High-Speed Spectroscopic Transient Absorption Imaging of Defects in

Subscriber access provided by UNIV OF OREGON

Communication

High-speed Spectroscopic Transient Absorption Imaging of Defects in Graphene Kai-Chih Huang, Jeremy McCall, Pu Wang, Chien-Sheng Liao, Gregory Eakins, Ji-Xin Cheng, and Chen Yang Nano Lett., Just Accepted Manuscript • DOI: 10.1021/acs.nanolett.7b05283 • Publication Date (Web): 17 Jan 2018 Downloaded from http://pubs.acs.org on January 18, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Nano Letters is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

High-speed Spectroscopic Transient Absorption Imaging of Defects in Graphene Kai-Chih Huang1,‡, Jeremy McCall2,‡, Pu Wang3,4,‡, Chien-Sheng Liao1, Gregory Eakins5, Ji-Xin Cheng1,6,*, and Chen Yang6,7,* 1

Department of Biomedical Engineering, Boston University, Boston, MA 02215, USA

2

Department of Chemistry, Purdue University, West Lafayette, IN 47907, USA

3

School of Biological Science and Medical Engineering, Beihang University, Beijing 100083,

China 4

Beijing Advanced Innovation Centre for Biomedical Engineering, Beihang University, Beijing

102402, China 5

Jonathan Amy Facility for Chemical Instrumentation, Purdue University, West Lafayette, IN

47907, USA 6

Department of Electrical and Computer Engineering, Boston University, Boston, MA 02215,

USA 7

Department of Chemistry, Boston University, Boston, MA 02215, USA

‡

These authors contributed equally to this work.

*Corresponding Author. E-mail: [email protected], [email protected]

Abstract: Graphene grain boundaries (GBs) and other nano-defects can deteriorate electronic properties. Here, using transient absorption (TA) microscopy, we directly visualized GBs by TA intensity increase due to change in density of state. We also observed a faster decay due to defect-accelerated carrier relaxation in the GB area. By line-illumination and parallel detection, we increased the TA intensity imaging speed to 1,000 frames per second, which is 6 orders of

1 Environment ACS Paragon Plus

Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 26

magnitude faster than Raman microscopy. Combined with a resonant optical delay tuner which scans a 5.3 picosecond temporal delay within 92 µs, our system enabled spectroscopic TA imaging, at a speed of 50 stacks per second, to probe and characterize graphene nano-defects based on the TA decay rate. Finally, we demonstrate real-time non-destructive characterization of graphene at a rolling speed of 0.3 m/min, which matches the fastest roll-to-roll manufacturing process reported. Keywords: graphene; grain boundary; ultra-high-speed microscopy; transient absorption microscopy.

Graphene, a 2-D single atomic layer nano-material, is attractive for several applications, including future electronics, based on its outstanding electrical, mechanical, and chemical properties.

1

Large-area graphene films are achieved by chemical vapor deposition (CVD)

2, 3

.

However, during CVD growth, nucleation of graphene occurs at multiple sites and each grows to island structures, consist of one or more crystal orientations distinct from its neighbors. The islands intersect and form grain boundaries (GBs), which often present a unique change in the band structure and are viewed as inevitable defects. Several methods have been developed for imaging GBs. Scanning tunneling microscopy (STM) has been reported as a reliable method for graphene GB analysis 4, whereas the nanoscale resolution requires billions of pixels for micrometer-scale grain imaging. Dark-filed TEM allows micrometer-scale grain imaging but with complicated sample preparation and requires specific substrates 5. Raman microscopy can characterize graphene with both global properties (i.e. coverage, layer number) and local properties (i.e. GBs, nano-defects) based on the molecular vibrational spectrum 6, but it necessitates an acquisition time of seconds per pixel. Thus, for in


Page 3 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

situ detection of nanoscale defects during the manufacturing process, a new modality that allows high-speed imaging of GBs and other defects in large-scale is needed.

Figure 1. Principle and setup of ultrafast TA imaging of nano-defects in graphene. (a) Mechanism of TA process of graphene. (b) Modulation transfer from pump to probe beam due to TA process. (c) Conventional laser dot-scan in a raster pattern. PD: photodiode. (d) Line illumination and parallel detection by TAMPs improve acquisition time. PD array: photodiode


Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 26

array. (e) Conventional optical delay scan by a mechanical stage. (f) Resonant delay tuning improves the temporal resolution from several seconds to ~92 µs. When measuring graphene, graphene electrons are excited by the pump from the valence band to the conduction band. The subsequent depletion of the valance band and pump-excited carriers inhibit the absorption of the sequential probe pulse due to ground state bleach and Pauli blocking, respectively, thus yielding a decreased absorption for the probe beam (Fig. 1a). Experimentally, the pump laser is modulated in MHz rate in order to reduce the 1/f laser noise. After interaction with the sample, the probe laser experiences intensity change at the same modulation frequency (Fig. 1b). This small change of probe beam intensity, which can be extracted by a lock-in amplifier, enables visualization of the carrier density in the excited state. Time-resolved TA signal, obtained as a function of the optical delay of the pump and probe, provides insights on carrier relaxation dynamics in graphene. Different time scale of the carrier dynamics in graphene has been studied by transient absorption measurement. First, the laser excites electrons from the valence band to conduction band as the “hot carriers”. Within several tens of femtoseconds, electron-electron (carrier-carrier) scattering7,

8

leads the system to quasi-equilibrium. In hundreds of femtoseconds, the

photoexcited carriers cool down by transferring energy to the lattice via intraband transitions by optical phonons, and interband transitions by radiative electron-hole (e-h)

9-11

. In the following

several picoseconds, the energy dissipates through acoustic phonon emission and the system arrives at thermal equilibrium12. However, the velocity of phonon is much slower than the velocity of electron. The momentum conservation constraint and velocity mismatch cause inefficient carrier relaxation through acoustic phonon emission.


Page 5 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

Recently, Song et al. predicts another carrier relaxation pathway by the defect-assisted acoustic phonon scattering as the “supercollision” model. In this model, the collision between a carrier and both an acoustic phonon and a defect relaxes the momentum constraint and provides a faster energy dissipation process compared with acoustic phonon emission. Several experimental results of graphene are consistent with the SC model

13 14

. Therefore, TA decay curves is a hot-

electron thermometer that records the cooling dynamics of hot electrons in graphene 15 and have been used to quantitatively characterize nano-scale defect density based on SC model 16. Here, we demonstrated that TA microscopy is able to visualize GBs, an important case of defects. We observed that, in the TA intensity image at zero-time delay, TA signal increases at GB sites. We also provided insights on image mechanism. We show that TA intensity at GB sites increases because of changes in the density of states induced by disorientation of inter-grains defects resulting in van Hove singularities. Thus, by matching the pump wavelength with resonances in the band structure, TA microscopy has the selectivity to image certain GB with different density of states. Meanwhile, in the TA decay mapping processed from time-resolved TA signal, the TA decay rate is faster at GB sites compared with single layer graphene area. The faster decay rate in GB area is consistent with the fact that defects in graphene accelerate the cooling of photoexcited carriers. We further reported an unprecedented imaging speed for TA microscopy to enable real-time characterization of nano-defects in large-area CVD graphene. First, in order to improve the spatial acquisition speed by the conventional raster-scanning (Fig. 1c), we deployed lineilluminated scanning scheme and parallel detection with a lab-designed tuned amplifier (TAMP) and photodiode array (Fig. 1d). Our method allows TA imaging at 1,000 frames per second at a specific temporal delay. At such speed, the surface coverage, degree of wrinkles, grain boundary,


Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

and layer numbers of graphene can be resolved with 35:1 contrast ratio of single layer graphene. Second, in order to improve the temporal delay tuning speed by conventional delay stage tuning method (Fig. 1e), we integrated the resonant delay tuner with an angle-to-delay convertor (Fig. 1f). We achieved acquisition of a time-resolved TA curve within 92 µs and decay mapping at the speed of 50 image stacks per second. Moreover, we are able to extract the density of nanodefects below the diffraction limit from the TA decay rate. Our technology enables high throughput characterization of graphene under the speed of 0.3m/min matching the state-of-art roll-to-roll manufacturing process.

Figure 2. TA imaging of graphene grain boundary. (a) TA intensity image of grain boundary at zero-time delay. (b) TA decay mapping image of grain boundary from time-resolved TA signal. The darker area indicates faster decay rate. (c) Raman map of the ratio of the D peak (~1,350 cm-1) to the G peak (~1,580 cm−1). (d) The grain boundary is not visible in the conventional wide-field optical image. Scale bars: 20 µm. TA imaging of graphene grain boundary. Fig. 2a demonstrates a TA intensity image where TA intensity signal at zero temporal delay between the pump and the probe is acquired at each pixel over a graphene sample synthesized by CVD method. Line features indicated by a higher TA intensity compared to that of single layer graphene areas are shown. Those inter-grain structures suggest GBs. Fig. 2b demonstrates TA decay mapping in the same area as in the Fig. 2a. The slow decay time constant τ2 in each pixel is extracted from a time-resolved TA signal


Page 6 of 26

Page 7 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

and fitted with bi-exponential decay function. Since defects in graphene assist relaxation process of photoexcited carriers, the darker line features with faster decay rate can be assigned as GBs, as a structural disorder in graphene. We confirmed the assignment of GBs by Raman spectroscopic imaging of the same area (Fig. 2c). We acquired Raman spectrum in each pixel and generated Raman image with the contrast of D peak to G peak intensity ratio. GBs were revealed with a higher D/G ratio, consistent with previous report6. The features highlighted by the red arrows in Fig. 2a and Fig. 2b co-localized with higher D/G ratio in Fig. 2c. Collectively, our results demonstrate that TA signal can be used for GB characterization with information comparable to Raman imaging. This GB site is not visible in a conventional wide-field optical image recorded from the same area of the same specimen (Fig. 2d). Via a detailed TA imaging study, we identified a few novel signatures of GBs. First, we noticed that GBs are distinct from other features which have similar dimensions, such as wrinkles. Specifically, Fig. S1a shows the TA intensity of wrinkles and GBs compared to the intensity from single layer graphene. Each of the features noted has a distinct change in intensity. Statistically, the increase of TA intensity at GB area is found to be 30 ± 9 %, much higher than the increase of 8± 2% at wrinkles, and much smaller than the increase of 93 ± 10 % at bi-layer graphene (Fig. S1b-c).


Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 3. Angular dependence of graphene grain boundary. (a) Schematic of orientations of intersecting hexagons. (left) misorientation angle of 30º. (middle) misorientation angle of 0º. (right) a representative fitting. (b) Histogram of intensity changes as a function of the misorientation angle. (c) Intensity map at zero-time delay (λPump = 1040 nm). Scale bar is 20 µm. (d) Plot of intensities at noted features with varying pump wavelength. Probe wavelength is 810 nm for all data points. We further observed that not all intersecting hexogen grains show a change in the TA intensity. We examined the visible GBs and measured the misorientation angle of the intersecting grains. In principle, two single crystal hexagon domains can intersect with an angle between 0 and 30 degrees (Fig. 3a). However, the histogram on visibility of GBs depending on misorentation angles shows that the intensity change at the intersection was only observable when the grains intersect at ~10 to 15 degrees (Fig. 3a-b). This observation suggests that TA intensity contrast at the GBs could be due to a change in the density of states occurring at the intersection of grains


Page 8 of 26

Page 9 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

through creating van Hove singularities

17, 18

. The energy of the van Hove singularities depends

on the misorientation angle of the grains. The van Hove singularity for an angle of 10-15 degrees occurs at ~2.1-2.6 eV and resonates with two-photon absorption process contributed by the 1040 nm pump beam. Therefore, states induced by 10-15 degree misorientation angles become two photons accessible through the 1040 nm pump, which results in a significant increase in signal being probed. We confirmed this mechanism by tuning the pump wavelength to resonate with the van Hove singularity energy. We compared the intensity profile of three GB sites in Fig. 3c at the pump wavelengths from resonance (1000 nm) to off-resonance (1200 nm). The TA signal from GBs became weaker when the pump wavelength tuned off the resonances (Fig. 3d). In sum, the results are consistent with our hypothesis that the increases of TA signal at GBs are because the pump wavelength resonates with the energy of van Hove singularities at specific misorientation angle of grains intersection and changes the density of states. Ultrafast TA microscopy. The speed is essential for large-scale imaging of graphene. Several factors limit the speed in current TA imaging methods. First, the spatial acquisition speed is limited by raster-mode laser scanning (Fig. 1c). Second, for temporal decay measurements, the frame-by-frame manner by a mechanical delay stage limits the temporal resolution to several seconds to minutes (Fig. 1e). In order to address the raster scanning limitation, we adopted a line-illuminated scanning scheme to improve the spatial acquisition speed

19

(Fig. 1d, Fig. S2).

By using higher laser power and spreading the laser energy into a line on the focal plane, the laser power density remains the same as the dot-scan scheme on samples, while the imaging speed is significantly improved at a rate corresponding to the number of pixels in the line. The remaining challenge lies in the parallel detection of the TA signals. The conventional way is to utilize a multichannel lock-in amplifier; however, commercially available multichannel lock-in


Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 26

amplifiers can only demodulate signals up to 50 kHz 20, which precludes its application in highspeed imaging. To address this issue, we built a tuned amplifier (TAMP) array to achieve lock-in free parallel detection of TA signals. In a previous study, we showed that signals modulated at MHz with modulation depth of as small as 10−6 dI/I can be extracted and amplified by a quartersized chip

21

. We further built a compact 16-channel TAMP array for parallel detection of

stimulated Raman signals at 32 µs pixel dwell time

22

. In this work, we further increased the

number of channels of the TAMP array to 32, and eliminated the inherent deviation in resonant frequencies between TAMP channels by incorporating tunable inductors (Fig. S3). Moreover, we improved the acquisition speed to 5 µs per pixel. This device allowed us to demodulate 32 TA signals at MHz frequency and acquire a TA image of 32×200 pixels within 1 millisecond (ms). In order to achieve fast acquisition of a decay curve at each pixel, we developed a microsecond delay tuning method by integrating a resonant mirror with a digital micromirror device (DMD) (Fig. 1f). By taking advantage of DMD as a wedged step reflector, we convert the angle scanning into delay tuning. Thus, a temporal delay scan is accomplished within a single period of the resonant mirror. In this way, we successfully improved the temporal resolution for timeresolved TA imaging from second scale to microsecond scale.


Page 11 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

Figure 4. TA image acquired by line-scan and dot-scan schemes. (a) Image of CVD graphene and air interface by dot-scan and line-scan with TA intensity as image contrast. Acquisition time is 32 ms and 1 ms for dot-scan and line-scan, respectively. For line-scan image, the lineillumination is in the vertical direction and scan in horizontal direction. (b) Intensity profile of the orange lines in (a). The single layer signal-to-noise ratio (SNR) is 35 for both cases. (c) Image of suspected wrinkles (in the left with the “V” shape) and cracks. (d) Intensity profile of the orange line in (c). The intensity height of suspected wrinkles is lower than the intensity height of single layer. Scale bar 5 µm (e) TA intensity image of grain boundary with line-


Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 26

scanning scheme (f) TA intensity image of grain boundary with dot-scanning scheme. (g) Raman map of the ratio of the D peak to the G peak. Scale bars are 20 µm. Ultrafast TA imaging of graphene by the line-illuminated scanning scheme. We first compared the performance our line-scanning scheme with the conventional dot-scanning scheme by imaging pristine graphene grown by the thermal CVD method (Fig. 4a). In both cases, the coverage and the number of layers were clearly visualized. The hexagonal shape of a second layer of graphene in the early stage of growth could also be resolved. The intensity profiles along the orange lines show the same signal-to-noise ratios (Fig. 4b), with value of 35:1 for single layer graphene. Moreover, other local features assigned as wrinkles and cracks were revealed in both images (Fig 4c-d). Moreover, our line-scanning clearly visualized GBs that exhibit higher TA intensity (Fig 4e-g). To note that no photodamage was observed based on the comparison between Raman spectra taken at the same location before and after TA imaging (Fig. S4) 23. We further demonstrated an imaging speed of 1,000 frames per second by recording a movie of manually moved graphene sample (Movie. S1). Collectively, our high speed TA method offers an imaging speed 32 times faster by line-illumination and parallel detection through acquisition of 32 channels compared to the conventional TA dot scanning microscopy. More significantly, our method enables ~6 orders of magnitude faster than Raman microscopy. With such speed, the layer number, coverage, degree of wrinkles, and GBs are clearly observed. Together, our TA imaging technology offers a characterization tool for graphene GBs in ultra-high-speed while provides similar information as Raman microscopy. Ultrafast time-resolved TA imaging of nanoscale defects by resonant delay-tuning. Of the characteristic features in graphene that can be measured with optical resolution, the density of nanoscale defects is a critical characteristic to evaluate the quality of graphene


24-26

. Raman

Page 13 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

microscopy/spectroscopy, with limited spatial resolution due to optical diffraction, is a gold standard to perform ensemble measurement of the density of nanoscale defects

6, 24, 27, 28

.

Nanoscale defects can also be resolved by time-resolved TA signal by mapping changes in the carrier relaxation rate caused by defect scattering, which resembles the information provided by Raman spectroscopy

16

. However, both Raman and TA spectroscopic measurements require

several second, thus making the 2D mapping of such characteristic formidable.

Here, we

decreased the time of measuring the density of nano-defects from second to microsecond scale by implementing a resonant delay tuning method for delay line scanning. In our method, the delay between the pump and probe beams is controlled by the angle of the resonant mirror. To quantitatively correlate the rotating time period of the resonant mirror to the temporal delay, and to compensate the nonlinear rotating pattern of the resonant mirror, we first established a calibration curve shown in Fig. S5a using TA signals from known pristine graphene (See Methods in supporting information). The sinusoidal fitting of the calibration curve shows a frequency of 10.8 kHz, consistent with the resonant frequency of the resonant mirror. To evaluated our microsecond-scale time-resolved TA system, we acquired TA signals from a graphene sample by our resonant delay tuning and compare to that by a conventional mechanical stage. Identical decay curves were obtained, as shown in Fig. S5b. More significantly, the decay constant τ2 obtained by bi-exponential fitting of TA spectra can be used to quantify the density of nano-defects in a graphene sample 15, 16. To test this capacity, we prepared graphene samples of various defect densities, induced by exposing to high power laser (from 1.2 MW/cm2 to 1.8 MW/cm2, See Methods in supporting information)

16

. We then

recorded time resolved TA images with delay from 0 ps to 5 ps within 18.4 ms, which is 100 times faster than the state-of-the-art point scan with mechanical delay tuning method


18

. TA

Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

images of graphene samples with four different defect densities were shown in Fig. 5a Specifically, in the right panel, a contrast was clearly observed between the area that was exposed under a laser beam of 1.8 MW/cm2 and the pristine area. To quantify the defects based on the reported method 29, Raman spectroscopy was performed at the boxed areas shown in Fig. 5b The Raman spectra showed D to G band intensity ratios of 0.45, 1.41, and 2.57 for boxed areas A, B, and C, respectively, consistent with the defect densities prepared. The average distance between defects ( ), a measure of the defect density, were estimated to 17.6 nm, 9.5 nm, and 6.3 nm for boxes A, B and C, respectively, based on the empirical model by Cancado et al

29

. Independently, the average TA spectra obtained from each boxed area and corresponding

fitting curves based on bi-exponential fitting are shown in Fig. 5c The Supercollision (SC) scattering rates A/α indicative of defect densities can be obtained using the SC model 16, 30 based on TA decay constants τ2 from the bi-exponential fitting of TA spectra. Fig. 5d shows that the derived SC rates are linearly correlated to independently obtained from Raman spectroscopy, as predicted by the SC model

16

. These results collectively suggest that our high-speed time-

resolved TA microscope enables ~4 orders of magnitude faster data acquisition while provides information on nanoscale defect density in graphene compared to Raman microscopy.


Page 14 of 26

Page 15 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

Figure 5. Quantification of nano-defect density in graphene by time-resolved TA signals. (a) TA images at zero-time delay of graphene with different induced defects highlighted by the boxes A~C. (b) Raman spectra measured at the boxed areas in C. The spectrum of the pristine area is also shown (Black). All the spectra are normalized to have the same G band intensity. (c) TA decay curves measured at the boxed areas and the corresponding bi-exponential decay fitting curves. (d) SC scattering rate A/α as a function of LD-1. LD-1 of pristine graphene is assumed to 0. In addition, the fitting error of the decay constant, which depends on the signal-to-noise ratio of TA spectroscopy, determines the minimal defect density that can be detected. Specifically, the TA decay time constant τ2 extracted from fitting of the pristine area and the box A area (Fig. 5a) are 1.88 ± 0.24 ps and 1.63 ± 0.19 ps, respectively. Thus, the fitting errors are comparable to the


Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

difference in the fitted constants. Such results suggest that our strategy is able to measure a minimal defect density at defect length of 17.6 nm under the condition of 0.24 MW/cm2 for pump and 0.17 MW/cm2 for probe on the sample, integrated area of 49 pixels, and TA spectra acquisition time of 92 µs.

Figure 6. Time-resolved TA images of pristine and defect graphene. (a) In the image contrast as TA intensity (upper row), the pristine graphene (up-left) and defect graphene (up-right) both show homogeneous intensity distribution. Whereas in the image contrast as TA decay time constant (bottom row), we are able to differentiate pristine graphene (bottom-left) and defect graphene (bottom-right) since defect one decays much faster. (b) The decay curve plotted from the green square area (25 pixels). The defect graphene decays faster. The decay time constant is 2.050 ± 0.146 ps for defect graphene, and 0.436 ± 0.045 ps for pristine graphene after the exponential fitting. (c) 5 graphene flakes were imaged within 150 ms at 0.3 m/min motorized stage speed with 0.5 µm step size. All scale bar 10 µm. Visualizing nanoscale defects by using TA decay rate as contrast. Our system is able to discriminate defective graphene and pristine graphene by using the TA decay time constant as


Page 16 of 26

Page 17 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

the image contrast. In the image based on the TA intensity at zero delay between pump and probe (Fig. 6a upper), both samples show homogeneous intensity distribution. The defects in nanoscale, which are below the optical diffraction limit, are not resolvable. However, because the time-resolved TA signals reflects carrier relaxation dynamics of the excited state, we are able to discriminate the defects by using the TA decay time constant as the contrast (Fig. 6a below). Since collision between a carrier and a structural defect accelerates the relaxation process by SC cooling, the defective graphene shows a much smaller decay time constant than the pristine graphene in the time-resolved TA image (Fig. 6b). On the contrary, the pristine graphene shows a homogeneously large decay time constant. The defective property of graphene was confirmed by Raman spectroscopy (Fig. S6). Since defects may happen in the roll-to-roll manufacturing process and could deteriorate the performance of graphene-related devices

31

, it is important to image graphene quality in real

time. For this purpose, we further demonstrated high-speed, large-scale scan of graphene in the condition similar to roll-to-roll process (Fig. 6c, Movie. S2). The illumination line is fixed and the motorized stage moves at constant speed. The imaging speed of 0.3 m/min with 0.5 µm spatial resolution is able to catch up the roll-to-roll process, which is 100 times faster than the state-of-art transient absorption microscopy under a similar field of view

18, 32

. These data

collectively show that our platform is able to detect nano-defects in graphene in a roll-to-roll production process. TA technique has been reported for characterization of other materials. Schnedermann et. al used wide-field transient absorption microscopy to image the GB of MAPbI3−xClx perovskite films, and observed that the TA curve of GB site is different from that in the grain area 33. Here, we used time-resolved TA microscopy and report the first visualization of GBs in graphene. We


Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 26

further propose the mechanism of changed density of states at GB sites, which is due to vHSs interaction depending on misorientation angle of the grains. The changes of local density of states at graphene grain boundaries have been studied in theory 34-36 and experimentally 37

38 39

.

Jiwoong Park’s group reported that in the twisted bilayer graphene (tBLG), the energy of the van Hove singularities (vHS) depends on the misorientation angle of the grains

40

. This interlayer

vHSs interaction creates new excitonic states and changes the optical conductivity of tBLG. Here, we found that TA intensity shows contrast at the GBs, which could be a result of change in the density of states occurring at the intersection of grains through creating vHSs. The vHS for an angle of 10-15 degrees occurs at ~2.1-2.6 eV and resonates with two-photon absorption process contributed by the 1040 nm pump beam. While tuning the wavelength to off-resonance, the TA signal decreases, as indicated in Fig. 3d. Our TA image results indicate that GBs locate in the intersection of two hexagonal grains, which is consistent with reported results by transmission electron microscopy (TEM) and Raman microscopy. Conventionally TEM is used to characterize the polycrystalline grain structures. By TEM, GB can be identified by the intersection of two grains with different selected area electron diffraction (SAED) pattern based on different crystal orientation 41. Along this direction, Yong P. Chen’s group used TEM to determine the location of GBs on the CVD graphene 6. Here, we imaged the CVD graphene provided by Yong P. Chen’s group and our TA microscopy technique also indicates that GBs locate in the intersection of two hexagonal grains as the TEM result shown in ref 6. We further confirmed our observation by Raman microscopy which is a gold standard to identify graphene GBs reported by multiple groups

6, 27, 28

. Graphene GB site can be

visualized by the increased signal in the Raman mapping of the D peak intensity. The GB


Page 19 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

assignment from TA image is consistent with Raman spectroscopic imaging at the same area (Fig. 2). Our TA image result (Fig. S1a) matches well with the reported Raman image data 6. We have pushed the acquisition speed of TA signal, which allowed mapping of graphene surface coverage, layer numbers, defect density and local defects in large-scale and in real-time. The imaging speed at one specific delay has reached 1,000 frames per second using the lineillumination and paralleled heterodyne detection method. To tune the optical delay and acquire time-resolved TA signals within a few microseconds, we developed a resonant optical delay tuner. Conventionally a motorized translational stage with retro-reflected mirrors is used for delay tuning at the speed of second-scale, limited by the waiting time for stabilization and communication18. The development of frequency-comb enables nanosecond optical delay tuning by two asynchronous laser pulses within a few milliseconds. Microsecond-scale optical delay tuning has been recently demonstrated by directing light to the edge of a tilted resonant mirror. Within 84 µs, up to 3 ps optical delay tuning has been reported. The acousto-optic programmable dispersive filter (AOPDF) has been reported to scan the optical delay of chirped femtosecond pulses for stimulated Raman imaging by Alshaykh et., al. A maximum delay of 7.9 ps was achieved within 12.8 µs, and the total acquisition time of one scan was ~30 µs 42. Our method scans the optical delay of femtosecond pulses with up to 29 ps delay tuning range at microsecond-scale speed. The acquisition time of 92 µs is determined by the resonant frequency of scanning mirror, which can be improved to ~50 µs by the state-of-art resonant mirror. The range optical delay scanning is flexible and determined by the numbers of pixel on the DMD. We performed 5.3 ps scanning range as a proof-of-concept demonstration, which corresponds to 346 pixels on the DMD. By using the full pixel number of our DMD (1920 × 1080 pixels), the maximum delay tuning range can be up to


Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

29.38 ps, which is sufficient for current advanced 2-D materials and biological samples. Moreover, since the DMD and resonant mirror are reflective components, the induced chirping effect to femtoseconds pulses can be minimized, which does not compromise the temporal resolution of time-resolve TA signals. Combined our resonant delay tuner with line-illuminated scanning scheme, we improved the speed of delay scanning to 92 µs temporal resolution and achieved a scan rate of 55 time-resolved TA stacks per second. Such high speed capability enabled visualization of graphene defects with a rolling speed of 0.3 m/min, which is comparable to state-of-art roll-to-roll process. There is still plenty of space for future optimization. Potentially the acquisition time can be further improved to 1.33 µs per line by using a multichannel data acquisition board with 750k samplings/second (Standard Generation Corporation, 18AI64SSC750K). The number of channel can be also improved through utilizing a photodiode array and multichannel data acquisition board with more channels. Notably, the strategy of reducing the integration time will reduce the signal-to-noise ratio and the sensitivity. In conclusion, the new platform reported here opens up new opportunities for study of graphene grain structure with respect to different synthesis strategies. For TA imaging in our system, a 1.0 cm by 1.0 cm substrate with 0.5 µm step size only takes ~1 minute, whereas the state-of-art TA microscope takes ~0.5 hour. In addition to graphene, our system is also applicable to studies of biological systems and resolve other samples, such as melanoma 43, thin films 44, other 2D materials, as well as their heteostructures (i.e., WSe2, MoS2, WS2) 45-47, largescale historic artwork 48. As an ultrafast label-free nonlinear method, it can also be applied to the other modality, such as superfast hyperspectral stimulated Raman scattering microscopy using chirped pulses 49.


Page 20 of 26

Page 21 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

ASSOCIATED CONTENT Supporting Information is available (file type: PDF) (Material and methods. Intensity change from features in graphene. Experimental setup. Photos of TAMP array. Raman spectrum of pristine graphene at the same location before and after the TA measurement. Calibration of acquired time-resolved TA signals. Raman spectrum of pristine graphene and defect graphene. Diagram of DMD as a wedged step reflector. Frequency response of 32 resonant amplifiers. Spontaneous Raman mapping of graphene with optically generated defects.) Movies of transient absorption intensity image of graphene by manually moving the sample stage. (file type: mp4) and high-speed large-scale scan of graphene by TA time-resolved microscopy. (file type: mp4). AUTHOR INFORMATION Corresponding Author *E-mail: [email protected], [email protected] Author Contributions K.H., P.W., C.L., and J.X.C. designed the experiments; K.H., J.M., and C.L. performed the experiments; J.M. provided the graphene samples; K.H. and J.M. analyzed the data; G.E. built the tuned amplifier array. K.H., J.M., P.W., C.L., C.Y., and J.X.C. wrote the manuscript; C.Y. and J.X.C. performed overall guidance of the project. All authors read and approved the manuscript. ‡These authors contributed equally. Funding Sources This work was supported by a Keck Foundation Science and Engineering grant. Notes


Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

J.-X.C. and P.W has a financial interest in Vibronix Inc. Acknowledgement The authors thank Dr. Yong Chen for help in sample preparation. Abbreviations TA, transient absorption; GB, grain boundary; CVD, chemical vapor deposition.

References 1.

Geim, A. K.; Novoselov, K. S. Nature materials 2007, 6, (3), 183-91.

2.

Li, X.; Cai, W.; An, J.; Kim, S.; Nah, J.; Yang, D.; Piner, R.; Velamakanni, A.; Jung, I.;

Tutuc, E.; Banerjee, S. K.; Colombo, L.; Ruoff, R. S. Science 2009, 324, (5932), 1312-4. 3.

Bae, S.; Kim, H.; Lee, Y.; Xu, X.; Park, J. S.; Zheng, Y.; Balakrishnan, J.; Lei, T.; Kim,

H. R.; Song, Y. I.; Kim, Y. J.; Kim, K. S.; Ozyilmaz, B.; Ahn, J. H.; Hong, B. H.; Iijima, S. Nature nanotechnology 2010, 5, (8), 574-8. 4.

Rasool, H. I.; Song, E. B.; Mecklenburg, M.; Regan, B. C.; Wang, K. L.; Weiller, B. H.;

Gimzewski, J. K. J Am Chem Soc 2011, 133, (32), 12536-12543. 5.

Huang, P. Y.; Ruiz-Vargas, C. S.; van der Zande, A. M.; Whitney, W. S.; Levendorf, M.

P.; Kevek, J. W.; Garg, S.; Alden, J. S.; Hustedt, C. J.; Zhu, Y.; Park, J.; McEuen, P. L.; Muller, D. A. Nature 2011, 469, (7330), 389-92. 6.

Yu, Q.; Jauregui, L. A.; Wu, W.; Colby, R.; Tian, J.; Su, Z.; Cao, H.; Liu, Z.; Pandey, D.;

Wei, D.; Chung, T. F.; Peng, P.; Guisinger, N. P.; Stach, E. A.; Bao, J.; Pei, S. S.; Chen, Y. P. Nature materials 2011, 10, (6), 443-9. 7.

Breusing, M.; Ropers, C.; Elsaesser, T. Physical review letters 2009, 102, (8), 086809.


Page 22 of 26

Page 23 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

8.

Tielrooij, K. J.; Song, J. C. W.; Jensen, S. A.; Centeno, A.; Pesquera, A.; Zurutuza

Elorza, A.; Bonn, M.; Levitov, L. S.; Koppens, F. H. L. Nature Physics 2013, 9, 248. 9.

Viljas, J. K.; Heikkilä, T. T. Physical Review B 2010, 81, (24), 245404.

10.

Kampfrath, T.; Perfetti, L.; Schapper, F.; Frischkorn, C.; Wolf, M. Physical review letters

2005, 95, (18), 187403. 11.

Sun, D.; Wu, Z.-K.; Divin, C.; Li, X.; Berger, C.; de Heer, W. A.; First, P. N.; Norris, T.

B. Physical review letters 2008, 101, (15), 157402. 12.

Bistritzer, R.; MacDonald, A. H. Physical review letters 2009, 102, (20), 206410.

13.

Graham, M. W.; Shi, S.-F.; Ralph, D. C.; Park, J.; McEuen, P. L. Nature Physics 2012, 9,

103. 14.

Laitinen, A.; Oksanen, M.; Fay, A.; Cox, D.; Tomi, M.; Virtanen, P.; Hakonen, P. J.

Nano Lett 2014, 14, (6), 3009-3013. 15.

Graham, M. W.; Shi, S.-F.; Wang, Z.; Ralph, D. C.; Park, J.; McEuen, P. L. Nano Lett

2013, 13, (11), 5497-5502. 16.

Alencar, T. V.; Silva, M. G.; Malard, L. M.; de Paula, A. M. Nano Lett 2014, 14, (10),

5621-4. 17.

Havener, R. W.; Liang, Y. F.; Brown, L.; Yang, L.; Park, J. Nano Lett 2014, 14, (6),

3353-3357. 18.

Patel, H.; Havener, R. W.; Brown, L.; Liang, Y. F.; Yang, L.; Park, J.; Graham, M. W.

Nano Lett 2015, 15, (9), 5932-5937. 19.

Palonpon, A. F.; Ando, J.; Yamakoshi, H.; Dodo, K.; Sodeoka, M.; Kawata, S.; Fujita, K.

Nature protocols 2013, 8, (4), 677-92.


Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

20.

Seto, K.; Okuda, Y.; Tokunaga, E.; Kobayashi, T. Journal of Physics D: Applied Physics

2014, 47, (34), 345401. 21.

Slipchenko, M. N.; Oglesbee, R. A.; Zhang, D. L.; Wu, W.; Cheng, J. X. J Biophotonics

2012, 5, (10), 801-807. 22.

Liao, C. S.; Slipchenko, M. N.; Wang, P.; Li, J. J.; Lee, S. Y.; Oglesbee, R. A.; Cheng, J.

X. Light-Sci Appl 2015, 4. 23.

Li, J.; Zhang, W.; Chung, T. F.; Slipchenko, M. N.; Chen, Y. P.; Cheng, J. X.; Yang, C.

Scientific reports 2015, 5, 12394. 24.

Chen, J. H.; Cullen, W. G.; Jang, C.; Fuhrer, M. S.; Williams, E. D. Physical review

letters 2009, 102, (23), 236805. 25.

Banhart, F.; Kotakoski, J.; Krasheninnikov, A. V. ACS nano 2011, 5, (1), 26-41.

26.

Yazyev, O. V.; Louie, S. G. Nature materials 2010, 9, (10), 806-9.

27.

Ferrari, A. C.; Meyer, J. C.; Scardaci, V.; Casiraghi, C.; Lazzeri, M.; Mauri, F.; Piscanec,

S.; Jiang, D.; Novoselov, K. S.; Roth, S.; Geim, A. K. Physical review letters 2006, 97, (18), 187401. 28.

Ferrari, A. C.; Basko, D. M. Nature nanotechnology 2013, 8, (4), 235-46.

29.

Cancado, L. G.; Jorio, A.; Ferreira, E. H. M.; Stavale, F.; Achete, C. A.; Capaz, R. B.;

Moutinho, M. V. O.; Lombardo, A.; Kulmala, T. S.; Ferrari, A. C. Nano Lett 2011, 11, (8), 31903196. 30.

Song, J. C. W.; Reizer, M. Y.; Levitov, L. S. Physical review letters 2012, 109, (10).

31.

Polsen, E. S.; McNerny, D. Q.; Viswanath, B.; Pattinson, S. W.; John Hart, A. Scientific

reports 2015, 5, 10257.


Page 24 of 26

Page 25 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

32.

Havener, R. W.; Ju, S. Y.; Brown, L.; Wang, Z.; Wojcik, M.; Ruiz-Vargas, C. S.; Park, J.

ACS nano 2012, 6, (1), 373-80. 33.

Schnedermann, C.; Lim, J. M.; Wende, T.; Duarte, A. S.; Ni, L.; Gu, Q.; Sadhanala, A.;

Rao, A.; Kukura, P. The Journal of Physical Chemistry Letters 2016, 7, (23), 4854-4859. 34.

Liu, Y.; Yakobson, B. I. Nano Lett 2010, 10, (6), 2178-83.

35.

Yazyev, O. V.; Louie, S. G. Physical Review B 2010, 81, (19), 195420.

36.

Mesaros, A.; Papanikolaou, S.; Flipse, C. F. J.; Sadri, D.; Zaanen, J. Physical Review B

2010, 82, (20), 205119. 37.

Tapasztó, L.; Nemes-Incze, P.; Dobrik, G.; Yoo, K. J.; Hwang, C.; Biró, L. P. Applied

Physics Letters 2012, 100, (5), 053114. 38.

Koepke, J. C.; Wood, J. D.; Estrada, D.; Ong, Z.-Y.; He, K. T.; Pop, E.; Lyding, J. W.

ACS nano 2013, 7, (1), 75-86. 39.

Adina, L.-M.; Jose, E. B.-V.; Jesper Toft, F.; Gabriel, A.; Aron, W. C.; David, S.;

Guohong, L.; Mads, B.; Oleg, V. Y.; Stephan, R.; Eva, Y. A. 2D Materials 2016, 3, (3), 031005. 40.

Havener, R. W.; Liang, Y.; Brown, L.; Yang, L.; Park, J. Nano Lett 2014, 14, (6), 3353-

3357. 41.

Kim, K.; Lee, Z.; Regan, W.; Kisielowski, C.; Crommie, M. F.; Zettl, A. ACS nano 2011,

5, (3), 2142-6. 42.

Alshaykh, M. S.; Liao, C. S.; Sandoval, O. E.; Gitzinger, G.; Forget, N.; Leaird, D. E.;

Cheng, J. X.; Weiner, A. M. Opt Lett 2017, 42, (8), 1548-1551. 43.

Matthews, T. E.; Piletic, I. R.; Selim, M. A.; Simpson, M. J.; Warren, W. S. Science

translational medicine 2011, 3, (71), 71ra15.


Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

44.

Kawashima, K.; Tamai, Y.; Ohkita, H.; Osaka, I.; Takimiya, K. Nature communications

2015, 6, 10085. 45.

Shi, H.; Yan, R.; Bertolazzi, S.; Brivio, J.; Gao, B.; Kis, A.; Jena, D.; Xing, H. G.;

Huang, L. ACS nano 2013, 7, (2), 1072-80. 46.

Cui, Q.; Ceballos, F.; Kumar, N.; Zhao, H. ACS nano 2014, 8, (3), 2970-6.

47.

Hong, X.; Kim, J.; Shi, S. F.; Zhang, Y.; Jin, C.; Sun, Y.; Tongay, S.; Wu, J.; Wang, F.

Nature nanotechnology 2014, 9, (9), 682-6. 48.

Villafana, T. E.; Brown, W. P.; Delaney, J. K.; Palmer, M.; Warren, W. S.; Fischer, M. C.

Proceedings of the National Academy of Sciences of the United States of America 2014, 111, (5), 1708-13. 49.

Fu, D.; Holtom, G.; Freudiger, C.; Zhang, X.; Xie, X. S. The journal of physical

chemistry. B 2013, 117, (16), 4634-40.

Table of Contents Graphic:


Page 26 of 26

High-Speed Spectroscopic Transient Absorption Imaging of Defects in

Recommend Documents