Poly(Boc-acryloyl hydrazide): The importance of temperature and RAFT agent degradation on its preparation

Poly(acryloyl hydrazide) is a versatile polymer scaffold readily functionalised through post-polymerisation modification with aldehydes to yield polymers for biological applications. However, its polymerisation is affected by nucleophilic degradation of the RAFT agent that leads to early termination, an issue often overlooked in the polymerisation of primary acrylamides. Here we report the effect of temperature on the RAFT polymerisation of N’-(tert-butoxycarbonyl)acryloyl hydrazide (1) and demonstrate that by carefully selecting this polymerisation temperature, a balance between rate of polymerisation and rate of degradation of the RAFT agent can be achieved. This way a greater control over the polymerisation process is achieved, allowing the synthesis of Boc-protected poly(acryloyl hydrazide) with higher degrees of polymerisation than those achieved previously, while still maintaining low dispersities. We believe our results should be of importance to those working on the RAFT polymerization of primary and secondary (meth)acrylamides and monomers with nucleophilic moieties. Abstract: Poly(acryloyl hydrazide) is a versatile polymer scaffold readily functionalised through post-polymerisation modification with aldehydes to yield polymers for biological applications. However, its polymerisation is affected by nucleophilic degradation of the RAFT agent that leads to early termination, an issue often overlooked in the polymerisation of primary acrylamides. Here we report the effect of temperature on the RAFT polymerisation of N’-( tert -butoxycarbonyl)acryloyl hydrazide ( 1 ) and demonstrate that by carefully selecting this polymerisation temperature, a balance between rate of polymerisation and rate of degradation of the RAFT agent can be achieved. This way a greater control over the polymerisation process is achieved, allowing the synthesis of Boc-protected poly(acryloyl hydrazide) with higher degrees of polymerisation than those achieved previously, while still maintaining low dispersities. We believe our results should be of importance to those working on the RAFT polymerization of primary and secondary (meth)acrylamides and monomers with nucleophilic moieties.

greatly broaden the scope of chemical functionalities used. Post-polymerization modification has normally relied on click chemistries, 12 and has now been greatly expanded through the use of oxime 13 and hydrazone chemistry, 14,15 reductive amination, 16 and epoxide ring opening. 17 A common limitation when developing synthetic polymers for biomedical applications is the need to screen large libraries of compounds which is costly and time consuming. In this regard, poly(acryloyl hydrazide) has been recently reported as a versatile platform for the synthesis and screening of polymers for biomedical applications. 14,[18][19][20] Functional polymers are obtained by simple incubation of poly(acryloyl hydrazide) with functional aldehydes, both under aqueous or organic conditions, 14 and this polymer has now been applied to the development of glycan arrays, 18 pH sensitive drug-delivery, 21 and nucleic acid delivery. 20,22,23 In our laboratories poly(acryloyl hydrazide) was prepared from Boc-protected precursor Boc-Px (Scheme 1) following deprotection with TFA. 14 Reversible Addition-Fragmentation (RAFT) polymerisation of N'-(tert-butoxycarbonyl)acryloyl hydrazide (1) resulted in a small library of polymers. However, control over the polymerisation was lost with increasing conversion and degree of polymerisation, possibly as a result of degradation of the RAFT agent through intramolecular nucleophilic attack. This degradation has been reported in the RAFT polymerisation of other acrylamide derivatives, 24,25 including closely related methacryloyl hydrazide, 26 with better control reported when the polymerisation is carried out at low temperatures. 25,27 This sidereaction is often overlooked in the polymerization of primary and secondary acryl-and methacrylamides, and makes synthesising highly functional polymers from this type of monomers inherently challenging. 28 The need for greater control over these materials is more significant when looking to understand better the nature of the structure-activity relationship throughout post-polymerisation modification and biological screening. Scheme 1: RAFT polymerisation of N'-(tert-butoxycarbonyl)acryloyl hydrazide (1) and potential degradation by-products.
Here, we report the effect of temperature and the decomposition rate of the initiator on the polymerisation of N'-(tert-butoxycarbonyl)acryloyl hydrazide (1), as a route to optimise the preparation of poly(acryloyl hydrazide). Polymerisations were carried out using 2,2'-azobis [2-(2-imidazolin-2-yl)propane]dihydrochloride (VA-044) as a low temperature initiator, so that the rate of generation of radicals could be readily modified as a function of temperature. Our results suggest that while increasing the temperature increases the polymerisation rate, it also speeds up RAFT degradation and thus, loss of control. Conditions have been identified for which the polymerisation "outperforms" this side reaction and polymers with good control over molecular mass and dispersities (Đ) can be obtained. More importantly, these conditions allowed us to prepare Boc-PX with higher degrees of polymerisation and lower dispersities (Đ), not accessible with our previous conditions. 14 We believe our results highlight the importance of balancing polymerisation kinetics and RAFT agent degradation in the polymerisation of monomers containing nucleophilic moieties such as acrylamides and methacrylamides. Moreover, this improved control over the polymerisation of Boc-protected poly(acryloyl hydrazide) will be of value when degree of polymerisation and dispersity may underpin future applications. were then left to react at a pre-set temperature (30-150 degrees ºC) for the required amount of time. The reaction was stopped by allowing the tube to cool using a water bath and exposing it to air. 100 μL aliquots of each timepoint were taken at this stage to calculate conversion (ρ) and for GPC analysis. NMR and GPC analysis of each timepoint was carried out from the crude mixture.
The natural logarithm of the inverse of the fractional concentration of monomer -ln(M0/Mt)was plotted against time, and the data fitted using GraphPad Prism version 6.0 for Mac Os X, GraphPad Software, La Jolla California USA, www.graphpad.com. The in-built segmental line regression was used to fit the data to two intersecting lines. This model was used to identify when a change in the polymerisation kinetics was observed (tdead).

Results and discussion
As reported, our initial efforts to optimise the polymerisation of Boc-protected acryloyl hydrazide 1 focused on reducing the temperature of the polymerisation. 14 RAFT polymerisation of acrylamides and methacrylamides often suffer from cleavage of the RAFT agent through intramolecular addition-elimination of the weakly nucleophilic amides to the trithiocarbonate group (Scheme 1). 25 In our previously reported conditions for the polymerisation of 1, a change in the rate of polymerisation was observed with increasing conversion ( Figure 1A), which we associated with this degradation of the terminal trithiocarbonate in the growing chain. It has been proposed that reducing the polymerisation temperature would significantly reduce the rate of this side reaction. 25 Thus, optimisation of the polymerisation was at that time carried out under the same conditions but using initiators with different 10 hour half-life decomposition temperatures (t10) ( Figure 1A). This way, rate of formation of radicals was kept as similar as possible for all polymerisations while reducing the temperatures to 50 ºC (V-65) or 44 ºC (VA-044). Despite the use of lower temperatures, in all cases, a change in the kinetics of the polymerisation was observed, although this change was not as obvious for the polymerisations performed at 44 ºC ( Figure 1A). To identify when this change in rate of polymerisation was occurring, the natural logarithm of the inverse of the fractional concentration of monomerln(M0/Mt) -was plotted against time, and the data fitted to a segmental line regression. This function fits the data to two different lines, before and after a breakpoint. In our case, we termed the breakpoint tdead because we think that after this point, side reactions have a predominant effect on the kinetics of the polymerisation resulting in an increasing number of dead polymer chains. This change in kinetics was reflected on the relatively high dispersity in molecular mass (Đ =1.38-1.95) obtained for the polymers prepared with these conditions. 14 Overall, no clear benefit from reducing the temperature was observed, with a tdead of approximately 4 and 4.5 hours for polymerisations at 50 ºC and 70 ºC respectively. Interestingly, tdead for the polymerisation performed at 44 ºC was observed at approximately 2.5 h, which would suggest degradation was occurring faster at this temperature. This was not expected and may suggest that other mechanisms beyond the simple degradation of the RAFT agent may be at play. Attempts to perform the polymerisation at an even lower temperature (30 ºC) using VA-044 as the source of radicals resulted in a very long induction period followed by a short period of linear increase of the fractional concentration of monomer until a change in kinetics was again evident ( Figure S1). The maximum conversion in this case was 50% -ln(M0/Mt) = 0.83, worse than that observed for the polymerisations performed at higher temperatures.
In order to determine if degradation of the RAFT agent was indeed possible at low temperatures, we attempted to synthesise a small molecule analogue which mimicked an n=1 polymer (Scheme S1). To this end, 2-bromopropionic acid (2)  where only traces of something that could resemble trithiocarbonate 4 could be identified ( Figure   S3). This observation was in line with our previous results, and suggested that hydrazide containing trithiocarbonates such as 4 were very amenable to intramolecular nucleophilic attack.
Attempts to isolate this trithiocarbonate 4 were unsuccessful, with the main isolated product of this reaction being tentatively assigned to a mixture of the 5-and 6-membered rings in a 6:4 ratio ( Figure S4).  These results were promising and we therefore explored decreasing the concentration of initiator in our polymerisations, in an attempt to suppress termination, increase the number of chains growing from the RAFT agent and thus optimising the dispersities. However, while dispersities were decreased, reducing the concentration of initiator in these polymerisations resulted in slower reactions, with no effect observed in tdead (Figure 2A). As a result, the maximum conversion obtained when the CTA:VA-044 ratio was increased to 10:1 or 15:1 (40% and 24% conversion respectively) was lower than in the previous case (70%).
We decided next to run the polymerisations at 150 ºC, in an attempt to further increase the concentration of radicals during early stages of polymerisation, and thus the rate of propagation.
However, these conditions not only resulted in lower conversions ( Figure S8) but a colour change of the reaction mixture from yellow to dark brown, suggesting that thermal decomposition of the trithiocarbonate group was ocurring. 33 Thermal decomposition of the RAFT agent was confirmed via 1 H-NMR where signals consistent with the β-elimination of the trithiocarbonate could be observed ( Figure S9). 33,34 Having identified conditions to run the polymerisation of 1 at 100 ºC, which resulted in similar conversions and dispersities to those previously reported, we decided to explore the use of these conditions to prepare polymers of higher Mw (Figure 3), which were harder to control using our priously reported method. 14 Three different DPs were targeted (i.e.   With encouraging results from the polymerizations at 65 ºC, we set out to probe the livingness of the polymer before and after tdead and whether this tdead was an indication of degradation of the RAFT agent. To this end, we isolated and purified two polymerisations of 1, one that had been stopped at intermediate conversions (ρ =47%, t=30 min), before tdead ( Figure S5 A) and one that was stopeed at maximum conversion (ρ =85%, t=120 min), after tdead (Figure S5 B). As expected, Boc-Px isolated before tdead was able to undergo complete chain extension with further addition of 1 and initiator ( Figure S5 A), thus demonstrating that at intermediate conversions the RAFT agent was still present in significant amounts. Boc-Px isolated after tdead showed no chain extension, instead showing a bimodal distribution of molecular mass and high dispersities (Figure S5 B) demonstrating that after tdead the RAFT group has been degraded. To probe if the RAFT agent degradation was temperature driven, we isolated and purified a second "living" Boc-Px at intermediate conversions (ρ =52% t=30 min) ( Figure S6). This polymer was then heated for 90 minutes under standard polymerisation conditions, but this time without addition of 1 and initiator. We anticipated that heating the polymer this way should result in degradation of the RAFT agent, a hypothesis that was confirmed upon attempting to chain extend this terminated Boc-Px. In this case, a shoulder was observed in the molecular mass and high dispersities indicating that the Boc-Px which had been subjected to further heating was "dead" ( Figure S6).
Additional evidence of the RAFT agent degradation was obtained from NMR spectroscopy, where the protons associated with both the R and Z end group of the polymer chain could be observed for the "living" Boc-Px whereas "dead" Boc-Px showed a loss of the Z group ( Figure S7).
At this point, we decided to evaluate if further improvement could be achieved by optimising the RAFT agent used. For an effective RAFT process where the majority of the polymer chains grow at the same rate, the reactivity of the propagating chain and the stability of the polymer-RAFT intermediate should be optimised such that the addition to the C=S and subsequent fragmentation has a higher rate than propagation. 28 Fast propagating monomers such as Seeing how running the polymerisations at 65 ºC using CTA1 gave the highest conversions (90%) at tdead of all the conditions evaluated, we decided to target different degrees of polymerisation using these conditions ( Figure 5). As before, targeting higher DPs resulted in slower rates of polymerisation, in particular for DP200 and DP300. While slower rates had a significant effect on the maximum conversion achieved (approx. 90%, 89%, 68% and 55% for DP 50, 100, 200 and 300 respectively), little effect was observed on the tdead, with most polymerisations "stopping" after 1 h ( Figure 5A). Under these optimised conditions, the polymerisations retained features of a controlled polymerisation, with the molecular mass of the polymers increasing linearly with conversion, narrow dispersities in molar mass ( Figure 5C) and good end group fidelity if isolated before tdead.
In all cases, the dispersities obtained were similar or lower to those reported previously. 14 This was particularly the case when targeting DPs of 100 and 200 with dispersities of <1.4 being observed at maximum conversion.

Conclusion
Here we have demonstrated the role of temperature and RAFT agent degradation in the polymerisation of N'-(tert-butoxycarbonyl)acryloyl hydrazide (1). Our results highlight that the polymerisation of acrylamides via RAFT can be severely hampered by the degradation of the chain transfer agent and that, under some circumstances, this degradation cannot be eliminated but rather outperformed if the rate of polymerisation is tuned. We demonstrate that by using a low-temperature initiator such as VA-044, optimal polymerisations conditions can be achieved at 65 ºC. This way, poly(N'-(tert-butoxycarbonyl)acryloyl hydrazide)s with high degrees of polymerisation could be obtained while still maintaining low dispersities. Finally, we demonstrate that for our system, no benefit is obtained when trithiocarbonates are replaced with dithioesters or trithiocarbamates, as the chain transfer agents.

Small molecule analogue of a DP= 1 of N'-(tert-butoxycarbonyl)acryloyl hydrazide (1).
Scheme S1: Attempted route for the synthesis of a DP= 1 analogue of N'-(tert-butoxycarbonyl)acryloyl hydrazide (1).   (20 ml) and was left stirring at room temperature for 10 minutes. CS2 (1.09 ml, 6.59 mmol) was then added and the reaction mixture was left for a further 10 minutes. tert-butyl 2-(2-bromopropanoyl)hydrazine-1-carboxylate (1) (1.6 g, 5.99 mmol) was added in one portion and the mixture left to react for 13 hours. The solvent was then removed under reduced pressure and HCl (100 ml, 1 M) was added to the crude of the reaction. The resulting mixture extracted into DCM (2 x 100 ml). The organic layer was then washed with water (2 x 100 ml) and brine (2 x 100 ml), dried with Na2SO4, filtered and the solvent removed under reduced pressure. The resulting orange oil was purified by column chromatography using a 7:3 ratio of diethyl ether and hexane, then dried under reduced pressure to leave a viscous orange liquid (0.12 g, 7 %) which consisted of two compounds, none of which is the title compound. a; 1H NMR (300MHz, CDCl3) d