Polyelectrolyte stabilized nanodiamond dispersions

Colloidal stability of negatively charged nanodiamonds (ND) has been realized with the help of double hydrophilic block copolymers poly(ethylene oxide)-block-poly(dimethylaminoethyl methacrylate)-dodecyl (PEO-bPDMAEMA-C12). The polymers were synthesized through RAFT polymerization of DMAEMA with a PEO macromonomer carrying trithiocarbonate and dodecyl end-groups. The NDs and the polymers were complexed by mixing them in different ratios. In addition to the amount of polymers, the effect of the detailed structure of the polymer was of interest and thus, also polymers with different lengths of the PEO-block were synthesized, as well as a block copolymer without the dodecyl end-group. The composition of the polymer, as well as the complexation conditions were varied to find the route to stable nanoparticles. The optimized complexation method gave colloidally stable ND particles with positively charged PDMAEMA coronas. The colloids were stable at room temperature and also in saline solutions up to 0.154M NaCl.


Introduction
Nanodiamonds (NDs) are an intriguing class of carbon nanomaterials with a diamond-like inner structure, rich surface chemistry and size in the submicron domain [1][2][3][4][5][6][7]. They have surfaced as a viable material to be used in place of and in conjunction with other carbon based nanomaterials especially in engineering [8,9] and biomedicine [10,11]. The useful properties of NDs for these applications include versatile and modifiable surface [5], electric insulation, high thermal conductivity [12], good mechanical-and chemical resistance [9], biocompatibility [13] and natural non-bleaching fluorescence [14,15]. These are a result of the diamond-like properties of the ND cores combined with the variable surface functionalities accessible by different ND production processes [16]. A number of different ways to prepare NDs are available, including so called static high pressure high temperature (HPHT), dynamic detonation methods, laser ablation, microwave plasma, chemical vapor deposition (CVD) and ultrasound cavitation [7]. Of these, the most often used are the HPHT and detonation methods.
Detonation methods are a cost effective approach for producing NDs in large scale [17]. In the most utilized detonation synthesis, two explosives having a combined oxygen balance under zero (mainly trinitrotoluene (TNT) and hexogen (RDX)) are detonated in a closed chamber yielding nanoscale sp 3 carbon soot. The soot is then collected and purified to yield the end product called detonation nanodiamond. With further processing the surfaces of the detonation NDs can be functionalized in multiple ways [5], leading to e.g. carboxylate-or amine functionalized diamonds. However, diamonds produced by detonation may also have very complex surface functionalities. This combined with their large surface to volume ratio leads to a tendency of NDs to bind to each other by different mechanisms forming tight core agglutinates, core aggregates, intermediate aggregates and agglutinates [18]. Aggregated NDs lose many of the attractive properties stemming from their small size and make the production of heterogeneous products or stable colloidal solutions a challenging task [2]. Studies of breaking the different aggregates of NDs have been reported by many groups [3,4,[18][19][20][21][22]. By either grinding [18], sonicating [21,22] or annealing [20,23], a colloidal dispersion of small ND particles is achievable. However, the long term stability [24], redispersibility and usability without further functionalization is difficult to achieve. To enhance these properties, methods to stabilize and functionalize NDs have been studied. Most of the approaches involve covalent functionalization of NDs with surface groups that enable longer term stability or bring additional functionality [25]. Azide-alkyne cycloaddition reactions on azide functionalized NDs have been shown to be a viable method to attach various moieties [26,27]. By attaching PEO or poly(methyl ether methacrylate) with NDs Zhang et al. demonstrated enhanced dispersibility in water [28].
The properties of NDs can also be enhanced by physical adsorption to the surface of NDs. For example doxorubicin has been complexed with negatively charged NDs in order to bind and release a drug upon varying the salt concentration [29]. Covalent and non-covalent approaches have also been combined interestingly by attaching proteins by electrostatic interactions to NDs followed by covalent protein immobilization [30]. Proteins bovine-and human serum albumin have been shown to complex fluorescent NDs and to prevent their aggregation in aqueous dispersions and be used for enhanced intracellular delivery [31]. According to the same report, block copolymers such as poly(ethylene glycol)-block-poly(dimethylaminoethyl methacrylate-cobutylmethacrylate) can also be used for this purpose. Hybrid ND:poly (ethylene imine) polyelectrolyte complexes have been utilized as vectors for siRNA delivery [32].
Polymeric dispersants can be often utilized to prevent colloid aggregation/instability. They function by adsorbing at the surface of the suspended particle and providing a protective layer that reduces particle-particle attraction and/or enhances solvent-particle interaction. PEO has been used to provide steric stabilization especially in the biomedical field [33,34]. Polyelectrolytes, on the other hand, are capable of providing electrostatic stabilization and may also enable charge inversion when oppositely charged moieties are used for complexation. PDMAEMA is a weak polycation that has been used with graphene to form colloidally stable multilayer structures sensitive to pH [35], to form pH-triggered antibiotic release systems when complexed with gelatine [36] and has also been used with carboxylated NDs to enhance their stability in biological fluids [31]. In addition, the polyelectrolyte is pH and temperature sensitive [37].
Knowing that amphiphilic PDMAEMA based block copolymers have been studied already [31], we wanted to take a step further in evaluating the effect of different structural units on the ND-polymer interaction. Multicomponent PEO-b-PDMAEMA-C12 polymers comprising of hydrophilic, hydrophobic and polyelectrolyte blocks are shown to enhance the stability and dispersibility of NDs in aqueous media. PEOblock lengths as well as the presence of the C12-group in the polymers were varied and their effect on the resulting polymer:ND complex properties and their colloidal stability were studied. The work demonstrates a promising approach for colloidally stable nanodiamond powder derived hybrid materials for bio-and materials applications.
For preparation of aqueous solutions and dispersions, deionized ultrapure water with~18 mΩ resistance was used. Glassware used to prepare light scattering samples and dilutions were carefully washed with methanol prior to measurement.

Synthesis of PEO-b-PDMAEMA block copolymers (I) and end-group removal (II)
The polymers were synthesized by RAFT polymerization reported earlier (Scheme 1, step I) [38]. Generally, 1 g of DMAEMA (6.35 mmol) and 46.2 mg of PEO-CTA (0.042 mmol) were weighed into a 25 mL round bottom flask followed by addition of 1 mL AIBN/dioxane solution (AIBN 0.0042 mmol/mL, 0.0042 mmol) and 4 mL of dioxane. Dissolved reagents were freeze-thawed three times under < 3 mbar vacuum to remove oxygen followed by addition of nitrogen atmosphere to the flask. The polymerization was initiated by immersing the flask into an 80°C oil bath with magnetic stirring. The reaction was left to stir in the bath for 6 h and was stopped by placing the opened flask into an icewater bath for 10 min.
The product was dialyzed in a MWCO 3500 g/mol cellulose membrane against deionized water for 2 days changing dialysis water at least 2 times a day. After dialysis, the product was freeze dried and stored in a freezer until used.
The removal of the dodecyl end-group was performed with a modified version of the method by Perrier et al. (Scheme 1, step II, Fig.  S1-2) [39]. 0.1 g of polymer a1 (end-groups 0.0016 g, 0.00624 mmol) and 0.0154 g of AIBN (0.0936 mmol) were weighed into a 25 mL round bottom flask. The flask was sealed and purged with Ar gas for 5 min under stirring. 5 mL of dried dioxane was added through a rubber septum and the mixture was left to stir with Ar bubbling for 35 min at room temperature. The reaction was started by immersing the flask into an 80°C oil bath with magnetic stirring. The flask was left to stir in the bath for 18 h and the reaction was stopped by placing the opened flask into an ice-water bath for 10 min. The product was purified by precipitation to cold n-hexane three times and dried in a vacuum. After purification, the dried product was dissolved in dioxane, freeze dried and stored in a freezer until used. Scheme 1. RAFT polymerization (I) and end-group removal (II).

Preparation of ND dispersions and polymer solutions
A Hielscher UP400S Ultrasonic Processor was used to disperse NDs and polymer:ND complexes using a H3 sonotrode at 400 W and a 0.5 duty cycle unless otherwise mentioned. The samples were placed into an ice-water-salt bath (IWS,~0°C) during sonication to prevent heating of the solutions.
Stock dispersions of ND were prepared by weighing the as-received ND powder into a 20 mL glass vial, adding deionized water, mixing by hand vigorously and sonicated for 30 min. All samples were refrigerated at +4°C when not used and sonicated at full duty cycle for 10 min prior use. The polymer solutions in water were made by dissolving polymer thoroughly into deionized water to 1 mg/mL concentrations and kept at +4°C prior use.

Polymer:ND complexation
For the complexation experiments, a Watson-Marlow solvent pump was employed to obtain a constant addition rate of polymer solvent (0.64 mL/min). Purification and fractionation of the complexes was performed with Sigma 2K15C centrifuge at 5000 RPM (3773g) at 20°C.
For the optimized procedure (see Supplementary data, Optimization of the complexation), 5 mL of refrigerated 10 mg/mL ND dispersion was measured into a 25 mL round-bottom flask with a large magnetic stirrer. The dispersion was sonicated in an IWS bath for 10 min at full duty cycle. Sonicated dispersion was moved into another IWS bath with magnetic stirring at 750 RPM. 15 mL of refrigerated 1 mg/mL polymer solution was added to the ND dispersion dropwise with a solvent pump. The mixture was left to stir for 24 h without replacing the IWS bath. After stirring, the dispersion was subjected to 120 min of sonication. For IR and TGA analysis, the complex was washed twice by centrifuging for 30 min, removing supernatant fraction and dispersing in fresh deionized water. Washed complex was dried in a 50°C oven and stored in a desiccator if not used.

Gel permeation chromatography (GPC)
The molar masses of the polymers were determined with a Waters Acquity APC -system equipped with Acquity APC XT 200 Å, 450 Å columns and UV-and RI-detectors. DMF + 1 mg/mL LiBr was used as an eluent and molar masses were compared to PMMA standards. Samples were prepared by measuring 2 mg of sample into 1 mL of eluent and letting the sample dissolve overnight on a shaking table at room temperature. Dissolved samples were filtered through 0.45 μm PTFE filters before measurement.

Nuclear magnetic resonance (NMR)
NMR-spectra were recorded with a Bruker Avance III spectrometer operating at 500 MHz for protons. For liquid state 1 H measurements, the samples were dissolved in deuterium oxide (D 2 O) to a concentration of 2 mg/mL. In the titration experiments a polymer sample in D 2 O (2 mg/mL) was titrated by ND dispersion in D 2 O (10 mg/mL), stirred and left to stabilize overnight. Stabilized sample was shaken by hand and sonicated in a bath for 30 min before measuring. Integrated solvent signal at 4.79 ppm was compared to signals at 3.71 ppm, 2.34 ppm and 1.32 ppm (Supplementary data, Fig. S1). The results were corrected to sample dilution.

Fourier-transform infrared (FTIR)
IR-spectra were collected with PerkinElmer Spectrum One using an attenuated total reflection (ATR) setup for polymers and transmission setup for ND-and complex samples. ND-and complex samples were prepared into a pellet by adding~5-10 wt% of sample to dried KBr, mixing with a mortar and pestle to a fine powder and pressed with 10 tons of pressure for 5 min. 4 scans were taken with a range of 450-4000 cm −1 .

Thermogravimetric-and evolved gas analysis (TGA-EGA)
TGA-EGA analysis was done using a Netzsch STA 449 F3 Jupiter simultaneous thermal analyzer connected to JAS-Agilent GC-MS (7890B/MSD5977A). 8-10 mg of sample was weighed into an Al 2 O 3 TG crucible and stabilized in the instrument at 40°C until there was no significant mass change. In the dynamic measurements done, the sample was heated from 40 to 1200°C with a heating rate of 5°C/min under a 40 mL/min helium flow. Half of the gas flow containing the evolved gases was led through a heated transfer line to JAS valve box, from where the flow was continuously sampled through a 60 cm long inert GC capillary acting as a pressure restriction to the MS detector (MSD5977A).

Transmission electron microscopy (TEM)
Transmission electron microscopy was conducted to complement the size data and visualize NDs and complexes. TEM images were taken with a Jeol JEM-2800. Samples with 0.1 mg/mL concentration in water were sonicated in a water bath for 90 min, blotted on an ultrathin CF300H-Cu-UL carbon support grid and dried at ambient atmosphere.

Dynamic light scattering (DLS) and Zeta Potential (ZP)
DLS and ZP measurements were made at 25°C with Malvern Zetasizer Nano DS instrument using a HeNe laser operating at 633 nm in a backscattering configuration. Samples were degassed and then prepared into folded capillary zeta cells with a concentration of 0.1 mg/ mL. ND and complex samples were prepared without filtering to prevent adsorption of samples into filters. Each sample was measured 5 times using the average of at least 3 measurements as the result. Measurements where the quality was low or cumulant-and distribution fit errors were too high, according to Zetasizer software, were not used. To give a full view on the particle size distributions, both number and intensity weighed distributions are presented and particle sizes are given as number mean values or Z-average values from 2nd order (cumulant) polynomial fit. Smoluchowski model was used in analysis of the electrophoretic mobility and zeta potential. Stability of the dispersions was measured from the samples stored in zeta cells at different periods of time.

pH
pH was measured to determine if the basic polymer (pKa ≈ 7.5) [40] and acidic carboxylic acid (pKa ≈ 4-5) [41] on the ND surface neutralize each other during complexation. pH of the solutions and dispersions (concentration 1-10 mg/mL in deionized water) were measured at room temperature with a VWR Phenomenal IS2100L using WTW or WVR electrodes calibrated within a day of the measurement with pH 4, 7 and 10 buffers.

Results and discussion
The PEO-b-PDMAEMA-C12 polymers were synthesized via RAFT polymerization and subsequent post-modification ( Table 1). The three polymers were synthesized to have the same PDMAEMA-block lengths, while the PEO-block length and presence of the hydrophobic C12 dodecyl end-group were varied. In order to investigate the interactions between NDs and polymers, the complexation process and the resulting complexes were first studied. Then complexation variables such as temperature, polymer to ND ratio and use of sonication were varied and their effect on the size distribution, pH and zeta potential of the complexes were monitored. Full details of the complex preparation are given in the Supplementary Data. Complexes with 20:1 (polymer:ND) molar ratio were chosen for further studies. These were then investigated to assess their size, surface charge and colloidal stability. For clarity, the polymers are labelled a1 (PEO17-b-PDMAEMA-C12), a2 (PEO105-b-PDMAEMA-C12) and b (PEO17-b-PDMAEMA) ( Table 1) and the corresponding complexes as C-a1, C-a2 and C-b (Table 2).

Complexation of polymers and NDs
The polymer:ND complexation was studied by 1 H NMR by adding ND dispersion to the polymer solution (Fig. S1). As the surface of NDs comprise mostly of carboxylic acids, especially the positively charged PDMAEMA-blocks are expected to interact with their surface. This would lead to lowered mobility of the polymer segments and consequently to suppressed NMR signal intensity. However, as the ND surfaces contain also other functional moieties capable of interacting with PEO or C12, their signals were monitored as well. The titration of polymers a1 and a2 by ND dispersion (Fig. 1) shows a decrease of signal intensities with increasing amount of ND added. By observing the change in signal intensities of block copolymer a1 (Fig. 1A), it can be concluded that all parts of the polymer lose mobility with the addition of NDs. When polymer a2 with longer PEO is studied (Fig. 1B) the PDMAEMA signal decreases similarly as with polymer a1, but the PEO and C12 signals are less suppressed compared to PDMAEMA. This shows that the electrostatic interaction between the ND carboxyls and PDMAEMA plays the main role in the complexation process. The less suppressed signal of PEO of a2 compared to a1 indicates that longer PEO-chains of a2 extend as dangling chains from the ND surface towards the solvent retaining their mobility. The situation is different for a1, where the short PEOs are embedded in the fuzzy interface. Analysis of the C12 signals in the case of a1 indicates that the dodecyl end-group is in very restricted environment due to the presence of hydrophobic interactions with the ND surface. For polymer a2, the signal from C12 suppresses less during titration, which can be related to the overall more hydrophilic nature of the a2 polymer due to the longer PEOblocks.
IR spectrum of the NDs (Fig. 2) reveals signals corresponding to OHgroups of absorbed water, (confirmed by measuring the sample again after heating at 130°C, Fig. S3), carboxyls and anhydrides [4,42]. The spectrum of the block copolymer is characteristic for PDMAEMA without obvious signals from PEO or dodecyl end-groups. For the purified and dried polymer:ND complex (C-a1) the signals of both polymer and ND are present. Complexes C-a2 and C-b show similar spectra as C-a1 (Fig. S4). Compared with the starting materials, the absorbances are shifted slightly indicating the binding of polymers onto the ND surfaces. Most clear indication of the interaction is the attenuation of the N-(CH 3 ) 2 signal at 2750 cm −1 , where the absorption band is nearly undetectable in the complex sample, as has been observed earlier for complexed PDMAEMA [43,44].
Thermal stability of NDs, polymer and complex were studied with TGA under He-atmosphere (Fig. 3). ND shows onset of degradation at 475°C and starting from this temperature up to 900°C. MS for the evolved gases shows a strong signal at m/z 44, assigned to CO 2 which may be assumed to form from degradation of carboxyl groups [45,46]. NDs retain 70% of their mass even when heated up to 1200°C but change their colour to dark black which can be related to high content of sp 2 carbon [47]. Polymer a1 shows a two-step degradation pattern starting at 155°C. Evolved gas analysis of the polymer below 300°C indicates nearly pure evaporation of DMAEMA until further heating leads to complex pyrolysis products and oxidized carbon.
The complex C-a1 shows a combination of the degradation patterns of pure ND and polymer with three distinguishable steps. Nearly all polymer degrades prior to the onset of ND degradation, enabling the estimation of the amount of polymer in the complex. The polymer mass of the purified complex according to TGA is around 20 wt%, which corresponds well to the mass ratio of the polymer in the feed (in this complex 23.2 wt%, Table S1). Thermal analysis data of complexes containing differential thermogravimetry curve (DTG) and extractedion chromatograms (EIC) can be found at Supplementary data (Fig. S5).
All in all, the IR, NMR and TGA data confirm successful complexation and that the purification does not selectively wash away either of the components.  S1).   Table 1. b Z-average size (Fig. S7). c Number mean size.

Complex characterization
When the as-received NDs are dispersed in water with stirring only, the particles remain as large aggregates and agglutinates. Sonication leads to narrower size distribution and decreased size (Fig. S6D). However, the dispersions sediment quickly due to re-formation of the larger structures. To minimize the complex size, achieve a neutral pH and maximize the surface charge of complexes for stability, the mixing ratios (polymer:ND), complexation temperatures and sonication times were varied (see Supplementary data for details, Optimization of the complexation). Generally, increasing the mixing ratio increases the size of the particles as well as their zeta potential. Initially negatively charged particles show positive zeta potential when enough polymer is added but aggregate when zeta potential is close to zero. Lowering the complexation temperature to 0°C from room temperature results in smaller complexes and increases the zeta potential to positive value at lower mixing ratios. Even at lowest mixing ratio studied (3:1) the pH is slightly basic compared to acidic ND dispersions. Sonication up to 120 min decreased the size of the complexes considerably.
The best conditions were chosen based on the smallest size, suitable final pH value and highest surface charge obtained. Complexes with a molar mixing ratio of 20:1 (polymer:ND), mixing temperature of 0°C and 120 min of sonication were chosen for further studies.
According to DLS the smallest complexes by number weighing are obtained by using polymer a1 with short PEO-block and C12-group ( Table 2, Fig. 4A, for particle size distributions obtained by intensity weighing see Fig. S7) showing similar particle size as the pure ND dispersion after sonication. The particle size then increases for complexes with polymer a2 with longer PEO-block, which is in line with the results from NMR experiments that suggest that dangling PEO-chains may extend towards the solvent from the complexes, increasing the observed hydrodynamic diameter. Complexes obtained from the polymer without the hydrophobic C12-group, C-b, show the largest size analysed by number weighing, which is also demonstrated in a morphological change in the TEM images discussed later. The importance of hydrophobic-hydrophilic balance has also been shown for dispersing carbon nanofibers by amphiphilic PDMAEMA copolymers [48] and for fullerene C60 with alkyne-functionalized poly(oxazoline)s [49]. In both cases polymers with hydrophobic moieties were better dispersants for these carbon materials leading to smaller sized aggregates.
Upon complexation, the acidic NDs and basic polymers neutralize each other leading to changes of pH compared to the original solutions (Fig. 4B). Acidic pH < 7 indicates free ND and basic pH > 8 (Table 1) indicates free, uncomplexed polymer. Average pH of the 20:1 complexes after sonication is around 7.4. Complexes C-a1 and C-a2 have a pH of 7.6 without sonication, a 2.5% difference to the sonicated solutions which can be regarded as measurement error. For these polymers, efficient complexation between ND and polymer without large mechanical stress applied to the solution takes place. As complexes C-a1 and C-a2 have different length PEOs, we can conclude that the length of PEO-block does not significantly change the binding between the polymer and ND. However, dispersion of complex C-b without the hydrophobic C12 has a pH of 8.5 before sonication. After sonication the  pH drops to 7.5. This shows that the binding of polymers to the ND surface is aided by the hydrophobic dodecyl end-group, which anchors the polymers to the hydrophobic sites on the NDs. After sonication the average pH of all of the 20:1 complexes is~7.4 which is a preferable pH for many applications.
All studied complexes have a ZP above +30 mV (Fig. 4C), in contrast to the pure ND dispersion having a ZP below −60 mV. The complexation thus induces a charge inversion of the particles. For all the dispersions sonication increases the ZP if the value is positive and decreases ZP if the value is negative. This is due to the breaking down of  the aggregates thus increasing the electrophoretic mobility of the particles being measured. The surface zeta potential > +40 mV of the sonicated complexes provides efficient electrostatic stabilization to the dispersed particles.
TEM imaging was used to visualize the NDs and their complexes (Fig. 5). The TEM images confirm large and dense aggregates for the pure ND samples, while the polymer:ND complexes are more finely and evenly dispersed at all magnifications. The complex C-a1 with the C12chain has smaller size and is clearly more dispersed than the complex Cb. This is in line with the DLS and zeta potential measurements for C-a1 and C-b highlighting the important role of the dodecyl end-group in the complexation process.

Colloidal stability
Stability of the ND and complex dispersions in deionized water was followed using samples stored at room temperature (Fig. 6). The pure ND dispersion aggregated considerably after 1 week of storage. The average diameter nearly doubled (Fig. 6A) and the ZP halved (Fig. 6B) on top of visual observation of sedimentation. The change in ZP can be explained through the size change. As the size increases, mobility decreases and surface to volume decreases causing smaller ZPs even though there is no chemical change in the charged groups. The polymer:ND complex samples showed a small increase in size and no significant changes in ZP over 5 weeks of observation time, apart from a small fraction of sediment seen by eye. This shows clearly how complexing NDs enhance the stability of the formed dispersion significantly and the composition of polymer has little to no effect on it.
Stability of NDs and complexes C-a1 and C-a2 in saline were measured up to isotonic concentrations. Samples prepared in different NaCl solutions were measured by DLS (Fig. 7A) and ZP (Fig. 7B) within 2 h of preparation. The ND dispersion showed increased size and greatly decreased ZP immediately when salt was present. The dispersion began to precipitate already at a concentration of 0.05 mol/L and was unmeasurable by DLS at higher salt concentrations due to complete sedimentation. Complexes C-a1 and C-a2 remained in solution upon increasing ionic strength. The size of C-a1 complexes increased when the salt concentration was above 0.05 mol/L. Complexes of C-a2 retained their size relatively unchanged over the whole concentration range studied, indicating that the increased length of the PEO-block improves the stability. ZP of both complexes decreased due to the screening of charges. The results show that the steric stabilization provided by PEOchains promotes the colloidal stability even though the charges are screened through increased ionic strength. Stability data according to Z-Average size is shown in Supplementary data (Fig. S8). In addition to the described benefits by complexation of NDs for use as dispersions, in some cases it would be preferable to store and use NDs as dry powder. However, dry ND powder is difficult to handle as it whirls easily as airborne small particulates and deposits on surfaces. Dried polymer:ND complexes yield redispersable (Fig. S9), non-powdery flakes that are easier and safer to work with.

Conclusions
Colloidally stable dispersions of negatively charged nanodiamonds (ND) were made with amphiphilic, cationic block copolymers. Combined with sonication, polymer:ND dispersions with large positive zeta potential, neutral pH, improved colloidal stability and particle sizes below 200 nm were obtained. Complexes prepared with block copolymers having different PEO-block lengths and/or with a hydrophobic C12-group resulted in differently sized aggregates with different morphology. More finely dispersed complexes were obtained with polymers bearing the C12-group, indicating the importance of hydrophobic interactions between the ND surface and polymers. The complex  dispersions retain their size and colloidal stability in water for extended periods of time. Their stability was also demonstrated in saline and found to be superior compared with pure ND dispersions that aggregate and sediment. Steric stabilization provided by the longer PEO-blocks further increased the colloidal stability of the dispersions against saline. Polyelectrolyte complexation of NDs thus provides a viable way to prevent the aggregation and clustering of their dispersions and provides a promising platform for functional, dispersible ND particles for a variety of material and biological applications.