Homoconjugated Acids as Low Cyclosiloxane-Producing Silanol Polycondensation Catalysts

Abstract

Polycondensation of α,ω-disilanols is a foundational technology for silicones producers. Commercially, this process is carried out with strong Brønsted acids and bases, which generates cyclosiloxane byproducts. Homoconjugated acids (a 2:1 complex of acid:base or a 1:1 complex of acid:salt), a seldom used class of silanol polycondensation catalysts, were evaluated for their ability to polymerize α,ω-disilanols while forming low levels of cyclosiloxane byproducts. Homoconjugated acid catalysts were highly active for silanol polycondensation, even when made from relatively mild acids such as acetic acid. Both the acid and base (or cation) component of the homoconjugated species was important for activity and avoiding cyclosiloxane byproduct formation. Stronger acids and bases were found to positively affect reactivity, and the pK_a of the acid was found to correlate with cyclosiloxane byproduct formation. The individual components of the homoconjugated species (the acid and base) were ineffective as catalysts by themselves, and compositions with fewer than 2 mol of acid to 1 mol of base were much less reactive. Homoconjugated trifluoroacetic acid tetramethylguanidinium and tetrabutylphosphonium complexes were found to be privileged catalysts, able to give high-molecular-weight siloxanes (M_n > 60 kDa) while generating less than 100 ppm of octamethylcyclotetrasiloxane byproduct. Finally, a mechanism has been proposed where silanols are electrophilically and nucleophilically activated by the homoconjugated species, leading to silanol polycondensation.

Introduction

Polycondensation of α,ω-disilanols to higher-molecular-weight silicones is one of the core polymerization technologies practiced by silicone producers.¹ At silicone manufacturing sites, various iterations of this technology are currently practiced, including homogeneous base-catalyzed condensation with KOH, heterogeneous condensation/equilibration with acid clays, and homogeneous acid catalysis with linear phosphonitrilic chlorides (Scheme 1).²⁻⁴ These commercial processes have historically met the silicone industry’s needs for medium- and high-molecular-weight silanol-terminated linear siloxanes. Each of these processes also makes cyclosiloxane byproducts in addition to silanol-terminated linear siloxanes, in amounts from parts per million to percent levels depending on the process and conditions via silanol/silanoate “back-biting”.^5,6 The volatile cyclosiloxanes, such as octamethylcyclotetrasiloxane (D4), are either present in low enough quantities post-polymerization not to warrant removal or, if present at higher levels, removed by vacuum stripping the polymers to afford low levels (<1000 ppm) of cyclosiloxane content.⁷

Scheme 1. Commercial Silanol Condensation Polymerization Catalysts.

Recently, there has been a push in the marketplace to offer silanol-terminated linear siloxanes with lower levels of cyclosiloxane byproducts than are currently being produced. This desire stems from both requests by customers and the desire for producers to remain in certain markets. Regarding customer requests, it is increasingly important for automotive and electronics customers to avoid volatilizing cyclosiloxanes during their manufacturing processes, thus their desire for a lower cyclosiloxane content in silicones. Similarly, the amount of volatile cyclosiloxane byproducts permissible in both leave-on and wash-off silicone personal care products continues to decrease due to new regulations in certain geographies.⁸ It is well known that the current suite of commercial catalysts, which are strong Brønsted acids (acid clays or phosphonitrilic chlorides) or bases (KOH), produce high levels of cyclosiloxanes under standard polymerization conditions (heating while removing water by applying vacuum, azeotropic distillation, or steam stripping).⁹ Thus, developing active and selective catalyst compositions based on milder acid/base complexes would represent a new avenue to address this issue.¹⁰

While surveying the literature for different catalyst compositions, a seldom used class of silanol polycondensation catalysts that are not strong acids or bases was noted: homoconjugated acids (Scheme 2).^11,12 Homoconjugation of acids—the formation of hydrogen bond between an acid and its conjugate base,¹³ also known as homoassociation, is known to increase the acidity of acids by stabilizing the corresponding conjugate bases—the case of HF is a well-known example.¹⁴ The homoconjugation constants of several acids have been measured in polar aprotic solvents, and in acetonitrile, homoconjugation is shown to increase the acidity of p-toluenesulfonic acid by a factor of 800 (2.8 pK_a units).^15,16

Scheme 2. Homoconjugated Acid Catalyst for Silanol Condensation Polymerization.

During the polymerization of siloxanes, cyclosiloxane byproducts are formed by several mechanisms, the most prominent being the “back-biting” mechanism where a silanol or silanoate attacks the siloxane backbone, forming a cyclosiloxane byproduct (Scheme 3). We hypothesized that homoconjugated acid catalysts may be able to selectively activate silanols vs the siloxane backbone, leading to less cyclosiloxane formation during silanol polycondensation. Homoconjugates of weaker acids, with enhanced acidity vs their component acids, can activate a silanol for condensation, whereas when attempting to protonate a siloxane backbone oxygen, the homoconjugate would be broken, resulting in the backbone being activated by a weaker acid (vs the homoconjugate) and a weak base (Scheme 4). This proposed catalyst system bears many similarities to cooperative catalysis, with both acids and bases being present in the reaction mixture, often in equilibrium between the individual components and the homoconjugated species. While the field of cooperative catalysis in polymerization is established,¹⁷ especially regarding ring-opening polymerization of cyclic esters and carbonates,¹⁸⁻²⁴ there are relatively few instances of cooperative catalysis to affect silanol polycondensation in the literature, and no mention of cyclosiloxane levels generated using these approaches.^11,12,25,26 Homoconjugated acid catalysts are cheap, simple to make, relatively mild, and commercially available. This provides ample room to develop new catalyst compositions and explore how the base and acid component of the homoconjugate relate to reactivity and selectivity.

Scheme 3. Backbiting in Acid-Catalyzed Silanol Polycondensation.

Scheme 4. Proposed Selective Silanol Condensation vs Cyclosiloxane Formation.

Results and Discussion

To begin investigating homoconjugated acids as silanol polycondensation catalysts, a screening setup was designed. Water removal from the polymer is critical to achieving high–molecular-weight polymers during silanol polycondensation, but this had to be balanced by ensuring any cyclosiloxane byproducts generated were not lost as measuring the levels of these byproducts was one of the key criteria by which catalysts would be evaluated. Thus, reactions were run in capped 40 mL scintillation vials with a small amount of polymer (5 g), allowing for a large headspace for water to accumulate and condense in, thus allowing the polymers to reach high molecular weights. The two values that were recorded for each polymerization reaction (each run was run in duplicate, with the values averaged) were M_w/M_n (and from this the dispersity, Đ) and the amount of D4 generated in parts per million. These were measured by GPC and headspace GC, respectively. A 100 ppm D4 standard was run with every headspace GC run to ensure the calibration was correct. D4 was chosen as the diagnostic cyclosiloxane byproduct to measure as it the main cyclosiloxane byproducts produced (along with decamethylcyclopentasiloxane, D5, and dodecylmethylcyclohexasiloxane, D6, and larger cyclosiloxanes). The intent of this investigation was to synthesize a viscous material (which is typically difficult to remove cyclosiloxanes from with stripping) with a low level of D4. As a metric of success, synthesizing a material with a viscosity of ∼60,000 cSt was chosen, which corresponds to a polymer with a degree of polymerization (DP_n) of >750. To achieve this DP_n, >95% conversion of the starting material, DOWSIL Q1-3563, a low viscosity α,ω-disilanol prepolymer with low D4 content (∼50 ppm) and a DP_n of ∼35 was required. Catalyst performance will be reported as % conversion of the initial amount of silanol groups in this report (see the Supporting Information for more detail).

The catalyst studies began by evaluating a catalyst that showed promising activity in the literature: tetramethylguanidine bis(2-ethylhexanoic acid), [TMG][EHA]₂ (Table 1).¹¹ All homoconjugates were prepared by simply mixing the desired equivalents of acid with one equivalent of base. Under the initial reaction conditions (1.3 mol % catalyst with respect to moles of starting silanol, 110 °C, 2 h, sealed vial), a polymer with a moderately high DP_n (373, 91% conversion of SiOH groups) and a very low amount of D4 (59 ppm generated with 50 ppm coming in with DOWSIL Q1-3563) was generated. Screening of other guanidines followed as Waymouth and others had shown major reactivity differences of guanidines (in their case, ring size in cyclic guanidine catalysts) in the amidation of esters and lactide polymerization.^19,27

Table 1. Initial Screen of 2:1 EHA:Guanidinium Homoconjugate Catalyst Compositions and Related Control Reaction in Silanol Polycondensation^e.

entry	catalyst	Mn (Da)	Mw (Da)	Đ	conversion (%)	D4 (ppm)	HB+ pKa28a
1	[TMG][EHA]2	27,900	51,100	1.8	91	109	23.3
2	[tBuTMG][EHA]2	6800	10,700	1.6	61	70	25b
3	[TBD][EHA]2	3900	7200	1.8	33	118	26
4	[7MeTBD][EHA]2	20,100	35,700	1.8	87	108	25.5
5	[G][EHA]2	3900	6300	1.6	33	41
6	TMG	3000	4600	1.5	14	54
7	TBD	5900	9600	1.6	55	661
8	7MeTBD	7700	13,400	1.7	66	2310
9	EHAc
10d	LPNC	38,500	75,700	2.0	93	3668

^apK_a values in MeCN.

^bpK_a value for MeTMG.

^cDid not polymerize.

^dCatalyst loading = 0.0013 mol %.

^eReactions run with 5 g of silanol in capped 2 oz. vials at 110 °C for 2 h with magnetic stirring.

Table 1 shows that the type of guanidine base in the homoconjugate has a large effect on catalysis. Tetramethylguanidine (TMG) shows by far the best activity when made into a homoconjugate with 2 equiv of 2-ethylhexanoic acid ([TMG][EHA]₂, entry 1) and stands in sharp contrast to tert-butyl tetramethylguanidine ([^tBuTMG][EHA]₂) and triazabicyclodecene [TBD][EHA]₂ (entries 2 and 3). 7-Methyl triazabicyclodecene ([7MeTBD][EHA]₂, entry 4) showed comparable reactivity as [TMG][EHA]₂. When moving to a weaker base such as guanidine ([G][EHA]₂, entry 5), very little reactivity was observed, but no clear pK_a trend is evident among the strong guanidine bases. As was hypothesized, homoconjugation is important to activity, with the control reactions (entries 6–9) showing that none of the individual homoconjugate components are particularly good catalysts. Cyclic guanidines themselves do show some reactivity (entries 7 and 8) but generate large amounts of D4. The importance of homoconjugation would also explain entry 5. The homoconjugation of an acid increases with increasing strength of the base, and guanidine is the weakest base included in the screen.²⁹ Finally, as a comparison, a commercial acidic homogeneous condensation catalyst, a linear phosphonitrilic chloride (LPNC), is shown in entry 10. While active at very low loadings, it generates unacceptable levels of D4 during polycondensation.

Having shown that homoconjugated acids indeed give good activity, further control reactions were performed by varying the stoichiometry of the homoconjugate (Table 2). A range of catalyst compositions varying the equivalents of acetic acid (AcOH) to TMG were screened. EHA and AcOH have essentially the same pK_a and acetic acid complexes were used going forward as they could be removed more easily via vacuum stripping post-reaction in a manufacturing setting. Using volatile carboxylic acids as removable catalysts is an established strategy in silicones manufacturing.³⁰ The 0.5:1 complex (entry 1) did not show appreciable activity. The 1:1 complex (entry 2) and 1.5:1 complex of AcOH:TMG (entry 3) show decreased reactivity vs the 2:1 homoconjugate (entry 4) and homoconjugates with additional acid (entries 5 and 6). As the formation of the homoconjugated acid is in equilibrium with the free acid and the base, it is hypothesized that with a weaker acid, such as acetic acid, some of the homoconjugated species is forming below the idealized 2:1 acid:base ratio, explaining the reactivity in entries 1–3. Entries 4–6 show that modest gains in conversion correspond by increasing the acid past the 2:1 ratio result in spiking levels of D4. Literature indicates that the acid strength has a strong impact on homoconjugate formation (as well as silanol condensation rate),³¹ and thus several homoconjugates across a broad pK_a range of acids were synthesized and screened (Table 3).

Table 2. Effect of Varying Equivalents of Acetic Acid (AcOH) vs Tetramethylguanidine (TMG) in a Homoconjugate Catalyst Composition in Silanol Polycondensation.

entry	AcOH:TMG	Mn (Da)	Mw (Da)	Đ	conversion (%)	D4 (ppm)
1	0.5:1	3100	6100	2.0	16	40
2	1:1	7600	13,700	1.8	66	76
3	1.5:1	12,200	23,700	1.9	79	121
4	2:1	19,000	37,500	2.0	86	149
5	2.5:1	19,600	39,300	2.0	87	154
6	3:1	22,800	28,500	1.3	89	168

Table 3. Effect of Varying the Acid Component of a 2:1 Acid:TMG Homoconjugate Catalyst in Silanol Polycondensation.

entry	catalyst	mol % cat.	Mn (Da)	Mw (Da)	Đ	conversion (%)	D4 (ppm)	acid pKa32a
1	[TMG][AcOH]2	1.3	18,900	37,500	2.0	86	149	4.76
2	[TMG][TFA]2	0.13	83,600	164,500	2.0	97	144	–0.25
3	[TMG][TsOH]2	0.13	40,200	74,000	1.8	93	252	–2.6b
4	[TMG][PFBSA]2	0.0013	9800	16,400	1.7	73	217	–14c
5	[TMG][AcOH][TFA]	1.3	3400	6700	2.0	22	41d

^apK_a values in water.

^bpK_a value for MeSO₃H.

^cpK_a value for CF₃SO₃H.

^dIt is possible that some D4 was lost during the reaction or analysis, leading to a value lower than that of DOWSIL Q1-3563 (50 ppm).

Table 3 shows the screen of a variety of acids across a wide pK_a range using TMG as the base in the homoconjugate. In this table, the acids are listed by decreasing pK_a and the catalyst loading was adjusted for more active catalysts so that the vials would remain stirring for the duration of the 2 h reaction time. A strong impact of pK_a relative to activity is observed, with trifluoroacetic acid (TFA) (entry 2) showing improved reactivity at a loading an order of magnitude lower than acetic acid (entry 1). Remarkably, this composition only produced 94 ppm of D4 over the 2 h run time (144 ppm total D4, including the 50 ppm present in DOWSIL Q1-3563), and this composition was able to make a polymer viscous enough to satisfy the initial criteria (60,000 cSt, >95% conversion). Further decreasing the pK_a of the homoconjugated acid resulted in somewhat decreased reactivity for p-toluenesulfonic acid (TsOH, entry 3) and perfluorobutanesulfonic acid (PFBSA, entry 4). These acids were used vs methanesulfonic acid and triflic acid, respectively, due to ease of handling and enhanced safety profile in the lab setting, although in manufacturing, methanesulfonic or triflic acid would likely be used due to cost and volatility (ease of removal using vacuum stripping). It should be noted that, even though [TMG][PFBSA]₂ is used at a much lower loading than other catalysts due to its high reactivity, it still generates a fair amount of D4, and when this catalyst is run at an order of magnitude higher loading (0.013 mol %), it generates ∼5000 ppm of D4.

Entry 5 shows that heteroconjugates between strong and weak acids will not give reactivity in between the reactivity of the respective pure homoconjugates. In this situation, TMG would deprotonate TFA preferentially, essentially rendering the catalyst as tetramethylguanidinium trifluoroacetate and free acetic acid (vs a homoconjugate, Scheme 5). As acidity influences homoconjugation and activity of the homoconjugated acids for silanol polycondensation, a screen varying equivalents of strong acid vs TMG was performed to see how a stronger acid would compare with a weak acid (AcOH). Here, the catalyst concentration was cut in half (vs Table 2) in hopes of showing larger differences between acid:base catalyst compositions. The results are shown in Table 4.

Scheme 5. Different Behavior of the Homoconjugate Catalysts When Complexing TMG with 2 equiv of AcOH vs 1 equiv of TFA and 1 equiv of AcOH.

Table 4. Effect of Varying Equivalents of Trifluoroacetic Acid (TFA) to Tetramethylguanidine (TMG) in a Homoconjugate Catalyst Composition in Silanol Polycondensation.

entry	TFA:TMG	Mn (Da)	Mw (Da)	Đ	conversion (%)	D4 (ppm)
1	0.5:1	2900	5200	1.8	6	51
2	1:1	2800	5100	1.8	7	51
3	1.5:1	7100	12,100	1.7	63	69
4	2:1	23,200	44,800	1.9	89	95
5	2.5:1	35,600	69,400	1.9	93	105
6	3:1	59,000	116,400	2.0	96	132

Unlike the case with acetic acid shown in Table 2, the base homoconjugate of TMG (entry 1) and the 1:1 complex (entry 2) are essentially inactive catalysts with TFA. This is due to the equilibrium between free acid/homoconjugate and salt lying heavily toward the salt due to the lower pK_a of trifluoroacetic acid. Entry 3, with a 1.5:1 ratio of TFA:TMG, displays modest reactivity but nowhere near some of the better candidates. Entries 4–6 show greatly increased reactivity vs entries 1–3, with excess acid beyond a 2:1 TFA:TMG ratio (entries 5 and 6) achieving higher conversion and only a modest increase in D4 production. Again, a polymer satisfying the viscosity specification was produced, this time at an even lower catalyst loading than used in Table 3. Figure 1 shows a comparison of Tables 2 and 4, where conversion has been plotted vs the equivalents of acid in each table. The trend observed here is like that shown in Scheme 5, where strong acid/strong base complexes are poorer catalysts when the ratio of acid:base is less than 2:1 due to the equilibrium between homoconjugate and salt complex/free acid lying heavily toward the salt complex/free acid. In contrast, with a weaker acid (acetic acid), some of the homoconjugate complex can form when less than 2 equiv is present with TMG, affording a better catalyst. This is shown in Scheme 6 when considering a 1:1 acid:base ratio. However, when a 2:1 ratio of acid to base is reached, the strong acid/strong base homoconjugate outperforms the weak acid/strong base homoconjugate.

Figure 1.

Conversion vs equivalents of acid for strong and weak acids.

Scheme 6. Weak and Strong acid Reactions with TMG Relating to Their Performance as Silanol Polycondensation Catalysts.

In addition to guanidinium homoconjugated acids, tetrabutylammonium homoconjugated acids have been show in the patent literature to be competent silanol polycondensation catalysts.^25,26 Now knowing that guanidinium homoconjugated acids generate very low levels of D4 during silanol polycondensation, cation-stabilized homoconjugates like tetrabutylammonium were investigated at the same catalyst loadings to see if the presence of a protonated conjugate base (tetramethylguanidinium) was playing a role in catalysis or D4 formation (Table 5). With a cation-stabilized homoconjugate, a strong acid at a 2:1 acid:cation molar ratio should exist exclusively as the homoconjugate as there would be no equilibrium between protonated base and free acid. If the homoconjugate was solely responsible for the catalysis, one would expect to see very similar reactivity for all cationic homoconjugates. Tetraalkylammonium and related cations have the added practical benefit of removal from the final polymer via thermally decomposition post-reaction (vs vacuum stripping). These catalysts are made by simply mixing an acid with an equivalent of its tetraalkylammonium salt. The acid (TFA) was held constant for these studies. When comparing TMG (entry 1) to tetrabutylammonium (TBA, entry 2), it was observed that the reactivity is very similar, but more D4 is generated in the case of TBA. Moving to tetrabutylphosphonium (TBP), one can see almost identical reactivity (vs TMG) but again more D4 formation. In the case of ethyl methylimidazolium (EMIM, entry 4) and Proton Sponge (PS, entry 5), decreased reactivity vs entries 1–3 is observed. From this comparison, one can conclude that the cationic component (either protonated base or aprotic soluble cation) plays a role in both reactivity of the homoconjugate and the formation of D4.

Table 5. Effect of Varying the Cation in a 2:1 TFA:Cation Homoconjugate Catalyst in Silanol Polycondensation.

entry	catalyst	Mn (Da)	Mw (Da)	Đ	conversion (%)	D4 (ppm)
1	[TMG][TFA]2	83,600	164,500	2.0	97	144
2	[TBA][TFA]2	61,200	115,700	1.9	96	307
3	[TBP][TFA]2	81,200	155,300	1.9	97	231
4	[EMIM][TFA]2	19,600	36,200	1.8	87	142
5	[PS][TFA]2	37,700	68,900	1.8	93	152

For homoconjugate-catalyzed silanol polycondensation, a mechanism with acid activation of silanols is proposed (Scheme 7). Wanting to observe the catalysts in the solid state, the crystal structure of the tetrabutylammonium acetate/acetic acid homoconjugate was found³³ and a crystal of a similar complex using tetramethylguanidine was grown (Figure 2). 4-Bromobenzoic acid was used instead of acetic acid as the [TMG][AcOH]₂ complex is a liquid. In the solid state, the guanidinium (Figure 2B) differs from the ammonium homoconjugate (Figure 2A) in that it binds the carboxylate homoconjugate in a 10-membered ring cyclic structure with participation from the guanidinium protons (see the Supporting Information for its molecular structure) vs the linear adduct. One can speculate this difference accounts for the drop in D4 production in guanidinium homoconjugates (vs the linear adducts) by preorganizing the silanols for condensation and proton transfer and thereby further disfavoring silanol back-biting. It should be noted that water, present in trace amounts due to silanol condensation and known to form higher order heteroconjugates with homoconjugated acids,³⁴ likely plays a role, which is complex and not currently understood.

Scheme 7. Proposed Reaction Mechanisms. (A) Unorganized Mechanism. (B) Guanidinium-Organized Mechanism.

Figure 2.

Crystal structures of (A) [TBA][AcOH]₂ and (B) [TMG][C₆H₅O₂Br]₂. Carbon = gray; oxygen = blue; nitrogen = red; bromine = gold; hydrogen = white.

In Scheme 7, two mechanisms are proposed: one organized by the guanidinium and one where only the homoconjugated acid plays a role. In the guanidinium system, both mechanisms may be occurring, and the 14-membered transition state depicted, though not without precedent,^35,36 may be one of several occurring with participation from the guanidinium. In each mechanism, it is proposed that a silanol is protonated by the homoconjugate (I), and a second silanol attacks the activated silicon, leading to siloxane bond formation and with the homoconjugate facilitating proton transfer, (II) resulting in generation of water and regeneration of the homoconjugate (III). While both mechanisms are speculative and water and silanols can also hydrogen bond with the acids and bases present as catalysts, data in Tables 2 and 4 and depicted in Scheme 6 show that it is necessary to have a ratio of acid (or cation) to base of 2:1 (a homoconjugate) or greater for good reactivity. This would indicate that a homoconjugated species, likely present in equilibrium with various silanol and water heteroconjugates, is playing an important role in the catalysis occurring. We also propose that homoconjugates selectively activated silanols vs the siloxane backbone as the latter breaks the homoconjugate, leaving weaker acids and bases to facilitate “back-biting” to afford cyclosiloxanes.

Conclusions

In conclusion, it has been shown that homoconjugated acids are an underutilized class of silanol polycondensation catalysts that are highly selective for silanol polycondensation vs cyclosiloxane formation to byproducts like D4. One can extract the following trends about homoconjugated acid-catalyzed polycondensation of silanols. Both the acid and the base (or cation) play a role in catalyst activity of the homoconjugate and the amount of D4 generated during polymerization. Homoconjugated acids are far more reactive than their respective acid or base components, and most reactive at compositions at or above a 2:1 acid:base ratio. Homoconjugated strong acids show increased activity vs weaker acids, and strong bases or soluble cations are also needed for high conversion. Acids with pK_a’s lower than trifluoroacetic acid and some bases (TBD) cause elevated D4 generation. Finally, homoconjugates of acids are better catalysts than heteroconjugates of acids (two different acids complexed with one base), and the data shows that homoconjugated acids with protonated conjugate bases (vs aprotic cations) generate less D4. Their highly tunable nature, ease of preparation, and unique reactivity warrant further investigation, both in the silicone field and other polycondensation reactions.

Experimental Section

Materials and Methods

All materials were used as received. 1,1,3,3-Tetramethylguanidine (99%), guanidine acetate salt (≥99%), 2-tert-butyl-1,1,3,3-tetramethylguanidine (≥97%), 1,5,7-triazabicyclo[4.4.0]dec-5-ene (98%), 7-methyl-1,5,7-triazabicyclo[4.4.0]dec-5-ene (98%), 2-ethylhexanoic acid (≥99%), trifluoroacetic acid (ReagentPlus, 99%), p-toluenesulfonic acid monohydrate (ACS Reagent, ≥98.5%), 4-bromobenzoic acid (98%), nonafluorobutane-1-sulfonic acid (97%), tetrabutylammonium acetate (97%), tetrabutylammonium iodide (reagent grade, 98%), Proton Sponge (99%), and 1-ethyl-3-methylimidazolium trifluoroacetate (for catalysis, ≥97.0% (HPLC)) were obtained from Millipore Sigma. DOWSIL Q1-3563 Fluid (lot #0008893890) and octamethylcyclotetrasiloxane (DOWSIL 244 Fluid) were obtained internally but available commercially from Dow. Toluene (HPLC grade), dichloromethane (Optima), and acetic acid (Glacial, certified ACS), and 19 Fluid Vacuum oil were obtained from Fisher. Tetrabutylammonium iodide (98%) was obtained from Alfa Aesar.

The molecular weight distribution of the polymers prepared by silanol polycondensation was determined by gel permeation chromatography (GPC) using an Agilent Technologies 1260 Infinity chromatograph and toluene as a solvent. The instrument is equipped with three columns: a PLgel 5 μm 7.5 × 50 mm guard column and two PLgel 5 μm Mixed-C 7.5 × 300 mm columns. Calibration was made using polystyrene standards. Samples were made by dissolving polymer in toluene (∼1 mg/mL), filtering through a Millipore Millex-FG, PTFE, 0.2 μm filter, and then immediately analyzing the material by GPC (50 μL injection, 1 mL/min flow, 35 °C column temperature, 25 min run time).

The amount of D4 in polymers was determined by headspace gas chromatography (GC) using a Hewlett-Packard 6890 GC with an RTX-1 column (30 m × 320 μm × 0.25 μm) with a Perkin-Elmer TurboMatrix 40 headspace unit. Samples were prepared by adding 1 mL of internal standard solution (0.01% dodecane by weight in Fisher Brand 19 fluid vacuum oil) to 20 mL headspace vial (with an Eppendorf repeater pipet). A 100 mg solution of D4 standard (100 ppm) or 100 mg of experimental sample is added to the headspace vial. Quantitation of the D4 content is by the single-point internal standard method. A relative response factor (RRF) of D4 relative to dodecane is established and updated every time a new batch of internal standard solution is prepared. The amount of D4 in the samples is determined within the Thermo Atlas data system according to an equation (eq 1) of the same type as the one below:

1

Experimental Procedures

General Screening Procedure

To a 40 mL scintillation vial(s) was added 5 g of DOWSIL Q1-3563 silanol fluid and a 1/8″ × 1/2″ stir bar. The vial(s) were placed in a “pie plate” heating block on a heated stir plate set to the desired temperature. Stirring was set at 500 rpm. After the temperature equilibrated to 110 °C, 100 μL of a methylene chloride catalyst solution was injected into the vial(s). The vial(s) were capped and stirred at 110 °C for 120 min. The vial(s) were then removed from the heating block, and a GPC sample was prepared. The vial(s) were then placed in a refrigerator until headspace D4 measurement was performed on the sample.

Screening Procedure Example (Table 2)

Five grams of DOWSIL Q1-3563 silanol fluid and a 1/8″ × 1/2″ stir bar was added to 13 × 40 mL scintillation vials. The vials were placed in a “pie plate” heating block on a heated stir plate set to 110 °C. Stirring was set at 500 rpm. After the temperature equilibrated to 110 °C, 100 μL of a catalyst solution was injected into each of the vials. Reactions were run in duplicate; thus, six catalyst solutions were screened, and catalyst was omitted from one vial as a control. The catalyst solutions consisted of a 0.5 M solution of the following complexes in dichloromethane: [TMG][AcOH]_0.5, [TMG][AcOH]_0.5, [TMG][AcOH]_1.5, [TMG][AcOH]₂, [TMG][AcOH]_2.5, and [TMG][AcOH]₃. The vials were capped after catalyst addition and stirred at 110 °C for 120 min. After 120 min, the vials were removed from the heating block and GPC samples were prepared. The vials were then placed in a refrigerator until headspace D4 measurement was performed on each sample.

General Catalyst Complex Procedure

Catalyst complexes were mixtures of previously reported and/or commercially available compounds. To a 20 mL scintillation vial was added 10 mmol of base or salt. Then, x mmol (y equiv) of acid was added slowly. The mixture was shaken on a speed mixer and stored at room temperature. All catalysts were prepared in this method or by otherwise known literature procedures except [TBP][TFA]₂.

Example Catalyst Complex Procedure ([TMG][TFA]₂)

To a 20 mL scintillation vial was added 1.15 g (10 mmol) of tetramethylguanidine. Then, 1.5 mL of trifluoroacetic acid (20 mmol, 2 equiv) was added slowly. The mixture was shaken on a speed mixer and stored at room temperature.

[TBP][TFA]₂

Tetrabutylphosphonium iodide (3.84 g, 10 mmol) was dissolved in 10 mL of acetone in a 40 mL scintillation vial. To this vial was added silver trifluoroacetate (2.2 g, 10 mmol). Immediately, a pale yellow precipitate formed. The contents of the flask were filtered through a disposable 20 μm frit and washed 2 × 10 mL with acetone.³⁷ To the filtrate was added trifluoroacetic acid (0.77 mL, 1.14 g, 10 mmol), and acetone was removed under reduced pressure. The resultant product was a low (room temperature) melting solid.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.0c03883.

• Supporting information including calculation of silanol conversion, spectra, and crystal structure data (PDF)

• Crystal data and structure refinement for 2017_07_19_17_BN6058_roberts_0 (CIF)

Author Contributions

The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript.

This work was internally funded by Dow Inc.

The authors declare no competing financial interest.