Some Products from C=O Condensations of Quinacridones


Quinacridone chemistry has been developed which provides a source of new compounds that have potential application in organic photovoltaic (OPV) devices. Phosphoryl chloride is used for the conversion of carbonyl groups in N,N′-dialkylquinacridones, generating reactive intermediates that enable selective condensation with either one or two nucleophilic molecules under mild conditions.



Introduction
With anthropogenic climate change potentially affecting all elements of the biosphere, efforts to mitigate the most significant of its effects by minimising carbon emissions are essential. Photovoltaic (PV) devices are a source of renewable energy that are critical to achieving this objective. Among the various PV device types that are being actively explored, solar cells based on organic materials demonstrate significant potential, due to their low production costs, high power-to-weight ratios, and the power conversion efficiencies (PCEs) of single junction devices now exceeding 18 %.
In particular, organic photovoltaics (OPVs) using soluble small molecules have attracted attention [1] due to their numerous favourable attributes, such as high purity, scalable synthesis, tuneable optoelectronic properties through molecular structural variation, and excellent stability. [2,3] As a result, a wide variety of planar small organic molecules have been evaluated for prospective utility in OPV devices. Both electron donor and electron acceptor molecules have been of interest, and while substantial progress has been made in the area of suitable donor molecules, promising acceptor molecules have been rather less abundant.
Fullerene-based compounds have long been among the most effective and popular single molecule electron acceptors since their first application. Gradual improvements have led to devices using this material giving PCEs in excess of 14 %. [4] Nevertheless, cost and other shortcomings have encouraged a continuing search for novel acceptor materials as alternatives for fullerenes. Several classes of well known commercial pigments, such as perylene diimides, phthalocyanines, and others, have shown some promise as n-type materials. The quinacridone (QA) nucleus (Fig. 1, Table 1, compounds 1a,  1c, 1f), which has previously been extensively utilised in the preparation of light fast pigments, has also been investigated for photovoltaic utility, both in a role as an acceptor or donor molecule. [5] Substantial efforts have been devoted to the synthesis and characterisation of various QA derivatives, however, somewhat less attention has been paid to specifically modifying its optoelectronic properties. One route to modification is via condensation of the relatively unreactive carbonyl groups at C-7 and C-14 to afford molecules with extended conjugation and possibly other desirable electronic properties. [6] Some success has been achieved in condensing QAs with arylamines in the presence of titanium tetrachloride to form N,N 0 -dialkyl quinacridoneimines. Photophysical and electrochemical properties and temperature dependent geometrical isomerism in alkyl quinacridoneimines have been documented, [7] but these products had relatively low efficacy in solar cell trials when tested as p-type materials.  While Wang and others have been able to condense the carbonyl groups of N,N 0 -dialkylquinacridones with selected highly activated methylene groups in compounds such as malononitrile in a Knoevenagel-type condensation under prolonged reflux in acetic anhydride, the synthetic diversity available by this route has been limited. [8,9] Subsequent developments employing TiCl 4 [10,11] to enable condensations have somewhat improved the accessibility of some condensed QAs, increasing the potential for producing candidate compounds for photovoltaic applications. Trials of a bis N-octylated dicyanoethylenated QA as an acceptor material in a solar cell with poly3-hexylthiophene (P3HT) as a donor gave encouraging conversion efficiencies up to 1.57 %.
While the power conversion efficiency achieved was relatively moderate, the study nevertheless confirmed the potential of a QA core, suitably condensed at the C=O groups, to provide a template worth exploring more extensively for designing improved acceptor molecules. [12] To this end, it was of interest to develop an alternative, more versatile, methodology for creating a wider variety of condensation products from diverse starting materials under mild conditions. In addition, if needed, several potentially useful QA variants bearing nuclear substituents such as 5,12-dichloro-(1c) and 5,12-dimethyl-(1e) are readily available commercially.

Results and Discussion
During a search for reagents to effect activation of an N-alkylated QA with relatively unreactive carbonyl groups into a more active species, thereby rendering the molecule more amenable to condensations, an excess of POCl 3 was added to an orange suspension of QA (1a) in 1,2-dichloroethane (DCE). When the mixture was stirred at 808C for 1 h it darkened, and the suspended material dissolved to form an emerald green solution which did not change with further heating. When a portion of this reaction mixture was stirred with an excess of water it recovered its original orange colouration and TLC indicated only starting material was present.
The nature of the compound contained in the green solution was examined more closely. With deuterochloroform as the reaction medium (containing a few drops of DMF as catalyst), a similar green solution was formed at room temperature over five days. The change from a symmetrical substitution pattern in the starting material (Fig. 1, Table 1, Fig. S1 (Supplementary Material), compound 1b) to an unsymmetrical reaction product (Scheme 1, Fig. S2 (Supplementary Material), Intermediate A) was inferred from a doubling in the number of resonances observed in the 1 H NMR spectrum. Particularly significant was the large downfield shift shown by one of the N-CH 2 groups from d 4.44 (in the starting QA 1b) to d 5.48, which was accompanied by a downfield shift of several of the 10 aromatic proton signals. This suggested a possible transformation to a quinacridinium salt (14-chloro-5,12-dimethyl-7-oxo-7,12-dihydroquinolino[2,3-b]acridin-5-ium chloride) (Intermediate A, Scheme 1). This inference is consistent with reactions observed subsequently with nucleophiles in the presence of an organic base. The absence of any indication of P-O-C(14) coupling in the 13 C NMR spectrum suggests that while a phosphate may have formed initially, it was readily converted into a reactive 14-chloro species in a manner resembling a vinylogous Vilsmeier reaction (Scheme 1).
From a consideration of the 1 H NMR spectrum of intermediate A, it seems likely that phosphoryl chloride (POCl 3 ) initially forms a mono quaternary salt which, as a consequence, now renders the molecule significantly more electron deficient and inhibits a similar conversion of the second carbonyl group. When the positive charge is discharged by condensation with a suitable nucleophile the ability to generate a reactive quaternary species is restored. This leads to a second activation step, and the second C=O is converted into a species amenable to further condensation and the formation of a bis-condensed product. Thus, even with a large excess of POCl 3 present it was feasible, by limiting the reaction time, to synthesise a mainly mono-condensed product which could be isolated, the second carbonyl group activated, and condensation effected with a different nucleophile.

Condensations with Activated -CH 2 -Groups
When a reaction was carried out with the green solution described above with an excess of methyl cyanoacetate at room temperature overnight in the presence of 2,4,6-collidine the bis product 2a was formed in 74 % yield. Several similar products (Scheme 2, Table 2) were obtained under comparable conditions with a variety of compounds, with the principal requirement for condensation being a methylene group sufficiently activated by flanking electron-withdrawing groups. An example of the conformation adopted by a representative of these bis-condensed products, 2b, in the solid state as determined by X-ray crystallography is shown in Fig. 2a Alternatively, in order to favour mono-condensation, the green solution was cooled to 08C and treated with an excess of methyl cyanoacetate followed by triethylamine. After 2 h at room temperature the mixture was quenched with ethyl acetate Similarly, a brief reaction with an excess of malononitrile also gave a mono-condensed product (3b, 85 %). Several organic bases, such as triethylamine, diisopropylethylamine, and 2,4,6-collidine, were effective in promoting these condensations, with 2,4,6-collidine being the most effective. In all cases the presence of a substantial excess of both POCl 3 and the nucleophilic component was tolerated and formation of the biscondensation product could be enhanced by extending the reaction time or by gently warming the reaction mixture.
Whether the routine addition of an excess of the nucleophilic component had any significant effect on the yield or rate of formation of the products was not investigated. Surprisingly, it was subsequently found that it was unnecessary to preprepare a solution of the green reactive intermediate before reaction with a nucleophile, and indeed the condensations proceeded quite readily when a mixture containing the QA, an excess of POCl 3 , the nucleophilic component, and an organic base was stirred at room temperature. In all reactions, progress was conveniently monitored by examining the TLC plate of a quenched aliquot, which showed the initial emergence of a major blue or violet component accompanied by a minor green spot that represented the bisadduct, which gradually increased in relative intensity at the expense of the mono product. In general, the condensations were surprisingly tolerant of the number of reaction equivalents of the reacting compounds. Thus, with one equivalent of QA or monocondensed QA, satisfactory condensation products were isolated from reaction mixtures containing from 3 to 5 equivalents of POCl 3 and . 1 equivalent of the nucleophilic coupling component. Progress of the condensation was readily monitored by quenching a small sample from the reaction solution and observing the elution of distinctive coloured spots, which were  easily identifiable as starting material, mono-condensed, and biscondensed entities. As expected, prolonged reaction favoured formation of bis-products, especially in the presence of 2,4,6-collidine as a base (Table 2).

Condensations with Activated -CH 3 Groups
Condensation of the reactive green QA intermediate could also be affected with activated methyl groups. Substrates sufficiently activated for such reactions were obtained by condensing acetophenones with malononitrile [13] to generate a product of the type ArC(CH 3 )=C(CN 2 ), in which the CH 3 group was sufficiently reactive for condensation to occur (Scheme 3). Several compounds bearing this moiety were successfully condensed (Scheme 3, Table 3). This procedure may be of some potential value to produce products with enhanced pi conjugation due to the greater planarity that could result from relatively less congestion at the newly formed methylene group. An example of the conformation of a representative of this group 4c as determined by X-ray crystallography is shown in Fig. 3.
Owing to their industrial applications as versatile pigments, several ring substituted QAs are commercially available. With the intention of evaluating the effect that a change in nuclear substitution might have on the photovoltaic properties of derivatives of the type previously described, a sample of Pigment Red 202 (1c, Colour Index 73907) was purchased and duly alkylated as received with n-octyl   bromide. Pigment Red 202 is usually assumed to be the 2,9dichloro derivative of QA. Rather surprisingly, the product from alkylation of this commercial sample was found to consist of two dioctylated species which were formed in comparable quantities and consisted of a mixture of 5,12-dioctyl-2,9-dichloroquinacridone (1d) and 5,12-dioctyl-2,9-dimethylquinacridone (1f). Evidently the QA sample as supplied was not a single entity as expected but in fact consisted of a blend of Pigment Red 202 (1c) and Pigment Red 122 (1e) in similar proportions. After separation, each of these analogues, in the presence of POCl 3 , readily gave products analogous to those observed with the unsubstituted parent (Table 5). Although the mixture as supplied (after N-alkylation) readily condensed with nucleophiles in the presence of POCl 3 to give the expected products as a mixture, it was more convenient to simplify subsequent purifications by

Condensation with a Different Second Component
Mono-condensed products, such as the initial product 3b obtained from malononitrile, were readily prepared and isolated from reactions employing short reaction times. Hence it became possible to activate the remaining C=O as a separate second step initiated by a second charge of POCl 3 . This enabled a stepwise introduction of a second substituent to yield a molecule with asymmetric electron flow characteristics, potentially enabling graduated tuning of various electronic properties. Notably, in an example of a product of this reaction that was characterised by X-ray crystallography, 6a, the QA moiety adopted a concave morphology (Fig. 5), which was in direct contrast to the planar or armchair morphologies observed for the products of the condensations with R-NH 2 groups (Fig. 4) and the bis-condensation with malononitrile ( Fig. 3), respectively.

UV-Visible Spectra
The spectra of representative compounds of structural type 2-6 were determined (Fig. 6). Significantly, each of each of these groups showed intense absorption in the range 650 to 750 nm, an energy rich range of the visible spectrum over which the popular n-type fullerene acceptors have low absorption. The newly synthesised QAs of structural type 4 showed particularly strong absorption in this region.

Conclusions
A new method has been developed which enables dialkylated QAs to be transformed under mild conditions to reactive intermediates which readily condense with a variety of nucleophilic components to yield condensed products. The reaction conditions can be modified to select either biscondensed products utilising both carbonyl groups at once or mono-condensed products in which the second carbonyl group can be activated and condensed with a second, different, nucleophile. This work provides a pathway to a variety of highly substituted quinacridone derivatives, allowing for customisation of their optoelectronic and physical properties, which is potentially useful in the fabrication of bulk heterojunction solar cells.

Methods and Materials
QA was purchased from Tokyo Kasei Kogyo Co. and a QA described as CI 202 was purchased from Winchem Industrial Co Ltd, China. Both compounds were used without further purification. N-(n-Octyl)-N 0 -(n-octyl)quinacridone was prepared by the literature method [14] with minor modifications. Analytical TLC was performed on Merck Kiesegel 60 F 254 silica aluminium backed sheets, and was visualised under visible or UV light. Column chromatography was performed on Merck Kiesegel (particle size: 0.04-0.063 mm) silica gel.
NMR spectra were recorded with a Bruker UltraShield Avance III 400 MHz Spectrometer running the TopSpin 2.1 software package at 293 K. CDCl 3 and DMSO-d 6 were used as solvents and as an internal lock. Chemical shifts are measured in ppm. 1 H NMR chemical shifts were referenced to d 7.26 for CDCl 3 and d 2.50 for DMSO-d 6 . Spectra were run routinely as a solution in deuterochloroform unless otherwise stated. 13 C NMR were run in CDCl 3 at 100.6 MHz unless otherwise stated; chemical shifts were referenced to d 77.0 for CDCl 3 and d 39.52 for DMSO-d 6 . Spectroscopic data were reported using the following format: chemical shift (ppm), integration, multiplicity assignment. Dichloromethane is abbreviated as DCM and 1,2-dichloroethane as DCE.
Melting points were determined using a Gallenkamp hotstage microscope and are uncorrected.
Atmospheric pressure chemical ionisation (APCI)/atmospheric sample analysis probe (ASAP) mass spectrometry, including high resolution experiments, were carried out on a Thermo Scientific Q Exactive FTMS employing ASAP/APCI probes.  n-Octyl bromide (19.3 g, 100 mmol) was added to a suspension of QA (10.41 g, 30 mmol) in a mixture of hexamethylphosphoramide (80 mL) and benzene (120 mL) followed by addition of sodium hydride (4.5 g, 60 % in oil) over 6 h. The blue mixture was stirred for 7 d at room temperature and the excess of hydride was destroyed with 50 % aq. iPrOH (5 mL). The benzene was evaporated and the mixture was diluted with a mixture of methanol (150 mL) and water (5 mL). The precipitate was filtered off the next day and purified by chromatography over alumina. Elution with dichloromethane containing ethyl acetate (up to 10 %) gave the product (14.7 g, 64 %) as orange needles, sufficiently pure for subsequent reactions.

General QA Condensation Methods
Method A (for Bis-condensed Products) 5,12-Dioctylquinacridone (1b) (0.54 g, 1 mmol) was suspended in DCE (15 ml) containing DMF (2 drops) and POCl 3 (3 mmol), and the mixture was stirred at 608C for 2 h. The substrate (4 mmol) and 2,4,6-collidine (8 mmol) were added to the clear green solution at room temperature and the reaction was continued for 48 h. The mixture was added to aqueous citric acid (15 %, 100 mL), stirred for 20 min, and extracted with ethyl acetate (2 Â 50 mL). The organic phase was separated, the volatile components removed by evaporation, and the residue was chromatographed over silica gel.

Method B (for Mono-condensed Products)
The green reactive intermediate was prepared from the quinacridone (1 mmol) as in Method A above. The substrate (2 mmol) and 2,4,6-collidine (8 mmol) were added to the clear green solution at room temperature and the reaction was continued for 2 h. The mixture was worked up as in Method A.

Method C (Malononitrile Mono-condensation Product 3b
Coupling with a Second Component) A mixture of the malononitrile monocondensation product (3b, 0.58 g, 1 mmol), DCE (15 ml), DMF (2 drops), and POCl 3 (0.45 g, 3 mmol) was stirred at 608C for 2 h and then cooled to room temperature. The nucleophilic component (2 mmol) was added followed by triethylamine (6 mmol) and the mixture was stirred for 24 h at room temperature. The mixture was worked up as in Method A.

Method D (All Components Combined at the Start of Reaction)
A stirred mixture of the 5,12-dioctylquinacridone (1 mmol), DCM (20 ml), and POCl 3 (0.6 g, 4 mmol) at 08C was treated with the nucleophilic component (4 mmol) added in one portion followed by disopropylethylamine or 2,4,6-collidine (10 mmol), and allowed to react overnight at room temperature. The next day the mixture was worked up as in Method A.

Supplementary Material
1 H and 13 C NMR spectra of characterised materials are available on the Journal's website.

Conflicts of Interest
The authors declare no conflicts of interest.