Kinked Bisamides as Efficient Supramolecular Foam Cell Nucleating Agents for Low-Density Polystyrene Foams with Homogeneous Microcellular Morphology

Polystyrene foams have become more and more important owing to their lightweight potential and their insulation properties. Progress in this field is expected to be realized by foams featuring a microcellular morphology. However, large-scale processing of low-density foams with a closed-cell structure and volume expansion ratio of larger than 10, exhibiting a homogenous morphology with a mean cell size of approximately 10 µm, remains challenging. Here, we report on a series of 4,4′-diphenylmethane substituted bisamides, which we refer to as kinked bisamides, acting as efficient supramolecular foam cell nucleating agents for polystyrene. Self-assembly experiments from solution showed that these bisamides form supramolecular fibrillary or ribbon-like nanoobjects. These kinked bisamides can be dissolved at elevated temperatures in a large concentration range, forming dispersed nano-objects upon cooling. Batch foaming experiments using 1.0 wt.% of a selected kinked bisamide revealed that the mean cell size can be as low as 3.5 µm. To demonstrate the applicability of kinked bisamides in a high-throughput continuous foam process, we performed foam extrusion. Using 0.5 wt.% of a kinked bisamide yielded polymer foams with a foam density of 71 kg/m3 and a homogeneous microcellular morphology with cell sizes of ≈10 µm, which is two orders of magnitude lower compared to the neat polystyrene reference foam with a comparable foam density.

. Schematic representation of the generalized process using supramolecular nanoobjects as heterogeneous cell nucleation agents for polymer foaming. The temperature and the relevant length scale, at which each individual step occurs, is indicated. The conceptual approach comprises: A) The preparation of a powder-powder mixture with the finely micro-sized supramolecular additive. B) Upon heating, the supramolecular additive is molecularly dissolved and both the additive as well as the added physical blowing agent is homogeneously dispersed in the polymer melt. C) Upon cooling, self-assembly of the supramolecular additive results into finely dispersed in-situ formed nanoobjects. D) Upon further cooling, these nanoobjects acts as heterogeneous nucleation sites for the foam cells, which subsequently grows during the foaming step on the nano and mesoscale. E) At room temperature, a solid foam with a microcellular foam morphology is obtained.

S2. Synthesis and characterization of kinked bisamides
General synthetic procedure 1 eq. of 4,4'-Diaminodiphenylmethane, 4,4'-Methylenebis(2,6-dimethylaniline) or 4,4'-Methylenebis(2,6-diethylaniline) was added under nitrogen to a mixture consisting of Nmethylpyrrolidone (NMP) or tetrahydrofurane (THF) and pyridine or triethylamine ( Figure S2). A small amount of LiCl was added and the solution was cooled to 0 -5 °C. 2.2 eq. of the respective acid chloride was added dropwise and the reaction mixture was heated to 20 -80 °C. The reaction mixture maintained at this temperature for several hours. After cooling, the reaction mixture was precipitated in ice water under vigorous stirring. The precipitate was filtered off and dried under vacuum for 12h. All compounds were further purified by recrystallization. Mass spectra (MS) were carried out on a FINNIGAN MAT 8500 spectrometer from Thermo-Fisher Scientific using electron spray ionization.
HPLC analyses were performed using an Agilent 1100 equipped with a C18 column Agilent Eclipse Plus (d = 3.5 µm, 4.6 x 150 mm). 0.1 mg -1.0 mg of the compounds were dissolved in 10 mL acetonitrile or NMP at 60 °C. 25 µL of the solutions were injected. As mobile phase an acetonitrile:water mixture (90:10 v/v %) was used with a flowrate of 0.3 mL/min. Detections were performed using a UV-detector at 250 nm and the peak values of the elution volume were reported. (4,1-

S3. Thermal properties of kinked bisamides
To reveal their thermal behavior at elevated temperatures and the phase transitions, the kinked bisamides were investigated by means of thermogravimetric analysis (TGA) and differential scanning calorimeter (DSC) under nitrogen atmosphere.

Thermal characterization methods
For TGA, an automated Mettler Toledo TGA/DSC 3+ operated under nitrogen atmosphere was used to determine the thermal behavior and the weight loss of the kinked bisamides. About 5 to 10 mg of the kinked bisamides were weighed in Al2O3 pans and heated from 50 °C to 700 °C at a rate of 10 K/min under nitrogen (60 mL/min). T-5wt.% value corresponds to the temperature at which 5 wt.% of the sample is lost.
Thermal phase behavior (melting points and crystallization behavior of the kinked bisamides) were determined by means of DSC. Here, an automated Mettler Toledo DSC 2/3 operated under nitrogen atmosphere was used. About 6 to 12 mg of the compounds were weighed into 30 µL high pressure crucible pans. The phase transitions for compounds 1a -1d, 2b, 2d, 3b and 3d were determined in a temperature range of 25 to 300 °C and for the compounds 2a, 2c, 3a and 3c in a temperature range of 25 to 350 °C, respectively. All measurements were performed with a heating and cooling rate of 10 K/min. Every heating and cooling scan was repeated three times.

Discussion of the thermal behavior
TGA experiments determine the temperature at which mass loss occurs. Mass loss can be either attributed to decomposition of the compounds or sublimation and evaporation, which should be avoided during polymer processing to ensure functioning of the self-assembly process and to keep the concentration constant. TGA reveals that all kinked bisamides show no significant mass loss at temperatures below 290 °C with T-5wt.% values being between 293 °C (3d) and 408 °C (1b). This finding renders all kinked bisamides in general suitable to be used as additives during processing of the polystyrene melt. Equally important are the temperatures at which phase transitions of the compounds in bulk occur, because the melting points as well as the crystallization temperatures can be regarded as a first approximation at which temperature the kinked bisamides dissolve in the polystyrene melt and at which temperature the supramolecular building blocks may self-assemble into nano-objects. In particular, the latter has to take place at a temperature before the foaming of the polymer melt is For kinked bisamide 3c it should be noted that the melting temperature is very close to the T-5wt.% value of 320 °C indicating also an evaporation or decomposition at this temperature. All kinked bisamides with phenyl substituents (b) exhibit reasonable high melting temperatures in the range of 225 °C to 258 °C without featuring a clear trend, which may be attributed to the partial conjugation over the amide group rendering a comparison complicated. The T-5wt.% values from TGA and the peak values of the endothermic transitions from the second heating and cooling run (DSC) are summarized in Table S1.

S4. PXRD investigation of kinked bisamide 3a
Recrystallization of kinked bisamide 3a from acetone (see section S1) yielded a crystalline white powder with sufficiently high order and was therefore subjected to powder X-ray diffraction (PXRD). A sample of 3a was filled in a capillary tube with a diameter of 0.5 mm and a tube wall thickness of 0.1 mm. PXRD measurements were carried out in Debye-Scherrer geometry on a STOE StadiP diffractometer using a Cu Kα1 radiation (λ= 1.5406 Å) and a curved germanium monochromator (oriented according to the 111 plane). The measurements were performed in the 2Θ range of 3 -50° with a step size of 0.015°.
Prior to the refinement of the crystal structure, the molecular structure of 3a was optimised using DFT methods with the CASTEP code [1]. Powder indexing and Pawley refinement was done with the programme TOPAS [2]. Simulated annealing with a rotational freedom of the ethyl substituents in ortho-position to the amide units and the cyclohexyl side-groups was performed with the Endeavour programme [3]. Moreover, additional rotational freemdoms were allowed for the CAr-CH2-CAr bonds.
The obtained solution was again geometry-refined with DFT methods. A final Rietveld refinement was done with TOPAS. An Rwp of 3.56 has been obtained, a preferred orientation of the 4 th order using spherical harmonics has been applied [4]. The x-ray diffractogram together with the Rietveld profile plot and the difference of the two for 3a recrystallized from acetone is shown in Figure S3. All relevant crystallographic data are summarized in Table S2. A CIF (crystallographic information file) is provided as additional supporting information. Figure S3. X-ray diffraction pattern (blue) together with the Rietveld profile plot (red) and the difference plot (grey) of 3a recrystallized from acetone.

S5. Solid-state NMR spectroscopy of kinked bisamide 3a
Solid-state NMR spectroscopy measurements have been performed on recrystallized samples from acetone (see section S1) and self-assembled samples from xylene of kinked bisamide 3a. To prepare the latter, 500 ppm of 3a was molecularly dissolved in xylene by refluxing the mixture for 30 minutes.
Subsequent cooling to room temperature yielded a turbid dispersion. Evaporation of the solvent under high vacuum yielded a white powder, which was used for the measurement.
Solid-state NMR spectroscopy has been performed on a Bruker Avance III spectrometer operating at a proton resonance frequency of 400 MHz. The two differently prepared samples were packed in 3.2 mm zirconia rotors and analysed in 3.2 mm triple resonance probe. High resolution proton NMR spectra have been recorded under Magic-Angle-Spinning (MAS) with a spinning frequency of 11.5 kHz using the DUMBO decoupling sequence [5]. The 13 C NMR spectra were acquired with a cross polarisation pulse programme at 12.5 kHz MAS frequency.  [7]. The CP contact time was set to 3 ms.
The number and chemical shifts of the resonances match the molecular structure and the fact that just one molecule is present in the asymetric unit. The sample that is recrystallized from xylene shows the same amount and position of the resonances as the one recrystallized from acetone ( Figure S4). This strongly hints that the crystal structure and packing are the same. However, all resonances are broader, indicating less ordering in the sample. Figure S4. Solid-state NMR spectroscopy of powder samples of 3a. 1 H solid state NMRs (A) of 3a prepared from acetone (blue) and xylene (red), respectively. 13 C solid state NMRs (B) of 3a prepared from acetone (blue) and xylene (red), respectively.

S6. FTIR spectroscopy of kinked bisamide 3a
Solid powder samples of kinked bisamide 3a for FTIR spectroscopy were prepared from acetone and xylene as described in section S1 and S4. FTIR spectra were measured with a PerkinElmer 100 FTIR spectrometer equipped with an ATR sampling accessory. Measurements were performed in a wave number range of 650 cm -1 to 4000 cm -1 with a resolution of 4 cm -1 .
The FTIR spectra of 3a from acetone (blue) and xylene (red) are shown in Figure S5

S7. Cell morphologies of polystyrene foams by batch foaming with kinked bisamides
Polystyrene foams with different supramolecular additve concentrations were prepared in a standard thermally-induced batch foam process as described in the experimental section comprising a master batch preparation, compounding and injection molding to polystyrene specimens, CO2-saturation of the specimens at room temperature and foaming at elevated temperatures. Each foam was characterized and evaluated in view of morphology, cell size, cell density and foam density.   Figure

Concentration-dependent evolution of the mean cell sizes of PS batch foams with kinked bisamides 1c, 2c and 3c
In Figure S8

Foam densities and volume expansion ratio of polystyrene batch foams with kinked bisamides
Foam densities were calculated by the water-displacement method in agreement with ISO 1183 based on the Archimedes principle using a Mettler Toledo XP 205 with density kit. For this, a small rectangle was cut out of the specimen and weighed in air (mair). Afterwards, the specimen was submerged in water and its apparent mass (mwater), which is reduced by the buoyant force, was measured. The resulting density (ρfoam) was calculated with equation 1.  Table S3.
All VER are summarized in Table S4.

Cell densities of polystyrene batch foams with kinked bisamides
Cell densities ρcell with respect to the unfoamed solid polymer were calculated according to equation 3, with Nc the number of cells in the selected area, AS the area of the selected section and the volume expansion ratio. The results of the cell densities are summarized in Table S5.

S9. Photograph of an extruded neat polystyrene foam and a polystyrene foam with 0.5 wt.% 3a
Foam extrusion experiments were carried out using a tandem extrusion line (Dr. Collin GmbH) comprising a twin-screw extruder with a 25 mm screw and a L/D ratio of 42 (A-Extruder) and a singlescrew extruder with a 45 mm screw and L/D ratio of 30 (B-Extruder) equipped with a slit die with a gap of 0.6 mm and a width of 30 mm. For the extrusion foaming with kinked bisamide 3a, a 5.0 wt.% masterbatch powder-powder mixture was used and foams were prepraed by diluting the masterbatch with neat PS granulates. 4 wt.% CO2 and 3 wt.% EtOH was used as physical blowing agent. As processing parameters, a screw speed of 8 rpm at the B-Extruder with an overall throughput of 4.5 kg h -1 was selected. The melt temperature in the A-Extruder was adjusted to 220 °C. The melt temperature near to the outlet of the B-extruder and the die temperature were selected between 110 -120 °C and 126 °C, respectively. The PS reference foam was prepared using neat PS granulates in the same manner. Figure S10 shows exemplarily photographs of non-calibrated extruded foams of a neat polystyrene reference foam and polystyrene foam with 0.5 wt.% 3a featuring almost the same foam density but very different cell sizes. Figure S10. Photograph of a neat extruded polystyrene foam (left) with a mean foam density of 61.6 ± 4.1 kg/m³ and a mean cell size of 1094 ± 377 µm and an extruded polystyrene foam with 0.5 wt.% 3a (right) with a mean foam density of 63.9 ± 1.9 kg/m³ and a mean cell size of 18.8 ± 8.5 µm, produced at very similar processing conditions.

S10. Photograph and light microscopy image of neat polystyrene foams by extrusion foaming
Extruded foams of neat polystyrene foam possess very large cell sizes, which cannot reliably be determined by means of scanning electron microscopy. Thus, photographs and light microscopy images were taken ( Figure S11) and evaluated. For this, a small rectangle was cut out of the neat polystyrene foam. The interface was colored black to make the cell morphology visible. Figure S11 A shows a photograph of an inhomogeneous cell morphology with cell sizes in the range between 0.1 mm and 2.0 mm. Figure S11 B shows a micrograph of a foam cell demonstrating that neat polystyrene foams feature comparable thick cell walls in the range between 50 µm and 150 µm.