Synthesis and self-assembly of high-χ poly(4-tertbutylstyrene)-block-poly(2-hydroxyethylmethacrylate)

Block copolymers (BCPs) are known to phase separate into morphologies with feature-to-feature spacings (pitch) as low as 5 nm.^1,2 By using chemically preferential (chemoepitaxy) or topographic (graphoepitaxy) prepatterns on a substrate, thin film BCP microphase separated features can be guided to form desired patterns.^3,4 By removing one of the blocks in the BCP via an etch, a template can be produced with an etch-transferable pattern to the substrate.⁵ Directed self-assembly (DSA) of BCPs is viewed as an extension of optical lithography since optical lithography would be used to pattern the underlayers necessary to direct the self-assembly of the BCP and the BCP's natural pitch would be small enough to make the patterns that conventional optical lithography could not. As of 2017, 10 nm node chips have become commercially available, meaning that the smallest printed feature on the chip is 20 nm in width.⁶ Future expected node sizes are 7 and 5 nm, meaning that the pitch between features will need to be 14 and 10 nm, respectively.⁶ DSA-BCPs have the potential to meet these size specifications through microphase separation of the BCP's two blocks. A BCP's pitch is dependent on the Flory–Huggins interaction parameter, χ, and the degree of polymerization, N.⁷ To reach low pitches, a low N yet a high enough χ is required such that the product χN remains above the order–disorder transition (ODT).⁸ 

In order to achieve this, considerable amounts of research have been conducted to increase the library of high-χ BCPs.^1,9,10 Ideally, a “high-χ” BCP will have a χ larger than that of polystyrene-b-polymethyl methacrylate (PS-b-PMMA, χ = 0.03–0.04),¹¹ both blocks will have relatively low glass transition temperatures (T_gs), and the BCP will not require a topcoat to form perpendicular features at its surface. High-χ BCP materials have previously been synthesized such as poly(cyclohexylethylene)-b-PMMA (PCHE-b-PMMA; χ = 0.17 at 160 °C),¹ poly(trimethylsilystyrene)-b-poly(d,l-lactide) (PTMSS-b-PLA; χ = 0.41 at 160 °C),¹² and poly(dihydroxystyrene)-b-PS (χ = 0.7 at 170 °C).¹³ While a high χ means that there is a strong driving force for phase separation to occur, if the BCP's chains are kinetically trapped, then phase separation may take extraordinarily long times. Below the T_g of a given polymer, the polymer chains are very stiff with little mobility, while above the T_g, those chains have high mobility throughout their bulk phase.¹⁴ Since most polymers have degradation temperatures near 300 °C, low T_gs are required so that the proper mobility is given to the BCP chains without the occurrence of degradation. A topcoat is a neutral layer at the free surface that is needed when one block has such a strong interaction with the free surface that it forms a thin layer near the free surface covering possibly perpendicular features beneath it.¹⁵ A topcoat is unattractive due to the increased number of processing steps to coat and remove the layer at the surface. However, as χ increases for a BCP, the difference in cohesive energy between the two blocks can become so great that one block wets the free surface, thus requiring a topcoat. Previously, it was observed that PS-block-poly(hydroxyethylmethacrylate) (PS-b-PHEMA) had an estimated χ = 0.37 and was able to form 15 nm pitch lamellae, visible via SEM without the need of a topcoat.¹⁶ In order to reach higher χ values, one strategy is to make the PS-block more hydrophobic, thus increasing the strength of the unfavorable interaction between it and PHEMA.¹⁷ This work will describe the synthesis, characterization, and self-assembly of poly(4-tertbutylstyrene)-block-PHEMA (PtBS-b-PHEMA) as a new high-χ BCP. Small angle x-ray scattering (SAXS) was used to discern the pitch of various molecular weight samples, and atomic force microscopy (AFM) and SEM were used to image thin film samples.

Monomers tertbutylstyrene (tBS) and hydroxyethylmethacrylate (HEMA) were purchased from Sigma Aldrich. Tertbutyl-dimethylsilyl chloride (TBDMS-Cl), 4,4′-dinonyl-2,2′-dipyridyl (dNbpy), ethyl bromoisobutyrate (EBiB), copper chloride, toluene, sec-butyl lithium (sec-BuLi), lithium chloride, tertbutyl ammonium fluoride, diphenylethylene, n-butyl lithium, dibutyl magnesium, hexamethyldisilizane, acetoxystyrene, propylene glycol monomethyl ether acetate (PGMEA), dimethyl formamide (DMF), and calcium hydride were also purchased from Sigma Aldrich. Unstabilized tetrahydrofuran (THF), dichloromethane, methanol, acetone, ethyl acetate, and magnesium sulfate were purchased from VWR. THF was run through a solvent purification still prior to use that lowered its water content to 30–50 ppm. Alumina (80–200 mesh) and triethylamine were purchased from Fischer Chemical. Hydrochloric acid (12M) was purchased from Acros Organics. The crosslinking agent for underlayers, 4-vinylbenzocyclobutene (BCB), was purchased from BOC Science. Deuterated chloroform and deuterated dimethyl sulfoxide (DMSO) for NMR solvents were purchased from Cambridge Isotope Laboratories. Triphenyl sulfonium hexafluoroantimonate (TPS-SbF₆) was purchased from Midori Kagaku, Ltd. All materials are used as received unless otherwise stated.

Prior to synthesis, a TBDMS protecting group was placed on the HEMA monomer to shield the propagating anionic chain end from the hydroxyl group of the HEMA monomer (Fig. 1). HEMA monomer was first run through an alumina column to remove its inhibitor and then dried. A typical protection reaction used 13.5 ml of HEMA, 20 ml triethylamine (1.29M equivalent to HEMA), 20.2 g of TBDMS-Cl (1.21M equivalent to HEMA), and 100 ml of dichloromethane. The reaction was allowed to stir for 24 h at room temperature. Afterwards, the product was washed with a 3% solution of hydrochloric acid three times, followed by three washings with deionized water. The solution was then poured over magnesium sulfate and dried using rotary evaporation leaving behind a dark yellow product. Figure 2 shows the NMR of the final product in deuterated chloroform with peaks at 0.15 ppm (h₁, six protons) and 0.98 ppm (h₂, nine protons) denoting the dimethyl and tertbutyl protons of TBDMS, respectively, while the alkyl peaks of HEMA are seen at 4.2 ppm (h₃, two protons) and 4.0 ppm (h₄, two protons).

For atom-transfer radical polymerization (ATRP) of PtBS-b-PHEMA, tBS was run through an alumina column prior to use to remove the inhibitor. First, the HEMATBDMS monomer was polymerized using the following procedure. HEMATBDMS (4.74 g), dNbpy (390 mg), toluene (2 ml), and EBiB (67.0 μl) were placed inside a 50 ml Schlenk flask with a small stir bar. Next, copper chloride (45 mg) was added and the flask was sealed with a septum seal and quickly placed in liquid nitrogen. Three freeze–pump–thaw cycles were given to the flask, backfilling with argon after each thaw. The flask was then placed in an oil bath at 75 °C and stirred for 3 h. Afterwards, the product was quenched in an ice bath and opened to the air. The product was then dissolved with a small amount of dichloromethane, run through an alumina column to remove the copper chloride, dried, and then precipitated in cold methanol. The product was then allowed to dry in a vacuum oven overnight at 50 °C.

A typical procedure for adding the second block, tBS, is as follows. PHEMATBDMS (3.10 g), tBS (2.00 g), dNbpy (390 mg), and toluene (9 ml) were added to a 50 ml Schlenk flask with a stir bar. Next, copper chloride (45 mg) was added to the flask and it was immediately septum sealed and placed in a liquid nitrogen bath for three freeze–pump–thaw cycles. The flask was then placed in an oil bath at 125 °C for 24 h. Afterwards, the reaction was quenched in an ice bath and opened to the air. The product was run through an alumina column to remove the copper chloride, dried, and then precipitated in cold methanol. Finally, the product was dried in a vacuum oven overnight at 50 °C. Typical final polydispersities (PDI) of BCPs via this method were approximately 1.3. Due to this, only a few samples were made via ATRP, with the majority of the BCP samples being synthesized via anionic polymerization.

Anionic polymerization requires stricter control over the level of impurities in its reactants. A typical procedure for monomer purification is as follows. All glasswares mentioned here are 100 ml round bottom flasks and, prior to being filled, have been flame dried and backfilled with argon three times. 4 ml of 1.0M dibutyl magnesium in heptane was injected into a distillation flask with a small stir bar. Vacuum was then pulled on the flask to remove the heptane. The tBS monomer (10 ml) was taken from its container, as is, and placed into a dry, clean distillation flask. The flask was given three freeze–pump–thaw cycles, backfilling with argon after each thaw. The tBS monomer was then vacuum distilled at 80 °C and 0.1 Torr pressure into the flask with dry dibutyl magnesium and allowed to stir for 30 min. Afterwards, the tBS monomer was vacuum distilled from the dibutyl magnesium flask and into a clean, dry flask. The purified monomer was then transferred to a flame dried, septum sealed vial for storage using a cannula and packed under argon. Purified tBS monomer was then stored in the fridge.

HEMATBDMS was purified using a slightly different procedure. HEMATBDMS (12 ml) was added to a flask with calcium hydride (about 2 g) with a small stir bar and allowed to stir overnight. The monomer was then vacuum distilled at 90 °C and 0.1 Torr into a clean, dry flask. The monomer was then transferred to a flame dried, septum sealed vial for storage using a cannula and packed under argon. Purified HEMATBDMS monomer was then stored in the fridge.

Prior to purification of diphenylethylene, 1.0M n-butyl lithium in heptane (approximately 3 ml) was injected into a dry, clean flask with a stir bar and vacuum dried to remove the heptane. Diphenylethylene (15 ml) was then injected into a dry, clean flask and given three freeze–pump–thaw cycles, backfilling with argon after each thaw. Diphenylethylene was then vacuum distilled at 95 °C and 0.1 Torr into the flask with n-butyl lithium. After stirring for 30 min, a dark red solution was present, signifying low levels of impurities. Diphenylethylene was then vacuum distilled into a clean, dry flask, transferred into a flame dried septum sealed vial, and packed with argon. Purified diphenylethylene was stored in the freezer.

A typical anionic polymerization procedure for PtBS-b-PHEMATBDMS is as follows. A 500 ml flask with a stir bar is placed in a vacuum oven and exposed to hexamethyldisilizane overnight at 220 °C. The reactor is then placed into a glovebox, charged with a small amount of lithium chloride, capped with a septum seal, and removed from the glovebox. The flask is then flame dried three times under vacuum, backfilling with argon between each cycle. Next, the reactor is attached to a purification still and THF is dispensed into it followed by one freeze–pump–thaw cycle. The reactor is then chilled in an acetone/dry ice bath (−78 °C), allowed to stir, and backfilled with argon. A 1.4M sec-BuLi solution in hexane was then removed via glass syringe from a container stored in the glovebox. sec-BuLi was then added to the chilled THF, turning the solution bright yellow. The color mentioned here is used to denote the presence of specific anions and thus is an indication of proper purity levels. The purified tBS monomer was then added via syringe, turning the solution bright orange, and allowed to stir for 60 min. Next, diphenylethylene was added via syringe, turning the solution bright red, and allowed to stir for 60 min. Finally, purified HEMATBDMS was added via syringe, the solution turned colorless, and was allowed to stir for 180 min. During this time, methanol was placed into a small vial and given three freeze–pump–thaw cycles. Methanol was then added to the solution at the end of the reaction to terminate the propagating chain. Typical PDIs of BCPs using this method were between 1.03 and 1.10.

Deprotection of PtBS-b-PHEMATBDMS into PtBS-b-PHEMA is shown in Fig. 3 and was done using the following procedure. PtBS-b-PHEMATBDMS was placed into a small round bottom flask with a stir bar and dissolved in dichloromethane. Tertbutyl ammonium fluoride was added such that it was in 2M excess compared to the amount of TBDMS in the BCP chain. The solution was then allowed to stir for 24 h at room temperature. Next, the solution was washed with deionized water three times and dried using a combination of rotary evaporation and a vacuum oven. In cases where dead PtBS homopolymer is present, the BCP is dissolved in methanol and then washed with hexane to extract out the majority of homopolymer.

Figure 4 shows the two different underlayers explored in the search for a neutral underlayer for thin films of PtBS-b-PHEMA to form perpendicular features on. For each random copolymer underlayer, BCB was used as a thermal crosslinking agent. Each underlayer was polymerized via radical polymerization.

The first underlayer explored was PHEMA-tetrahydropyran-r-BCB (PHEMA-THP-r-BCB) [Fig. 4(a)]. THP-protected HEMA was synthesized by reacting HEMA with a 2M excess amount of 3,4-dihydro-2H-pyran and 0.2M equivalent of trifluoroacetic acid in 5 ml of ethyl acetate at room temperature for 18 h. The NMR of the protected HEMA-THP monomer is shown in Fig. 5(a) where the alkyl protons (a₂ and a₃, total of 4) of the HEMA are shown at 3.8 ppm (two) and 4.3 ppm (two) and the single proton between the oxygen of HEMA and the oxygen of THP (a₁) is found at a shift of 4.6 ppm. PHEMA-THP-r-BCB was synthesized by dissolving HEMA-THP, BCB, and azobisisobutyronitrile (AIBN) in THF in a Schlenk flask with a stir bar. The vessel was given three freeze–pump–thaw cycles, backfilling with nitrogen after each thaw. The solution was then placed in an oil bath at 75 °C and stirred for 2 h. Afterwards, the solution was quenched, opened to the air, and precipitated in cold methanol. Figure 5(b) shows an example of the NMR for a typical PHEMA-THP-r-BCB underlayer with the alkyl protons (a₂ and a₃, four total protons) at 3.8 ppm (two) and 4.3 ppm (two) of PHEMA-THP. The phenyl protons of BCB are found at a chemical shift of 6.5–6.9 ppm (a₄, three total protons). These peaks are used to determine the relative amounts of PHEMA-THP and BCB in the random copolymer.

The second underlayer explored was isopropyloxycarbonyl-r-BCB [iPOC-r-BCB; Fig. 4(b)]. Typical synthesis for iPOC-r-BCB is a three-stage reaction. In the first step, acetoxystyrene, BCB, and AIBN are dissolved in toluene, given three freeze–pump–thaw cycles (backfilled with nitrogen), and then stirred at 75 °C for 24 h. The reaction was then quenched, precipitated in cold hexane, and dried in a vacuum oven. NMR spectra identify acetoxystyrene by the acetyl protons (2.2 ppm, three protons) while the BCB is identified by the butene protons (3.0 ppm, four protons). The product was then dissolved in dioxane and hydrazine hydrate was added slowly, dropwise, while stirring for 12 h to deprotect the poly(acetoxystyrene-r-BCB) into poly(hydroxystyrene-r-BCB) (PHOST-r-BCB). The product was then precipitated in deionized water and dried in a vacuum oven. NMR spectra of PHOST-r-BCB identify the PHOST by the broad hydroxyl proton at 8 ppm (one proton) and the BCB by its four protons at 3.0 ppm as previously mentioned. PHOST-r-BCB was then dissolved in ethyl acetate with triethylamine. Next, iPOC-Cl was added dropwise to the solution and allowed to stir at room temperature for 24 h. The product was then washed with 3% hydrochloric acid twice, poured over magnesium sulfate, and then precipitated in cold hexanes. Figure 6 shows an NMR of iPOC-r-BCB where the isopropyl protons of iPOC are shown at 1.4 ppm (b₃, six protons) and the butene protons of BCB are shown at 3.0 ppm (b₂, four protons).

Characterization of the bulk BCP was done using gel permeation chromatography (GPC), NMR, infrared (IR), and SAXS. For NMR, while the protected PtBS-b-PHEMATBDMS was able to dissolve in deuterated chloroform, PtBS-b-PHEMA was not due to the unfavorable interactions between the PHEMA block and chloroform. Instead, a mixture of chloroform and DMSO (75:25 by volume) was needed to dissolve PtBS-b-PHEMA. SAXS data were measured using the Xenocs Xeuss 2.0, Line Eraser and a Pilatus 300 K SAXS detector. Bulk SAXS samples were annealed at 150 °C for 12 h prior to their measurement. Infrared spectroscopy was taken using a Thermo Nicolet IS10 under attenuated total reflectance mode using an MCT detector.

Thin film thickness was measured via ellipsometry. Underlayer preference was measured using the water contact angle of PtBS, PHEMA, and the underlayer. BCP thin films were annealed under vacuum at temperatures between 150 and 210 °C. Imaging of thin film morphology was done using a Zeiss Ultra60 FE-SEM and a Veeco ICON AFM with an AFM tip radius of 8 nm. Etch contrast between the blocks was done using a Harrick Plasma Cleaner. Samples were placed in a chamber, put under vacuum, and O₂ plasma was generated from the air at a radio frequency (RF) of 10.5 W.

Figure 7 shows the structure of PtBS-b-PHEMA. The four phenyl protons of PtBS (a₁) and the four alkyl protons of PHEMA (b₁ and c₁) are used to determine the relative amounts of each block in the BCP. Since these three sets of protons do not alter due to the deprotection, their chemical shifts are relatively unchanged in the NMRs of PtBS-b-PHEMATBDMS and PtBS-b-PHEMA. Figures 8 and 9 show the GPC traces and NMRs of PtBS-b-PHEMATBDMS and PtBS-b-PHEMA, respectively. The peaks for a₁ are found along the chemical shift of 6.2–7.3 ppm, while the shifts for b₁ and c₁ are found at 4.1 and 3.9 ppm, respectively. All PDIs of PtBS-b-PHEMA BCPs were measured as less than 1.13. The volume fraction of PHEMA in the BCP samples ranged from 0.28 to 0.52.

Infrared spectroscopy, under attenuated total reflection mode, of the protected and deprotected BCP was done to ensure that the TBDMS group had in fact been deprotected and a hydroxyl group had formed in the sample. Figure 10(a) shows the IR of PtBS-b-PHEMATBDMS with the absence of a hydroxyl peak and the appearance of silicon-carbon bonds, albeit in the fingerprint region of the scan. Figure 10(b) shows the IR of PtBS-b-PHEMA with a large peak at roughly 3500 cm⁻¹ denoting the presence of hydroxyl groups in the BCP. There is a disappearance of the silicon-carbon peaks in the fingerprint region as well. The combination of NMR and IR, and the GPC showing a relatively unchanged molecular weight distribution, is evidence that the deprotection of PtBS-b-PHEMATBDMS to PtBS-b-PHEMA was a success.

SAXS measurements were taken of bulk BCP samples to determine the morphology and the pitch of microphase separated features. Bulk samples were prepared by first creating a mold for a disk with a 5 mm diameter and a 1 mm thickness. BCP powder was placed in the mold and dissolved with a few drops of solvent. The mold was then placed in a vacuum oven at 90 °C for 30 min to remove the majority of the solvent. The temperature was then raised to the desired annealing temperature (150–210 °C) and allowed to anneal for 12 h. The BCP disk was then peeled off the aluminum mold and SAXS was performed. Examples of the SAXS profiles of PtBS-b-PHEMA are shown in Fig. 11.

The position of the primary peak (q*) of a scattering profile gives information on the pitch of the phase separated features, while the reflections of the q* give information on the morphology. The majority of samples showed a morphology of either hexagonally close packed cylinders (q* reflections at 3^0.5·q*, 2·q*, 7^0.5·q*, 3·q*, etc.) or lamellae (q* reflections at 2·q*, 3·q*, 4·q*, etc.).¹⁸ The more reflections present, the more ordered the features are. Figure 11(a) shows a cylinder forming sample (φ_PHEMA = 0.36) with a pitch of 10 nm and reflections at 3^0.5·q* and 2·q* while Fig. 11(b) shows an example of a lamellae forming sample (φ_PHEMA = 0.52) with a pitch of 6.9 nm and a single reflection at 2·q*. The sample in Fig. 11(b) represents the lowest molecular weight sample produced with an N = 22. Despite this, the BCP was still able to phase separate, making it unable to determine this BCP's χ. However, the fact that the PtBS-b-PHEMA was able to form features with a half-pitch less than 3.5 nm is a clear indication of a high-χ material. A lower limit for the estimate of χ of PtBS-b-PHEMA, using the ODT as χN = 10.5, would be approximately 0.45, giving PtBS-b-PHEMA a greater driving force for phase separation when compared to other high-χ BCPs like PCHE-b-PMMA (Ref. 1) or PTMSS-b-PLA.¹² 

Figure 12 shows a plot of the experimentally measured pitch of lamellae forming PtBS-b-PHEMA samples annealed at 150 °C for 12 h versus their degree of polymerization. A trendline was fit to these data (dashed line in Fig. 12) to determine how the pitch scales with N. The exponent calculated by the trendline was 0.802. Given the weak reflection in Fig. 11(b) of the N = 22 sample, it should be expected that N = 22 is in the proximity of the ODT. This would mean that the points shown in Fig. 12 are expected to be in the intermediate segregation regime where the expected scaling of the pitch is N^0.8–N^1.0.^19,20 This matches well with the scaling calculated by the trendline in Fig. 12.

The crosslinking procedure used was the same for all underlayers. After coating, the films were heated to 250 °C for 10 min causing the BCB units to crosslink with one another.^21,22 The water contact angle for PtBS and PHEMA was measured as 104° and 38°, respectively. A “neutral” underlayer here is described as one that provides perpendicular orientation for lamellae forming BCPs and thus has no strong preference for either PtBS or PHEMA. First, estimates would suggest that a neutral underlayer would have a water contact angle near 71°, the average of PtBS and PHEMA's water contact angle. A random copolymer, PtBS-r-PHEMA-r-BCB, was originally synthesized but showed little variation in water contact angle with mole fraction of PHEMA in the BCP. Random copolymers such as PtBS-r-PHEMA-r-BCB would require synthesizing a new underlayer for every unique surface energy desired. Given that this BCP is expected to have a very high-χ, the range of underlayer contact angles expected to be neutral to this BCP is likely narrow. Therefore, a faster method of varying the contact angle would be desirable compared to synthesis of a new underlayer for any change in contact angle.

PHEMA-THP-r-BCB was developed as an acid sensitive underlayer with the ability to change its surface energy by using a photoacid generator. The random copolymer was dissolved in PGMEA with 15 wt. % TPS-SbF₆ and filtered. The solution was spin coated onto a clean silicon wafer piece and the BCB units were crosslinked. Next, the thin film was exposed to a specific dose of 248 nm wavelength light from an Oriel Instruments deep ultra-violet (DUV) exposure tool (Model #87530-1000) and then baked at 100 °C for 2 min. The exposure to 248 nm light causes an acid to form from the TPS-SbF₆ and the postexposure bake allows this acid to diffuse through the film and deprotect PHEMA-THP into PHEMA.²³ By varying the dose of light the film receives, the amount of deprotection can be controlled. After the postexposure bake, the film is then rinsed with PGMEA and baked at 90 °C to remove any remaining solvent. Figure 13 shows the variance in water contact angle of PHEMA-THP-r-BCB with exposure dose. With no deprotection, PHEMA-THP-r-BCB shows a water contact angle of about 66°. The random copolymer however is very sensitive to the deprotecting acid, dropping to around 55° after an exposure of only 5 mJ/cm².

The random copolymer iPOC-r-BCB was developed similar to PHEMA-THP-r-BCB in that its surface energy is also sensitive to acid deprotection of the iPOC into PHOST. The iPOC-r-BCB was dissolved in PGMEA with 20 wt. % TPS-SbF₆, filtered, and then coated onto a clean wafer piece. After crosslinking the BCB, the film was exposed to 248 nm wavelength light and then given a postexposure bake at 170 °C for 2 min. Figure 14 shows the water contact angle of iPOC-r-BCB is well controlled by the exposure dose and varies between 70° and 85°.

Thin films of PtBS-b-PHEMA were dissolved in THF, filtered, and coated onto silicon wafer pieces precoated with the crosslinked underlayers mentioned in Sec. III C. The thickness of the crosslinked underlayers used was typically in the range of 20–40 nm. To promote perpendicular orientation of the BCPs morphology, BCP thin films were coated at incommensurate thicknesses with their pitch. Thicknesses ranged between 1·L_o and 3·L_o. For each underlayer, a lamellae forming sample and a PHEMA-cylinder forming BCP sample were coated on a silicon wafer piece. While lamellae forming BCPs prefer a neutral underlayer in order to orient perpendicular to the substrate, PHEMA-cylinder forming BCPs would prefer an underlayer slightly preferential to PHEMA in order for the PHEMA-cylinders to orient perpendicular. The three BCP samples used on each underlayer were a lamellae forming BCP (L1) with a pitch of 15 nm (φ_PHEMA = 0.45) and an N = 58, a cylinder forming BCP (C1) with a 10 nm pitch (φ_PHEMA = 0.36) and an N = 29, and a suggested perforated lamellae forming BCP (PL1) with a 24 nm pitch (φ_PHEMA = 0.34) and a N = 88. While the SAXS for PL1 at first glance appears to have a lamellar morphology, after thin film annealing of the sample, perpendicularly oriented cylindrical features were observed. Due to the conflicting results from SAXS and thin film samples, it is suggested that this sample is actually forming a metastable perforated lamellae morphology. These samples were chosen due to their large pitches relative to the other samples synthesized, making them the easiest to focus on in the SEM. After the BCP film was coated onto the crosslinked underlayer, the sample was baked at 90 °C for 2 min to remove the casting solvent. The BCP thin film was then placed under vacuum (<25 Torr), the temperature was increased to 160 °C, and the films were annealed for 12 h.

Table I gives an overview of the thin film results as examined by SEM where the symbol for each underlayer/BCP pair represents the orientation of the features as parallel or perpendicular. Depending on the underlayer preference, islands and holes may form when the film thickness is incommensurate with the BCP's pitch.²⁴ For a neutral underlayer, these islands and holes will not form at incommensurate film thicknesses. These islands and holes would be visible by SEM. Topography was seen for all C1 samples with featureless SEM scans when zoomed in and focused. This suggests that either the windows for perpendicularly oriented features is very narrow or that a more PHEMA-preferential (the minority block of the cylinder forming BCP) underlayer is required. Figure 15(a) shows the peak error force of an AFM scan of C1 on an underlayer with a 75° water contact angle. The peak error force is a measure of the overshoot by the AFM tip when it approaches a region with topography. Cylinders are oriented parallel to the substrate giving the appearance of striped patterns over the scan. The stripes displayed in Fig. 15(a) implies that PHEMA-cylinders buried under a PtBS matrix are causing a slight topography in the underlayer. The inset of Fig. 15 shows the 2D-fast Fourier transform (2D-FFT) of the striped pattern, and a profile of the peak error force along several parallel cylinders is shown in Fig. 15(b). Both the 2D-FFT and the profile along the cylinders show an average pitch of about 12 nm, in close agreement with the bulk pitch of 10 nm determined by SAXS.

Thin film samples of L1 were featureless in the SEM and AFM. Parallel oriented lamellae stack in layers would appear featureless from top down scans of either imaging techniques. No large-scale topography of islands and holes though were seen for L1 samples on underlayers shown in Table I. The perforated lamellae sample (PL1) showed some island/hole topography only while on a 50° water contact angle underlayer. For perforated lamellae, the only instance when features should not be present at the surface is when the thickness is such that the perforating block's lamellae is at the surface. SEMs for PL1 thin films atop 66° and 80° water contact angle underlayers displayed a mixture of feature and featureless regions at the surface which may be explained by variations in local film thickness. However, at a 75° contact angle, perpendicular cylindrical features (⊥c in Table I) were seen as shown in Fig. 16. Due to the presence of these cylinders conflicting with the lamellar morphology SAXS profile, this sample is suggested to be perforated lamellae.²⁵ The 2D-FFT inset measures an average pitch between perforations of about 25 nm, in good agreement with the bulk pitch measurement of 24 nm by SAXS. The presence of these cylinders means that neither the PtBS nor the PHEMA block prefers to be at the free surface so greatly that a wetting layer forms there. Despite this, one block may still prefer the free surface more than another to cause the apparent parallel orientations for cylinder and lamellae forming samples. Previous work showed that the surface energy difference between PS and PHEMA was not so great that a top coat was needed to form perpendicular features.¹⁵ However, the surface energy difference between PtBS and PHEMA may have crossed some threshold that now requires a more neutral top coat in order to form perpendicular features.

To be useful as a lithographic mask, BCPs must be composed of blocks with an etch contrast, meaning one block needs to be able to be etched faster than the other. To determine the etch contrast for PtBS-b-PHEMA, an etch study on thin films of homopolymer PtBS and PHEMA was conducted. Typically, plasma etchers use specific recipes and conditions for etching a particular BCP; however, as a rough estimate for the etch contrast between PtBS and PHEMA, O₂ plasma generated from the air will be used in this study. The Ohnishi parameter (O.N.) can be used to determine the expected etch contrast between two polymers based on the ratios of carbon atoms in each polymer's repeat unit.^26–28 Equation (1) shows that the Ohnishi parameter is calculated as the ratio of the total number of atoms N_total in a repeat unit to the difference in the number of carbons (N_C) and oxygen (N_O) atoms

For PtBS, O.N._PtBS = 2.3, while for PHEMA, O.N._PHEMA = 6.3. This means that PHEMA is expected to etch roughly three times faster than PtBS. Homopolymers PtBS and PHEMA were dissolved in PGMEA and DMF, respectively, and coated on clean silicon wafer pieces. One at a time, the films were placed into an O₂ plasma etcher, put under vacuum, and then etched (RF: 10.5 W) in 10 s intervals. Figure 17 shows the amount etched from the two films at each time interval with the slopes representing the etch rate. PHEMA homopolymer was found to etch 60% faster than PtBS under these conditions. It should be noted that the conditions inside the etcher may need to be optimized as well with a proper concentration of O₂ plasma rather than just using O₂ from the air that leaks into the system to generate plasma. This may explain the difference between the etch contrast measured (1.6) and the etch contrast expected by the Ohnishi parameter (3.0).

However, even if PHEMA did have a low etch contrast with PtBS, an alternative route does exist to increase that contrast. Due to PHEMA having a native hydroxyl group, sequential infiltration synthesis could be used to grow layers of titanium or aluminum oxides in the PHEMA region, increasing the etch contrast between the two BCPs.²⁹ 

PtBS-b-PHEMA, a new high-χ BCP, was synthesized by anionic polymerization and ATRP with low PDIs. Bulk SAXS profiles of PtBS-b-PHEMA show that the BCP can form lamellae with a sub-7 nm pitch and cylinders with a 10 nm pitch. While the χ of the BCP is still unknown, estimates show that it has a lower limit of 0.45, a greater value compared to other high-χ BCPs such as PCHE-b-PMMA, PTMSS-b-PLA, and PS-b-PHEMA. Radical polymerized random copolymer underlayers with the ability to change their surface energy via acid-catalyzed reactions were used for searching a neutral underlayer for PtBS-b-PHEMA. PHEMA-THP-r-BCB and iPOC-r-BCB were able to successfully modulate their surface energies over a water contact angle range of 50°–66° and 70°–85°, respectively. Perpendicularly oriented features were only found for a perforated lamellae sample implying that a wetting layer of one block does not inherently form at the free surface. However, a top coat may be needed to produce a sufficiently neutral-free surface to allow for perpendicular features to form. Further thin film studies are needed to find an underlayer as well that will perpendicularly orient features to the substrate for lamellae and cylinder forming samples of PtBS-b-PHEMA. Due to the high-χ of PtBS-b-PHEMA, a narrow window in surface energies may exist for these to occur. An etch study was conducted for PtBS and PHEMA homopolymers showing that PHEMA etches about 60% faster than PtBS under O₂ plasma. This etch contrast is expected to increase by using a more well controlled etcher.

Special thanks to Tingting's group at Beijing University of Chemical Technology and Nan Zheng from South China University of Technology for providing SAXS. Research on directed self-assembly materials for advanced lithography project was supported by Beijing Municipal Science & Technology Commission (Grant No. Z161100002116001). This paper is based upon work supported by the National Science Foundation (NSF) (Grant Nos. CMMI-1534461 and CBET-1512517). Any opinions, findings, and conclusions or recommendations expressed in this paper are those of the authors and do not necessarily reflect the views of the NSF.