Substituted 1,3,5-Triazine Hexacarboxylates as Potential Linkers for MOFs

Hexacarboxylates are promising linkers for MOFs such as NU-109 or NU-110, which possess large values for surfaces and pore volumina. Starting from 2,4,6-tris(bromoaryl)-1,3,5-triazines, palladium-catalyzed cross coupling reactions (Suzuki-Miyaura, Sonogashira-Hagihara) form elongated hexacarboxylate linkers. Eight new 2,4,6-tris(biphenyl) and 2,4,6-tris(phenylethynylphenyl) 1,3,5-triazines have been prepared in quantities ranging from 40 mg to 1.1 g. In five cases, one of the arms of the linker carries an additional functionality (NO2 or OMe).

The properties of MOFs can be altered by the introduction of additional substituents. There are two general approaches: substituted linkers can be used, or the additional functionality is introduced post-synthetically. However, post-synthetic modifications are rarely quantitative and are frequently accompanied by decomposition especially when the material possesses a high porosity. Therefore, the use of already functionalized linkers will lead to more homogeneous MOFs.
In this work, we describe the syntheses of mono-functionalized hexadentate ligands. In contrast to the linkers used in NU-109 and NU-110, we have chosen a 1,3,5-triazine as the central aromatic ring because triaryltriazines are more planar than triarylbenzenes [12,13]. A variety of related tridentate and monofunctionalized triazine linkers 1 and 2 have been synthesized (see Figure 1) [14]. Some of them have already been used in the syntheses of MOFs, yielding PCN-6 analogues that contain additional functionalities such as NO 2 and NH 2 [13][14][15][16].  [13,14]. The elongated tricarboxylic acids 2 were synthesized from the respective tribromides 3a-c using Suzuki-Miyaura couplings.

Results and Discussion
For the extension of 3a-c, Suzuki-Miyaura couplings and other transition metal-catalyzed cross-couplings can be utilized, for instance the Sonogashira-Hagihara reaction. To finally obtain a hexacarboxylate, the coupling partners must contain two carboxylic acid functions. Hence, isophthalic acid derivatives have to be used, many of which are commercially available or have been described in the literature. For the syntheses of the hexadentate linkers 6 and 11, 5-boron-and 5-iodo-substituted isophthalic derivatives 4 and 5, respectively, were needed ( Figure 2). Boronate 4 is commercially available but was synthesized in this work from dimethyl isophthalate via its 5-bromoderivative by palladium(0)-catalyzed boronation with bis(pinacolato)diboron [17]. Anhydrous conditions are necessary to avoid coupling of the brominated starting compound with the product to give an undesired biphenyl derivative carrying four ester groups.
Diethyl iodoisophthalate 5 was synthesized from 5-aminophthalic acid. After esterification, the iodo function was introduced by a Sandmeyer analogous iodination following a procedure from the literature [18].

2,4,6-Tris(biphenyl)-1,3,5-triazine Hexacarboxylates
The Suzuki-Miyaura reaction is a well-established method to connect aromatic rings. Consequently, unsubstituted and nitro-and methoxy-substituted 2,4,6-tribromo-1,3,5-triazines 3a-c were coupled with boronate 4 (Figure 3). The respective hexamethyl hexacarboxylates 6 could be isolated in an 81 to 88% yield. The last step in the synthesis of the functionalized hexadentate linkers 7, hydrolyses of the esters 6, was performed with lithium hydroxide in a mixture of water and THF in an 89 to >99% yield. However, it should be noted that the esters can also be used in MOF syntheses, as long as hydrolytic conditions are used. In these cases, the esters are hydrolysed yielding the corresponding carboxylates as the actual linkers.

2,4,6-Tris(phenylethynylphenyl)-1,3,5-triazine Hexacarboxylates
The reason to use a triazine core for linkers with three arms rather than a benzene ring -i.e. using for instance TATB (1,3,5,-triazine-2,4,6-tribenzoate) instead of BTB (1,3,5-benzenetribenzoate), see above and ref. [12,13] -is the more pronounced planarity of the aromatic rings in the triaryltriazine system. However, in the hexadentate linkers 6 and 7, biphenyl substructures have been generated by the cross-coupling reaction. The repulsion of the ortho hydrogen atoms in the biphenyls will lead to twists in the "arms." This problem will even be larger if each arm contained more aryl rings-for instance, if it was a p-terphenyl. We have therefore chosen an alternative structure by exchanging the central aromatic ring of a potential p-terphenyl by an alkyne. NU-109 also contains aryl-alkyne-aryl subunits instead of aryl-aryl ones as present in NU-110 [10]. In terms of coupling chemistry in the syntheses of the linkers, Sonogashira-Hagihara couplings have to be performed instead of Suzuki-Miyaura reactions ( Figure 4).
Starting from the tribromides 3a-c, a first triple Sonogashira-Hagihara coupling with trimethylsilyl-protected ethyne 8 gave 2,4,6-tris(trimethylsilylethynylphenyl)-1,3,5-triazines 9 in 72 to 94% yield. The TMS protecting groups in 9 were easily cleaved off by treatment with potassium carbonate in methanol [19]. The unsubstituted and the methoxy-substituted linkers 10a and 10c could be isolated in a >99% yield, while the nitro compound 10b could not be purified sufficiently. Therefore, the final step, a triple Sonogashira reaction of the triynes with diethyl 5-iodoisophthalate 5, was carried out with 10a and c, and the hexaethyl hexacarboxylates 11a and 11c were obtained in yields of 79 and 81%. These yields correspond to >92% yield for each single coupling step.
While the hexaesters 6a-c could be hydrolyzed to yield the corresponding hexaacids 7a-c in analytically pure form (see above), we were not able to isolate the hexaacids derived from 11a or c in a sufficiently pure form. Nevertheless, esters can be employed in MOF syntheses as well because most solvothermal reactions conditions hydrolyze esters anyway.

Experimental Section
General Remarks: 1,1 -Bis(diphenylphosphine)ferrocene-palladium(II) chloride (99.9%, ABCR, Karlsruhe, Germany), bis(triphenylphosphine)palladium(II) dichloride (98%, ABCR), tetrakis(triphenylphosphine)palladium(0) (99%, ABCR), and trimethylsilylethyne (8, 98%, ABCR) were purchased and used without further purification. Dry solvents were obtained using suitable desiccants. Other solvents were distilled before use. Melting points were measured with a Gallenkamp MPD350.BM2.5 instrument. NMR spectra were recorded with a Bruker DRX 500 or Avance 600 instrument at 300 K (Billerica, MA, USA). Assignments are supported by COSY, HSQC, and HMBC. Even when obtained by DEPT, the type of 13 C signal is always listed as singlet, doublet, etc. All chemical shifts are referenced to tetramethylsilane or the residual proton or carbon signals of the solvent. 1 H and 13 C-NMR spectra of compounds 6, 7, 9, 10, and 11 can be found in the Supplementary Materials. HRMS-EI mass spectra were recorded with JEOL AccuTOF GCV 4G (Tokyo, Japan). MALDI-TOF mass spectra were recorded with a Bruker-Daltronics Biflex III (Billerica, MA, USA) with Cl-CCA (4-chloro-α-cyanocinnamic acid) as matrix. IR spectra were recorded with a Perkin-Elmer Spectrum 100 spectrometer (Waltham, MA, USA) equipped with a Golden Gate Diamond ATR unit A-531-G. Elemental analyses were carried out with a Euro EA 3000 Elemental Analyzer from Euro Vector (Pavia, Italy). Traces of solvent originated from the purification step. The analytical sample of 11a was obtained from an NMR solution.