Mechanical enhancement of ripples and dimples in CaCO3/low-density unsaturated polyester resin composites

Incorporation of different fine grain calcium carbonate into CaCO3/low-density unsaturated polyester resin (LDUPR) composites was studied and evident mechanical enhancement of CaCO3 on composites was investigated. Preliminary experiment results indicated that proper content of CaCO3 was less than 30.00 phr (parts per hundreds of resin), suitable preparation temperature range was from 72.0 °C to 80.0 °C, and initiator content was 1.80 phr. Optimal preparation conditions of CaCO3/LDUPR samples were obtained with the presence of 25.00 phr CaCO3 and 2.50 phr NH4HCO3 at 76.0 °C based on preliminary experiments. The lowest apparent density of A-CaCO3/LDUPR composite was 0.53 ± 0.02 g · cm−3 with a compressive strength of 20.27 ± 0.51 MPa · g−1 · cm3, and the highest specific compressive strength of the sample was 38.25 ± 1.43 MPa · g−1 · cm3. It is attributed to the hindrance to cross-linking between unsaturated polyester and styrene, and to the decrease of exothermic heat of the polymerization, which was caused by the existence of CaCO3. Unusual matrix microstructure with regular ripples and dimples formed by CaCO3, and the particular mechanical enhancement of regular ripples and dimples in composites were explored. ‘CaCO3 reefs’ concept, reefs-induced ripples, dimples of streams flowing, and resolution of external force with major force further being consumed models comprised the regulated mechanical enhancement of CaCO3 in CaCO3/LDUPR composites. This particular polymerization retarding and mechanical strengthening were obvious for the finest grain CaCO3.


Introduction
With increasing demand of economic efficiency and environmental protection, low cost and high specific strength product has received attention lately as for unsaturated polyester resin. In the case of the tendency, lowdensity unsaturated polyester resin (LDUPR) and its composite materials, which present less resin cost, low apparent density and high specific compressive strength, have been the growing highlight of studies for UPR composite materials [1][2][3].
The preparation of LDUPR and its composite materials have progressed over the past decade. Guo et al [4], Zhang et al [5] and Ji et al [1] presented new organic foaming methods or novel mechanisms of LDUPR preparation. Xu et al prepared LDUPR utilizing the cooperation of sodium bicarbonate and azodiisobutyronitrile [2]. At high temperature, Zhu et al explored the LDUPR preparation with the presence of sodium carbonhydrate/azodicarbonamide mixture [6]. Novel mild-thermal fabrication of chopped glass fiber/ LDUPR composite via NH 4 HCO 3 was presented by Guo et al [3].
Besides researches of LDUPR and fiber reinforced LDUPR, it is essential to incorporate fillers, such as calcium carbonate, silica, talc, alumina hydrates into resin to reduce the cost of composites and to improve the mechanical properties of composites [7,8]. Calcium carbonate is the most commercial inorganic filler exhibiting high volume, low cost, good stability, nontoxic, and non-black specialties [9][10][11]. Gupta et al [12] investigated the effect of CaCO 3 filler on fiber-reinforced polymer composites (FRP), and proposed a hybrid filler composition to optimize mechanical characteristics. He et al [13] pointed out that the toughening of vinyl Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI. ester resin composite, which was with the presence of calcium carbonate in different particle sizes, was 27%-45% higher than that of pure resin. The dominant feature of CaCO 3 in LDUPR and LDUPR composites is crucial for the versatility, the economics efficiency, and the mechanical properties of composites. Moreover, the addition of CaCO 3 might provide CaCO 3 /LDUPR composites with competitive performances in case of using LDUPR alone. However, the effects of CaCO 3 on CaCO 3 /LDUPR composites preparation, on the curing of UPR during the process, and on the microstructural specialties reported so far are incomplete or shallow.
In this study, three kinds of fine grain CaCO 3 in different particle sizes were used to prepare CaCO 3 /LDUPR composites. The effects of fine grain CaCO 3 on the preparation of CaCO 3 /LDUPR, on the cross-linking between unsaturated polyester and styrene, on the microstructure evolution of CaCO 3 /LDUPR, and on the improvement of CaCO 3 /LDUPR mechanical properties were investigated and innovated strengthening mechanism was presented. Tert-butylperoxy benzoate (TBPB, 98 wt%) was the initiator for UPR polymerization, which was produced from Akzo Nobel N.V, Netherland.
PMR-EZ, as an external release agent, was produced from CHEM-TREND Co., Ltd., USA. Dioctyl phthalate (DOP) was the solvent applied in volumetric gas-burette experiments obtained from Shanghai Ling Feng chemical reagents Co., Ltd., China, and its content 99 wt%.

Viscosity testing
Referring to the ISO 3219.1993 standard ('Unsaturated polyester resin viscosity measurement') the viscosity of resin glue was measured by use of digital rotatory viscometer NDJ-79A, which was made by Shanghai Changji Geological Instrument Co., Ltd. China. The sensor accuracy was 2% F.S. and the relative humidity was 80%.

Decomposition of NH 4 HCO 3 measurement
The decomposition of NH 4 HCO 3 was detected through a volumetric gas-burette experiment [14,15]. Decomposition temperature-bleed volume curve of NH 4 HCO 3 was heated from ambient temperature (20.0°C) to 100.0°C at the heating rate of 2°C·min −1 , and bleed volume curves of NH 4 HCO 3 decomposition at a certain temperature were also analyzed.

Gel time determination
GT-2a gel time meter (Lin'an Fengyuan Electronics Co., Ltd., China, with a temperature accuracy of ±0.1°C, and working temperature 230°C) was applied to detect the gel time of UPR glue according to the standard ISO 2554:1997 ('Test Methods of Unsaturated Polyester Resins').

Properties detection of CaCO 3 /LDUPR composites
Apparent density (ρ) of sample was tested according to standard ISO 845-2006. Compressive strength (P) of sample was measured according to standard ISO 844:2014 at an ambient temperature of 23±2°C, and relative humidity was 50±5% utilizing electronic universal testing machine (WDW3100, produced by Changchun Xinke Instrument Co. Ltd., China, the accuracy was 0.5%, and the maximum pressure was 100 kN). A parallel experiment including five replicated samples was processed for each formulation.

Thermal analysis
Non-isothermal differential scanning calorimetry (DSC) of samples was examined by using NETZSCH DSC 204 (NETZSCH Scientific Instruments Trading Ltd., Germany), where the nitrogen atmosphere of samples sealed in an aluminum crucible was kept at 30 ml·min −1 , the temperature accuracy of the instrument was ±0.1°C, the heating rate of the experiment was 10°C·min −1 , and the mass of the sample was 5-10 mg.

Scanning electron morphology
Fracture surfaces of samples were observed by SEM (JEOL JSM-6510, Japan) at an accelerating voltage of 15 kV in high-vacuum mode. Surface sections of samples were sputtered with gold to enhance the electronic conductivity.

Results and discussion
3.1. Effect of CaCO 3 on the viscosity of resin glue and on the decomposition of NH 4 HCO 3 CaCO 3 affected the viscosity of resin glue. The maximum content of CaCO 3 was suggested to be controlled within 65.00 phr for UPR [16]. In this study, the content of CaCO 3 was controlled no more than 35.00 phr, considering the motion and the distribution of bubbles (which were caused by decomposition of NH 4 HCO 3 ) in resin glue. Viscosity changes of UPR glue, which contained variant contents and different particle size CaCO 3 , were illustrated in figure 1. Figure 1 shows that, with the presence of CaCO 3 30.00 phr, the viscosity of UPR glue is higher than 4000 mPa s accompanied with difficult distribution of CaCO 3 in resin glue.
Based on the above viscosity experiment, 10.00 phr, 20.00 phr, and 30.00 phr CaCO 3 was applied separately to examine the effect of particle size of CaCO 3 on the thermal decomposition of NH 4 HCO 3 (see figure 2). Figure 2(a) indicates that the temperature range of the maximum thermal decomposition rate of the NH 4 HCO 3 was from 70°C to 86°C, for three different particle sizes of CaCO 3 . The gel time of resin glue in the preparation of CaCO 3 /LDUPR should be controlled in the range of 25.0 min to 35.0 min [17][18][19], where, the gel time of resin glue lasted 4.0 min [20,21]. Figure 2(b) shows that the bleed time of NH 4 HCO 3 was 39.3-28.9 min in the temperature range of 72.0-80.0°C. Therefore, the most suitable preparation temperature range of CaCO 3 /LDUPR sample was set from 72.0°C to 80.0°C.
Considering the dosage of initiator should be controlled no more than 3.00 phr [1,18] in practice, the adding amount of initiator was tried to set as 1.50 phr, 2.00 phr, and 2.50 phr to detect the change of gel time. Under the condition, the gel time of resin glue at 72.0°C, 76.0°C, and 80.0°C was detected separately, where, both CaCO 3 (with the presence of 10.00 phr, 20.00 phr, and 30.00 phr separately) and NH 4 HCO 3 (with the presence of 1.00 phr, 2.00 phr, 3.00phr separately) existed in resin glue. These gel time experiment results are shown in table 1. Table 1 illustrates that proper adding amount of initiator should be between 1.50 phr and 2.00 phr. After that, 1.80 phr initiator was set as the optimal content of initiator, and gel time experiment was repeated. It is concluded that the most suitable gel time was 25.3-34.2 min corresponding to 1.80 phr initiator.    3 foaming agent (X) should be controlled no more than 3 phr [3,22]. Values of Y and T were determined in terms of preliminary experiments above, which Y value was no more than 30.00 phr and T was in range from 72.0°C to 80.0°C. Under the condition, CaCO 3 /LDUPR samples were prepared through orthogonal experiment. Orthogonal experiment is a kind of multifactor and multilevel test design method, which can substantially reduce the number of tests under the condition of ensuring the representativeness of the test [23,24]. In this study, an orthogonal experiment of L 25 (5 3 ) (where L is the code name, 25 is the number of test, 3 is the number of different factors, and 5 is the level number for one factor [3,5]) was designed for the preparation of CaCO 3 /LDUPR. X, Y, and T were three different factors. Level numbers for X were varying from 1.00 phr to 3.00 phr with an interval of 0.50 phr. Level numbers for Y were from 10.00 phr to 30.00 phr with an interval of 5.00 phr. Lever numbers for T were set from 72.0°C to 80.0°C with an interval of 2.0°C. CaCO 3 /LDUPR samples were cured at a preparation temperature for 2 h, and then were cooled down to room temperature and got by demolding. In the orthogonal experiment, a parallel experiment including five replicated samples was processed for each formulation. The orthogonal experiment results are listed in table 2.
Effects of three experimental factors on the ρ and the P of CaCO 3 /LDUPR samples were analyzed, and results are summarized in table 3. In table 3, K a is the average of ρ and K s is the average of P, which were calculated separately from five values at a level of one single factor. R a is the range of K a between K a max and K , a min and R s represents the range of K s between K s max and K .
s min As shown in table 3, the R a value of ρ is 0.10 g·cm −3 for the preparation temperature of CaCO 3 /LDUPR at 76.0°C. Effects of the content of A-CaCO 3 and that of NH 4 HCO 3 on ρ, P, and on P S of A-CaCO 3 /LDUPR samples at 76.0°C are illustrated in figure 3.

Mechanical enhancement mechanism 3.3.1. Retardation of CaCO 3 to the cross-linking of LDUPR
The effects of three kinds of CaCO 3 on the curing process of UPR were characterized by non-isothermal DSC experiments and is shown in figure 4.
With the content of CaCO 3 increase, the broad exothermic peak of cross-linking between unsaturated polyester and styrene widen and the cross-linking curing rate decreases, accompanied with lower curing reaction heat. Therefore, it is revealed that the cross-linking of UPR was retarded by CaCO 3 , which could improve the mechanical properties of CaCO 3 /LDUPR samples [25,26].

Mechanical enhancement of CaCO 3 particles
A fundamental principle, which dislocation motion was under constraint of grain boundary resistance with an external force, was put forward [27][28][29]. Under the condition, mechanical strengthening of fine grain for sample became more pronounced. Hall-Petch formula [30,31] describes the strengthening principle as: Where ΔR G is the strengthening increment, K G is the constant, and d represents the particle diameter. It is evident that smaller particle size produced stronger mechanical strengthen for sample. Therefore, combined with orthogonal experiment results above, it is deduced that A-CaCO 3 , which owned the smallest particle size, performed the strongest mechanical strengthen for CaCO 3 /LDUPR composites compared with B-CaCO 3 and C-CaCO 3 .

Ripples formation and mechanical enhancement
Sections of UPR, LDUPR, and A-CaCO 3 /LDUPR cured samples are shown in figure 5. Figures 5(a)-(c) illustrate that cured UPR presents flat morphology with micro cracks and micro voids. Figures 5(d)-(f) shows that there are few ripples in the resin matrix of LDUPR sample as there is no CaCO 3 in LDUPR. Figures 5(g), and (h) shows that there are homogeneously distributed bubbles and regular ripples in resin matrix of the cured LDUPR sample with 25.00 phr A-CaCO 3 , where the specific compressive strength of the sample was 41 percent higher than that of LDUPR sample. The specific formation of ripples and the enhancement of ripples to CaCO 3 /LDUPR composite, which were not declared before, were explored. During the curing process of UPR without the presence of CaCO 3 , the viscosity of resin glue increased accompanying the cross linking between unsaturated polyester and styrene, self-polymerization of styrene, and the vaporization of styrene. Homogeneous resin glue gelled and then cured into solid exhibiting micro voids and micro cracks in the matrix of sample.
Fine particles of CaCO 3 immerged and heaped in resin glue, which could be described as 'CaCO 3 reefs' in resin glue. These 'CaCO 3 reefs' moved freely in resin glue but became obstacles of glue fluid when resin glue viscosity had suddenly increased. During the process, resin glue fluid swashed against the 'CaCO 3 reefs', and tiny )/ ρ i is an apparent density for a single factor, where i is level among level 1 to level 5. R a : the range between K max and K min . K s the mean of P calculated from five values under one level for a single factor, ) / P i is a compression strength for a single factor, where i is level among level 1 to level 5. R s : the range between K max and K min . The superscripts 'pt', 'ab' and 'cc' represent preparation temperature, ammonium bicarbonate and calcium carbonate, respectively.
waves appeared around these 'CaCO 3 reefs' during the stick-flow motion of resin glue. After that, ripples generated in the resin matrix till resin cured (see figure 6(a)). Ripples were achieved in the cured resin matrix of sample and they improved the mechanical property of sample. More 'CaCO 3 reefs' were in resin glue with the content of CaCO 3 increasing, and more ripples generated in cured resin matrix. Figure 6(b) indicates that exceed content of CaCO 3 acted as more 'CaCO 3 reefs' in resin glue for 30.00 phr A-CaCO 3 /LDUPR composite sample, where ripples were extruded and broken to 'piece ripples'. It is disadvantageous to the mechanical enhancement of CaCO 3 in composite. Therefore, the specific compressive strength of 30.00 phr A-CaCO 3 /LDUPR composite was 25.15±1.20 MPa·cm 3 ·g −1 . These changes  and gradually moved slowly. After that, dimple with sharp edge flow generated along with resin glue being cured. During the process, a constant stream moved ahead and mingled with another stream caused by another particle of CaCO 3 . Two constant streams flowed through the resin glue gently with distinguished boundaries left, and dimples generated (shown in figure 5(i)). Therefore, the specific dimple exhibited an image of branch divided from trunk of a tree in the cured resin matrix. In accordance with this pattern, one main dimple mingled with several other dimples and generated a typical ductile dimple pattern in which a main dimple being branched into several small dimples.
In light of these extraordinary microstructures of ripples and dimples, major force transmitted along the microstructure, changed its propagation direction against the along dimples. Accordingly, resolution of external force occurred, and the major force was consumed by regular ripples and dimples in resin matrix, resulting in the improvement of compressive strength of the sample. This innovative force resolution mechanism is described in figure 8.

Conclusions
Three kinds of fine grain CaCO 3 in different particle sizes were used to prepare CaCO 3 /LDUPR composites in this study and their evident mechanical enhancement of on CaCO 3 /LDUPR composites was investigated. The finest grain A-CaCO 3 performed the most outstanding actions, including the lowest viscosity effect, the best bubble distribution and the most excellent mechanical strengthening among three kinds of fine grain CaCO 3 . Under the optimal conditions of 25.00 phr A-CaCO 3 and 2.50 phr NH 4 HCO 3 at 76.0°C, A-CaCO 3 /LDUPR exhibited the lowest apparent density was 0.53±0.02 g·cm −3 , the compressive strength of 20.27±0.51 MPa, and the highest specific compressive strength of 38.25±1.43 MPa·g −1 ·cm 3 .
The cross-linking between unsaturated polyester and styrene was hindered by CaCO 3 and the exothermic heat of the polymerization was decreased. Unusual matrix microstructure with regular ripples and dimples in CaCO 3 /LDUPR composites was presented and explored. It is deduced that resin glue fluid swashed against the 'CaCO 3 reefs' and tiny waves appeared around these 'CaCO 3 reefs' during the flow motion of resin glue. In smaller scale, resin glue fluid streamed around particles of CaCO 3 and particles parted the fluid stream. Several constant streams flowed through the resin glue gently and mingled with other streams exhibiting an image of branch divided from trunk of a tree in the cured resin matrix. With the extraordinary microstructure of ripples and dimples, major force transmitted along and changed its propagation direction against the ripples starting and was resolved into several components along dimples. Under the regulation mechanism, resolution of external force was carried out and major force was consumed by regular ripples and dimples, which resulting in an improvement of mechanical enhancement of CaCO 3 /LDUPR composites.