# Introduction Some papers (review), structures and project targets. Experimental techniques and theory. >[!Todo] >- Spin coating parameters. >- Gelation of BTO solution and rheological properties. >- Rietveld refinement. >- Draw the map. # Experiments ## General lab info (draw the lab layout) >[!Notice] >Workflow and time estimation: ## Reactions We use a transition metal alkoxides sol-gel reaction to form a gel layer and pattern to desired structures. > [!Note]- > The classical, water aid reaction would be as following. For metal $\ce M$, the reaction is > 1. Hydrolysis > $\ce{M(OR)_{n} + mH2O -> M(OR)_{n-m}(OH)_{m} + mROH}$ > 2. Condensation > $\ce{2M(OR)_{n-m}(OH)_{m} -> (OR)_{n-m}M-O-M(OR)_{n-m} + H2O}$ or > $\ce{(OR)_{n-1}M-(OR) + (HO)-M(OR)_{n-1} -> (OR)_{n-1}M-O-M(OR)_{n-1} + ROH}$ > 3. Crosslinking > $\ce{(OR)_{n-m}M-O-M(OR)_{n-m} -> [3D cross-linked network]}$ > > For titanium acetate and our system, the reaction would be > 1. Hydrolysis > $\ce{Ti(OCH(CH3)2)4 + H2O -> Ti(OCH(CH3)2)3(OH) + HOCH(CH3)2}$ > 2. Condensation (Pathway 1: via water elimination) > $\ce{2Ti(OCH(CH3)2)3(OH) -> (OCH(CH3)2)3Ti-O-Ti(OCH(CH3)2)3 + H2O}$ > Condensation (Pathway 2: via alcohol elimination) > $\ce{Ti(OCH(CH3)2)3(OH) + Ti(OCH(CH3)2)4 -> (OCH(CH3)2)3Ti-O-Ti(OCH(CH3)2)3 + HOCH(CH3)2}$ > 3. 3D cross-linking > $\ce{(OCH(CH3)2)3Ti-O-Ti(OCH(CH3)2)3 -> [3D cross-linked Ti-O-Ti network]}$ > The weak acidic and nonaqueous environment provided by acetic acid decreases the first hydrolysis step. To further slow this process down, chelate agent acetylacetone is added. This helps to maintain the solution stability at room temperature. But it is suspected even at room temperature, the condensation and gelation will happen through some other routes. >[!Note] >The viscosity of acetic acid is 1.22 mPas, close to water and the value is quite low. The reaction is described in [[Influence of Precursor Chemistry on the Formation of MTiO3 (M = Ba, Sr) Ceramic Thin Films]]. ## BTO solution preparation ### Acetic acid and barium acetate mixing 1. Take the flasks for BTO solution mixing out. If they are not `BTO mixing specific` wash them with acetone and isopropanol. Find the acetic acid and barium acetate and take it out. 2. Use the pipette to transfer 10 mL acetic acid into the beaker. Add a stir bar inside. Once finished, use parafilm to seal the flask. The acetic acid is volatile, try not release the odor outside the fume hood. 3. Transfer the entire setup to the fume hood with analytical/micro balance. Take a weighting paper as the tray and try not contaminate the balance. Use a small spectacular to weight 0.511 g barium acetate, then **carefully** transfer the barium acetate into the flask and not getting them sticking on stuck on the neck of the flask. (I could try to use spatula to transfer the powder this time) 4. Once finished, use the parafilm to seal the flask again, then start stirring at 300 rpm, 60 degrees of Celsius for 1 hour. A conventional method for higher concentration of BTO solution is decreasing the acetic acid amount. In this part we just mixed the acetic acid and acetate, there is no actual reaction happening. ### Titanium isopropoxide preparation and sol mixing 1. Prepare ice bath in advance. Due to our small amount of reactants, we could just find a tray and fill it with DI water and ice. 2. Take a plastic bottle for later reactions and solution storage. **Carefully** transfer the stir bar from the flask to the plastic bottle. Use the plastic bar to perform the transfer. Avoid touching considering the corrosive nature of acetic acid. Rinse the stir bar with isopropanol, then dry with nitrogen flow, finally drop it inside the plastic bottle. 3. Add 425 microliter of acetylacetone, start stirring at 100 rpm. Then add 529 microliter of titanium isopropoxide drop by drop. This is an exothermal reaction, after this one could remove the ice bath. 4. Transfer the solution obtained in [[#Acetic acid and barium acetate mixing]] into the bottle, drop by drop. Adjust the stir bar to 3000 rpm, seal with parafilm. Leave it for further reaction. 5. After transfer all the solution into the plastic bottle, rinse the flask with acetone and isopropanol. Take special care on the neck part. >[!Note] >The concentration of the filtered solution is to be determined. This is directly related to the spin coating parameters. An easy way to see is take a glass slide, use the pipette to drip 5 mL solution onto it, place on the hot plate for drying and weight again. For estimation the temperature should be ~ 150 degrees of Celsius. For compensation variation, use XRD to determine the e ## Spin-coating of layers ### Substrate (wafer) cleaning 1. Find the wafer for substrate for spin coating and carefully take substrates out with (plastic) tweezer. Transfer the substrate into the sample box and label them properly. The label could be printed with the tag printer. 2. Take the sample box into the fume hood. Wash the substrate with acetone first, to remove the residual layer on them. This is done by placing the substrates inside the small beaker labeled with acetone and isopropanol, respectively. Fill the beaker with corresponding washing agent, place the substrate in and transfer the beaker into the sonicator. Sonicate for five minutes with acetone, rinse the substrate with isopropanol, then another five minutes with isopropanol. Take the substrates out after finishing and dry with the nitrogen flow. 3. Put the wafer back to the sample box for further processing. The waste liquid goes into the labeled beaker, if no other people uses it, pour the waste chemicals (for washing) into the waste liquid bucket (with the proper label). ### Plasma cleaning 1. Take one or more glass slides and place inside the plasma cleaner. Transfer the samples (substrates) from the sample box into the chamber. 2. Make sure the valve is closed (clockwise to close). Turn on the vacuum pump, then hold the door until the vacuum is established, which can be felt. 3. Once the vacuum is established, turn to maximum power. Check the vacuum chamber for plasma glow. If no glow can be seen, switch off the power knobs and turn it on again. Turn the cleaner off after 5 mins (?). Use the timer for time control. 4. Switch the pump off before opening the valve. After that open the door. Place the cleaned substrate back into the sample box for further use. >[!Note] >The plasma cleaning is for cleaning the attached hydrocarbons, as well as increase the surface energy of the substrate. This makes it easier to deposit layers on the substrate. ### Spin coating of layers 1. Turn the spin coater on and check its parameters. Select `Layers`. It should have a ramp up rate of 2000 rpm/s, then coating at 3500 rpm for 40 s, getting down with 2000 rpm/s. 2. Transfer the BTO solution with target concentrations into the vials with pipette. Use the syringe for this procedure. Choose the small syringe (1 mL) and the 220 nm filter. When transfer the solution from the bottle to the vial, try avoid touching (or contacting) to avoid contaminations. 3. The typical amount is 40 microliters per layer, prepare the pipette in advance. Place the substrate onto the spin coater. Turn the vacuum on and examine whether the vacuum is established. Move the substrate to the center and close the lid. During this procedure, always use the metallic tweezer, since the temperature could be significantly higher than the maximum that a plastic tweezer could resist. 4. Insert the pipette into the spin coater. Pipe the BTO solution onto the substrate and instantly press the start button. Wait until stop, then transfer the substrate onto a petri dish to eliminate the contaminations. Dry for 5 mins to have gelation happened before depositing next layer or annealing. The layer should have the same color if the layer thickness uniformity is good. >[!Note] >It is worth to examine the spin coating parameters (thickness and rpm, etc) and theory of spin coating. >The thickness is independent with the liquid amount as long as we have excessive solutions. ### Annealing with the hot plate The hot plate heats very fast, it takes only 2-3 mins to reach ~ 400 degrees of Celsius. 1. After every two layers (after drying), place the substrates onto the preheated hot plate. Use glass slides to avoid contamination. The annealing time is (at least) 5 mins. 2. After annealing, transfer the glass slides somewhere else and wait for 1 min before placing the substrate back to the sample box or petri dish since they are both plastic. ### Annealing in the furnace 1. Check the furnace setup. Log in use the password written on the tag. The ramp up rate should be 20 °C/min. The target temperature is 800 degrees of Celsius, then cool down with the same rate. 2. Place the substrate directly into the quartz tray, use the wire to move the tray to proper site. Start the furnace and anneal the samples. The whole process will take around 14 hours. ### Layer thickness for different concentrations | concentrations | per layer thickness (nm) | total thickness (nm) | number of layers | solution required (microliter) | | -------------- | ------------------------ | -------------------- | ---------------- | ------------------------------ | | 0.2 M | 31.5 | 504 | 16 | 640 | | 0.4 M | 58 | 464 | 8 | 320 | | 0.6 M | (90) | (540) | 6 | 240 | | 0.8 M | (120) | (480) | 4 | 160 | | 1.0 M | | | | | ## PDMS preparation ### Fluorinate master molds (only once) Follow the fluorination recipe. #### Clean master molds (optional, only when noticeable dirt exists) Rinse the molds with acetone, clean with isopropanol if needed. ### hPDMS preparation The mixing is done in the plastic sample box. 1. Almost all PDMS preparation related chemicals are stored in the specific drawer and not in the cabinet. Take agent A and B into the fume hood with the holder. They are typically stored inside the centrifuge tubes. Take required amount out with plastic pipettes (3 mL). The minima amount is 0.5 g each. Choose one grid for mixing and properly label it to avoid later misuse. The weighting is done on the electric balance (less accurate one). 2. Stir with a spatula for around 3 mins. 3. Degas in the vacuum oven. (It does not have to be turned on since we need no heating, just close the valve and turn the pump on) Turn the pump off after reaching 200 mbar and wait for another 2 mins. 4. Transfer the PMDS into the fume hood. Remove air bubbles with (small) pipette. Degas one more time if necessary. ### hPDMS spin coating 1. Turn on the new spin coater (specific for PDMS coating). Check the parameters (1000 rpm, 1000 rpm/s, 40 s). 2. Put a drop of hPDMS onto the center of master mold with plastic pipette (3 mL or 1 mL), wait for 1 min and use the pipette to ensure full coverage on the substrate, degas (release once 200 mbar vacuum is established) and remove air bubbles with pipette. 3. Transfer the master mold with hPDMS drop onto the spin coater. Start coating, after finished place the master mold onto the hot plate at 70 degrees of Celsius. The heating should be ~ 5 mins and no longer than 30 mins. 4. Clean the spin coater by wrapping, check the hole for vacuum. ### sPDMS preparation 1. Use the plastic bottle (the bottle could be reused, or if there are noticeable flakes or contaminations, find a new one). Add the base (does not mean it is a base!) agent first, which is typically stored inside the white plastic bottle (if refilling is required, use the metallic bottle inside the cardboard box), then add the curing agent (in the box, inside the small glass bottle), the ratio should be 10:1. For proper mixing, the minimum amount should be 30 g and 3 g, but for more substrates, follow the following table (does not have to match exact amount as long as ensure 1:1) | # of substrates | Base agent (g) | Curing agent (g) | | :-------------: | :------------: | :--------------: | | 1 | 30 | 3 | | 2 | 40 | 4 | | 4 | 50 | 5 | | 5 | 60 | 6 | 2. Have the bottle well closed and mix the two agents by hand. Then vortex it for (at least) 15 s. Repeat this procedure several times. This should be more than three times if the PMDS inside the bottle is for more than a single master mold. 3. Degas with the vacuum oven. The time should be ~ 4 mins. Do not exceed 5 mins. 4. Move it to the fume hood and remove the bubbles with a (small) pipette. This could take long, if residual air was trapped underneath, degas again for ~ 1 min. 5. Pour a big "drop" of sPDMS on the plastic petri dish, after dropping it should have a diameter of ~ 3 cm. Put the spin coated master mold on it (the coated part towards the bottom). Then press it with two pipettes. 6. Pour the rest of PMDS in and make sure the master mold is covered with ~ 4 mm PMDS. Degas again and remove bubbles. After that press the master mold again to ensure good contact. >[!Notice] >Thicker PDMS would facilitate later imprint since it is easier to hold. While too thick may be difficult to cut. 1. After finished, place the petri dish inside the oven (low temperature one, 65 degrees of Celsius) for around 24 h. ### Mold separation Wear a mask before separating the PDMS mold. This may help avoid contamination to the patterns. 1. Cut the PMDS **through** with the knife at four edges. There should be 5-10 mm distance between the cutting line and the master mold. This distance could be narrower if the master mold does not locates at the center. 2. Use a tweezer holding a corner of PDMS with the master mold underneath, ensure it will not slide or sketch. Carefully lift four cut edges off from the petri dish. 3. Massage the PDMS (now only the center part left on the petri dish) outside the master mold. Try to separate the PDMS. If it would not go down by itself, hold the PDMS by fingers and lift it off gently. 4. Now the PDMS should get separated. Use the blade to cut at four edges of the master mold as close as possible (< 1 mm). Then cut off the residual PDMS at the backside of the substrate. Carefully remove the PDMS flakes sticking on the substrate with tweezer. 5. Deform the PDMS a little bit and detach the master mold with tweezer. Place it back to the sample box. If there are PDMS flakes, remove them. 6. Trim the edges of the newly obtained PDMS mold. Put it to an empty petri dish and properly label them. ## Nanoimprint Before nanoimprint, degas the PDMS mold for half an hour. Since it consume quite some time, do this before coating the layer or cleaning of the substrate. Take the PDMS mold and the substrates after plasma cleaning **after** all other preparations are completed. 1. Ensure that the substrates are washed and the PDMS molds are placed inside the vacuum oven for degassing. For waveguide or any structures with layer underneath, the plasma cleaning should be 3 mins instead of 5 mins. 2. Prepare the spin coater. Turn it on and turn the vacuum pump on. Prepare the solution, filter required amount (the syringe is 1 mL) of BTO solution into the vial. 3. Select `imprint`, this recipe should have the parameters set as ramp up rate 1500 rpm/s and coating at 1000 rpm for 5 s. The time is very short, it is better do a run-through before actual operation. 4. After having everything prepared, take the PDMS from the oven and the substrate from the plasma cleaner. Put the substrate on the spin coater and turn the vacuum on. Get required amount of solution (25 microliter for waveguide, 20 microliter for metasurface). 5. Once the spin coating starts, pick the PDMS mold and prepare to imprint. Open the coater and align the mold to the substrate, then place the patterned surface on top of the substrate. Press a bit to ensure full contact (but if this is difficult, like gas bubbles exist, do not press more and just leave as it is). Then turn off the vacuum and transfer the substrate to the 70 degrees of Celsius hot plate for 2 hours. Find one (or more, depends on structures) copper weight and put it on the top of PDMS. ### Annealing with the hot plate 1. After 2 hours, take them off and release the weight and PDMS. Be careful if these three parts stick together. Use fingers (not tweezer) to separate the weight, which should be easier. For PDMS and the substrate, use tweezer if hard to remove. Bend the PDMS a little bit. Try not destroy the pattern or break the substrate. 2. Turn the programmable hot plate (Thermo scientific) on, check the heating profile (5 mins 250 degrees of Celsius; 5 mins 350 degrees of Celsius; 10 mins 405 degrees of Celsius) and start annealing. 3. After 20 mins, remove the slide glass from the hot plate and transfer to somewhere else (to cool it down). After that transfer the substrates to the sample box or for further annealing. ### Annealing in the furnace Same as the previous one. ## XRD The data is save in the server. ## SEM The data is saved in the server. ## Solution concentration and stability From the XRD for 0.2 and 0.4 M solutions annealed at 600 degree Celsius, we already see that we failed to form crystalized BTO. It could be possible that our samples are too thick. But this only happen at 600 degrees while 800 degrees are crystalized BTO. This could also be aging of solutions. The 0.2 M and 0.4 M solutions are more than 20 days before layer preparations. It is fairly possible that some carbon contents formed during this period while they are thermodynamically stable at 600 degrees, or failed to fully decompose at such temperature. The reason are to be investigated. Therefore, a new batch of solutions are prepared, and will be coated for less layers. In the same time, the same solution will be kept for 20 days, with the same amount of layers (8 layers for 0.2 M BTO) repeat the coating. I still do not know whether there will be particles larger than 220 nm. It is possible to try, but the accuracy of the scale might not support such operation. ## Spin-coating parameters adjustment Current issue is that the structures printed are too thick. Due to uneven thermal expansion rate and shrinkage during solvent lose and recrystallization, inside the material internal stress is induced and cracks are formed. This in theory can be optimized by controlling the thickness of spin-coated layers. Current spin coating parameters are: 1000 rpm for 5 s, ramp up rate 1500 rpm/s. The solvent in our solution is acetic acid, whose viscosity is low. The evaporation rate is tested and should be small compared with our coating time (~ 10 seconds). Therefore, the layer thickness should only be affected by the spin rate. An empirical relation between angular velocity $\omega$ and layer thickness $h$ is $h \propto \omega^{-\beta}$ This $\beta$ value ranges from 0.5 to 0.7, typically. This makes the thickness having the trend like the figure below. ![[spin_coating_thickness_vs_speed_small.png]] Here we assumed the $\beta=0.6$. To ensure the consistency, during the adjusting there should also be a *fake* PDMS block on the coated later, and go through the following procedures. (?) >[!Notice] >Ensure the spin coating is not too bad for imprinting. So check the morphology before putting the PDMS. The volume should match the metasurface result (20 mL), and the pressing weight should be 3 blocks. | Name | Mold | Concentration | Spin speed | thickness (actual) | thickness (normalized) | Comments | | ---- | ----------- | ------------- | ---------- | ------------------ | ---------------------- | --------------------------------------------------------------------------------------------------------------------------------- | | | Empty | 0.2 M | 1000 | | | Bad imprinting. Possibly due to dirty PDMS. Fast cooling after removal from programmable hot plate at around 350 degrees Celsius. | | | WG v2 (old) | 0.6 M | 1000 | | | Not very uniform. Cracks are noticeable (by eyes). Fast cooling. | | | MS v4 | 0.6 M | 1500 | | | Uniform at the center. Cracks are noticeable. Fast cooling. | | | Empty | 0.6 M | 2000 | | | | | | WG v1 | 0.6 M | 2000 | | | | | | Empty | 0.6 M | 2500 | | | The layer right after spin coating is not very uniform. | | | Empty | 0.6 M | 3000 | | | It seems that the layer dries faster than those of lower spinning rate. | The cracks are noticeable when removing from high temperature hotplate. This is possibly due to fast cooling of the sample with relative thick layers. The internal stress is large and the cracks are formed. For samples with higher spinning rate. I will try to cool down slowly on the HP, to check whether we still have those cracks. ## Evaporation test This was done for 0.2 M BTO solution coated on a $1.5\times 1.5$ p-Si substrate. The time and the weight of the solution were noted after first 10 seconds. The solution volume is 20 microliters (same as MS) and the substrate is plasma-treated. No spin coating, only drop casting. >[!Note]- >The result is shown as following. > > | Time (s) | Mass ($\mu g$) | Time (s) | Mass ($\mu g$) | Time (s) | Mass ($\mu g$) | Time (s) | Mass ($\mu g$) | > | -------- | ---------------- | -------- | ---------------- | -------- | ---------------- | -------- | ---------------- | > | 10 | 22.31 | 260 | 14.06 | 510 | 7.06 | 760 | 1.95 | > | 20 | 22.14 | 270 | 13.81 | 520 | 6.83 | 770 | 1.79 | > | 30 | 21.85 | 280 | 13.52 | 530 | 6.57 | 780 | 1.64 | > | 40 | 21.5 | 290 | 13.18 | 540 | 6.36 | 790 | 1.54 | > | 50 | 21.14 | 300 | 12.85 | 550 | 6.13 | 800 | 1.38 | > | 60 | 20.77 | 310 | 12.59 | 560 | 5.86 | 810 | 1.22 | > | 70 | 20.46 | 320 | 12.3 | 570 | 5.68 | 820 | 1.13 | > | 80 | 20.1 | 330 | 11.99 | 580 | 5.4 | 830 | 1 | > | 90 | 19.8 | 340 | 11.66 | 590 | 5.19 | 840 | 0.9 | > | 100 | 19.42 | 350 | 11.42 | 600 | 4.95 | 850 | 0.79 | > | 110 | 19.06 | 360 | 11.14 | 610 | 4.74 | 860 | 0.69 | > | 120 | 18.71 | 370 | 10.83 | 620 | 4.53 | 870 | 0.61 | > | 130 | 18.37 | 380 | 10.56 | 630 | 4.32 | 880 | 0.54 | > | 140 | 18.01 | 390 | 10.25 | 640 | 4.1 | 890 | 0.46 | > | 150 | 17.64 | 400 | 9.99 | 650 | 3.9 | 900 | 0.39 | > | 160 | 17.36 | 410 | 9.74 | 660 | 3.69 | 910 | 0.34 | > | 170 | 17 | 420 | 9.44 | 670 | 3.51 | 920 | 0.28 | > | 180 | 16.7 | 430 | 9.13 | 680 | 3.34 | 930 | 0.23 | > | 190 | 16.31 | 440 | 8.84 | 690 | 3.15 | 940 | 0.19 | > | 200 | 15.98 | 450 | 8.6 | 700 | 2.97 | 950 | 0.17 | > | 210 | 15.7 | 460 | 8.36 | 710 | 2.78 | 960 | 0.12 | > | 220 | 15.34 | 470 | 8.08 | 720 | 2.6 | 970 | 0.08 | > | 230 | 15.06 | 480 | 7.82 | 730 | 2.42 | 980 | 0.07 | > | 240 | 14.72 | 490 | 7.59 | 740 | 2.28 | 990 | 0 | > | 250 | 14.44 | 500 | 7.35 | 750 | 2.12 | 1000 | 0 | > These results indicate for our short operation time (without considering the difference for spin coating), typically less than 30s, the solution will have 97% mass remaining. This should not have no significant result on layer thickness. ## Experiment schedules The target is to create as many samples as possible. The entire fabrication process can be separated into five relatively independent parts. 1. **PDMS molds preparation**. This could be done without solution and any layer characterizations, but master molds are required. (~3 h and overnight in 65 degrees oven) 2. **Layers**. This can be done without PDMS and master molds, only solution is needed. Only one sample is required for each concentration. For high concentration solutions, this is better be finished right after the solution is prepared. (~1 h and overnight for annealing) 3. **Nanoimprints**. This requires PDMS, layer characterization and solutions. Similarly, for high concentration solutions, this is better be finished right after the solution is prepared. (~4 h and overnight for annealing) 4. **Solution preparation**. This is independent to all other procedures. But for high concentration solutions, they should be used right after the synthesis. Also, I should analyze the content of the formed solution. (~2 h and overnight for stirring) 5. **Characterizations**. This include XRD and SEM. Characterizations should first be done for layers, XRD first, and then SEM. After SEM of layers and confirming that the layer quality is okay (no much cracks), nanoimprint should be done. But for high concentration solutions, this should be done together to avoid aging. (depends on the amount of samples and data points required) > [!Note]- > | | Monday | Tuesday | Wednesday | Thursday | Friday | > | --------- | ------------------- | ------------------------------------------- | ------------------------------------------------ | ------------------------------------------- | ------------------------------------------------ | > | Morning | PDMS (WG v1, MS v2) | *(XRD of 0.6 M layers)* | *(XRD/SEM of 0.6 M layers)* | *(XRD/SEM of 0.6 M layers)* | *(XRD/SEM of 0.6 M layers, SEM of metasurfaces)* | > | Afternoon | | Master mold separation, PDMS (WG v1, MS v2) | Nanoimprinting<br>(4-6 substrate, 0.6 M), layers | Master mold separation, PDMS (WG v1, MS v2) | (Take the sample out from the furnace) | > Week 4 > > | | Monday | Tuesday | Wednesday | Thursday | Friday | > | --------- | ------------------------------------------------ | ------------------------------------------------------------- | ------------------------------------------------ | ---------------------------------------------------------- | ---------------------------------------------------------- | > | Morning | *(XRD/SEM of 0.6 M layers, SEM of metasurfaces)* | *(XRD/SEM of 0.6 M layers, SEM of metasurfaces)* | *(XRD/SEM of 0.6 M layers, SEM of metasurfaces)* | *(XRD/SEM of 0.6 M and 0.8 M layers, SEM of metasurfaces)* | *(XRD/SEM of 0.6 M and 0.8 M layers, SEM of metasurfaces)* | > | Afternoon | Master mold separation, PDMS (WG v1, MS v2) | Master mold separation, PDMS (WG v1, MS v2), Solution (0.8 M) | Layers, nanoimprint (4-6 substrate, 0.8 M) | | (Take the sample out from the furnace) | > Week 5 > # Data analysis ## XRD I'll try to use Fullprof to refine the data. But before that I'll first find the proper `.cif` file for our structures. `.cif` is ready. I may perform the refinement on Nov 28. It is difficult to obtain reasonable data, I will retry on Dec 03. Rietveld refinement was done for 0.6M 800 degree Celsius. The proportion of cubic and tetragonal phases are obtained. Some contents cannot be identified appear in the diffraction profile and introduce peaks near $2\theta=28\degree$. Some papers suggests they are Ba2Ti2O5CO3. We do not have the `.cif` file or any crystallographic information about this content. Next step: add different results as patterns, since they shear the same instrumental data. ## SEM I'll use `trainable Weka segmentation` to find pores and calculate porosity (there could be a factor, check this later). It would be great if using contrast/edges or/and other machine learning segmentations to find grain boundaries and get quantitative distribution of grains. From the dielectric constant measurements one may get the porosity too, this may be used for crosscheck for SEM value. The code for solving it is `maxwell_garnett_porosityeval.py`. SEM for 0.6 M, 800 degree Celsius metasurfaces are done. Cracks appear on the substrate BTO layer (namely, between the pillars and the silicon substrate). The target now is to control the morphology of the imprinted samples and ensure that they do not crack and form uniform structures without scratches and cracks. This will first be optimized by [[#Spin-coating adjustment|adjusting the spin coating parameters]]. ![[MS_BTO_003_02.jpg]] Here is an image for r390 area, 800 degrees Celsius annealed, 0.6 M BTO solutions. Besides the noticeable cracks, pillars are distorted and no longer have circular shapes, but diamond shapes. The shrinkage for 0.2M and 0.4M solutions are 50% and 42%, respectively. For this 0.6 M sample, the shrinkage is ~38%, if take the short diagonal as the diameter.