Molecular Recognition with 2,4-Diaminotriazine-Functionalized Colloids

ARTICLE pubs.acs.org/Langmuir

Molecular Recognition with 2,4-Diaminotriazine-Functionalized Colloids Frank M. Bayer, Mingxue Tang, Rolf Michels, Claudia Schmidt, and Klaus Huber* Department of Chemistry, Physical Chemistry, University of Paderborn, Warburger Straße 100, 33098 Paderborn, Germany

bS Supporting Information ABSTRACT: New polymeric colloids functionalized with 2,4diaminotriazine residues have been prepared. The functionalities provide a triple hydrogen bond motif with a donor
acceptor
donor (DAD) pattern. The colloids are based on cross-linked poly-4-methoxymethyl styrene and are polymerized by means of surfactant-free emulsion polymerization. The reaction pathway including five steps was successfully tracked and verified via 13C CP/MAS solid-state NMR. Characterization of the colloids was done by combined static and dynamic light scattering and indicates a compact spherical particle shape. In solvents with the appropriate polarity, intercolloidal hydrogen bonding was enabled, including colloidal aggregation. In highly dilute solutions of THF, this aggregation was recordable by means of time-resolved static light scattering experiments. If THF was saturated with uracil, then aggregation could be completely inhibited. Uracil bears a triple hydrogen bond motif of the form acceptor
donor
acceptor (ADA) and is a direct antagonist of 2,4diaminotriazine. The charging of the colloids with uracil via hydrogen bond formation as a typical molecular recognition mechanism could be confirmed by IR spectroscopy.

’ INTRODUCTION Hydrogen bonds are highly specific and directional1,2 and hence offer new routes to shape 3D structures in chemical and biological systems.3,4 Lawrence et al.5 pointed out that biological materials with reversible binding patterns can provide advantages compared to materials exclusively based on covalent bonds. The formation of hydrogen bonds, for instance, leads to new properties of the resulting supramolecular assemblies.6
8 On a microscopic scale, the presence of hydrogen bond recognition elements in certain molecular building blocks generates specific secondary structures (e.g., molecular filaments, polymers with noncovalent main chains and side chains, and self-assembled dendritic macromolecules9). The resulting structures exhibit new material properties such as higher strength and/or improved elasticity.6
8 In comparison to covalent bonds, which are stable under normal conditions and can be broken only by providing sufficient energy, hydrogen bond formation is reversible, and the strength of hydrogen bonds depends on the chemical environment, such as the solvent or the temperature.5 Stadler et al.10,11 modified the properties of polybutadienes with a narrow molecular weight distribution by hydrogen bond interaction between urazole groups attached to the polymer backbone. The hydrogen bonds between urazole units form a thermoreversible network. Because of the labile hydrogen bond linkages, no rubber elastic equilibrium network modulus is observed in these systems. Thus, stabilization due to hydrogen bonding can provide an interesting alternative contribution to the design of new polymer architectures and polymer properties. In particular, triple hydrogen bond motifs inserted in linear r 2011 American Chemical Society

polymers have been studied extensively in the past by Zimmermann et al.,12 Whitesides et al.,13,14 and Meijer et al.15 Multiple hydrogen bond motifs on the surfaces of particles and complementary multiple hydrogen bond motifs on linear polymer chains are applicable to the reversible cross-linking of polymer
colloid blends. This would establish a way to fabricate new blends with thermoreversible elastomeric properties.16 Another interesting feature of hydrogen bond interactions is that they offer new strategies for the crystallization of binary colloidal systems toward a superlattice. In contrast to the crystallization of simple homogeneous colloids, the crystallization of binary systems requires a more complex pattern of interaction among the components, which in its simplest mode is also provided by the entropically driven formation of a binary superlattice of hard spheres occurring within a very narrow concentration and composition regime.17 However, the formation of such binary crystals may be considerably assisted by specific interactions among the colloids from the two components. In this sense, H bonds may be a good choice because they are capable of providing complementary motifs. This concept has been successfully applied in assembling nanoparticles and colloids into largescale structures such as colloidal crystals18 by covering the particle surfaces with DNA strands. The concept was introduced by Mirkin et al.19 and by Alivisatos et al.20 in 1996. Both groups covered gold nanoparticles with DNA strands and succeeded in Received: July 14, 2011 Revised: September 16, 2011 Published: September 20, 2011 12851

dx.doi.org/10.1021/la202685e | Langmuir 2011, 27, 12851–12858

Langmuir

ARTICLE

Figure 1. Synthesis route to DAD-labeled colloids.

reversibly self-assembling these particles into aggregates. Only recently, an excellent introduction to the field was provided by the review from Geerts and Eiser.21 However, the resulting selfassembly process, which is usually performed in aqueous systems, is controlled by many factors such as the temperature, pH, ionic strength, and the selection of a particular DNA bonding pattern. DNA bonding patterns apply various strategies such as those supplying two colloidal fractions with two cDNA strands or adding a duplex DNA with two sticky ends to a binary colloidal mixture where each colloid is equipped with a DNA sequence complementary to one of the duplex ends. In such DNA bonding patterns, the length of the DNA strands attached to the particles or used as connectors is a tunable factor. DNA functionalization in colloids is achieved via three alternative routes: the attachment of DNA to neutravidin-coated colloids via avidin
biotin complexation with the DNA strands covalently bound to biotin,22 the binding of amino-labeled DNA via carboxydiimide chemistry to polystyrene colloids,23,24 and the swelling/deswelling-based trapping of poly(ethylene glycol) chains (PEG) to colloids with the DNA bound to PEG.25 Aside from the complexity of tuning the mutual interactions, furnishing the particles with an appropriate amount of DNA strands is a formidable task, notwithstanding the fact that DNA strands on their own are complex systems compared to simple, small molecular functionalities. Therefore, we introduced an alternative concept that is based on cross-linked polystyrene colloids with a low (but tunable) percentage of comonomers with either an H-bond donor (4-vinyphenol) or an H-bond acceptor (4-vinylpyridine).26 With this system, corresponding to the most simple H-bonding motif available, we succeeded in self-assembling two binary colloidal fractions, differing in particle size, into binary aggregates and in controlling this aggregation by means of the addition of phenol as a competing H-bond donor.26 Direct crystallization might be supported in an even more efficient way by complex multiple

H-bond patterns such as ADA and DAD. The feature of strong, highly directional bonds becomes particularly interesting if bonding is reversible. Hence, the aim of this work is to extend our alternative concept26 and design colloids furnished with a DAD triple hydrogen bond pattern. The pattern is provided by 2,4-diaminotriazine functionalization. Such functionalization provides a route to two attractive properties: (i) it may enable heteroaggregation among complementarily functionalized components and (ii) it may cause the colloids to act as carriers of active ingredients such as drugs. In this context, it is of major relevance first to analyze whether colloidal homoaggregation by the DAD moieties takes place because this would interfere with the anticipated heteroaggregation. If homoaggregation among the colloids occurs, then molecular recognition of a low-molecular-weight antagonist with an ADA pattern (e.g., uracil) may help to control this homoaggregation. Moreover, the successful binding of uracil is considered to be a proof of principle for the loading of colloids with specific molecules via molecular recognition. Our preparation of the DAD-functionalized polystyrene colloids requires a procedure with five synthesis steps. Despite this effort, we consider the concept of colloids with an H-bond motif bearing comonomers as an interesting alternative to DNAfunctionalized particles. The reasoning is as follows: (i) Undesired homoaggregation between the colloids with equal types of DNA strands may be more pronounced than homoaggregation among colloids equipped with a fraction of small H-bondexerting comonomers. (ii) DNA is an elongated strand and thus causes undesired steric interactions in addition to H bonding. (iii) DAD colloids as well as the simple donor
acceptor system presented in a preceding paper26 are less expensive and less laborious to prepare than the DNA-functionalized colloids, exert H-bond interactions that can be tuned with simple lowmolecular-weight antagonists, are based on covalently bound 12852

dx.doi.org/10.1021/la202685e |Langmuir 2011, 27, 12851–12858

Langmuir

Figure 2. 13C CP/MAS NMR spectra of monomer 1-(methoxymethyl)4-vinylbenzene M-2 and the resulting polymer colloids. The spectra were obtained at a spinning rate of 5 kHz. Peak labels refer to Figure 1. The stars indicate spinning side bands.

functionalities and are equally applicable in nonaqueous solvent systems and in aqueous systems.

’ EXPERIMENTAL SECTION AND DATA EVALUATION Materials. 4-Vinylbenzyl chloride (90%) and boron trichloride (1.0 M in methylene chloride) were purchased from Aldrich, U.K. 1,4-Divinylbenzene (DVB), sodium hydroxide, potassium hydroxide, potassium peroxodisulfate (KPS), sodium cyanide (NaCN), 1-cyanoguanidin, methanol, carbon tetrachloride, N,N-dimethylformamide (DMF), 2-propanol, tetrahydrofuran (THF), and uracil were purchased from Merck KGaA, Germany. All monomers were destabilized prior to use via a flush column charged with basic aluminum oxide (Woelm B - Super I from Woelm Pharma, Germany). Doubly distilled water was used for the synthesis. Synthesis of 2,4-Diaminotriazine-Functionalized Poly-4methoxymethyl Styrene Colloids. Deans et al.27 published the synthesis of linear 2,4-diaminotriazine-functionalized random polystyrene copolymers in solution in order to analyze the chain folding due to intramolecular hydrogen bonding between the triazine moieties, followed by an unfolding of those polymeric globules enforced by specific docking of an added antagonist or by changing the solvent polarity. We transformed the polymerization published by Deans et al. into the preparation of colloids via surfactant-free emulsion polymerization (SFEP) of 4-methoxymethyl styrene including divinylbenzene as a cross-linker. This method provided access to quite monodisperse particles and avoided surfactants, which may have interfered in further applications. Furthermore, this method enabled us to control the particle size, for example, by varying the monomer concentration. 4-Methoxymethyl styrene colloids (PMS-M-I) were successively transformed into 2,4-diaminotriazine-functionalized colloids. A complete scheme of the synthesis route is shown in Figure 1. Detailed information on all synthesis steps can be found in the Supporting Information. The complete synthesis was tracked via solid-state 13C CP/MAS NMR. The spectra are shown in Figure 2. Magic Angle Spinning (MAS) 13C NMR. The spectra were recorded at room temperature with a Tecmag Apollo NMR spectrometer at the 13C resonance frequency of 75.47 MHz. The samples were placed in 4 mm zirconia rotors with Kel F caps. A cross-polarization (CP) sequence with a 90
pulse of 3.5 ms, a cross-polarization contact time of 2.0 ms, and a relaxation delay of 5 s was used. The spectral width was set to 20 000 Hz, and 8192 scans were accumulated. Spectra at spinning rates of 5 and 8 kHz were recorded in order to identify spinning side bands (SSB). No overlap between center peaks and spinning side

ARTICLE

bands was observed when the samples were spun at 5 kHz. Therefore, the spectra obtained at 5 kHz are reported and analyzed in this article. The resonance of the secondary carbon atom in adamantane at 38 ppm was used as a reference for chemical shifts. IR Measurements. Proof of the successful functionalization of PMS-Triazin-IV with 2,4-diaminotriazine moieties and proof of the binding of molecular uracil on the colloids were achieved by FTIR spectroscopy. We used a Hyperion 2000 IR/NIR microscope coupled with an Equinox 55 IR/NIR device (Bruker, USA). For the analysis of PMS-Triazin-IV colloids, a few droplets of a 1 g/L colloid solution in pure THF were placed on a blank metal substrate. After the gentle evaporation of THF and 1 h of storage under high vacuum at room temperature, the reflectance of the dry colloids was analyzed. The occurrence of hydrogen bonds between the 2,4-diaminotriazine-functionalized colloids and the uracil antagonist was proven with samples of PMS-Triazin-IV in THF saturated with uracil (1.33 g/L, 12 mM). To this end, a solution of 1 g/L 2,4-diaminotriazine-functionalized colloids in uracil-saturated THF was sonicated for 10 min before storing it for 1 h. After two centrifugations in pure THF (20 min at 4000 rpm), the colloids were dispersed in pure THF via ultrasonic treatment and stored for 1 h. A few drops of the colloid suspension were successively placed on a neat metal substrate. After the gentle evaporation of THF and 1 h of storage under high vacuum at room temperature, the dry mixture was analyzed in reflectance again. The wavenumber is denoted as λ
1 (cm
1). Scanning Electron Microscopy (SEM). Samples have been prepared by placing a few drops of colloidal dispersions of PMS-M-I and PMS-Triazin-IV in THF on a clean aluminum stub and gently evaporating the solvent. The SEM pictures were taken with a NEON 40 focused ion beam scanning electron microscope equipped with an EDX (Zeiss, Germany). The sample was coated with a 3-nm-thick gold film (Bal-Tec SCD 500, U.K.) to avoid charging effects. The accelerating voltage for the SEM imaging was 2 kV. Scattering Experiments and Data Evaluation. The PMS-M-I and PMS-Triazin-IV colloids were characterized in water via combined static (SLS) and dynamic light scattering (DLS) in very dilute aqueous solutions. Characterization was performed with a model 5000e compact goniometer system (ALV-Laser Vertriebsgesellschaft, Germany), which allows the simultaneous recording of SLS and DLS. A 100 mW Nd:YAG laser (Soliton, Germany) operating at a wavelength of 532 nm was used as the light source. Cylindrical quartz glass cuvettes with an outer diameter of 20 mm (Hellma, Germany) served as scattering cells. A C25 Haake thermostat was used to set the temperature to 25
C with a precision of 0.01
C. SLS and DLS were recorded in an angular range of 15 e θ e 150
with angular increments of 5
. This corresponds to a q regime of 0.0045 e q e 0.03294 nm
1. The parameter q is the momentum transfer as defined in eq 1. All samples were filtered with a 1.2 μm PET syringe filter (Macherey-Nagel, Germany) prior to the experiment. An investigation of the aggregation kinetics of PMS-Triazin-IV colloids was performed with time-resolved static light scattering (TR-SLS). The scattering instrument was a home-built goniometer28 equipped with a 41 mW He
Ne laser (Coherent, U.S.) operating at a wavelength of 633 nm. It allowed the simultaneous recording of static scattering twice at 19 fixed angles symmetrically arranged along both sides of the primary beam, covering an angular range of 25.84 e θ e 143.13
(0.0064 e q e 0.02720 nm
1). A C25 Haake thermostat was used to set the temperature to 25
C. Cylindrical quartz glass cuvettes with an outer diameter of 25 mm (Hellma, Germany) were used as scattering cells. The aggregation kinetics of PMS-Triazin-IV were investigated in THF and uracil-saturated THF. To this end, 8 mL of THF in the first case and 8 mL of THF saturated with uracil (1.33 g/L, 12 mM) in the second case were injected into the scattering cell by filtering through a 0.2 μm PET syringe filter (Macherey-Nagel, Germany) prior to the 12853

dx.doi.org/10.1021/la202685e |Langmuir 2011, 27, 12851–12858

Langmuir

ARTICLE

experiment. In both cases, 50 μL of an ultrasonically treated stock solution with 0.5 g/L colloid content was added to THF or THF/uracil with Eppendorf micropipets under ultrasonic treatment, which resulted in a final colloid concentration of 3.2 mg/L. The scattering cell was immediately placed into the goniometer, and the TR-SLS experiment was started. The moment when the cell was removed from the ultrasonic bath was considered to be time t = 0. One scattering curve was recorded every 30 s. Processing of Scattering Data. SLS experiments yield the excess Rayleigh ratio ΔR(q) of the solute as a function of the scattering vector q¼

4πn θ sin λ0 2

ð1Þ

where λ0 is the laser wavelength in vacuum, n is the refractive index of the solvent, and θ is the scattering angle. According to the Guinier approximation,29 a plot of ln(ΔR(q)) versus q2 becomes a function of the form lnðΔRðqÞÞ ¼ lnðΔRðq ¼ 0ÞÞ


Rg 2 2 q 3

ð2Þ

at qRg < 2. A linear fit over the corresponding q regime thus gives access to the excess Rayleigh ratio of the solute at zero scattering angle ΔR(q = 0) as well as to the z-averaged squared radius of gyration of the colloids, Rg2. Generally, the z average for any quantity X uses the following weighting procedure, with Ni and Mi being the number and mass of particles from fraction i. X̅ z ¼

∑Ni Mi 2 Xi ∑i Ni Mi 2

ð3Þ

ΔR(q = 0) is proportional to the weight-averaged molar mass of the solute. For the characterization of PMS-Triazin-IV in H2O, a Guinier fit29 has been applied to the scattering data within an angular regime of 30 e θ e 70
(0.0081 e q e 0.0181 nm
1). An analysis of the aggregation experiments of PMS-Triazin-IV, including reference experiments, was performed by Guinier fits in the range of 25.84 e θ e 72.5
(0.0064 e q e 0.0165 nm
1) for each scattering curve. In addition, a model fit for a polydisperse sphere with the distribution of the sphere volume V has been applied to the whole scattering curve in the case of the characterization of PMS-M-I and PMS-Triazin-IV in H2O. The polydispersity has been taken into account by means of the Schulz
Zimm distribution30
 kV
k
1 kk exp
V Vw Wsz ðV , Vw , kÞ ¼ ð4Þ Vw Vw ΓðkÞ In eq 4, Vw denotes the weight-averaged volume of the dispersed particles and σ2 = 1/k is the normalized variance of wSZ(V, Vw, k) with the gamma function Γ(k). Equation 4 represents a volume-fractionweighted distribution of particle volume V. The form factor was modeled with the Rayleigh equation31 !2 3ðsinðqRÞ
qR cosðqRÞÞ Psphere ðq, RÞ ¼ ð5Þ ðqRÞ3 which corresponds to compact spheres with an outer radius R. If the particle volume is assumed to be proportional to the particle mass, the z-averaged scattering curve can thus be calculated as ΔRðq, Vw , kÞ ¼ K

Z ∞ 0

Psphere ðq, RÞ wSZ ðV , Vw , kÞ V dV

ð6Þ

where the constant K considers the scattering contrast of the particles. The sphere radius is related to the sphere volume according to R = (3V/(4π))1/3. The modeling of PMS-Triazin-IV scattering data by means of the sphere form factor was not successful. Therefore, the form

factor of a dumbbell Pdumbbell ðq, RÞ ¼

3ðsinðqRÞ
qR cosðqRÞÞ ðqRÞ3

!2

! 1 sinð2qRÞ þ 2 4qR ð7Þ

consisting of two attached spheres with radii R was used instead of Psphere in eq 6 to model the PMS-Triazin-IV scattering data. DLS provides the field intensity correlation g1(τ) as a function of the relaxation time τ. According to the cumulant analysis,32 a plot of ln(g1(τ)) versus τ can be described by a function of the form μ lnðg1 ðτÞÞ ¼ C þ Γτ þ 22 τ2 ð8Þ 2Γ where C is a constant and Γ is the mean inverse relaxation time of diffusive modes with ÆDæz = Γ/q2. DLS measurements were evaluated by a linear fit of the apparent diffusion coefficients in the regime of 30 e θ e 70
(corresponding to 0.0081 e q e 0.0181 nm
1) to give the z-averaged diffusion coefficient ÆDæz at q = 0. The z-averaged translational diffusion coefficient ÆDæz can be transformed into a mean effective hydrodynamic radius Rh of the particles via the Stokes
Einstein equation: Rh ¼

kB T 6πηÆDæz

ð9Þ

In eq 9, kB is the Boltzmann constant, T is the temperature, and η is the dynamic viscosity of the solvent. The ratio of the two size parameters Rg and Rh obtained from SLS and DLS F¼

Rg Rh

ð10Þ

is a shape-sensitive quantity. For homogeneous spheres, F = 0.775.33 Apart from the z-averaged diffusion coefficient, the fit of a polynomial of second order in τ based on eq 8 gives access to the quantity μ2 ÆD2 æz
ÆDæz2 2 ¼ Γ ÆDæz2

ð11Þ

which corresponds to a normalized variance of the diffusion coefficient based on an intensity-weighted distribution. A quantity varSZ that is comparable to μ2/Γ2 can be calculated with the Schulz
Zimm distribution obtained from eq 5 from SLS as    2 1 1
R2 z R z ð12Þ varSZ ¼  2 1 R z whereupon z averages Æ1/Ræz and Æ1/R2æz are calculated analogously to ΔR(q, Vw, k) in eq 6 by replacing KPsphere(q, R) with 1/R and 1/R2, respectively. Aside from the cumulant analysis, the field intensity correlation function can be modeled by a series of exponentials. This is achieved with the so-called CONTIN analysis,34 which yields an intensity-weighted distribution of effective hydrodynamic radii wCON (Rh).

’ RESULTS AND DISCUSSION 2,4-Diaminotriazine-functionalized colloids (PMS-TriazinIV) provide a system capable of intra- and intermolecular selfassembly through a triple-hydrogen-bonding motif. The preparation of this system was first obtained by carrying out a SFEP with 90 mol % 4-methoxymethylstyrene with 10 mol % 1,4-divinylstyrene as a cross-linker. The methoxy groups on the PMS-M-I 12854

dx.doi.org/10.1021/la202685e |Langmuir 2011, 27, 12851–12858

Langmuir colloid were successively substituted for chlorine, followed by reaction of the chloromethyl groups of PMS-Cl-II colloids with excess sodium cyanide. The reaction of cyano-functionalized colloids PMS-CN-III with dicyandiamide finally provided 2,4diaminotriazine-bearing product PMS-Triazin-IV. For steric reasons, it can be assumed that the final 2,4-diaminotriazine functionalizations are primarily on the surfaces of the PMSTriazin-IV colloids. The whole reaction was successfully tracked by solid-state 13C CP/MAS NMR. The 13C CP/MAS NMR spectra of all synthesized compounds are shown in Figure 2. Notice that the peak intensities are not proportional to the respective number of carbon atoms because different carbon atoms have different cross-polarization time constants. The M-2 spectrum of 1-(methoxymethyl)-4vinylbenzene), depicted at the top of Figure 2, shows two peaks at 113 and 116 ppm (labeled a and b), which are assigned to the CdC double bond of this monomer. In the spectrum of the 4-methoxymethyl-functionalized colloid (PMS-M-I) (Figure 2, row 2), two new peaks, a0 and b0 , appear at 39 and 46 ppm, respectively. They can be assigned to the polymer backbone. The peaks at 113 and 116 ppm are absent in the spectrum of PMS-MI, indicating that the polymerization was complete. The peaks of most carbon atoms not involved in the reaction stay at the same chemical shift except for the peak of the aromatic carbon atom connected to the main chain. This peak is shifted from 138 to 144 ppm. The polymer peaks are much broader than those of the monomer because different constitutions and conformations of the amorphous atactic polymer lead to a spread in chemical shifts. The third spectrum in Figure 2 was obtained from the 4-chloromethyl-functionalized colloid (PMS-Cl-II). The peak from the methoxy group of PMS-M-I (carbon atom e) at 58 ppm has vanished in PMS-Cl-II. Furthermore, because of the stronger negative inductive effect of chlorine compared to oxygen, the methylene peak at 74 ppm (d) is shifted to 45 ppm (d0 ). This shows that colloid PMS-M-I was successfully chlorinated to form colloid PMS-Cl-II. The spectrum of PMS-Cl-II is the first one showing spinning sidebands, indicating a lowering of the glasstransition temperature due to the change in functionalities. Row 4 in Figure 2 shows the spectrum of the 4-cyanomethyl-functionalized colloid (PMS-CN-III). When chlorine is replaced by the nitrile group, the methylene peak formerly at 45 ppm (d0 ) shifts further to 22 ppm (d00 ). An additional small peak appears at 118 ppm, which belongs to the carbon atom in the nitrile group (f). Peaks c and c0 at 136 ppm in polymers PMS-M-I and PMSCl-II have disappeared in the spectrum of PMS-CN-III. The peak of the corresponding carbon atom is shifted upfield and thereby hidden under the large peak of the aromatic carbon atoms at 127 ppm. The spectrum of the final (2,4-diaminotriazine)methylfunctionalized colloids PMS-Triazin-IV is shown at the bottom of Figure 2. The two distinguished peaks at 166 and 177 ppm indicate amino-substituted carbon atoms h and i and carbon atom g of the triazine group. Two additional peaks are visible at 45 and 135 ppm, corresponding to carbon atoms d000 and c000 , respectively. Their larger chemical shift values compared to the precursor can be attributed to the more electron-rich heterocycle compared to the nitrile group. However, the remaining peak at 22 ppm shows that the last reaction was not complete. This is not surprising because an analogous polymer reaction involving a sequence of substitution reactions cannot be expected to show complete conversion because steric hindrance might interfere. To verify the last synthesis step, two IR spectra of PMS-CN-III

ARTICLE

Table 1. Summary of Particle Properties Including the Hydrodynamic Radius (Rh) from the Experimental Data, the Radius of Gyration (Rg) from the Experimental Data, the Shape-Sensitive Parameter F = Rg/Rh, the Extent of Polydispersity μ2/Γ2 Calculated from the Cumulant Analysis of DLS Data, and the Extent of Polydispersity varSZ Calculated from Form Factor Fits Based on Equation 6 by Applying the Polydisperse Sphere Model with a Schulz
Zimm Distribution to the Experimental SLS Data Rg (nm)

Rh (nm)

F

μ2/Γ2

varSZ

PMS-M-I

128

178

0.72

0.08

0.06

PMS-Triazin-IV

165

200

0.83

0.32

0.27

colloid sample

Figure 3. Excess Rayleigh ratios ΔR(q) of (top) precursor colloid PMSM-I and (bottom) 2,4-diaminotriazine-functionalized PMS-Triazin-IV colloids measured in aqueous solution. Solid lines give the best fits with a polydisperse sphere model based on a Schulz
Zimm distribution of the volume V according to eq 6. The dashed line in the lower plot shows the best fit with a polydisperse dumbbell according to eq 7.

and the final PMS-Triazin-IV have been included in the Supporting Information. After having established the successful preparation of PMSTriazin-IV along the path outlined in Figure 1, we now discuss the colloidal properties of PMS-Triazin-IV in comparison to its PMS-M-I precursor. The z-averaged diffusion coefficient from dynamic light scattering yields a hydrodynamic radius of 178 nm for PMS-MI. The evaluation of the DLS data of PMS-Triazin-IV yields an averaged hydrodynamic radius of 200 nm. In agreement with this increase in the hydrodynamically effective size, the radius of gyration Rg also adopts a larger value of 165 nm for PMS-TriazinIV compared to Rg = 128 measured for PMS-M-I. A larger size of the PMS-Triazin-IV colloids can be caused by partial swelling due to the better solubility of 2,4-diaminotriazine moieties in water compared to the methoxy functionalization of the PMS-MI colloids or alternatively due to the tendency to form oligomeric aggregates. Nevertheless, the structure-sensitive factor F for PMS-M-I colloids as well as for PMS-Triazin-IV colloids indicates homogeneous and compact spheres in both cases. Table 1 12855

dx.doi.org/10.1021/la202685e |Langmuir 2011, 27, 12851–12858

Langmuir

ARTICLE

summarizes the characteristic parameters of colloids PMS-M-I and PMS-Triazin-IV. The excess Rayleigh ratios ΔR(q) of the precursor colloid PMS-M-I and of the final colloid PMS-Triazin-IV, obtained from SLS, are represented in Figure 3. As can be seen from the fits, the scattering experiment for PMS-M-I can be perfectly described with the model calculations for a polydisperse sphere with a Schulz
Zimm distribution of the colloid volume (top Figure 3, solid line). However, 2,4-diaminotriazine-functionalized colloids PMS-Triazin-IV show a variation in the scattering form factor, and a fit to the same model used for the PMS-M-I colloids is not as good in the case of PMS-Triazin-IV colloids (bottom Figure 3, solid line). This may serve as a hint of the aggregation of functionalized PMS-Triazin-IV colloids. If dimerization or trimerization takes place, then the model for a hard sphere no longer reveals reliable results. Further support for the oligomerization of PMS-Triazin-IV colloids is provided by an alternative fit based on polydisperse dumbbells (bottom of Figure 3, dashed line), which is able to reproduce the experimental data perfectly. A comparison of the distribution functions revealed from the CONTIN analysis (Figure 4) also indicates a larger polydispersity of the PMS-Triazin-IV sample than of the PMS-M-I sample. The polydispersities varSZ (Table 1) established from form factor fits with eq 6 and from the second cumulant analysis μ2/Γ 2 (Table 1) of the correlation functions also result in a larger polydispersity of the PMS-Triazin-IV sample, hence supporting the hypothesis of the oligomerization of PMS-Triazin-IV.

However, we can unambiguously demonstrate that the colloids survive the three-step chemical modification as colloidal entities. This confirmation is achieved by a comparison of SEM images from PMS-M-I precursors and from PMS-Triazin-IV colloids. A comparison of both SEM pictures, which do not show any significant variation, is presented in Figure 5. In addition, the obtained size values of the PMS-M-I colloids from light scattering are in good agreement with the size values indicated by the SEM image. The tendency to form aggregates is expected to be tunable by decreasing the colloid concentration. Thus, highly dilute solutions may offer the opportunity to analyze the aggregation kinetics by means of time-resolved static light scattering. We first had to select a suitable solvent for such an aggregation experiment. THF turned out to be suitable. An analysis of the aggregation of the pure PMS-Triazin-IV colloids shall be compared with the interplay of the same colloids with uracil. Uracil is an antagonist to 2,4-diaminotriazine. The graph of Figure 6 shows the characteristic parameters resulting from the time-resolved static light scattering of an aggregation experiment of 2,4-diaminotriazine-functionalized colloids in pure THF at a colloid concentration of 3.2 mg/L. The radius of gyration increases gradually from 136 to 156 nm within 3000 s. This is an increase of 20 nm. In line with the increase in Rg, a growth of ΔR(q = 0) has been recorded. The

Figure 4. Colloid size distributions calculated from the CONTIN analysis at an angle of 30
. The lines indicate (—) functionalized PMS-Triazin-IV colloids and (- - -) precursor PMS-M-I colloids.

Figure 6. TR-SLS experiment of a solution of triazine-functionalized PMS-Triazin-IV colloids in pure THF. The colloid concentration is 3.2 mg/L.

Figure 5. (Left) SEM image of the PMS-M-I colloids directly after surfactant-free emulsion polymerization. (Right) SEM image of the PMS-Triazin-IV colloids. 12856

dx.doi.org/10.1021/la202685e |Langmuir 2011, 27, 12851–12858

Langmuir

ARTICLE

Figure 7. TR-SLS experiment of a solution of 2,4-diaminotriazinefunctionalized PMS-Triazin-IV colloids in pure THF. The colloid concentration is 1.6 mg/L. Figure 9. FTIR spectra of (a) pure PMS-Triazin-IV colloids, (b) pure uracil, and (c) a dried mixture of both components after three cycles of centrifugation and the addition of pure THF. All samples were analyzed in the dried state in reflection mode.

Figure 8. TR-SLS experiment of a solution of triazine-functionalized PMS-Triazin-IV colloids in THF saturated with uracil (12 mM). The colloid concentration is 3.2 mg/L.

latter parameter is a measure of the weight-average particle mass multiplied by the particle concentration. The decrease in ΔR(q = 0) beyond 1250 s is due to the sedimentation of the colloidal aggregates, which is further evidenced by a slight precipitation appearing after a few days on the bottom of the cuvette. This sedimentation is caused by a density difference with THF, which is 20% smaller than the density of the polystyrene-based PMSTriazin-IV colloids. In an experiment with a concentration of only 1.6 mg/L shown in Figure 7, we did not observe any more sedimentation and detected an increase in size of only 10 nm within the same time regime as the one shown in Figure 6. The aggregation indicated by TR-SLS for the present system is considered to be caused by hydrogen bond interactions. In a combined SLS and DLS experiment, Borkovec et al.35 analyzed the onset of heteroaggregation of spherical colloids where the aggregation was induced by opposite electrical charging. They succeeded in catching dimers of complementary charged colloids as mutual intermediates in similarly low-concentration solutions such as those applied in the present work. Uracil with its ADA hydrogen bond motif is a direct antagonist of 2,4-diaminotriazine and should be able to bind by hydrogen bond interaction preferentially on the triazine functionalities. Thus, if the 2,4-diaminotriazine-functionalized colloids are dissolved in THF saturated with uracil (12 mM), we expect a strong modulation of the aggregation if it is driven by H-bond formation. To verify this expectation, a time-resolved experiment

has been carried out in a THF solution saturated with uracil (12 mM) but otherwise under the same conditions as the aggregation experiment shown in Figures 6 and 7. The results shown in Figure 8 clearly indicate that aggregation is indeed fully inhibited now. The values for Rg and ΔR(q = 0) stay constant over the whole time range. As a consequence, this reference experiment is considered to be the final proof that hydrogen bonds are the origin of the aggregation of 2,4-diaminotriazine colloids in THF. It is possible to prevent this aggregation by completely capping the triazine moieties with the antagonist uracil. The strong tendency of the PMS-Triazin-IV colloids to form hydrogen bonds is further indicated by FTIR analysis. To this end, we dissolved the 2,4-diaminotriazine-functionalized colloids in THF saturated with uracil (12 mM). On the basis of the dramatic impact exerted by uracil on the aggregation of PMSTriazin-IV, it is expected that uracil binds considerably to the 2,4diaminotriazine-functionalized colloid surface. Figure 9a shows the spectrum of the pure 2,4-diaminotriazine-functionalized PMS-Triazin-IV colloids with the characteristic N
H bands at 3300
3500 and 1600 cm
1. Additionally, a broad absorption regime is observable between 2500 and 3500 cm
1. This feature is considered to be characteristic of the self-associated 2,4diaminotriazine residues. Rossotti36 and Elmsley37 found a similar broadening in the IR absorbance in systems of selfassociated pyrazoles and imidazoles. As expected, comparable behavior was also revealed for pure uracil in the spectrum of Figure 9b. Uracil is also able to form intermolecular hydrogen bonds. In a mixture of uracil and PMS-Triazin-IV colloids recovered from a solution in THF saturated with uracil, we found a combined IR pattern of both components even after two centrifugation cycles and the successive addition of pure THF (spectrum in Figure 9c). Hence, the uracil could not be removed by even two washing cycles, which unambiguously indicates the successful surface functionalization of the colloids and the recognition and strong binding of uracil upon those functionalizations.

’ SUMMARY We accomplished the synthesis of new colloids functionalized with a triple hydrogen bond motif by means of surfactant-free 12857

dx.doi.org/10.1021/la202685e |Langmuir 2011, 27, 12851–12858

Langmuir emulsion polymerization followed by three further synthesis steps. The functional group is a 2,4-diaminotriazine residue, which bears a donor
acceptor
donor (DAD) pattern. The successful reaction was verified by 13C CP/MAS solid-state NMR. A detailed characterization of the PMS-M-I precursor colloid via combined static and dynamic light scattering and SEM suggests a compact spherical shape of the colloids. SEM imaging confirmed the conservation of the colloidal entities throughout chemical modification. Because of their functionalization, the colloids were able to aggregate in suitable solvents. THF, for instance, turned out to be a convenient medium in which to dissolve 2,4-diaminotriazinefunctionalized colloids while exhibiting controlled and gradual aggregation. This was analyzed in very dilute solution by timeresolved static light scattering in a concentration regime as low as a few milligrams per liter. If THF was saturated with uracil (12 mM), colloid aggregation was completely inhibited. This result unambiguously proves that colloidal aggregation is caused by H bonds formed among the 2,4-diaminotriazine groups. Uracil with its ADA hydrogen bond pattern is the direct antagonist of 2,4diaminotriazine and could thus successfully be used to block those functionalities on the colloids completely. Uracil could not be removed from the PMS-Triazin-IV colloids even with two washing cycles. Using IR spectroscopy, it could be demonstrated that the uracil residues stick to the PMS-Triazin-IV colloids. These promising results give rise to the hope that controlled heteroaggregation among complementary functionalized components, such as the present colloids on one side and any tailormade second colloid component or linear polymer with a complementary functionalization on the other side, becomes possible, thereby offering new ways to form hierarchical structures such as binary superlattices. At the same time, PMS-TriazinIV turned out to act as a carrier for specific molecules as soon as they were tagged with an antagonistic H-bond motif.

’ ASSOCIATED CONTENT

bS

Supporting Information. Detailed information on all synthesis steps and IR spectra of the last synthesis step (colloids PMS-CN-III and PMS-Triazine-IV). This material is available free of charge via the Internet at http://pubs.acs.org.

’ AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected]. Phone: (+49) 5251 602125. Fax: (+49) 5251 604208.

’ ACKNOWLEDGMENT This work was supported by a grant from the Deutsche Forschungsgemeinschaft (DFG) within the Research Training Group “Micro- and Nanostructures in Optoelectronics and Photonics” (GRK 1464). Support of SEM by Dr. Karl Hiltrop is gratefully acknowledged.

ARTICLE

(5) Lawrence, D. S.; Jiang, T.; Levitt, M. Chem. Rev. 1995, 95, 2229. (6) Whitesides, G. M.; Mathias, J. P.; Seto, C. T. Science 1991, 254, 1312. (7) Bowden, N. B.; Weck, M.; Choi, I. S.; Whitesides, G. M. Acc. Chem. Res. 2001, 34, 231. (8) Sharma, C. V. K. J. Chem. Educ. 2001, 78, 617. (9) Hill, D. J.; Mio, M. J.; Prince, R. B.; Hughes, T. S.; Moore, J. S. Chem. Rev. 2001, 101, 3893. (10) Stadler, R.; Burgert, J. Makromol. Chem. 1986, 187, 1681. (11) Stadler, R.; Freitas, L. L. Macromolecules 1987, 20, 2478. (12) Zimmermann, S. C.; Corbin, P. S. Struct. Bonding (Berlin) 2000, 96, 63. (13) Whitesides, G. M.; Simanek, E. E.; Mathias, J. P.; Seto, C. T.; Chin, D. N.; Mammen, M.; Gordon, D. M. Acc. Chem. Res. 1995, 28, 37. (14) MacDonald, C.; Whitesides, G. M. Chem. Rev. 1994, 94, 2383. (15) Beijer, F. H.; Sijbesma, R. P.; Vekemans, J. A. J. M.; Meijer, E. W.; Kooijman, H.; Spek, A. L. J. Org. Chem. 1996, 61, 6371. (16) S€ontjens, S. H. M.; Renken, R. A. E.; van Gemert, G. M. L.; Engels, T. A. P.; Bosman, A. W.; Janssen, H. M.; Govaert, L. E.; Baaijens, F. P. T. Macromolecules 2008, 41, 5703. (17) Bartlett, P.; Pusey, P. N. Phys. Rev. Lett. 1992, 68, 3801. (18) Biancaniello, P. L.; Kim, A. J.; Crocker, J. C. Phys. Rev. Lett. 2005, 94, 058302. (19) Mirkin, C. A.; Letsinger, R. L.; Mucic, R. C.; Strorhoff, J. J. Nature 1996, 382, 607. (20) Alivisatos, A. P.; Johnsson, K. P.; Peng, X.; Wilson, T. E.; Loweth, C.; Bruchez, M.; Schultz, P. Nature 1996, 382, 609. (21) Geerts, N.; Eiser, E. Soft Matter 2010, 6, 4647. (22) Milam, V. T.; Hiddessen, A. L.; Crocker, J. C.; Graves, D. J.; Hammer, D. A. Langmuir 2003, 19, 10317. (23) Soto, C. M.; Srinivasan, A.; Ratna, B. R. J. Am. Chem. Soc. 2002, 124, 8508. (24) Rogers, P. H.; Michel, E.; Bauer, C. A.; Vanderet, S.; Hansen, D.; Roberts, B. K.; Calvez, A.; Crews, J. B.; Lau, K. O.; Wood, A.; Pine, D. J.; Schwartz, P. V. Langmuir 2005, 21, 5562. (25) Kim, A. J.; Manoharan, V. N.; Crocker, J. C. J. Am. Chem. Soc. 2005, 127, 1592. (26) Bayer, F. M.; Hiltrop, K.; Huber, K. Langmuir 2010, 26, 13815. (27) Deans, R.; Ilhan, F.; Rotello, V. M. Macromolecules 1999, 32, 4956. (28) Becker, A.; Schmidt, M. Macromol. Chem. Macromol. Symp. 1991, 15, 249. (29) Guinier, A.; Fournet, G. Small Angle Scattering of X-rays; WileyVCH: New York, 1955. (30) Schulz, G. V. Phys. Chem. 1939, 43, 25. (31) Rayleigh, J. W. Proc. R. Soc. A 1914, 90, 219. (32) Koppel, D. E. J. Chem. Phys. 1972, 57, 4818. (33) Schweins, R.; Huber, K. Eur. Phys. J. E 2001, 5, 117. (34) Provencher, S. W. Comput. Phys. 1982, 27, 213. (35) Galletto, P.; Lin, W.; Mishchenko, M. I.; Borkovec, M. J. Chem. Phys. 2005, 123, 064709. (36) Anderson, D. M. W.; Duncan, L.; Rossotti, F. J. C. J. Chem. Soc. 1961, 140, 2165. (37) Elmsley, J. Chem. Soc. Rev. 1980, 9, 91.

’ REFERENCES (1) Jeffrey, G. A. An Introduction to Hydrogen Bonding; Oxford University Press: New York, 1997. (2) Israelachvili, J. Intermolecular and Surface Forces; Academic Press: London, 1992. (3) Philip, D.; Stoddart, J. F. Angew. Chem. 1996, 108, 1242. (4) Philip, D.; Stoddart, J. F. Angew. Chem., Int. Ed. Engl. 1996, 35, 1154. 12858

dx.doi.org/10.1021/la202685e |Langmuir 2011, 27, 12851–12858