Green and Rapid Hydrothermal Crystallization and Synthesis of Fully Conjugated Aromatic Compounds

Abstract Highly fused, fully conjugated aromatic compounds are interesting candidates for organic electronics. With higher crystallinity their electronic properties improve. It is shown here that the crystallization of three archetypes of such molecules—pentacenetetrone, indigo, and perinone—can be achieved hydrothermally. Given their molecular structure, this is a truly startling finding. In addition, it is demonstrated that perinone can also be synthesized in solely high‐temperature water from the starting compounds naphthalene bisanhydride and o‐phenylene diamine without the need for co‐solvents or catalysts. The transformation can be drastically accelerated by the application of microwave irradiation. This is the first report on the hydrothermal generation of two fused heterocycles.

Organic colorants are molecules that show strong color when interacting with light. Aside from their oldest application, that is,c oloring other substances,i nr ecent years they have become of interest for optoelectronic applications in, for example,organic solar cells [1] and as organic transistors. [2] One important class of organic colorants are the so-called carbonyl dyes,which are typically highly resistant to heat, solvents,and weathering. [3] Perinone,t hat is,n aphthalene tetracarboxylic acid bisbenzimidazole,i ss uch ac arbonyl dye.T he name perinone is generally used for mixtures of cis and trans isomers ( Figure 1).
These mixtures are of intense red color.T he pure, separated isomers show cleaner, individual colors: cis-perinone has ab lueish-red hue,w hile the industrially more important trans-perinone has ab rilliant orange-reddish color. [4] To the best of our knowledge,t here is not as ingle direct synthetic approach that allows for obtaining exclusively either of the two isomers.T his is because the isomers cocrystallize:I ndeed, the two isomers form as olid solution, which is structurally highly tolerant as reflected by the fact that the two isomers can be combined in ab road range of ratios by positional and rotational disorder. [5] As in the last decades,r esearch on perinone has been dominated by the development of isomer separation techniques rather than syntheses,t here are only ah andful of synthetic procedures towards perinone.T hese include: 1) refluxing the starting compounds in conc. acetic acid for several hours; [6] 2) condensation of the starting compounds in imidazole using Zn(OAc) 2 as catalyst; [7] this is analogous to the Langhals method for generating perylene bisimides; [8] 3) rapid condensation (15 min) in H 3 PO 4 at 190 8 8C; [9] 4) precondensation to a-amino-imidazoles or a-carboxylic acid imides in H 2 Oa tr eflux, followed by solid-state condensation to perinone. [10] Clearly,a ll of the these syntheses except the latter two-step route reported by Mamada et al. are far from being green approaches. [10] With this contribution, we have set out to prepare perinone hydrothermally using nothing but high-temperature water (HTW) and the starting compounds in astoichiometric ratio.T his work was prompted by our recent reports on the fully green hydrothermal (HT) synthesis of perylene and naphthalene bisimides, [11] and polyimides. [12][13][14] While others have used near-critical water (250-350 8 8C) for the preparation of benzimidazoles, [15] and supercritical water ( % 400 8 8C) for various benzazoles, [16] our HT method requires only HTW (typically 180-250 8 8C). Consequently,e nergy consumption is lower and the required safety measures are reduced. However,todateHTcondensation has only been reported for the formation of single heterocycles and noncyclic amides. [12][13][14]17] In the latter cases,t he action of HTW has been shown to generate superior crystallinity.T herefore,w ew ere intrigued to expand the scope of HT synthesis and crystallization for the very first time towards fused heterocycles.
In an initial experiment, we investigated the general feasibility of hydrothermally generating perinone from the starting compounds o-phenylene diamine (o-PDA) and naphthalene bisanhydride (NBA), as shown in Figure 1. Therefore, o-PDAand NBA(2:1molar ratio at an equivalent concentration c eq = 0.01 mol L À1 )w ere suspended in deionized H 2 Oa nd heated to 200 8 8Ci nanon-stirred batch autoclave (see the Supporting Information (SI) for experimental details). After areaction time t R of 16 h, the autoclave contained two distinct phases:ared, flocculent solid as as ediment at the bottom of the glass liner, and an orange, translucent aqueous supernatant phase (see Figure 1). The sediment, whose red color was already indicative for the formation of perinone,was collected, dried and characterized without further purification. Attenuated total reflectance Fourier transform infrared (ATR-FTIR) analysis ( Figure 2; SI for details) of the crude product revealed the absence of modes characteristic for the starting compounds.Instead, we found several intense modes indicative for perinone.
Theamount of obtained crude perinone corresponded to approximately 90 %o ft he theoretical yield. In order to further ascertain the formation of perinone,w ep erformed solution 1 HNMR analysis. 1 HNMR measurements (Figure 3; SI) revealed, that the crude perinone is composed of amixture of cis and trans isomers. [10,18] However,t he two multiplets labeled as Aa nd Bi nF igure 3i ndicated the presence of as mall amount of byproduct. Attempts to remove this byproduct via extraction were not successful. Therefore,w e turned our attention to selective precipitation, at echnique commonly used to purify or fractionate polydisperse polymers. [19] By selective precipitation from trifluoroacetic acid (TFA) with H 2 Oi tw as finally possible to purify the target compound and isolate the byproduct. Thecombined results of ATR-FTIR, 1 HN MR, and laser desorption/ionization highresolution mass spectrometry (LDI-HRMS) measurements (see SI) served to identify the byproduct as naphthalene tetracarboxylic acid monoanhydride monobenzimidazole (NMM). Consequently,i tw as also possible to calculate the amount of NMM in crude perinone.A ccording to 1 HN MR analysis,N MM accounts for approximately 7mol %o ft he unpurified product mixture (see SI for calculation). From these measurements the cis/trans ratio in the crude product as well as in reprecipitated perinone was determined to be approximately 2:3(see SI for details).
Afurther set of experiments revealed that 1) aminimum of t R = 12 hwas necessary to yield maximum conversion of the starting compounds and 2) formation of NMM could not be avoided at any tested t R (up to 48 h). Moreover,i tb ecame evident that NMM is formed as an intermediate during the reaction and its amount steadily decreases with t R until acertain threshold is reached (after t R ! 12 h).
TheHTmethod for the condensation of o-PDAand NBA described clearly qualifies as agreen synthesis,since 1) no cosolvents or catalysts are required and 2) H 2 Oi sn ot only the sole reaction medium but also the sole condensation byproduct. Nonetheless,w ew ere interested in making the transformation even greener and more energy-efficient. We therefore turned our attention towards microwave (MW)-assisted HT condensation. [20,21] We conducted aseries of experiments in aMWoven equipped with astirrer, where it was possible to exactly control the reaction temperature T R ,h eating time t H (time until T R is reached), and t R (reaction time at T R ). When we performed these experiments at T R = 200 8 8Ca nd t H = 10 min, we found that already after t R = 2min the starting compounds had completely reacted to form perinone.E vidently,not only efficient heating via MW irradiation, but also the application of stirring is highly beneficial for the reaction to proceed rapidly.H ence,t hrough use of this MW setup it was possible to decrease the total reaction time at T R = 200 8 8C from 12 ht o1 2min. Lowering T R to 180 8 8Cd rastically increases t R ,w hereas at the highest tested T R of 250 8 8Ci ti s not even necessary to maintain this T R for acertain time.This tremendous reduction of t R upon increasing T R can be attributed, among other factors,t otemperature-dependent,  significant changes in the physicochemical properties of H 2 O. [22] With increasing T R the static dielectric constant decreases steadily,w hich allows for better dissolution of organic compounds that are completely insoluble in H 2 Oa t RT.F urthermore,a ne levation of T R leads to an increasing ionic product of H 2 Ou ntil am aximum is reached at 250 8 8C. This phenomenon makes H 2 Oi tself ap owerful acid/base catalyst in the HT regime,w hich is indeed beneficial for organic condensation reactions.
Interestingly,a sl ong as full conversion of o-PDAa nd NBAi sa chieved, the NMM content as well as the cis/trans ratio ( % 2:3) remained constant independent of the reaction conditions.T he typical amount of % 7mol %o fN MM was always formed. Neither extending t R ,i ncreasing t H ,v arying the pH, nor lowering c eq affected the ratio of isomers and the byproduct content. Forthe HT condensation of perylene and naphthalene bisanhydride with monoamines to yield bisimides it had been shown that the presence of an onnucleophilic base such as Hünigsb ase had ap ositive effect (reduction of t R ,i ncrease in yield) on the formation of some derivatives. [11] Unfortunately,for the reaction of NBAwith o-PDAthe addition of Hünigsbase did not suppress or reduce the formation of NMM. One could furthermore speculate that NMM always forms due to al ack of o-PDA, which could be consumed by the well-known oxidative polymerization of aromatic amines. [12] When we concentrated the liquid, orange supernatant phase,w hich was always found after the HT reaction, to dryness,asmall amount of dark purple solid was obtained. FTIR-ATR, 1  Thef act that the formation of NMM is not suppressed under any of the tested conditions,i ncluding the addition of acondensation promoter and the use of excess o-PDA, points to the incorporation of NMM into the crystal lattice of perinone as an alternative explanation. We suspect that the similar size and the planarity of NMM as well as its electronic similarity to cis-a nd trans-perinone facilitates its incorporation into perinonesc rystal lattice through positional and rotational disorder.O nce NMM is incorporated, it can be considered as "entrapped" and consequently it is not available for further condensation. As for the curious fact that we always find % 7mol %ofNMM in the crude perinone, we speculate that perinone crystals incorporate the amount of NMM that the crystal lattice can maximally tolerate.
Therefore,a lthough the HT condensation of o-PDAa nd NBAwas found to be highly robust and proceeds rapidly at T R ! 200 8 8C, selective precipitation from TFAwas essential to remove traces of NMM. Interestingly,p owder X-ray diffraction (PXRD) measurements revealed that this purification technique significantly decreased crystallinity (Figure 4). In fact, reprecipitation with an on-solvent rapidly quenches the perinone solution, thereby generating as trongly disordered liquid-crystalline (LC) phase.W henw ea ttempted to index the reflections to arectangular lattice,wefound that all peaks fitted relatively well (see SI for detailed information) to acolumnar rectangular LC phase.
Theu nexpectedly high disorder in reprecipitated perinone provided an excellent opportunity for us to attempt the very challenging task of its HT crystallization. This task is ambitious,because perinone is much more hydrophobic than its precursors NBAa nd o-PDA. Hence,g reater difficulties regarding perinonesdissolution must be expected. At 250 8 8C polarity of H 2 Oiscomparable to that of ethanol at RT; [23] thus H 2 Oi sc ertainly not apolar enough to expect it to be ag ood solvent for perinone.
To test the feasibility of HT crystallization, perinone was simply dispersed in H 2 Oa nd heated to the HT regime. Fortunately,P XRD patterns showed that the HT treatment tremendously increased crystallinity ( Figure 4). Furthermore, scanning electron microscopy (SEM) measurements clearly revealed that the HT treatment also improved the morphology (see SI): Ther eprecipitated roundish particles (0.1-0.5 mm) with ar ough, random surface texture were transformed into agglomerates of fine needles (1-2 mminlength). When commercially available trans-perinone was subjected to HT conditions,t he morphological improvement was even more significant:b eautiful needles (approximately 5-10 mm in length) with ar ather narrow size distribution featuring smooth crystal facets were obtained ( Figure 5).
Hence,H Tc rystallization of perinone is indeed possible. These drastic morphological changes can only arise from HT dissolution of reprecipitated/commercially purchased perinone and its subsequent crystallization. Considering the polarity of H 2 Oa t2 50 8 8C, it is very surprising that perinone dissolves.W es uspect that the increased ionic product of HTW allows for the protonation of perinone,which is indeed the mechanism for its dissolution in some strong acids such as TFA. Additionally,t he gain in lattice energy upon passing from the LC to the fully crystalline state could be am ajor driving force.
Thrilled by these findings,wedecided to attempt the HT crystallization of other,fully conjugated aromatic compounds, also of interest for organic electronics and for which crystallinity matters.S pecifically,w ec hose indigo (2,2'bis(2,3-dihydro-3-oxoindolyliden)) and pentacenetetrone (pentacene-5,7,12,14-tetrone). [24,25] Like perinone,b oth of these compounds are insoluble in H 2 Oa tr oom temperature due to their ability to strongly p-stack. [26,27] To test the feasibility of their HT crystallization, we applied the optimized reaction conditions found for crystallizing reprecipitated perinone.A TR-FTIR and 1 HN MR measurements revealed that both substances are stable under HT conditions (see SI). Furthermore,P XRD results indicated that dissolution and crystallization of these compounds is indeed possible under HT conditions,s ince reflections became sharper and more pronounced (see SI). SEM analysis (Figure 6; SI) additionally showed as triking improvement of the morphology for both compounds.
In summary,w eh ave shown that perinone can be synthesized hydrothermally solely from its starting compounds in stoichiometric ratio.T he synthesis is highly robust and remains unaffected by various reaction parameters (T R , t R , t H , c eq ,p H, O 2 )a nd does not require condensation promotors or catalysts.T he obtained product only contains minor amounts of NMM byproduct. Moreover,w ed emonstrate that it is possible to crystallize intentionally amorphisized perinone as well as recrystallize commercially obtained perinone.Furthermore,wecould show that HT crystallization can also be successfully applied to other fully conjugated aromatic compounds.T hese crystallization experiments set the basis for our strong conviction that HTW is ap romising, cheap,a nd environmentally benign medium for crystallizing temperature-stable low-molecular-weight compounds.