Dealkylation of Poly(methyl methacrylate) by TiCl4 Vapor Phase Infiltration (VPI) and the Resulting Chemical and Thermophysical Properties of the Hybrid Material

This study examines the chemical reaction pathways for vapor phase infiltration (VPI) of TiCl4 into poly(methyl methacrylate) (PMMA). VPI is a processing method that transforms organic polymers into organic–inorganic hybrid materials with new properties of interest for microelectronic patterning, technical textiles, and chemical separations. Understanding the fundamental chemical mechanisms of the VPI process is essential for establishing approaches to design the chemical structure and properties of these hybrid materials. While prior work has suggested that TiCl4 infiltration into PMMA does not disrupt the polymer’s carbonyl bond, a clear reaction mechanism has yet to be proposed. Here, we present a detailed X-ray photoelectron spectroscopy study that presents evidence for a concerted reaction mechanism that involves TiCl4 coordinating with the PMMA’s ester group to dealkylate the methyl side group, creating a chloromethane byproduct and primary chemical bonds between the organic and inorganic components of the hybrid material. Additional spectroscopy, quartz crystal microbalance gravimetry, and thermophysical and chemical property measurements of this material, including solubility studies and thermal expansion measurements, provide further evidence for this chemical reaction pathway and the subsequent creation of inorganic cross-links that network these TiOx–PMMA hybrid materials.


■ INTRODUCTION
−18 The properties of these hybrids are dependent upon the quantity of infiltrated inorganic material, the chemical bonding between the organic and inorganic components, and the spatial distribution of the inorganic component. 2,8,19,20nfiltrated inorganic precursors become entrapped in polymers by chemically reacting with or adducting to the functional groups of the polymer.Adducted species may subsequently react with a sequentially delivered coreactant species to form a nonvolatile inorganic product that remains entrapped but possibly weakly bound to the polymer itself. 1,2,21or example, McGuinness et al. 2 have shown that TMA infiltrated into PMMA can occur in both adducted and chemically reacted states, and this chemical binding state along with the inorganic's spatial distribution directly impacts the resultant material properties.For example, at low temperatures, TMA sorbs into PMMA and forms a metastable reversible adduct, leading to weakly bound inorganic species that can actually be dissolved out of the hybrid by immersion in water, while at higher temperatures, the TMA is more likely to form primary chemical bonds to the carbonyl's oxygen forming a strongly cross-linked network hybrid. 22The difference in this bonding chemistry leads to differences in the chemical stability of these hybrids in various solvents. 2n prior work, we have demonstrated that TiCl 4 infiltration into PMMA occurs via a reaction-limited mechanism. 23Unlike TMA infiltration, TiCl 4 infiltrates uniformly, albeit more slowly, within the entire depth of a PMMA thin film�that is to say, the TiCl 4 rapidly diffuses into the PMMA, but entrapment is limited by the reaction rate.While this prior study clarified the rate-limiting step for the VPI process kinetics�the chemical reaction�the exact chemical mechanisms for this reaction process were not clarified.−26 However, Biswas et al. have shown with in situ IR spectroscopy measurements that only weak interactions occur between TiCl 4 and the carbonyl during both the TiCl 4 and water dose. 24Thus, the exact chemical mechanisms for TiCl 4 entrapment in PMMA during VPI are still unclear.
Although the entrapment mechanism is not fully understood, TiCl 4 infiltrated polymers have been shown to have altered chemical, thermal, optical, and electrical properties. 5,7,12,16,17Additionally, PMMA/TiO 2 nanocomposites formed from TiO 2 nanoparticles have showed enhanced properties such as increased thermal stability, increased refractive index, and increased UV absorption. 27,28Thus, these TiO x /PMMA hybrids likely have potential interest for numerous applications.
In this study, systematic X-ray photoelectron spectroscopy (XPS) studies are used to track the behavior of PMMA's ester group after varying times of TiCl 4 infiltration.The changes in the oxygen spectrum provide particularly compelling evidence for a concerted chemical reaction that leads to dealkylation of the methoxy group and C−O−Ti bond formation.Additional chemical solubility data and thermal expansion measurements provide further evidence that a chemically cross-linked hybrid material is formed.

Materials.
Poly(methyl methacrylate) (PMMA) with a molecular weight of approximately 75 kDa was acquired from Polysciences, Inc.A 5 wt % solution of the polymer in toluene (99.8% anhydrous, Sigma-Aldrich) was prepared.This solution was spun-cast onto silicon substrates at 3000 rpm for 30 s, resulting in films with a nominal thickness of 160−200 nm.Thicker films necessary for Fourier transform infrared (FTIR) spectroscopy were made using 15 wt % PMMA and cyclohexanone as the solvent.This solution was spun-cast onto double-side-polished silicon substrates at 6000 rpm for 60 s to achieve films of 1−1.2 μm.All films were then placed on a hot plate and heated to 150 °C for 1 h to remove any remaining solvent.
Vapor Phase Infiltration.PMMA films were infiltrated in a custom-built reactor having a 28 L chamber and operated with decision-tree-based control software. 29PMMA was infiltrated at 90, 120, and 150 °C with TiCl 4 .The TiCl 4 precursor was infiltrated with overpressures of ∼1 Torr.All pressures in the reaction chamber were measured with a Baratron capacitance manometer.All VPI processes used a single precursor−co-reactant cycle static hold scheme.The general process sequence was (1) ultrahigh purity N 2 gas was flowed into the reactor to purge the system for 5 min, (2) the system was pumped down to base vacuum (30 mTorr) for an hour for full removal of water, (3) the chamber was isolated, (4) the TiCl 4 precursor valve, which is connected directly to the chamber, was opened for 5 s to reach a vapor pressure of about 1 Torr of TiCl 4 , (5) the TiCl 4 was then held in the chamber for between 1 and 48 h, (6) the system was then pumped to base vacuum for 5 min, (7) the water co-reactant valve, which is also connected directly to the chamber, was opened for 1 s to give a vapor pressure of 1.8 Torr in the chamber, and (8) the water was held in the chamber for 1 h before purging the system for 60 s and venting to the atmosphere. 23n Situ Quartz Crystal Microbalance (QCM) Gravimetry Measurements.QCM experiments were performed in a hot-walled custom-built VPI reactor described elsewhere. 4,30The QCM used is a Phoenix high-temperature film thickness sensor PC-based system purchased from Colnatec.A polished gold 6 MHz RC quartz crystal was used as the substrate for the poly(methyl methacrylate).The crystal and the surrounding walls were heated to 150 °C under vacuum and flowing nitrogen to determine its baseline resonance frequency.Subsequently, 5 wt % solution of the PMMA in toluene was spun-cast onto the gold crystal substrate at 3000 rpm for 30 s, resulting in a film of ∼270 nm.The PMMA-coated crystal is then placed in the reactor and heated to 150 °C under vacuum and flowing nitrogen to determine the resonance frequency with just the bare polymer.VPI is then conducted as described above except for three key differences: (1) the TiCl 4 precursor valve, which is connected directly to the chamber, was opened for 0.5 s to reach a vapor pressure of about 3.2 Torr of TiCl 4 due to the smaller size of this reactor, (2) the film is pumped to base vacuum for 24 h between the TiCl 4 dose and the water dose to ensure all unreacted precursor and byproducts have sufficient time to escape the polymer, and (3) the water co-reactant valve, which is also connected directly to the chamber, was opened for 0.5 s to give a vapor pressure of 12 Torr in the chamber.The crystal frequency was recorded every second during this time and was exported and converted to mass via the Sauerbrey equation. 31esidual Gas Analyzer (RGA).An EXtorr XT Series residual gas analyzer operated on a PF70 turbomolecular vacuum pump was attached to the hot-walled, custom-built VPI reactor described above via a capillary tube to sample the reaction atmosphere.The mass/ charge (m/z) was determined with a scan speed of 48/s from 1 to 200 atomic mass units.The gas atmosphere was sampled by (1) opening a valve that transmits the gas in the reactor to the RGA and (2) turning on the filament inside the RGA and analyzing the gas atmosphere.
Fourier Transform Infrared Spectroscopy.Attenuated total reflectance Fourier transform infrared (ATR-FTIR) spectroscopy was performed on 1 μm films of PMMA on double-sided polished, lowconductivity silicon.The spectra were collected on a Thermo Scientific Nicolet iS5 FTIR spectrometer with an iD7 ATR accessory and a diamond crystal.Spectra were collected with a resolution of 4 cm −1 and are the average of 64 scans.
X-ray Photoelectron Spectroscopy (XPS).XPS was performed by using a Thermo K-alpha system using a monochromatic Al Kα Xray source (1486.6 eV) with a 60°incident angle and a 90°emission collection geometry.Survey scans were conducted at a pass energy of 200 eV and for binding energies from −10 to +1350 eV.For the elemental analysis, the following elements at the following binding energies were collected: Ti 2p (448−475 eV), O 1s (525−545 eV), C 1s (279−298 eV), Cl 2p (190−210 eV), and Si 2p (95−110 eV).Films were etched with a raster size of 400 × 400 μm 2 , an ion gun voltage of 2000 V, a high current, and for 65 s, yielding an approximate rate of 25 nm per etch level.At each level the elemental analysis and survey scan was performed.Three survey scans were conducted for all elements at each etch level.A Shirley background subtraction was used to determine the atomic percentage of Ti while a simple background subtraction was used for determining other atomic percentages.
Spectroscopic Ellipsometry and Measurement of CTE and Glass Transition Temperatures.Film thicknesses were measured with an Alpha-SE spectroscopic ellipsometer (J.A. Woollam) at a 70°a ngle over a spectral range of 340−900 nm.The film was measured before and after infiltration to determine the starting and ending film thicknesses and refractive index (at a wavelength of 632.8 nm).A Cauchy model was used to ascertain the film thickness and refractive indices.The A and B Cauchy coefficients and thicknesses were fitted.This layer was stacked on a 2.3 nm layer of native SiO 2 on a silicon substrate.
The coefficient of thermal expansion (CTE) and the glass transition temperature (T g ) were determined by measuring the polymer thickness over a range of temperatures (40−150 °C).Details of this measurement setup can be found in prior studies. 8,32,33easurements were taken at 5 °C intervals while the film was kept at a constant temperature for at least 10 s.No discernible differences in film thickness occurred for repeated measurements beyond 10 s at a given temperature.Films remained stationary throughout each temperature excursion, such that the exact spot was continuously measured.The volumetric CTE was determined from the slope of the thickness versus temperature.Assuming constrained isotropic behavior, the volumetric CTE value was divided by 3 to determine the more commonly reported linear CTE.
Dissolution in Toluene.Films were placed in 10 mL vials of toluene, which is a good solvent for the pure PMMA polymer.Films were removed from solution at various time intervals, placed in a fume hood for 5 min to dry, and then had their thickness measured with spectroscopic ellipsometry to determine the extent of film dissolution.
■ RESULTS AND DISCUSSION Chemical Mechanism.Previously, we have used the spatiotemporal profiles of the infiltrated TiO x inorganics to determine that the TiCl 4 −PMMA VPI process proceeds via a reaction-limited mechanism. 23−36 Above this spectrum are the spectra for the infiltrated hybrid at varying process temperatures.Difference spectra are also reported in Figure S1 of the Supporting Information.In all cases, only minor changes are observed in the infiltrated spectra.Specifically, the carbonyl stretch shows little to no bleaching upon infiltration, suggesting little to no direct reaction of this functional group with TiCl 4 and an end product that contains carbonyls.This observation of undisturbed carbonyl groups when infiltrating with TiCl 4 is consistent with the in situ IR data previously reported by Biswas. 24Additionally, the absorption for the C−O−R stretch is also largely unchanged.This result is significantly different from what is understood for trimethylaluminum (TMA) infiltration into PMMA, where the carbonyl stretch is completely quenched and the C−O−R stretch shows a significant frequency shift at infiltration temperatures above ∼100 °C, indicative of a chemical reaction between the TMA and ester group. 1,2,24,37,38While the small amount of C�O quenching and C−O−R vibrational shifts detected here in the difference spectra could indicate some amount of direct reaction between TiCl 4 and PMMA's carbonyl group at the higher process temperatures, these results largely suggest that some other chemical reaction mechanism is dominant for the TiCl 4 /PMMA VPI chemistry.
Figure 2a−e presents XPS spectra for PMMA films infiltrated with TiCl 4 for 0 (neat PMMA), 1, 12, 24, and 48 h at 150 °C.Note that these are collected from the film's surface.Surface scans were used because the surface is assumed to be the most saturated portion of the infiltrated polymer. 1,4,21s expected, no Ti is detected in the uninfiltrated PMMA film with zero counts observed (Figure 2a).Upon infiltration (Figure 2b−e), a clear doublet emerges at 458.7 and 464.45 eV that is consistent with Ti in its 4+ oxidation state.The intensity of these Ti emissions also generally increases with TiCl 4 exposure time, consistent with the expected increase in the titanium loading.The second column of Figure 2 shows the evaluation of the O 1s spectrum with TiCl 4 exposure time; these data are particularly indicative of the chemical mechanism.For neat PMMA, near equal amounts of the methyl−oxy (C−O) at 533.6 eV and carbonyl (C�O) at 530.5 eV are observed as expected.However, with infiltration, the methyl−oxy bond decreases in intensity while the carbonyl bond maintains its intensity.This C−O intensity continues to reduce as exposure time continues to increase.Concurrently, a metal−oxide (Ti−O) emission emerges at 530.5 eV and increases with TiCl 4 exposure time, consistent with the formation of infiltrated inorganics.These changes in the O 1s spectra suggest that with increasing TiCl 4 exposure times, the C−O species are being consumed, while Ti−O bonds are being created.A similar relative reduction of the C−O species is observed in the C 1s spectra at 286.3 eV, particularly evident at exposure times of 24 and 48 h.At these long hold times, the ratio of C−O to C�O (288.5 eV) significantly reduces, again suggesting a consumption of C−O bonds.Lastly, the Cl 2p emission is also reported.Negligible Cl is detected until 24 h of TiCl 4 exposure, at which point the signal is still relatively weak (0.12−0.60 at.%) and difficult to deconvolute.We can roughly attribute the Cl to possibly a Ti−Cl and/or a C−Cl state, which have been reported to occur at 198.5−199 and 200 eV, respectively. 40These may suggest a subsequent reaction that occurs after a long time or an inability to fully hydrolyze some of TiCl 4 when inorganic loading becomes high.
To further confirm these chemical trends, XPS spectra were also collected for hybrids created at varying VPI process temperatures and a fixed TiCl 4 exposure time of 24 h, comparable to the FTIR data reported in Figure 1.Presumably, the temperature would increase the reaction kinetics for such a reaction.Figure 3 plots the O 1s XPS spectra for this series of films.For neat PMMA (Figure 3a), equal amounts of the methyloxy (C−O) and carbonyl (C�O) are observed, as expected.At the lowest process temperature (90 °C, Figure 3b), the C−O emission decreases, while Ti−O emission increases.This trend then continues with increasing temperature (120 and 150 °C), similar to what is observed as a function of exposure time in Figure 2.
Overall, these spectroscopic observations suggest a chemical reaction that disrupts the methoxide side group of PMMA but leaves the carbonyl functionality intact.Both the reduction of the intensity of the O−C peak in the O 1s spectra and the reduction of the intensity of the C−O peak in the C 1s spectra suggest the loss of the methyl group.Metal chlorides have been shown to readily form adducts with ester groups via bidentate binding between the metal and the ester.These adducts can then drive concerted reactions. 41Scheme 1 proposes a similar interaction between PMMA and TiCl 4 that is consistent with these spectroscopic observations.The proposed mechanism follows a concerted dealkylation pathway between the TiCl 4 and the ester moiety of PMMA, resulting in a chloromethane byproduct.Because TiCl 4 is a known Lewis acid, TiCl 4 can conceivably adduct with the lone pair electrons from the ester group, creating a TiCl 4 :PMMA complex (ii).We speculate that this complex may place the Ti in an octahedral coordination, which is common for Ti, bonding to its four chlorine ligands and the two oxygens of the ester group.In this configuration one of the chlorines in the basal plane of the oxygens would be aligned with the methyl group, making it susceptible to reaction.Subsequently, this Cl − ion attacks the Me−O bond of the ester group, dealkylating it as a chloromethane byproduct and leaving a negatively charged carboxylate group adducted to a positively charged TiCl 3 + (iii).These two charged species presumably form a more permanent primary bond (iv) between Ti and one of the oxygens, although a resonance between (iii) and (iv) and a switching between which O the Ti is bonded to cannot be precluded.Interestingly, this proposed mechanism results in a residual carbonyl group, consistent with IR and XP spectroscopic observations, and the small to negligible shift in the C−O−R vibration in IR spectroscopy is consistent with the similar vibrational frequencies reported for both C−O−R and C−O−M species. 39−45 This process commonly occurs by acidor base-catalyzed hydrolysis of esters following a concerted dealkylation mechanism.Reports exist for alkyl cleavage of esters via strong Lewis acids such as AlCl 3 , TiCl 4 , or other metal halides as a catalyst in the presence of another coreactant. 41,46However, to our knowledge, no reports exist for dealkylation of esters in polymers using gaseous metal halides in low vapor pressure environments.Here we posit that it is this dealkylation reaction that acts as the rate-limiting step for the TiCl 4 −PMMA VPI process.To further validate this proposed mechanism, Figure 4 presents quartz crystal microbalance (QCM) gravimetry analysis of this TiCl 4 + H 2 O − PMMA VPI process collected at 150 °C.While an in-depth analysis of this data is beyond the scope of this paper, a brief quantitative and qualitative assessment provides further evidence supporting the chemical pathway proposed in Scheme 1.The QCM data are separated into five regions: (0) preinfiltration pumping, (1) TiCl 4 exposure (24 h), (2) TiCl 4 removal via vacuum pumping (24 h), (3) H 2 O exposure (1 h), and (4) H 2 O removal via vacuum pumping.In the initial stage of TiCl 4 exposure, a rapid increase in mass of approximately 4.00 × 10 −6 g is observed consistent with sorption of TiCl 4 into PMMA.Interestingly, though, within 5 min, a maximum mass uptake is reached, and the mass begins to decrease.This decrease in mass suggests the loss of a high-mass byproduct, such as chloromethane, consistent with the chemical pathway proposed in Scheme 1.After about 20 h of TiCl 4 exposure, mass loss nearly stops, and a constant mass is reached, suggesting the reaction has ended.
The total mass loss during the hold step is approximately 2.40 × 10 −6 g, which we attribute to chloromethane byproduct loss.Subsequently, upon TiCl 4 removal via vacuum pumping, the mass decreases again quite rapidly, presumably due to the desorption of any unreacted TiCl 4 that is dissolved only in the polymer and not chemically bound to the polymer.The total mass of desorbed species is 1.28 × 10 −6 g.Removal of additional byproducts may also happen in this step.However, we assume that only unreacted TiCl 4 desorbs.This removal of species occurs quickly, with most mass loss occurring within less than 1 h of pumping.Then upon H 2 O exposure, a mass rise is observed, presumably due to sorption of water.However, interestingly, upon water removal via vacuum pumping, the mass returns to nearly the same value as that prior to the water exposure; only a small amount of mass is gained, ∼0.25%.While this result may occur because a reaction species has the same mass as a byproduct species, this explanation seems unlikely given that chlorine-containing byproducts have considerably higher atomic masses than   4) H 2 O removal via vacuum pumping.The mass uptake is normalized to the original mass of the polymer (3.24 × 10 −05 g) to provide a percentage of mass added to the polymer via infiltration.All masses are calculated from the Sauerbrey equation. 31The inset shows in situ RGA data collected during step 1 that provides the gaseous composition of the atmosphere of the VPI chamber after 1 min, 30 min, and 1 h of TiCl 4 exposure.hydroxyl or oxide linkages.Thus, the more reasonable explanation appears to be that the water is just fully desorbing from the hybrid film without reacting with anything inside the film.This suggests that the TiCl 4 is "fully reacted" with the polymer prior to the water exposure step, and no further hydrolysis occurs upon water exposure.The lack of observable hydroxide groups in the FTIR spectra of Figure 1 provides further evidence that few hydrolysis reactions are occurring in this step. 47sing the masses that we described above, the mass of TiCl 4 that is consumed during the hold step and remains entrapped in the hybrid is 2.72 × 10 −6 g, which equates to reacting 1.43 × 10 −8 mol of TiCl 4 .Assuming that the mass loss is chloromethane with a molar mass of 50.5 g/mol, this equates to a molar loss of 4.75 × 10 −8 mol of CH 3 Cl.The molar ratio of TiCl 4 reacted to the CH 3 Cl released is 1:3.32.This result indicates that each of the entrapped TiCl 4 molecules reacts with on average ∼3.3 methoxy functional groups or forms ∼3.3 bonds to PMMA chains.This value approaches the full reaction extent of 4 bonds, and we suspect steric hindrances limit any further reaction under the process conditions explored.For a more detailed explanation of these calculations, refer to Figure S3.
Additionally, in preliminary efforts to sample the gas atmosphere during TiCl 4 vapor phase infiltration with a residual gas analyzer (RGA), we have been able to detect a signal at a mass-to-charge ratio of 50 units, which is consistent with the fractionation behavior expected for a chloromethane byproduct.This signal, shown in the inset of Figure 4, is observed to increase with the infiltration time, providing further evidence for Scheme 1.We do note that we cannot preclude the involvement of the HCl byproducts in this reaction mechanism, and that it is possible that HCl may serve as a catalyst for the proposed mechanism.However, at least in preliminary testing we did not find any direct reaction of PMMA with aqueous HCl solutions at elevated temperatures nor did attempts to reduce the HCl byproduct in the chamber during VPI alter the amount nor rate at which inorganic infiltration occurred.
Chemical and Thermophysical Properties of TiO x − PMMA Hybrids.We have subsequently interrogated the chemical and thermophysical properties of the TiO x −PMMA hybrid materials because these properties often give physical insights into the hybrid's chemical structure.These properties often provide evidence of whether chemical cross-links, via primary chemical bonds, are formed between the organic and inorganic constituents.Specifically, we have investigated the TiO x −PMMA hybrids prepared at infiltration temperatures of 120 and 150 °C at varying hold times because these films are expected to be fully infiltrated throughout the entire film depth. 23Figure 5 presents XPS depth profiling data for 120 °C (Figure 5a) and 150 °C (Figure 5b) infiltration into 200 nm PMMA films at varying exposure times.The abscissa axis has been normalized to the total film thickness based upon the silicon substrate signal to improve comparisons among each film.At 120 °C (Figure 5a) Ti is uniformly present throughout the entire depth, and the concentration increases from ∼1.5 to ∼3 at.% as the TiCl 4 exposure time is increased from 1 to 12 h.Above 12 h of exposure time, the inorganic loading approximately saturates. 23Figure 5a corroborates the reactionlimited profile observed in our previous paper at 135 °C; however, we observe a lower overall titanium concentration due to the lower processing temperature. 23A similar trend is observed at a VPI process temperature of 150 °C (Figure 5b).However, notably, only 1 h of TiCl 4 exposure is needed at 150 °C to reach the same inorganic loading as is achieved after 12 h of exposure at 120 °C (∼3.5%).At 12 and 24 h of exposure time, inorganic loading saturates at ∼6.5 at.% Ti.To establish structure−property relationships, we calculated average Ti 2p atomic percentages from the depth profiles of Figure 5. Figure 5c plots these average percentages for TiCl 4 infiltration at 120 and 150 °C as a function of exposure times.
Figure 6 plots the dissolution behavior of these hybrid materials as a function of the immersion time in toluene, a good solvent for the pure PMMA polymer.The total  immersion time here is 4 weeks.As expected, the untreated PMMA films fully dissolve within 10 min.In Figure 6a, we observe that films infiltrated at 120 °C have enhanced chemical stability at all exposure times.Hybrids synthesized from 1 h of VPI exposure, which is ∼1.5 at.% Ti, take >100 min before dissolution begins, but these films do approach full dissolution after ∼4 weeks of immersion.Hybrids prepared at 120 °C and 12 h of VPI exposure show no evidence for dissolution for up to 4 weeks of immersion in toluene, indicating good chemical stability.These hybrids, especially those prepared at shorter infiltration times (12 and 24 h), do show some swelling at long immersion times, which is indicative of a cross-linked organogel.Higher degrees of cross-linking would be expected to reduce the extent of swelling observed. 8Figure 6b plots the dissolution behavior in toluene for TiO x −PMMA hybrids prepared at 150 °C.Here, all hybrids remain stable through 4 weeks of immersion, with minimal swelling, suggesting even more substantial cross-linking.
To further test for evidence of chemical cross-linking, the thermophysical properties of these TiO x −PMMA hybrids are investigated.Using a heated-stage spectroscopic ellipsometer, we quantify both the glass transition temperature (T g ) and coefficient of thermal expansion (CTE) of these materials above and below the T g . 8,32,33Raw data from these measurements are included in Figure S2a,c.Figure 7 summarizes these data by plotting the T g and CTE as a function of average inorganic loading for the TiO x −PMMA hybrids.Figure 7a shows that increasing TiO x concentration increases the T g .Pure PMMA films exhibit a T g of about 95 °C, consistent with reports in the literature. 32,48,49With increasing TiO x loading, the T g exceeds 120 °C.−52 As cross-link density increases, the polymer's chain mobility becomes more restricted, increasing T g.We can further assess the consistency of T g increasing due to increased cross-link density by comparing to the Fox−Loshaek model. 8,50Assuming the TiO x concentration is correlated to the hybrid's cross-linking density, the dotted line in Figure 7a shows the best fit to the Fox−Loshaek model.Here, the fitting parameter, K c , which represents the cross-link's mechanical stiffness, is found to be 4.2, similar to the value of 3 reported for the inorganic cross-links in the AlO x −PS-r-PHEMA hybrid studied by Bamford et al. 8 While this value is higher than common organic cross-links, this result may suggest that we are actually forming multiple cross-links from a single TiCl 4 molecule.TiCl 4 reacting with multiple ester groups forming bonds to multiple PMMA chains could explain why no chloride ligands remain for the reaction with water, as indicated by the QCM gravimetry data in Figure 4.
Figure 7b plots the linear CTE of these hybrids as a function of the TiO x loading.Above T g the CTE decreases dramatically with increasing inorganic loading, from a value of 283 ppm/°C for pure PMMA films to 77 ppm/°C for the highest TiO x loading fraction.This decrease in CTE is approximately linear with the inorganic loading fraction.This reduction in CTE is again consistent with the formation of chemical cross-links that prevent chain motion and, hence, thermal expansion of the hybrid material.Below T g the reduction in CTE is much more modest.In the glassy state, the motion of polymer chains requires a sufficient free volume for atomic motion.Below the glass transition temperature, the density increases, and it becomes increasingly difficult for an atom to find sufficient free volume for motion to occur on a reasonable time scale. 33,53herefore, it is more difficult for the CTE to change substantially below T g .Additionally, in the glassy state, the thermal energy is already so low that it is difficult to overcome secondary bonds, so it is not until significant quantities of primary cross-links are formed that a noticeable drop in CTE occurs.However, at the highest loadings, CTE still drops by ∼50% compared to the neat polymer.
The chemical and thermal properties observed in PMMA− TiO x hybrid materials suggest that true chemical cross-links form during infiltration.These observations are consistent with the formation of primary chemical bonds between the organic and inorganic components as evidenced earlier.However, to achieve cross-linking, the final product of Scheme 1 must continue to react.Scheme 2 presents one posited cross-linking pathway where the bound TiCl 3 species continues to act as a Lewis acid that can undergo a subsequent concerted dealkylation reaction with a neighboring PMMA ester group, presumably on a nearby polymer chain, thus forming a chemical cross-link.While we cannot rule out the possibility of these −TiCl 3 groups hydrating with water and then forming C−O−Ti−O−Ti−O−C linkages via condensation with nearby metal hydroxides, the apparent lack of reaction with water in the QCM data and the lack of hydroxyls in the FTIR spectra suggest that this mechanism is unlikely.Currently, we suspect that the product in Scheme 2 may continue to react until all chloride ligands react with nearby PMMA chains, such that the single TiCl 4 molecule actually cross-links with up to four different PMMA chains.This higher-order cross-linking ■ CONCLUSION TiCl 4 infiltration into PMMA appears to occur via a reactionlimited mechanism.This paper elucidates the chemical mechanism of this rate-limiting reaction.Spectroscopically, a C−O−Ti bond forms but not with the PMMA's carbonyl group, which is undisturbed.Instead, the TiCl 4 appears to react with the ester group's methoxy oxygen, via a concerted dealkylation of the methyl side group, removing this methyl group.XPS spectroscopy, especially for O 1s, suggests that the ester O−Me group is consumed during the TiCl 4 VPI reaction and not the carbonyl.Additionally, QCM gravimetry reveals that after the initial sorptive mass uptake a consistent mass loss occurs during TiCl 4 infiltration, consistent with the loss of a massive byproduct, presumably chloromethane.Measuring the thermophysical properties of the resultant TiO x −PMMA hybrids provides further evidence that primary chemical bonds are forming between the organic and inorganic constituents of this material.For example, these hybrids are insoluble in organic liquids that are good solvents for the pure PMMA polymer.Moreover, the glass transition increases with infiltration, and the thermal expansion is reduced.All of these thermophysical observations are consistent with the formation of chemical cross-links.These results thus introduce a new reaction pathway between vapor-phase-infiltrated inorganic precursors and organic polymers.While prior studies have shown direct reactions between precursors and the carbonyl, this is the first time a clear reaction is shown with the methyl− oxy group, leading to cleavage and byproduct generation.The knowledge reported here about the reaction between TiCl 4 and PMMA will likely inform future investigation for infiltrating a variety of similar polymers with comparable ester functional groups.

* sı Supporting Information
The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.chemmater.3c02446.FTIR difference spectra showing PMMA infiltrated with TiCl 4 at varying temperatures (90, 120, and 150 °C) with 24 h precursor exposure time; plots of raw data collected for hybrid film thickness as a function of temperature used to evaluate the glass transition temperature and the coefficient of thermal expansion of these materials; plot of mass uptake versus time collected via in situ QCM gravimetry during VPI of TiCl 4 + H 2 O into PMMA at 150 °C and subsequent analysis of reaction and byproduct stoichiometries 31 (PDF) Coordinated Infrastructure (NNCI), which is supported by the National Science Foundation (ECCS-2025462).

Figure 1 presents
data collected from IR studies completed before and after VPI using a 24 h TiCl 4 exposure step at 90, 120, and 150 °C.The FTIR spectrum for neat PMMA shows the expected C�O stretch at 1729 cm −1 , a C−C−O stretch at 1240 cm −1 , a C−O−R stretch at 1145 cm −1 , a symmetric C−H stretch of CH 3 at 2954 cm −1 , a C−H stretch of CH 2 at 2994 cm −1 , and a C−H bending of CH 3 and CH 2 at 1440 and 1480 cm −1 .

Scheme 1 .
Scheme 1. Proposed Reaction Mechanism for TiCl 4 Infiltration of PMMA a

Figure 4 .
Figure 4. Mass uptake versus time plot collected via in situ QCM gravimetry during VPI of TiCl 4 + H 2 O into PMMA at 150 °C.The plot is separated into five temporal regimes: (0) preinfiltration pumping (vacuum base pressure), (1) TiCl 4 exposure (3.2 Torr), (2) TiCl 4 removal via vacuum pumping, (3) H 2 O exposure (12 Torr), and (4) H 2 O removal via vacuum pumping.The mass uptake is normalized to the original mass of the polymer (3.24 × 10 −05 g) to provide a percentage of mass added to the polymer via infiltration.All masses are calculated from the Sauerbrey equation.31The inset shows in situ RGA data collected during step 1 that provides the gaseous composition of the atmosphere of the VPI chamber after 1 min, 30 min, and 1 h of TiCl 4 exposure.

Figure 5 .
Figure 5. XPS depth profiles collected from PMMA films infiltrated with TiCl 4 + H 2 O at varying precursor exposure times from 0 h (neat PMMA) to 48 h.(a) Processes were performed at 120 °C.(b) Processes were performed at 150 °C.(c) Average Ti 2p atomic percentage for infiltration of TiCl 4 + H 2 O occurring at infiltration temperatures of 120 and 150 °C.The average Ti 2p atomic percentage was calculated using the weighted average of XPS atomic percentage values from (a) and (b).All films are nominally 200 nm thick, but the depth is normalized to the silicon substrate signal (not shown for clarity).

Figure 6 .
Figure 6.Dissolution plots for PMMA films infiltrated with TiCl 4 + H 2 O at varying precursor exposure times from 0 h (neat PMMA) to 48 h.Tracking the % thickness remaining at different time points over 4 weeks.(a) Processes were performed at 120 °C.(b) Processes were performed at 150 °C.The x-axis is presented in log scale.

Figure 7 .
Figure 7. Thermophysical properties of PMMA films infiltrated with TiCl 4 + H 2 O at 120 and 150 °C: (a) Glass transition temperature and (b) linear coefficient of thermal expansion (CTE) above and below the glass transition temperature.The average TiO x concentration is the weighted average of the XPS Ti 2p atomic % throughout the depth of the film.The black "+" shows the values for the untreated PMMA film.