Bimetallic Nickel Complexes for Selective CO 2 Carbon Capture and Sequestration

fast, selective, tight CO 2 Exploiting the affinity of the cavitand towards azides, CO 2 release Herein we report on the selective CO 2 uptake from air by a dinickel azacryptand complex and its capability to act as a reversible CO 2 storage system.

Herein we report a dinickel azacryptand complex that enables fast, selective, and tight CO 2 binding from air. Exploiting the affinity of the cavitand towards azides, CO 2 release was observed. Despite the stability of the azido complex, UV irradiation under atmospheric conditions proved to be a suitable pathway for N 3 replacement by CO 2 .
The continuous emission of carbon dioxide from combustion processes is a major reason for the increase of the global greenhouse effect. Thus CO 2 capture and sequestration technologies (CCS) are required to lower its concentration in our atmosphere and allow for storage and transportation. 1 In addition, reversible systems are needed that allow for its successive release and application as a C1 building block in industrial processes. In industry, aqueous ethanolic amine solutions are commonly applied for CCS but show high energy consumption and severe corrosion problems. 2 Furthermore, amino acids, 3 metal oxides, 4 and organic polymers 5 were evaluated for this purpose. Additionally, metalorganic frameworks show good CO 2 uptake capabilities, 6,7 and are also reported to allow for CO 2 reduction. [8][9][10] However, although promising, they do not allow for a selective, non-competitive guest absorption / uptake of CO 2 versus e.g. H 2 , N 2 , or CH 4 and increasing the CO 2 affinity results in difficult desorption. 11 Furthermore, water, a significant component of flue gases, was found to predominantly decrease CO 2 capacities. 12 In contrast, dinuclear metalloenzymes are known to selectively bind CO 2 and catalyze hydrolysis as well as redox reactions at ambient conditions. Among those, the nickel containing ureases 13  few systems exist that selectively bind CO 2 between two nickel centers. [18][19][20][21][22] General problems associated with such systems are either loose binding of the generated carbonate to nickel or formation of oligomers with bridging carbonates rendering the systems not applicable for reversible CCS and transportation. Contrary to those complexes, cryptands comprise well-defined metal binding cavities with tuneable metal-metal distances that allow for a strong binding of specific anions. Nelson and coworkers showed the uptake of carbonates into the cavity of dinickel azacryptands affording a stable dinickel complex with an anti-anti η 1 ,η 1 -bound carbonate (Scheme 1). 23 Furthermore, Sullivan and coworkers showed that in presence of H 2 , CO 2 can be reduced to afford CO and CH 4 . 24 However, neither CO 2 uptake kinetics nor selectivity studies were reported in order to apply azacryptands for CCS. Additionally, Lu as well as Malthouse and coworkers showed evidence for a direct atmospheric CO 2 uptake by dinuclear copper 25 and zinc cryptands. 26 The uptake capability of these bimetallic cryptands and of dinuclear metalloenzymes as well as the high stability of the obtained carbonate complexes within those frameworks pointed our interest towards the application of dinickel azacryptands as highly selective CO 2 binding and transportation platform under ambient conditions. Reaction of the azacryptands 1 or 2 with Ni(ClO 4 ) 2 comprising the non-coordinative perchlorate anion afforded complexes 3 and 4 under inert atmosphere as crystalline material in 56 % and 52 % yield, respectively (Scheme 2). The composition of both complexes Scheme 1. Carbonate uptake by heterobimetallic complexes.
was unequivocally confirmed by structure analysis (Figure 1, Figure  S1). Both complexes 3 and 4 are isostructural and reveal, although possessing two identical metal coordination sites, two nickel centers with different coordination modes. One nickel atom is coordinated by four amines and an additional chloride ligand. The chlorido ligand herein most likely originates from the reduction of a perchlorate catalyzed by a Ni-O species, 27 and multiple synthetic attempts with independently generated starting materials gave similar results. In contrast, the second nickel center is coordinated by four amines as well as an additional acetonitrile and a water molecule. The Ni-O distances of 2.044 Å (in 3) and 2.035 Å (in 4) are in good agreement with literature data for nickel-bound water and are significantly longer than Ni-O distances observed for Ni-OH moieties. 28 Charge balance is provided by three perchlorate anions.
To exclude nickel oxidation due to reductive perchlorate decomposition within the complexes we performed SQUID measurements on 4, which indicate two nickel centers with S = 1, consistent with two isolated Ni(II) centers ( Figure S2). Notably, whereas solutions of 4 were stable under atmospheric conditions, solutions of 3, kept in acetonitrile in air rapidly changed color ( Figure 2). The resulting species was thus isolated and revealed a significantly altered IR spectrum ( Figure S3). Crystals obtained by concentrating an acetonitrile solution of 3 under air revealed the molecular structure of the resulting compound 5 as depicted in Figure 1. It is noteworthy, the cavity of the azacryptand is occupied by a bicarbonate bridging the two nickel-centers with a Ni-O CO3 distance of 1.991 Å and 2.019 Å respectively. The Ni-Cl distance is significantly increased from 2.326 in 3 to 2.783 Å in 5 and the chloride is held in place by a hydrogen bond to the coordinated bicarbonate. As highlighted in Figure 1, an internal hydrogen bond to an attached water molecule further stabilizes the bicarbonate ion between the two nickel(II) centers. Although a magnetic communication between both nickel centers can be expected, SQUID measurements unequivocally revealed two isolated nickel(II) centers in complex 5 with S = 1.
In order to prove that the formation of 5 is the result of a selective uptake of CO 2 from air, we performed binding experiments under  both IR spectra did not allow for a clear assignment of the carbonate vibrations since the bicarbonate bands coincide with the azacryptand ligand vibrations. Mass spectrometry, however, confirmed the successful uptake of the respective isotopicallyenriched substrates ( Figure S4). Notably, no reaction of complex 3 with SO 2 , NO, N 2 O, NO 2 , nor CO was observed by UV-vis spectroscopy. The kinetics of the CO 2 fixation process was furthermore followed by stopped-flow spectrophotometry at various temperatures and applying different concentrated CO 2 solutions in acetonitrile obtained from saturated stock solutions ( Figure S5). 29 The reaction was monitored by the occurring absorbance changes at 470 nm. The reaction of 3 and CO 2 is of first-order in CO 2 concentration, yielding k obs , and the apparent association rate constant k 2, 298K = 0.067 ± 0.005 M -1 s -1 is obtained from the slope of the plot of k obs vs. CO 2 concentration ( Figure S5c). Only a small increase of k 2 with increasing temperature is observed and an apparent enthalpy of the activation of ΔH ‡ = 7.55 ± 1.07 kJ mol -1 is obtained from the Eyring plot ( Figure S6). This small value reflects that the association of CO 2 proceeds in more than one step. Therefore pre-equilibrium and activation enthalpies compose the apparent ΔH ‡ value.
Contrary to 3, an uptake of CO 2 into complex 4 was not possible. This fact demonstrates the different uptake capacities of both species. As no significant structural differences are observed at the metal centers in 3 vs. 4, we believe that the hydrophilic bicarbonate is repelled by the interaction with the hydrophobic tert-butyl group. Thus the CO 2 uptake can easily be adjusted by different bridging elements. Attempts to release the bicarbonate by purging solutions of 5 with nitrogen gas or heating up the same solutions did not result in the re-generation of 3 ( Figure S7). Azacryptands are well known to support the binding of different anions with varying binding strength. 30 (Figure S8), 32 or other, unidentifiable, decomposition products. However, when N 3 or SCN was allowed to react with 5, the formation of a single well-defined product was observed via UV-vis spectroscopy (Figure 2, Figure S9). The identity of the N 3 reaction product was elucidated as   Figure S10) and shows that the cavity is not intrinsically blocked by the tert-butyl groups. We wondered about the general nature of the released "CO 2species". We therefore treated 13 CO 2 -5 with the different substrates and investigated changes by 13 C NMR spectroscopy, measured in an inert tube in deuterated acetonitrile, and using gas chromatography. This treatment unequivocally showed the release of CO 2 as the only C1-species in all cases ( Figure S12). Similar reactivity was observed upon treatment of complex 5 with NaHCO 3 and provides evidence for a reversible CO 2 binding within the cryptand. Unfortunately, complex 6 did not allow for further ligand substitution reactions. It is, however, well known that Ni-N 3 complexes decompose under UV irradiation. 33 We therefore irradiated an acetonitrile solution of 6 with a 150 W Xe-lamp under air. This treatment revealed a change of the initial UV-vis bands and complete conversion was achieved within 2 hours ( Figure 2). Inspection of the UV-vis spectrum of the obtained compound showed close resemblance to the spectrum of complex 5 ( Figure 2). Scheme 3. CO2 binding and release utilizing dinickel azacryptands 5 and 6.
Furthermore, the obtained IR and MS spectra are in good agreement with the formation of compound 5 through UVirradiation of 6 ( Figure S13). Although we are at this point not able to provide an exact mechanism for the uptake of CO 2 , its expelling by N 3 and the subsequent cleavage of the resulting free azide, an overall pathway is presented in Scheme 3.
In conclusion, we have shown for the first time that nickelazacryptands are superior compounds for highly selective CO 2 capture under ambient conditions. Due to their defined CO 2 -hosting cavity, typical problems associated with gas selectivity and low binding strength in the presence of moisture can be avoided as compound 5 is thermodynamically stable in the presence of excess SO 2 , NO, N 2 O, NO 2 , CO, or CS 2 . As shown with complex 4, the CO 2 binding affinity can be easily adjusted by manipulation of the bridging elements connecting both coordination cavities. Complex 5 showed high stability even at higher temperatures. We were able to show a selective release of CO 2 upon treatment with pseudohalogenides. Especially interesting is the reaction of 5 with NaN 3 to give 6. This reaction proved to be very useful since irradiation with a standard UV Xe-lamp emptied the cryptands' cavity and again allowed for subsequent binding of CO 2 .
The well-defined cavity of azacryptands provides a highly sophisticated system for carbon capture and sequestration in a reversible manner. In addition, it shows that azacryptands can serve as molecular reactors allowing specific reactions to proceed in an elaborate, protected environment. Further investigations concerning the mechanism of N 3 uptake, photocleavage, as well as manipulation of the substitution kinetics and reduction experiments are currently under way.