an illustrative system for multi-modal optical characterization of sample regions similar to that shown in FIGs. 1-3 are now described. The characterization of an array of samples proceeds by a multi-stage process in which in a first step, an n x m array of samples inside a well plate are simultaneously probed by a diffractive optic (D.O.) microscope for initial screening; in a second step, promising candidates from the initial screening are probed by a high-resolution (H.R.) microscope which can simultaneously provide spectrally-resolved transient-absorption (SR-TA), transient-reflection (SR-TR), PL, or TRPL on the same sample at a resolution of ~400 nm for linear and ~200 nm for nonlinear optical measurements (note, resolution depends on smallest detectable change in signal as well as the spot size); and in a third step, samples from the second step are removed from the microscope and subjected to ex situ correlative characterization using polymer registers (e.g., from using a patterned substrate) by atomic force microscopy (AFM), transmission electron microscopy (TEM), scanning electron microscopy (SEM), or Raman microscopy. These steps are described in more detail below.

[0049] Samples may be prepared by drop casting solutions in an n x m grid on a glass cover slide. In the first step, a wide-field microscope may be used by combining pulse shaping and diffractive optics. The input to the pulse shaper may be an ultrabroadband pulse from a non-colbnear optical parametric amplifier (NOPA), compressed to < 6 fs using custom chirp mirrors and a deformable mirror (DM), which corrects for higher-order dispersion. A portion (blue side) of the compressed pulse may be used as the pump source, while the remainder (red side) may serve as the probe. Separation of these two components may be done either inside the pulse shaper or through dichroic optical elements. The pulse shaper can control the timing between the pump and probe to better than 0.1 fs by applying a phase mask along a spatial light modulator (SLM) in a 4f geometry. A custom DO may be placed at the focus of the shaped pulse, generating an n x m beam pattern (i.e. 8 x 8 or 64 total beamlets), whose intensity may be characterized prior to the experiment. A lens may be used to collimate the beam pattern, and a reflective objective (or parabolic mirror) may be used to focus each beamlet to a unique position at the sample array. The spot size of each beam is -10 microns, while the pitch is ~l mm. Other arrangements can be made depending on the details of the sample pattern, focal lengths of lenses/mirrors, and specifications of the DO. After blocking the pump light, the probe beams may be focused using a micro-lens array into an n x m array of fiber optic cables. The cables may terminate in a linear pattern on the slit of a high-resolution imaging spectrometer attached to an sCMOS detector. In this way, each of the 64 probe beams is dispersed onto a row on the 2D array detector. The camera can operate in cropped mode, using ~4 rows per fiber optic line, so that -256 rows are sufficient to capture the entire data set. The camera can operate at -lkHz, so that chopping the pump beam at 500 Hz, allows for differential absorption or reflection measurements to be acquired. A full data set, including sampling at -20-50 pump-probe delays will take less than 10 minutes using this type of parallel acquisition scheme. The data measured using the DO microscope can include: absorption and emission spectra, bleach recovery time, hot-carrier relaxation rate, bandgap energy, Fermi level, exciton/free -carrier ratios, sub-bandgap energy, sub-bandgap rise and decay times, carrier (electron/hole/exciton/polaron) temperature and density, and exciton binding energy.

[0050] In the second step, candidates from the first step may be analyzed in further detail using a second high-resolution (HR) microscope with enhanced capabilities. The HR microscope may use a confocal geometry so that much higher spatial resolution may be achieved, -200 nm. The sensitivity of this system is also much higher than the DO microscope because the detector can operate at up to 10 kHz, meaning faster chopping of the pump beam and better signal isolation from laboratory noise sources. Most importantly, the HR microscope provides a means to peer underneath the inhomogeneously broadened spectra caused by sample heterogeneity including chemical and non-chemical effects, such as the presence of unreacted species, poorly passivated surface sites, edge states, grain boundaries, local strain, surface interactions, etc. Such heterogeneities can mask minority species. This microscope can also incorporate shaped pulses in order to do two-dimensional spectroscopy so that further reduction in line widths are achievable owing to the photon echo effect. That is, even within the spot size of the microscope, inhomogeneous broadening is separated from homogeneous broadening. Performing these experiments on the microscope offers the advantage of greatly reduced scattering which oftentimes plagues ensemble 2D methods. The signal may be isolated by rapid phase cycling using an acoustic optic modulator operating at 100 kHz and a synchronized linear array detector.

[0051] Since the properties of many materials are dictated by many factors including chemical and non-chemical properties, next, additional characterization methods are carried out on the exact same regions of space as analyzed by the optical measurements in the first two steps. This may be accomplished by using a correlative characterization technique enabled by the sample registration achieved via substrate patterning. For example, samples exhibiting the most desirable properties in the first and/or second step are subjected to correlative characterization techniques such as Raman microscopy, e.g., to assess high- frequency modes of organic linkers and cations. This provides information about the chemical composition of the sample, as well as unreacted species and the surface passivation. Low-frequency modes may be analyzed by 2D spectroscopy. In addition, AFM may be used to characterize the topography, while SEM and TEM may be used to characterize the surface, edge, and grain boundaries at high-resolution (<10 nm). More advanced correlative characterization techniques may be used, such as near-field scanning optical microscopy (NSOM) if higher spatial resolution is required.

[0052] Thorough optical characterization of materials has traditionally been a rate limiting step in the sense that it can take days or weeks to acquire the necessary spectroscopic information and to interpret the measurements in terms of parameters that summarize the key performance metrics of a particular material, such as its photoluminescent quantum yield or its electronic coherence lifetime. Acquisition and data reduction using the present methods and systems are explained in the following table.

[0053] In a fourth step, once the multi-modal data has been acquired, it is analyzed by a statistical technique. A variety of approaches may be used such as descriptive statistics (e.g. measures of central tendency, normal or skewed distributions) or inferential statistics (e.g. p- value), parametric and non-parametric tests (e.g. t-test, analysis of variance (ANOVA), sign test, Spearman rank order, etc.). Nonlinear regression analysis reveals connections between different material descriptors which may be coupled by some underlying physical principle or mechanism. For instance, the presence of a defect may alter the initial carrier temperature through screening effects, as well as the rate of carrier relaxation through disruption of phonons or as an additional energy sink. (See, e.g., Example, below.) Therefore, the initial carrier temperature and carrier relaxation rates may be correlated, which would be revealed in a statistical analysis of thousands of different spatial locations.

[0054] As further described below, the Example demonstrates how an illustrative system and method may be used to measure the following from different spatial locations within a single perovskite particle: (i) mapping carrier density, (ii) carrier temperature, (iii) thermalization rates, (iv) single-particle statistics, and (v) dynamic bandgap shift, (vi) Fermi level, and (vii) sub-bandgap state energies. Next, these measurements are correlated to independent, but referenced scanning probe imaging (viii) of the perovskite particle. This generation of multi-modal data is illustrated in FIG. 4. The left-most image shows an SEM image of a perovskite particle. Spectrally-resolved TA at each point is measured as shown in the two middle panels. The SEM image identifies structural features within single perovskite particles, such as grain boundary (GB) area and grain interior (GI) area. As shown in the six right-most images, Fermi level, carrier relaxation, carrier density, and initial carrier temperature are mapped by analyzing the TA signal from multiple locations in the particle.

Sub-bandgap state is identified (arrow in top middle image) and may be analyzed separately. In the six right-most images, the GBs between crystallites are marked using dashed curves (the dense ones within the inner core of the particle are not marked). Next, statistical analysis is applied to the multi-modal data. This is illustrated in FIG. 5, showing a two-dimensional scatter plot of the amplitude of a sub-bandgap state (y-axis) and the initial carrier temperature (x-axis) at grain boundaries (GB) and grain interiors (GI) at two different times (0.1 ps and 2.0 ps). Triangles show the positions of median values. From FIG. 5, one can see that the correlation between the descriptors of the sub-bandgap state and the initial carrier temperature is different among the two different populations (GB, GI), especially at later times (2.0 ps versus 0.1 ps). This indicates that the role of the sub-bandgap state is different in each spatial region. This type of information is unique to the present system and methods. The statistical analysis is further described in the Example and is also illustrated in FIGs. 7A- D and 9B.

[0055] Samples which may be characterized by the present methods and systems are not particularly limited. Illustrative samples include semiconductors, including those for use in photovoltaic or light-emitting diode applications or those for use as room-temperature qubits. The present methods and systems, which may be used to generate and process multi-modal data from an extremely large number of sample regions (e.g., thousands), address the slow speed and cost of analyzing molecules/materials using conventional approaches. Thus, the present methods and system provide a faster, cheaper and more systematic approach to the discovery of new molecules and materials.

EXAMPLE

Introduction

[0056] Metal halide perovskites have received intense attention as active layers in solar cells for their high power conversion efficiency and the relatively low technical barrier in fabrication. Contributing to their idiosyncratic carrier behavior, like low recombination rates, low cooling rates, and long diffusion lengths, are the unique band structures perovskites may have that arise from different film morphologies. The impact of the local nanostructure, such as at grain boundaries (GBs), on carrier properties has been widely studied using other microscopic methods, mainly photoluminescence (PL) microscopy and atomic force microscopy (AFM), along with ensemble pump-probe spectroscopy by comparing samples with different bulk morphologies. However, the nature of the GBs and the associated defect states remain unclear. Many previous studies either performed ensemble measurements averaging over tens of microns in films, thus convolving signals from many individual grains, or studied other properties like the local PL lifetime at the bandgap to indirectly probe the dynamics at the defect state. Until now, there has not been a direct probe of the transition energy and population dynamics in a locally precise way so that only sections of films (e.g. interior, near edge, boundary, etc.) or single grains are probed, while simultaneously measuring hot carrier dynamics. Further, comparison of carrier properties between different morphologies has not, until now, been supported by large-data statistics.

[0057] This Example presents the first realization of scanning electron microscopy (SEM) correlated spectrally-resolved transient absorption microscopy (TAM) on

polycrystalline perovskites, which directly maps the formation of a sub-bandgap state at GBs. The dynamics of these states are investigated together with their influence on the local carrier behaviors, suggesting the electrically-charged nature and the potential role of GBs as centers for fast carrier cooling. Furthermore, the microscope setup and the specially fabricated films allowed for the capability of studying hundreds of data points in a single scan, with multiple film morphologies included. A thorough statistical analysis performed on such a big dataset enables the drawing of a reliable connection between the formation of the sub-bandgap state, the doping effect at GBs, and the accelerated carrier cooling process. The findings described here indicate that, contrary to the conventional wisdom, the shallow defect state in perovskite grain boundaries is benign and the carrier cooling dynamics are strongly affected by additional energy acceptors formed by these states, which speeds up the process and leaves limited time for hot carrier extraction. The outcome of this effect, which is an accumulation of carriers at the band edge, may play a beneficial role in conventional photovoltaics, as no evidence of a higher recombination rate is observed at grain boundaries.

Materials and Methods

Transient absorption microscopy (TAM)

[0058] The block diagram of the home-built TAM is shown in FIG. 8. The fundamental l030-nm beam was generated from an Yb:KGW amplifier system (Light Conversion, PHAROS) operating at 200 kHz with a pulse duration of 190 fs. The beam was split into pump and probe arms. The pump was focused into a beta barium borate (BBO) crystal to produce second harmonic generation light centered at 515 nm (2.41 eV) with a maximum average power of 15 m\¥. The pump was then delayed relative to the probe with a high resolution motorized linear stage (Aerotech). The white-light probe (1.50-1.87 eV) was generated by focusing the fundamental beam into an yttrium aluminum garnet (Y AG) crystal and then compressed with a pair of chirped mirrors (DCM, LayerTec). Both beams were recombined and focused collinearly onto the sample with a 74x reflective objective (NA 0.65, Beck). The beam spot size of 540 nm was determined by measuring the power passing through a pinhole and then calculated using a Gaussian beam model. The sample was taped on a piezo-driven XYZ stage with a resolution of 0.2 pm (PiezosystemJena, Newport XPS). The sample box was purged with nitrogen gas to eliminate sample damage caused by laser heating. The signal was collected using a lOOx refractive objective (NA 0.70, Leica), spatially filtered through a 300-pm pinhole, and then spectrally dispersed in a spectrometer (Horiba, IHR-320). The TA signal was detected using a high-speed CCD camera (Andor,

Ixon Ultra 897).

[0059] Linear absorption of probe and photoluminescence (PL) maps were measured within the same setup.

Correlated scanning electron microscopy (SEM)

[0060] The CEENEEPbL· perovskite film studied here was deposited on a marked glass coverslip (Coming, No 1.5 cover glass). More details of the perovskite synthesis process could be found in Nah, S. et al. Nat. Photonics 11, 285-288, doi: l0T038/nphoton.20l7.36 (2017). Optical images (not shown) were first taken along with the transient absorption (TA) measurement, and further calibrated using a probe transmission map (not shown). The SEM images of same regions (not shown) were later taken on Hitachi S-4800 SEM (NUANCE, Northwestern University) at 5 kV, with 6 nm Au/Pd coating (Denton Desk IV, NUANCE, Northwestern University). More high-resolution SEM images of perovskite particles on the films of same chemical composition are also included, showing the two different

morphologies within one particle (not shown).

Fitting methods: Band-filling model and single exponential decay

[0061] The model was constructed based on parabolic -band hypothesis and the band- filling effect. The change in absorption is given by

where A (£; E_g) and A (£; E_g— E_g) represents the absorption before and after the shift of the bandgap. The Fermi distribution function represents the carrier distribution characterized by the quasi-Fermi energy, Ef, and the quasi-temperature, T. This equation only holds true when the effective masses of electrons and holes are similar, as is the case in CFFNFFPbL·.

[0062] The fit is more weighted within 1 eV above the bandgap (1.59-1.73 eV), where the parabolic model should hold well compared to the high-energy range. R-square values at different times were mapped (not shown). Only points with an r-square value higher than 0.995 were later used. The fitting algorithm used was the“lsqnonlin” function in MATLAB. The“fit” function was also tried and gave the same results. Unweighted fit was also performed, and the sub-bandgap state was still present.

[0063] A single-exponential fit of each temperature -time curve gave the cooling rate.

Results with a confidence level over 0.9 were later used. A single-exponential fit of sub- bandgap state intensity as a function of time was performed to extract the population lifetime.

Initial carrier density

[0064] The overall initial carrier density was estimated to be 8.38 ^c 10¹⁹ cm ³ using the following equation:

where j is the input pump power density calculated using the input power and beam size, R and T are the reflectivity and transmittance at the pump wavelength, /if is the energy of a pump photon, and a is the absorption coefficient. The values of R. T, and a in Yang, Y. et al, Nat. Photonics 10, 53-59, doi: K).l038/nphoton.20l5.2l3 (2015) were used for the estimation.

[0065] Due to the difficulty of collecting overall reflectance and transmission within the microscope setup, to get the carrier density map, the intial carrier density at every point given was calculated by

[0066] A common way to deal with the integration in the equation above is to approximate the Fermi-Dirac distribution for energies 3 hT away from Ef by a Maxwell- Boltzmann distribution. However, this condition does not hold well in perovskite systems, but it is still wildly used. Herein, the Joyce-Dixon approximation was used, which provided an analytical solution to the Fermi integration. The parameters extracted at 100 fs from the band-filling model were used to calculate the initial carrier density. Note that the static bandgap value was used instead of the shifted one to represent E_c in the equation to better simulate the initial condition. It causes a possible loss in the density as the population below the static bandgap is not counted, but the values are still within the same order of magnitude (10¹⁹ cm ³) as the value calculated using pump power.

[0067] Also, the carrier density is insufficient for complete band-filling. The power used here corresponds to 0.6 pW in that figure, which falls into the regime dominated by bimolecular recombination, with no obvious high-order process present.

Results and Discussion

[0068] Thin films of 8% HI-pretreated methylammonium lead iodide (CFFNFFPbE) deposited on marked glass coverslips by spin-coating perovskite precursor solution were studied. It is worthwhile to mention that the only difference between fabricating these particles and continuous films made for devices is the concentration of HI additive in the first step of synthesis, which is meant for grain size control only and does not change the chemical composition (proven by energy-dispersive x-ray spectroscopy, EDS, data not shown) and does not introduce any impurity, as G is part of the CFFNFFPbE lattice. The films fabricated this way form large poly crystalline particles with two distinct morphologies: small crystallites (<0. l pm²) in the center and large crystallites (>l pm²) in the outer ring (FIG.

6A). The grain boundaries between crystallites are marked using dashed curves (the dense ones in inner circle not marked). While vastly different particle morphologies form depending on film coverage and fabrication conditions, these small polycrystalline particles provide a unique platform to study many morphologies and perform statistical analysis (L_''=806) within a small region of interest. Further, the transient absorption (TA) mapping and direct comparison between points may be performed within a single particle to eliminate the possible variation in chemical composition.

[0069] The TA spectra across the whole particle within the solid curve (FIG. 1A) are measured with a single-wavelength pump at 2.41 eV and a broadband white-light probe from 1.50 to 1.87 eV, which covers spectral ranges both below and above the bandgap at 1.63 eV. The TA spectra of two representative points picked from the grain interior (GI) and GB areas are displayed in FIG. 1B-1E. The conspicuous difference between the two spectra are the sub-peak below the bandgap at the GB (FIG. 1D) which shows up as a growing feature in the time-resolved spectra (FIG. 1E). To extract this sub-bandgap peak, a band-filling model was constructed (see above), which took into account the three prominent features and the related photophysics observed in previous ensemble measurements and the GI TA spectra here: (1) the growing positive ground state bleaching (GSB) representing the increasing amount of band edge carriers; (2) the high-energy tail of the GSB representing the population and cooling of hot carriers; (3) the negative photo-induced absorption (PIA) below the bandgap representing the renormalized bandgap due to the presence of photo-excited carriers. All three features are considered in the band-filling model and later subtracted from the spectra to extract the sub-bandgap state. The sub-bandgap peak centered at ~l .57 eV, and the dynamics were then extracted (FIG. 1F-G). Notably, the high-energy tail is much broader in GI (FIG. 1C) compared to that in GB (FIG. 1E) at 2 ps, indicating a higher quasi-temperature of the carriers in GI. Different fit algorithms were used and all gave similar results.

[0070] The band-filling model was then applied across the whole particle to extract the sub-bandgap state and basic properties embedded in the model, describing the distribution of carriers above the bandgap. Correlated SEM images were obtained and maps of integrated sub-bandgap state intensity, quasi-Fermi energy, and quasi-temperature at representative time points were generated (not shown, but see FIG. 4). The differences between GB and GI are prominent. In GB areas, a stronger intensity of the sub-bandgap state was observed, accompanied by a higher quasi-Fermi energy. The increased population at the sub-bandgap state along with the elevated quasi-Fermi energy suggest the GB is mainly doped by one kind of carrier. Otherwise, if both electrons and holes are trapped or simply recombine, the quasi- Fermi level should decrease accordingly. The relatively uniform carrier density across the whole particle also indicates that the grain boundaries were not acting as fast recombination centers, as no notable loss of carriers was seen, which agrees well with previous reports on the benign nature of GBs in carrier recombination in perovskites. The quasi-temperature maps suggest that the carrier cooling is much faster in GB areas, which is more clearly shown in the carrier cooling rate map. A temporary conclusion was drawn that the GBs are related to the formation of a sub-bandgap state, doping effects, and faster cooling rates. Though sub- bandgap absorption is also observed in GIs close to the edge of the particle, similar changes in the Fermi-energy level and carrier cooling dynamics were not observed. The GB states can be distinguished from the states formed at particle edges by mapping the depth of the state into bandgap. [0071] To beter explore the correlation between the sub-bandgap state and the local carrier cooling dynamics, two-dimensional scater plots and histograms were analyzed (FIGS. 7A-7D show The particle was divided into two separate regions (GB (circles) and GI (x)), and statistical analysis was performed on each, with histograms projected from each axis and median values marked by triangles. The first thing to note is that the sub-bandgap state intensity is obviously higher in GBs after 0.8 ps, both shown by the histograms and the median values, which aligns with the phenomenon described above, that the sub-bandgap state tends to form in GBs. The temperature of GBs starts from a slightly lower value compared to the GIs but cools down more rapidly ( Mtnu,, - 2.9 ps ¹) to a narrower distribution centered at -300 K, with a small sub-peak at -425 K. The GI region, on the other hand, cools down at a slower rate (1 /XM* - 2.6 ps ¹) and forms a similar two-peak

distribution, with stronger intensity at 425 K. A strong positive correlation between the sub- bandgap peak intensity and temperature cooling rate was observed (p<0.05). It is also important to point out again that the sub-bandgap state is only accelerating carrier cooling, not carrier recombination, which are two different steps in photo voltaics. The high temperatures in GIs at 2.0 ps and the slow cooling rates suggest a latice heat-up effect exerted by the hot carriers. It is also expected that longitudinal optical (LO)-phonon scatering is more prominent in GIs where complete or high-quality latices tend to form. For GBs, the temperature mostly relaxes rapidly to the surroundings near 300 K, denoting a negligible heating up of the latice. Also, by applying the LO-phonon emission model to the temperature cooling curve, the time constant x_ave could be extracted, characterizing the energy loss via phonon-emission, which represents the strength of the phonon-botleneck effect, and an obvious increase in the GB region was not seen. Therefore, it is hypothesized that the GIs may be predominated by the widely-reported phonon-botleneck effect in perovskites, while in GBs, there must be additional cooling pathways besides LO-phonon scatering. This implies that other species may form at GBs to accept the extra energy from hot carriers and relieve the strain on phonons as the major energy acceptors. One of the new species is most likely to be the charged sub-bandgap state, the formation of which is reflected in the increased quasi-Fermi energy level. As is shown in the histograms of FIG. 7B(i), the initial Fermi energy across the whole particle was uniformly distributed with no clear favor for GI or GB. A higher Fermi energy was observed in GBs only later, most clearly seen after 0.8 ps, and found to be positively correlated to the sub-bandgap state intensity, suggesting that the doping effect of one kind of carrier is prominent in GBs where the sub-bandgap state shows up. FIG. 7D then further links this charged sub-bandgap state to the carrier cooling process, where a positive correlation between the averaged Fermi level and the carrier cooling rate was observed.

[0072] To summarize, positive correlations were observed among the formation of sub- bandgap, increased cooling rates, and elevated Fermi energy. These three linked effects are all more pronounced in GB regions. Clearly, there is some overlap in the data points or histograms from GB and GI, but this may be attributed to the sub-ensemble averaging effect within the 540-nm spot size, which means small contributions from GIs in GB points and vice versa are inevitable. However, the overall statistics help facilitate a more reliable conclusion compared to other studies on GBs, the dataset size of which are usually limited to single points or, at most, a single line scan over the GB.

[0073] The same data analysis was performed on another particle on a film prepared using the same method. Similar trends were observed. The grain boundaries (GBs) show stronger absorption in the sub-bandgap state, elevated Fermi-level, and faster temperature relaxation. Sub-bandgap absorption was also observed at particle edges, but was later well separated from the GB sub-bandgap absorption using the depth of sub-bandgap state as well.

[0074] A clear link between the carrier temperature and the distribution of the sub- bandgap state was observed from the scatter plots and the histograms (FIG. 9A). The positive correlation between sub-bandgap state intensity and temperature cooling rate was even stronger here (p = 0.0010) (FIG. 9B). compared to the particle analyzed above (particle 1). The bi-peak distribution was not as clear, though. This is because this particle is smaller and there are more grain boundaries per unit area, which also suggests a relatively faster cooling rate (centered at 2.0 and 2.1 ps ¹) compared to particle 1 (centered at 2.6 and 2.9 ps ¹).

[0075] Finally, the probable nature of the sub-bandgap state at GBs and its role in carrier cooling process are discussed. The measurements reveal that the sub-bandgap state (1) shows up ~60 meV into the static bandgap; (2) is a transient feature growing after initial excitation; (3) is relevant to the fast temperature relaxation of carriers; and (4) likely traps only one kind of carrier. It is not an exciton, as the depth of the states falls out of the range of newly- reported binding energy, which would form immediately after excitation instead of growing on the order of ps. It is suggested that the origin of this state is from I-I bond at GBs, which is a shallow but electrically-charged defect state (illustration not shown). Due to first- principles calculations done by Yin et.ak, a shallow defect state (within tens of meV) above the valence band minimum (VBM) tend to form from the I-I bond in certain types of grain boundaries in CH3NH3PbL·. (Yin, W.-J. et al., Adv. Electron. Mater. 1, 1500044, doi: 10. l002/aelm.201500044 (2015).) The mechanism of accelerated carrier cooling could be that negatively charged states induce a built-in potential which quickly localizes the holes. A dynamic equilibrium then formed between the trapped holes and the hot carriers, during which the holes lost energy when they got trapped into this sub-bandgap state; they regained energy absorbed from hot carriers and came back to the valence band. At early times, the sub- bandgap states offered another pathway in GBs for carrier cooling by quickly exchanging energy with the hot carriers, while at later times, most carriers cooled down to ambient temperature and the holes started piling up in this sub-bandgap state, giving rise to a stronger bleach signal.

[0076] A more sophisticated kinetic model was tested to study the detailed kinetic processes involved in cooling, but because of the limited time resolution (0.1 ps) and relatively weak signal from the sub-bandgap state, it failed to give more information than a simple exponential fit. Further study with finer time resolution and stronger pump intensity or increased data collecting efficiency may be done to justify the proposed mechanisms.

However, the statistical analysis performed in this Example strongly supports the hypothesis that a charged sub-bandgap state is formed in GBs of perovskites and offers a faster carrier cooling pathway. These charged stats will harm the hot carrier extraction process by assisting the energy dissipation of hot carriers. The higher cooling rate will lead to an accumulation of carriers at the band edge, as the rates of reactions generating band edge carriers are increased; however, a significant decrease in carrier density coming from carrier recombination was not seen. Along with the fact that carrier separation is easier at GBs, GBs might still benefit the conventional solar cells which harvest carriers at the band edge.

[0077] Conclusion

[0078] In conclusion, SEM-correlated transient absorption microscopy was used to identify a transient sub-bandgap state located only in the grain boundaries of CFFNHiPbli perovskites. The findings described herein were based on statistical analysis of many hundreds of spatial locations and enhance the fundamental understanding of the GBs, which behave like shallow charged states accelerating carrier cooling. They also directly impact the fabrication of perovskite active layers in hot carrier devices, as they provide direct evidence that sub-bandgap states in GBs offer a fast temperature relaxation pathway, which harms the extraction of hot carriers. Contrary to the common belief that the shallow defect states in GBs are benign, they actually cast a strong impact on the carrier cooling dynamics and should be eliminated in hot carrier applications. The impact of GBs in conventional solar cells seems to be benign, as no enhanced carrier recombination is observed and the sub-bandgap state probably helps with carrier separation.

[0079] Data not shown and additional supplementary information may be found in U.S. Application No. 62/745,636, which is hereby incorporated by reference in its entirety. The word "illustrative" is used herein to mean serving as an example, instance, or illustration.

Any aspect or design described herein as "illustrative" is not necessarily to be construed as preferred or advantageous over other aspects or designs. Further, for the purposes of this disclosure and unless otherwise specified, "a" or "an" means "one or more.”

[0080] The foregoing description of illustrative embodiments of the disclosure has been presented for purposes of illustration and of description. It is not intended to be exhaustive or to limit the disclosure to the precise form disclosed, and modifications and variations are possible in light of the above teachings or may be acquired from practice of the disclosure. The embodiments were chosen and described in order to explain the principles of the disclosure and as practical applications of the disclosure to enable one skilled in the art to utilize the disclosure in various embodiments and with various modifications as suited to the particular use contemplated. It is intended that the scope of the disclosure be defined by the claims appended hereto and their equivalents.

Claims

WHAT IS CLAIMED IS:

1. A system for multi-modal optical characterization of sample regions, the system comprising:

a wide-field microscope apparatus configured to illuminate a pattern of sample regions with a patterned light beam to induce a patterned optical signal and to detect the patterned optical signal to generate a first set of data comprising optical values and associated sample region values;

a confocal microscope apparatus configured to illuminate a subset of sample regions of the pattern of sample regions with a light beam under confocal point illumination to induce optical signals and to detect the optical signals to generate a second set of data comprising optical values and associated sample region values;

a microscope apparatus configured to conduct a correlative characterization technique on the subset of sample regions to generate a third set of data comprising chemical or structural values and associated sample region values; and

a controller comprising a processor; and a non-transitory computer-readable medium operably coupled to the processor, the computer-readable medium comprising instructions that, when executed by the processor, cause the controller to

receive multi-modal data comprising at least the second and third sets of data, perform a statistical analysis on the multi-modal data to generate correlation data; and

output the correlation data.

2. The system of claim 1, wherein the correlative characterization technique is Raman microscopy, atomic force microscopy, scanning electron microscopy, transmission electron microscopy, near-field scanning optical microscopy.

3. The system of claim 1, wherein the non-transitory computer-readable medium further comprises instructions that, when executed by the processor, cause the controller to determine the subset of sample regions by comparing optical values of the first set of data to predetermined optical values.

4. The system of claim 1, wherein the non-transitory computer-readable medium further comprises instructions that, when executed by the processor, cause the system to illuminate the pattern of sample regions with the patterned light beam, illuminate the subset of sample regions with the light beam under confocal point illumination, conduct the correlative characterization technique on the subset of sample regions, or a combination thereof.

5. A method for multi-modal optical characterization of sample regions using the system of claim 1, the method comprising:

performing an initial screening by illuminating the pattern of sample regions with the patterned light beam via the wide-field microscope apparatus;

performing a first subsequent screening by illuminating the subset of sample regions with the light beam under confocal illumination via the confocal microscope apparatus;

performing a second subsequent screening by conducting the correlative

characterization technique on the subset of sample regions via the microscope apparatus; and performing the statistical analysis on the multi-modal data to generate the correlation data via the controller.

6. A system for multi-modal optical characterization of sample regions, the system comprising:

a processor; and

a non-transitory computer-readable medium operably coupled to the processor, the computer-readable medium comprising instructions that, when executed by the processor, cause the system to

receive multi-modal data comprising at least a set of data comprising optical values and associated sample region values generated from a confocal microscope apparatus configured to illuminate a subset of sample regions of a pattern of sample regions with a light beam under confocal point illumination to induce optical signals and to detect the optical signals; and an additional set of data comprising chemical or structural values and associated sample region values generated from a microscope apparatus configured to conduct a correlative characterization technique on the subset of sample regions;

perform a statistical analysis on the multi-modal data to generate correlation data; and

output the correlation data.

7. The system of claim 6, wherein the non-transitory computer-readable medium further comprises instructions that, when executed by the processor, cause the system to determine the subset of sample regions by comparing optical values of an initial set of data to predetermined optical values, the initial set of data comprising optical values and associated sample region values generated from a wide-field microscope apparatus configured to illuminate the pattern of sample regions with a patterned light beam to induce a patterned optical signal and to detect the patterned optical signal.

8. The system of claim 6, wherein the non-transitory computer-readable medium further comprises instructions that, when executed by the processor, cause the system to illuminate the pattern of sample regions with a patterned light beam, illuminate the subset of sample regions with the light beam under confocal point illumination, conduct the correlative characterization technique on the subset of sample regions, or a combination thereof.

9. A non-transitory computer-readable medium comprising instructions that, when executed by a processor, cause a system for multi-modal optical characterization of sample regions to

receive multi-modal data comprising at least a set of data comprising optical values and associated sample region values generated from a confocal microscope apparatus configured to illuminate a subset of sample regions of a pattern of sample regions with a light beam under confocal point illumination to induce optical signals and to detect the optical signals; and an additional set of data comprising chemical or structural values and associated sample region values generated from a microscope apparatus configured to conduct a correlative characterization technique on the subset of sample regions;

perform a statistical analysis on the multi-modal data to generate correlation data; and

output the correlation data.

10. The non-transitory computer-readable medium of claim 9, further comprising instructions that, when executed by the processor, cause the system to determine the subset of sample regions by comparing optical values of an initial set of data to predetermined optical values, the initial set of data comprising optical values and associated sample region values generated from a wide-field microscope apparatus configured to illuminate the pattern of sample regions with a patterned light beam to induce a patterned optical signal and to detect the patterned optical signal.

11. The non-transitory computer-readable medium of claim 9, further comprising instructions that, when executed by the processor, cause the system to illuminate the pattern of sample regions with a patterned light beam, illuminate the subset of sample regions with the light beam under confocal point illumination, conduct the correlative characterization technique on the subset of sample regions, or a combination thereof.

12. A method for multi-modal optical characterization of sample regions, the method comprising:

receiving, by a processor, multi-modal data comprising at least a set of data comprising optical values and associated sample region values generated from a confocal microscope apparatus configured to illuminate a subset of sample regions of a pattern of sample regions with a light beam under confocal point illumination to induce optical signals and to detect the optical signals; and an additional set of data comprising chemical or structural values and associated sample region values generated from a microscope apparatus configured to conduct a correlative characterization technique on the subset of sample regions;

performing, by the processor, a statistical analysis on the multi-modal data to generate correlation data; and

outputting, by the processor, the correlation data.

13. A method for multi-modal optical characterization of sample regions, the method comprising:

performing an initial screening by illuminating a pattern of sample regions with a patterned light beam to induce a patterned optical signal and detecting the patterned optical signal to generate a first set of data comprising optical values and associated sample region values;

performing a first subsequent screening by illuminating a subset of sample regions of the pattern of sample regions with a light beam under confocal point illumination to induce optical signals and detecting the optical signals to generate a second set of data comprising optical values and associated sample region values; performing a second subsequent screening by conducting a correlative characterization technique on the subset of sample regions to generate a third set of data comprising chemical or structural values and associated sample region values; and

performing a statistical analysis on multi-modal data comprising at least the second and third sets of data to generate correlation data.

14. The method of claim 13, wherein more than one patterned light beam is used in the initial screening step, more than one light beam under confocal point illumination is used in the first subsequent screening step, or both.

15. The method of claim 13, wherein the correlative characterization technique is Raman microscopy, atomic force microscopy, scanning electron microscopy, transmission electron microscopy, near-field scanning optical microscopy, or a combination thereof.

16. The method of claim 13, wherein the pattern of sample regions comprises a plurality of different samples arranged in an array, each sample region corresponding to one of the plurality of different samples.

17. The method of claim 13, wherein the sample regions are composed of one or more types of semiconductors.