Platinum-Catalyzed Regioselective Dehydrogenative Homocoupling of Dimethyl Phthalate for the Direct Synthesis of Sym-BPTT

Abstract

A novel synthesis method for sym-BPTT, an important synthetic intermediate for electronic materials, is reported. This method features a platinum-catalyzed regioselective dehydrogenative homocoupling of dimethyl phthalate. Unlike previous palladium-catalyzed methods, this reaction proceeds with high regioselectivity without requiring any ligands when an appropriate reoxidation system is employed, thus simplifying the purification process. While the yield is currently lower than that of existing methods, this study demonstrates the potential of platinum catalysts for dehydrogenative homocoupling reactions. A preliminary continuous-flow system for this homocoupling reaction was also developed, although the yield was low, and regioselectivity was not well-controlled.

Keywords

Platinum catalyst - Dehydrogenative homocoupling - Continuous-flow synthesis

Significance

This work reports a novel platinum-catalyzed regioselective dehydrogenative homocoupling of dimethyl phthalate to produce sym-BPTT, a key intermediate for electronic materials. Notably, this method eliminates the need for ligands, unlike previous palladium-catalyzed approaches, simplifying the synthesis and purification process. Although the current yield is lower than that of existing methods, this study highlights the potential of platinum catalysts for dehydrogenative homocoupling reactions and opens new avenues for further development.

Introduction

Multisubstituted biphenyls are essential compounds used as drug substances, electronic materials, and other highly functionalized materials, as well as their synthetic intermediates. Among these, 3,3′,4,4′-biphenyltetracarboxylic acid tetramethyl ester (3,3′,4,4′-BPTT, sym-BPTT, 2) is a key intermediate in the synthesis of the polyimide monomer 3,3′,4,4′-biphenyltetracarboxylic acid dianhydride [1], [2]. Various catalytic systems have been explored for its production [3]. Notably, the direct Ar-H/Ar-H dehydrogenative homocoupling reaction between two dimethyl phthalates has garnered significant attention due to its green chemistry principles, as it does not require the introduction of additional functionalities such as halogen, boron, or other metallic elements [4]. However, the competitive coproduction of the unsymmetrical regioisomer 2,3,3′,4′-biphenyltetracarboxylic acid tetramethyl ester (2,3,3′,4′-BPTT, asym-BPTT, 3) pose a challenge. For example, in the palladium-catalyzed dehydrogenative homocoupling of dimethyl phthalate reported by Shiotani et al. [4d], regioselectivity was heavily influenced by the choice of ligands. Notably, 1,10-phenanthroline was found to effectively promote the selective formation of the desired sym-BPTT ([Scheme 1]).

Scheme 1 Palladium-Catalyzed Dehydrogenative Homocoupling of Dimethyl Phthalate (ref. [3d]).

Meanwhile, other noble metal-based catalysts, such as heterogeneous gold or platinum, have been explored for the synthesis of sym-BPTT. While gold preferentially formed sym-BPTT, the combined yield of both sym- and asym-BPTT was only 1% with platinum, and regioselectivity data were not reported [5]. To meet the growing demand for polyimide materials with a sustainable and cost-effective supply, it is crucial to develop an alternative method to access sym-BPTT. In this study, we focused on platinum catalysis, which has been underinvestigated in both homogeneous and heterogeneous systems, to assess its viability. Although there is substantial research on Ar-H/Ar-H dehydrogenative homocoupling with palladium catalysis, platinum-catalyzed examples are notably scarce, despite both being group 10 transition metals [6], [7]. We hypothesize that combining platinum compounds with an appropriate oxidant may be a key to achieving a catalytic cycle for this oxidative coupling reaction.

Results and discussion

Initially, the activity of typical platinum compounds was investigated for this reaction. A dehydrogenative homocoupling of 1 was conducted using Pt(PPh₃)₄ in the presence of Cu(OAc)₂･H₂O and 1,10-phenanthroline as a ligand at 190 °C in air ([Table 1], entry 1). The desired sym-BPTT was obtained in a 0.08% yield (TON = 0.05) with a regioselectivity (sym/asym) of 5.74. K₂PtCl₄ exhibited poor solubility in 1, resulting in a similar yield and regioselectivity (entry 2). Na₂PtCl₆･6H₂O, with higher solubility, improved both yield and regioselectivity, providing 0.25% sym-BPTT and a sym/asym selectivity of 7.3 (entry 3). This suggests that the initial oxidation state of platinum is not a critical factor for catalytic activity. Other metal compounds, such as nickel, cobalt, rhodium, and iridium, did not give a detectable amount of the product. Rh(CO)₂(acac) produced sym-BPTT in a 0.08% yield but with lower regioselectivity (entry 9). None of these metals surpassed the catalytic performance of palladium (entry 10). However, we observed that platinum exhibited regioselectivity comparable to or exceeding that of palladium, promoting further investigation.

Table 1 Catalyst survey on dehydrogenative homocoupling of dimethyl phthalate.^a

Entry	Metal compound	Yieldb [%] (TONc)	sym/asym b
a Conditions; Dimethyl phthalate (6.1 mmol), catalyst (0.061 mmol), Cu(OAc)2･4H2O (0.61 mmol), 1,10-phenantholine･H2O (Phen, 0.061 mmol), air, 190 °C, 2 h. bDetermined by GC analysis. cTON was calculated from the total yield of sym-BPTT and asym-BPTT. dNot detected.
1	Pt(PPh3)4	0.08 (0.05)	5.7
2	K2PtCl6	0.09 (0.06)	5.8
3	Na2PtCl6•6H2O	0.25 (0.15)	7.3
4	Ni(OAc)2•4H2O	NDd	–
5	Ni(acac)2•nH2O	NDd	–
6	Co(OAc)2•4H2O	Trace	–
7	Co(acac)2	Trace	–
8	Ir(acac)3	Trace	–
9	Rh(CO)2(acac)	0.08 (0.05)	2.2
10	Pd(OAc)2	3.9 (2.2)	6.3

As mentioned above, the addition of 1,10-phenanthroline is known to favor sym-BPTT formation [8] and is presumed to contribute to the comparable selectivity observed in the platinum system. However, a more detailed investigation of ligand effects in the platinum system is warranted, given the numerous unknowns. We then screened various ligands listed in [Table 2]. Surprisingly, the absence of ligands led to increased yield and regioselectivity (entry 1). While the effect of 1,10-phenanthroline (A) in palladium catalysis has been studied extensively using synchrotron radiation [9], suggesting a significant contribution to the cis-configuration of PdAr₂ species, the deviation from this trend in the present platinum system indicates the existence of other effective ligands, potentially even monodentate ones. 2,2′-Bipyridyl (B) largely suppressed activity, whereas terpyridine (C) showed a trend similar to 1,10-phenanthroline (entries 3–4). A multidentate bipyridine-diol ligand, known as Bolm’s ligand (D) [10], improved regioselectivity (entry 5), suggesting that nitrogen was not essential as a donor element. We then tested phosphorus-based ligands, including diphenylphosphine oxide (E), 1,3-di-tert-butyl-1,3,2-diazaphospholidine 2-oxide (F), 4,4,5,5-tetramethyl-1,3,2-dioxaphosphoran 2-oxide (G), and 5,5-dimethyl-1,3,2-dioxaphosphorinane 2-oxide (H) (entries 6–9). These generally resulted in similar or lower yields compared to ligand-free conditions. Bidentate phosphine oxide-based ligands tended to increase regioselectivity, albeit with similar or lower yields. For example, 1,1′-ferrocenbis(diphenylphosphine oxide) (K) yielded sym-BPTT in 0.93% with a high regioselectivity of 30:1 (entry 12). Testing β-diketone and diol ligands revealed their effectiveness for regioselectivity but at the expense of yield (entries 14–19).

Table 2 Screening of ligands.^a

Entry	Ligand	Yieldb [%] (TONc)	sym/asym b	Entry	Ligand	Yieldb [%] (TONc)	sym/asym b
a Conditions; Dimethyl phthalate (6.1 mmol), catalyst (0.061 mmol), Cu(OAc)2･4H2O (0.61 mmol), ligand (0.061 mmol, otherwise noted), air, 190 °C, 2 h [8]. bDetermined by GC analysis. cTON was calculated from the total yield of sym-BPTT and asym-BPTT. dNot detected. e2 mol % of ligand was used. fPerformed with 0.073 mmol of K2CO3.
1	A	0.25 (0.15)	7.3	11	J	1.7 (0.91)	24
2	without	2.0 (1.0)	12	12	K	0.93 (0.50)	31
3	B	NDd	7.3	13	L	1.7 (0.92)	26
4	C	0.26 (0.16)	4.6	14f	M	0.87 (0.48)	19
5	D	0.97 (0.53)	25	15f	N	0.75 (0.43)	27
6	E e	Trace	–	16f	O	0.91 (0.51)	22
7	F e	0.79 (0.41)	15	17f	P	1.5 (0.85)	16
8	G e	2.0 (1.1)	7.5	18f	Q	0.62 (0.36)	20
9	H e	1.9 (1.1)	12	19f	R	0.92 (0.47)	22
10	I	1.6 (0.81)	20

Based on the above investigation, several ligands were identified that improved regioselectivity compared to the Pd/A system. However, all led to decreased activity compared to the ligand-free conditions. Ligands I, J, and L showed significant potential, as they markedly enhanced selectivity with only a minor reduction in yield. Despite this potential, the general trend of lowering yield upon ligand addition remained a critical challenge. Therefore, we chose to further investigate reoxidation systems under ligand-free conditions, recognizing the potential for future discovery of more efficient ligands.

Given the limited information on the reduction/reoxidation catalytic cycle for platinum catalysis [11], we hypothesized that certain metal compounds or oxidants could facilitate the reoxidation of Pt(0) to Pt(II), or Pt(II) to Pt(IV). Focusing on air as the terminal oxidant, we investigated the effect of various metal compounds, as well as organic and inorganic oxidants, to enhance catalyst turnover ([Table 3]). CuCl₂, Cu(acac)₂, FeCl₃, MnO₂, AgOAc, hydrogen peroxide, Oxone®, 1,4-benzoquinone, and PhI(OAc)2 were tested as alternatives for Cu(OAc)₂·H₂O; however, all of these resulted in a significant decrease in the yield of sym-BPTT (entry 1 vs. 2-10). The insufficient performance of two Cu(II) compounds suggests that copper carboxylates are essential for this reaction. Consequently, we screened a range of copper carboxylates. Among Cu(OAc)₂, Cu(OCOEt)₂ (entry 12), Cu(OPiv)₂ (entry 13), and Cu(OCOnPr)₂ (entry 14), Cu(OCOEt)₂ afforded sym-BPTT with the highest yield, turnover number (TON), and regioselectivity, although a clear trend among the carboxylates was not observed.

Table 3 Effects of reoxidation system.^a

Entry	Oxidation reagent	Yieldb [%] (TONc)	sym/asym b
a Conditions: Dimethyl phthalate (6.1 mmol), Na2PtCl6•6H2O (0.061 mmol), oxidation reagent (0.61 mmol, otherwise noted), air, 190 °C, 2 h [8]. bDetermined by GC analysis. cTON was calculated from the total yield of sym-BPTT and asym-BPTT. d30 mol % Reagent was employed. eNot detected.
1	Cu(OAc)2･H2O	2.0 (1.0)	12
2	CuCl2	0.21 (0.06)	0.45
3	Cu(acac)2	0.25 (0.15)	2.8
4	FeCl3	0.03 (0.03)	0.43
5	MnO2	0.34 (0.24)	2.9
6d	AgOAc	0.66 (0.37)	18
7d	H2O2	NDe	–
8d	Oxone®	NDe	–
9d	1,4-Benzoquinone	NDe	–
10d	PhI(OAc)2	Trace	–
11	Withiout	0.10 (–)	4.2
12	Cu(OCOEt)2	4.0 (1.9)	20
13	Cu(OPiv)2	1.1 (0.59)	18
14	Cu(OCO n Pr)2	2.8 (1.7)	16

Acetic acid has occasionally been employed as a promoter in dehydrogenative homocoupling reactions [5]. Therefore, we anticipated improvements in the current Na₂PtCl₆/Cu(OAc)₂ or Na₂PtCl₆/Cu(OCOEt)₂ systems ([Table 4]). While the addition of AcOH to the Na₂PtCl₆/Cu(OCOEt)₂ system (previously the highest yielding) resulted in lower yields (entry 5), in the case of the Na₂PtCl₆/Cu(OAc)₂ system, the yield of sym-BPTT increased to 5%, albeit with reduced regioselectivity (entry 3). This decrease in regioselectivity could be mitigated by using a larger amount (30 mol %) of Cu(OAc)₂ and Na₂PtCl₆ (3 mol %) than previously employed (entries 4 and 5). Based on these results, we conclude that the Na₂PtCl₆/Cu(OAc)₂/AcOH system under air is optimal [12].

Table 4 Effect of acetic acid.^a

Entry	Cu salt, AcOH	Yieldb [%] (TONc)	sym/asym b
a Conditions: Dimethyl phthalate (6.1 mmol), Na2PtCl6•6H2O (0.061 mmol), Cu salt (0.61 mmol), AcOH (0.1 ml, otherwise noted), air, 190 °C, 2 h [8]. bDetermined by GC analysis. cTON was calculated from the total yield of sym-BPTT and asym-BPTT. d1.8 mmol (30 mol %) of Cu(OAc)2•H2O was employed. e0.18 mmol (3 mol %) of Na2PtCl6•6H2O was employed fIsolated yield. gDetermined by 1H NMR analysis.
1	A, w/o AcOH	2.0 (1.0)	12
2	B, w/o AcOH	4.0 (1.9)	20
3	A, w/ AcOH	5.2 (2.8)	8.7
4	A d, w/ AcOH	5.4 (2.8)	14
5	A d,e, w/ AcOH	3.2 (0.54)	18
		3.3f (0.55)	17g
6	B, w/ AcOH	2.6 (1.5)	7.4

Concurrently, we explored a preliminary continuous-flow system for this homocoupling reaction [13]. Continuous-flow synthesis of sym-BPTT offers substantial industrial advantages, including reduced spatial and labor demands, scalable productivity, and smaller manufacturing footprints. As a foundational study, we chose to employ a heterogeneous platinum catalyst immobilized on silica gel using a standard impregnation method. In this system, a copper salt dissolved in acetic acid was pumped along with 1 (acetic acid:1 = 1:1) into a column reactor packed with the catalyst. Recognizing the importance of air for the catalytic cycle, it was introduced into the reactor using an air compressor, though without precise mass flow control. The feedstock solution and air were introduced separately through a tube-in-tube type column inlet unit. A circulation-flow system was implemented, with the outlet solution collected in a reservoir flask ([Scheme 2]). After 16 h of circulation, we detected sym-BPTT in the solution, albeit in low yield and with poorly controlled regioselectivity. Although we were unable to replicate the batch reaction results in the initial continuous-flow setup, we are actively working to improve performance through the examination of platinum-immobilized catalysts and ligands.

Scheme 2 Examination of continuous-flow synthesis of sym-BPTT.

Conclusions

In conclusion, we have investigated the dehydrogenative homocoupling of dimethyl phthalate as a direct synthetic route to sym-BPTT, a crucial synthetic intermediate for electronic materials. This study specifically focused on the potential of platinum catalysis, which has been underexplored in the literature, including patent sources. We discovered that the reaction proceeds with high regioselectivity in the absence of additional ligands when an effective reoxidation system comprising copper carboxylate, acetic acid, and air is employed. Future work will involve a detailed study of the reaction mechanism, expanding the substrate scope, and implementing our findings in the continuous synthesis of this valuable compound through intermolecular Ar-H/Ar-H dehydrogenative coupling reactions catalyzed by platinum.

Contributors’ Statement

Data collection: S. Ishikawa; Design of the study: H. Ishitani, S. Kobayashi, K. Iseki; Analysis and interpretation of the data: S. Ishikawa, H. Ishitani, S. Kobayashi; Drafting the manuscript: S. Ishikawa, H. Ishitani. S. Kobayashi; Critical revision of the manuscript, H. Ishitani, S. Kobayashi.

Conflict of Interest

The authors declare that they have no conflict of interest.

• References

• 1a Numata S, Oohara S, Fujisaki K, Imaizumi J. Kinjo N. J. Appl. Polym. Sci. 1986; 31: 101
• 1b Saraf RF, Tong H.-M, Poon TW, Silverman BD, Ho PS, Rossi AR. J. Appl. Polym. Sci. 1992; 46: 1329
• 1c Chen ST, Wagner HH. J. Electron. Mater. Lett. 1993; 22: 797
• 1d Hasegawa M, Matano T, Shindo Y, Sugimura T. Macromolecules 1996; 29: 7897
• 2a Ree M, Goh WH, Park J.-W, Lee M.-H, Rhee S.B. Polym. Bull. 1995; 35: 129
• 2b Yue Z, Cai Y.-B, Xu S. J. Polym. Res. I 2014; 21: 463
• 2c Wang Y, Tao L, Wang T, Wang Q. RSC Adv. 2015; 5: 101533
• 2d Ishikawa A, Mitsui H, Tanaka Y, Sasaki K. JP Patent No. 4048689 2007
• 2e Inoue H. Kobunshi 1997; 46: 566
• 3a Kuroboshi M, Waki Y, Tanaka H. Synlett 2002; 4: 637
• 3b Kuroboshi M, Waki Y, Tanaka H. J. Org. Chem. 2003; 68: 3938
• 3c Wu XE, Gao CL, Meng X, Zhang SB, Gao LX. Chin. Chem. Lett. 2004; 15: 787
• 3d Wu X.-L. Huaxue Shiji 2017; 39: 1015
• 4a Itatani H, Yoshimoto H. J. Org. Chem. 1973; 38: 76
• 4b Yoshimoto H, Itatani H. J. Catal. 1973; 31: 8
• 4c Shiotani A, Yoshikiyo M, Itatani H. J. Mol. Catal. 1983; 18: 23
• 4d Shiotani A, Itatani H, Inagaki T. J. Mol. Catal. 1986; 34: 57
• 5 Ishida T, Aikawa S, Mise Y, Akebi R, Hamasaki A, Honma T, Ohashi H, Tsuji T, Yamamoto Y, Miyasaka M, Yokoyama T, Tokunaga M. ChemSusChem 2015; 8: 695
• 6 Yang Y, Lan J, You J. Chem. Rev. 2017; 117: 8787 and references therein
• 7a Zhang X.-W, Shen S.-C, Hidajat K, Kawi S, Yu LE, Simon Ng KY. Catal. Lett. 2004; 96: 1
• 7b Maphoru MV, Heveling J, Pillai SK. Eur. J. Org. Chem. 2016; 331
• 7c Maphhoru MV, Pillai SK, Heveling J. J. Catal. 2017; 348: 47
• 7d Matsumoto K, Yoshida M, Shindo M. Angew. Chem., Int. Ed. 2016; 55: 5272
• 7e Fujimoto S, Matsumoto K, Iwata T, Shindo M. Tetrahedron Lett. 2017; 58: 973
• 7f Nabavizadeh SM, Hosseini FN, Park C, Wu G, Abu-Omar MM. Dalton Trans. 2020; 49: 2477
• 8 We confirmed that contamination amount of Pd in 9.1 mg of Na₂PtCl₆ was under detection limit. We also tested the reaction using 1.0 mol % of Pd(OAc)₂ as a catalyst in the presence of 10 mol % of Cu(OAc)₂ without using ligands: this resulted in 2.5% yield with 0.40 of sym/asym regioselectivity
• 9a Hirano M, Sano K, Kanazawa Y, Komine N, Maeno Z, Mitsudome T, Takaya H. ACS Catal. 2018; 8: 5827
• 9b Kanazawa Y, Mitsudome T, Takaya H, Hirano M. ACS Catal. 2020; 10: 5909
• 10a Maphoru MV, Heveling J, Pillai SK. ChemPlusChem 2014; 79: 99
• 10b Maphoru MV, Heveling J, Pillai SK. Eur. J. Org. Chem. 2016; 2016: 331
• 10c Maphoru MV, Pillai SK, Heveling J. J. Catal. 2017; 348: 47
• 10d Fujimoto S, Matsumoto K, Iwata T, Shindo M. Tetrahedron Lett. 2017; 58: 973
• 10e Zhang X.-W, Shen S.-C, Hidajat K, Kawi S, Yu LE, Simon Ng KY. Catal. Lett. 2004; 96: 87
• 10f Nabavizadeh SM, Hosseini FN, Park C, Wu G, Abu-Omar MM. Dalton Trans. 2020; 49: 2477
• 10g Wagner AM, Hickman AJ, Sanford MS. J. Am. Chem. Soc. 2013; 135: 15710
• 10h Chen M, Rago AJ, Dong G. Angew. Chem., Int. Ed. 2018; 57: 16205
• 11a Ishikawa S, Hamada T, Manabe K, Kobayashi S. Synthesis 2005; 13: 2176
• 11b Bednářová E, Malatinec S, Kotora M. Molecules 2020; 25: 958
• 12 A typical procedure for a batch reaction was described using Entry 3 of Table 4, for example: To an autoclave containing a magnetic stirring bar, dimethyl phthalate (6.1 mmol), Na₂PtCl₄ (0.061 mmol), Cu(OAc)₂ (0.61 mmol), and acetic acid (1.0 mL) were added, and the autoclave was heated at 190 °C to start the reaction. After 2 h, the reaction mixture was diluted with EtOAc and filtered through celite. The obtained solution was then concentrated and subjected to analysis by GC-FID
• 13 Typical procedure for a continuous-flow reaction: A thoroughly mixed composite of Pt/SiO₂ (0.4 g, Pt 0.25 mmol) and Celite was packed into a stainless steel column (Φ10 × 100 mm) equipped with a filter and a column head at one end. A double-inlet column head, capable of independently introducing gas and liquid into the reactor, was attached to the other end of the column. Dimethyl phthalate (110 mmol), copper(II) acetate monohydrate (1.6 mmol), and acetic acid (25 mL) were combined and stirred at 90 °C to form a homogeneous substrate solution, and the reservoir temperature was maintained at 90 °C using a plate heater. Toluene was initially fed into the reactor. Subsequently, dry air flow was initiated, and the reactor was heated to 200 °C. The feed was then switched from toluene to the substrate solution, and the outflow was connected to the reservoir to create a recirculation system. This recirculation system was driven for 16 h, and a sample of the solution was taken to be subject to analysis by GC-FID

Publication History

Received: 21 October 2024

Accepted after revision: 10 March 2025

Accepted Manuscript online:
12 March 2025

Article published online:
02 April 2025