Exploiting Visible Light Triggered Formation of Trans-cyclohexene for the Contra-thermodynamic Protection of Alcohols

Abstract

We report herein a method for the contra-thermodynamic protection and deprotection of alcohols in which all reagents are returned to their original state. This is accomplished by use of visible light photochemical energy to drive the formation of a highly strained trans-(Z)-cyclohexene. At STP the product ethers contain more potential energy than the starting materials, and thus, can be catalytically returned to the starting materials- effectively realizing a protection-deprotection scheme paid for with an energy currency.

Graphical Abstract

An important distinguishing feature of photochemistry (or photocatalysis) is the ability to pump energy into the system and molecular motion that is generally inaccessible on the ground state surface- recognized in the early days of the 20^th century.^1–7 Chemists have shown that visible light can drive alkene classes away from their thermodynamic equilibrium.^8–11 During the 1960’s and 1970’s Kropp, Bonneau, and Dauben studied the geometrical isomerization of cyclic alkenes which enabled a number of unique transformations that were due to captured strain.^12–18 Our recent efforts have focused on exploiting visible light to generate such strain and using it to accomplish synthetic objectives.^19–23

We are particularly interested in the development of molecules that can effectively, and repeatedly, convert photochemical energy to useable ground state potential- an energy “currency”.⁶ To this end, we are drawn towards cyclohexene, which is predicted to harvest a large portion of the available photochemical energy, approximately 52 kcal/mol.^24–28 Indeed, while transient, its existence has been established.¹⁶ The strained alkene was shown to undergo protonation with strong acids to form a cationic intermediate followed by nucleophilic solvolysis.^{12, 29, 30} While strong acids were used, we postulated that much of the strain energy would manifest itself in the form of enhanced basicity, and that trans-cyclohexenes could be protonated using much weaker acids, which could be exploited to accomplish synthetic work.

Seeking an application, we honed in on protecting group chemistry. As is the way, protection necessitates deprotection and two separate sets of reagents.^31–33 In both reactions, it is generally the choice of chemical reagents that drive the reaction forward. In contrast, we speculated that we could exploit contrathermodynamic photocatalysis to form a higher energy, but kinetically stable product in the form of an alcohol protected as a cyclohexyl ether. When desired, this could be reverted with acid catalysis to recover both the starting alcohol and the cyclohexene (TOC).

Indeed, performing DFT calculations on phenylcyclohexene, ethanol, and 3, using the B3LYP functional and 6-311+G(2d,p) basis set yielded minimized structures. A comparison of their energies revealed the reaction to be exothermic with a ΔH = −3.8 kcal/mol, but at room temperature, the free energy is dominated by the entropy term and the process is endergonic by +10.3 kcal/mol. Thus, we anticipated that this would be an ideal reaction to explore this concept as the forward reaction was unfavorable at temperatures above −205 °C.

After an optimization campaign (see SI for details), we settled on condition A, with which we were able to isolate the product of the model reaction in high yield (96%, 3a Table 1) and began exploring the scope. Electronic assessment of the phenyl ring of arylcyclohexene indicated that electron withdrawing substituents were less efficient. Thus, for the remainder of our study we limited ourselves to arylcyclohexenes with substituents that are slightly electronegative (m-F), to neutral (unsubstituted, p-F, and p-Me), which displayed only minor differences in reactivity. We next focused on alcohols which displayed substantial differences in reactivity. We subjected a diverse range of primary, secondary, tertiary alcohols and phenols to the reaction conditions which typically afforded the desired ethers in very good to excellent yields (Table 1). A 5 mmol-scale reaction performed with ethanol gave an excellent yield of the isolated ether, 3a (95%, 1.05 g). 2-Methoxy ethanol performed brilliantly affording the diether product 3b in 94% yield. Nucleophile sensitive alcohols that contain primary bromides and chlorides (3c, 3d) performed well. Acid sensitive alcohols (3f and 3g) were protected efficiently with no evidence of rearrangement. The lack of elimination, regiochemical- and geometrical-isomerization of 3e-g (or their starting alcohols) highlight the advantage of the mildly acidic conditions. Hydroxyacetone, which is prone to polymerization,³⁴ benzyl alcohol and 4-ethynylbenzyl alcohol were also smoothly protected with excellent yields (3h-3j).

Table 1.

Substrate Scope^a,b

^aReaction condition: alcohol (0.2 mmol), alkene (1.5 equiv), lr(dFppy)3 (0.25 mol%), pyridiniumhexafluorophosphate (10 mol%), DCM (0.4 M), argon atmosphere, blue LEDs;

^bIsolated yields;

^calcohol (5 mmol), alkene (1.3 equiv), DCM (2.5 M);

^d1 mmol scale;

^eAlkene (3 equiv);

^fA3 instead of A.

^gWithout A.

^h72 h.

Secondary alcohols (3k-3q) isopropanol, hexafluoro isopropanol and indanol underwent proficient etherification, giving 81-94% yields (3k-3m). The reaction of isopropanol was scaled to the 1 mmol scale without appreciable difference (3k). 1,2-propanediol having two adjacent primary and secondary hydroxyl groups can be doubly protected with excellent yield by doubling the equivalents of alkene (3p). Next, we protected the hydroxyl group of dehydroepiandrosterone (DHEA)- an endogenous steroid hormone precursor, which took place cleanly and without addition to, or transposition of, the alkene (3q). Of note, this substrate required reduced temperature (5 °C) and increased LED flux to achieve the protection in high yield.

We next turned our attention to tertiary alcohols. Initially, their protection failed. However, we found that the key to success was to decrease the temperature to 5 °C and to increase the LED flux. In doing this, we were able to synthesize tert-butyl ether, 3r, in 87% yield, as were 1-methylcyclopentanol and 1-adamantanol (3s, 3t). One limitation appears to be tertiary benzylic alcohols such as trityl alcohol where competitive cleavage of C–O bonds is possible.

Next, we turned to phenols which proved to be excellent substrates. We found that no external acid catalyst is required. Instead, phenols themselves are sufficiently acidic to protonate the trans-cyclohexene. While etherification of phenols does take place in the presence of the external nitrogenous acid, we found that higher yields were obtained in its absence. Different phenolic substitutions including an ester, ketone, aldehyde, and bromide were all well tolerated- illustrating the excellent functional group tolerance of the reaction (3u-3z).

We next investigated cinnamyl alcohol with its excitable double bond which could compete with the cyclohexene. The E-Z isomerisation of cinnamyl alcohol itself was more facile and faster transformation observed in the reaction. But prolonged irradiation lead to building up of the desired product (3aa) although in lower yield (56%).

We wanted to further assess the selectivity of the reaction. Modeled after Glorius‘ study,³⁵ we performed an intermolecular robustness screening (see SI for details). In general, we found the reaction is tolerant to electrophilic additives, but electron rich nucleophilic additives had a tendency to hinder the reaction by competitively reacting with 2a or the acid. However, most of the additives gave a high recovery- highlighting the mild conditions of the reaction.

We wanted to better understand the relative reactivities of the alcohols. Therefore, we performed competition studies to assess the relative rates of alcohol protection (Scheme 1). When ethanol, isopropanol and alkene 2a were taken in equimolar amounts and the mixture was subjected to the otherwise standard reaction conditions, we observed that an initially modest selectivity (Scheme 1a) in favor of the ethyl ether became highly selective over 20 h. From this experiment we concluded that, this remarkable selectivity arises not only from the forward reaction but also from the reverse reaction which prefers elimination from the more hindered ether. Stated differently, under acid catalysis both ether products can undergo thermodynamically favored elimination, but the rate of elimination from 3k is significantly greater than from 3a. This complements the relatively poor initial selectivity of the forward reaction, and as a result, 3a builds up in the reaction with time. A similar result was observed when an intramolecular competition study was performed with 1,2-propanediol (Scheme 1b). Only the primary hydroxyl group is protected with a 75% NMR yield (3ab) in just 2.5 hours which did not change with extended irradiation.³⁶ Moreover, a set of competition experiments between ethanol and phenol 1x was performed in the presence and absence of external acid A (Scheme 1c). With A, we observed a slight initial preference for the phenolic ether 3x which gradually switched to the alkyl ether (3a). Given that we had already observed that phenols were sufficiently acidic to mediate their own addition, we repeated the competition experiment sans external acid and at half concentration- anticipating that ion pairing could be exploited to enhance selectivity. Indeed, we observed an increased early selectivity for the phenolic ether (6.2:1). Thus, we could effectively protect either the alkyl alcohol- by the addition of external acid A and increasing the reaction time, or the phenolic alcohol by decreasing the reaction time and performing the reaction free of external acid under dilute conditions.

Scheme 1.

Competition studies

Next, we investigated the deprotection of the ether product which traditionally requires such forcing conditions that its utility is limited. Hindered dialkyl ethers are generally relatively difficult to synthesize via conventional methods, and potentially for this reason, their cleavage has remained underexplored in the literature.³⁷ We found that the alcohol could be smoothly liberated using catalytic amount of tosic acid (10 mol%), 10 equivalents of hexafluoroisopropanol (HFIP), or 2 equivalents of trifluoroacetic acid, all of which resulted in excellent yield of both the ethanol and the alkene, 2a (Scheme 2a). In addition to acid mediated deprotection, it also proved susceptible to Pd-catalyzed hydrogenolysis using 1 atm of H₂. While the alkene is saturated, it may be useful to have a non-acidic path to deprotection. Next, to establish the selectivity of deprotection between primary and secondary alcohol, we took a 1:1 mixture of ethers 3a and 3k, which were made from ethanol and isopropanol respectively, and studied their propensity to deprotect in presence of catalytic tosic acid (Scheme 2ci). After 1 hour, we observed the formation of isopropanol in 73% NMR conversion, while we could only detect trace amounts of ethanol. Similarly, doubly protected propylene glycol ether 3p, which contains both primary and secondary carbinols, underwent deprotection primarily at secondary carbinol center to yield 3ab (Scheme 2cii). Thus, deprotection of cyclohexylarylethers generally favors elimination of the more hindered aliphatic alcohol. Taking the competitive addition and elimination experiments together, indicates that the protection and deprotection are taking place dynamically and synergistically to enhance selectivity with time. Stated differently, the thermal back reaction is in competition with the photochemically driven protection. While the photochemical protection step is marginally selective for less hindered alcohols (~ 2:1, Scheme 1), the rate of deprotection is greater for the more hindered ether (~9:1, Scheme 2ci).Thus, since the minor ether product preferentially eliminates and undergoes repeated etherification, the initial selectivity improves with time. We expect that arylcyclohexyl ethers will provide a valuable complementary strategy to that of silyl ether deprotection (Scheme 2d). In general, the deprotection of silyl ethers tends to occur more readily from a less bulky parent alcohol site (i.e., primary over secondary).³⁸ In contrast, use of the arylcyclohexyl ethers allows selective deprotection of the bulkier carbinol. Because of this complementary selectivity, it may find utility in complex multistep synthesis. Importantly, the rate of the deprotection step, is extremely temperature sensitive. Meanwhile, we have seen that the formation of trans-cyclohexene is less thermally sensitive,⁸ and thus, lowering the reaction temperature is expected to accelerate the protection step.

Scheme 2.

Exploring the deprotection

Finally, we demonstrated the tolerance of the ethereal protecting group by subjecting protected substrates to a variety of chemical transformations and conditions. Specifically, we investigated transformations that were incompatible with the alcohol functionality (see SI for details). We performed Grignard and Suzuki–Miyaura reactions which contained highly basic organometallic groups, wet, and elevated temperatures with boronic acids. We also performed a Wittig reaction with refluxing toluene, suggesting that simple thermal ionization does not take place at these temperatures. We also performed reduction with sodium borohydride in protic solvents, suggesting that the ether is stable to hydrides and protic solvents. In all cases, the ether group survived. Conveniently, unlike the common protecting group tetrahydropyran, the symmetric nature of the cyclohexyl group means that the protected alcohols create no new stereocenters, and thus remain chromatographically and spectroscopically simple to handle.

Based on the experimental results and previous reports, we propose following mechanism for the protection-deprotection strategy of alcohol (Scheme 3). The triplet excited photocatalyst formed by blue LED excitation undergoes a Dexter energy transfer with the ground state of arylcyclohexene to form their triplet excited state and regenerate the ground state photocatalyst. The excited cyclohexene then undergoes rotation about the former C=C double bond and undergoes inter-system crossing back onto the ground state surface to form a twisted and highly strained trans-isomer or unproductively relax back to the cis-starting material. Meanwhile, the strained trans-cyclohexene is readily protonated by weak acids (pyridinium or phenols) to form the carbenium intermediate which is intercepted by alcohol to yield the corresponding arylcyclohexyl ether. This ether can be isolated and subjected to other reactions before returning it to the cycle. The overall cycle is possible because the ether possesses more potential energy (+10 kcal/mol at STP) than the starting materials, allowing the catalytic return of the (modified)-alcohol and arylcyclohexene.

Scheme 3.

Plausible Mechanism

In conclusion, we have explored the use of arylcyclohexenes as energy transducers and shown that they can be effectively used to protect alcohols- despite the endergonic nature of the reaction. With mild and operationally simple reaction conditions to both protect and deprotect, we expect that this will be a useful reaction given the complementary selectivity to existing protection strategies.

Data Availability Statement

The Supporting Information is available free of charge on the ACS Publications website.