Tetrafluorophenylene-Containing Vinylbenzyl Ether-Terminated Oligo(2,6-dimethyl-1,4-phenylene ether) with Better Thermal, Dielectric, and Flame-Retardant Properties for Application in High-Frequency Communication.

In an integrated circuit, signal propagation loss is proportional to the frequency, dissipation factor (D f), and square root of dielectric constant (D k). The loss becomes obvious as we move to high-frequency communication. Therefore, a polymer having low D k and D f is critical for copper-clad laminates at higher frequencies. For this purpose, a 4-vinylbenzyl ether phenoxy-2,3,5,6-tetrafluorophenylene-terminated OPE (VT-OPE) resin was synthesized and its properties were compared with the thermoset of commercial OPE-2St resin. The thermoset of VT-OPE shows a higher T g (242 vs 229 °C), a relatively high cross-linking density (1.59 vs 1.41 mmole cm-3), a lower coefficient of thermal expansion (55 vs 76 ppm/°C), better dielectric characteristic at 10 GHz (D k values of 2.58 vs 2.75, D f values of 0.005 vs 0.006), lower water absorption (0.135 vs 0.312 wt %), and better flame retardancy (UL-94 VTM-0 vs VTM-1 with dropping seriously) than the thermoset of OPE-2St. To verify the practicability of VT-OPE for copper-clad laminate, a laboratory process was also performed to prepare a copper-clad laminate, which shows a high peeling strength with copper foil (5.5 lb/in), high thermal reliability with a solder dipping test at 288 °C (>600 s), and the time for delamination of the laminate in thermal mechanical analysis (TMA) at 288 °C is over 60 min.

Introduction

In an integrated circuit, signal propagation speed (Vp) is inversely related to the square root of dielectric constant (Dk). The signal propagation loss (L) is directly related to the frequency, dissipation factor (Df), and square root of the dielectric constant.[1]In eq , C is a constant and f is the frequency. The propagation loss increases proportionally with the frequency. A polymer network with low Dk and Df values can reduce the signal loss in high-frequency communication. In addition, a polymer network with a low Dk value can increase signal propagation speed. Many low-dielectric polymers have been reported in the literature.[2−19] Among them, poly(2,6-dimethyl-1,4-phenylene oxide) (PPO)[8,20−30] has been extensively studied due to its special characteristics such as low polarity, high Tg, low dielectric properties, and so forth. However, PPO also has some disadvantages, for example, properties like high melting point, limited solubility, and high viscosity, leading to their limited applications in copper-clad laminate industry.

To alleviate the shortcomings of PPO, oligo(2,6-dimethyl-1,4-phenylene ether) also known as OPE was established.[20−22,26,31−38] To compensate for the loss in thermal and mechanical properties owing to the low molecular weight of OPE, it should be cured with epoxy resins by terminal hydroxyl.[39] However, secondary hydroxyl groups with high polarity would be generated by the ring-opening reaction of epoxide, sacrificing the dielectric characteristic. To balance the dielectric and thermal properties, a typical commercial bifunctional OPE with end groups (vinylbenzyl ether) was developed by Mitsubishi Gas Chemical (MGC), named OPE-2St[40,41] (Figure ). Using this resin, a self-cross-linking process could be achieved without creating polar groups, via radical initiated polymerization of vinyl groups. The makers are recommending these materials for high-frequency printed circuit boards because they have excellent thermal qualities without losing the necessary dielectric properties. However, one OPE-based material with all of the necessary features such as low dielectric constant, high glass-transition temperature, good dimensional stability, excellent thermal stability, minimal moisture absorption, high contact angle, and strong flame retardancy is rare.

Figure 1

Structures of OPE-2St and VT-OPE.

It has been reported that incorporating fluoro-containing moieties can reduce the dielectric constant and enhance hydrophobicity.[3,5,42−57] Note that hydrophobicity is important for electronic materials since the Dk value of water is about 80 at room temperature, therefore moisture absorption of a thermoset has a negative effect on dielectric constant. Among the fluoro-containing moieties, we focus on the symmetric tetrafluorophenylene moiety that will not increase the dipole moment due to an equal distribution of polarizable C–F bonds. Therefore, our strategy is to incorporate the tetrafluorophenylene moieties at both ends to enhance the overall properties of OPE thermosets. For this purpose, a 4-vinylbenzyl ether phenoxy-2,3,5,6-tetrafluorophenylene-terminated OPE (VT-OPE) resin was first prepared. We found that cured VT-OPE (C-(VT-OPE)) shows better dielectric constant, hydrophobicity, and flame retardancy than the cured commercial resin OPE-2St (C-(OPE-2St)). Not only the detailed synthesis and curing behavior of VT-OPE resin but also the thermal, moisture absorption, and flame-retardant properties of the thermoset are discussed thoroughly. To verify the practicability of VT-OPE for copper-clad laminate, a process was performed to prepare a copper-clad laminate. Table S1 shows the overall performance data of different resins reported in the field of low-dielectric constant materials at high frequencies (in the GHz range).

Experimental Section

Materials

SA90 was purchased from SABIC (Saudi Basic Industries Corporation). Vinylbenzyl ether-terminated OPE named OPE-2St was purchased from MGC. t-Butyl cumyl peroxide (TBCP) was purchased from Sigma-Aldrich. Hexafluorobenzene and 4-benzyloxyphenol from Alfa Aesar, potassium carbonate from SHOWA, and 4-vinylbenzyl chloride and palladium on activated carbon (10% Pd/C) were purchased from Acros Organics. All of the solvents are of ACS or high-performance liquid chromatography (HPLC) grades.

Synthesis of OPEs

OPEs were synthesized according to the chemistries depicted in Schemes –3, and detailed synthetic procedures are shown below.

Scheme 1

Attempted Synthesis of HT-OPE

Scheme 3

Synthesis of VT-OPE from HT-OPE

Synthesis of Pentafluorophenyl-Terminated Oligo(2,6-dimethyl-1,4-phenylene ether) (PF-OPE)

SA90 (2.0 g, 1.25 mmol), hexafluorobenzene (1.16 g, 6.25 mmol), K2CO3 (0.43 g, 3.125 mmol), and DMAc (15 g) were added into a three-neck reactor equipped with a thermocouple, magnetic stirrer, and nitrogen flow. The solution was then kept for stirring at 80 °C for 36 h; after cooling to 25 °C, the crude suspension was poured into methanol/water (1/1, v/v), and the turbid solution was stirred overnight. The precipitate was filtered and dried in a vacuum oven for 12 h at 60 °C to get an off-white powder (yield = 86%). 1H NMR (ppm, CDCl3, 400 MHz): δ = 1.70 (6H, H1), 2.09 (Ar–CH3, H3, H5), 6.47 (Ar–H, H4), and 6.97 (4H, H2). 13C NMR (ppm, CDCl3, 100 MHz): δ = 16.77 (C3, C5), 30.95 (C1), 114.40 (C4), 127.35 (C2), 135.61 (C8), 137.24, 138.89 (C6, C10), 139.16, and 140.80 (C7, C9). 19F NMR (ppm, CDCl3, 376 MHz): δ = −159 (F at C6, C10), −157 (F at C8), and −155 (F at C7, C9). Fourier transform infrared (FTIR) spectra (KBr, cm–1): ν = 991 (C–F), 1197, 1308 (ether C–O), and 1606 (C=C). Gel permeation chromatography (GPC) (NMP): Mn = 3714 g/mol, and Mw = 5512 g/mol.

Synthesis of 4-(Benzyl ether)phenoxy-2,3,5,6-tetrafluorophenylene-Terminated Oligo(2,6-dimethyl-1,4-phenylene ether) (BT-OPE)

PF-OPE (1.0 g, 0.51 mmol), 4-benzyloxyphenol (0.2247 g, 1.122 mmol), K2CO3 (0.155 g, 1.122 mmol), and DMF (10 g) were taken into a 100 mL three-neck reactor equipped with a thermocouple, magnetic stirrer, and nitrogen flow. The solution was then kept for stirring for 12 h at 120 °C. After cooling to 25 °C, the reaction mixture was poured into methanol/water (1/1, v/v), and the turbid solution was stirred overnight. The precipitate was then filtered and dried in a vacuum oven for 12 h at 60 °C to get an off-white powder (yield = 85%). 1H NMR (ppm, CDCl3, 400 MHz): δ = 1.69 (6H, H1), 2.08 (Ar–CH3, H3, H5), 5.01 (4H, H12), 6.46 (Ar–H, H4), 6.89 (Ar–H, H10, H11, H15), 6.97 (4H, H2), and 7.39 (Ar–H, H13, H14). 13C NMR (ppm, CDCl3, 100 MHz): δ = 16.77 (C3, C5), 30.97 (C1), 70.53 (C12), 114.40 (C4), 115.76 (C11), 116.60 (C10), 127.35 (C2, C15), 128.01 (C14), and 128.58 (C13). 19F NMR (ppm, CDCl3, 376 MHz): δ = −159 (F at C6, C10), and −155 (F at C7, C9). FTIR (KBr, cm–1): ν = 991(C–F), 1191, 1302 (C–O–C), and 1601(Ar. C=C). GPC (NMP): Mn = 4587 g/mol, and Mw = 6132 g/mol.

Synthesis of 4-Hydroxylphenoxy-2,3,5,6-tetrafluorophenylene-Terminated Oligo(2,6-dimethyl-1,4-phenylene ether) (HT-OPE)

BT-OPE (5.0 g, 2.15 mmol), Pd/C (0.25 g), and THF (50 mL) were taken in a 100 mL Parr reactor, equipped with a temperature sensor and hydrogen pressure regulator. The solution was stirred at 140 psi, for 24 h at 50 °C. After that, the reaction mixture was passed through Celite, and the filtrate was then poured into methanol/water (1/1, v/v). The precipitate was filtered and dried in a vacuum oven for 12 h at 60 °C to get a yellow powder (yield = 72%). 1H NMR (ppm, CDCl3, 400 MHz): δ = 1.70 (6H, H1), 2.09 (Ar–CH3, H3, H5), 5.30 (2H, OH), 6.48 (Ar–H, H4), 6.89 (Ar–H, H10, H11), and 6.97 (4H, H2). 13C NMR (ppm, CDCl3, 100 MHz): δ = 16.76 (C3, C5), 30.92 (C1), 114.41 (C4), 116.61 (C11), 116.82 (C10), and 127.36 (C2). 19F NMR (ppm, CDCl3, 376 MHz): δ = −158 (F at C6, C10), and −154 (F at C7, C9). FTIR (KBr, cm–1): ν = 991(C–F), 1191, 1308 (C–O–C), 1606 (Ar. C=C), and 3463 (−OH). GPC (NMP): Mn = 4225 g/mol, and Mw = 5404 g/mol.

Synthesis of 4-Vinylbenzyl Ether Phenoxy-2,3,5,6-tetrafluorophenylene-Terminated OPE (VT-OPE)

HT-OPE (1.0 g, 0.4673 mmol), K2CO3 (0.3229 g, 2.337 mmol), 4-vinylbenzyl chloride (0.3922 g, 2.57 mmol), and DMAc (10 g) were taken in a 100 mL three-neck reactor equipped with a temperature sensor, nitrogen flow, and magnetic pellet. The solution was stirred at 100 °C for 24 h. After cooling to 25 °C, the reaction mixture was poured into methanol/water (1/1, v/v). The precipitate was filtered and dried in a vacuum oven for 12 h at 60 °C to get a white powder (yield = 82%). 1H NMR (ppm, CDCl3, 400 MHz): δ = 1.68 (6H, H1), 2.08 (Ar–CH3, H3, H5), 4.99 (4H, H12), 5.25 (2H, H16), 5.72 (2H, H16′), 6.46 (Ar–H, H4), 6.95(H2, H10, H11, H15), 7.34 (4H, H13), and 7.40 (4H, H14). 13C NMR (ppm, CDCl3, 100 MHz): δ = 16.77 (C3, C5), 31.03 (C1), 70.30 (C12), 114.41 (C4), 115.79 (C11), 116.61 (C10), 126.40 (C14), 127.36 (C2), 127.64 (C13), and 136.35 (C15). 19F NMR (ppm, CDCl3, 376 MHz): δ = −159 (F at C6, C10), and −155 (F at C7, C9). FTIR (KBr, cm–1): ν = 905 (C=C–H), 991 (C–F), 1191, 1302 (ether C–O), and 1601 (ar. C=C). GPC (NMP): Mn = 4069 g/mol, and Mw = 5194 g/mol.

Thermosets Preparation via Self-Curing of OPEs

As shown in Scheme , self-cross-linking of vinyl groups of OPEs was performed using tert-butyl cumyl peroxide (TBCP) as an initiator to get the thermosets. The curing of VT-OPE exemplifies thermoset synthesis. VT-OPE (1.0 g) and TBCP (0.01 g) were dissolved in 4.04 g of xylenes (is commonly used solvent for the CCL industry) and taken in an aluminum mold. The solvent was kept for drying at 80 °C for 12 h and then cured at 120 °C (2 h), 180 °C (2 h), 200 °C (2 h), and 220 °C (2 h) under a nitrogen atmosphere. The thermosets of VT-OPE and OPE-2St are named as C-(VT-OPE) and C-(OPE-2St), respectively, where C refers to cured.

Scheme 4

Proposed Curing Chemistry of VT-OPE

Preparation of Copper-Clad Laminate

VT-OPE was dissolved in toluene to make a solution with a solid content of 40 wt %. E-glass cloth (AST#106 L glass) was immersed in the solution for 30–50 s and subsequently dried at 70 °C for 10 min to prepare the prepreg. The copper-clad laminate was manufactured by stacking six pieces of prepregs with copper foil (18 μm) on each side. The heating and pressing process is 130 °C/15 min, 150 °C/0.15 MPa/15 min, 170 °C/0.45 MPa/30 min, 170 °C/0.75 MPa/1.5 h. The laminate was post-cured at 200 °C for 2 h to form a copper-clad laminate.

Characterization

NMR measurements: Varian Inova 400 NMR in DMSO-d6 (or in CDCl3).

Differential scanning calorimetry (DSC): PerkinElmer DSC 8000, the heating rate was 10 °C/min. Thermogravimetric analysis (TGA): PerkinElmer Pyris1 at a heating rate of 20 °C/min under a nitrogen atmosphere.

Dynamic mechanical analysis (DMA): PerkinElmer Pyris Diamond DMA, programmed heating rate of 5 °C/min with a frequency of 1 Hz. Storage modulus and tan δ were observed with respect to temperature.

Dielectric properties: Rhodes Schwarz vector network analyzer at 10 GHz, on the thin film with a dimension of 9 cm × 13 cm. The exact thickness of the film was measured for each test (5–100 μm).

Water absorption tests: The samples were dipped into boiling water for a few minutes (5 min), the gained weight was measured, and wettability was calculated using the formula (W −W0)/W0 × 100 (%), where W is the weight of the sample after placing in water and W0 is the weight of the sample before placing in water.

Thermal mechanical analysis (TMA): A SII TMA/SS6100 with a heating rate of 5 °C/min was used.

Gel permeation chromatography (GPC): A Hitachi-L2400 (NMP as solvent) gel permeation chromatogram was used at 40 °C with a flow rate of 1.0 mL/min. The instrument was calibrated using polystyrene as standards (Mn = 4130, 10 730, 200 000, and 300 000 g/mol).

Contact angle: The data were collected using an FTA1000B through the Sessile drop method using water as a testing liquid.

FTIR spectra: A PerkinElmer RX1 infrared spectrophotometer was used to scan the wavenumber ranging 400–4000 cm–1.

Flame retardancy of the OPE thermosets was measured by a UL-94 VTM vertical thin-film test. In that test, an 8″ × 2″ sample was wrapped around a 1/2″ mandrel and then taped on one end. The mandrel was removed, leaving a cone-shaped sample that was relatively rigid. The two flame applications were applied, 3 s for each ignition to perform the UL-94 VTM vertical thin-film tests. After the first ignition, the flame was removed and the time for the polymer to self-extinguish (t1) was recorded. Cotton ignition was noted if polymer dripping occurred during the test. After cooling, the second ignition was performed on the same sample and the self-extinguishing time (t2) and dripping characteristics were recorded. If t1 plus t2 was less than 10 s without any dripping, the polymer was considered to be a VTM-0 material. If t1 plus t2 was in the range of 10–30 s without any dripping, the polymer was considered to be a VTM-1 material. In another scenario, if the material can self-extinguish in <10 s with dripping liquid surrounded by fire, then such material is considered as VTM-1 grade. Pressure cook test: A standard procedure to understand the reliability of CCL. The CCL is put in a pressured tank with water and heated to 120 °C for 2 h to absorb saturated water. After that, the CCL was put in a solder at 288 °C. The position of the CCL in the solder can be horizontal (floating test) or vertical (dipping test). If the CCL remains intact without delamination for 3 min, it is called “pass”. In this work, the CCL based on VT-OPE can pass the solder floating test and the solder dipping test after the pressure cook test.

Results and Discussion

Synthesis and Characterization of HT-OPE

The design of this work is to incorporate the symmetric tetrafluorophenylene moiety into OPEs to simultaneously enhance properties such as thermal, dielectric property, hydrophobicity, and flame retardancy. SA90 is a fully developed hydroxyl-terminated OPE from SABIC. It is suitable for preparing telechelic OPE derivatives through an elementary reaction, such as nucleophilic addition or nucleophilic substitution. Therefore, we proposed a synthetic route to prepare a fluoro-containing OPE with hydroxyl end groups (HT-OPE) using SA90 as a starting material (Scheme ).

Firstly, PF-OPE was successfully prepared from SA90 with hexafluorobenzene at around 80 °C. When PF-OPE was further reacted with excess hydroquinone (about 4 equiv), gelation was observed (Scheme ). It is probable that the pentafluorophenyl moieties of PF-OPE exhibited great reactivity, causing the undesirable polymerization or tri-substituted cross-link.[58] To obtain the critical precursor HT-OPE for further functionalization, a two-step synthetic strategy has been proposed in Scheme . PF-OPE reacted with 4-benzyloxyphenol at 120 °C to obtain BT-OPE in the first step. After deprotecting the benzyl group from hydroxyl of BT-OPE by catalytic (Pd/C) hydrogenation, we successfully received HT-OPE.

Scheme 2

Synthesis of HT-OPE via Deprotection Procedure

Figure S1 shows the 1H NMR spectra of SA90, PF-OPE, BT-OPE, and HT-OPE in CDCl3. In Figure S1a,b, the broad hydroxyl signal at 4.5 ppm in SA90 was not observed in PF-OPE, indicating that the reaction has already been completed. In Figure S1c, the Ar–H signals at 6.89–7.39 ppm and the methylene signal of benzyl ether (H12) at 5.01 ppm support the formation of BT-OPE. For Figure S1d, the appearance of hydroxyl signal (H12) at 5.30 ppm and the disappearance of methylene signal of benzyl ether support the formation of HT-OPE. Figure S2a shows the 19F NMR spectra of PF-OPE; the characteristic −F atom bonded with C8 is located at −157 ppm as a triplet with a lower intensity than the rest of the −F atom bonded with C6, C7, C8, and C9. In Figure S2b, the intensity of this triplet (−F at C8) is negligible suggesting the presence of benzyloxyphenyl ether at −para position. Figure S3 shows the FTIR spectra of SA90, PF-OPE, BT-OPE, and HT-OPE. For Figure S3a,b, the disappearance of the hydroxyl signal at 3572 cm–1 and the appearance of the C–F bond at 991 cm–1 indicate PF-OPE formation. For Figure S3d, the hydroxyl signal at 3563 cm–1 and the C–F bond at 991 cm–1 indicate HT-OPE formation. Figure S4 shows the matrix-assisted laser desorption ionization–time-of-flight (MALDI-TOF) mass spectrum of HT-OPE. Various m + n values corresponding to the HT-OPE structure have been listed in the figure, further confirming the structure.

Synthesis of VT-OPE

As shown in Scheme , VT-OPE was prepared by reacting HT-OPE with 4-vinylbenzyl chloride in the presence of K2CO3. Figure shows the 1H NMR spectrum of VT-OPE in CDCl3. We can see the characteristic signals at 5.3, 5.7 ppm for vinyl (H16 and H16′) and 5.0 ppm for vinylbenzyl ether (H12). Similarly, no characteristic signal of the hydroxyl group at 5.30 ppm was observed, indicating a complete nucleophilic substitution. Figure displays the FTIR spectra of HT-OPE and VT-OPE. For VT-OPE, the appearance of a small vinyl peak at 905 cm–1 is also accompanied by the disappearance of hydroxyl. MALDI-TOF mass spectra of VT-OPE are provided in Figure S5, further confirming the individual structure.

Figure 2

1H NMR spectrum of VT-OPE in CDCl3, as per 1H NMR the m + n is approximately 6 pronouncing VT-OPE as an oligomer.

Figure 3

FTIR spectra of (a) HT-OPE and (b) VT-OPE.

GPC and Solubility Data of OPEs

GPC curves of (i) SA90, (ii) PF-OPE, (iii) BT-OPE, (iv) HT-OPE, and (v) VT-OPE are shown in Figure S6a, GPC curves of (i) blank NMP and (ii) OPE-2St are shown in Figure S6b. The molecular weights (Mn) of the OPE-based resins are between 3500 and 4500 g/mol with a PDI of 1.2–1.4 Đ. The Mn of VT-OPE is slightly lower than OPE-2St (with Mn = 4336 g/mol). Table lists the solubility data of SA90, VT-OPE, and OPE-2St. Since the Mn values of all of the OPE-based resins are nearly similar, their solubility is also quite similar. The addition of the tetraphenylene moiety has no negative effect on the solubility of VT-OPE; therefore, we expect that VT-OPE would follow the same processing pattern as that of commercial OPE-2St so that it is convenient to produce commercial copper-clad laminates.

Table 1

Solubility of SA90, VT-OPE, and OPE-2St

sample code	DMAc	DMF	NMP	THF	MEK	Toluene	CHCl3	DCM	PM
SA90	++a	++	++	++	++	++	++	++	+–
VT-OPE	++	++	++	++	++	++	++	++	+h
OPE-2St	++	++	++	++	++	++	++	++	+–

++, soluble at room temperature; +h, soluble on heating (80 °C); +–, slightly soluble on heating (80 °C). Solubility tests were conducted by taking 5 mg of sample in 2 mL of test solvent.

Curing Behavior of VT-OPE

Figure shows DSC thermograms of VT-OPE with and without 1.0 wt % TBCP. For Figure a, a non-notable exothermic characteristic of neat VT-OPE occurred at 244 °C. However, as shown in Figure b, a big exothermic peak temperature of VT-OPE with 1 wt % TBCP reduced remarkably to around 190 from 244 °C, and the corresponding exothermic enthalpy increased quite substantially. Therefore, we suggest that the radical polymerization of vinylbenzene could be initiated successfully by TBCP between 150 and 220 °C, impelling us to use TBCP as an initiator to cure VT-OPE. To check the structures of thermosets, FTIR analyses were performed. Figure S7 shows the FTIR spectra of (a) VT-OPE and (b) C-(VT-OPE). For C-(VT-OPE), the disappearance of vinyl characteristic peak at 905 cm–1, and the intensity reduction of the styrene −CH=CH2 peak at 2932 cm–1 supports the radical polymerization of vinyl groups (Figure S7). Besides those mentioned above, there are no other remarkable variations in FTIR spectra. However, some literature has disclosed an unexpected curing chemistry of tetrafluorophenylene units. Kobrina et.al. reported that the reactions of hexafluorobenzene with alkyl radicals would generate the mixtures of fluoro-containing dimers (see Scheme in ref (59)). AlliedSignal, Inc. disclosed that bifunctional radicals from bistriazenes would cross-link with a fluorinated poly(arylene ether).[60] This type of chemistry was studied very less in the literature; to shed some light on it, we have verified the real possibility of an unexpected reaction (Scheme S1) of fluorinated moieties with radicals (from TBCP) by doing some model study. We have performed DSC thermograms on PF-OPE, BT-OPE, and HT-OPE with 1.0 wt % TBCP (Figure S8). As shown in Figure S8a, all of the tetrafluorophenylene moiety-containing samples showed exothermic peaks even though they do not contain any cross-linking unit (4-methyl styrene). To further explain the reason for the existence of the exothermic peak, we have tested DSC of TBCP alone (Figure S8a), and surprisingly, the exothermic peak is indeed caused by the dissociation of the TBCP, rather than the dimerization of tetrafluorophenylene moieties shown in Scheme S1. Figure S8b shows DSC thermograms of PF-OPE, BT-OPE, and HT-OPE without the addition of TBCP catalyst, and no exothermic peaks were observed in any of the samples. As a result, the proposed curing chemistry of VT-OPE is only radical polymerizations of active double bonds of styrene moieties (Scheme ).

Figure 4

DSC thermograms of (a) neat VT-OPE and (b) VT-OPE with 1.0 wt % of TBCP.

DMA and TMA Thermograms

Figure S9 shows the photograph of C-(VT-OPE); it can be folded without cracking. Figure and Table show the DMA thermograms and thermal data of C-(VT-OPE). Post Tg, stable modulus plateaus (E′ > 107 Pa after Tg) were observed for both the samples, indicating that they belong to highly cross-linked network. The Tg value, defined by the peak maximum temperature of the tan δ curve, is 242 °C for C-(VT-OPE), which is 13 °C higher than the commercial C-(OPE-2St). The modulus at Tg +40 °C (the rubbery modulus) is 22 MPa for C-(VT-OPE), whereas the modulus at Tg +40 °C is 19 MPa for the commercial C-(OPE-2St). The cross-linking density, according to E′/3RT, in which E′ is the storage modulus in the rubbery plateau,[61] is listed in Table .

Figure 5

DMA thermograms of C-(VT-OPE) and C-(OPE-2St).

Table 2

Thermal Properties of C-(VT-OPE) and C-(OPE-2St)

sample code	Tg (°C) (DMA)a	E′ at Tg +40 °C (MPa)	cross-linking density (mmole/cm3)b	Tg (°C) (TMA)c	CTE (ppm/°C)d
C-(VT-OPE)e	242	22	1.59	202	55
C-(OPE-2St)e	229	19	1.41	199	76

Measured by DMA at a heating rate of 5 °C/min. Defined by the peak temperature of the tan δ curve.

The cross-linking density accords with E′/3RT, in which E′ is the storage modulus in the rubbery plateau (E′ at Tg +40 °C).

Measured by TMA at a heating rate of 5 °C/min. Defined by the onset temperature of TMA thermograms.

Coefficient of thermal expansion, defined at 50 to 150 °C by TMA thermograms.

Mn of VT-OPE is 4069 g/mol, and Mn of OPE-2St is 4336 g/mol (NMP as the eluting solvent) the respective results are depicted in Figure S6a,b.

Polymeric material with a low coefficient of thermal expansion (CTE) is essential to use in electronic applications. Figure S10 displays the TMA thermograms of C-(OPE-2St) and C-(VT-OPE). The Tg value (onset temperature of dimensional change), is 202 °C for C-(VT-OPE), which is 26 °C higher than that of C-(OPE-2St). The CTE value is 55 ppm/°C for C-(VT-OPE), which is much better than that of the C-(OPE-2St) (76 ppm/°C).

Thermal Stability and Flame Retardancy

Figure and Table display the TGA thermograms and data of C-(VT-OPE) and C-(OPE-2St). The decomposition temperatures Td5% and Td10% of C-(VT-OPE) are 420 and 455 °C, respectively, which are at least 25 °C higher than commercial C-(OPE-2St). The char yield is 38 wt % for C-(VT-OPE), which is 11 wt % higher than that of the commercial C-(OPE-2St). The enhanced char yield for C-(VT-OPE) can be ascribed to the higher cross-linking density, which can promote the carbon content, in other words char formation.[62,63]Table also listed the flame-retardant data of C-(VT-OPE) and C-(OPE-2St). The sum of two burning times is 2.6 for C-(VT-OPE) without any fire dripping. Therefore, it belongs to UL-94 VTM-0 grade since the sum of the two burning times is less than 10 s and without any dripping. However, C-(OPE-2St) could not obtain VTM-0 quality because of severe fire dripping during the test.

Figure 6

TGA thermograms of C-(VT-OPE) and C-(OPE-2St).

Table 3

Thermal Stability and Flame-Retardant Properties of C-(VT-OPE) and C-(OPE-2St)

sample code	Td5% (°C)a	Td10% (°C)b	CY (%)c	1st burning time (s)	2nd burning time (s)	dripping	UL-94
C-(VT-OPE)	420	455	38	2.6	0	no	VTM-0
C-(OPE-2St)	394	427	27	3.3		seriously	VTM-1

Temperature corresponds to 5% weight loss by thermogravimetry at a heating rate of 10 °C/min.

Temperature corresponds to 10% weight loss by thermogravimetry at a heating rate of 10 °C/min.

Char yield (residual weight % at 800 °C).

Water Absorption, Contact Angle, and Dielectric Properties

Hydrophobicity of polymeric material is as important as ultralow dielectric property for high-frequency transmission printed circuit boards. Moisture absorption of the polymeric material can have a negative impact on Dk. Table lists the results of water absorption tests for C-(VT-OPE) and C-(OPE-2St). The saturated water absorption is around 0.135 wt % for C-(VT-OPE), whereas the saturated water absorption is about 0.312 wt % for C-(OPE-2St). Therefore, a 57% reduction in water absorption was observed through the incorporation of tetrafluorophenylene moiety. Figure S11 shows the contact angle of C-(VT-OPE) and C-(OPE-2St). The contact angles are 87 and 82° for C-(VT-OPE) and C-(OPE-2St), respectively, showing better hydrophobicity of the C-(VT-OPE). Table also lists the dielectric properties of OPE thermosets. The Dk values are 2.58 and 2.75 for C-(VT-OPE) and C-(OPE-2St), respectively. A 6.2% reduction in the Dk value was observed. These dielectric results indicate the bulky tetrafluorophenylene moiety, which packs less densely and provides larger free volume, has an advantage in reducing the dielectric constant. We also hypothesize that the symmetric tetrafluorophenylene moiety can reduce polarizability within the chain because all of the electronegative fluorine atoms are distributed evenly and contribute to lower Df characteristic.

Table 4

Water Absorption, Contact Angle, and Dielectric Properties of C-(VT-OPE) and C-(OPE-2St)

	water absorption (%) at
sample code	24 h	48 h	72 h	96 h	contact angle (°)	Dk (10 GHz)	Df (10 GHz)
C-(VT-OPE)	0.135	0.135	0.135	0.135	87.0	2.58	0.005
C-(OPE-2St)	0.231	0.312	0.312	0.312	82.8	2.75	0.006

Practicability of OPE Thermosets for Copper-Clad Laminate

The derivative CCLs were obtained by a simplified immersion process. The properties of CCL based on the VT-OPE are listed in Table . The properties of CCL based on OPE-2St are also presented to conduct a comparison. The CCL based on VT-OPE can pass the solder floating test and the solder dipping test after the pressure cook test at 120 °C for 2 h. No abrupt dimensional change was observed for the laminate in a TMA measurement at 288 °C, which means no delamination of the laminate during the test. It is worth noting that even incorporating fluoro-containing moieties, the CCL based on VT-OPE still exhibits good peeling strength with copper foil, which can be attributed to the relatively low fluorine content for VT-OPE compared with perfluorinated polymers. Additionally, the CCL based on VT-OPE shows a higher peeling strength with copper foil than OPE-2St. As a result, VT-OPE has the potential for application in high-frequency copper-clad laminates.

Table 5

Data of CCL Based on VT-OPE and OPE-2St

OPE for CCL	floating (s)a	peeling (lb/in)b P	dipping (s)c	TMA288 °C (min)d
VT-OPE	>600	5.5	>600	>60
OPE-2St	>600	4.7	>600	>60

Time for the laminate to delaminate at solder floating test at 288 °C.

eeling strength with copper foil (lb/in).

Time for the laminate to delaminate at solder dipping test at 288 °C after the laminate experience pressure cook test at 120 °C for 2 h.

Time for an abrupt dimensional change for the laminate in a TMA measurement at 288 °C, that is, the time for delamination of the laminate during the test.

Conclusions

We have prepared a tetrafluorophenylene-containing vinylbenzyl ether-terminated oligo(2,6-dimethyl-1,4-phenylene ether) (VT-OPE) as a resin for copper-clad laminate to apply in high-frequency communication. The detailed structural investigation of VT-OPE resin and the thermal and mechanical properties of thermoset C-(VT-OPE) were discussed. C-(VT-OPE) thermoset shows a higher Tg (242 °C), a lower CTE (55 ppm/°C), a higher contact angle (87°), a lower dielectric constant (Dk = 2.58, Df = 0.005 at 10 GHz), and excellent flame-retardant property (VTM-0) compared to commercial C-(OPE-2St). These results demonstrate that incorporating the tetrafluorophenylene moiety into OPE is a promising strategy to enhance the overall performance of the commercial C-(OPE-2St).