**2. Results**

The reaction between MHMH and water leads first to disiloxanol **1** and then to the tetrasiloxane formation **2** (Figure 3A,B); both reactions are rapid and lead to the concomitant formation of hydrogen at the water droplet interface. *Note: caution must be taken, as this can lead to a pressure spike of this flammable gas*. Repetition of these processes lead to PDMS polymer initially terminated with SiH and then SiOH groups (Figure 3C–E). The processes were followed using 1H-NMR (particularly the relationship between integrated peak areas of Si-*H* (~4.7 ppm vs. Si-C*H*3 ~0.1–0.2 ppm [34,35])) and 29Si-NMR spectra that allow clear differentiation between cyclic silicone monomers, linear silicone polymers and HSiterminated linear materials [35], as well as gel permeation chromatography (GPC) and Fourier-transform infrared spectroscopy (FTIR; Supplementary Material, SM).

The objective of the research was to identify simple, efficient processes that would lead to linear PDMS, optionally terminated with SiH groups. Preliminary experiments examined the impact of reaction parameters, including catalyst concentration, solvent and, in particular, the effect of homogeneity between water and silicone phases during the course of hydrolysis. In addition, the impact of "one-shot" addition of reagents was compared to a titration in which water and/or additional BCF were added when reaction ceased.

Simply mixing M H M H, B(C6F5)3 stock solution in dry toluene with bulk water led to linear silicones (entries 1–3, Table 1). Unsurprisingly, reactions were faster with more catalyst, but satisfactory rates were already achieved with only 0.02 mol% of this not inexpensive catalyst (<30 min, see below). Adding water in excess to the stoichiometric requirement did not lead to an improved outcome: cyclics were formed as a byproduct (see below, Supplementary Materials). Dilution with solvent, initially dichloromethane (containing 72.5 ppm water), demonstrated that much higher molecular weight polymers were accessible with even lower levels of catalyst (entries 4–6, Table 1); however, D4 was also a byproduct of these reactions (Supplementary Materials). Thus, if high molecular weight is most desired, using small amounts of organic solvents is beneficial (entries 5, 6, Table 1); if suppression of cyclics is key, use slightly more BCF (entry 1, Table 1). Note that the use of a completely homogeneous reaction was disadvantageous for practical and kinetic reasons. For example, the reaction of 1 g of M H M H would require 14.3 mL of "wet" toluene (saturated with sufficient water to complete SiH hydrolysis), so scale up would not be practicable. When we did attempt the reaction on a small scale under these conditions, it proceeded for only a small extent over 5 h and then remained unchanged even after 24 h (see Supplementary Materials).

**Table 1.** High-molecular-weight PDMS (polydimethylsiloxane) preparation using hydrolysis.

$$\begin{array}{ccccc} \mathsf{H}^{\mathsf{A}} & \mathsf{H}^{\mathsf{C}} & \mathsf{S}^{\mathsf{C}} & \mathsf{S}^{\mathsf{C}} & \mathsf{S}^{\mathsf{C}} \\ & & & & \mathsf{H}^{\mathsf{C}} & \mathsf{O}^{\mathsf{C}} & \mathsf{O}^{\mathsf{C}} & \mathsf{O} \end{array} \quad \begin{array}{ccccc} \mathsf{C}^{\mathsf{A}} & \mathsf{O}^{\mathsf{A}} & \mathsf{O}^{\mathsf{A}} & \mathsf{O}^{\mathsf{A}} \\ & & & \mathsf{H}^{\mathsf{C}} & \mathsf{O}^{\mathsf{C}} & \mathsf{O}^{\mathsf{C}} \\ & & & & \mathsf{H}^{\mathsf{A}} & \mathsf{O}^{\mathsf{C}} & \mathsf{H}^{\mathsf{C}} \\ \end{array}$$


a Stoichiometric addition of bulk water except for entries 5, 6. Reactions were quenched between 2.5 and 22 h (SM). Most reactions were complete within 30 min (see below), conditions under which the catalyst remains highly effective [16]. b O = one-shot reaction; P = reagents were added portion by portion, each time bubble evolution ceased. c DCM contained 72.5 ppm water. Toluene was dry. Note: experiments in Table 1 were completed in DCM to optimize reaction conditions. We then switched to the greener solvent toluene. d Fraction of D units in D3 + D4 + D5 or PDMS based on integration in 29Si-NMR. e Mass % of non-volatiles. f Cold trap (volatiles) contained 23.9% (total mass balance 97.7%), which consisted of MHMH 50%, Me2SiH2 42%. g Cold trap contained 10.5% (total mass balance 99.6%), which consisted of MHMH 66%, Me2SiH2 26%.

#### *2.1. High-Molecular-Weight PDMS Preparation Using Hydrolysis*

The preparation of telechelic PDMS polymers from M H M H terminated with either SiOH or SiH groups was straightforward by adding water in a one-shot process in the presence of 0.02% B(C6F5)3. The two-phase reaction of water/silicone was surprisingly rapid; reaction times of less than 30 min led to polymers and concomitant formation of

H2. The hydrolysis/condensation reactions in 50 wt% toluene were ye<sup>t</sup> more rapid, as judged by the rate of build of the polymer molecular weight (entries 1–9, Table 2). We have not determined if the hydrolytic processes in toluene/water involve only homogeneous or a mixture of homogeneous/heterogeneous steps. The efficiency of chain extension will decrease as the living polymer increases in size, particularly above the entanglement limit of about 29,000 <sup>g</sup>·mol−<sup>1</sup> (note: in the literature, reported entanglement limits ranges from about 15,000–35,000 <sup>g</sup>·mol−1. Here we use data from the seminal study of Mrozek et al. [36]). This was clearly observed here, as the final Mn were 31–45 kg·mol−<sup>1</sup> (entry 9, Table 2); complete consumption of SiH groups at higher conversion terminates polymerization by forming HOSi-terminated polymers. Re-initiation of such "dead" polymers is straightforward; however, addition of small quantities of MHMH caps the SiOH groups leading to dimerization or higher homologues of the existing polymer (entry 10, Table 2, Figure 3F). That is, a beneficial consequence of this process is that if HSi-terminated polymers are desired, one need only add excess MHMH to cap (Silicone-Me2SiOH → Silicone-(Me2SiO)2Me2SiH) and, if desired, grow the polymers before quenching the catalyst. This observation demonstrates that the process is living (entry 10, Table 2) [26]. Note that any residual water will compete for the silanol chain ends such that, if high molecular weights are desired, it is advantageous before capping to remove using distillation the small quantities of low molecular weight materials present, including residual water.

**Table 2.** Molecular weight versus reaction time using hydrolysis of MHMH.


aReactions performed with 50 wt% MHMH dry toluene + liquid water. b Calculated based on Si*H* peak area versus SiC*H*3 (integrated to 100) region in the 1H-NMR. c The observation at 6 min. is considered an outlier. d Starting polymer entry 4, Table 1 contained no cyclics, and none formed by 29Si-NMR; mass balance 99.9%.

#### *2.2. Managing Cyclics*

Our objective was to develop simple, practicable polymer syntheses that avoided the need for inert gas blankets; a septum with a bubbler was used to control pressure. One of the challenges presented by MHMH is its high volatility, which was problematic with or without solvents. The evolution of cyclics was followed by a combination of gravimetric analysis for volatile products (trapped in a cold trap that permitted cogenerated H2 to escape) and, for the polymerization mixture, 29Si-NMR, which is particularly sensitive to subtle differences in the chemical environment of D units; it is straightforward to distinguish D3 (−8.3 ppm), D4 (−19.1 ppm) and D5 (−21.5 ppm) from D units in linear polymers (−21.6 ppm) [37]. MHMH was the main constituent captured in a cold finger, with small amounts of MHDMH, Me2SiH2 and D3 (entries 7,8, Table 1).

The hydrolysis/condensation of MHMH in dilute, homogenous toluene solution was very slow. However, two-phase reactions (water/silicone or water/silicone+toluene) were very rapid reactions and completed in <30 min. Reactions can occur at the interface or within the organic fluid. In the case of the water/silicone mixture, the polymer yield was near 80%, with competing growth in D3 and small amounts of D5; D4 was only inefficiently formed (Figure 4A). By contrast, even with the small dilution provided in the 50% water/silicone+toluene system, much less polymer was formed at the expense of D3 and D5 production, again with little D4 (Figure 4B).

**Figure 4.** (**A**) Conversion to polymer without solvent. (**B**) Conversion to polymer in 50 wt% toluene. Rate of reaction (loss of HSi) during hydrolysis of MHMH was monitored using 1H-NMR and silicone constituents using 29Si-NMR. Integrations assumed identical sensitivity for all D units and are normalized to 100%.

Cyclic monomers are important starting materials for silicone synthesis. As noted above, the ring strain in D3 makes it an attractive starting material for ring-opening polymerizations. On the other hand, there is an interest for a variety of reasons in making cyclic-free silicones, particularly from an inexpensive starting material like MHMH that polymerizes so rapidly. We had hoped to be able to fine tune the polymerization to create high molecular polymers in the absence of cyclics.

After dimerization, MHMH → MHDDMH **2**, the tetramer can undergo hydrolysis to give **3** then further chain extension, or cyclization to give D4 (Figure 3D vs. Figure 3G). The generation of D4 during polymerization under equilibrating conditions is favored both enthalpically and entropically, possessing virtually zero ring strain, and the SiO bond strength in cyclics is similar to those in linear chains and larger number of molecules than in linear polymers [1]. In neither of the cases examined was D4 a significant product (Figure 4, Table 1). Thus, under these conditions of chain extension, **3** outcompetes cyclization (Figure 3D,K vs. Figure 3G). Adding small amounts of toluene led to an increase in D4 product only from 1% to 3%.

More problematic, with respect to cyclics, was the formation of D3 and, to a lesser extent, D5. The most concentrated (neat) solution produced 13% D3, but in the diluted sample 33% of the product mixture was D3, in addition to linear polymer (entries 7,8, Table 1). These data show that, even in the presence of water, the Chojnowski metathesis to give D3 and Me2SiH2 from **2** is highly competitive with hydrolysis (Figure 3H). The presence of D5 in such high quantities is consistent with chain extension from **3** to **4** and then a different Chojnowski metathesis leading to D5 and Me2SiH2 (e.g., Figure 3K,L. We thank a referee for this suggestion). It is believed that the formation of MHDMH, found in the cold trap, can be ascribed to reactions with Me2SiH2 with disiloxanes in the presence of B(C6F5)3 (Figure 3M).

Previous experience with polymer chain extension of HSi-telechelic polymers with water did not lead to the formation of cyclics. Under these conditions with low catalyst concentrations (0.02%) (BCF/H2O and B(C6F5)3·OH2 [38]), redistribution reactions are similarly not efficient. While the reactions here were complete in 30 min, there was no change in the cyclics profile between 30–180 min (SM); redistribution would favor the formation of D4 which, of the cyclics characterized, was formed in the lowest concentration. Under equilibration in the absence of solvents, the normal concentration of D4 is ~15% [4]. Similarly, redistribution cannot explain the higher concentrations of D3 and D5.

The formation of cyclics vs. linears is therefore a consequence of the kinetics of intravs. intermolecular reactions within a silicone/organic solvent, or at the water interface. When done in 50wt% toluene, more cyclics, particularly D3, were produced than when the only solvent for the silicone was M H M H and silicone products themselves, i.e., than when the system was more concentrated (Figure 4B vs. Figure 4A). This shows the power of Chojnowski metathesis. The conversion of **2** → D3 (and **4** → D5) effectively competes with hydrolysis of **2** → **3** when the reaction is slightly diluted, and is an important reaction even when done neat. The formation and then disappearance in the 29Si-NMR of a maximum of 1.4% of Me2SiH2 was observed based on D-units (for full details of all intermediates from Figure 4, see Supplementary Materials). M HDM H, observed in the cold trap, could arise from reaction of Me2SiH2 with **3** and there may be additional homologation reactions that consume this reactive material. We cannot distinguish between polymerization involving hydrolysis/condensation (Figure 3B–E) vs. the reaction between D3 and hydrosiloxanes, as shown by Chojnowski et al. (Figure 3I,J) [33]. However, the rapid rate and high molecular weight support the hydrolytic/condensation process as dominant.

Molecular weight is, of course, an important determinant of polymer properties. One advantage of this process is the relatively narrow dispersities achieved in the higher molecular weight polymers (Table 2). They do not match dispersities of polymers formed from ring-opening polymerization of D3 but, apart from the challenge of managing hydrogen generation, are easier to perform. In part, the objectives of the work were met. It was possible to obtain rapid polymerization of M H M H with water to ge<sup>t</sup> medium-molecular-weight linear silicones terminated, if desired, with SiH groups. High concentrations facilitated the rates of polymerization. The process is living and any premature suppression of polymerization by complete hydrolysis to SiOH groups can be overcome by the addition of small amounts of M H M H, preferably after removing any residual water, a process that allows much longer polymers to form. However, unlike the hydrolytic process with starting materials that have a higher MW (e.g., DP > 6) [26], polymerization competed with undesired cyclic formation. Distinctions between these in previous studies are mostly related to concentrations. The important precedent work of Ganachaud showed that with very high water and higher catalyst concentrations, complex equilibration reactions occur under less controlled conditions [32]. On the other hand, with exceptionally low concentrations of water and catalysts, Chojnowski et al. showed metathesis of **2** dominates the process; one can elect to capture and then use D3 for polymerization [21,22]. This work demonstrates the power of Chojnowski metathesis. Even without a solvent, efficient formation of D3 is observed, up to 33%, made worse even by small amounts of solvent.

#### **3. Experimental Section**
