Abstract
A personal history of sumanene chemistry: from the encounter with Prof. Mehta’s first report to the synthesis of pristine triazasumanene is described.
1. Introduction (Prologue)
It was in the end of October, 1993 when autumn was deepening and people began to miss the summer warmth day by day, I was in the library of the Chemistry Department, at the University of Tokyo. There was a sofa in the section of newly arrived journals, and on the side a precision glasswork of the chemistry building1 by Mr. Mitsugi Ohno2 was displayed. The sunlight shining through the window made it warm and drowsy. Unlike now, neither electronic journals nor Web search tools were available, and even if you were using CAS online, it had to be accessed at midnight to save the cost of an international phone call, therefore it was still common for a literature survey to take half a day in the library piling up a large number of Chemical Abstract Collective Volumes. In other words, it was a good excuse to go to the library not only for the literature survey, but also for relaxing or changing moods. Although I was supposed be busy preparing for the preliminary examination of my doctoral degree, I visited there for a short break, sat on the sofa, and randomly looked over the journals. Since I had studied chromium acyl/carbene complexes during my Ph.D. work,3 my major interest was, of course, organometallic chemistry. However, whenever one looked over the bound journals, one could not help seeing papers that seemed not directly related to one’s own studies. On that day, I paused over a manuscript in JCS Chem. Commun. written by Prof. Mehta’s group (Figure 1).4
Prof. Mehta’s first report on sumanene from J. Chem. Soc. Chem. Commun.1993, 1006; with permission from Royal Society of Chemistry.
I recognized Prof. Mehta as a giant in Indian Organic Chemistry, so I was very surprised that the paper sounded like a kind of “failure” report for the synthesis of sumanene (1) (eq 1). I also remembered the recent reports on the preparation of corannulene by Profs. Scott and Siegel (eq 2).5 Since I was not so aware of the strain in these molecules, at least judging from a look at the structure of sumanene, it was mysterious for me why even Prof. Mehta could not achieve the synthesis of sumanene. This was the dawn of my sumanene chemistry as well as the start of my history with π-figuration.
I do not want to demonstrate a full account of my sumanene chemistry here, but rather show how I established myself in sumanene and azasumanene chemistry from 1993 to 2014. Comprehensive reviews of sumanene and the related buckybowls can be found in previous articles.6 Having said that, I should add a brief introduction to buckybowls for the readers. After the discovery of C60,7 the so-called buckyball, the corresponding partial structure, the buckybowl, has also been studied for many reasons including as a potential precursor for the chemical bottom-up synthesis of C60, and also for the unique chemical and physical properties stemming from its “bowl” shape. Fullerenes consist of the combination of hexagons with 12 pentagons in accordance with Euler’s formula, hence the buckybowls consist of hexagons and less than 11 pentagons (buckybowls possessing 6 pentagons are defined as cap structures of carbon nanotubes).8 Two pristine frameworks of buckybowls are known: one is pentagon-centered (C5 symmetric) corannulene, and another is hexagon-centered (C3 symmetric) sumanene (1) (Figure 2). Indeed, the history of corannulene is much older than that of fullerene, and its first synthesis by Barth and Lawton was reported in 1966.9 Corannulene chemistry revived after the discovery of C60, followed by the innovative flash vacuum pyrolisis (FVP) syntheses in 1991.5 Later on Prof. Siegel developed a kilogram-scale synthesis.10 Now corannulene is commercially available and its chemistry has been expanded tremendously.11 On the other hand, Prof. Mehta’s failure report had a serious impact in delaying sumanene chemistry behind corannulene chemistry; the only successful report was the FVP synthesis of trithiasumanene by Prof. Ohtsubo (eq 3).12
Pristine buckybowls, corannulene and sumanene (1).
2. Results and Discussion
2.1 Retrosynthetic Analysis
It is impossible to remember all of my own history, but my research notebooks record what I thought at that moment. My memorandum about sumanene suddenly appeared from October 1993, immediately after I encountered Prof. Mehta’s paper. The copy shown in Figure 3 is the very earliest stage of my retrosynthetic consideration of sumanene. I was fortunate that (now I can easily recognize that) I had already figured out my principal idea about the final precursor of 1, hexahydrosumanene (2), so I did not go in the wrong direction. Moreover, I was fortunate not to start the sumanene chemistry at that moment because the rest of the idea, leading to 2, was too awkward to realize. I also knew that my plan was not yet ready for investigation. Indeed, I kept pondering the retrosynthesis occasionally, but a bright idea did not appear for a long time. I mostly spent my time in the investigation of organochromium13 and related organometallic chemistry.
The first retrosynthesis of 1 (October 1993). The number of the compounds 1, 2 and A was added.
The next break-through came out when I was a JSPS postdoc in Prof. Chuck Casey’s group at the University of Wisconsin-Madison for the purpose of honing my experimental techniques and gaining deeper insights into organometallic chemistry (from April 1996 to March 1998).14 During the postdoc time, I felt free and was able to spend a lot of time considering my future chemistry projects, one of which was the synthesis of sumanene. Three years after the encounter with Prof. Mehta’s paper, the memorandum on October 1996 (Figure 4) indicated that I noticed the cyclotrimerization of norbornadiene useful for this purpose (I do not remember how it came, but probably I happened to see a paper of Prof. De Lucchi15 at the time). Then 3 would be converted to 1 through ozonolysis, pinacol coupling, and the dehydrative aromatization. It was amazing that I formally solved the stereoisomer problem of hexaformyltriindane (A and A′), which could not be solved in 1993.
The second retrosynthesis of 2 (October 1996). The number of the compounds 1, 3 and A′ was added.
Soon after, I had an opportunity to attend a lecture by Prof. Grubbs. To be honest, I recognized that Grubbs metathesis was an excellent organometallic reaction, but could not imagine how it would go on to become a powerful tool for practical synthetic chemistry. In this sense, his presentation struck me as shown in my memo on February 1997 (Figure 5).
The memo about the idea for the tandem metathesis (February 1997).
Since I felt I had gotten together all the pieces for the synthesis of 1 (I learned the importance of continued thinking), I decided to start sumanene chemistry after returning to the University of Tokyo as an assistant professor in April 1998. Because I still belonged to Prof. Narasaka’s group, and also for the sake of future grants, I needed to carefully defend my idea for the new project. Then I wove a logical sequence of significance for sumanene synthesis for the purpose of developing a step-wise synthesis, leading to the novel heterofullerene synthesis, and I anticipated mainly two routes, ozonolysis/McMurry and tandem ROM/RCM (Figure 6).
Part of a research proposal for sumanene chemistry (June 1998) (the numbers of compounds are different from those in this manuscript). Translation written in Japanese: Research Report 1998 June 17. Hidehiro Sakurai. 1. Synthesis of novel aromatic and heteroaromatic compounds directed toward the synthesis of heterofullerene. Heterofullerene: hetero-atom substituted fullernene. In particular, molecules having a pore structure such as 2, have potential as ligand or host molecule. For the approach to the synthesis of them, it is necessary to develop a traditional bottom-up synthesis of C60 and C70, which can be arranged to the heteroatom-substituted derivatives. Therefore, the retro-synthesis of C60 is proposed below (retro synthesis part 1 is omitted). Retro-synthesis part 2: synthetic route to sumanene.
After publishing our first paper in 2003,16 Professor De Lucchi kindly sent us a letter of celebration, and realized that he was also considering the same synthetic plan.17
2.2 Synthesis of 1.16
The research began from the cyclotrimerization of the norbornene unit based on the retrosynthesis. Two methods for the preparation of 3 were already reported by De Lucchi’s group. One involves the direct trimerization of a α-bromocuprio-norbonadiene species from the corresponding deprotonated intermediate by the Schlosser base.15 The other was to trap the same intermediate as a Me3Sn derivative, followed by the Cu(2-thiophenecarboxylate) (CuTC)-mediated cyclotrimerization.18 For both cases, the undesired anti isomer must be the major product, and the separation might be a serious problem; therefore, first, I desired to develop a new method for the syn selective cyclotrimerization. As a result, I consumed nearly two years attempting the acid-catalyzed trimerization of ketone; most of the known methods19 cannot be applied to the trimerization of norbornanone derivatives, probably due to steric problems. The trials were continued after I moved to Osaka University (from April 2000). The first student in this project, Mr. Takanori Kawakami, an undergraduate student, joined for a year, and his year was completely dedicated to this cycrotrimerization, and as a result, he had to give a no-data presentation for the final year report. The next year (April, 2001), the next student, Mr. Taro Daiko, was assigned as this project investigator. Since Taro’s final report for his Bachelor’s was also unsuccessful, we could not ignore the criticism and the pressure to terminate this project.20 Therefore, we changed tactics to prepare 3 by the reported methods when Taro started his Master’s project. Because we were restricted in handling Me3Sn derivatives due to their high toxicity, Me3Sn was replaced by Bu3Sn to get 4 in initially 75% yield (now improved to 92%).21 The CuTC-mediated cyclotorimerization also worked using 4, giving the syn/anti mixture of 3 (first reported at 38% yield, then improved to 88% yield). The syn/anti isomers were carefully isolated by GPC for the next step of the investigation.22
When I considered this route (Scheme 1), I imagined that the most difficult and tricky process would be the conversion from 3 to 2 (in reality, the most difficult process was the first cyclotrimerization!). Among two retrosynthetic analyses, we expected that the ozonolysis-McMurry coupling route might be more promising. However, to our surprise, 3 was totally intact under ozonolysis conditions. Therefore, we chose the tandem ROM-RCM route. Because 3 possesses three benzonorbornadiene units, which are highly reactive toward the ROMP, we decided to carry out the reaction under ethylene atmosphere to avoid the ROMP and to trap a hexavinyl intermediate, and also to use the first generation Grubbs catalyst to avoid the heat to generate the reactive coordinatively unsaturated species. Unlike the previous ozonolysis approach, the spot of 3 immediately disappeared after treatment with the Grubbs catalyst. The crude NMR also surprised us: we did not find the expected patterns of terminal vinyl protons, instead, the direct formation of hexahydrosumanene 2 even under the ethylene atmosphere! This strongly indicated that 2 must possess a thermodynamically favored and less-strained structure. So, we achieved the 3 to 2 conversion through tandem ROM-RCM route after only a couple of attempts. The final aromatization process was rather easy. When 2 was treated with DDQ, a new product was detected. The 1H NMR showed only one singlet peak and the doublet pairs, strongly indicating the formation of sumanene 1.
Synthetic route of sumanene (1).
The main problems in the manuscript preparation were the reproducibility and the scale of the reactions. In particular, the yields of the first two reactions were not stable, and we always obtained 1 only in several mg. Purification of 1 was another problem because of the small scale, therefore 1 did not pass elemental analysis (EA). And 1 gave very thin cotton-like crystals, which were impossible to measure by single crystallographic analysis. Anyhow, we wanted to publish in a hurry, so we submitted a simple communication involving the synthesis and the preliminary bowl inversion analysis to Journal of the American Chemical Society; however, it was ultimately rejected mainly due to the lack of the evidence of 1 such as an EA and/or a crystal structure. We needed to get further confirmation of 1, but judging from the preparation scale, it would take time to obtain EA/XRD data; we decided to submit this manuscript to Science while collecting data. Things always move in unexpected directions: we received a letter from the editorial office of Science that it was acceptable if we did not mind changing it to a “Bravia”.16
Ten years had already passed since I read the paper of Prof. Mehta, and even five years since I myself started the experiments. I was so lucky to study at such a slow pace, probably because not so many synthetic chemists were engaged in π-figuration chemistry, totally unlike the current highly competitive environments.
2.3 The Crystal Structure of 1
Mr. Taro Daiko was a very self-disciplined person. When Taro joined this project, he claimed to me that his ultimate goal was to be a patent attorney. Hence, we made an agreement that he should terminate any experiments before 5 pm to keep time go to prep school. Indeed, he lived such a dual life for two and a half years. In the summer of 2003, after he passed the examination to enter the Japan Patent Office, Taro changed his manner to concentrate on the research for the final half year, and he carried out several important experiments such as a generation of benzylic anions.23 A few weeks after his successful master thesis presentation, he brought the final (revised) version of his thesis, and I was completely surprised to find that he had added totally new data about the single crystal structure. He was satisfied with my reaction, and said that he had been secretly investigating with the help of Dr. Toshiyuki Moriuchi, a senior colleague in the same Hirao group, and he wished to surprise me at the final moment.
The data was what we published in 2005.23 Just at a first glance, I was so impressed with the packing structure (Figure 7). Unlike the packing of the pristine corannulene, that of sumanene exhibited a perpendicular columnar structure, and in addition, all of the column was aligned in the same direction, which indicated that this crystal intrinsically possesses a large dipole moment. This unique packing structure of 1 might have increased the interest in sumanene chemistry, and led to extensive work such as that on electron conductivity,24 the thermoelectric properties,25 and the liquid crystal mesogen.26
Crystal packing of 1.
2.4 Synthetic Plan of Triazasumanene 5.27
In 2003, I was given an opportunity to interview for the PI position at the Institute for Molecular Science (IMS), Okazaki, Japan. IMS seemed to be an arcadia for the young chemists in Japan: all the associate professors are guaranteed the independence, sufficient lab space and grants, and professional support, the only problem being that it was hard to get students due to the lack of an undergraduate program. It was also said that the committee would evaluate the candidates by their proposal during the interview, and previous achievement was not so seriously considered. Indeed, when I was invited to the interview, my sumanene paper was still under review, which meant I had no career (at least in my publication list) in this field. Therefore, I gave a research proposal without any publication background. I was extremely grateful for the committee who chose me even at this stage, and I was also relieved and I felt I had given back to IMS when our first paper was accepted.
My IMS position formally began in October 2003, but I still needed to take care of the students in Osaka including Taro, so I was only fully engaged at IMS from April, 2004. I was given a vacant room in a brand-new building, and I started alone from the construction of the laboratory. Two months later, Dr. Shuhei Higashibayashi joined me to start research together. Shuhei was also selected through the open recruitment. He had been engaged in the total synthesis of natural products, and before joining IMS he was a postdoc in Prof. Kishi’s group at Harvard. We already decided to pause for the first several years to investigate simple functionalization starting from 1 because we wanted to maximize the merit of Shuhei’s tremendous talent for synthetic chemistry; indeed, as I commented above, there were not so many synthetic chemists involved in physical organic chemistry (in this sense I was also thankful for Shuhei’s decision to join this field).
As you can see in my hand-written proposal at the beginning of sumanene chemistry (Figure 6), one of the perspectives to develop the step-wise, conventional synthetic route was to be applied to the corresponding heteroatom-substituted framework construction. Hence, we decided some of our target molecules should be heterobuckybowls including triazasumanene 5.27
The retrosynthetic route is shown in Scheme 2. Based on the same approach to sumanene, the precursor of 5 should be the hexahydro derivative 6′, which would be approached from the cyclic tri-imine 7′ through the imine metathesis. However, this route looked like it might be impossible because 7′ would be too strained a structure to synthesize. Therefore, a detour had to be considered, and one of the practical and synthetically accessible functional groups for an imine substitute is amide. Although multiple steps might be necessary to convert 6 to 5, the cyclic lactam 7 may not be such a strained compound thanks to the single bond character of C-N, and the amide-metathesis might be possible through the hydrolysis-condensation process. For the preparation of 7, we were able to find a similar cyclotrimerization method from the lactam 8 derivatives.
Retrosynthetic analysis of triazasumanene 5.
However, the cyclotrimerization process became a problem for us again. In the case of sumanene synthesis, the cyclotrimer 3 is an achiral molecule; therefore, when starting from racemic 4, the syn/anti ratio is distributed statistically 1:3, even the desired syn isomer is a minor product. If a similar cyclotrimerization was performed for the preparation of chiral (unsymmetrical) cyclotrimer like 7, six possible isomers might be obtained and yield of the desired syn-/symmetrical-isomer would be sparse. To achieve the syn-selective and homochiral cyclotrimer, two requirements must be satisfied: 1) the reactant must be enantiomerically pure, 2) there must be perfect cross coupling at X and Y partners and no mismatched coupling. Therefore, a new method for the cyclotrimerization was urgently needed (Figure 8).
Requirements for C3-symmetric homochiral trimerization.
2.5 Synthesis of C3-Symmetric Chiral Trimethylsumanene 9.28,29
To simplify and to make the method more versatile, we decided to investigate in parallel the synthesis of C3-symmetric chiral sumanene derivatives. That is, C3-symmetric tri-substituted sumanene 928 has an inherent chirality30 (bowl chirality) stemming from the bowl structure. To achieve the asymmetric synthesis of 9, the same cyclotrimerization method needed to be developed. In addition to the first cyclotrimerization problem, two more problems needed to be solved. One was to make the tandem ROM-RCM conditions suitable for the tri-substituted alkene, which is sometimes very different from the di-substituted alkenes. Another was that the final oxidative aromatization must be carried out at low temperature. This aromatization process means quite significantly that the sp3 chirality is converted to bowl chirality. When the molecule possesses sp3 chiral carbon(s), the racemization must not happen, but once these sp3 chiral carbons are oxidized to the (pyramidized) sp2 carbons, bowl-to-bowl inversion equal to racemization might start. The activation energy of bowl-to-bowl inversion can be estimated by DFT calculation in advance of the experiment,5a,28,31 indicating that the half-life racemization at 20 °C, 0 °C, and −20 °C is 9 min., 2 h, and 2 days, respectively. This indicates that a temperature at least below 0 °C might be necessary to observe the racemization process, and −20 °C for the isolation as an enantiopure or enantioenriched species to carry out the aromatization reaction, followed by the isolation process.
Fabris and De Lucchi’s group already reported some examples of the CuTC-mediated cyclotrimerization from enantiopure α-bromo-Me3Sn-norbornene derivatives, and indeed, some derivatives gave excellent syn selectivity (eq 4).18,32 However, the syn-selectivity was not general, unfortunately, and even anti-selective results were obtained. In addition, such enantiopure tin compounds are not so easy to prepare synthetically. Therefore, an alternative method was desired, and if possible, the starting reactant should be easy to synthesize. De Lucchi reported an example of the Mizoroki-Heck-type cyclotrimerization of bromo-benzonorbornadiene (eq 5).33 Although the reaction started from the racemate and the perfect anti-selective reaction took place, the ease of the preparation of the reactant as well as the high conversion (95% yield) fascinated us and inspired us to investigate the Mizoroki-Heck-type reaction.
Shuhei started to screen the reaction conditions using an enantiopure iodonorbornene (1S, 2R, 4S)-10, derived from the corresponding chiral norbornanone, but changing the phosphine ligands to Pd did not affect the reaction. At the same time, our lab was also engaged in metal nanoparticle catalysis in collaboration with Prof. Tsukuda’s group at the same institute,34 which might have given him a hint to develop the ligand-free, in-situ generated Pd nanoparticle-catalyzed conditions. Addition of Bu4NOAc was indispensable for successfully improving the dispersibility and the stability of the finely generated Pd particles. The yield of the cyclotrimer 11 was around 50–60%, but with syn selectivity and enantiomerically pure form.35 This cyclotrimerization could be applied to construct several chiral C3-symmetric compounds (eq 6).36
The proposed mechanism (Scheme 3) involves the repetitive insertion of vinylpalladium into the iodoalkene, followed by HI elimination, and from the thus-formed trienyl-Pd species, electrocyclization might occur to form the syn Pd-H intermediate, which spontaneously suffered elimination to afford 11 with syn selectivity. Since the entire process occurs in highly sterically crowded space, ligandless and naked Pd need to be supplied to the reaction mixture. The key role of the Pd nanoparticle is both to deliver the naked Pd species and to be a reservoir of such a highly reactive Pd(0) species to avoid the formation of Pd black (bulk solid), the so-called boomerang machanism.37 Later, the precise reaction mechanism was further investigated by DFT calculation with the help of Prof. Yamanaka, leading to the realization that the mechanism must be more complicated.38 Here I would like to emphasize one finding from these calculations, that is, the electrocyclization of triene is significantly accelerated in the case of Pd derivatives in contrast with the non-metallated triene.
Proposed mechanism of Pd nanoparticle-catalyzed cyclotrimerization.
Thanks to the new cyclotrimerization method, we were free from the syn/anti problem of the cyclotrimer. Starting from (1S, 4S)-12, the desired syn-13 without contamination of anti-isomer was obtained in 55% yield, which was converted to the methyl-substituted precursor of the metathesis reaction 14. The tandem ROM-RCM reaction of tri-substituted alkene was carried out either by the combination of Grubbs 1st (ROM) and 2nd (RCM) catalysts,28 or by changing the sacrificial alkene from ethylene gas to (Z)-oct-4-ene.29 Shuhei struggled with the final aromatization and the isolation of the enantiopure (or enriched) trimethylsumanene 16. The suitable oxidant for the aromatization at such a low temperature was very rare, and moreover the following (low-temperature) isolation was expected to be easy. Finally, he found that DDQ oxidation occurred very quickly (1 min.) when it was carried out at 0 °C, then the reaction was immediately cooled down and quick separation was done by column chromatography at −20 °C to avoid the bowl-to-bowl inversion (Scheme 4).28
Asymmetric synthesis of (C)-9.
We were delighted that the decay of the CD spectra unambiguously determined the bowl-to-bowl inversion rate. Previous methods for the determination of the inversion rate rely on the presence of a diastereotopic proton in the molecule, because 1H NMR techniques such as DNMR or EXSY were utilized. In contrast, the CD method does not require the existence of the diastereotopic proton, but instead prepares the chiral bowl by synthesis or resolution (Figure 9).28,29,39
(a) Decay of CD spectra of (C)-9 in CH3CN at 10 °C. (b) Decay of CD spectra of (C)-9 at 247 nm in CH3CN at 10 °C.
2.6 Synthesis of Triazasumanene 5.27,40
It seemed that we were ready to start the synthesis of triazasumanene. Indeed, Shuhei had already advanced the synthesis of the chiral iodoazanorbornene (1S, 4R)-17, a precursor for the cyclotorimerization, and luckily the Pd-catalyzed cyclotrimerization also proceeded without problem to provide the syn-cyclotrimer 18 at 57% yield. He spent a long time trying to find suitable conditions for the amide metathesis, but finally succeeded in optimizing at 59% yield in two steps. As we knew from the start, the final conversion from the bowl-shaped trilactam (19) to the final triazasumanene was extremely difficult. Once we probably obtained the trichloro triazasumanene derivative through the α-chloroimine intermediate, but we failed to achieve reproducibility, and the intermediates were all susceptible. Therefore, we gave up that route.
In August 2010, Dr. Qitao Tan joined our group as a JSPS postdoctoral fellow, and he succeeded in this project. After a long struggle, he decided to use the thioamide for the further conversion. Initially, he hesitated about this thioamide route because it would have required longer steps to the final triazasumanene; however, it was practically useful because the conversion to the MeS-substituted hexahydro form (21) proceeded relatively smoothly with good reproducibility. Contrary to the sumanene synthesis, the aromatization process from 21 to 22 was not so easy. DDQ or related quinone-type oxidants did not work, but trityl cation-type oxidants were suitable. Finally, the use of a sterically bulky base, 2,6-di-t-butyl-4-methylpyridine, maximized the yield to 77% (Scheme 5).40
MeS-triazasumanene (22) synthesis.
Because of the highly-strained structure, we wondered whether we could measure the bowl-to-bowl inversion behavior of 22, which might be competitive with the thermal decomposition. The inversion barrier was estimated to be approx. 40 kcal/mol, indicating that conditions over 200 °C were required. We were pleased that 22 allowed for measurement of the racemization process simply by chiral HPLC as 42.2 kcal/mol.
Our next interest was to obtain the crystal structure, because we did want to see such a deep bowl structure directly, and also because we anticipated a similar unique packing structure like in sumanene derivatives. Unfortunately, the MeS derivative 22 had poor crystallinity, whereas tiny crystals were obtained from the sulfone derivative 23.40 As expected, we can visualize the deep bowl structure with the highest POAV angle as 10.8°, but its featureless packing structure disappointed us (Figure 10). It was apparent that the sulfone group significantly affected the packing structure, hence reconfirming our desire to synthesize the pristine triazasumanene 5.
Crystal packing structure of (C)-23.
After completing data collection for the first paper on the synthesis of 22/23, Dr. Tan spent almost half of his time in his two-year stay at IMS on the synthesis of 5. It seemed only one step left to accomplish the synthesis was just to remove either the MeS or MeSO2 groups; however, this was extremely difficult because 22/23 was too strained a structure to induce many side reactions, including the decomposition or reduction of the bowl framework, and in addition the expected product 5 was very unstable. For example, typical desulfurization methods such as Raney Ni, Red-Al, etc. did not work at all. He finally found that a Liebeskind-type41 silane reduction42 mediated by Pd/Cu gave 5 (eq 7).
He also struggled to get good crystals suitable for X-ray analysis. Finally, he obtained very thin yellow needle crystals of enantiopure (C)-(−)-5 (which was considered to be good news because most of the sumanene derivatives with columnar packing were also obtained as similar needle forms), however, regrettably its structure could not be solved by the in-house machine. The structure was analyzed by my friends since my time at UW-Madison, Prof. Masaki Kawano and Dr. Yumi Yakiyama (later she joined our group at Osaka U.) using synchrotron X-ray analysis at PF-AR (NW2A beam line) of the High Energy Accelerator Research Organization (KEK). We were so excited to see the structure (Figure 11 left): there was also a columnar packing, but beyond our expectations, with a unique helical fashion induced by the inherent chirality of (C)-5. All the columns show a homogenous righthanded helix; one helical pitch including three molecules is 11.4 Å.27
Crystal packing structure of the homochiral (C)-5 (left) and the racemic 5 (right).
Just after obtaining the crystal structure, Dr. Tan completed his JSPS fellowship period and returned to China.43 Therefore, we tried to publish a paper using the above results, but it seemed too preliminary. Most of the referees were eager to compare the packing structure between the homochiral crystal and the racemate. Ms. Patcharin Keawmati (Ing), a Ph.D. student, continued this project. Since Ing did not have enough experience in organic synthesis, she spent a long time in training to reproduce such a long synthesis of 5, but finally she was able to prepare both enantiomers in significant amounts. At last, she got the crystals of the racemic 5 the old-fashioned way: she mixed an equal amount of each enantiomer by monitoring the disappearance of the CD signal of the solution, and then the crystals grown from the thus-obtained racemate solution showed different platelet shapes. The packing structure of the racemate 5 quite resembled that of the pristine sumanene 1 (Figure 11 right).27
3. Summary (Epilogue)
I have only described the synthesis of sumanene and triazasumanene here. It was true that the reason why I started this chemistry was simple curiosity, and I was fascinated with sumanene’s simple but beautiful structure. Furthermore, its single crystal structure and the columnar packing structure made me excited again. However, I believe that the most attractive point of sumanene is the property derived from its unique structure, in addition to the ready functionalization at the peripheral aromatic positions, the benzylic positions, and even at the internal carbon,44 which greatly increase the potential of the sumanene family as a unique organic material. Now sumanene is commercially available, so I expect many researchers to be engaged in sumanene chemistry.
Acknowledgment
I was able to continue this sumanene chemistry thanks to the efforts by many people. First of all, I would like to express appreciation for Emeritus Professor Koichi Narasaka, at the University of Tokyo, to enabled me to start this project and also encouraged me to continue at Osaka. I also thank Emeritus Professor Toshikazu Hirao who welcomed me to Osaka and studied this sumanene chemistry together with me. Thanks to the effort of Prof. Hirao, Prof. Amaya, and their team, sumanene chemistry has greatly expanded. Furthermore, I greatly appreciate IMS members, including the (late) Director-General Professor Koji Kaya, who recruited me as a PI even though there was no publication by me in this field.
Finally, all the works described in this account could not be achieved without the contribution of the co-authors, especially Prof. Shuhei Higashibayashi.
This work was first supported by Kinki Invention Center even before the success of the synthesis of sumanene, indeed I could not have continued this project without this grant. Then, it was also supported by KAKENHIs including two Grant-in-Aid for Scientific Research on Innovative Areas, π-Space project and π-System Figuration project, JST PRESTO, Tokuyama Foundation, Asahi Glass Foundation, Ishikawa Carbon Foundation, and Sumitomo Foundation.
References
a)
b)
e)
c)
a)
b)
b)
a)
b)
c)
d)
e)
a)
b)
c)
b)
c)
a)
b)
a)
b)
b)
a)
b)
c)
a)
b)
c)
Hidehiro Sakurai
Hidehiro Sakurai received his Ph.D. degree from The University of Tokyo in 1994 under the direction of Professor Koichi Narasaka. After studying with Professor Narasaka as an Assistant Professor in Tokyo, and with Professor C. P. Casey as a JSPS Postdoctoral fellow at the University of Wisconsin–Madison, he joined Osaka University in 2000. In 2004, he moved to the Institute for Molecular Science (IMS). In 2014, he returned to Osaka University as a Professor at Division of Applied Chemistry. His current research interests include the science of buckybowls, nonplanar π-aromatic compounds, and the development of nanocatalysts.
