Acetylenic Carbon-Containing Stable Five-Membered Metallacycles

: Due to the linear property around an acetylenic carbon, the introduction of such an atom to a small cycle would result in high ring strain. Currently, the smallest isolated rings are ﬁve-membered, including metallacycloalkynes and metallapentalynes. Both types contain at least one unusual small bond angle around the acetylenic carbon, thus exhibiting abnormal reactivities. This feature article gives a comprehensive overview on these two kind complexes. The synthesis and reactivities are extensively described, the source of stability is presented, and the future prospect is discussed. The article aims to provide a better development for the chemical diversity of ﬁve-membered metallacycloalkynes and metallapentalynes.


Introduction
Due to the sp hybridization of acetylenic carbon, the X≡C-Y moiety is normally linear. If such a moiety is introduced into a small cycle, large angle strain would be caused, which increases with decreasing ring size [1,2]. To date, the smallest purified unsubstituted carbocyclic alkyne is cyclooctyne with an eight-membered ring (1, Figure 1) [3], although cycloheptyne, cyclohexyne, and cyclopentyne could exist as transient reactional intermediates [4]. One of the methods to stabilize such small cyclic compounds is by the introduction of bulky substituent(s) into the α-position(s) of the X≡C bond. The bulky group(s) can prevent the dimerization of the unsaturated bond(s), and also the attack from other molecules [2]. For example, cycloheptyne is highly unstable and has a half-life of 1 h at −78 • C [5], however after the introduction of four methyl groups into the α-positions, the resulting compound 3,3,7,7-tetramethylcycloheptyne could be purified and is stable at room temperature (2, Figure 1) [6]. The substitution of carbon atoms in the ring by heteroatoms sometimes also stabilizes the strained carbocyclic cycles. Cyclohexyne is not stable above −100 • C and could only be characterized by infrared (IR) spectroscopy at low temperature [7]. In comparison, the six-membered tetrasilacyclohexyne is isolated and stable in boiling toluene even after one day (3, Figure 1) [8,9]. The stability of tetrasilacyclohexyne is due to the reduction of the angle strain by the longer Si-C bonds, which make the C≡C-Si angles closer to linear.
The incorporation of a metal center is another elegant method to stabilize the strained acetylenic carbon-containing compounds. For example, cyclopentyne is too unstable to be trapped [10]; in contrast, two types of five-membered metallacycles containing at least one acetylenic carbon atom have been developed (metallacycloalkyne A and metallapentalyne B, Figure 2) [11][12][13][14][15][16]. These five-membered metallacycles are interesting not only due to their abnormal structures and unusual reactivities, but also because five-membered metallacycles are intermediates in many transition-metal-catalyzed organic reactions [11][12][13][14][15][16][17][18][19]. Surprisingly, there is no review paper covering both types. Considering the similarity of A and B, and the emergence of more and more novel chemistry, this review will discuss them together for the first time. Through the summarization and comparison on the synthesis and reactivities of A and B, it is favorable for the development of their chemical diversity, which is the aim of this account.

Synthesis of Five-Membered Metallacycloalkynes
The first five-membered metallacycloalkynes with the structure of A shown in Figure 2 were reported by Suzuki et al. in 2002 [20]. Treatment of CpZrCl 2 with 2 equivalents of n-BuMgX at low temperature generated a low-valent zirconocene, which further reacted with (Z)-1,4-disubstituted-1,2,3-butatriene (4a: R = SiMe 3 , 4b: R = t-Bu) for 1 h at room temperature to provide two 1-zirconacyclopent-3-ynes 5a and 5b, existing as cisand trans-isomers (Scheme 1). n-BuMgX can be replaced by n-BuLi or EtMgBr, while in the latter case, a seven-membered metalacyclic alkyne would be afforded as the by-product [21,22]. Initially, cis-configurations were dominant for both 5a and 5b. However, after 48 h at room temperature in solution, the ratios of cis/trans isomers of 5a and 5b changed to 36/64 and 12/88, respectively. The isomerization from cis to trans resulted in the isolation of trans-isomers as good single crystals only. Later on, complex 5c with a cis/trans ratio of 50/50 was prepared in a similar procedure (Scheme 1) [23]. This strategy could also be extended for the preparation of titanium and hafnium analogues [23,24]. Scheme 1. Synthesis of metallacyclopentynes 5a-5c.
Complex trans-5a is selected as the example to illustrate the structures of these metallacyclopentynes. The five atoms in the cyclopentyne ring are almost coplanar, with an interior angle sum of the five-membered zirconacycle as 539.5 • , close to an ideal planar pentagon ( Figure 3). The bond length of C2≡C3 is 1.195(7) Å, slightly shorter than that of cyclonoyne (1.21 Å) determined by gas phase electron diffraction [25], and at the longer end of those in normal acyclic alkynes (1.167-1.197 Å) [26]. The Zr-C2 and Zr-C3 distances are 2.291(5) and 2.286(6) Å, respectively, comparable to the Zr-CH 3 lengths in Cp 2 Zr(CH 3 ) 2 (2.280(5) and 2.273(5) Å) [27], and even shorter than those of Zr-C1 (2.469(5) Å) and Zr-C4 (2.456(7) Å) in trans-5a. Therefore, one may think the C2≡C3 bond is coordinated with the metal and it is actually a metallacyclopentene complex. However, in normal zirconocene-alkyne complexes, the lengths of coordinated triple bonds are in the range for double bonds [28], which is contradictory to the C2≡C3 distance in trans-5a. Besides, the IR absorption band for C2≡C3 bond at 2014 cm −1 , which is in a sharp contrast to those in Zr-alkyne complexes (i.e., 1611 cm −1 ) [29], further supports its triple bond nature. The C1-C2≡C3 and C2≡C3-C4 angles are 154.2(6) • and 154.4(7) • , respectively, which largely deviate from the standard angle around sp-hybridized carbon. Thus, there must be a high ring strain, and how can this kind of complexes be stable? Jemmis and coworkers proposed that the stability of 5 is due to the Lewis structure of 5B (η 2 -σ,σ + η 2 -π coordination mode), in which the C≡C triple bond is coordinated by the Zr atom ( Figure 4) [30][31][32]. Lin et al., on the other hand, proposed that the cumulene coordinated resonance form 5C is the major contribution for their stability (Figure 4, η 4 -π,π mode) [33]. Through an electron density analysis based on X-ray diffraction data, Hashizume, Suzuki, and Chihara, however, stated that the mode is a resonance hybrid, with the major contributor as the η 2 -σ,σ form (5A, Figure 4), and the minor one as 5C [34]. The existence of 5C explains the interconversion of cis/trans isomers. In 2004, the Suzuki group developed a more convenient route, in which butatriene derivatives are not necessary, for the construction of 1-metallacyclopent-3-ynes. Reduction of Cp'ZrCl 2 and 1,4-dichlorobut-2-yne by Mg in THF resulted in the formation of 5d and 5e [35], and 1-titanacyclopent-3-yne 6 was generated similarly (Scheme 2) [23,24]. The synthetic route was also applicable for the synthesis of 1-hafnacyclopent-3-yne 7, although it could not be purified and was only characterized by nuclear magnetic resonance (NMR) spectroscopy (Scheme 2) [23]. Note that a variety of strain five-membered metallacyclocumulene complexes have been reported by the group of Rosenthal, while all of them contain substituents at the metallacycle [36,37]. The "non-substituted" metallacyclocumulenes might be too instable to be isolated, in agreement with general statements in the introduction. Therefore, it is surprising that the "non-substituted" 5d, 6 and 7 can be stable at room temperature. Notably, after the comparation of the bond distances and bond angles in the metallacycles, the authors proposed that the η 4 -π,π mode in the titanium complex 6 has a larger contribution in relative to those in similar Zr and Hf complexes. Scheme 2. Synthesis of "non-substituted" metallacyclopentynes 5d, 5e, 6 and 7.
In fact, an alternative synthetic route to target the titanium complex 6 was developed by Rosenthal and coworkers in prior to that shown in Scheme 2. When ClCH 2 C≡CCH 2 Cl was added into two equivalents of Cp 2 Ti(η 2 -Me 3 SiC≡CSiMe 3 ) in hexane, complex 6 was formed rapidly, accompanied by the generation of Me 3 SiC≡CSiMe 3 and Cp 2 TiCl 2 (Scheme 3) [38]. The same group later synthesized 1-zircona-2,5-disilacyclopent-3-yne 8, through reduction of Cp 2 ZrCl 2 and ClMe 2 SiC≡CSiMe 2 Cl by Mg (Scheme 4). The structural and spectral parameters, together with natural localized molecular orbital (NLMO) analysis, suggest 8 is a Si-substituted metallacyclopentyne with a weak interaction between the triple bond and the metal center, rather than a metallocene-stabilized 1,4-disilabutatriene complex [39]. The results are consistent with Suzuki's all-C metallacycloalkynes 5 [20,34]. The other type of metallacyclopentynes, with an alkylidene at each α-carbon of the C≡C bond, was designed by the group of Suzuki, either.

Formation of Alkyne-Coordinated Complexes
As mentioned previously, Suzuki and coworkers regarded the major contributor of metallacycloalkynes 5 as the η 2 -σ,σ form 5A (Figure 4) [34]. 5A contains a free C≡C bond, which might act as a 2e − ligand to be coordinated with other metals. Thus, the reactions of 5d with Cp 2 Zr(η 2 -CH 2 =CHEt)(PMe 3 ) or (PR 3 ) 2 Ni(η 2 -CH 2 =CH 2 ) (R = Ph or Cy) were carried out, and as expected, the alkyne-coordinated products 14, 15a, and 15b were obtained, respectively (Scheme 7) [35,45]. It is noteworthy that, after comparing the structural data between 15a and 15b with 5d, 14, and the Ni(0) zirconacyclocumulene complex Cp 2 Zr[µ-(η 4 -PhC 4 Ph)]Ni(PPh 3 ) 2 , the authors pointed out that in both of 15a and 15b, the interactions of the C≡C bond with Zr center are stronger than those in 5d and 14 [45]. On the other hand, if the starting metal precursors do not contain phosphine ligand, symmetric bimetallic complexes would be produced. For example, upon treatment of 5d with Cp 2 ZrCl 2 and Mg, or Cp 2 Zr(n-Bu) 2 , complex 16 was afforded, and the alkyne-coordinated intermediate was not detected (Scheme 8) [46]. 16 can be regarded as a µ-trans-butatriene complex, with a "zig-zag" C4 ligand connecting the two zirconocenes. Alternatively, 16 could be synthesized directly from ClCH 2 C≡CCH 2 Cl with two equivalents of Cp 2 ZrCl 2 and three equivalents of Mg, or from 14 with BEt 3 [46]. Two analogues 17 and 18 were produced similarly (Scheme 8) [38,47].

Reactions with Electrophiles
Organic alkynes can react with electrophiles to give alkenes, in contrast, the reactions between five-membered metallacycloalkynes and electrophiles normally result in ring-opening, due to their high ring strain. For example, treatment of 5d in THF with excess HCl in diethyl ether (1.0 M) gave a mixture of 1,2-butadiene and 2-butyne in NMR yields of 83% and 3%, respectively (Scheme 9) [48]. To investigate the reaction mechanism, the amount of HCl in ether was decreased to 0.3 equivalent, and the protonated intermediate 19 (R = H) was observed. Together with a deuterium labeling experiment, it was proposed that 19 is in equilibrium with 19A, which is the originality of 2-butyne (Scheme 9) [48]. When 5d was replaced by 5a or 5b, the major products would be the corresponding alkynes rather than allene derivatives (Scheme 9) [20], probably because the equilibrium favors 19A under the effect of the bulky R group [48]. Scheme 9. Protonation of 5d 5a and 5b.
Similar to the formation of 19, the group of Rosenthal found that B(C 6 F 5 ) 3 could activate one of the α-C-Zr bonds in their five-membered zirconacycloalkynes as well. When 20 and 21 were treated with B(C 6 F 5 ) 3 , two zwitterionic boranate complexes 22 and 23 were generated, respectively (Scheme 10) [49,50]. In each of complexes 22 and 23, one of the ortho-F atoms has certain interactions with the Zr center. These two complexes are analogous with the products of metallacyclocumulenes with B(C 6 F 5 ) 3 [51]. 1-Titanacyclopent-3-yne 6 displayed a similar reactivity, except that no F-Ti interaction is found in the final product [49]. In the presence of a proton source, metallacycloalkynes could further react with aldehydes. For example, treatment of 5d and triethylamine hydrochloride with benzaldehyde or propylaldehyde gave 2-methyl-1-phenyl-2,3-butadien-1-ol and 1-ethyl-2-methyl-2,3-butadien-1-ol, respectively. The intermediate was also supposed as 19 (R = H), which further attacks the carbonyl group of aldehyde, accompanied by the formation of O-Zr bond and the cleavage of Zr-C bond, to form an allene moiety. After the quenching by H + , the final products are generated (Scheme 11) [48]. Scheme 11. Reactions of 5d with aldehydes in the presence of NHEt 3 Cl.

Isocyanide Insertion Reactions
As discussed in Scheme 5, 10b can be converted back from 11b with PMe 3 [41]. Tert-butylisocyanide has a strong coordination ability, thus it might behave similarly as PMe 3 . Indeed, the formation of 24 was achieved expectedly upon treatment of 11b with 1 equivalent of t-BuNC. Interestingly, excess of t-BuNC would result in the carbon insertion into the three-membered zirconacycle to form intermediate 25, which could further rearrange into 26 with a cyclobutene structure (Scheme 12) [41,43]. Actually, similar reactivities of the "non-substituted" zirconacyclocumulenes 5d and 20 with t-BuNC had been reported earlier [52].

Formation of 1-Metallacyclopenta-2,3-Dienes
By the treatment of the five-membered zirconacycloalkyne 11a with two equivalents of Li (or KC 8 ), a two-electron reduction occurred, generating a dianionic complex. Addition of catechol produced the major product 27, and a minor zirconacycloallene 28. However, the addition of catechol is not a good method for the synthesis of zirconacycloallene, because the yield of 28 was very low (18%).

Synthesis of Metallapentalynes
The first metallapentalynes with the structure of B shown in Figure 2 were reported by Xia group in 2013. Treatment of 31 with terminal alkynes, including methyl propiolate, ethyl propiolate, and tert-butyl propiolate, produced the cationic osmapentalynes 32-34, respectively (Scheme 14) [54]. On the other hand, when internal alkynes dimethyl but-2-ynedioate or dimethyl but-2-ynedioate were used, the neutral products 35 and 36 were afforded (Scheme 14) [55]. The X-ray single crystal structure of complex 32 is shown in Figure 5. The Os center is in an octahedral geometry. The nearly coplanar metallabicycle is composed of seven carbon atoms and one osmium atom, with the mean deviation from the least-squares plane as 0.0415 Å. Besides the three carbon atoms, there are one Cl and two PPh 3 groups linking directly with the Os center. The C-C bond lengths in the two fused rings are with small alternation (between 1.377 and 1.402 Å). The Os-C1 triple bond distance is 1.845 Å, slightly longer than those in acyclic osmium carbynes (1.671-1.841 Å). The most remarkable feature is the Os≡C1-C2 angle (129.5 • ), which is much smaller than those around the sp-carbons in trans-5a as discussed previously (154.2(6) • and 154.4(7) • ) [20], and represents the smallest one around a carbyne carbon at that time (the smallest angle now is 127.9(5) • in complex 35) [55]. How can such a small angle exist in a stable organometallic complex? To answer that question, density functional theory (DFT) calculations optimized at the B3LYP level were performed. Compared to the estimated angles around acetylenic carbons in organic cyclopentyne (116.0 • ), the 129.5 • in 32 indicates the great release of angle strain. In fact, due to the bridgehead Os atom, the computed strain energy (24.3 kcal mol −1 ) on the basis of a chain model molecule is much smaller than that of cyclopentyne (71.9 kcal mol −1 ). Besides, according to a model in which the PPh 3 ligands are simplified to PH 3 , the authors stated that complex 32 is a cyclic eight-center eight-electron Craig-type Möbius aromatic complex, which is supported by the calculated values of NICS(0) zz (−11.1 and −10.8 ppm in each ring) and isomerization stabilization energies (ISE). Moreover, the down-fielded proton chemical shifts (H3, 8.32; H5, 9.27; H7, 14.25 ppm), and delocalized C-C bonds in the osmapentalyne cycle as described in the previous paragraph, further confirm its aromaticity. Thus, the stability of metallapentalynes is attributed to both of the reduced ring strain and the aromaticity. Notably, organic pentalyne is anti-aromatic, hence the incorporation of a metal into organics might change their aromaticity [54].
Alternatively, metallapentalynes can be synthesized directly from organics with metal precursors. Xia and coworkers found that triynes with a hydroxyl group reacted with MCl 2 (PPh 3 ) 3 (M = Os or Ru) in the presence of PPh 3 to give a series of metallapentalynes 37-42 (Scheme 15) [56,57]. When X = CH 2 CH 2 , the alkyne-coordinated intermediate was isolable [56]. It is worth pointing out that complexes 41 and 42 are the first metallapentalynes based on a second-row transition metal [57]. Sometimes, metallapentalenes [58,59], which is another kind of metallaaromatics, could act as the starting materials for the preparation of metallapentalynes. For example, in the presence of HBF 4 , osmapentalene 43 reacted with allenylboronic acid pinacol ester to afford osmapentalyne 44. Similarly, treatment of 43 with phenylacetylene and HBF 4 led to the formation of osmapentalyne 45. Tetracyanoethylene (TCNE) was active, too, and under the existence of NaPF 6 , osmapentalyne 46 was generated (Scheme 16) [60].

Formation of Metal Carbyne-Coordinated Complexes
Similar to five-membered metallacycloalkynes, in which the alkyne can act as a 2edonor to be coordinated with other metals, the metal carbyne moiety in metallapentalynes is somehow "alkyne-like", thus can also react with an external metal precursor to form metal carbyne-coordinated complexes [57,62].

Reactions with Electrophiles
As mentioned in Section 2.2.2, Brönsted acids are active and can easily break the metallacycle of five-membered metallacycloalkynes. Metallapentalynes are also sensitive to acids, however the fused rings are always reserved, indicating their higher stability owing to the aromaticity.
Treatment of osmapentalyne 32 with HBF 4 ·Et 2 O at room temperature for 30 min resulted in the formation of osmapentalene 54 [54,63]. Complex 54 is not stable in solution due to its 16echaracter and is easily converted to the new 18eosmapentalyne 55 (Scheme 19). During the process from 54 to 55, there are two possibilities: Elimination of H a or H b . DFT calculations optimized at the B3LYP level in a model complex suggest that the deprotonation of H b is more favorable both kinetically (5.1 kcal mol −1 difference in energy) and dynamically (0.9 kcal mol −1 difference in energy), thus forming 55 selectively [63]. In comparison to the structures between 32 and 55, we can see clearly that the addition of HBF 4 ·Et 2 O resulted in the shift of Os≡C from one cycle to the other. Some other metallapentalynes, such as osmapentalynes 33 and 34, and ruthenapentalyne 42, display similar metal carbyne bond shift reactivity [54,57]. Complex 55 could further react with other electrophiles. Upon the addition of ICl or Br 2 , 55 was converted to iodocarbene 56 and bromocarbene 57, respectively (Scheme 20) [64]. 56 and 57 are the first metallaiodirenium ion and metallabromirenium ion, which have similar structures as the proposed intermediates in the halogenation of alkynes. The results further support the "alkyne-like" properties of M≡C in metallapentalynes. Interestingly, nucleophilic substitution of 56 by (n-Bu) 4 NBr was observed, and 57 was produced. However, when (n-Bu) 4 NCl was used, the regeneration of 55 rather than the formation of an analogous metallachlorirenium ion was discovered. This is probably because of the smaller atomic size of the Cl atom compared to Br and I, which makes the three-membered metallachlorirenium ring unstable. In addition, Se was another electrophile to be active with 55, and osmapentaloselenirene 58 was given (Scheme 20) [65]. 58 was regarded as the first complex that contains an unsaturated Se-containing ring, in which the aromaticity is dominated by σ-type. Alkynes could act as electrophiles to react with metallapentalynes, either. Treatment of osmapentalyne 59 with phenylacetylene in the presence of HCl·Et 2 O resulted in the formation of 60, in which a trans-alkenyl is positioned in the α-position, accompanied by the C≡C bond shift (Scheme 21). This reaction is highly efficient, and a series of other alkynes are also applicable [66]. Scheme 21. Electrophilic addition reaction of osmapentalyne 59 with phenylacetylene.

Reactions with Nucleophiles
Although there is no report on five-membered metallacycloalkynes with nucleophiles, such reactivities have been discovered for metallapentalynes. For example, under CO atmosphere, complex 32 reacted with CH 3 ONa or CH 3 SNa to give osmapentalenes 61 and 62, respectively (Scheme 22) [63]. On the other hand, when CO was replaced by O 2 , in the presence of CH 3 OH, 32 was transformed to the osmium-peroxo complex 63, which is active for alcohol dehydrogenation reactions (Scheme 22) [67]. Given that alkynes sometimes can act as 2edonor analogous to CO, Xia et al. further treated 32 with CH 3 OH and Cs 2 CO 3 in the presence of different alkynes, and several osmafulvenallenes 64-67 were afforded (Scheme 22) [68]. It was proposed that the first step is the nucleophilic attack of methoxide to the Os≡C bond, the same as that from 32 to 61. After the coordination of alkyne to Os center, an isomerization occurs to form a vinylidene intermediate, which can be further subjected to addition of CH 3 OH. In the end, the ring opening followed by the ester coordination provides the final product. It is noteworthy that complexes 64-67 are the first metallafulvenallenes. Isocyanides are another kind of reactive nucleophiles. Treatment of 32 with t-BuNC or 1-adamantanyl isocyanide resulted in the formation of the C-C coupling complexes 68 and 69, where the isocyanides only connect with the α-carbon of the resulting osmapentalene ring. On the other hand, the η 2 -iminoketenyl products 70-74 were isolated, respectively, when tosylmethyl isocyanide, p-nitrophenyl isocyanide, cyclohexyl isocyanide, 2-naphthyl isocyanide, or p-anisyl isocyanide was used (Scheme 23) [69]. For 68 and 69, they do not adopt the η 2 -iminoketenyl coordination mode, probably because the bulky tert-butyl and 1-adamantyl groups inhibit bending at their isocyanide nitrogen atoms. It is worth mentioning that η 2 -iminoketenyl species were often proposed as the intermediates in nucleophile-induced carbyne-isocyanide coupling processes, while they had never been isolated before the publication reported by Xia and coworkers. Later on, the η 2 -iminoketenyl complexes were extended to ruthenium, reported by the same group [70]. The metal carbyne bond shifted complex 55 is also active to nucleophiles. For example, the addition of NaSH produced 75, and the reaction with PhCH 2 NH 2 in the presence of Cs 2 CO 3 led to the formation of 76 (Scheme 24) [54]. Interestingly, when PhNH 2 was selected as the nucleophile, under basic conditions, the product was a polycyclic aromatic complex 77, formed involving C-H activation. PhOH exhibited a similar reactivity as PhNH 2 and 78 was afforded (Scheme 24) [71]. Complexes 77 and 78 are the first metal-bridged tricyclic aromatic complexes, in which the three five-membered rings are all aromatic. By using the same strategy, a series of polycyclic aromatics [71][72][73], including one with six and one with seven fused aromatic rings, derived from 1-aminopyrene and 9-phenanthrenol, respectively, had been developed [71]. Notably, when the lactone-fused osmapentalyne 79 [74] was mixed with 3-(methylsulfanyl)-benzenol in the presence of K 2 CO 3 , the resulting polycyclic osmaaromatic complex 80 possesses high charge transport ability (Scheme 25) [75]. These results indicate that metallapentalynes and their derivatives have a very broad application prospect in the field of material chemistry.

Cycloaddition Reactions
In principle, cycloaddition of the C≡C bond in five-membered metallacycloalkynes would greatly reduce their ring strain, while surprisingly, such reactions have not been reported up to date. In contrast, many cycloaddition reactions involving the M≡C bond in metallapentalynes have been discovered.
The first [2+2+2] cycloaddition reaction of an alkyne with a late transition metal carbyne occurred between the neutral osmapentalyne 35 and HC≡COEt, generating complex 81 (Scheme 26) [55]. The three fused rings in 81 are not planar, with the carbene carbon of Os=C(OEt) almost perpendicular to the two five-membered rings. It was proposed that the reaction underwent a [2+2] intermediate, which could not be detected.  As just mentioned, [2+2] cycloaddition products are found between 55 and alkynes, such as HC≡COEt and HC≡CCOOH. Unexpectedly, when aryl alkynes were reacted with 55 in the presence of O 2 and H 2 O, the products were totally different. For example, complexes 87 and 88 were afforded, respectively, when phenylacetylene or 3-ethynylthiophene was selected (Scheme 28) [74,78]. 87 and 88 are the first α-metallapentalenofurans, and DFT calculations optimized at the B3LYP level suggest that their Möbius aromaticity is derived from d π -p π π-conjugation. Deuterium labeling experiments indicate the oxygen in the furan ring comes from H 2 O. Interestingly, the presence of HBF 4 induced the formation of two lactone-fused metallapentalynes 89 and 79, accompanied by the migration of Os≡C bond (Scheme 28) [74]. It is worthy to compare the formation of 89 and 79 with that of 60 shown in Scheme 21: During the formation of 60, H 2 O and O 2 are absent, and the reactant 59 does not contain a methyl ester substituent. Although [2+2+2] cycloaddition has not been reported between 55 and alkynes, it has been found between 55 and two molecules of 2,2-diphenylacetonitrile. The product is metallapentalenopyrazine 90, which contains three fused aromatic rings. 90 is the first aromatic complex with a metallapyrazine fragment (Scheme 29) [79]. Recently, the Xia group found a new type of "alkyne-like" reactivity. The authors used the osmapentalyne derivative 91, which already contains three fused aromatic rings, as the starting material to react with azides, and several polycyclic products 92-96 were produced (Scheme 30) [80]. They are "metalla-click" type reactions, and the products consist of four fused five-membered rings. Moreover, these rings are all aromatic. Before the publication, the maximum number of fused aromatic rings sharing with one bridgehead metal was three, thus these results broke through the record.

Conclusions and Perspectives
Challenging the limits is the driving force of social progress. Chemists have devoted to the isolation of small cyclic compounds with acetylenic carbon(s), for more than half a century. After continuous efforts, the number of atoms in the ring has been decreased to as small as five. Such complexes can be divided into two types: Metallacycloalkynes and metallapentalynes. For the former type, the angles around the acetylenic carbons are around 155 • , representing the smallest ones before the discovery of the latter type (around 130 • ). These angles are far away from 180 • , thus leading to extremely high ring strain and unusual reactivities.
For metallacycloalkynes, four main classes of reactions have been investigated, including the reactions with external metal precursors to form alkyne-coordinated complexes, the reactions with electrophiles, isocyanide insertion reactions, and the formation of 1-metallacyclopenta-2,3-dienes. The first two classes of reactions have been discovered for metallapentalynes as well, although for the reactions with electrophiles, the products were totally different. The addition of an electrophile, such as H + or B(C 6 F 5 ) 3 , into metallacycloalkynes normally resulted in ring opening of the metallacycles. In contrast, when H + , ICl, Br 2 or Se was added into metallapentalynes, the metallacycles could be preserved. The results might be attributed to the higher stability of metallapentalynes resulted by their aromaticity. On the other hand, the last two classes of reactions for metallacycloalkynes have not been discovered for metallapentalynes, whereas two other "alkyne-like" reactivities, namely nucleophilic addition and cycloaddition reactions, had been explored.
Compared to the abundant chemistry of alkynes, the research on five-membered metallacycloalkynes and metallapentalynes is still limited. Except one example based on thorium (complex 13), five-membered metallacyclopentynes are restricted on group 4 metals; metallapentalynes have only been discovered for ruthenium and osmium complexes. Especially, in consideration that five-membered metallacycles are regarded as intermediates in many transition-metal catalyzed reactions, and the reactivity studies of metallapentalynes have given a variety of novel metallapentalene derivatives, which showed a great application prospect in various areas, such as near-infrared dyes [54], photothermal therapy [82][83][84], phototherapy [85], and self-healing materials [86], the chemistry on these two kinds of five-membered metallacycles is still far more than anticipated. Moreover, as special organometallic complexes, the application prospect of metallacycloalkynes and metallapentalynes in catalytic transformations should not be underestimated.

Conflicts of Interest:
There are no conflicts to declare.