Synthesis, Post‐Functionalization and Properties of Diphosphapentaarenes

Abstract

Linearly‐fused polyarenes are an important class of compounds with high relevance in materials science. While modifying the shape and size represents a common means to fine‐tune their properties, the precise placement of heteroatoms is a strategy that is receiving an increasing deal of attention to overcome the intrinsic limitations of all‐carbon structures. Thus, linearly‐fused diphosphaarenes recently emerged as a novel family of molecules with striking optoelectronic properties and outstanding stability. However, the properties of diphosphaarenes are far from being benchmarked. Herein, we report the synthesis, phosphorus post‐functionalization and properties of new diphosphapentaarene derivatives. We describe their synthetic limitations and unveil their potential for optoelectronic applications.

Shortening the core. Diphosphapentaarenes, the shorter analogs of diphosphahexaarenes, are stable linearly fused systems containing two six‐membered phosphorus heterocycles. They are fluorescent and their phosphorus atoms can be readily post‐functionalized to modify their optoelectronic properties.

Introduction

Polyaromatic molecules (PAs) are an important class of molecules because they are pivotal not only for the development of organic electronics but also for the fundamental understanding of superior carbon nanoforms: graphene and carbon nanotubes. Typical means to control the properties of PAs consist in perturbing the conjugation from the carbon p‐orbitals by modifying the size and shape of the structure.[ ¹ , ² , ³ , ⁴ , ⁵ ] The incorporation of heteroatoms into PAs emerged in the last years as an elegant strategy to overcome the limitations from the intrinsic properties of carbon atoms.[ ¹ , ³ , ⁶ , ⁷ , ⁸ , ⁹ , ¹⁰ ] Thus, a myriad of heteroatomic derivatives have been developed whose features – luminescence, coordinating behavior, electrochemical properties, etc. – clearly eclipse those of all‐carbon architectures.[ ¹¹ , ¹² , ¹³ , ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ , ¹⁹ , ²⁰ ]

The synthesis of acenes; i. e. linearly‐fused hexagonal carbon rings, and their derivatives have received a great deal of attention in the last decades. Since their first discovery by Clar in the 1930s, ^[21] pentacenes evolved from simple curiosities to become the hallmark of hole transporting semiconductors for organic field‐effect transistors.[ ²² , ²³ , ²⁴ ] The development of synthetic methods for the precise placement of heteroatoms into the framework of pentacenes, however, still represents a major challenge nowadays. The importance of this matter is reflected, for example, in the hexacenes; i. e. pentacenes analogs, which present a relative stability. The inclusion of two phosphorus atoms to provide the linearly‐fused diphosphahexaarenes (compounds A1–A3 and B, Figure 1a) contributed to enhance the photo‐ and chemical stability of hexacenes, led to high fluorescence quantum yields (QY) and a series of facile, chemical post‐functionalizations through which modify both their optical properties and their molecular organization.[ ²⁵ , ²⁶ ] Thus, linearly‐fused arenes containing phosphorus centers have recently emerged as novel class of compounds with uncharted possibilities.[ ²⁶ , ²⁷ , ²⁸ , ²⁹ , ³⁰ , ³¹ ] Notwithstanding, the impact of embedding phosphorus atoms into linearly‐fused pentaarenes has not been yet investigated.

Figure 1.

a) Reported syn‐(compounds A1–A3) and anti‐(compound B) diphosphahexaarene derivatives.[ ²⁵ , ²⁶ ] b) Targeted diphosphapentaarenes C and 1–3.

As a part of our program on the development of π‐extended systems based on six‐membered phosphorus heterocycles, we targeted the synthesis of diphosphapentaarene derivatives in syn (compound C) and anti (compounds 1–3) arrangements (Figure 1), i. e. with the phosphorus atom at the same or opposite sides of the pentaarene core, respectively. Our ultimate goal was to benchmark the properties of the linearly‐fused phosphaarenes. Thus, we systematically investigated the impact of reducing size of the π‐core into the physicochemical properties of the linearly‐fused arenes and explored the impact of modifying the phosphorus atom with oxygen, sulfur and methyl – compounds 1, 2 and 3, respectively (Figure 1) – into the optoelectronic properties of the diphosphapentaarenes. In particular, we targeted to assess whether the physicochemical properties could be modified compared to those of the diphosphahexaarenes, while keeping the access to post‐functionalization reactions.

Results and Discussion

To prepare compound C, we designed strategy I (Scheme 1). It consists in a double Stille cross‐coupling between the triflate derivative I a and stannane I b to yield the tetrabrominated compound D, i. e. the suitable reactant for the final double phosphorus cyclization via lithiation and reaction with dichlorophenylphosphane. ^[25] Compound I a was expected to allow for the selective coupling of I b via the triflate moieties. However, the reaction was not selective. We only detected monocoupled derivatives to both the triflated and the brominated positions. Modification of the conditions (see Supporting Information, Table S1) did not lead to the targeted molecule D. To improve the selectivity, we used the 1,3‐dibromo‐4,6‐diiodobenzene II a instead of the triflate derivative I a via strategy II (Scheme 1); it involves a Suzuki‐Miyaura cross coupling but between II a and the pinacolborane derivative II b. Unfortunately, this approach was also unselective and only afforded monocoupled products.

Scheme 1.

Synthetic routes I and II for the synthesis of diphosphapentaarene C.

In views of the difficulties to prepare D, we turned to the synthesis of the anti derivative 1 (Figure 1). To this end, we employed the synthetic route depicted in Scheme 2. It is based on the Suzuki‐Miyaura coupling of the 1,4‐dibromo‐3,6‐diiodobenzene with the pinacol derivative II b. Here, the 1,4‐dibromo‐3,6‐diiodobenzene underwent selective double coupling to yield E (Scheme 2) (see Supporting Information for details). Finally, the treatment of E with ^t BuLi followed by the reaction with dichlorophenylphosphane and the in situ oxidation of the trivalent intermediate with hydrogen peroxide afforded compound 1. It is important to remark that, as seen for the diphosphahexaarenes, the cyclization reaction led to a mixture of cis‐ and trans‐isomers (Scheme 2), i. e. with the phenyl substituents of the phosphorus atoms at the same or opposite plane of the molecule. Thus, the reaction with hydrogen peroxide led to isomers 1 a and 1 b. Compound 1 did not show any sign of decomposition; this allowed 1 a and 1 b to be isolated via regular column chromatography. As well known for π‐systems with two phosphorus heterocycles,[ ¹⁸ , ²⁵ , ²⁶ , ³² ] the trans‐derivatives appear to be less polar than the cis‐analog and thus elute first from column chromatography. This is further confirmed by the ³¹P NMR; the trans‐molecules exhibit a singlet at lower fields with respect to the cis‐ ones. In particular, 1 a displays a ³¹P NMR at 8.9 ppm, while 1 b at 9.4 ppm (Figure S1, Table S2).

Scheme 2.

Synthetic route designed to afford diphosphapentaarene derivatives 1–3.

On the other hand, the reaction of the trivalent phosphapentaarene intermediate with elemental sulfur, even at high temperatures, did not afford derivative 2 (Scheme 2). This was surprising since the reaction of trivalent phosphorus center with elemental sulfur at room temperature is commonly a quantitative reaction to provide P=S derivatives.[ ²⁶ , ³³ , ³⁴ , ³⁵ , ³⁶ ] Nevertheless, this synthetic challenge was overcome by reacting 1 with the Lawesson's reagent to afford sulfur derivative 2 as a mixture of isomers. Again, the stability of 2 allowed us to separate the isomers 2 a from 2 b by column chromatography. Compound 2 a is characterized by a ³¹P NMR signal at 22.1 ppm and derivative 2 b by a singlet at 22.3 ppm (Figure S1, Table S2).

Another remarkable fact was that 1 showed reluctance to get reduced to its trivalent form. Standard reduction with HSiCl₃ did not afford the desired compound. Refluxing 1 in THF with more than 150 equivalents of HSiCl₃ during 8 days did not change the result. Again, this is in stark contrast to diphosphahexaarenes; the reduction of the latter was completed within 3 days after addition of 90 equivalents of HSiCl₃. ^[26] Finally, we could obtain the diphosphapentaarene derivative 3 with two quaternized phosphorus centers necessarily by in situ treatment of the trivalent intermediate with methyl triflate at room temperature. The separation of the isomers via column chromatography turned to be tedious; it allowed to isolate the pure isomer 3 b; however, 3 a could only be obtained in traces. ³¹P NMR signal of 3 b appeared at 1.2 ppm (Figure S1, Table S2).

The solubility of PAs commonly decreases when increasing the surface of the π‐conjugated system due to π‐stacking. Surprisingly, the solubility of both the precursor E (Scheme 2) and the diphosphapentaarenes was much lower than for the diphosphahexaarenes. Whereas the cis‐isomer 1 a was soluble in chloroform and dichloromethane, the trans‐isomer 1 b only dissolved in 1,2‐dichloroethane upon heating at 140 °C and was partially soluble in chlorobenzene, 1,2‐dichlorobenzene, 1,4‐dioxane, and 1,1,2,2‐tetrachloroethane at room temperature. In turn, the sulfide 2 showed a partial solubility in chloroform, DCM, EtOAc, Et₂O, EtOH, MeOH and MeCN and was insoluble in pentane and hexane. Again, the cis‐isomer 2 a showed a higher solubility than the trans‐isomer 2 b. The quaternization of the phosphorus atoms significantly improved the solubility. Compound 3 dissolved in acetone, acetonitrile, MeOH and water. It also showed a partial solubility in DCM, CHCl₃ and EtOAc, and was insoluble in pentane, toluene and Et₂O.

First insight into the electronic properties of 1–3 came from DFT calculations at the B3LYP/6‐311+G(d) level of theory (Figures 2, S2–S4 and Tables 1 and S3). Structurally, the disphosphapentaarenes present a distorted core due to the repulsion from the hydrogen atoms of the central core (see Figure 2a for molecules 1 a, 2 a and 3 a). This results in a framework with a certain degree of helicity. In contrast to the diphosphahexaarene oxide analog (compound B, Figure 1), ^[25] all diphosphapentaarenes derivatives present a contribution of the phosphorus atoms in the LUMO; the sulfur atoms are included in the HOMO of derivatives 2 (Figures 2b, S2 and S3). Regarding the energy of the molecular levels (Tables 1 and S3), in line with the diphosphahexaarene derivatives, the differences between the cis‐ and trans‐derivatives are negligible. Overall, the optical band gap of the diphosphahexaarenes[ ²⁵ , ²⁶ ] is smaller than for the diphosphapentaarenes (Table S3) due to their larger π‐extended system. However, both HOMO and LUMO are relatively lower in energy for the diphosphapentaarenes in line with lower intramolecular electronic repulsions. In particular, the HOMO is reduced from −6.08 eV for 1 and 2, to −6.92 eV for 3 (Figure S4, Tables 1 and S3). In turn, the LUMO decreases from −2.58 eV for 1, −2.62 eV for 2 to −3.51 eV for 3 (Figure S4, Tables 1 and S3). In other words, the electron accepting properties follow the trend: 1<2<3. This is not surprising taking into consideration the dicationic nature of 3. Thus, the electron accepting properties of the disphosphapentaarenes can be readily modulated by the functionalization of the phosphorus atom. This was further corroborated experimentally by electrochemical techniques. In common with other π‐systems containing six‐membered phosphorus heterocycles; i. e. phosphaphenalenes, diphosphapyrenes and diphosphahexaarenes, the diphosphapentaarenes present oxidation and reduction processes (here both irreversible) with the exception of the quaternized derivative 3 (Table S4). In particular, compound 1 and 2 display an oxidation at 1.89 and 1.9 eV and a first reduction at −1.37 and −1.35 eV, respectively (Table S4, Figure S5). In turn, the quaternized 3 presents only two reversible reduction processes at −0.89 and −1.27 eV (Table S4, Figure S5), which convert it in a remarkable electron accepting molecule. All‐in‐all, the diphosphapentaarenes are thus better electron accepting molecules although worse electron‐donating systems compared to the diphosphahexaarenes; the latter present an oxidation potential at 1.79 eV and a first reduction process at −1.41 eV ^[25] (Table S4, Figure S5).

Figure 2.

a) Optimized structures of 1 a, 2 a, and 3 a. Red lines indicate repulsion between hydrogen atoms. b) Optimized structure and HOMO/LUMO levels of 1 a, 2 a, and 3 a obtained by DFT calculations (see Supporting Information for details). c) Steady‐state spectroscopic properties of 1 a, b, 2 a, b, and 3 b from dichloromethane solutions (*aqueous solutions).

Table 1.

Photophysical properties of diphosphapentaarenes 1–3.

Compound	Abs. λ max [nm]	Em. λ max [nm]	Log ϵ	τ 1; τ 2; τ 3 [ns][c]	Φ [%][d]	HOMO [eV][e]	LUMO [eV][f]	Eg [eV][g]
Diphospha‐hexaarene oxide [25]	373, 388(sh)	420, 442	4.4	6.2 (100 %)	80	−5.842	−2.408	3.43
1a [a]	347, 376, 395	408, 430, 455 (sh)	4.48	0.60 (7 %), 1.97 (93 %)	37	−6.08	−2.582	3.50
1b [a]	347, 376, 395	409, 430, 455 (sh)	4.55	0.64 (12 %), 1.66 (88 %)	28	−6.077	−2.588	3.49
2a [a]	354, 379, 397	408, 430, 455(sh)	4.38	0.06 (35 %), 1.74 (62 %), 4.55 (3 %)	0.9	−6.083	−2.605	3.48
2b [a]	354, 378, 397	409, 430, 455(sh)	4.47	0.08 (21 %), 1.75 (74 %), 4.44 (5 %)	0.7	−6.076	−2.606	3.47
3b [a]	352, 396, 415	435, 454, 488(sh)	4.06	2.56 (90 %), 4.50 (10 %)	31	−6.922	−3.514	3.41
3b [b]	350, 391, 407	432, 448, 483(sh)	4.10	0.88 (5 %), 3.09 (95 %)	76	n.d.	n.d.	n.d.

^[a] Recorded from DCM solutions. ^[b] Recorded from aqueous solutions. ^[c] Fluorescence lifetimes. Relative amplitude in brackets. ^[d] Fluorescence quantum yields relative to quinine sulfate in 0.1 M H₂SO₄, Φ=0.54. ^[e] Calculated highest occupied molecular orbital energy. ^[f] Calculated lowest occupied molecular orbital energy. ^[g] Energy band gap. Note that the diphosphahexaarene oxide is the cis/anti isomer B. ^[25]

On the other hand, as a result of the changes in the frontier molecular orbitals, DFT calculations anticipated that the optical band gap decreases from 3.5 eV for 1, 3.48 eV for 2 and 3.41 eV for the dicationic compound 3 (Table 1 and S3). This was confirmed by steady‐state spectroscopy (Figures 2c, S6–S12). Indeed, the difference in the absorption spectra between 1 and 2 are very small and 3 displays relatively red‐shifted absorption and emission maxima (Figure 2c). In particular, the absorption spectrum of 1 maximizes at 347, 376 and 395 nm and the one of 2 maximizes at 354, 378 and 397 nm. They are red‐shifted compared to the diphosphahexaarenes (Table 1); i. e. decreasing the size of the π‐system leads to a relative smaller optical band gap. Note that the absorption properties of both cis‐ and trans‐isomers are identical. This is in line with the observations found for the diphosphahexaarenes.[ ²⁵ , ²⁶ ] In 3, the absorption spectrum is red‐shifted to maxima at 352, 396 and 415 nm in DCM, which are surprisingly hypsochromically shifted to 350, 391 and 407 nm in water (Figure 2c). This reflects a negative solvatochromism in which the ground state is more polar than the excited state.

In line with the absorption properties, the emission spectra of the oxide and sulfide derivatives 1 and 2 are alike; both maximize at 409 and 430 nm and present a shoulder at 455 or 482 nm, respectively (Figure 2c). In turn, compound 3 displays a significantly red‐shifted emission with maxima at 435 and 454 nm and a shoulder at 488 nm in DCM; in water the absorption spectrum of 3 maximizes at 432, 448 and presents a shoulder at 483 nm (Figure 2c). Regarding the fluorescence lifetime, in contrast to the diphosphahexaarenes oxides that present a mono‐exponential decay of 6.2 ns, 1 presents a bi‐exponential decay with a main component below 2 ns and a very short one of around 0.6 ns, whereas 2 exhibits a tri‐exponential decay with lifetimes of around 4.5, 1.7 ns and a very short one below 0.1 ns (Table 1). This is indicative of a more sophisticated excited‐state deactivation mechanism, which also correlates with the fluorescence quantum yields (QYs). In general, the QYs of the disphosphapentaarenes are lower than those of the diphosphahexaarenes (Table 1).[ ²⁵ , ²⁶ ] The bi‐exponential decay of 1 leads to QYs of 37 and 28 % for 1 a and 1 b, whereas the tri‐exponential decay of 2 results in very low QYs; i. e. 0.9 and 0.7 % for 2 a and 2 b, respectively. As for other sulfur‐functionalized six‐membered phosphorus heterocyclic derivatives,[ ²⁶ , ³⁷ ] the very low QY of the sulfur‐containing phosphapentaarenes derivatives is attributed to the heavy atom effect from the sulfur atoms. Finally, compound 3 with the bi‐exponential decay in DCM (τ ₁=2.6 ns; τ ₂=4.5 ns) presents a QY of 31 %, whereas in water an also bi‐exponential decay (τ ₁=0.9 ns; τ ₂=3.1 ns) results in a remarkable QY of 76 %. This may be indicative of a potential applicability in bio‐imaging.

Conclusions

In summary, herein we report the preparation, post‐functionalization and properties of a new class of molecules; the diphosphapentaarenes. Their synthesis is based on a divergent approach from the 1,4‐dibromo‐3,6‐diiodobenzene through a selective, double Suzuki‐Miyaura coupling to the iodo groups followed by a double phosphacyclization via treatment with ^t BuLi and subsequent reaction with dichlorophenylphosphane. Diphosphapentaarene intermediates with two trivalent phosphorus centers can be easily post‐functionalized; they react quantitatively with hydrogen peroxide to render the oxidized pentavalent analogs. However, the reduction back to the trivalent states requires harsher conditions than for the diphosphahexaarene derivatives. The sulfur derivatives can be easily obtained by treatment of the oxidized derivative with the Lawesson's reagent. In turn, the dicationic species is obtained by reacting the trivalent intermediate with methyl triflate. Structurally, all derivatives present a contorted core due to the repulsion of the central hydrogen atoms and they are less soluble in organic solvents than the diphosphahexaarene analogs, although the dicationic diphosphapentaarene presents an outstanding solubility in water. All derivatives are fluorescent; nevertheless, the phosphorus functionalization strongly affects the optoelectronic properties of the diphosphapentaarene. The oxide derivatives present a significant QY of 28–37 %, while in the sulfur‐containing molecules the QY is drops below 1 %. Remarkably, the dicationic species exhibits a prominent QY of 76 % in water. All‐in‐all, the investigations described herein represent a cornerstone towards the fundamental understanding of diphosphaarene molecules. Investigations on superior phosphorus architectures are currently underway in our laboratories.

Experimental Section

Experimental details, synthetic procedures, theoretical calculations, UV/Vis spectra, electrochemical investigation and NMR spectra are included in the Supporting Information.

Conflict of interest

The authors declare no conflict of interest.

Supporting information

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.

Supporting Information

Regulska E., Hindenberg P., Espineira-Gutierrez A., Romero-Nieto C., Chem. Eur. J. 2023, 29, e202202769.

Data Availability Statement

The data that support the findings of this study are available in the supplementary material of this article.