Synthesis of Zn-based 1D and 2D coordination polymer nanoparticles in block copolymer micelles

Nanoparticles of the 1D and 2D coordination polymers [Zn(OAc)2(bipy)]n and [Zn(TFA)2(bppa)2]n were prepared, employing polystyrene-block-poly(4-vinylpyridine) diblock copolymers with different weight fractions of the 4-vinylpyridine (4VP) block and comparable overall molecular weights of Mn ≈ 155 kg mol−1 as template (SV-15 and SV-42 with 15 and 42 wt% 4VP, respectively). [Zn(OAc)2(bipy)]n nanoparticles were successfully synthesised within the 4VP core of SV-42 micelles, showing a core size of Dcore = 47 ± 5 nm and a hydrodynamic diameter of Dh = 157 ± 46 nm, determined by transmission electron microscopy (TEM) and dynamic light scattering (DLS). The crystallinity of the composite is quite low, showing only low intensity reflexes in the powder X-ray diffraction (PXRD) pattern with the highest particle load. No indications for larger microcrystals were detected by scanning electron microscopy (SEM), proving the successful integration of the coordination polymer nanoparticles within the micellar cores. Nanocomposites of the 2D coordination network [Zn(TFA)2(bppa)2]n were synthesised using both diblock copolymers. The particle core sizes (from TEM) and hydrodynamic diameters (from DLS) correlate with the 4VP fraction of the micelles, resulting in Dcore = 46 ± 6 nm for SV-42 and 15 ± 2 nm for SV-15 and Dh = 340 ± 153 nm and 177 ± 57 nm, respectively. The successful synthesis was proven by PXRD and SEM images, confirming the absence of larger crystallites. Hence, it is possible to synthesise nanocomposites of Zn-based 1D and 2D coordination polymers by a direct approach utilizing diblock copolymer micelles as template.


Introduction
Devices built up from functional molecular materials are an interesting approach to realize new functionalities for new elds of applications. Examples for promising molecule-based systems are porous coordination networks (MOFs, metal organic frameworks), Prussian blue based materials or molecular magnetic materials including spin crossover coordination polymers. [1][2][3][4][5][6][7][8][9] Nanoparticles and nanocomposites of such materials are oen considered to play a key role in future device engineering. [10][11][12][13][14][15] However, the synthesis of well-dened, stable nanoparticles or nanocomposites of molecule-based materials is a highly demanding task, as a wide range of techniques successfully used for solid state materials (e.g. the reduction of metal salts [16][17][18][19][20] or the hydrothermal synthesis [21][22][23] ) are inapplicable. For molecular materials, some synthetic procedures like the inverse micelle technique [24][25][26][27][28] or micro-uidic approaches using fast precipitation [29][30][31] have already been established to achieve that task. However, each new material has its needs regarding the reaction conditions (e.g. reaction temperature, solvent, reactant solubility, air or moisture sensitivity). Furthermore, some of the approaches have limitations regarding the size limits that can be reached. This makes a netuning of the reaction conditions indispensable to not only achieve a successful synthesis of the nanomaterial of the desired size, but also to preserve the desired properties. Furthermore, some synthesis procedures have been proven more suitable for the formation of functional materials than others, because they allow for example the even distribution of the nanomaterial or nanocomposite on surfaces or prevent the aggregation of the formed nanoparticles. 32,33 Nanoparticles of 2D [34][35][36] or 3D 24,37,38 coordination networks (CNs) have been prepared with a wide range of bridging ligands and metal ions. However, the formation of 2D and 3D CN nanoparticles directly in the core of block copolymer micelles is quite rare. To the best of our knowledge, only 6 examples of 2D or 3D CN nanoparticles formed in a polymer matrix can be found in the literature. [39][40][41][42][43][44] A more commonly used technique is the immobilization of pre-formed nanoparticles in block copolymer micelles or polymer matrices (bulk polymers, gels, etc.), [45][46][47][48][49][50][51][52][53][54][55] in some cases even size-selective employing polymer cages. 56 We have previously shown that the use of polystyrene-blockpoly(4-vinylpyridine) (PS-b-P4VP) diblock copolymers (BCPs) is ideal for the size-controlled synthesis of 1D Fe(II) spin crossover (SCO) coordination polymer (CP) nanoparticles with core sizes of 16 AE 2 nm and 48 AE 4 nm. It was possible to retain the SCO properties with hysteresis at both particle sizes. Thermal treatment of the 16 nm particles triggers a conned crystallization of the NPs leading to SCO properties comparable to those of the bulk material. 57,58 In other cases, the synthesis in connement results in different morphologies for NPs and bulk material and therefore different SCO properties. 59 Herein, we report the successful adaptation of our general synthetic concept to a completely new type of CPs and for the rst time to a 2D CN to illustrate its general applicability. The double-stranded 1D CP [Zn(OAc) 2 (bipy)] n 60 (bipy ¼ 4,4 0 -bipyridine) and the layer-like 2D CN [Zn(TFA) 2 (bppa) 2 ] n 61 (TFA ¼ triuoroacetic acid, bppa ¼ 1,3-di(4-pyridyl)propane) were used for the formation of Zn-CP/CN-BCP nanocomposites. The nanocomposites were synthesised using two PS-b-P4VP diblock copolymers (SV-15 and SV-42) as templates, which have an almost identical molecular weight but differ in the weight fraction of the 4VP blocks (see Table 1).

Synthetic procedures
The synthesis procedure was adapted from the literature and adjusted to the requirements of the Zn-based CPs (Scheme 1). 58 Dissolving the diblock copolymer in THF leads to the formation of BCP micelles due to the signicantly lower solubility of the P4VP block compared to the PS block. Thus the less-soluble P4VP core, where the synthesis of the NPs takes place, is surrounded by soluble PS corona chains. The nanocomposite samples containing the 1D CP [Zn(OAc) 2 (bipy)] n were synthesised employing SV-42 diblock copolymer micelles in THF (  (Table 2). Here, the synthesis protocol had to be adapted due to the very low solubility of the desired 2D CN. The respective BCPs were dissolved under reux in THF to trigger the self-assembly to micelles, [Zn(TFA) 2 ]$H 2 O was added and the mixture was heated to reux for 1 h to initiate the coordination of the zinc(II) precursor at the pyridine units in the P4VP core of the micelle.

Nanoscale Advances Paper
To avoid a precipitation of the CN and to decelerate its formation, the bridging ligand bppa was dissolved in THF and added dropwise to the reaction solution over 15 min, followed by a 1 h reux. The solvent was removed by rotary evaporation and subsequent drying in vacuo to yield light-yellow samples 5 and 6 (1 cycle each). The reaction procedure can be repeated to yield samples 7 and 8 (2 cycles each) with a higher complex loading. The formation of nanocomposites with higher cycle counts (>2) was tested, but the formation of microcrystals was observed by SEM (see Fig. S3 †). Therefore, no further addition of reactants was conducted aer the second addition of bppa (for experimental details see Experimental section).

Characterisation of nanocomposites
In total, eight different nanocomposites have been synthesised (Table 2), of which four contain the 1D CP [Zn(OAc) 2 (bipy)] n (samples 1-4) and another four the 2D CN [Zn(TFA) 2 (bppa) 2 ] n (samples 5-8). All nanocomposite materials were characterised by transmission electron microscopy (TEM) and dynamic light scattering (DLS) to evaluate the particle sizes in the dry state and in dispersion. Furthermore, the nanocomposites were analysed by elemental analysis (C, H, N), infrared spectroscopy (IR), powder X-ray diffraction (PXRD), and scanning electron microscopy (SEM). IR measurements were supported by computational calculations.
[Zn(OAc) 2 (bipy)] n nanocomposites IR measurements of the starting material [Zn(OAc) 2 ]$2H 2 O, the bulk material [Zn(OAc) 2 (bipy)] n and the samples 1-4 are displayed in Fig. 1A. The nanocomposites show a characteristic band at 1598 cm À1 , which increases in intensity relative to other bands when higher cycle counts are reached. This is in excellent agreement with the spectrum of independently synthesised bulk [Zn(OAc) 2 (bipy)] n , which features a band at 1600 cm À1 . Thus, this band can be safely assigned to the C]O stretching mode of the neat CP. Peak assignment in the ngerprint area between 1400 cm À1 and 1800 cm À1 proved valuable to identify the nature and purity of the nanocomposites, which was further supported by numerical frequency calculations of optimized model structures. The CP was approximated as binuclear [Zn 2 (OAc) 4 (py) 4 ], whereas the H-bond network of the precursor was taken into account in pentanuclear [Zn(OAc) 2 (OH 2 )] Â 4 [Zn(OAc) 2 (OH 2 )] (see Experimental section for computational details, animations of diagnostic modes are given in the ESI, † anim_1-6). In fact, the calculated C]O stretching mode in the CP model [Zn 2 (OAc) 2 (py) 2 ] n is located at 1601 cm À1 , almost identical to samples 1-4 and the bulk material. This is a distinct difference to the C]O band of the precursor [Zn(OAc) 2 ]$2H 2 O, which is experimentally found at 1549 cm À1 (computed value: 1534 cm À1 ). The formation of single-stranded [Zn(OAc) 2 (bipy)] n can be similarly ruled out, as C]O based stretching modes computed for the model [Zn(OAc) 2 (py) 2 ] are predicted at 1500 cm À1 , proving the successful synthesis of the 1D CP in the P4VP core of the SV-42 micelles.
Further proof is given by the PXRD patterns of the samples 1-4. Samples 1-3 are highly amorphous as indicated by the powder diffraction patterns. Only sample 4 with ve reaction cycles shows ve reexes that also correspond to the dominant reexes of the bulk material (Fig. 1B) indicating a successful formation of the CP inside the micellar core.
Exemplary for all nanocomposites with the [Zn(OAc) 2 (bipy)] n CP, the TEM and DLS measurements of sample 4 are displayed in Fig. 2. The corresponding core diameter and hydrodynamic diameter of all samples are summarized in Table 2. The DLS measurement shows narrowly distributed nanocomposite particles with a hydrodynamic diameter of D h ¼ 157 AE 46 nm. As the electron-rich [Zn(OAc) 2 (bipy)] n CP is incorporated inside the micelle core of the BCP, only the core of nanocomposite particles is clearly visible in TEM, resulting in notably smaller diameters compared to DLS. The TEM image of sample 4 shows spherical particle cores with a core size of D core ¼ 47 AE 5 nm (Fig. 2). In line with the results for other coordination polymers reported so far, 57-59 particles core sizes and hydrodynamic diameters of samples 1-3 are nearly identical and slightly increased compared to the empty template. Respective data of  Nanoscale Advances all samples conrming these results can be found in Fig. S4 and S5 † together with the autocorrelation function of sample 4 ( Fig. S6 †). The samples 1-4 were also characterised by SEM, revealing the absence of microcrystals on the sample surface (Fig. S7 †). Thus, the CP is regioselectively formed inside the cores of the BCP micelles.
[Zn(TFA) 2 (bppa) 2 ] n nanocomposites IR measurements were also performed for the four nanocomposites containing the [Zn(TFA) 2 (bppa) 2 ] n CN (samples 5-8, Fig. 3A 4 ], similarly gives two bands at 1668 cm À1 and 1662 cm À1 . For samples 5-7 only one band was detected at 1690 cm À1 , which is exactly between the two bands of the bulk CN. For sample 8 two bands for the CN were determined at 1699 cm À1 and 1684 cm À1 , being in good agreement with the bulk material. Again, a relative increase in intensity of the carbonyl band is detectable with higher cycles. Thus, it was possible to incorporate the 2D CN into the 4VP cores of both micellar templates (SV-15 and SV-42). In line with the PXRD results of the [Zn(OAc) 2 (bipy)] n CP nanocomposites, the samples 5 and 6 (one loading cycle) are completely amorphous as represented by the diffraction patterns. Nevertheless, samples 7 and 8 (two loading cycles) already show some reexes at positions that match with the bulk material, indicating the successful formation of the desired CN inside the BCP micelles (Fig. 3B).
Particle sizes of the nanocomposites were also analysed by TEM and DLS (Fig. 4, 5; S9-S11 †). While sample 5 only shows spherical particles with core sizes of D core ¼ 13 AE 1 nm (Fig. S9 †), sample 7 shows spherical particles which, however, tend to form chain-like aggregates (Fig. 4). This behaviour was also observed in other samples of the same nanocomposite (Fig. S12 †). In fact, the formation of spherical particles rather than worm-like micelles in THF would be expected in THF due to the low 4VP fraction of the utilised SV-15 diblock copolymer. 62,63 The presence of the anisotropic 2D CN together with the limited space available in the P4VP core of the highly asymmetric SV-15 BCP micelles (D core ¼ 15 AE 2 nm, D h ¼ 177 AE 57 nm for sample 7) could trigger the formation of chain-like structures, even at comparably low 4VP fractions. This may be an effect that occurs during drying of the sample on the TEM grid, since the hydrodynamic diameter distribution of sample 7 is rather narrow (Fig. 4C) and D h is only slightly increased compared to that of sample 5 ( Table 2).
The particle core sizes of samples 6 (D core ¼ 49 AE 4 nm,  Fig. 2 and S4 †), which were synthesised using the same BCP (SV-42, Table  2). This underlines the fact that the BCP determines the size of the nanocomposites. Again, the formed nanocomposite particles tend to form chain-like structures for sample 8 (2 loading    The SEM images for the samples 5-8 (Fig. 6, and S13 †) show the absence of microcrystals at the surface of the nanocomposites, proving that the 2D CN is incorporated inside the P4VP cores of the BCP micelles. However, if samples with more than 2 reaction cycles were synthesised, the very low solubility of the 2D CN [Zn(TFA) 2 (bppa) 2 ] n led to a fast precipitation of the CN, thus, resulting in the formation of microcrystals on the polymer surface and in the reaction solution. Consequently, the formation of truncated cuboctahedron crystals on the nanocomposite surface was observed by SEM (Fig. S3 †).

Conclusions
The synthesis of well-dened 1D, 2D and 3D coordination polymer (CP) and network (CN) nanoparticles is highly challenging. Self-assembled polymeric micelles derived from block copolymers (BCPs) that offer coordination sites inside the micellar core may be an elegant and generally applicable concept for the direct synthesis of these CP and CN nanoparticles (NPs). We were able to show that our established synthetic approach can be adapted to other 1D CP like the double-stranded [Zn(OAc) 2 (bipy)] n and more importantly to the 2D CN [Zn(TFA) 2 (bppa) 2 ] n . Employing micelles of the BCP SV-42 as template it was possible to achieve spherical NPs of the 1D CP [Zn(OAc) 2 (bipy)] n and the 2D CN [Zn(TFA) 2 (bppa) 2 ] n with nanocomposite core sizes of D core ¼ 47 AE 5 nm and D core ¼ 46 AE 6 nm, respectively. The average hydrodynamic diameter was determined to D h ¼ 157 AE 46 nm for the [Zn(OAc) 2 (bipy)] n and to D h ¼ 340 AE 153 nm for the [Zn(TFA) 2 (bppa) 2 ] n nanocomposites. Moreover, even smaller composite NPs of the 2D CN [Zn(TFA) 2 (bppa) 2 ] n were successfully prepared in SV-15 micelles, having a core size D core as small as 15 AE 2 nm and a hydrodynamic diameter of D h ¼ 139 AE 39 nm. No microcrystals were found on the nanocomposite surface as proven by SEM measurements. The crystallinity of the nanocomposite samples increases with the loading cycles, showing characteristic reexes in the PXRD at positions identical to the bulk materials. Since it was possible to synthesise NPs of the doublestranded 1D CP [Zn(OAc) 2 (bipy)] n and particularly the 2D CN [Zn(TFA) 2 (bppa) 2 ] n , we are convinced that our synthetic approach can be adapted to a wide range of other 1D, 2D, and even 3D CP and CN nanoparticles, which will be investigated in future work. The two polystyrene-block-poly(4-vinylpyridine) diblock copolymers (SV-15 and SV-42) were synthesised by sequential anionic polymerization of styrene and 4-vinylpyridine according to our previously published method. 57 For gel permeation chromatography (GPC) in N,N-dimethylformamide (DMF) with lithium bromide (5 g L À1 ), GRAM columns (300 Â 8 mm, 10 mm particle size, PSS Mainz) with 100 and 3000Å pore sizes were used. The samples were measured on a SEC 1260 Innity system (Agilent Technologies) at a ow rate of 0.5 mL min À1 at 23 C, using a refractive index detector (Agilent Technologies). The calibration was done with narrowly distributed polystyrene standards (PSS calibration kit) and toluene (HPLC grade) was used as internal standard.

Experimental section
MALDI-ToF MS (matrix-assisted laser desorption/ionization time-of-ight mass spectrometry) measurements were performed on a Reex III (Bruker) equipped with a N 2 Laser (l ¼ 337 nm). An acceleration voltage of 20 kV was used in linear mode and the samples were prepared according to the dried droplet method. Matrix (trans-2-[3-(4-tert-butylphenyl)-2-methyl-2-propenylidene]malononitrile, DCTB, 10 g L À1 in THF), analyte (10 g L À1 in THF) and salt (silver triuoroacetate, 10 g L À1 in THF) were dissolved and mixed in the ratio of 20 : 5 : 1 and 0.5 mL of the mixture was placed and dried on the target plate. This journal is © The Royal Society of Chemistry 2020 Nanoscale Adv., 2020, 2, 4557-4565 | 4561

Paper
Nanoscale Advances 1 H-NMR spectra were acquired with a Bruker Ultrashield 300 spectrometer using CDCl 3 as solvent.
Transmission electron microscopy (TEM) was conducted on a Zeiss CEM902 electron microscope (Zeiss, Oberkochen, Germany). Samples were dispersed in THF at a concentration of 2 g L À1 . The unltered solution was dropped on a carbon coated copper grid (mesh 200, Science Services, Munich). Electron acceleration voltage was set to 80 kV. Micrographs were taken with a MegaView III/iTEM image acquiring and processing system from Olympus So Imaging Systems (OSIS, Münster, Germany) and an Orius 830 SC200W/DigitalMicrograph system from Gatan (Munich, Germany). Particles size measurements were done with "ImageJ" image processing soware developed by Wayne Rasband (National Institutes of Health, USA).
Scanning electron microscopy (SEM) micrographs were taken on a Zeiss LEO 1530 GEMINI. The acceleration voltage was set to 3 kV and the sample was sputter-coated with a 1.3 nm platinum layer.
Dynamic light scattering (DLS) measurements were done with an AntonPaar Litesizer 500 in quartz glass cuvettes from Helma at 25 C in backscattering mode (175 ). One measurement consists of six consecutive runs. Samples were dispersed in THF at a concentration of 2 g L À1 . The unltered solution was used. A cumulant t was used for tting the experimental data.
Room temperature powder X-ray diffraction (PXRD) data were collected with a STOE StadiP X-ray diffractometer in transmission geometry between 5 and 30 2Q for all samples, which were placed on at surfaces. Cu-K a1 radiation (l ¼ 1.541 A) was used for the measurements together with a Mythen 1K detector.
For elemental analysis, the carbon, nitrogen and hydrogen contents were determined with a Vario EL III (Elementar Analysensysteme GmbH) with acetanilide as standard or at a Unicube (Elementar Analysensysteme GmbH) with sulfanilamide as standard. The samples were placed in tin boats and measured at least twice. The average of the measurements was used.
Transmission infrared spectra (IR) were collected on a Perkin Elmer Spectrum 100 FT-IR (ATR). The samples were measured directly as solids.

Computation setting
Theoretical structure calculations on the zinc(II) precursor complexes and coordination polymer/network models have been performed through density-functional theory (DFT) methods using the ORCA program package. 64 For all optimizations triple-x-valence TZVP 65 basis sets were used with the generalized gradient approximated functional BP86. 66 Grimme's third generation D3 correction of dispersion was used. 67,68 Medium effects were included in a dielectric continuum approach (COSMO), parameterized for acetonitrile; 69 the inclusion of a stationary dielectric background proved benecial for the match between experimental and theoretically observed structures. Optimized structures have been identied as stationary points through the absence of imaginary modes in harmonic frequency calculations; spurious low-frequency imaginary modes in some calculations due to -CH 3 rotations are typical artefacts of DFT-based numerical frequency scans. Coordinates of the computed structures are assembled in the ESI, Tables S1-S4. † Graphical presentation of the vibrational modes are also available (anim_1-6.gif).
Phenomenological approach. The input structure of [Zn(TFA) 2 (OH 2 ) 4 ] was extracted from the available XRD data. 70 In order to reduce computational cost, we have approximated the coordination polymers and the bulk [Zn(OAc) 2 ]$2H 2 O precursor as truncated model complexes. Thereby we have put emphasis on the conservation of the rst coordination sphere of the zinc centres. The bidentate bridging ligands were mimicked as monodentate pyridine ligands. In particular we employed the settings: [Zn(TFA) 2 2 ] n nanocomposites (samples 5-8). 50 mg of the diblock copolymer SV-15 were placed in a 50 mL ask tted with a magnetic stir bar. 20 mL THF were added and the polymer was dissolved under reux until complete dissolution. The polymer solution was cooled down to rt and 2.0 mg (6.5 mmol, 1 eq.) [Zn(TFA) 2 ]$H 2 O were added and the solution was reuxed again for 1 h. Subsequently, the reaction solution was cooled down to rt. 2.8 mg (14 mmol, 2.2 eq.) 1,3-di(4-pyridyl)propane (bppa) were dissolved in 10 mL THF and the solution was added dropwise over 15 min. Aer the addition of the ligand solution, the reaction mixture was reuxed again for 1 h. The synthesis can be stopped by removal of the solvent by rotary evaporation to obtain sample 5 (1 cycle). Alternatively, the synthesis procedure can be repeated exactly as before to synthesise sample 7 (2 cycles). The resulting light yellow solids were dried in vacuo.
Besides the adjustment of the reactants, the synthetic procedure for samples 6 and 8 using the diblock copolymer SV-42 is identical to samples 5 and 7, respectively. 5

Conflicts of interest
There are no conicts to declare.