Cubanes, Fenestranes, Ladderanes, Prismanes, Staffanes and Other Oligocyclobutanoids
Published Online: 15 DEC 2009
Copyright © 2009 John Wiley & Sons, Ltd. All rights reserved.
Patai's Chemistry of Functional Groups
How to Cite
Hopf, H., Liebman, J. F. and Perks, H. M. 2009. Cubanes, Fenestranes, Ladderanes, Prismanes, Staffanes and Other Oligocyclobutanoids. Patai's Chemistry of Functional Groups. .
- Published Online: 15 DEC 2009
Choose one or more boxes to highlight terms.
- Top of page
- Preparation of Oligocyclobutanoid Systems
- From Rigid to Flexible Oligo- and Polymeric Cyclobutanes
- References and Notes
- 2Preparation of Oligocyclobutanoid Systems
- 3From Rigid to Flexible Oligo- and Polymeric Cyclobutanes
- Top of page
- Preparation of Oligocyclobutanoid Systems
- From Rigid to Flexible Oligo- and Polymeric Cyclobutanes
- References and Notes
The cyclobutane ring system, readily available by numerous ring-formation processes1, is, as the present chapter shows, an interesting, versatile and unique building unit for the construction of complex organic molecules. (The effects of cyclobutyl groups on molecular properties are discussed by Charton in chapter Structural Effects of the Cyclobutyl Group on Reactivity and Properties of this volume). Depending on the type of conjoining the four-membered rings organic compounds of vastly different structural properties, shapes and functions may be constructed. Provided that these reactions can be extended to the preparation of oligomeric or even polymeric derivatives, we anticipate that these will possess very interesting physical properties; whereas, for example, some representatives will be very flexible, other will be very rigid organic materials.
As shown in Scheme 1, one of the simplest ways to connect cyclobutane rings is by way of a single bond, i.e. the two ring units have no carbon atom in common.
Beginning the aufbau with cyclobutane () via bicyclobutyl (2), the linear structures 3 are obtained. Of course, the third cyclobutyl ring must not be bonded opposite to the first cyclobutyl unit (in 3-position), but can also be connected at the geminal or vicinal position, giving rise to 1,1- (4) or 1,2-dicyclobutylcyclobutane (5), i.e. to angular oligocyclobutyls. It is easily seen that by this set of simple building rules an endless number of structures can be obtained—including cyclic ones such as 6 or branched (dendritic) structures—even if one makes use of only a small number of distinct cyclobutyl ‘tiles’. (cis- and trans-Substitution should be distinguished. Stereochemical and spectroscopic aspects of this are discussed in chapters Stereochemical Aspects—Conformation and Configuration and NMR Spectroscopy of Cyclobutanes by Berg, and Seidl and Diaz, respectively, in this volume.)
Of course, in principle the number of other atoms or bonds between the cyclobutane rings can be selected freely, again generating a huge (largely so far not realized) structural variety. For example, if vinylcyclobutane (7) is polymerized, the cyclobutyl substituted polyethylene 8 (see below, Section 3.2) is obtained (Scheme 2).
When bifunctional monomers such as 1,3-divinylcyclobutane (9) are employed as a substrate, this could lead to cross-linked polymers such as 10. Again, the monomeric building unit and the connectivity of the polymeric products can be made much more complex.
When the two four-membered rings share one carbon atom, spiro structures result, in the simplest case spiro[3.3]heptane (11). And again the building process can be carried out in many different directions: linearly as in 12, cyclically as in 13 or in 14, the former combination making use either of two neighboring carbon atoms as spiro centers or having these also in a 1,3-arrangement. One would expect smaller members of this series to be highly strained and, when substituents are introduced, many stereoisomers can be produced. A sterically interesting situation is illustrated by structure 14 (Scheme 3) in which a molecular helix is generated by the cyclobutane rings. And structure 15 illustrates the building of a rotane molecule, as a special case of a polyspiro compound from five cyclobutane subunits.
Fusing two cyclobutane rings with one bond (sharing of two cyclobutane carbon atoms) leads to bicyclo[2.2.0]hexane (16), from which—again—branching in different directions can occur. Proceeding with linear annelation leads to the ladderanes, 17, and continuing in angular or circular fashion provides fenestranes, 18 and 19. In fact, these cases are not so simple, as illustrated in Scheme 4, since the stereochemical situation at the common bonds has been neglected (see below).
Ladderane annulation leads to interesting results when extended to cyclic derivatives, which can be represented in symbolic general form by structure 20. In this category the simplest representative is prismane (21), followed by the higher homologs prismane (22, cubane), prismane (23), all the way up to oligomers such as the isomeric dodecaprismanes 24 and 25, which have also been called helvetane and israelane for obvious reasons, both hydrocarbons so far having been suggested as target molecules in a spoof paper only2. It should be noted that the prismanes themselves could also be used as building blocks for more complex polycyclobutane structures again. If these are conjoined by their faces, rod-like structures result (which are novel, highly strained forms of carbon if extend ‘to infinity’); if, on the other hand, they are connected via their edges, various branched structures can be designed (see Section 2.5).
A hydrocarbon in which two four-membered rings share three carbon atoms is bicyclo[1.1.1]pentane (26). When this is used as a monomer and connected to other bicyclo[1.1.1]units by single bonds, the so-called staffanes, 27, result (Scheme 5). As we shall see in Section 2.5, 26 and its derivatives are obtained from an even more strained hydrocarbon, [1.1.1]propellane.
Again, the connected units (spacers XY in 28) may by saturated—polymethylene chains, for example, or consist of functional groups—producing staffanes (which may also be cyclic, see Section 2.5)—of different rigidity and chemical reactivity. In modern aromatic chemistry as well as for the preparation of substructures of novel carbon allotropes, the acetylene group has often been used, since it can be introduced readily and can be dimerized or cyclooligomerized by various metal-mediated reactions3. Applying this approach to cyclobutanes could yield linear structures such as 29 or sheet-like oligomers such as 30, keeping in mind again that in both designs the stereochemical situation of the cyclobutane ring has not been considered.
A completely different polycyclobutanoid situation arises when the thymine units of DNA photodimerize, as shown in highly abbreviated form in Scheme 64. For a more complete discussion see chapter Cyclobutane Pyrimidine Dimers as UV-Induced DNA Lesions by Friedel, Gierlich and Carell in this volume.
In principle, the [2 + 2]photoaddition 31 32 can take place numerous times in a DNA double helix, giving rise to an oligo or polycyclobutanoid system whose four-membered rings are separated by highly functionalized heteroorganic spacer units. In fact, derivatives such as 32 are not the only polycyclobutane derivatives that occur in nature, several ladderane fatty acids having been isolated and characterized recently (see Section 2.5). In passing, we note that photodimerized DNA is detrimental to the organism's health while the ladderane fatty acids are seemingly essential to its health.
Before beginning to describe the synthesis of oligomeric cyclobutane systems, some general remarks on the preparation of four-membered rings are in order. (While our discussions that follow deal almost exclusively with the hydrocarbon derivatives, functionalized species are discussed in more detail in chapters Synthesis of Cyclobutanes and The Application of Cyclobutane Derivatives in Organic Synthesis by Lee-Ruff, and Fu, Chen and Wong, respectively, in this volume.) Probably by far the most often used method to prepare four-memberd rings is the photochemical [2 + 2]cycloaddition of olefins5. (For a more complete discussion of photochemical aspects of cyclobutane chemistry see chapters Photochemistry of Cyclobutanes: Synthesis and Reactivity and Solvent-Free Photosynthesis of Cyclobutanes: Photodimerization of Crystalline Olefins by Horspool, and Natarajan and Ramamurthy, respectively, in this volume). However, at least one of the olefinic precursors can be replaced by other substrates providing two carbon atoms (and often appropriate functional groups) to the future cyclobutane ring, ketenes and acetylenes being used most often. In the latter case a hydrogenation step has to follow, of course, if saturated rings are the target. If both double bond precursors are replaced by alkynes, the [2 + 2]cycloaddition formally leads to a 1,3-cyclobutadiene6, and the very high reactivity of such an intermediate can also be exploited for the preparation of polycyclobutane derivatives (see Section 2.5 for further discussion of these derivatives. The reader should also note that “1,3” is not superfluous—1,2-cyclobutadiene and other highly unsaturated species are discussed by Johnson in chapter Highly Unsaturated Cyclobutane Derivatives of this volume). Other popular routes to cyclobutanes use various ring-contraction methods and 1,4-cyclization reactions. These different approaches will be discussed in detail for the specific examples mentioned below. If a cyclobutane ring has neighboring functional groups, in principle, intramolecular reactions can take place between them. A particular important example in the context of the present chapter is the readily occurring (thermal) ring-opening of cis-divinylcyclobutanes by Cope rearrangements leading to eight-membered-ring systems. Since this side reaction is to be avoided here, the appropriate three-dimensional orientation of these functional groups should also be avoided if one is interested in building extended polycyclobutanoid compounds. The inherent strain of the cyclobutane ring (Es ca 27 kcal mol−1)7 is usually no reason for a particular instability. (For a discussion of strain energy and thermochemical and physical chemical considerations of cyclobutanes see chapters Cyclobutane—Physical Properties and Theoretical Studies and Thermochemistry of Cyclobutane and Its Derivatives by Wiberg, and Liebman and Slayden, respectively, in this volume. For related discussions of aromaticity and antiaromaticity see chapters Antiaromaticity and Aromaticity in Carbocyclic Four-Membered Rings and Cyclobutarenes by Maksic and Maksic, and Stanger, respectively, in this volume).
2 Preparation of Oligocyclobutanoid Systems
- Top of page
- Preparation of Oligocyclobutanoid Systems
- From Rigid to Flexible Oligo- and Polymeric Cyclobutanes
- References and Notes
2.1 Cyclobutane Rings Connected by Single Bonds
One of the simplest hydrocarbons containing more than one cyclobutane ring is bicyclobutyl (2); it was first prepared by either CuCl2-mediated dimerization of cyclobutyl magnesium bromide (34), itself obtained as usual from cyclobutyl bromide (33) or from cyclobutene () via oxidation of the organoborane 36. In the former case, 2 is produced as a component (43% GC-yield) in a hydrocarbon mixture; the second approach yields the pure product in 22% yield (Scheme 7)8, 9.
Whereas neither 1,3- (3, n = 1) or 1,2-dicyclobutylcyclobutane (5) seem to be known, their isomer, 1,1-dicyclobutylcyclobutane (4), was obtained by the route summarized in Scheme 810. Treatment of the acid chloride 37 with triethylamine in ether caused dehydrochlorination and in situ dimerization of the generated ketene to the tricyclic diketone 38, itself an oligocyclobutane (of the spiro type to be discussed below in Section 2.3). When this is subjected to base treatment, the central ring is cleaved and the resulting intermediate decarboxylated to provide dicyclobutyl ketone (39) in good yield. Wittig olefination with the ylid 40 leads to the alkene 41 that, on epoxidation with meta-chloroperbenzoic acid (MCPBA), furnishes the expected spirocyclic oxirane. Ring-enlargement/rearrangement of the latter yields the ketone 42, that by Wolff–Kishner reduction finally gives the desired tercyclobutane 4 (Scheme 8). (Rearrangements figure prominently in cyclobutane chemistry as discussed in chapters Rearrangements of Cyclobutanes and Cyclobutyl, Cyclobutyl-Substituted and Related Carbocations by Tanko and Siehl, respectively, in this volume.)
To prepare various analogues of 4, different routes were investigated. For example, the Grignard coupling used for the preparation of 2 could in principle also be applied here, provided suitable precursors such as the halides 47 would be available (Scheme 8). Although its precursor carboxylic acid 46 could be prepared readily from the phosphonium salt 43 via the intermediate 44 (itself a hydrocarbon containing two cyclobutane rings) and produced from 43 by Wittig reaction between cyclobutanone and the non-oxidized cyclic ylid generated from 439, 11a-11d and the ketene adduct 45, all methods (inter alia Hunsdiecker degradation, Barton bromodecarboxylation) failed to yield 47. Instead, in all these experiments ring-expanded compounds of the general structure 48 and products derived therefrom were isolated. If, alternatively, the triketone 49 is cleaved by barium hydroxide treatment, the resulting diketone 50 could not be transformed to the diolefin 51 by the Wittig reaction with 40. Rather, the degradation product 41 was obtained (Scheme 9).
Still, several promising precursors en route to quinquecyclobutane, such as 52,53 and 54, were obtained in the course of these investigations. According to MM3(92) calculations of the septicyclobutane 55 (R1 = R2 = CH3), these oligomers adopt a helical conformation as shown in 56 as the thermodynamically most stable conformation10.
2.2 Cyclobutane Rings Connected by a Common Carbon Atom—The Spiro Oligocyclobutanes
The synthesis of linear structures in which the cyclobutane units are connected by spiro centers makes use of the oldest reaction in small-ring chemistry and the first cyclobutane derivative ever to be prepared: Perkin's 1,1-diethoxycarbonylcyclobutane (57, Scheme 10)12.
When 57 is reduced with lithium aluminum hydride, the diol 58 is obtained which can either be converted into the dibromide 59 (X = Br) or the ditosylate 59 (X = OTs). Both alkylating reagents react as expected with diethyl malonate (60) under basic condition and provide the spiro diester 61, ready to be subjected to the same sequence of steps again13.
To reduce 61 to the parent hydrocarbon spiro[3.3]heptane (11), 61 was first saponified and decarboxylated to the monocarboxylic acid 6313. Iododecarboxylation to the iodide 64 and reduction of the latter with lithium in tert-butanol then afforded the target hydrocarbon14. Later workers found the decarboxylation via the acid chloride 65 and the perester 66 preparatively more rewarding, although it also led to a small amount of the monoolefin 67, this route also being superior to an alternative employing Hunsdiecker degradation of 63 and subsequent reduction of the resulting bromide with tri-n-butyltin hydride15.
Treatment of 65 with triethylamine causes dehydrochlorination to a ketene intermediate () again, which dimerizes to the 1,3-cyclobutandione 69. After converting this to the corresponding bis-thioketal, Raney nickel induced desulfurization in benzene/ethanol readily provided tetraspiro[18.104.22.168.22.214.171.124]hexadecane (70)16. Since in the first step of the whole sequence other α,ω-dibromoalkanes can be used for the bis-alkylation of 60, other terminal rings can be introduced into these oligomeric cyclobutanes. Thus the decaspirane 71, the so-far longest spiro compound of this type, was prepared from 1,1-diethoxycarbonylcyclohexane16. Clearly, this truly general route deserves further attention, especially as far as the preparation of chiral representatives of this structural type is concerned, and these materials also have not been investigated from the material science viewpoint either.
Dispiro[126.96.36.199]decane (12), incidentally, was obtained from diketone 38 by converting it to the bis-thioketal and reducing this to the hydrocarbon, as described for the conversion of 69 into 7017.
Interestingly, often chiral hydrocarbon structures arise when the spiroannulation process is not continued in linear but in angular fashion, dispiro[188.8.131.52]decane, with its two cyclobutane rings in vicinal position at a four-membered ring being the parent compound. A case in point is provided by trispiro[184.108.40.206.2.2]tridecane (73), recently prepared from the readily available18 bicyclobutylidene 72 as shown in Scheme 1119.
Addition of dichloroketene to 72 followed by dechlorination furnished a 2:1 mixture of the two ketones 74 and 75 in very good yield (89%). When this was subjected to Wolff–Kishner reduction, the racemic trispiro hydrocarbon 73 was obtained readily. This hydrocarbon is chiral and its (M)-enantiomer, 77, the first hydrocarbon with a helical primary structure of four-membered rings, was prepared by reducing the cyclobutanone 74 with bakers yeast. The generated (5S,10S)-alcohol 76 was subsequently reoxidized with PCC and the optically active ketone deoxygenated by a Wolff–Kishner reduction again. Compared to the analogous hydrocarbon consisting of three-membered rings only20, (M)-trispiro[220.127.116.11.1.1]nonane, the specific rotation of 77 is significantly smaller (about a third of the cyclopropane system). According to molecular mechanics calculations, this could be caused by the greater flexibility of 77 plus the fact that the cyclobutane-composed hydrocarbon describes a distinctly shorter section of a helix than its three-membered-ring relative.
In a later synthesis, the preparation of derivatives such as 74 and 76 could not only be improved significantly by a novel cyclobutane-ring forming reaction, but this methodology could also be used to prepare the next higher homolog of 77, tetraspiro[18.104.22.168.22.214.171.124.2]hexadecane (M-83, P-83), as summarized in Scheme 1221.
In the crucial step of this evidently general approach, the four-membered-ring system is produced by the addition of a keteniminium salt to a double bond. These salts, compound 79 being an example, are conveniently obtained when carboxylic acid amides such as 78, itself prepared from the acid chloride 37 and piperidine, are treated with trifluoromethanesulfonic acid anhydride followed by 2,4,6-collidine22a-22c. Cycloaddition of 79 to the dispiro olefin 80, followed by hydrolysis, lead to a mixture of diastereomeric cyclobutanones 81 and 82 in 30% yield in which the former predominated by far (product ratio 92:8). After resolution of 81, the two enantiomeric ketones were Wolff–Kishner reduced to the pure (M)- and (P)-configurated hydrocarbons M-83, P-83.
Connecting the spiro carbon atoms not in a linear fashion as above but in a cyclic array leads to a new rotane family, the [n.4]rotanes23, the first analogues of which having recently been prepared. According to this nomenclature cyclobutane 15 (Scheme 3) would be [4.4]rotane24.
All syntheses are based on bicyclobutylidene (44), again demonstrating the overwhelming importance of this olefin in this area of hydrocarbon chemistry. To prepare the first member of the series, cyclobutylidene () was generated from 1,1-dibromocyclobutane (84) by treatment with methyl lithium in ether at low temperatures. Addition of the carbene to 44 furnished [3.4]rotane (86) directly (Scheme 13)23.
To prepare the next higher ‘cyclobutanolog’ 15, ketene was generated from the acid chloride 37 as described above and intercepted by 44 to yield the trispiroketone 45. Its spiroalkylation with 1-lithio cyclopropyl phenyl thioether (88) provided the tertiary alcohol 89, which by acid-catalyzed isomerization/elimination gave the expected ketone 90. This, finally, was reduced to 15 by Wolff–Kishner reduction23.
The syntheses of [5.4]- (96) and [6.4]rotane (99) followed practically identical paths. Homologization of 45 via a β-hydroxyselenide intermediate made the ring-expanded ketone 91 available, the ketone function of which was transformed into the missing (fifth) cyclobutane ring by the following protocol. The first of the still lacking carbon atoms was introduced by Wittig olefination. The resulting semicyclic alkene 92 underwent addition of the carbene produced by copper-catalyzed decomposition of methyl diazoacetate, yielding the cyclopropane ester 93 in moderate yield (Scheme 14; the yield is even lower in the six-membered-ring case—see below—indicating the poor accessibility of the double bond in both cases)23.
By reduction, reoxidation and reaction with tosylhydrazine, 93 was converted into the tosylhydrazone 94. After salt formation with sodium methoxide, a cyclopropylcarbene–cyclobutene rearrangement was initiated providing the cyclobutene derivative 95 which, by catalytic hydrogenation, was cleanly transformed to [5.4]rotane (96). Starting with the higher homolog of 91, the ketone 97, and submitting it to the same sequence of steps, furnished the rotane 99 with six consecutive spiroannelated cyclobutane rings passing the cyclobutene 98 en route. With the exception of 86, for all these new rotanes low-temperature X-ray structural investigations provided insights into the detailed geometric features of these interesting polycyclic hydrocarbons. Furthermore, with the help of axially labeled [1-13C]-99, temperature-dependent NMR studies allowed the determination of the free energy of activation for the chair-to-chair interconversion of this unusual cyclohexane derivative. With ΔG‡487 = 37.5 kcal mol−1, this is the highest inversion barrier ever reported for a cyclohexane derivative23.
2.3 Cyclobutanes Sharing Two and More Carbon Atoms—The Ladderanes and Fenestranes
The simplest hydrocarbon in which two cyclobutane units share two carbon atoms, i.e. are conjoined by a single common bond, is bicyclo[2.2.0]hexane (16), the parent system of the ladderanes (17, Scheme 4). It was originally prepared either from norbornan-2-one (100)25 or from 1,5-hexadiene (101)26 as shown in Scheme 15. (For a more complete discussion of bicyclopentanes and hexanes see chapter Bicyclo[2.1.0]pentanes and Bicyclo[2.2.0]hexanes by Carpenter in this volume.)
Today, many other approaches are known, most of them involving [2 + 2]cycloaddition steps (see below)27. Formally, 16 is the bis-hydrogenation product of another classic hydrocarbon, Dewar benzene (104, bicyclo[2.2.0]hexa-2,5-diene). To arrive at this valence isomer of benzene, phthalic acid was Birch-reduced to the cyclohexadiene diacid 102, which was photoisomerized to the bicyclic intermediate 103. Decarboxylation of the latter with lead tetraacetate or electrolytically finally provided the target compound 104 (Scheme 15)28.
In the first one, cyclobutenylaluminate salts, 106, stable intermediates (Lewis acid and Lewis base complexes) produced by aluminum trichloride-catalyzed dimerization of acetylenes such as 2-butyne (), are trapped with reactive olefins such as dimethyl maleate (107, Scheme 15) at low temperature: the resulting adducts 108 are produced in fair yields (35%) but high stereochemical integrity, dimethyl fumarate yielding the trans-isomer of 10829a, 29b. (Aspects of Bronsted acid and base chemistry of cyclobutane derivatives is discussed at greater length by Quintanilla, Davalos, Abboud and Alkorta in chapter Acidity and Basicity of Cyclobutanes of this volume.)
Other cyclobutadiene equivalents such as various cyclobutadiene tricarbonyliron complexes lead to comparable results. For example, liberating cyclobutadiene from metal complex 109 by oxidation with ceric(IV) ammonium nitrate (CAN) in acetone at 0°C and trapping this highly reactive intermediate with the strain-activated cyclopropene () furnished the two cyclo adducts 111 and 112, with the endo-isomer 112 slightly predominating (Scheme 16)30. (For a more complete discussion of organometallic chemistry see chapter Organometallic Derivatives by Butenschön in this volume.)
Analogously, the cyclobutadiene generated from 113 furnished the three cycloadducts 114, 115, 116 with 31. Only the endo-adducts 117 and 119 were obtained on oxidation of 118 in the presence of 107 or maleic anhydride (MA), respectively32, a stereochemical outcome attributed to secondary orbital interactions. That, after these observations, isolable cyclobutadiene derivatives such as 121 can also be used in such cycloaddition experiments to form 120 and 122 is not surprising (Scheme 16)33.
That cyclobutadienes are also excellent precursors for ladderanes really deserving this name has been known for many years. When 1,3-cyclobutadiene () is generated by treating cis-3,4-dichlorocyclobutene (123) with sodium amalgam, a reaction taking place via the metalated intermediate , it dimerizes to the syn-diolefin 128, a hydrocarbon readily hydrogenated to the saturated syn-ladderane 130 (Scheme 17)34.
Alternatively, the trans-isomers 129 and 131 are obtained via and , which, in turn, are generated by subjecting 123 to lithium amalgam treatment34. Note that in both isomers the ring-junctions are cis-configurated. Although this would produce very high strain, trans-fusion is also possible in principle.
To prepare the next higher analogue of 129, the bisanhydride 132 was prepared by photoaddition of maleic anhydride to acetylene, and its anhydride rings were converted to tetrahydrothiophene units of 133, by the steps summarized in Scheme 1835.
Chlorination and oxidation of 133 subsequently yielded the bissulfones 134, which by a Ramberg–Bäcklund ring contraction furnished the hydrocarbon 135, already containing four annelated four-membered rings. To prepare derivatives of this parent hydrocarbon, an abbreviated synthesis beginning with metal complex 109 was developed36a-36d. Cyclobutadiene () set free from it by Ce(IV)-oxidation again (CAN) was trapped with various acetylenes 136 (R = CH3, C6H5, COOCH3) to yield the 2:1 adducts 137, called pterodactyladienes because of their resemblance to the extinct flying reptiles Pterodactyla; that these olefins yield the ladderanes 138 on catalytic hydrogenation was to be expected (Scheme 18).
Considering the high reactivity of cyclobutadiene and its derivatives and the dimerization of to ladderanes (see above), it should in principle be possible to obtain still higher ladderanes by letting cyclobutadienes react with themselves. Indeed, when the dimethyl ester 139 is oxidized with CAN at low temperatures, a product mixture is obtained in 55% yield containing the ladderane derivatives 140, 141, 142 in 3:2:1 ratio (Scheme 19)37.
Later, this approach was extended to the preparation of - (143) to ladderanes (144) carrying different ester substituents38, 39. A related tandem cycloaddition approach is exemplified in Scheme 2040a-40d.
The norbornane-fused cyclobutene-3,4-diester 145 is reacted at 0°C with cyclobutadiene (), generated in situ by CAN oxidation of the iron tricarbonyl complex 109. This cycloaddition yields adduct 146 as the major isomer, which on treatment with excess dimethyl acetylenedicarboxylate (DMAD) in the presence of a Ru(0) catalyst at 50°C in benzene furnishes the [2 + 2]cycloadduct 147 in 52% yield. Repetition of these steps first led to the ladderane derivative 148 and finally to the ladderane hexaester 149 (Scheme 20). The shown exo-stereochemistry was proven by X-ray structure analysis of selected intermediates and reference compounds.
In a remarkable recent discovery, the tetramethylcyclobutadiene aluminium trichloride complex 106 was treated with iron pentacarbonyl in the hope that this in situ exchange reaction would open up a new route to the known iron tricarbonyl complex of tetramethylcyclobutadiene (see above). Actually, the desired iron complex, if formed, undergoes rapid decomposition to yield a mixture of hydrocarbons (Scheme 21).
According to spectroscopic analysis this mixture consists of various ladderanes, the so far longest, the ladderane derivative 150, being constructed of 14 2-butyne units41.
Since the most general route to cyclobutanes consists in the photodimerization of alkenes5, 42, the question may be asked whether ladderanes could not be prepared by multiple [2 + 2]photoadditions between the double bonds of a di- or oligoene. Usually, these systems find other more favorable reaction channels on photoexcitation (cis/trans-isomerization, cyclization, photo-Diels–Alder addition etc. depending on chain length and substitution pattern)43. However, provided that the oligomers units are fixed with respect to each other and held in parallel arrangement by a suitably constructed spacer unit, multiple [2 + 2] photoaddition can indeed be performed and ladderanes be obtained by this approach.
Such a spacer is provided by the [2.2]paracyclophane system, which is not only sufficiently rigid but also displays an inter-ring distance of only 3.1 Å, short enough to allow close approach between two parallel double bond systems for [2 + 2] photoaddition to take place44. As shown in Scheme 22, intramolecular addition takes place when the diester 151 is irradiated, yielding the cyclobutane derivative 152 not only in quantitative chemical yield but also with the highest quantum yield (ca 0.8) ever observed for a trans-cinnamic ester photodimerization (Scheme 22)45.
Extension of the chromophore by standard methodology converts 151 into its higher vinylogs 153 and 155. When these diesters are irradiated, the intended multiple cycloadditions in fact take place and provide the - and ladderanes 154 and 156, respectively46. Only from four consecutive double bonds on, the stereochemistry controlling influence of the [2.2]paracyclophane unit begins to break down and the irradiation of the corresponding diester does not yield a ladderane anymore47.
Recently, this polyene ladderane approach has not only been extended to other cyclophane spacers which hold the polyolefinic substituents in proper orientation48, but also to preorganized systems that owe the required parallel orientation of the reacting double bonds to supramolecular bonding effects.
Thus when trans-bis(4-pyridyl)polyenes (158) are co-crystallized with 5-methoxyresorcinol (157), a template is produced in which the two polyene molecules are fixed in parallel orientation by hydrogen-bonding interactions (Scheme 23)49.
When powdered samples of the templated polyenes are irradiated with UV light, ladderanes such as 159 are produced49.
Remembering that cyclobutane has a strain energy of 27.4 kcal mol−1 7, the ladderanes must be highly strained organic compounds and one might assume that they are available under laboratory conditions only. Surprisingly, this is not the case, ladderanes having been discovered recently as the dominant membrane lipids of two anaerobic ammonium-oxidizing bacteria50. These ladderane lipids with structures such as 160 and 161 generate an exceptionally dense membrane, helping to contain toxic intermediates such as hydrazine and hydroxylamine; their biosynthesis is so far unknown.
In a recent theoretical development, [n]ladderanes have been suggested as starting materials for so-called ‘shiftamers’51. Assuming that the ladderane 162 could undergo a [2 + 2]cycloreversion, this would lead to a structure with a local ‘defect’ consisting of two parallel double bonds, 163. This diene could isomerize via the boatlike transition state to 165, which for the parent system would be equivalent to 163 (Scheme 24). If this process could be continued along the parallel chains, it would lead to a pair of double bonds shifting along the polymer chains, hence shiftamers. According to B3LYP/6-31G(d) calculations, the activation barrier for the Cope rearrangement is low enough to make the shiftamers 163 fluxional at room temperature. Provided the four-membered rings in ladderanes are all-cis-fused—as is the case in the prismanes discussed in Section 2.5—polycyclic hydrocarbons would result in which appropriate double bonds could move around the perimeter in circles.
As already mentioned above (Scheme 4), the annelation of the cyclobutane rings must not necessarily continue linearly but can also take place in an angular fashion. In this latter case, tricyclo[4.2.0.01, 4]octane (19) would be the first representative from which the building process could continue in different reactions. By adding a methylene group between the first and the third ring, for example, the fenestrane molecule 18 can be constructed. Although this hydrocarbon so far is unknown (see below), its immediate formal precursor 19 has been prepared as described in Scheme 2552.
Starting with the bicyclic ketoester 166, obtained by photoaddition of ethylene to the appropriate cyclopentenone, a contraction of its five-membered ring was first carried out by employing the often used and well established sequence of ketone activation by formylation diazo group transfer photolytic Wolff rearrangement ester formation by ketene trapping with methanol. The resulting diester 167 readily underwent Dieckmann cyclization and the produced ketoester after saponification decarboxylated to the tricyclic ketone 168, as expected. When this was subjected to another ring-contraction protocol, the ester derivative 169 of the target molecule was obtained as a mixture of isomers. Saponification and treatment of the resulting isomeric acids with methyl lithium yielded the methyl ketone 170, which in a Baeyer–Villiger oxidation with m-chloroperbenzoic acid was converted to the acetate 171. When this was photochemically reduced by irradiation with 254 nm light in aqueous hexamethyl phosphorous triamide (HMPTA), the hydrocarbon 19 was obtained in fair yield.
Among the ‘unbroken’ fenestranes, the smallest parent system to be prepared so far is the [126.96.36.199]species, tetracyclo[4.4.1.03, 11.09, 11]undecane (178), and since it contains two annelated cyclobutane rings, i.e. is a derivative of ‘ladderane’ (16), its synthesis will also be discussed here (Scheme 26)53a, 53b.
The bicyclo[2.2.0]hexane core of the desired [188.8.131.52]fenesterane is produced early in the synthesis by a [2 + 2] photoaddition again, this time, however, by an intramolecular one, by which 173 is generated from the cyclopentenone 172. After the keto group has been protected as a dioxolane, reduction with lithium aluminum hydride subsequently provides a hydroxymethyl function and lithium treatment in ammonia removes the chlorine substituent. To construct the second five-membered ring, 174 is first oxidized to the acid 175, which is then chain-extended by treatment with oxalyl chloride. Reaction of the resulting acid chloride with diazomethane gives the diazoketone, which in the presence of rhodium(II) acetate provides a ketocarbene intermediate set up to insert in a proximal C,H-bond, thus providing 176 containing the complete carbon skeleton of 178. This hydrocarbon was finally obtained by two reduction sequences involving LAH treatment, conversion of the resulting alcohol to a tosylate and its reduction by a second LAH-attack, with the protective group having been removed after the ketal 177 had been reached from 176 as an intermediate54a-54c.
Hydrocarbons such as 18 and 19 are of interest with respect to the problem of creating a planar tetravalent carbon atom. In the absence of stabilizing substituents, planar methane has been calculated to be ca 150 kcal mol−1 less stable than tetrahedral methane55a-55c. On the basis of models and molecular mechanics calculations, the C1C8C5 angle in 18 should be about 130° and the strain energy on the order of 180 kcal mol−1. For 19, this angle has been estimated to be around 125° and the strain energy to be ca 90 kcal mol−1. The introduction of the fourth methylene group hence causes a drastic increase in strain52.
2.4 Cyclic Hydrocarbons Consisting of cis-Fused Cyclobutane Rings Only—The [n]Prismanes
The preparation of the prismanes - (21) to prismane (23) has been described and reviewed many times56a, 56b, (most notably by Bashir-Hashemi and Higuchi in chapter Chemistry of Cubane and Other Prismanes of this volume) a brief summary in this Chapter is hence sufficient in our view.
Thus beginning with prismane (21), this archetypical prismane, also known as Ladenburg benzene, was prepared from one of its (and benzene's) (CH)6-valence isomers, benzvalene (179, Scheme 27)57a-57c.
The ‘isomerization’ of 179 to 21 is initiated by the cycloaddition of N-phenyltriazolindione (180, NPTD) to 179. This causes a deep-seated reshuffling of the carbon atoms and provides the 1:1 adduct 181 in 50–60% yield, a process most likely taking place via polar intermediates. To close the last cyclobutane ring, 181 was first hydrolyzed and the resulting diacid oxidized to the azo compound 182, the yield being acceptable again. When this intermediate is irradiated with ultraviolet light, nitrogen is split off and a complex photolysate is produced from which small amounts of prismane (21) could be isolated (percentage range).
The synthesis of the next higher prismane, cubane (22), which is considerably less strained than 21, has been improved several times through its long history, but the basic idea—building precursors by cycloaddition and rearrangement reactions having more carbon atoms than the target molecule and then ‘chiseling’ these atoms away after they have fulfilled their purpose—has remained the same, since cubane was first synthesized forty years ago58a, 58b. Thus as illustrated by the optimized cubane synthesis in Scheme 27, the 1,4-dicarboxylic acid 188 can be prepared from cyclopentanone (183) in just five steps in about 25% over-all yield, allowing the preparation of cubanes in multi-kilogram batches. After ketalization of 183 followed by a bromination step, the tribromide 184 is obtained, which on base-treatment loses two equivalents of hydrogen bromide and provides the ethylene ketal of 2-bromocyclopentadienone, , as a reactive intermediate. This undergoes spontaneous dimerization to the Diels–Alder adduct 186, in which the two double bonds are so close that they can participate in an intramolecular [2 + 2] photoaddition to a product that, on acid treatment, loses its protective group and furnishes the bishomocubanedione 187. Ring contraction is subsequently readily accomplished by two Favorskii rearrangement steps and the resulting diacid 186 is best decarboxylated via 189 to the parent hydrocarbon by photolysis in the presence of AIBN and the hydrogen atom donor tert-butyl mercaptan59.
For the synthesis of prismane (23), the least strained of all prismanes known to date, several of the concepts used successfully for the preparation of cubane were exploited again (Scheme 28) where the hydrogen atom donor is 2,4,6-triisopropylbenzene60a, 60b.
Beginning with the [2 + 4]cycloadduct between benzoquinone and the ketal of perchlorocyclopentadienone, 190, a first photoaddition yielded the saturated diketone 191. This was dechlorinated and reduced by treatment with lithium in liquid ammonia, and from the resulting diol the iodo tosylate 192 was prepared. When this was reacted with base, rather than the intended bridge formation between the functionalized secondary carbon atoms, a ring-opening to the hypostrophene derivative 193 took place. Photochemical recyclization quickly returned three cyclobutane rings, and after removal of the protective group, the ketone 194, formally already very close to the target prismane, was at hand. Still, to remove the bridging carbonyl group required nine further steps, i.e. involved more than half of the whole synthesis, the main reason being that a direct functionalization of a bridgehead—prerequisite of, e.g., a ring contraction step—is very difficult because of the ‘protection’ of these positions by Bredt's rule. Therefore, 194 was Baeyer–Villiger oxidized, followed by another oxidation step with ruthenium dioxide/sodium periodate to yield a keto acid which was esterified with diazomethane to 195. Bridge closure could then be effected by an acyloin condensation and, when the resulting diol was oxidized, a homopentaprismane derivative was available with a functional group at a bridgehead, a hydroxyl group. Converting this to a tosylate set the stage for a Favorskii ring contraction which lead to 196, with all five cyclobutane rings completed. Decarboxylation via a perester pyrolysis in a hydrogen-donating solvent finally gave prismane (23).
Although ‘in principle’ there is no reason why the methodology used so successfully for the preparation of 22 and 23 should not be extended to prismane (199), all experiments have met with failure so far. A particularly disappointing one is an approach which lead—via the intermediate 197—to secoprismane (198, Scheme 29). Unfortunately, all attempts to dehydrogenate/oxidize this to 199, the face-to-face dimer of benzene, failed61a, 61b.
The value of n in a prismane represents the order of the hydrocarbon, and can in principle be between 3 and ∞. According to molecular mechanics and MO calculations, only the prismanes with up to 12 cyclobutane rings have been predicted to be planar with a Dnh symmetry. Higher members of the series should possess puckered structures displaying reduced angle strain and less-pronounced nonbonded hydrogen interactions62.
2.5 Building with Oligocyclobutanoid Precursors—The [n]Staffanes and the Oligo[n]cubyls
Hydrocarbons containing already several cyclobutane rings or those constructed from cyclobutane units completely are interesting building blocks for the creation of higher polycyclobutanoid oligomers. To be used for this purpose they have to be available in sufficient amounts, i.e. by straightforward and efficient synthetic protocols. Two particularly well studied cases in this context are bicyclo[1.1.1]pentane (26) and cubane (22), the monomers of the [n]staffanes and the oligo[n]cubyls.
Since all [n]staffanes can ultimately be traced back to [1.1.1]propellane (202, tricyclo-[1.1.1.01, 3]pentane), a brief description of its synthesis appears to be justified (Scheme 30).
In the original synthesis63, the dicarboxylic acid 200 was first converted into the dibromide 201 by Hunsdiecker degradation. When this intermediate was debrominated with tert-butyllithium in pentane, the propellane 202 was generated as the sole hydrocarbon. Since 200 is a rare chemical, the development of the chemistry of 202 had to await the discovery of an efficient synthesis for this smallest possible propellane. The breakthrough was accomplished by addition of dibromocarbene to the commercial product 203 which furnished the tetrahalide 204 in acceptable amounts. When this is dehalogenated with n-butyllithium in ether/pentane, 202 is produced in such amounts as to make the preparation of gram quantities easy64.
Staffane synthesis from 202 usually starts with the addition of a radical R· to the central (most reactive) bond of 202, yielding the bicyclo[1.1.1]pentyl radical (Scheme 31).
This can either react with an R′X system (which may also be a bicyclo[1.1.1]pentane derivative) to furnish the derivative 206 in a controlled fashion, or initiate the oligomerization of 202 leading to the staffyl radical . In a last step the latter then stabilizes itself by insertion into another R′X bond. Of course, this oligomerization route leads to mixtures of staffanes that have to be separated if the pure target compounds are desired. Since, however, today a wide variety of bicyclo[1.1.1]pentanes—usually obtained from 202—is known65, coupling reactions of these precursors provide a third, controlled entry to the staffanes (see below).
To prepare the parent systems, a mixture of argon and hydrogen or ammonia was passed through a microwave discharge and subsequently over a stirred solution of 202 in pentane at −110°C for 10 h. After 202 had been consumed, distillation and gradient sublimation yielded the homologous series of staffanes with n up to 6 (Scheme 32).
The hydrocarbons 209 were characterized by their spectroscopic and analytical data and several of them also by X-ray structure analysis, which confirmed the anticipated linear molecular shape and staggered conformations66. Besides these oligomers, 14% of nonvolatile polymers insoluble in the common organic solvents was also produced.
Functionalized staffanes, such as the ester 210, were prepared by irradiation of a mixture of 202 and methyl formate in pentane using dibenzoyl peroxide as the free radical initiator (Scheme 33).
The oligomerization also takes place under ionic conditions, as shown by the conversion of 202 to the oligomeric esters 211 in the presence of n-butyllithium67.
The propellane nucleus can also be a part of a more complex framework, as demonstrated by the oligomerization of 212 to 214 via the lithiated intermediate 21368.
A typical bridgehead-to-bridgehead coupling process takes place when 1-iodo-3-phenylbicyclo[1.1.1]pentane (215), itself obtained by photochemical insertion of 202 into iodobenzene69, is first metalated with tert-butyllithium and the resulting metal organic intermediate subsequently subjected to metal-mediated dimerization to the diphenyl derivative 216 (Scheme 34).
Today, staffanes constitute a widely investigated class of cyclobutane ring containing rod-like molecules, that are, inter alia, of interest for the preparation of novel liquid crystalline materials or the investigation of electron and energy transfer processes across the saturated spacer70a, 70b.
Incidentally, the bicyclo[1.1.1]pentane core has also been incorporated as a spacer unit in a completely different class of compounds. Thus the homologous series of bicyclo[1.1.1]-pentane-1,3-dicarboxylate esters 217 featuring ω-alkenols of differing chain length undergo ring-closing metathesis with Grubbs catalyst to provide various paddlanes with monomeric, 218, to tetrameric structures 219 (Scheme 35)71.
Just as propellane 202 is the central starting material in staffane synthesis, the [n]oligocubyls can all be traced back to two diiodocubanes: 1,2- (220) and 1,4-diiodocubane (225, see below), respectively, both readily available by the various methods for preparing functionalized cubanes developed over the years by Eaton and coworkers58b.
When 220 is treated with tert-butyllithium, deiodination takes place and dehydrocubane 222 is generated as a highly reactive (though trappable) intermediate via the organolithio compound 221 (Scheme 36).
The pyramidalized olefin reacts with a further molecule of 221 to provide the lithiated dimeric iodide 223, which—after further metalation and quenching with methanol—yields cubylcubane (224) as the simplest oligocubyl.
More effective routes to 224 and its derivatives involve another highly reactive intermediate generated from the isomer of 220, 1,4-diiodocubane (225) by phenyllithium treatment: 1,4-dehydrocubane or cubane-1,4-diyl (227, Scheme 37)72.
Reaction of 227, the structure of which was determined by matrix isolation infrared spectroscopy73, with intermediate 226 provides the lithio derivative 228 as a precursor for the bisiodide 229 (reaction with iodobenzene). From 229, the hydrocarbon 224 could again be prepared by metalation/methanol quench.
Intermediate 230 initiates a ‘living polymerization’ leading to the series of intermediates 231, 232, 233 which, by trapping with the 1,4-diiodide 225, furnishes the p-[n]cubyls 234 (n = 2–4). These are not only useful precursors for the preparation of numerous other derivatives but can also be converted to the parent hydrocarbons.
Since other aryl lithium reagents can take the role of the phenyl lithium in these transformations—or can be exchanged altogether for Grignard reagents—numerous oligocubyl systems have been synthesized, as illustrated by several representative examples, 235, 236, 237, 238.
As expected, the solubility of the [n]cubylcubanes decreases rapidly with growing degree of oligomerization. By introducing solubilizing alkyl groups, this drawback can be overcome and truly polymeric molecular rods may be obtained. Thus when 7,7-di-n-hexyl-1,4-diiodocubane (239) is subjected to the above coupling conditions, a polymer 240 with a molecular weight of ca 10,000 is isolated, corresponding to a polycubyl rod containing ca 40 cubane building blocks (Scheme 39)75.
In all the above examples, the connections between the cubane blocks are established by single bonds. Other modes of sticking these cubes together are conceivable, though. By introducing additional single bonds as shown for the cubylcubane 224 and its higher ‘cubylog’ 241 in Scheme 40, edge-fused oligocubyls such as 243 and 244 may be generated. And when the cubane monomer units 22 are fused at their faces, columnar structures such as 242 and 245 result (Scheme 40).
Of course, hybrid types containing different prismanes—as shown for prismane and prismane in structure 246—can also be imagined, and when the ring size is not restricted to a particular size, skyscraper structures such as 247 result. None of the molecules 242, 243, 244, 245, 246, 247 has been prepared so far, but at least DFT calculations on several of the poly[n]prismanes—such as 242 and 245—have appeared recently76. Although these structures must contain highly distorted tetracoordinated carbon atoms, the calculations suggest relative stable Dnh-structures for the poly[n]prismanes. The main factor for the stability of these hydrocarbons is the πσ–πσ orbital interaction between the parallel rings.
3 From Rigid to Flexible Oligo- and Polymeric Cyclobutanes
- Top of page
- Preparation of Oligocyclobutanoid Systems
- From Rigid to Flexible Oligo- and Polymeric Cyclobutanes
- References and Notes
Nearly all of the oligocyclobutanoid compounds described so far are characterized by their more or less pronounced rigidity. And for their synthesis, usually a stepwise approach has been employed, i.e. one involving the addition of one cyclobutane ring(s) containing subunit after the other. We now turn to ‘real’ cyclobutane-derived polymers and it is not surprising that for their preparation, classical cationic, anionic and radical chain-polymerization techniques have been widely applied. We will divide these polymers into three categories: In the first one, both the monomers and the polymeric products contain four-membered rings. In the second one, the polymers no longer contain a cyclobutane ring although the starting material is a cyclobutyl derivative. And, finally, the third category will present transformations in which cyclobutane rings are generated in the polymer although the monomers have none.
3.1 From Cyclobutyl Monomers to Polymers Containing Cyclobutane Rings
It is not unexpected that traditional alkene additions and condensations of carboxylic acid derivatives comprise two principal modes of polymerization of cyclobutane-containing monomers to cyclobutane polymers.
Probably the most traditional approach to polymers in this category is ester and amide condensation polymerization of cyclobutane-containing bifunctional monomers. An appreciable variety of cyclobutane polymers has been prepared by this means. As a precursor to multiple bifunctional monomers, the dicyanocyclobutane dimer of acrylonitrile (248) is an exceedingly versatile core synthon77. The dimerization produces principally the 1,2-compound as a mixture of cis- and trans-isomers, but both hydrolysis and reduction proceed to the trans-diacid 249 and the trans-diamine 252, respectively. Polyamides prepared as melts directly from 249 and n-methylenediamines are isomerized unevenly back to the cis-form, but the milder conditions of interfacial polymerization with the diacid chloride 250 and the same diamines yielded more uniform crystalline polyamides 251 as shown in Scheme 41.
Polyamides of 252 and adipic acid and sebacic acid, the polymers 253, can be prepared by melt methods without isomerization, as also shown in Scheme 41.
The diphenyl ester prepared from 250 yields several polyesters with diols such as bisphenol A, and the ester and amide permutations of bifunctional cyclobutanes are completed by lithium aluminum hydride reduction of 250 to trans-1,2-bis(hydroxymethyl)cyclobutane (254) and subsequent polymerization with diacids exemplified by terephthalic acid (255) to provide the polyesters 256 (Scheme 42).
A more highly functionalized cyclobutane-containing copolyamide has been synthesized that contains both cyclobutane ring and conjugated double bond in the main chain78. The prototype reaction—yielding the complex polymer 260—is shown in Scheme 43 for bis(p-nitrophenyl) β-truxinate (257) reacting with di(p-nitrophenyl) p-phenylenebis(acrylate) (258) and 1,3-di(4-piperidyl)propane (259).
Another elaborate cyclobutane-containing polyamide, poly[p-phenylene-trans-3,cis-4-bis(2-hydroxyphenyl)-1,trans-2-cyclobutane dicarboxamide] (262), can be made by ring-opening condensation polymerization of p-phenylenediamine with the anti-cyclobutane coumarin dimer 261 (Scheme 44)79. A spare cyclobutane-free poly(p-phenylenefumaramide) polymer 263 results from nearly quantitative asymmetric photolytic extrusion of trans-2-hydroxy-2′-hydroxystilbene from the cyclobutane precursor 26279.
Maleic anhydride is photodimerized to 1,2,3,4-cyclobutanetetracarboxylic dianhydride (264) which is shown to have a cis,trans,cis-configuration by X-ray crystallography. The dianhydride can be condensed with various aromatic diamines, such as p,p′-diaminodiphenyl ether, to prepare polyimides such as 265, as shown in Scheme 4580.
Five-substituted 2-oxabicyclo[2.1.1]hexan-3-ones (266, R1 = H, CH3, R2 = H, CH3 and R1 = CH3, R2 = H, CF3) were synthesized from the corresponding 3-chlorocyclobutanecarboxylates obtained by addition of hydrochloric acid to the cycloadduct of allene to acrylonitrile or methacrylonitrile81. These bicyclic lactones resemble β-lactones in acid- or base-catalyzed polymerizations to high-molecular-weight and high-melting 267 polyesters (Scheme 45).
Another conspicuously straightforward approach to cyclobutane polymers is through conventional alkene addition polymerizations. Polymers have been prepared from cyclobutane-containing monomers with an olefin as a pendant vinyl group, exo to the ring or as cyclobutene. Thus, vinylcyclobutane (7) was prepared by the synthesis shown in Scheme 46 from cyclobutanecarboxylic acid (268)82. A highly crystalline polymeric material as identified by crystallographic d-spacings, compound 8, was obtained after initiation with triisobutyl aluminum and titanium tetrachloride (Scheme 46).
Now consider the olefin moiety contracted closer to the cyclobutane ring in the methylene-cyclobutane 269 made from allene and various acrylates en route to the bicyclic lactones 266 described above. Free-radical-initiated homopolymerization with azobisisobutyrylnitrile failed for methylenecyclobutane and all 3-substituted cyano, carboxylic acid, methyl, and phenyl derivative and the 3-methyl-3-methoxycarbonyl analogue83. However, the same methylene-cyclobutanes could be made to copolymerize with acrylonitrile, methyl methacrylate, vinyl acetate, vinyl bromide and styrene to furnish polymers such as 270. In Scheme 46 a process is illustrated in which the mole percent ratios of methylenecyclobutane and vinyl bromide in the starting monomer mixture and in the final polymer are approximately equal.
As the double bond is positioned closer to the cyclobutane ring in this subcategory of alkene additions, we finally arrive at the polymerization of cyclobutenes. Spontaneous polymerization of 1-cyclobutenecarboxylic acid 271 was first observed during an attempt to purify the compound84. It had been prepared by dehydrohalogenation of 1-bromo-1-cyclobutane carboxylic acid obtained by free-radical bromination of 1-cyclobutanecarboxylic acid (268). Polymerization could be deliberately induced by UV irradiation in the presence of 2,2-dimethoxy-2-phenylacetophenone. 13C NMR spectroscopy confirmed the absence of ester carbon atoms that would result from possible polymerization by Michael addition and the absence of alkene carbon that would result from possible ring-opening polymerization. The polymer 272 is a noteworthy member of a small population of 1,2-di-, tri- and tetra-substituted alkene homopolymers that presumably issues in this case from relief of cyclobutene ring strain as well as the protection against radical termination afforded by three substituents on the olefin.
A unique polymerization route in this category is radical polymerization of bicyclobutane derivatives. There are many examples of free-radical homopolymerization of bicyclobutanes bearing bridgehead electron-withdrawing groups and copolymerization with vinyl monomers. 1-Bicyclobutanecarbonitrile and C-2 and C-4 methyl substituted 1-bicyclobutanecarbonitriles (278) (R1 = R2 = H; R1 = H, R2 = CH3; R1 = R2 = CH3) readily polymerized by radical and anionic initiation. Copolymers, e.g. 279, of a great variety of the 1-bicyclobutanecarbonitriles with each other and with standard vinyl monomers were also realized (Scheme 47)85.
2,2,4,4-Tetramethyl-1-bicyclobutanecarbonitrile (278 R1 = R2 = CH3) was synthesized as shown in Scheme 47 starting from 1,1-dimethylallene (273) and acrylonitrile. The initially obtained [2 + 2]cycloadduct 274 was transformed to the target systems 278 via intermediates 275, 276, 277 following the protocol summarized in Scheme 47. Other 1-bicyclobutanecarbonitriles were synthesized similarly86.
A more controlled and improved radical polymerization route, atom transfer radical polymerization (with CuBr, methyl 2-bromopropanoate and 4, 4′-dinonyl-2, 2′-bipyridyl (dNbyp)), has recently been applied to methyl 1-bicyclobutanecarboxylate (280) as shown in Scheme 4887. Stereochemistry of the polymer 281 is 66% trans-configuration of the methyl ester group to the cyclobutane polymer chain. The monomer 280 is prepared by treatment of methyl 3-chloro-1-cyclobutanecarboxylate with sodium hydride. The controlled step-growth atom transfer radical polymerization method is conveniently extended to the synthesis of block copolymers of methyl 1-bicyclobutanecarboxylate with styrene. The same investigators have also reported atom transfer radical homo- and block copolymerization of methyl 1-cyclobutenecarboxylate and methyl 1-bicyclobutanecarboxylate88.
Functional polyester, polyamide and polyisoprene polymers with transient cyclobutane formation induced by irradiation have been reported by many investigators89. Scheme 48 depicts a norbornadiene incorporated into the polymer chain along with a carbazole sensitizer in a donor–acceptor combination, 282. Later photoisomerization of the norbornadiene unit to a metastable quadricyclane, 283, qualifies the polymer as one comprising a cyclobutane system. The energy stored in the quadricyclane can be released in the presence of catalytic amounts of (5,10,15,20-tetraphenyl-21H,23H-porphyrinato)-cobalt(II) (Co-TPP) and makes the combination polymer an attractive scheme for storage of radiation energy.
3.2 From Monomers Containing Cyclobutane Rings to Polymers Without Cyclobutane Units
Just as the strained 1,3-bridge bond of bicyclobutanes invites a radical dissociation accompanied by polymerization (see also the examples discussed above with propellanes), a strained cyclobutane itself can be exploited for polymerization by ring opening, although not necessarily through a radical route.
Cyclobutane adducts 284 of tetracyanoethylene and ethyl, butyl, isobutyl vinyl ethers and 2,3-dihydrofuranyl and 3,4-dihydro-2H-pyranyl vinyl ethers are readily polymerized anionically via intermediate with tetrabutylammonium iodide as shown in Scheme 4990. The authors present convincing evidence that the polymerization process to 286 includes, in addition to the stepwise mechanism of Scheme 49, a linking of individually growing chains by nucleophilic substitution for halide of one chain by the anion of another chain. Other anionic initiators, Lewis acids and tertiary amines have also been reported to polymerize 284.
In contrast to the radical vinyl-type polymerizations described above, anionic coordination catalysts are shown to promote ring-opening polymerization of trans-1,2-divinylcyclobutane (287) and cyclobutene (). Scheme 50 shows that 287 is opened to a linear polymer, 28891.
Anionic-coordinated polymerization of cyclobutene () proceeds to macromolecules of a polycyclobutane structure, 289, similar to the 1-cyclobutenecarboxylic acid polymers (272) above or to macromolecules with the structure of 1,4-polybutadienes 290, depending on the catalyst92. Choosing VCl4 and trihexylaluminum yields the polycyclobutane 289, while TiCl4 and Al(C2H5)3 as the catalyst gave pure samples of the latter 1,4-polybutadienes 290 (Scheme 50), all in heptane as solvent. A final gratifying manipulation—in the presence of δ-TiCl3 (solid solution of TiCl3, AlCl3 and Al(C2H5)3 in heptane—produced a simultaneous crystallization of distinct trans-1,4-polybutadiene and polycyclobutyl polymers but apparently no copolymer containing both-CH2CHCHCH2- and 1,2-cyclo[(CH)2(CH2)2] subunits.
3.3 From Monomers Without Cyclobutane Rings to Polymers Containing Cyclobutanes
Highly fluorinated cyclobutanes show unique chemistry (see chapter Fluorinated Cyclobutanes and Their Derivatives by Lemal and Chen in this volume). Polymeric derivatives demonstrate features of this uniqueness. For example, perfluorocyclobutyl polymers 293 are prepared by free-radical mediated thermal cyclodimerization reactions of aryl trifluorovinyl ethers 291 as shown in Scheme 51, the process presumably involving diradical intermediates of type 29293. A recent modification employs a phenylphosphine oxide group of two varieties as the ether backbone, 294.
A unique cation radical chain cycloaddition mechanism accounts for new polymer structures from bifunctional propenyl (295, drawn in abbreviated form as 296) or vinyl monomers94. The reaction is initiated by catalytic amounts of a stable cation radical salt, tris(4-bromophenyl)aminium hexachloroantimonate. The mechanism, involving the intermediate formation of the radical cations and , is shown in Scheme 52 for the prototypical bifunctional bis-1,2-[4-(1-propenyl)phenoxy]ethane 295/296, which produces trans,anti,trans-cyclobutane units along the polymer backbone, 299. A crucial attribute of the monomer is efficient electron transfer from the cyclobutane radical cation cycloadduct 299 to the remote end of the formed oligomer for continued chain growth. (For a more complete discussion of cyclobutane ion chemistry see chapters Mass Spectrometry and Gas-Phase Ion Chemistry of Cyclobutanes and Cation Radicals in the Synthesis and Reactions of Cyclobutanes by Kuck and Bauld, respectively, in this volume).
Traditional alkene polymerizations discussed above in Section 2.1 are again seen in the radical polymerizations of 1,1-disubstituted 2-vinylcyclopropanes (300, R = Cl, EtO2C, CN). Although the expected 1,5-ring-opened products 301/302 comprised the bulk of the polymeric material, cyclobutane products (303/304) are strongly suggested as well (Scheme 53)95.
References and Notes
- Top of page
- Preparation of Oligocyclobutanoid Systems
- From Rigid to Flexible Oligo- and Polymeric Cyclobutanes
- References and Notes
- 1Houben-Weyl, Methods of Organic Chemistry, Carbocyclic Three- and Four-membered Ring Compounds (Ed. A. de Meijere), Vol. E 17e, Cyclobutanes, Georg Thieme Verlag, Stuttgart 1997.
- 2Organic Chemistry: The Name Game, Chap. 6, Pergamon Press, New York, 1987, p. 87.and ,
- 3Modern Acetylene Chemistry (Eds. P. J. Stang and F. Diederich), Chap. 12, VCH, Weinheim, 1995, p. 415.and , in
- 4Bioorganic Photochemistry (Ed. H. Morrison), Vol. I, Wiley Interscience, New York, 1990, p. 1.and , in
- 5Organic Photochemistry and Photobiology (Eds. W. M. Horspool and P.-S. Song), CRC Press, Boca Raton, 1995, p. 29., in
- 6Methods of Organic Chemistry, Carbocyclic Three- and Four-membered Ring Compounds (Ed. A. de Meijere), Vol. E 17f, Cyclobutenes and Cyclobutadienes, Georg Thieme Verlag, Stuttgart 1997, p. 1051.and , in Houben-Weyl,
- 7Strained Organic Molecules, Academic Press, New York, 1978, pp. 66 and 94.and ,
- 9Bull. Acad. Sci. USSR, Div. Chem. Sci. (Engl. Transl.), 22, 1397 (1973); Izv. Akad. Nauk SSSR, Ser. Khim, 22, 1429 (1973)., , , , , , and ,
- 11aSynthesis, 299 (1987). The preparation of from is probably the preparatively best route to this olefin, providing it in multigram quantities. However, many other syntheses of bicyclobutylidene are known, inter alia: J. Org. Chem., 39, 1650 (1974).and ,
- 11bDokl. Chem., 232, 98 (1977).,
- 11cRecl. Trav. Chim. Pays-Bas, 112, 457 (1993)., , , , and ,
- 11dJ. Chem. Soc., Chem. Commun., 642 (1972). Chem. Ber., 117, 277 (1984).and ,
- 24For the preparation of the ‘classical’ [n.3]rotanes, see the summary in Classics in Hydrocarbon Chemistry, Chap. 11.6, Wiley-Interscience, New York, 2000.,
- 27Many routes leading to bicyclo[2.2.0]hexane derivatives have been summarized by Methods of Organic Chemistry, Carbocyclic Three- and Four-membered Ring Compounds (Ed. A. de Meijere), Vol. E 17f, Cyclobutenes and Cyclobutadienes, Georg Thieme Verlag, Stuttgart, 1997, p. 831., in Houben-Weyl,
- 36aChem. Eng. News, 43, 38 (1965)., , and ,
- 42Methods of Organic Chemistry, Carbocyclic Three- and Four-membered Ring Compounds (Ed. A. de Meijere), Vol. E 17e, Cyclobutanes, Chap. 1.3, Georg Thieme Verlag, Stuttgart, 1997, p. 85., in Houben-Weyl,
- 43Elements of Organic Photochemistry, Plenum Press, New York, 1976.and ,
- 44As shown by Schmidt in his famous studies on topochemical reaction control, in the solid state cinnamic acid undergoes photodimerization only when it crystallizes in a lattice in which the distance between the reacting (parallel) double bonds does not exceed 3.5–4.0 Å; for a summary, see Monographs in Modern Chemistry. Solid-State Photochemistry, Verlag Chemie, Weinheim, 1976.,
- 45J. Chem. Soc., Chem. Commun., 1075 (1994)., , , , and ,
- 53aTetrahedron, 42, 1549 (1986). Agosta and coworkers have also prepared derivatives of [184.108.40.206]fenestrane (i.e. hydrocarbon with one methylene group replaced by an ethano bridge). Since this chapter concentrates on the syntheses of parent hydrocarbons, the routes leading to these smallest and most highly strained of the tetracyclic fenestranes will not be discussed here. The synthetic approach to these derivates is comparable to the synthesis of summarized in Scheme 25; cf. J. Chem. Soc., Chem. Commun., 293 (1984)., and ,
- 56aFor recent reviews on prismanes, see: Carbocyclic Cage Compounds (Eds. E. Osawa and O. Yonemitsu), VCH Publ., New York, 1992, p. 183.and , in
- 56bClassics in Hydrocarbon Chemistry, Chap. 4, Wiley-VCH, Weinheim, 2000, p. 41.,
- 58aJ. Am. Chem. Soc., 86, 962 and 3157 (1964).and ,
- 61aInd. J. Chem., 22B, 959 (1983)., and ,
- 62Strain and Its Implications in Organic Chemistry (Eds. A. de Meijere and S. Blechert), Vol. 273, NATO, ASI Series, Kluwer Academic Publ., Dordrecht, 1989, p. 259., , and , in
- 70aReviews: Advances in Strain in Organic Chemistry, Vol. 4, JAI Press Inc., Greenwich. CT, 1995, p. 283.and , in
- 7558b., and , unpublished results, cited in Reference
- 86We recognize bicyclobutane as a highly strained molecule. From Greenberg and Liebman, op. cit. (Reference 7), p. 72, we find the strain energy of bicyclobutane to be 66.5 kcal mol−1. As defined by the exothermicity of the cyano transfer reaction, Bicyclobutane + cyanocyclobutane 1− cyanobicyclobutane + cyclobutane, 1-cyanobicyclobutane is less strained by ca 6.5 kcal mol−1 than the parent hydrocarbon. 1-Cyanobicyclobutane is thus strained by ca 60 kcal mol−1, a still impressive amount. (All thermochemical data were taken from , and , Thermochemical Data of Organic Compounds (2nd edn.), Chapman & Hall, New York, 1986.)