SPECIAL FEATURE: PERSPECTIVE Dihydrogen complexes as prototypes for the coordination chemistry of saturated molecules Gregory J. Kubas† Chemistry Division, Los Alamos National Laboratory, Los Alamos, NM 87545 Edited by Jay A. Labinger, California Institute of Technology, Pasadena, CA, and accepted by the Editorial Board February 26, 2007 (received for review November 14, 2006) The binding of a dihydrogen molecule (H2) to a transition metal center in an organometallic complex was a major discovery because it changed the way chemists think about the reactivity of molecules with chemically ‘‘inert’’ strong bonds such as H–H and C–H. Be- fore the seminal finding of side-on bonded H2 in W(CO)3(PR3)2(H2), it was generally believed that H2 could not bind to another atom in stable fashion and would split into two separate H atoms to form a metal dihydride before undergoing chemical reaction. Metal- bound saturated molecules such as H2, silanes, and alkanes (␴-complexes) have a chemistry of their own, with surprisingly varied structures, bonding, and dynamics. H2 complexes are of increased relevance for H2 production and storage in the hydrogen economy of the future. ihydrogen (H2) and hydrocar- bons are vital in chemical processes such as hydrogena- tion and conversions of Dorganic compounds. Catalytic hydro- genations are the largest-volume chemi- cal reactions: all crude oil is treated Sketch 1. Eq. 1. with H2 to remove sulfur/nitrogen, and Ͼ100 million tons of ammonia fertilizer buckminsterfullerene (C ) (6). That its are produced annually to support much 60 existence remained hidden for so long cyclohexyl) (9). Its unusual purple color of the world’s population. The H mole- 2 adds to the lore, and our unexpected changed instantly and reversibly to yel- cule is married together by a very strong revelation of metal–H2 complexes has low on exposure to N2 and H2 in both two-electron H–H bond but is only use- some commonality. Metal dihydrides ful chemically when the two H atoms solution and solid states, signifying ad- formed by oxidative addition (OA) of duct formation (Eq. 1). divorce in controlled fashion. This also the H–H bond to a metal center had applies to other strong ␴-bonds such as Crystallography later revealed a phos- been known early on to be a part of phine C–H bond weakly occupying the C–H in alkanes. However, the mecha- catalytic cycles (7), as documented in sixth binding site in W(CO)3(PCy3)2 nism at the molecular level by which the a 1980 retrospective on catalytic hydro- union splits was established only rela- (10). This type of ‘‘agostic’’ interaction genation by a pioneer in the field, Jack (11) relieves electronic unsaturation in tively recently because such electroni- Halpern (8). Although some form of coordinatively unsaturated complexes cally saturated molecules were never metal–H interaction was assumed to 2 and is entropically favorable because it caught in the act of chemically binding participate in dihydride formation, it to a metal or other ‘‘third party,’’ usu- was thought to be unobservable. We is ‘‘intramolecular.’’ ‘‘Intermolecular’’ ally the first step in breaking apart a were fortunate to observe it in the com- binding of a C–H bond as in an alkane strong bond. The discovery by Kubas ␴-complex (often also termed ‘‘agostic’’) plex W(CO)3(PR3)2(H2), as detailed by and coworkers (1) in 1984 of coordina- this author (2, 3). This was the first mo- is less stable. Irrefutable evidence for tion of a nearly intact H2 molecule to a lecular compound synthesized and iso- H2 binding in Eq. 1 came slowly because ϭ metal complex (LnM; L ligand) lated entirely under ambient conditions pinpointing H positions crystallographi- caught this in intimate detail and led to that contained the H2 molecule (albeit cally is difficult, even by neutron diffrac- a new paradigm in chemistry (1–4) (see ‘‘stretched’’) other than elemental H2 tion. A consultant, Russ Drago, sug- Sketch 1). itself. The H–H bond length in gested an experiment elegant in its ␩2 The H2 binds side-on ( ) to M pri- W(CO) (PiPr ) (H ) (0.89 Å) is simplicity: synthesize the HD complex ␴ 3 3 2 2 marily through donation of its two stretched Ϸ20% over that in H2 (0.74 and look for a large HD coupling con- electrons to a vacant metal orbital to Å), showing that H2 is not physisorbed stant in the proton NMR that would form a stable H2 complex. It is remark- but rather chemisorbed, where the bond show that the H–D bond was mostly able that the electrons already strongly is ‘‘activated’’ toward rupture. Like H2, intact. It worked beautifully: the 1H bonded can donate to a metal to form a i other saturated molecules such as al- NMR of W(CO)3(P Pr3)2(HD) showed a nonclassical 2-electron, 3-center bond, kanes were thought to be inert to such 1:1:1 triplet (deuterium spin ϭ 1) with as in other ‘‘electron-deficient’’ mole- binding, although their C–H bonds JHD ϭ 33.5 Hz, nearly as high as in HD cules such as diborane (B2H6). M–H2 somehow could also be broken on met- ␴ gas, 43.2 Hz. Observation of JHD higher and other ‘‘ -complexes’’ (3, 5), encom- als. The ‘‘somehow’’ is why the finding Ͼ ␴ than that for a dihydride complex ( 2 passing interaction of any -bond (C–H, of an H2 complex was important: it is Si–H, etc) with a metal center, are the the prototype for activation of all major theme of this special feature. ␴-bonds. Author contributions: G.J.K. wrote the paper. i This discovery of W(CO)3(P Pr3)2(H2) The author declares no conflict of interest. Introduction and Historical Perspective ensued the serendipitous synthesis of its This article is a PNAS Direct Submission. J.A.L. is a guest Certain discoveries and how they came novel, ‘‘unsaturated’’ 16-e precursor, editor invited by the Editorial Board. about are fascinating sagas, e.g., that for M(CO)3(PCy3)2 (M ϭ Mo, W; Cy ϭ †E-mail: [email protected]. www.pnas.org͞cgi͞doi͞10.1073͞pnas.0609707104 PNAS ͉ April 24, 2007 ͉ vol. 104 ͉ no. 17 ͉ 6901–6907 Downloaded by guest on September 25, 2021 Eq. 2. Scheme 1. Hz) became the premier criterion for an and Heinekey. This quartet has since H2 complex. performed elegant synthetic, reactivity, more characteristic of a dihydride, which One reason that H2 complexes were and NMR studies on H2 and silane it was initially believed to be (47). Com- so well hidden was the notion that such complexes (5, 25–30) and was eventually plexes containing only H2O (48) or CO complexes could not be stable relative joined by Ͼ100 investigators worldwide. (17, 18) coligands are also known, but to classical dihydrides, as exemplified by Remarkably, several complexes initially are marginally stable (Scheme 1). the controversy over our initial findings. believed to be hydrides were revealed to Determining the presence of a H2 ligand and its d is nontrivial because This paralleled the discovery of metal– be H2 complexes by Crabtree and Ham- HH dinitrogen complexes by Allen and Se- ilton in 1986 (5, 31), by using as criteria even neutron diffraction has limited ap- noff, whose seminal paper was initially the short proton NMR relaxation times plicability and can give foreshortened rejected (12). At the time of our finding, of H ligands (T Ͻ 100 msec). Particu- dHH because of rapid H2 rotation/libra- 2 1 1 tion (49). JHD is the best criterion, and spectroscopic evidence for unstable larly interesting was RuH2(H2)(PPh3)3 values determined in solution correlate M–H2 interactions was found by photol- first reported in 1968 (32); it possessed well with dHH in the solid state through ysis of Cr(CO)6 in the presence of H2 at unusual H2 lability that Singleton in Eqs. 3 and 4 (50, 51). low T (13–16). Cr(CO)5(H2) was postu- 1976 commented was characteristic of lated based on IR CO stretching fre- ‘‘H2-like bonding’’ (33). However, at- ϭ Ϫ ͓ ͔ dHH 1.42 0.0167JHD Å Morris quencies but could not be conclusively tempts to prove H2 binding here was 1 demonstrated; only recently has its H problematic, even long after H2 binding [3] NMR spectrum been observed at low T was established (34).‡ ϭ (17, 18). Even theoretical bases for in- More than 600 H complexes are dHH 1.44 ␴ 2 teraction of H2 and other -bonds with known (most of them stable) for nearly Ϫ 0.0168J Å ͓Heinekey͔. a metal was still in its infancy at the every transition metal and type of coli- HD time of our discovery. Ironically, a com- gand and are the focus of 1,500 publica- [4] putational paper by Saillard and Hoff- tions, dozens of reviews, and three mann (19) in 1984 on the bonding of H2 monographs (2–5, 20, 25–30, 35–43). Data include dHH from crystallography and CH4 to metal fragments such as The view on H2 complexes has shifted and also solid-state NMR measurements Cr(CO)5 was published shortly after our from significance in basic science to a by Zilm and Millar (52) that gave the publication (1) of the W–H2 complex, more practical bent, e.g., H2 fuel pro- most accurate dHH (direct measure of without mutual knowledge. Such inter- duction and storage. Two frequent internuclear HH separation). For i ϭ play between theory and experiment has questions after their discovery were as W(CO)3(P Pr3)2(H2), JHD 34 Hz, giv- ϭ continued as one of the most valuable follows.