380 likes | 643 Views
pBTTT Processing for OFET Devices. Data & Anlaysis : Brian A. Collins & Eliot Gann (Ade Group) Sample Prep & Device Testing : Justin E. Cochran ( Chybanic Group) Nov. 8, 2010. Motivation. FET Basics Voltage on Gate creates E-Field
E N D
pBTTT Processing for OFET Devices Data & Anlaysis: Brian A. Collins & Eliot Gann(Ade Group) Sample Prep & Device Testing: Justin E. Cochran (Chybanic Group) Nov. 8, 2010 2010-09 pBTTT Bilayer Processing Study
Motivation FET Basics Voltage on Gate creates E-Field E-Field controls carrier concentration (CC)at Dielectric/Channel interface CC controls Source-Drain Current Gate Electrode Source Drain • Organic Field-Effect Transistors (OFET) could open up new applications for low-cost devices on flexible substrates • Field-Effect occurs at interface between conducting channel and dielectric layers • Interface and bulk morphology of conducting polymer very important • Processing conditions can be used to tune morphology • We study morphology with x-ray microscopy and scattering with UCSB preparing samples and studying device properties Dielectric pBTTT “Channel” Substrate 2010-09 pBTTT Bilayer Processing Study
What are the nominal layer thicknesses? Sample Batches As understood by Brian – Justin: please adjust as necessary Two batches made at separate times (same solutions?) • Batch 1 (orig. for STXM): Some on TEM grids and others on SiN windows (partial coverage) • Made first with some difficulty • Batch 2 (orig. for Scattering): some on Si substrates others on SiN windows (complete coverage) • Second set made after gained experience • Also had differentially cast bilayers (didn’t investigate) Single Layers Spincast PBTTT Annealing here is “Pre Annealed” OTS (self assembled monolayer) Si Lift-off w/PDMS Both Batches on OTS? PDMS PBTTT OTS (self assembled monolayer) Si Transfer onto PMMA PDMS PBTTT PMMA Annealing here is “Post Annealed” Si or SiN window How was PMMA deposited? 2010-09 pBTTT Bilayer Processing Study
Data Sets • Brian took STXM on Batch 1 at 5.3.2 • Samples: Both single layer PBTTT and bilayers • Energies: 285.4eV (PBTTT π*), 288.4eV (PMMA σ*), 390eV (thickness, no dichroism?) • Hongping took STXM on Batch 2 at 11.0.2 • Samples: Just bilayers • Energies: 285.4eV, 287.5eV, 320eV (thickness, no dichroism?) • Polarizations: Circular, Horizontal and Vertical Linear • Brian took Scattering on Batch 2 at 11.0.1 • Energies: 280eV (no dichroism?), 285.4eV, 287.5eV • Polarizations: Circular and Horizontal Linear • CCD postions: one higher in detector theta not used, yet • Eliot took some Scattering on both batches at 11.0.1 under different conditions (different slits & possibly different sample positions) • Similar energies and polarization • CCD positions: Moved CCD perpendicular to theta (“CCDY”) 2010-09 pBTTT Bilayer Processing Study
pBTTT NEXAFS PMMA σ* @ 288.4eV PBTTT σ* @ 287.5eV PBTTT π* @ 285.4eV 2010-09 pBTTT Bilayer Processing Study
STXM Dataon Batch 1 Brian’s Data at 5.3.2 Method Order Sorting Aperture Sample Zone Plate Diffractive Focusing Linear Detector xyz-stage 2010-09 pBTTT Bilayer Processing Study
UCSB PBTTT: Single Layers (285.4 eV) Annealed on SiO2 Annealed on OTS Different area • Difference between annealing on different substrates is subtle • Domains on As Cast seem less defined as expected 500nm 500nm 500nm As Cast on OTS 500nm 2010-09 pBTTT Bilayer Processing Study
Direct Comparison – Single Layers E=285.4 eV Annealed on SiO2 Annealed on OTS As Cast on OTS Didn’t see larger dark blobsas in March samples Sept 2010 Batch/Data 500nm 500nm 500nm March 2010 Batch/Data 500nm 500nm Kept a constant contrast ratio=1.4 (white/black) 2010-09 pBTTT Bilayer Processing Study
UCSB PBTTT/PMMA Bilayers As Cast (1D) 285.4 eV (PBTTT peak) 288.4 eV (PMMA peak) 390 eV (Thickness) 285.4 eV PMMA pBTTT/PMMA My Interpretation of the different areas 10μm 1μm 1μm 1μm 285.4 eV 2 x pBTTT/PMMA Spectra across edge Preannealed(1J) PMMA is here 500nm PBTTT is here too 285.4 eV 1μm 2010-09 pBTTT Bilayer Processing Study
UCSB PBTTT/PMMA Bilayers Pinholes in PMMA under-layer due to annealing? (No pinholes in unannealedbilayer – see prev. slide) Postannealed(1K) Zoom of previous images Focus Damage 1μm 1μm 500nm 500nm 285.4 eV 288.4 eV 285.4 eV 288.4 eV Pre & Post (1L) 1μm 1μm 500nm 1μm 285.4 eV 288.4 eV 285.4 eV 390 eV 2010-09 pBTTT Bilayer Processing Study
UCSB Bilayer Comparison As Cast (1D) Pre (1J) • Color scale the same for equal contrast • Post annealing the bilayer clearly affects the morphology more than annealing the PBTTT first • Postannealedbilayer has the most contrast and largest domains • Basically no domains in the unannealedbilayer 1μm 1μm Pre & Post (1L) Post (1K) 1μm 1μm 2010-09 pBTTT Bilayer Processing Study
STXM Dataon Batch 2 Hongping’s Data at 11.0.2 2010-09 pBTTT Bilayer Processing Study
Samples • Scattering samples (Silicon Native Oxide Substrate ~5-10nm) • A: Pre Annealed • B: Pre & Post Annlealed • C: As Cast • D: Post Annealed Goals • Observe the domains in bilayers • Observe the sensitivity of domains to the polarization of soft x-ray • Get chemical compositions of the samples Setup • 11.0.2: the zone plate they have for best resolution, slits: X/25/25 • 5.3.2: 240μ/25nm/95μ zone plate , slits: 50/25/25 microns, Nitrogen filter in. Energies: 285.4 eV(π*), 287.5 eV(σ*), 320 eV (min dirchroism) 2010-09 pBTTT Bilayer Processing Study
Comparing Samples (285.4 eV, Horiz. Pol.) As Cast (2C) Pre Annealed (2A) • As Cast Sample • Just transmission data Horiz-Pol (no thickness correction) • No obvious Features • Other Samples are ratio of Horiz/VertPolariazation to accentuate the domains • Post Annealed samples has largest/best-defined domains 500nm 500nm Post Annealed (2D) Pre & Post (2B) 500nm 500nm 2010-09 pBTTT Bilayer Processing Study
Comparing Sample Batches (285.4 eV) Preannealed As Cast Batch 1 Batch 2 Batch 1 Batch 2 1D 2C 2A 1J 500nm 500nm Post Annealed Pre & Post Annealed Batch 1 Batch 2 Batch 1 Batch 2 1K 1L 2D 2B Correlated domains? 500nm 500nm NOTE: Batch 1 samples not normalized to thickness variations. 2010-09 pBTTT Bilayer Processing Study
Scattering Data Method Sample Increasing q (theta) Decreasing size scale CCD Detector 2010-09 pBTTT Bilayer Processing Study
Data Processing • Sample-detector distance measured via AgBpwdr. diff. at 1.5keV • Beam Position imaged directly on CCD (detuned undulator & narrow entrance Hslit) • Random ~60ct bkgbtwn frames removed from each image (something in software?) • Normed to top left corner of image (red bar) • Readout, hot pixel, and residual light bkgsubtracted via dark image • Normed to I0 via Au grid upstream from slits • Can’t compare abs. intensities if change slits • Masked beamstop and CCD edges as well as any parasitic scatter from sample (black) -> • Integrated all data around full arc • To accentuate peaks, multiplied data by q^2 • Linear Dichroism: Took ratio of scattering data linear/circular polarization Frame Bkg Mask (Also tried right side & bottom) Masked Data 2010-09 pBTTT Bilayer Processing Study
Brian’s Scattering Data (Batch 2) • Improper Dark image & uncertainty in frame bkg makes data below 50nm unusable • Definite trend in Peak intensity with annealing conditions • Larger, more intense (pure) features with annealing 2010-09 pBTTT Bilayer Processing Study
Comparison to Eliot’s Data (Batch 2) • Eliot’s data had proper dark image (same dwell time) • Used same frame mask for norm • Beam ~10x dimmer probably due to slit cutoff • Smaller data range due to differnet beam stop/slits • Strange difference in the slope of the two datasets • Intensity feature agrees between the data sets (see next slide for quantitative comparison) 2010-09 pBTTT Bilayer Processing Study
Size at scattering max shown for Circular Polarizaton @ 285.2, (.4) 129nm 144nm As Cast 575nm 254nm Post Annealed Sample Batch 2 (Best Bkg Sub) Comparing All 360nm 338nm Sample Batch 1 (Poor Bkg Sub) Pre Annealed 379nm 252nm 240nm Pre & Post 2010-09 pBTTT Bilayer Processing Study
Results from the Comparisons • Different beamline conditions on the same sample gives fairly reproducible results • Same feature sizes • Exception is slope of data (unknown origin) • Two sample batches show significantly different feature sizes • Need to determine the origin and impact of this!! 2010-09 pBTTT Bilayer Processing Study
Batch 1: Scattering – FFT of STXM • Quick and dirty FFT of STXM images w/const. noise floor subtraction • STXM data at 285.4 eV (from slide 11) • Any film edges etc. not removed • Decent Correlation with Scattering data (not definitive) 575nm 144nm 379nm 2010-09 pBTTT Bilayer Processing Study
Scattering Ratio (Batch 2) • Data Normedto 1 at 100 nm • Reproducibility of trends very good between beamline setups on same sample within range of 80-600 nm • Definitely see size correlations toward ~1 micron 2010-09 pBTTT Bilayer Processing Study
Scattering Ratio: Batch Comparison • Different Sample batches don’t agree • Batch 1 shows very little polarization contrast!! 2010-09 pBTTT Bilayer Processing Study
Discussion • See increased dichroism>500 nm size scales in Batch 2 • Will attempt to take above 2μm size scale next time • 280eV potentially no low enough to eliminate dichroism • Will calculate cmplx refractive index based on NEXAFS & do at lower energy next time • Background overtakes data at 50 nm and below • Readout error causing data fall-off definitely the cause • Not sure best way to eliminate this problem (try tiled CCD images, longer dwells, more dark images, fix firmware/software data acquisition errors?) – working on all these routes • Batch 1 results don’t agree w/Batch 2 • Need to discuss differences in sample prep • Need a tie-breaker set & more firmly correlate mobility w/structure of the same samples • Bulk Structure measured here may not play main role in device performance – need to look at interface • Use Index-matched grazing incidence scattering 2010-09 pBTTT Bilayer Processing Study
Next Beamtime • New mechanically stabilized & computer controlled slits allow lower q-range (larger size scales) • New order sorter will purify beam energy • New software enhancements & constant two-person team will allow more efficient & successful data acquisition • Beam at bottom corner of CCD to extend high q-range (allow better frame norm too?) • Work on tiling CCD images to get even higher-q (need data at 30nm size scale – predicted domain size) • Multiple sample copies to check for reproducibility of data • Will Complete Index-Matched Grazing Incidence scattering to asses structure of interface 2010-09 pBTTT Bilayer Processing Study
Appendix 2010-09 pBTTT Bilayer Processing Study
Absolute Scattering Ratio • Trends similar • Significant difference in abs. ratio between the two datasets • Different slits block diff amounts of beam? 2010-09 pBTTT Bilayer Processing Study
Single Layers: Larger Range Gray Scale is normalized to dwell time Annealed on OTS Annealed on SiOx Domains here seem larger than on OTS! 1μm Don’t see dark blotches as seen in March samples (see Annealed OTS samples in March report below) 2010-09 pBTTT Bilayer Processing Study
Chabinyc PBTTT: Imaging domains using linear dichroism Brian Collins ALS Run March 8-13, 2010 2010-09 pBTTT Bilayer Processing Study
Goals/Samples • PBTTT is highly crystalline • Linear dichroism: domains absorb differently depending on crystal orientation • Use abs. contrast to image domain size (potentially 10s of nm) • Samples • Thin As Cast • Thin Annealed • OTS As Cast • OTS Annealed → What is OTS? 2010-09 pBTTT Bilayer Processing Study
Initial Spectra to find peaks S4: OTS Annealed 390eV 287.5eV 285.4eV 2μm 285.4eV Absolutely no damage Due to 2μm manual defocus 2μm 2010-09 pBTTT Bilayer Processing Study
S1: Thin As Cast Grid 1 390 eV 285.4 eV 285.4 eV Much more uniform But has some features Due to thickness variation 1μm 1μm 500nm 285.4 eV 285.4 eV 285.4 eV 5ms dwell Gives Error 500nm 200nm 200nm 1ms dwell Motor doesn’t allow for longer dwell times 2010-09 pBTTT Bilayer Processing Study
S1: Thin As Cast Grid 2 5μm 1μm Grid 3 500nm 500nm 1μm 2010-09 pBTTT Bilayer Processing Study
S2: Thin Annealed 390 eV 285.4 eV Again basically uniform 1μm 1μm 285.4 eV 285.4 eV 500nm 500nm 2010-09 pBTTT Bilayer Processing Study
S3: OTS As Cast Dwell Time = 4ms Could be stretching 285.4 eV 285.4 eV 285.4 eV Contrast Ratio (Hi-Low)/Avg = 0.21 500nm 500nm 200nm CR= 0.28 287.5 eV CR = 0.12 390 eV NOTE: Refocused for each energy 500nm 500nm 2010-09 pBTTT Bilayer Processing Study
S4: OTS Annealed 285.4 eV 285.4 eV 285.4 eV 1μm 500nm 200nm 285.4 eV 287.5 eV 390 eV After refocusing (Repeatable defocusing) 500nm 500nm 500nm 2010-09 pBTTT Bilayer Processing Study