Matthias Müller1, Magnus R Buchner1. 1. Anorganische Chemie, Nachwuchsgruppe Berylliumchemie, Fachbereich Chemie, Philipps-Universität Marburg, Hans-Meerwein-Straße 4, 35032, Marburg, Germany.
Abstract
Aldehydes play a key role in the human metabolism. Therefore, it is essential to know their reactivity with beryllium compounds in order to assess its effects in the body. The reactivity of simple aldehydes towards beryllium halides (F, Cl, Br, I) was studied through solution and solid-state techniques and revealed distinctively different reactivities of the beryllium halides, with BeF2 being the least and BeI2 the most reactive. Rearrangement and aldol condensation reactions were observed and monitored by in situ NMR spectroscopy. Crystal structures of various compounds obtained by Be2+ -catalyzed cyclization, rearrangement, and aldol addition reactions or ligation of beryllium halides have been determined, including unprecedented one-dimensional BeCl2 chains and the first structurally characterized example of an 1-iodo-alkoxide. Long-term studies showed that only aldehydes without a β-H can form stable beryllium complexes, whereas other aldehydes are oligo- and polymerized or decomposed by beryllium halides.
Aldehydes play a key role in the human metabolism. Therefore, it is essential to know their reactivity with beryllium compounds in order to assess its effects in the body. The reactivity of simple aldehydes towards beryllium halides (F, Cl, Br, I) was studied through solution and solid-state techniques and revealed distinctively different reactivities of the beryllium halides, with BeF2 being the least and BeI2 the most reactive. Rearrangement and aldol condensation reactions were observed and monitored by in situ NMR spectroscopy. Crystal structures of various compounds obtained by Be2+ -catalyzed cyclization, rearrangement, and aldol addition reactions or ligation of beryllium halides have been determined, including unprecedented one-dimensional BeCl2 chains and the first structurally characterized example of an 1-iodo-alkoxide. Long-term studies showed that only aldehydes without a β-H can form stable beryllium complexes, whereas other aldehydes are oligo- and polymerized or decomposed by beryllium halides.
It is known that hard Lewis acids are able to catalyze the conversion of glucose into 5‐(hydroxymethyl)furfural and subsequently into humins.1, 2, 3, 4 Hence, inhaled beryllium species should perform in a similar way in lung tissue, in which hexoses are abundant.5 To what extend the Be2+ cation is able to catalyze the conversion of carbonyle groups is therefore a subject of immense interest with regards to the alleged toxicity of beryllium in general and especially to understand the localization of beryllium‐associated diseases to the lung.6, 7Lewis acids—especially aluminum chloride species—promote a wide variety of reactivities in aldehydes and are important for C−C bond formation, such as the carbonyl en reaction.8, 9, 10 Furthermore, Lewis acidic metal chlorides are able to induce aldol condensation11, 12, 13 and rearrangement reactions14 or cyclization into trioxanes.11, 12, 13 The latter is also well catalyzed by metal oxides or Me3SiCl.15, 16, 17 Among the various Lewis acids, the activity of which towards aldehydes was examined, beryllium species, which are among the hardest known Lewis acids, have never been investigated. Only one beryllium aldehyde compound—a benzaldehyde adduct towards BeCl2 of which no reactivity had been investigated—was known so far.18 The formation of BeCl2 adducts towards carbonyl groups is also observed in esters18, 19 and ketones,20 but to the best of our knowledge no experimental studies regarding a potentially altered reactivity of beryllium‐coordinated carbonyl groups had been performed so far. Here, we present our findings regarding the reactivity of various simple aldehydes towards beryllium halides.
Results and Discussion
All our experiments with BeF2, which is in theory the hardest Lewis acid of the investigated beryllium halides, showed that it does neither react with the investigated aldehydes nor dissolve in them or any of the used solvents. This is presumably caused by the significantly higher lattice energy of BeF2 relative to those of the other beryllium halides21 and is in line with the fact that only reactions of BeF2 in water or liquid ammonia are known.22, 23Contrarily, BeCl2 readily reacts in a wide variety of organic solvents. We started our investigations with pivaldehyde that has neither a proton in α position nor the direct ability for an aldol condensation reaction and was assumed to be the least reactive aliphatic aldehyde. Therefore, a reaction of BeCl2 was performed with pivaldehyde in various stoichiometries on NMR scale. A brief summary of the obtained results is illustrated in Scheme 1. When few milligrams of BeCl2 were dissolved in several milliliters of pivaldehyde the reaction mixture solidified within hours. This colorless solid can be sublimated in vacuo to yield centimeter‐long single crystals of 2,4,6‐tris‐tert‐butyl‐1,3,5‐trioxane (1 a, Figure 1) that crystallized in the monoclinic space group C2/c. This cyclization reaction is also known to be catalyzed by Lewis acids such as Me3SiCl, FeCl3, or metal oxides. BeCl2 as a Lewis acid is therefore no exception. Subsequent reactions of pivaldehyde with catalytic amounts of BeBr2 and BeI2 led also to the formation of trioxane 1 a, which was analyzed by X‐ray diffraction, IR and NMR spectroscopy. The latter revealed its instability in polar solvents. Although stable in C6D6, it decomposes in CDCl3 to pivaldehyde. The same decomposition occurs when beryllium halides were added in stoichiometric amounts. Therefore, we used 1 a instead of pivaldehyde as an aldehyde precursor for the following experiments because of its supreme purity and ease of use compared to the free aldehyde.
Scheme 1
Schematic illustration of the reaction pathways for the formation of compounds 1 a, 2 a,b, 3 a,b, 4 a,b, and 5 a,b depending on the amount of pivaldehyde (a: X=Cl; b: X=Br; MIPK=methyl isopropyl ketone).
Figure 1
Molecular structures of compounds 1 a, 2 a, 5 a, and 3 a (from top to bottom). Ellipsoids are depicted at 70 % probability at 100 K. Hydrogen atoms are omitted and carbon atoms are shown as wire frame for clarity (see Supporting Information Figures 1, 2, 4, 5 for molecular structures with all non‐hydrogen atoms represented as ellipsoids). Subfigures are not to scale with each other.
Schematic illustration of the reaction pathways for the formation of compounds 1 a, 2 a,b, 3 a,b, 4 a,b, and 5 a,b depending on the amount of pivaldehyde (a: X=Cl; b: X=Br; MIPK=methyl isopropyl ketone).Molecular structures of compounds 1 a, 2 a, 5 a, and 3 a (from top to bottom). Ellipsoids are depicted at 70 % probability at 100 K. Hydrogen atoms are omitted and carbon atoms are shown as wire frame for clarity (see Supporting Information Figures 1, 2, 4, 5 for molecular structures with all non‐hydrogen atoms represented as ellipsoids). Subfigures are not to scale with each other.If BeCl2 was treated with two or more equivalents of pivaldehyde in CDCl3 a clear colorless solution was obtained. 1H and 13C NMR spectroscopic experiments revealed that the pivaldehyde moiety was stable for several weeks, whereas the different chemical shifts of the signals, in comparison to the free ligand, indicated a coordination towards BeCl2 (see also Table 1). In the 9Be NMR spectrum we observed one sharp singlet at 5.8 ppm (ω1/2=4.9 Hz) that is typical for tetrahedral coordinated Be nuclei.24 By removing the solvent, crystals of dichlorobis(2,2‐dimethylpropanal)beryllium(II) (2 a) could be obtained as colorless blocks. Compound 2 a crystallizes in the monoclinic space group P21/c. In this compound BeCl2 is coordinated by two pivaldehyde molecules leading to a tetrahedral coordination sphere around the beryllium atom, which is a typical structural motive for BeCl2 coordinated by electron‐donating functional groups.18, 19, 25 The molecular structure of 2 a is illustrated in Figure 1. The Be−Cl atomic distances of 1.9595(18) and 1.9717(17) Å correspond to those in the BeCl2 benzaldehyde adduct with 1.9623(19) and 1.982(2) Å. The Be−O atomic distances in 2 a of 1.6819(18) and 1.6925(19) Å are, however, significantly longer than those in the benzaldehyde complex [1.664(2) and 1.665(2) Å].18 This might be caused by the larger steric bulk of the pivaldehyde. In the IR spectrum of 2 a we observed a bathochromic shift for the C=O‐stretching frequency of 61 cm−1 relative to that of non‐coordinated pivaldehyde, indicating a weakening of the C=O double bond because of the coordination towards BeCl2. This is in agreement with our findings on the benzaldehyde adduct.18
Table 1
Selected chemical shifts and coordination shifts of the investigated compounds. Δδ(1H): shift of aldehyde proton (CH(O)) relative to that of free ligand, δ(9Be): chemical shift of Be nucleus, Δδ(13C): shift of aldehyde carbon atom (CH(O)) relative to that of free ligand.
Δδ(1H) [ppm]
δ(9Be) [ppm]
Δδ(13C) [ppm]
2 a
0.11
5.8
17.2
2 b
0.30
5.4
18.6
2 c
0.47
2.7
20.0
3 a
0.25
5.0
17.8
3 b
0.33
2.6
19.2
4 a
–
4.3
22.1
4 b
–
2.7
20.9
4 c
–
−1.0
20.8
5 a
–
4.7
18.8
5 b
–
3.4
20.1
5 c
–
2.7
24.1
6
0.47
4.1
20.0
11
–
3.0
–
14
0
5.3
9.6
Selected chemical shifts and coordination shifts of the investigated compounds. Δδ(1H): shift of aldehyde proton (CH(O)) relative to that of free ligand, δ(9Be): chemical shift of Be nucleus, Δδ(13C): shift of aldehyde carbon atom (CH(O)) relative to that of free ligand.When BeCl2 was treated with only one equivalent of pivaldehyde we also obtained a clear and colorless solution. The in situ NMR spectra recorded a few hours after the reaction was initiated, showed the same set of peaks as that of compound 2 a though with different chemical shifts. A broad singlet in the 9Be NMR spectrum at 5.0 ppm (ω1/2=21.1 Hz) speaks for an increase in electron density around the beryllium nucleus relative to that in 2 a, whereas the significantly broader line width is indicative of a more fluxional compound in solution. When the solvent of a solution with 1:1 ratio of BeCl2/ pivaldehyde was removed in vacuo immediately after the starting materials had dissolved, single crystals of poly[dichloro‐μ2‐2,2‐dimethylpropanalberyllium(II)]⋅C6H6 (3 a) were obtained from a colorless oil. A detail of its crystal structure is illustrated in Figure 1. It features one‐dimensional μ2‐Cl linked BeCl2 chains coordinated by one pivaldehyde per beryllium atom. The Be−Cl atomic distances of the terminal chloride atoms are 1.946(4) and 1.952(4) Å, comparable with those of compound 2 a, whereas the μ2‐Cl−Be atomic distances range from 2.056(4) to 2.066(4) Å and are therefore significantly longer. These distances are comparable to those reported for [Be2Cl2(μ‐Cl)2(PCy3)2].26 The Be−O atomic distances are 1.643(4) and 1.646(4) Å; significantly shorter than the corresponding distances for the aldehyde adducts mentioned above. This stronger interaction is expected because of the coordination of only one aldehyde towards this BeCl2 chain, which results in less steric repulsion. This is also supported by the larger coordination shifts observed in the 1H and 13C NMR spectra, and the broad line widths of the NMR signals suggest that various oligomers are present in solution that exchange relatively rapidly on the NMR time scale. However, in the IR spectra of the dry residue of the reaction mixture of 3 a we observed a set of bands nearly identical to that of 2 a, and the C=O‐stretching frequency of 1662 cm−1 equals that of compound 2 a.When the solution with a 1:1 ratio of BeCl2/pivaldehyde was stored for several weeks, the NMR spectra changed (see also Supporting Information Figures 10–12). The former singlet in the 9Be NMR spectrum at 5.0 ppm shifted within six weeks to 4.3 ppm (ω1/2=13.9 Hz) indicating the formation of a new, less fluxional species with increased electron density at the beryllium nucleus. We also observed a new signal set in the 1H and 13C NMR spectra, which could be assigned to 3‐methyl‐2‐butanone (methyl isopropyl ketone, MIPK) coordinated to BeCl2 (4 a). To verify this we treated BeCl2 with one equivalent of MIPK and obtained matching NMR spectra. To figure out if MIPK can be converted to pivaldehyde by BeCl2 we monitored a 1:1 reaction mixture of MIPK and BeCl2 in CDCl3 over time. However, even after several weeks we did not observe any change in the NMR spectra, which leads to the conclusion that the rearrangement reaction is nonreversible in the presence of BeCl2. In an additional control experiment BeCl2 was treated with two equivalents of MIPK. In the 1H and 13C NMR spectra we observed only signals for MIPK and even several weeks later no change was observed. In the 13C NMR spectra we noticed a downfield shift of 18.8 ppm for the carbonyl carbon atom relative to that of the free butanone, which is similar to the coordination shift of compound 2 a relative to that of free pivaldehyde (see Table 1). This indicates a depletion of electron density at the carbonyl group. In the 9Be NMR spectrum one singlet is observed at 4.7 ppm (ω1/2=8.0 Hz) and therefore shifted downfield in relation to the signal of compound 4 a. This shift difference is comparable to the one observed between compounds 2 a and 3 a and indicates that the electron density at the beryllium nuclei is higher in complexes in which only one carbonyloxygen atom is coordinated. By removing the solvent in vacuo, we obtained rod‐shaped crystals of dichloridobis(3‐methyl‐2‐butanone)beryllium(II) (5 a). This butanone adduct of BeCl2 crystallizes in the monoclinic space group C2/c and shows the same coordination around the Be atom as that described for 2 a with Be−Cl atomic distances of 1.983(3) and 1.987(3) Å, which are slightly longer than those of 2 a. On the other hand, the Be−O atomic distances of 1.659(3) and 1.669(3) Å are slightly shorter than the corresponding distances in 2 a suggesting that MIPK coordinates stronger to BeCl2 than pivaldehyde. This is in agreement with the upfield shift of the 9Be NMR signal. However, in the IR spectrum of 5 a the bathochromic shift for the C=O‐stretching frequency of 61 cm−1 is identical to that of 2 a.If BeBr2 was treated with two or more equivalents of pivaldehyde a clear brownish solution was obtained. The 1H and 13C NMR spectra recorded initially showed a signal set for pivaldehyde shifted downfield from that of free pivaldehyde in a scale comparable to that of 2 a. In the 9Be NMR spectrum we observed a singlet at 5.4 ppm (ω1/2=3.0 Hz). After removal of the solvent in vacuo we obtained colorless needle‐shaped crystals of dibromobis(2,2‐dimethylpropanal)beryllium(II) (2 b, Supporting Information, Figure 2) from an oily residue. Compound 2 b crystallizes isotypic to 2 a in the monoclinic space group P21/c. The Be−O atomic distances are 1.653(5) and 1.657(6) Å and therefore shorter than those in 2 a but within the standard deviation identical to those in 3 a. This indicates a stronger interaction between pivaldehyde and BeBr2 relative to BeCl2 and is supported by the more pronounced NMR coordination shifts of the aldehyde proton and the carbonyl carbon nuclei in 2 b relative to 2 a. The Be−Br atomic distances are 2.118(5) and 2.141(5) Å, comparable to those in the acetonitrile compound [BeBr2(CH3CN)2] [2.146(5); 2.147(5) Å]27 but shorter than those in adduct [BeBr2(Et2O)2] [2.170(8); 2.176(8) Å].28 The IR spectrum of 2 b is similar to that of 2 a with a comparable coordination shift of the C=O‐stretching frequency of 66 cm−1.When BeBr2 was treated with only one equivalent of pivaldehyde we also observed the initial formation of the mono‐aldehyde adduct to BeBr2 (3 b) as evident from the downfield coordination shifts of the aldehyde proton and carbonyl carbon signals in the 1H and 13C NMR spectra, which are comparable to those of 3 a. Compound 3 b gives rise to a 9Be NMR signal at 2.6 ppm (ω1/2=13.5 Hz). After six weeks we also observed the quantitative conversion of pivaldehyde to MIPK as well as the emergence of a new singlet at 2.7 ppm (ω1/2=8.7 Hz) in the 9Be NMR spectrum, leading to the conclusion that the same rearrangement reaction occurs like that described for BeCl2. This was also verified by the reaction of one equivalent of BeBr2 with one equivalent of MIPK, which resulted in matching NMR spectra. Reaction of BeBr2 with two equivalents of MIPK in CDCl3 led, in analogy to BeCl2, to the formation of the bis‐ketone adduct 5 b with a 9Be NMR signal at 3.4 ppm (ω1/2=4.1 Hz). Solutions of complex 5 b proved to be stable for over six weeks such as those of 5 a. However, it was neither possible to obtain single crystals of the butanone adduct 5 b nor of the mono‐pivaldehyde adduct 3 b similar to 3 a. In both cases a brownish oil was obtained.Finally, we investigated how BeI2 interacts with pivaldehyde and trioxane 1 a. Although it behaved like BeCl2 and BeBr2 when applied in catalytic amounts, it showed a distinctively different behavior when applied in equimolar amounts. When BeI2 was treated with three equivalents of pivaldehyde in CDCl3 we obtained a brownish solution that darkened within two weeks to give a brown opaque solution. The 1H and 13C NMR spectra of a freshly prepared solution showed primarily signals for pivaldehyde coordinating towards beryllium iodide (Figure S18). Therefore, we assume that the formation of diiodobis(2,2‐dimethylpropanal)beryllium(II) (2 c) analogous to 2 a and 2 b occurred. However, the 1H and 13C NMR spectra recorded directly after the reagents were dissolved already showed additional signals to those assigned to 2 c. In the 9Be NMR spectrum three signals at 4.1, 2.7, and 4.9 ppm were observed. On the basis of the assignment of the signals from 2 a and 2 b the singlet at 2.7 ppm (ω1/2=36.3 Hz) was assigned to 2 c. Over the course of a week, the intensity of the 9Be NMR signal at 4.1 ppm increased while the other two declined accordingly (Figure S19). In the 1H and 13C NMR spectra the intensity of the signals of pivaldehyde also decreased and a variety of new signals could be observed. After CDCl3 was removed in vacuo a brown oil was obtained that contained rod‐shaped single crystals of bis(2,2‐dimethylpropanal)bis(μ2‐[1‐iodo‐2,2‐dimethylpropan‐1‐ol])diiododiberyllium(II) (6, Figure 2), which crystallizes in the orthorhombic space group Pna21. This is the compound from which the 9Be NMR signal at 4.1 ppm originates. Compound 6 features a four‐membered ring of two beryllium atoms that are linked through two μ2‐oxygen atoms of two corresponding 1‐iodo‐2,2‐dimethyl‐propan‐1‐olate (7) molecules, which is, to the best of our knowledge, the first structurally characterized example of an 1‐iodo‐alkoxide. Each beryllium atom is also coordinated by an iodine atom and an oxygen atom of a pivaldehyde molecule to complete the tetrahedral coordination sphere around the beryllium atom. The Be−O atomic distances inside the ring in 6 range from 1.56(4) to 1.70(4) Å and those to the pivaldehyde molecule are 1.63(4) and 1.68(4) Å, which are within the standard deviation identical. The Be−O atomic distances between the beryllium atoms and the pivaldehyde molecules are within the standard deviation also identical to those of the adducts 2 a and 2 b because of the low precision of the data obtained. The Be−I atomic distances are 2.30(3) and 2.36(4) Å, thus comparable to the Be−I atomic distances found in BeI2 that range from 2.403(1) to 2.446(1) Å.29 A few crystals were separated from the oil and an IR spectrum was recorded subsequently. The carbonyl stretching frequency of 1664 cm−1 is similar to that of compounds 2 a and 2 b. The formation of 1‐iodo‐2‐methyl‐propan‐1‐ol could additionally be verified by NMR spectroscopy. In the reaction mixture of 6 a singlet at 6.71 ppm in the 1H NMR spectrum was assigned to the tBu‐CHOI proton because of its 3
J
CH coupling to a tert‐butyl group observed in the HMBC NMR spectrum and no coupling in the COSY NMR spectrum as expected. The singlet at 1.08 ppm in the 1H NMR spectrum that is also present in the 1H NMR spectrum recorded initially was assigned to the tert‐butyl group. This means that the conversion of the pivaldehyde to 1‐iodo‐2‐methyl‐propan‐1‐ol takes place rapidly. However, we did not observe complete conversion even after several weeks. A plausible reaction pathway for this conversion is illustrated in Scheme 2; higher pivaldehyde concentrations also led to the formation of 1‐iodo‐2‐methyl‐propan‐1‐ol. After coordination of two aldehydes one iodo ligand is displaced by a third aldehyde because of the weak Be−I bond generating the cationic tris‐aldehyde complex 8.30 Because of the increased Lewis acidity of the cationic complex subsequent nucleophilic attack of iodide at one of the coordinated aldehydes is then facilitated. This leads to the formation of alcoholate 9. After dissociation of one aldehyde, compound 6 is formed, which is presumably due to the superior coordination abilities of the negatively charged alcoholate compared to those of the aldehyde. We therefore assume that the signal at 4.9 ppm in the 9Be NMR spectrum, which was observed in the beginning of the reaction, is caused by cationic species 8.
Figure 2
Molecular structures of compounds 6 (bottom) and 11 (top), including their formula. Ellipsoids are depicted at 70 % probability at 100 K. Hydrogen atoms are omitted and carbon atoms are shown as wire frame for clarity (see Supporting Information Figures 6 and 7 for molecular structures with all non‐hydrogen atoms represented as ellipsoids). Subfigures are not to scale with each other.
Scheme 2
Schematic illustration of the reaction pathway for the formation of compound 6.
Molecular structures of compounds 6 (bottom) and 11 (top), including their formula. Ellipsoids are depicted at 70 % probability at 100 K. Hydrogen atoms are omitted and carbon atoms are shown as wire frame for clarity (see Supporting Information Figures 6 and 7 for molecular structures with all non‐hydrogen atoms represented as ellipsoids). Subfigures are not to scale with each other.Schematic illustration of the reaction pathway for the formation of compound 6.Furthermore, reaction mixtures of BeI2 with two or less equivalents of pivaldehyde showed a completely different reactivity, which is summarized in Scheme 3. When a reaction of pivaldehyde with BeI2 was performed in a one‐to‐one ratio, one signal set for beryllium‐coordinated pivaldehyde was observed by 1H and 13C NMR spectroscopy, indicating the formation of mono‐aldehyde adduct 3 c. However, three broad singlets at 4.9, 4.1, and 2.7 ppm, which were also present in the three‐to‐one reaction mixture and belong to compounds 8, 6, and 2 c, respectively, were observed by 9Be NMR spectroscopy. This was presumably caused through an initial excess of aldehyde that was due to undissolved BeI2. While the signal at 4.9 and 4.1 ppm disappeared within one week and the one at 2.7 ppm decreased, two additional signals appeared at 0.9 and −1.0 ppm. During this reaction an additional organic intermediate was formed, which was identified by 1H and 13C NMR spectroscopy as well as mass spectrometry as 5‐hydroxy‐2,6,6‐trimethylheptan‐3‐one (10). This is the aldol addition product of pivaldehyde and MIPK. After two weeks, only the signal at −1.0 ppm (ω1/2=3.7 Hz) remained and only one signal set for MIPK was observable in the 1H and 13C NMR spectra with no pivaldehyde left; therefore, poly[diiodo‐μ2‐3‐methyl‐2‐butanoneberyllium] (4 c) was formed (see also Supporting Information, Figures 20–22). To confirm this assumption we performed a reaction of BeI2 with MIPK in a one‐to‐one ratio in CDCl3 and obtained the same signals in the 1H, 13C, and 9Be NMR spectra. Thus, BeI2 shows a higher potential in promoting the rearrangement reaction of pivaldehyde into MIPK compared to BeCl2 and BeBr2. However, no single crystals of 4 c were obtained. Compound 10 together with MIPK was also obtained when two equivalents of pivaldehyde were treated with one equivalent of BeI2 in CDCl3. In situ 9Be NMR reaction monitoring revealed the formation of 8 and 2 c within the first hour of reaction because of an initial excess of aldehyde caused by undissolved BeI2 (see also Supporting Information, Figures 23–25). After four weeks one major signal at 3.0 ppm, which we assign to a beryllium species (11) coordinated to 10, together with three smaller signals at 4.1 (6), 2.7 (5 c), and 0.9 ppm were observed. The signal assignments to 5 c (2.7 ppm, ω1/2=7.3 Hz) were confirmed through the reaction of BeI2 with two equivalents of MIPK. Additionally, HI was observed in the 1H NMR spectra evident from a sharp singlet at −11.33 ppm. We assumed that the 9Be NMR signals at 0.9 ppm arose from the presence of BeI2 coordinated to one aldehyde and one ketone 12 c and that this species was responsible for the condensation reaction. Therefore, we performed a reaction with one equivalent of pivaldehyde, MIPK, and BeI2 each in CDCl3 to promote the aldol addition reaction (see also Supporting Information, Figures 26–29). The recorded NMR spectra of the obtained brownish solution after four weeks of storage at ambient temperature showed that 10 was obtained as main organic moiety. After six weeks of storage colorless cubic crystals swam on top of the solution. The X‐ray structure analysis revealed bis(μ2‐[5‐hydroxy‐2,6,6‐trimethylheptan‐3‐one])diiododiberyllium(II) (11) as the main complex in solution. Compound 11 crystallizes in the triclinic space group P
and features a four‐membered Be−O ring similar to that in compound 6. Like in 6 the beryllium atoms are coordinated by one iodine atom and three oxygen atoms each. The hydroxy group acts as μ2‐link while the oxygen atom of the carbonyl group completes the tetrahedral coordination around the Be atom (Figure 2). The Be−O atomic distances within the four‐membered ring are 1.591(4) and 1.616(4) Å, which is comparable to the values found in 6 and those reported for tetrakis(μ2‐t‐butoxy)‐dichloro‐tri‐beryllium [from 1.540(2) to 1.635(2) Å].31 The Be−O atomic distance to the carbonyl group is 1.655(4) Å, thus longer than that reported for Be(acac)2 [1.62(1) Å].20 The Be−I atomic distance is 2.390(4) Å, thus within the standard deviation identical to those of 6. Compound 11 is obtained by cleavage of HI after the aldol addition and subsequent coordination towards “BeI+”. A similar increase of the acidity is described for diketones upon the coordination to BeCl2.32 While 11 represents the major product, 4 c, 5 c, and an unknown compound were also observed in the 1H and 13C NMR spectra. EXSY NMR experiments showed that there is a fast exchange between the ketones coordinated to compounds 4 c and 5 c on the NMR time scale, whereas no NOE cross‐peaks between the signals of ligand 10 and any MIPK moiety were observed. This leads to the deduction that complex 11 does not dissociate in solution. Furthermore, solutions of di‐ketone 5 c also proved to be unstable and led to the formation of a mixture of compounds 11, 4 c, and 5 c after several weeks. Therefore, in contrast to BeCl2 and BeBr2, BeI2 is able to convert MIPK into pivaldehyde. The analogous reaction of one equivalent of pivaldehyde and MIPK with BeCl2 showed signals of both ligands in 1H and 13C NMR spectroscopy and one signal in the 9Be NMR spectrum at 5.1 ppm, which is in between the signals observed for 2 a and 5 a as expected. This solution was stable for weeks and did not show any conversion to condensation product 10. We therefore deduced that the corresponding mixed aldehyde ketone adduct of BeCl2 (12 a) is stable and that a cationic intermediate is necessary for the condensation, which is only available from BeI2 adducts because of the weak Be−I bond.30
Scheme 3
Schematic illustration of the reactions of BeI2 with one and two equivalents of pivaldehyde and proposed pathway for the formation of compound 11.
Schematic illustration of the reactions of BeI2 with one and two equivalents of pivaldehyde and proposed pathway for the formation of compound 11.In comparison to pivaldehyde, isobutyraldehyde features a proton at the α‐C atom that enables it to form enols that can react further with the aid of a Lewis acid such as BeCl2. If isobutyraldehyde was exposed to catalytic amounts of BeCl2, the trimerization product 2,4,6‐tris‐isopropyl‐1,3,5‐trioxane (1 b) was obtained (Scheme 4). The same applies to BeBr2 and BeI2. Trioxane 1 b was subsequently used as precursor for isobutyraldehyde because of the easier handling. When BeCl2 was treated with two equivalents of isobutyraldehyde in CDCl3 or C6D6 a brownish solution was obtained within a few hours. The 1H NMR spectra showed a broad multiplet, which ranges from 0.5 to 3.0 ppm. In the 13C NMR spectra we were not able to observe discrete signals. After the solvent was removed in vacuo we obtained a brown waxy residue that did not show any discrete bands in the IR spectrum. The same results were obtained for other stoichiometries and with BeBr2 and BeI2. Therefore, we assume that isobutyraldehyde undergoes undirected polymerization.
Scheme 4
Schematic illustration of the reaction pathways of isobutyraldehyde with beryllium halides (X=Cl, Br, I); xs.=excess.
Schematic illustration of the reaction pathways of isobutyraldehyde with beryllium halides (X=Cl, Br, I); xs.=excess.For butyraldehyde we did not obtain the trimer when it was exposed to catalytic amounts of beryllium halides. Instead, an opaque beige solution of undefined composition was obtained. Because of this unselective reaction behavior, the decomposition of isobutyraldehyde and the complex reactivity of pivaldehyde with respect to BeI2, we limited our studies to the stochiometric reactivity of butyraldehyde to BeCl2 (Scheme 5). If two equivalents of butyraldehyde were treated with BeCl2 in low concentrations in CDCl3 the solution remained clear. The recorded 1H and 13C NMR spectra showed discrete signals for the butyraldehyde and its aldol condensation product 2‐en‐2‐ethyl‐1‐hexanal (13) in a one‐to‐four ratio. The signals of the carbonyl as well as of the allylic carbon atom are shifted downfield by 9.5 and 16.6 ppm, respectively, indicating a coordination of 13 towards BeCl2. In the 9Be NMR spectrum a singlet at 5.3 ppm (ω1/2=12.5 Hz) is observed, which is in a similar region as the signals of adducts 2 a and 5 a and evidence for a fourfold‐coordinated beryllium species in solution. We assumed that this is a simple adduct of two ligands 13 to BeCl2 (14). This was confirmed by the NMR spectra of a solution of two equivalents of 13 with one equivalent of BeCl2, which was prepared as a control experiment, and showed an identical set of signals with the same chemical shifts. However, if these solutions were stored for more than two weeks at ambient temperature they turned black and an unknown black solid precipitated, which indicates polymerization or decomposition. Therefore, attempts to obtain single crystals of that compound remained unsuccessful.
Scheme 5
Schematic illustration of the reaction pathways of butyraldehyde with BeCl2.
Schematic illustration of the reaction pathways of butyraldehyde with BeCl2.
Conclusion
We observed a rising reactivity among the halides from BeF2 to BeI2 when aldehydes were exposed to them. The aldehydes also showed distinctly different reactivities. The presence of a β‐H was the decisive factor for different behaviors towards the beryllium halides. We were able to catalyze the formation of trioxans 1 a and 1 b through exposure of pivaldehyde and isobutyraldehyde, respectively, towards minute quantities of BeCl2, BeBr2, or BeI2, whereas BeF2 stayed inert in all the experiments we carried out. This cyclization was reversed when the trioxanes were exposed to higher quantities of beryllium halides (Cl, Br, I). If two equivalents of pivaldehyde were treated with BeCl2 or BeBr2, the adducts 2 a and 2 b were obtained, whereas use of the same amounts of BeI2 resulted in the rearrangement reaction of pivaldehyde to MIPK and subsequent aldol addition to 10, which then coordinates towards BeI2 to form compound 11 under evolution of HI. The respective isomerization to methyl isopropyl ketone could also be achieved when pivaldehyde was treated with equimolar amounts of BeCl2 or BeBr2, and the corresponding adducts 4 a and 4 b were also obtained by direct reaction of one equivalent of MIPK with BeCl2 or BeBr2, respectively. In addition, the adducts of two ketones to BeCl2 (5 a), BeBr2 (5 b), and BeI2 (5 c) were synthesized and characterized as well as mono‐aldehyde adducts 3 a, 3 b, and 3 c. The solid‐state structure of compound 3 a could be determined, which is the first example of a linear beryllium halide chain. While the reaction of BeCl2 or BeBr2 with three or more equivalents of pivaldehyde led to the formation of the bisaldehyde adducts 2 a and 2 b, 1‐iodo‐alkoxide 6 was formed in the case of BeI2. This compound is generated through the formation of the cationic tris‐aldehyde complex 8 and subsequent nucleophilic attack of iodine at one of the coordinated aldehydes. This leap in reactivity from BeBr2 to BeI2 is caused by the weak Be−I bond, which is cleaved easily and leads to the generation of cationic beryllium species. This proves on the one hand the potential of beryllium compounds to catalyze a vast amount of reactions as one of the hardest Lewis acids. On the other hand are complexes 6 and 11 model systems for transition states that have been proposed for proton‐induced conversions of carbonyl groups.33 Therefore, Be2+—which is significantly smaller than any other metal ion—can be used as a proton substitute to isolate and structurally characterize transition‐state‐related compounds.Polysaccharide binding to glycan‐binding proteins is essential for immune recognition34 and even slight changes of these polysaccharides can lead to an immune response.35 Therefore, it is plausible that beryllium‐induced alterations of the glycocalyx lead to an attack against the body's own cells by the immune system. Thus, the various observed conversion reactions of aldehydes and ketones enable the evaluation of potential pathways, in which carbohydrates are altered or decomposed when exposed to beryllium compounds in the human body, especially in lung tissue, in which hexoses are abundant. Accordingly, an examination of the glycocalyx in the lung tissue of berylliosispatients or corresponding animal models with use of biochemical and immunologic techniques seems highly relevant.
Experimental Section
Caution! Beryllium and its compounds are regarded as toxic and carcinogenic. As the biochemical mechanisms that cause beryllium‐associated diseases are still unknown, special (safety) precautions are strongly advised.7
General experimental techniques
All manipulations were performed either under solvent vapor pressure or dry argon by using glovebox and Schlenk techniques. Chloroform was dried by heating to reflux over CaH2, and benzene over sodium and subsequent distillation under argon. CDCl3 was dried with CaH2 and C6D6 with Na/K‐alloy and vacuum‐transferred directly into the J. Young NMR tubes. BeF2, BeCl2, BeBr2, and BeI2 were prepared according to literature procedures36, 37 and the aldehydes and ketones were purchased from VWR. Because of the expected extreme toxicity of the obtained compounds no elemental analysis or mass spectrometry could be performed of the beryllium‐containing species. The purity of the compounds was therefore determined by NMR and IR spectroscopy.
NMR spectroscopy
1H, 9Be and 13C NMR spectra were recorded with Bruker Avance III HD 300 and Avance III 500 NMR spectrometers. The latter was equipped with a Prodigy cryo‐probe. 1H NMR (300/500 MHz) and 13C NMR (76/126 MHz) chemical shifts are given relative to the solvent signal for C6D6 (7.16 and 128.1 ppm)/CDCl3 (7.26 and 77.2 ppm) and 9Be (42 MHz) used 0.43 m BeSO4 in D2O as an external standard. NMR spectra were processed with the MestReNova software package.38
IR spectroscopy
IR spectra were recorded with a Bruker alpha FTIR spectrometer equipped with a diamond ATR unit in an argon filled glovebox. Processing of the spectra was performed with the OPUS software package39 and OriginPro 2017.40 Either single crystals of the compounds (when possible) or bulk material were used for the IR spectroscopic measurements.
Single‐crystal X‐ray diffraction
Single crystals were selected under exclusion of air in perfluorinated polyether (Fomblin YR 1800, Solvay Solexis) and mounted by using the MiTeGen MicroLoop system. X‐ray diffraction data were collected with use of either the monochromated CuKα radiation of a Stoe StadiVari diffractometer equipped with a Xenocs microfocus source and a Dectris Pilatus 300 K detector or monochromated MoKα radiation of a Stoe IPDS II equipped with an image plate detector. The diffraction data were reduced with the X‐Area software package.41 The structures were solved by using direct methods (SHELXT‐2018/2) and refined against F
2 (SHELXL‐2018/3) with use of the ShelXle software package.42, 43, 44 All atoms were located by Difference Fourier synthesis and non‐hydrogen atoms refined anisotropically. Hydrogen atoms were refined isotropically.
Mass spectrometry
Mass spectrometry was performed with a Thermo Fischer Scientific LTQ‐FT Ultra equipped with a combined ESI, APCI, and nanospray ionization source.
NMR and IR spectroscopic data of purchased compounds
Beryllium‐containing compounds were dissolved in chloroform. The chloroform solution was treated with water to extract the Be2+ cations from the organic phase. The chloroform solution was then diluted with dichloromethane to obtain a 1 mm solution that was subsequently analyzed by mass spectrometry. MS (ESI+): m/z (%): 173.2 [10
+] (35), 195.2 [10+Na+] (100).
Conflict of interest
The authors declare no conflict of interest.As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re‐organized for online delivery, but are not copy‐edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.SupplementaryClick here for additional data file.
Authors: Onyekachi Raymond; Penelope J Brothers; Magnus R Buchner; Joseph R Lane; Matthias Müller; Nils Spang; William Henderson; Paul G Plieger Journal: Inorg Chem Date: 2019-04-09 Impact factor: 5.165
Authors: Paul G Plieger; Kevin D John; Timothy S Keizer; T Mark McCleskey; Anthony K Burrell; Richard L Martin Journal: J Am Chem Soc Date: 2004-11-10 Impact factor: 15.419
Authors: Dominik Naglav; Briac Tobey; Kevin Dzialkowski; Dieter Bläser; Christoph Wölper; Georg Jansen; Stephan Schulz Journal: Dalton Trans Date: 2018-09-18 Impact factor: 4.390