Preventing the coffee-ring effect and aggregate sedimentation by in situ gelation of monodisperse materials

Microscale uniformity and long-range cohesion in multi-functional films assembled through drop-casting is realized by in situ gelation of monodisperse building blocks.

Scheme S2: Synthesis of the third generation poly(amido amine) (PAMAM G3) dendrimer with thioacetate functionalized viologen terminal groups (2), based on the reaction of the amino head group of commercially available PAMAM G3 dendrimers with a viologen monomer terminated by both isothiocyanate and thioacetate groups (1-(3-(acetylthio)propyl)-1'-(3isothiocyanatopropyl)-[4,4'-bipyridine]-1,1'-diium iodide bromide, 1). The synthesis and full characterization of the viologen derivative 1 was reported previously. 1 Quantification of the film thickness distribution from the electroanalytical method. 2 The electroanalytical method developed and used for the characterization of redox film preparations is based on analysis of the peak currents obtained from the Linear Sweep Voltammograms (LSV). The model is based on a Langmuir assumption (no interaction of neighboring redox active mediator molecules). However, the LSVs obtained in our experimental system (i.e. Figure S4B) include contributions related to interactions between the redox mediator molecules, according to the Frumkin isotherm. In this case, the model is still applicable, but requires conversion of the peak currents to their Langmuir equivalents before further data processing.
Frumkin isotherm correction. In the case of the viologen-based dendrimer, the interaction between the redox centers requires the use of the Frumkin isotherm 3 to describe the current response in the surface confined case. Equation S1 includes a correction factor (ψ p ) based on the Frumkin isotherm: The value of ψ p can be easily derived from Equation S1 based on the slope of a plot of i p versus ν with known parameters of n, F, R, T, A and Γ T for CVs in the surface confined case (the surface coverage of Γ T is calculated by integration of the charge). For the viologen-based dendrimer on Au electrode, the value of the factor 1/ψ p is 2.06.
In the next step, with the use of the other relevant experimental parameters (such as concentration, electron diffusion coefficient, etc.), the peak currents (y-axis) and the scan rates (x-axis) are normalized by application of the expressions shown in the respective axes (i.e. Figure S4C,D). In the expression for peak current normalization, i p is the current density.
Next, the value of the dimensionless current at the reference line (where the value of the dimensionless current is equal to 2) is read from the dimensionless plot. Using this value, and a correlation plot, the information related to the film thickness distribution is directly obtained (Weibull Distribution shape factor). This value is then used to generate a plot of the probability distribution function (normalized histogram) for the underlying film thickness (i.e Figure   S4E,F).
In order to visualize what the resulting surface could possibly look like, for comparison with AFM or optical microscope images, the probability distribution function is transformed into a three-dimensional representation (i.e. Figure S4G,H). When comparing the 3D representation from the electrochemical method with images from AFM, care must be taken to regard the limiting and complementary nature of the information provided by these two methods. For microscopy type methods such as AFM, for example, the obtained images are highly dependent on the sampling area, which is very small when compared to the entire film. For the electrochemical method, although the entire film is represented, the number of points used to represent the film is arbitrarily chosen (in Figure S4G,H this number is 100), and the locations of the film sections are also arbitrarily chosen, since this information is not accessible using the electrochemical method. Therefore, the electrochemical and AFM methods provide complementary information. When the results of these methods are in good general agreement, a high degree of assurance regarding the underlying thickness distribution and roughness of the film preparation is obtained.

Figure S6: Determination of the film thickness distribution for a dendrimer film prepared on a gold substrate with a horizontal orientation (0°) using the electrochemical method A)
CVs of viologen-modified dendrimer films at different scan rates B) CVs at slow scan rate from A). C) Variation of dimensionless peak currents with dimensionless scan-rate [(nF/RT)vd 2 /D e )] 1/2 : Theoretical dimensionless current for a perfectly smooth film (red squares) and for the experimental peak currents extracted from (A) (blue squares). The parameters used for the non-dimensionalization of the experimental data are the same as those used in the caption of Figure S4. D) Probability distribution function. The diameter of the Au electrode was 2 mm, and the measurements were performed in phosphate buffer (100 mM, pH 7.2) under anaerobic conditions.

Figure S7: Determination of the film thickness distribution for a dendrimer film prepared on a gold substrate with a vertical orientation (90°) using the electrochemical method. A)
CVs of viologen-modified dendrimer films at different scan rates B) CVs at slow scan rate from A). C) Variation of dimensionless peak currents with dimensionless scan-rate [(nF/RT)vd 2 /D e )] 1/2 : Theoretical dimensionless current for a perfectly smooth film (red squares) and for the experimental peak currents extracted from (A) (blue squares). The parameters used for the non-dimensionalization of the experimental data are the same as those used in the caption of Figure S4. D) Probability distribution function. The diameter of the Au electrode was 2 mm, and the measurements were performed in phosphate buffer (100 mM, pH 7.2) under anaerobic conditions.

Figure S8: Determination of the film thickness distribution for a dendrimer film prepared on a gold substrates with an upside down orientation (180°) A)
CVs of viologen-modified dendrimer films at different scan rates B) CVs at slow scan rate from A). C) Variation of dimensionless peak currents with dimensionless scan-rate [(nF/RT)vd 2 /D e )] 1/2 : Theoretical dimensionless current for a perfectly smooth film (red squares) and for the experimental peak currents extracted from (A) (blue squares). The parameters used for the non-dimensionalization of the experimental data are the same as those used in the caption of Figure S4. D) Probability distribution function. The diameter of the Au electrode was 2 mm, and the measurements were performed in phosphate buffer (100 mM, pH 7.2) under anaerobic conditions.

Determination of the apparent electron diffusion coefficient (D e ) in redox active films.
Interdigitated array electrodes (IDA) were employed for the determination of D e in viologenmodified dendrimer films. IDA electrodes allow for the calculation of D e from the limiting current (i L ) obtained from generator-collector mode, and from the charge (Q) required for the reduction of all viologens obtained from generator-generator mode (and at slow scan rate).
Then, D e can be calculated from Q and from i L by: Where N is the number of IDA fingers, d is the gap between the fingers and p is the center-tocenter width of a two-finger segment.  Table S1). The calculated charge slightly increases with decreasing scan rate due to increased inaccuracy in the background current at the lower scan rates. Therefore, the average value of these scan rates was used in the determination of D e .
In generator-collector mode, the generator electrode was swept toward -0 The average value of D e for the dendrimer films is (1.15 ± 0.1) 10 −8 cm 2 s −1 . This value was independent of the dimensions and the geometry of the IDAs. Additionally, the D e value was independent of electrolyte concentration, demonstrating that the counter ion transfer associated with electron transfer is non-limiting. In comparison, the D e value obtained for viologenmodified polymer films, which consist of the same viologen moieties, was (4.7 ± 1.7) 10 −9 cm 2 s −1 , 5 which is about 2 times lower than for the viologen-modified dendrimer hydrogel. The pseudo-reference electrode was modified with a Ag/AgCl ink. Table S1: Charge obtained from integration of the area under the cathodic curve of the CVs given in Figure S9A at slow scan rate. Additional phosphate buffer (100 mM, pH 7.0) was added to each preparation so that total volume would be 4µl.
Figure S11: Catalytic current density for NADPH oxidation. Cyclic voltammograms for FNR immobilized in films of A) viologen-modified dendrimer and B) viologen-modified polymer in presence (red traces) and absence (black trace) of NADPH. C) Comparison of catalytic current densities of FNR immobilized in viologen-modified dendrimer and in viologen-modified polymer films respectively. The electrodes were prepared by drop-casting 2 µL of a viologen-modified dendrimer mixture (1µL, 7.14 mg/mL) or viologen-modified polymer mixture (1µL, 9.55 mg/mL) with FNR (1 µL, 0.1 mM). The standard deviations used in the error bars are obtained from two replicates from two individual preparations under the same conditions. Figure S12: Linear film thickness dependency on the amount of drop-cast dendrimer. The film thicknesses (l) were obtained from the viologen surface coverage (Γ) and from the viologen concentration (C) according to l = Γ / C. 5 The values of Γ and C were, respectively, extracted from the integration of the CV wave at slow scan rates and from the peak currents of the CVs at fast scan rates according to the Randles Sevcik equation and using the value of D e (1.15 · 10 −8 cm 2 s -1 ). The inset is the enlargement of dotted-line area.