Ultralow dielectric, high performing hyperbranched epoxy thermosets: synthesis, characterization and property evaluation

Abstract

In the present report, low viscosity hyperbranched epoxy resins are synthesized using a simple A₂ + B₄ polycondensation reaction between pentaerythritol and the in situ prepared diglycidyl ether of bisphenol-A, with variation of the reaction time and amount of B₄ reactant. The structural features and degree of branching (DB) of the hyperbranched epoxy resins are determined by spectroscopic analyses, such as FTIR, ¹H NMR and ¹³C NMR studies. The highest DB (0.78) with the lowest epoxy equivalent (394 g eq⁻¹), viscosity (2.99 Pa s) and specific gravity (1.03) values is found for the resin formed after a reaction time of 4 h with 10 wt% pentaerythritol. The thermoset of the same resin also exhibits the highest tensile strength (51 MPa), elongation at break (37.5%), toughness (1432 MPa), adhesive strength (3429 MPa) and thermal stability (∼300 °C), as well as the lowest dielectric constant (1.8), dielectric loss (0.009) and moisture absorption (0.09%). The results are also compared with a linear diglycidyl ether of bisphenol-A based epoxy (without the addition of pentaerythritol) prepared under the same conditions. The study also shows that the hyperbranched epoxy is superior in terms of physical properties as well as performance (especially, the toughness value is 800% higher) compared to the linear analog. Thus the synthesized hyperbranched epoxy thermoset can solve the genuine problem of the brittleness character of the conventional epoxy. This epoxy thermoset can also be used efficiently as a low dielectric adhesive material in the field of electronics.

---

Introduction

Over the last 25 years, different hyperbranched polymers have been utilized in several fields, where benefits arise from the structural uniqueness of these polymers. However, hyperbranched epoxy resins are still not gaining importance over the conventional linear diglycidyl ether of bisphenol-A based epoxy. But over the last 10 years, research has been ongoing to substitute linear commercial epoxy with hyperbranched epoxy, to explore the interesting properties and easy processing of the latter. Thus a variety of hyperbranched epoxy resins have been prepared using various synthetic approaches including end group modification of suitable hyperbranched polymers,^1–3 polycondensation reactions,^4–6 functionalization of poly(methyl acrylate) using atom transfer radical polymerization (ATRP),⁷ proton transfer polymerization,^8,9 etc. However, in most cases the products possess high viscosity and thus would need dilution with organic solvents, which is detrimental to the environment. In most of the cases, poor tensile strength was also found, mainly due to a low crosslinking density. Thus, an epoxy with the desired high performance is rarely found in the literature. On the other hand, a diglycidyl ether of bisphenol-A based commercial epoxy thermoset is widely used because of its overall good performance, like high mechanical properties, good weather and chemical resistance, low shrinkage, high adhesive strength, etc.^10,11 However, many advanced applications are restricted due to its inherent brittleness and low toughness character. To impart toughness or flexibility to this commercial epoxy, different modifications have been intensively investigated throughout the world. These include toughening by blending with different flexible polymers like rubbers,^12,13 hyperbranched polyesters,^14,15 etc. However, due to processing difficulties, insignificant improvements and low cost-performance ratios, commercial exploration is limited. Thus, a unique structural design may impart adequate toughness, as well as other desired properties, to address the shortcomings of the conventional epoxy. The chemical composition, structural architecture and reaction conditions are the key parameters for tuning the properties of such novel epoxy resins.

Another important proposition in this regard is the synthesis of a low dielectric constant epoxy thermoset. Such low dielectric constant thermosets are preferred for several electrical and electronic applications, especially printed circuit boards, semiconductor devices, microelectronic devices, electronic packages, etc.¹⁶ The use of a low dielectric epoxy thermoset is advantageous, in part, because they adhere well to different substrates, including metals, compared to conventional dielectric materials. Polyethylene and poly(tetrafluoroethylene), even though they possess a low dielectric constant (approximately 2.1), cannot be used for the above purposes due to their low adhesive strength or high chemical inertness. Therefore, epoxy thermosets which can provide a low dielectric constant are essential. However, most of the conventional epoxy thermosets possess a dielectric constant in the range of 4–5. Even the filled systems, as well as the blends of different conventional epoxy thermosets, possess a high dielectric constant (>3). In this vein, Chellis et al. teaches about a microsphere filled low dielectric epoxy,¹⁷ Lin et al. used a silica filled epoxy for a printed circuit,¹⁸ Wojnarowski et al. used an epoxy/polyimide copolymer blend for layered circuits¹⁹ and Viswanath et al. studied the electrical properties of an epoxy/hyperbranched polyester blend.²⁰ Many researchers have also tried to synthesize low dielectric constant epoxy thermosets from different types of reactants. Lin et al. synthesized a 2,6-dimethyl phenol based epoxy thermoset which exhibited a dielectric constant of 3.1 with a dielectric loss of 0.065.²¹ Hwang et al. synthesized dicyclopentadiene-containing poly(phenylene oxide) novolac and the dicyclopentadiene containing epoxy cured thermoset showed a dielectric constant of 2.6 with a dielectric loss of 0.0104.²² In our earlier report, a triethanol amine and bisphenol-A based hyperbranched epoxy thermoset was synthesized with a dielectric constant of 3.16 and a dielectric loss of 0.01.²³ However, obtaining ultralow dielectric constant materials (<2) with a high performance and low power consumption is the key challenge in microelectronic devices. The authors, therefore, wish to report the synthesis, characterization and property evaluation of a hyperbranched epoxy resin with the desired toughness, flexibility, thermostability, dielectric constant and adhesive strength to address all such drawbacks of the commercial epoxy resin. This study also showed that the reaction time of the polycondensation reaction and the amount of branch generating moiety have a strong influence on the ultimate performance of the thermosets.

Experimental

Materials

Pentaerythritol (Sigma-Aldrich, Germany) was used after re-crystallization from ethanol. Bisphenol-A and epichlorohydrin were purchased from G. S. Chemical, India. Bisphenol-A was re-crystallized from toluene before use. The poly(amido-amine) hardener (HY840, amine value 5–7 eq kg⁻¹) was obtained from Ciba Giegy, India. All other reagents used in the present investigation were of reagent grade.

Preparation of the hyperbranched epoxy resin

The hyperbranched epoxy resin was synthesized by a polycondensation reaction of bisphenol-A and pentaerythritol (5, 10 and 15 wt% separately with respect to bisphenol-A) with epichlorohydrin (1 : 2 molar ratio with respect to the hydroxyl group) at 110 °C with continuous stirring. Amounts of 10.0 g of bisphenol-A, 1.0 g of pentaerythritol (for 10 wt%) and 21.64 g of epichlorohydrin were placed in a two necked round bottom flask equipped with a water condenser and a dropping funnel. The reaction mixture was stirred with a magnetic stirrer. To this reaction mixture, a 5 N aqueous solution of NaOH (4.67 g, equivalent to the hydroxyl group) was added slowly from the dropping funnel. The addition of the NaOH solution to the reaction mixture was started at 60 °C, and it took about 30 min to reach the temperature 110 °C. After the desired time period, the reaction was allowed to settle in a separating funnel and the aqueous layer was separated out from the desired organic layer. Then the organic layer was washed with 15% aqueous sodium chloride solution followed by distilled water 2–3 times. Then the epoxy resin was dried at 70 °C under vacuum with the help of a rotary evaporator (Eyela, Tokyo, Japan) to remove the excess epichlorohydrin and entrapped water. The drying process was performed until all of the epichlorohydrin was removed from the resin, by checking for a constant weight. The epoxy resins from three different reaction times, viz. 2 h, 4 h and 6 h, were synthesized and coded as PHE2h, PHE4h and PHE6h (yield: ∼95%). Epoxy resins with three different wt% of pentaerythritol, viz. 5, 10, 15 wt%, were synthesized similarly using a 4 h reaction time and coded as PHE5, PHE4h and PHE15, respectively (as PHE4h was also synthesized at the same composition under the same conditions). An epoxy resin (linear, diglycidyl ether of bisphenol-A) with an absence of pentaerythritol was also prepared using the same procedure, for comparison purposes, and coded as PHE0.

Curing of the resins

The hyperbranched epoxy resins and PHE0 were cured by homogenous mixing with 50 phr of poly(amido amine) hardener, with respect to the resins, in a glass beaker at room temperature in the presence of a small amount of THF solvent (2–3 drops per g). Then the mixtures were uniformly coated on glass plates (75 × 25 × 1.3 mm³) and steel plates (150 × 50 × 1.6 mm³). The solvent and other volatiles were removed under vacuum over 24 h and the plates were cured at 100 °C for the specified time interval. Optimization of the curing reaction was done by determining the swelling values of the cured films, using THF, after being kept for 48 h in glass bottles.

Characterization

The FTIR spectrum was recorded using a Nicolet FTIR spectrophotometer (Impact-410, Madison, USA) using KBr pellets. The ¹H NMR and ¹³C NMR spectra of the resins were recorded using D₆-DMSO as the solvent and TMS as the internal standard on a 400 MHz, Jeol FTNMR spectrometer. The dilute solution viscosity of the resins was measured using a suspended level Ubbelodhe viscometer, using a 0.5 wt% solution in DMAc at 25 °C. The physical properties, such as the epoxy equivalent, hydroxyl value, density, solubility and swelling value of the resins, were measured using the standard test methods. The determination of the shear viscosities of the resins was done using a Bohlin rheometer CVO100 (Malvern, UK). The tensile strength (standard ASTM D 882) and lap-shear tensile adhesive strength of the hyperbranched epoxy thermosets were measured using a Universal Testing Machine (UTM, WDW10, Jinan, China). The tensile test was performed on rectangular film samples (size: 60 × 10 × 0.3–0.4 mm³) with a 500 N load cell (crosshead speed of 10 mm min⁻¹). The lap-shear adhesion test (standard ASTM D4896-01) was carried out on metal–metal (M–M), wood–wood (W–W) and plastic–plastic (polypropylene sheets) (P–P) adherents (the area of the overlapping zone was 25 × 25 mm² and the thickness of the zone was 0.02–0.03 mm) with a 10 kN load cell (crosshead speed of 50 mm min⁻¹). The lap-shear tensile strength (MPa) (calculated as maximum load per unit bonded area) was obtained directly from the UTM. A scratch hardness test (standard ASTM G171) was carried out using a scratch hardness tester (Sheen instrument Ltd., UK) on the surface of a glass slide coated with hyperbranched epoxy thermoset (area: 75 × 25 × 0.3 mm³). The impact strength of the steel plate coated thermosets (area: 150 × 50 × 1.6 mm³) was measured with an Impact tester (S. C. Dey Co., Kolkata) using the standard falling weight (ball) method (ASTM D 1709). A bending test of the thermosets (thickness of the films were 0.3–0.4 mm) was done using a mandrel with a diameter of 1 to 100 mm (standard ASTM D 522). The flexibility of the PHE4h thermoset was also checked by multi-folding the thin film (thickness 0.4 mm). All of the tests for the measurement of mechanical properties were repeated five times and average values were taken. Thermogravimetric analysis of the thermosets was done using a PerkinElmer TG4000, US, with a nitrogen flow rate of 30 mL min⁻¹ and a heating rate of 10 °C min⁻¹, from 30 to 700 °C. The curing study and measurement of the glass transition temperature (T_g) of the PHE4h and PHE0 thermosets were done by differential scanning calorimetry (DSC) analysis using a PerkinElmer DSC6000, US. For the measurement of T_g, the sample was heated from −20 to 140 °C in a heating–cooling–heating cycle at a rate of 10 °C min⁻¹ under a nitrogen flow rate of 30 mL min⁻¹. The dielectric properties of the thermoset films were measured using an LCR Hitester instrument (Hioki 3532-50), up to a frequency of 1 MHz at 25 °C. The dielectric constants were calculated using the equation C = ε₀ε_rA/d, where C is the capacitance, ε₀ is the permittivity of the vacuum, ε_r is the permittivity or dielectric constant of the polymer sample, A is the area of the electrode plates and d is the distance between the top and bottom plates i.e. the thickness of the films (0.3–0.4 mm).²³ The water absorption test was done by immersing the thermosets (size: 15 × 15 × 0.4 mm³) in distilled water at 25 °C and the percentange weight gain was determined after 48 h. The chemical resistance test was done in various chemical environments. Aqueous HCl (10% by volume), aqueous NaOH (5 wt%), aqueous NaCl (10 wt%), aqueous ethanol (20% by volume) and tap water were used to investigate the effect of these chemicals on the thermoset films. The films were cut into small pieces (size: 15 × 15 × 0.3–0.4 mm³) and kept in 100 mL glass bottles containing the aforesaid media at an ambient temperature (25 °C). The percentage weight loss was measured after 30 days of testing. The transparency of the PHE4h thermoset was checked using printed words covered with a thin film (thickness 0.4 mm). The molecular weights of the resins were determined using gel permeation chromatography (Waters Corporation, USA) with the help of a refractive index detector-2414 and polystyrene–polydivinylbenzene GPC column, using THF as the solvent.

Results and discussion

Synthesis and characterization of the hyperbranched epoxy resins

Hyperbranched epoxy resins were synthesized using an A₂ + B₄ polycondensation reaction. First, the diglycidyl ether of bisphenol-A (A₂) was formed in situ during the reaction, as the reactivity of the hydroxyl groups of bisphenol-A is higher than for pentaerythritol due to the higher acidic nature of the phenolic protons compared to the aliphatic alcohol protons. The addition of an aqueous solution of NaOH to the reaction mixture was started at 60 °C and it took about 30 min to reach the temperature 110 °C in the procedure. During this period, it is expected that bisphenol-A reacted with epichlorohydrin to produce the diglycidyl ether of bisphenol-A, rather than pentaerythritol. Then the hyperbranched epoxy was formed by reaction between the diglycidyl ether of bisphenol-A and pentaerythritol (B₄) along with the reaction of epichlorohydrin with the free hydroxyl groups of pentaerythritol, as shown in Scheme 1. The structural features of the synthesized epoxy were characterized by FTIR and NMR studies. The stretching vibrations (ν_max/cm⁻¹) in the FTIR spectrum (Fig. 1) are attributed to the following features: 916 (oxirane), 3453 (O–H), 3055 (aromatic C–H), 2973 (aliphatic C–H), 1610 (aromatic C=C), 1251 (C–O) and 1037 (C–C).^4,5 The ¹H NMR (Fig. 2) spectrum δ_H (ppm) indicate the following structural features: 3.3 (1H, oxirane), 2.7 and 2.8 (2H, oxirane),^4,5 3.7–3.8 (2H, 4CH₂ of the substituted and un-substituted pentaerythritol), 1.5 (3H, CH₃), 6.8 (4H, Ph), 7.1 (4H, Ph), 4.2 (1H, CHOH), 4.0 (2H, CH₂–oxirane), 3.5 (2H, CH₂–pentaerythritol unit), 4.1 (2H, CH₂–bisphenol-A unit), 5.4 and 5.6 (1H, two types of OH). In the ¹³C NMR spectrum (Fig. 3), δ_C (ppm): 43 (CH₂, oxirane), 49 (CH, oxirane),^4,5 46–47 (central C of pentaerythritol unit), 114, 127, 143 and 156 (4C, Ph), 31 (CH₃, bisphenol-A unit), 41 (C, isopropylidiene of bisphenol-A unit), 68 (CH₂, pentaerythritol unit), 51 (CH₂–oxirane) and 62–67 (CH₂–O units and CHOH unit).^4,5 The degree of branching (DB) of the hyperbranched epoxy with variation of the time and amount of B₄ reactant was determined from the ¹³C NMR spectra of the synthesized resins. A hyperbranched polymer should contain three different types of unit: dendritic (D), linear (L) and terminal (T) in its structure. DB is the ratio of the sum of the integration of the D and T units to the sum of the integration of the D, L, and T units, i.e., DB = (D + T)/(D + L + T).^4,5 For a hyperbranched polymer, this value should be greater than 0.5. In the case of the synthesized epoxy resins, the units with three or four, two and one hydroxyl group(s) of pentaerythritol substituted by the diglycidyl ether of bisphenol-A moiety are dendritic (D), linear (L) and terminal (T) units respectively, as shown in Scheme 1. In the ¹³C NMR spectrum (Fig. 3), the central carbon atom of pentaerythritol for these four units were observed at δ = 46.75, 46.55, 46.40 and 46.20 ppm, respectively. The calculated DB from the integration values of these peaks for the hyperbranched epoxy resins with variations in the time and amount of B₄ unit are given in Table 1. The highest DB was found for a 4 h reaction time with 10 wt% pentaerythritol (with respect to bisphenol-A), due to the fact that under these conditions the highest number of dendritic and terminal units were formed (in this case the highest number of terminal units were formed, as shown in Table 1). In the case of a 2 h reaction time, a smaller number of dendritic and terminal units were formed compared to the linear units (the highest number of linear units were formed). This is due to the incomplete growth of the structure of the hyperbranched epoxy resin. Whereas, at a 6 h reaction time the lowest number of terminal units were formed (Table 1). The epoxy resin with 10 wt% pentaerythritol provides the highest DB among the studied compositions (5–15 wt%). This may be due to the presence of an appropriate ratio of reactants for the formation of the hyperbranched structure of the epoxy resin at this composition. This resulted in both a higher percentage of tri- and tetra-substituted branched units, as well as greater chain end substitution to form terminal epoxides, compared to the other two wt% of pentaerythritol (as shown in Table 1). Whereas a decrease in the amount of pentaerythritol (5 wt%) in the epoxy resulted in the lowest numbers of tri- and tetra-substituted internal branched units, as well as terminal epoxides, due to incomplete growth of the hyperbranched structure at this composition. On the other hand, with an increase in the amount of pentaerythritol (15 wt%) in the epoxy, the generation of branching units might be prohibited due to the congested nature of the pentaerythritol moiety.^4,8,9 Thus, PHE4h exhibited the lowest epoxy equivalent (highest number of epoxy groups), hydroxyl, viscosity (solution as well as shear) and density values, as shown in Table 2. With an increase of the linear units, the viscosity and density values increase as the hydrogen bonding and other intermolecular forces are increased. The synthesized hyperbranched epoxy resins are soluble in most of the common organic solvents, like methanol, ethanol, acetone, THF, DMF, DMAc, DMSO, CHCl₃, CH₂Cl₂, toluene, xylene, ethyl acetate, etc., due to the hyperbranched architecture and the presence of a large number of functionalities, along with a combination of aliphatic and aromatic moieties. Whereas the prepared linear epoxy resin (PHE0) is not soluble in toluene, xylene or ethyl acetate. It also possessed high viscosity (both solution and shear) and density values. The epoxy equivalent and hydroxyl value of PHE0 are lower than for the hyperbranched epoxy because of the linear structure of the former. PHE0 was prepared using the same procedure as PHE4h, for comparison purposes, to show the effect of a branch generating unit. The preparation procedure and structure of PHE0 thus are different from the commercial diglycidyl ether of bisphenol-A based epoxy and hence the results were slightly different.

Scheme 1 Synthesis of the hyperbranched epoxy resin and its possible general structure.

Fig. 1 FTIR spectrum of PHE4h resin.

Fig. 2 ¹H NMR spectrum of PHE4h resin.

Fig. 3 ¹³C NMR spectrum of PHE4h resin.

Table 1 The values of the D, L and T units (%) and the degree of branching (DB) of PHE2h, PHE4h, PHE6h, PHE5 and PHE15

Hyperbranched epoxy resin	D units (%)	L units (%)	T units (%)	DB
PHE2h	39.82	53.59	6.59	0.46
PHE4h	62.62	21.61	15.77	0.78
PHE6h	64.04	32.02	3.94	0.68
PHE5	41.76	47.40	10.84	0.53
PHE15	59.12	29.0	11.88	0.71

---

Table 2 Physical properties of PHE2h, PHE4h, PHE6h, PHE5, PHE15 and PHE0

Parameter	PHE2h	PHE4h	PHE6h	PHE5	PHE15	PHE0
Epoxy equivalent (g eq^−1)	688	394	583	499	427	254
Hydroxyl value (mg KOH per g)	156	100	220	102	170	98
Solution viscosity (dL g^−1) at 25 °C	0.046	0.028	0.033	0.038	0.030	0.089
Shear viscosity (Pa s) at 25 °C	9.40	2.99	6.12	7.88	5.03	18.92
Specific gravity at 25 °C	1.19	1.03	1.12	1.16	1.08	1.31
Mn	1274	1430	1666	1217	1474	496
Mw	1279	1433	1672	1243	1532	511
PDI	1.0035	1.0025	1.0039	1.0216	1.0389	1.0302

---

The molecular weight (number average, M_n, and weight average, M_w) and poly-disparity index (PDI) of the resins are given in Table 2. The molecular weight of the resins increases with the increase of reaction time as well as the amount of branch generating moiety. However, the real molecular weight of the resins must be higher than the obtained value, as linear polystyrene is used as the standard in GPC analysis. The molecular weight distribution curves, with the results for the PHE4h resin obtained from three sets of analyses, are given in Fig. S4.‡

Curing study

According to the epoxy equivalent of PHE4h, 50 phr poly(amido-amine) hardener is required to cure it (as epoxy eq. wt : amine value = ∼400 : 200). However, in the case of the other hyperbranched resins, a slightly lower amount of hardener was needed. But, for a better comparison, herein the same amount of hardener was used. Curing of PHE2h, PHE4h, PHE6h, PHE5, PHE15 and PHE0 was done under the same conditions for the specified period of time (Table 3). PHE4h took the lowest amount of time for curing with the poly(amido-amine) hardener to obtain a swelling value of 20–30% (optimum cure), as shown in Table 3. This is due to the fact that PHE4h exhibited the highest DB as well as the lowest epoxy equivalent. Thus it has the highest number of chain end substitution epoxide groups to react with the poly(amido-amine) hardener. Thus it formed the best compact three-dimensional network structure, which resulted in the lowest swelling value. The hyperbranched epoxy resins can also be cured at room temperature (25 °C) and to obtain the optimum swelling value, PHE4h takes 2–3 days after mixing with the same hardener. After curing, the oxirane ring stretching frequency (916 cm⁻¹) of the PHE4h thermoset was completely diminished in the FTIR spectrum (Fig. S1‡). No unreacted epoxide groups were found in the spectrum. The other peaks remained intact in the spectrum. The densities of the cured PHE4h and PHE0 films are 0.98 and 1.02 g cm⁻³, respectively. The difference in density is due to the hyperbranched structure of PHE4h, which provides a high free volume between the molecules due to the confined geometry of the structure. A curing study of the resins was also performed using DSC analysis. However, they took less time to reach optimum curing compared to the normal curing process, as the curing was performed inside the closed chamber of the DSC sample pan. Whereas, for the achievement of high performance films for analysis, the resins were cured by a hot air oven process and the optimization of the curing reaction was done by determining the swelling values. In the case of normal heating, a glass plate was used so more time was required, as glass is a thermal insulator and the heat exchange capacity is very low and thus more energy is required to cure the resin which results in a higher curing time. The DSC curing curve for PHE4h is shown in Fig. S2.‡ The glass transition temperature (T_g) of the PHE4h and PHE0 thermosets is shown in Fig. 4. From the figure, it is seen that PHE0 exhibited a higher T_g value compared to the PHE4h thermoset, which demonstrated that it possessed higher rigidity or brittleness compared to the PHE4h thermoset.

Table 3 Performances of PHE2h, PHE4h, PHE6h, PHE5, PHE15 and PHE0 thermosets

Parameter	PHE2h	PHE4h	PHE6h	PHE5	PHE15	PHE0
Curing at 100 °C (min)	60 ± 3	35 ± 2	70 ± 5	50 ± 2	45 ± 2	90 ± 4
Swelling value (%)	29 ± 0.5	22 ± 1.0	31 ± 2.0	28 ± 1.2	24 ± 0.5	26 ± 0.5
Tensile strength (MPa)	38.5 ± 0.5	51 ± 1.5	37 ± 1.0	42 ± 2.0	47 ± 0.5	40 ± 2.0
Elongation at break (%)	26 ± 2.0	37.5 ± 1.6	49 ± 2.0	16 ± 1.2	28 ± 0.8	6 ± 0.8
Toughness (MPa)	785	1432	1308	511	961	181
Modulus (MPa)	411	345	278	665	359	3360
Impact strength (cm)	>100	>100	>100	>100	>100	70
Scratch hardness (kg)	8.0 ± 0.5	>10.0	9.5 ± 0.3	8.5 ± 0.5	>10.0	6.5 ± 0.2
Bending (mm)	<1	<1	<1	>1	>1	3
Initial degradation temperature (°C)	289	296	291	290	292	291
Dielectric constant at 1 MHz (at 25 °C)	3.5 ± 0.06	1.8 ± 0.09	2.5 ± 0.11	3.2 ± 0.04	2.1 ± 0.05	4.2 ± 0.08
Dielectric loss at 1 MHz (at 25 °C)	0.025	0.009	0.015	0.019	0.012	0.033
Moisture absorption at 25 °C (%)	0.32	0.09	0.34	0.19	0.11	0.26
Adhesive strength (M–M) (MPa)	2436 ± 24	3429 ± 17	2258 ± 12	2670 ± 22	2919 ± 18	1028 ± 19
Adhesive strength (W–W) (MPa)	>2840	>2911	>2812	>2865	>2902	1114 ± 8
Adhesive strength (P–P) (MPa)	597 ± 9	789 ± 6	419 ± 11	663 ± 9	708 ± 5	256 ± 12

---

Fig. 4 DSC thermograms of PHE0 and PHE4h thermosets for measurement of T_g.

Mechanical properties

The mechanical properties, like tensile strength, elongation at break, toughness, impact resistance, scratch hardness and bending value, for the cured PHE2h, PHE4h, PHE6h, PHE5, PHE15 and PHE0 are given in Table 3. From the results it was found that the PHE4h thermoset exhibited the highest tensile strength value. This is due to PHE4h possessing the highest DB value, which resulted in the highest number of chain end substitution epoxide groups to form the best compact three-dimensional network structure after curing with the poly(amido-amine) hardener, as stated above. It also exhibited the highest value of elongation at break. This is because of the appropriate combination of aliphatic and aromatic moieties as well as the ideal hyperbranched structural architecture.⁴ This architecture helps to increase the free volume (the unoccupied space) between the molecules in the three-dimensional network due to the confined geometry of the structure.^4,5,24 Generally, in the literature, two types of epoxy thermosets have been reported, one is a glassy epoxy thermoset which exhibited high tensile strength (>40 MPa) with a very low elongation at break (<5%), and the other is a rubbery epoxy thermoset which exhibited low strength (<10 MPa) with a high elongation at break (>40%).²⁵ However, the combination of both these types is rarely reported.⁵ Commercially available epoxy like EPON 828 thermoset also exhibited high brittleness though it showed a high tensile strength value (∼70 MPa).²⁶ Whereas the PHE4h thermoset exhibited both high tensile strength (51 MPa) and elongation at break (38%) due to the unique structural combination. Thus it also possessed the highest value of toughness (measured using the area under the stress–strain curves from Fig. 5) as well as impact resistance, scratch hardness and bending among the other studied thermosets. However impact resistance and bending value differences for all of the hyperbranched epoxy thermosets could not be measured as the values reached the highest limit of the instruments for impact resistance (100 cm) and flexibility evaluation (1 mm bending diameter of the mandrel). Another reason for the high impact resistance and flexibility of the hyperbranched thermosets, besides their aromatic–aliphatic combination and structural architecture, is the presence of different flexible moieties (hydrocarbon chain of the hardener and ether linkages of the epoxy).^4,5,27 Optical images of the PHE4h thermoset films are also shown in Fig. 6, to allow observation of its transparency and flexibility. The scratch hardness value difference between the thermosets of PHE4h and PHE15 also could not be measured as these values also reached the highest limit of the instrument for scratch hardness (10 kg). On the other hand, the linear epoxy (PHE0) thermoset exhibited the lowest values of elongation at break, as well as toughness, though its tensile strength value is high. This is due to its linear structure as well as it containing more of the rigid aromatic bisphenol-A moiety in its structure, which also resulted in its higher T_g value in Fig. 4. Thus it also possessed the lowest value of impact resistance, scratch hardness and bending. Due to the high brittleness and rigidity, PHE0 also exhibited an extremely high tensile modulus compared to the hyperbranched epoxy thermosets, as shown in Table 3.

Fig. 5 Stress–strain profiles of the thermosets.

Fig. 6 Optical images of PHE4h thermoset films for observation of (a) transparency and (b) flexibility.

Adhesive strength

The lap-shear tensile adhesive strength values for the PHE2h, PHE4h, PHE6h, PHE5, PHE15 and PHE0 thermosets are given in Table 3. The PHE4h thermoset exhibited the highest value of adhesive strength among all the other thermosets and this is due to the same reasons as for the tensile strength. It has the highest number of end terminal polar groups (due to the highest value of DB), which helps to adhere the substrates by strong polar–polar, H-bonding, etc. interactions (W–W) and by physical interlocking (M–M and P–P). However, in the case of the W–W substrate, all of the hyperbranched epoxy thermosets showed almost equal adhesive strength due to the failure of the substrate. On the other hand, the PHE0 thermoset possessed poor adhesive strength values for all of the three types of substrates compared to the hyperbranched thermosets.

Thermal stability

TGA thermograms of the PHE2h, PHE4h, PHE6h, PHE5, PHE15 and PHE0 thermosets are shown in Fig. 7. The initial decomposition (5% weight loss) temperatures of the thermosets are given in Table 3. The initial and midpoint (50% weight loss) degradations of the thermosets were found at around 300 and 400 °C, respectively. The thermosets are degraded mainly by two step patterns, where the first stage (∼300 °C) is related to the degradation of the aliphatic moieties of the hyperbranched epoxy (ether and pentaerythritol moieties) as well as the poly(amido-amine) hardener (long chains of the fatty acid moieties), and the second stage (∼400 °C) is related to the degradation of the bisphenol-A aromatic moieties of the hyperbranched epoxy. The degradation patterns and the degradation temperatures of the thermosets were almost the same as their structural units are almost same. However, due to the highest crosslinked and compact structure, the initial decomposition temperature (296 °C) of the PHE4h thermoset was found to be the highest. On the other hand, as the PHE2h thermoset contains more aromatic bisphenol-A moieties in its structure compared to the aliphatic moieties, its second stage degradation temperature (420 °C) and char residue (8%) were found to be slightly higher compared to the hyperbranched thermosets.

Fig. 7 TGA thermograms of the thermosets.

Electrical properties

The dielectric constant and dielectric loss values at a 1 MHz frequency and 25 °C for the PHE2h, PHE4h, PHE6h, PHE5, PHE15 and PHE0 thermosets are given in Table 3. The variations of the dielectric constants with frequency (0.1–1.0 MHz) at 25 °C of the thermosets are shown in Fig. 8. From the results, it was found that PHE4h exhibited the lowest dielectric constant due to the highest crosslinked and compact structure resulting from its ideal hyperbranched architecture. Here the authors are happy to announce that the synthesized hyperbranched epoxy thermoset possessed a comparatively low dielectric constant compared to any other conventional epoxy thermoset, like EPON 828, novolac, etc., as well as PHE0 (4.2), because of the unique hyperbranched structural architecture. Many publications have reported that the dielectric constant of a polymer can be reduced by increasing its porosity or free volume, or by decreasing its polarity.^21,28 From the density difference of the PHE4h and PHE0 films, as well as from the SEM micrograph (Fig. S3‡) of the fracture surface of the PHE4h film, it was found that there are no bubbles or porosity inside the film. In this case, the hyperbranched structural architecture can led to high free volume among the molecules even after curing.^21,29 The free volume of the thermoset is, therefore, large and the dielectric constant reached a low value (<2.0). This unique architecture may also help to prohibit the hydroxyl groups from restricting the molar polarization at a higher frequency, which decreases the dielectric constant as well as the dielectric loss. The moisture absorption value (Table 3) of the hyperbranched thermoset is very low because of the compact crosslinked structure, which is advantageous for advanced electronic applications.

Fig. 8 Dielectric constants of the thermosets with the variation of frequency.

Chemical resistance

The results of the chemical resistance (percentage of weight loss) for the thermosets in different chemical environments after 30 days of exposure are shown in Table 4. An excellent chemical resistance was shown by the thermosets in all the tested media, this is due to the compact and crosslinked structure of the hyperbranched thermosets. No weight loss was found for the PHE4h and PHE15 thermosets in any of the studied chemical environments. Only marginal weight loss was found for the PHE2h, PHE6h and PHE5 thermosets in basic and acidic media, due to their less compact structure compared to the PHE4h and PHE15 thermosets. On the other hand, due to the least compact structure of PHE0, the base and acid resistance were also found to be the lowest.

Table 4 Weight loss (%) of PHE2h, PHE4h, PHE6h, PHE5, PHE15 and PHE0 thermosets in different chemical media after 30 days of exposure

Chemical medium	PHE2h	PHE4h	PHE6h	PHE5	PHE15	PHE0
Aq. NaOH (5%)	0.64	0	0.30	0.58	0	0.77
Aq. HCl (10%)	1.92	0	0.54	1.69	0	2.34
Aq. NaCl (20%)	0.09	0	0	0.08	0	0.11
Aq. EtOH (20%)	0	0	0	0	0	0
Water	0	0	0	0	0	0

---

Conclusions

In this study industrially important low viscosity hyperbranched epoxy resins were synthesized using a single step A₂ + B₄ polycondensation reaction, with variation of the reaction time and amount of B₄ moiety. The structural architecture of the resins was successfully characterized using spectroscopic analyses. High tensile strength, good elongation at break, excellent toughness, high flexibility, good thermal stability, an ultralow dielectric constant, excellent adhesive strength and outstanding chemical resistance were achieved for the poly(amido-amine) cured hyperbranched epoxy resin obtained from a 4 h reaction time with 10 wt% B₄ moiety. Thus the unique hyperbranched architecture, with a combination of aromatic–aliphatic moieties, can offer high performance tough epoxy thermosets to overcome the shortcomings of commercial epoxy thermosets. The study also showed that the nature and amount of the branch generating moiety of the hyperbranched epoxy strongly influence the ultimate performance of the thermosets. The studied thermoset could also find a unique position as an ultralow dielectric constant adhesive with low moisture absorption in microelectric industries.

Notes and references

1. P. Huang, A. Gu, G. Liang and L. Yuan, J. Appl. Polym. Sci., 2012, 123, 2351–2359.
2. D. Zhang, D. Jia and Z. Zhou, Macromol. Res., 2009, 17, 289–295.
3. A. Sidorenko, X. W. Zhai, F. Simon, D. Pleul and V. V. Tsukruk, Macromolecules, 2002, 35, 5131–5139.
4. B. De and N. Karak, J. Mater. Chem. A, 2013, 1, 348–353.
5. B. De, K. Gupta, M. Mandal and N. Karak, ACS Sustainable Chem. Eng., 2014, 2, 445–453.
6. S. Barua, G. Dutta and N. Karak, Chem. Eng. Sci., 2013, 95, 138–147.
7. V. Coessens, J. Pyun and P. J. Miller, Macromol. Rapid Commun., 2000, 21, 103–109.
8. T. Emrick, H. T. Chang and J. M. J. Frechet, Macromolecules, 1999, 32, 6380–6382.
9. I. Gadwal, S. Binder, M. C. Stuparu and A. Khan, Macromolecules, 2014, 47, 5070–5080.
10. J. Yu, X. Huang, L. Wang, P. Peng, C. Wu, X. Wu and P. Jiang, Polym. Chem., 2011, 2, 1380–1388.
11. T. G. Kotch and A. J. Lees, Chem. Mater., 1992, 4, 675–683.
12. N. Chikhi, S. Fellahi and M. Bakar, Eur. Polym. J., 2002, 38, 251–264.
13. G. Qi, X. Zhang, B. Li, Z. Song and J. Qiao, Polym. Chem., 2011, 2, 1271–1274.
14. S. M. Shin, D. K. Shin and D. C. Lee, J. Appl. Polym. Sci., 2000, 78, 2464–2473.
15. M. Frigione and E. J. Calo, Appl. Polym. Sci., 2008, 107, 1744–1758.
16. H. S. Nalwa, Handbook of low and high dielectric constant materials and their applications, Academic Press, Boston, 1999.
17. L. N. Chellis, R. M. Japp, W. J. Summa, W. J. Rudik and D. W. Wang, US Pat. 5126192, 30 June 1992.
18. P. Y. Lin, G. P. Rajendran and G. E. Zahr, US Pat. 7658988 B2, 9 February 2010.
19. R. J. Wojnarowski and C. W. Eichelberger, US Pat. 5108825, 28 April 1992.
20. G. R. Viswanath, R. Thangaraj and S. Guhanathan, Int. J. Polym. Anal. Charact., 2009, 14, 493–507.
21. C. H. Lin, Z. R. Jiang and C. S. Wang, J. Polym. Sci., Part A: Polym. Chem., 2002, 40, 4084–4097.
22. H. J. Hwang, S. W. Hsu, C. L. Chung and C. S. Wang, React. Funct. Polym., 2008, 68, 1185–1193.
23. B. De and N. Karak, J. Chem. Sci., 2014, 126, 587–595.
24. J. P. Yang, J. K. Chen, G. Yang, S. Y. Fu and L. Ye, Polymer, 2008, 49, 3168.
25. T. Lan and T. J. Pinnavaia, Chem. Mater., 1994, 6, 2216–2219.
26. HEXOIN Speciality Chemicals, Product Data Bulletin, EPON™ Resin 828, Re-issued: September 2005, RP: 3075.
27. S. Pathan and S. Ahmad, ACS Sustainable Chem. Eng., 2013, 1, 1246–1257.
28. S. Devaraju, M. R. Vengatesan, M. Selvi, J. K. Song and M. Alagar, Polym. Compos., 2013, 34, 904–911.
29. J. Lv, Y. Meng, L. He, X. Li and H. Wang, Chin. J. Polym. Sci., 2012, 30, 493–502.

---

Footnotes

† Parts of the results in this manuscript are patented as an Indian Patent, application no. 786/KOL/2013 with complete specification on 19.02.2014.

‡ Electronic supplementary information (ESI) available: FTIR spectrum, DSC curing curve, SEM micrograph of the fracture surface and GPC results with molecular weight distribution curves for PHE4h. See DOI: 10.1039/c5ra04248h

---