Graphene transistors for real-time monitoring molecular self-assembly dynamics

Mastering the dynamics of molecular assembly on surfaces enables the engineering of predictable structural motifs to bestow programmable properties upon target substrates. Yet, monitoring self-assembly in real time on technologically relevant interfaces between a substrate and a solution is challenging, due to experimental complexity of disentangling interfacial from bulk phenomena. Here, we show that graphene devices can be used as highly sensitive detectors to read out the dynamics of molecular self-assembly at the solid/liquid interface in-situ. Irradiation of a photochromic molecule is used to trigger the formation of a metastable self-assembled adlayer on graphene and the dynamics of this process are monitored by tracking the current in the device over time. In perspective, the electrical readout in graphene devices is a diagnostic and highly sensitive means to resolve molecular ensemble dynamics occurring down to the nanosecond time scale, thereby providing a practical and powerful tool to investigate molecular self-organization in 2D.


S3
In all the curves, a stationary state is reached in which the photo-induced SP → MC isomerization is balanced by the thermal MC → SP isomerization. We note that at high concentration (C0 > 0. 75     We highlight that the minor ΔI recorded for the top left panel in (c) corresponds to a situation in which graphene is covered only by the solvent (1-phenyloctane), see also     Importantly, the initial increase in IDS recorded in the three cycles is extremely similar, implying that the dynamics of formation of the self-assembled adlayer are the same in the three cases.
Measurements performed at a SP initial concentration C0 = 4 mM.

Origin of the molecular induced gating effect
We previously reported that a MC adlayer at the solid-air interface introduced n-type doping on graphene, which was understood on the basis of the nanoscale molecular arrangement. 2 In particular, molecular dynamics simulations showed that MC isomers in the assembly are slightly tilted with respect to the basal plane of the substrate. Such tilt gives rise to a non-zero out-ofplane component of each molecular electrical dipoles, which sum up to generate an electric field effect capable of shifting the work function of graphene, as a gate terminal. Since the nanoscale assembly of MC is the same at the solid/liquid and at the solid/air interface, a similar effect is measured in this work. We also stress that the ordered structure framed by self-assembly ensures that each molecule contributes with the same nanoscale field to the overall molecular gating effect. Therefore, the intensity of the overall field effect will be proportional to the number of molecules on the graphene surface (see next section), that is, ΔI scales linearly with the covered area. The same field-effect on graphene is induced by different crystalline domains of the molecular adlayer with lamellae oriented along the equivalent crystallographic directions of graphene. Indeed, the gating effect originates from the out-of-plane component of the electrical dipoles, which does not depend on the in-plane orientation of the lamellar assembly. In this regard, we consider that the maximum shift in charge neutrality point is achieved when the graphene layer is fully coveredso that the field effect acts all over the graphene surface. If instead part of the graphene surface is left uncovered, it maintains its original charge neutrality point and it does not contribute to the ΔI increase.
We highlight that a linear change of the electrical current with the surface coverage is not a specificity of our system. Rather, the same scaling is expected for every situation in which a selfassembled adlayer introduces a field effect in graphene via ordered alignment of permanent dipoles. Additionally, a very similar situation is expected for all molecular processes which generate a variation of ordered (out-of-plane) dipoles at the graphene surface. In this regard, we anticipate that our method and our machinery for data analysis might allow monitoring the S18 dynamics not only of other self-assembled adlayers, but also of other interfacial phenomena, such as 2D polymerization.
On the contrary, the absence of an ordered assembly for the SP isomer at the solid/liquid interface implies that molecular dipoles in the vicinity of the surface are randomly oriented at any moment in time, hence they do not introduce a strong field effect. We highlight that at the solid/air interface we were able to visualize a self-assembled adlayer for the same SP derivative, in which alkyl chains were imaged with high resolution, while the SP head groups gave rise to fuzzy regions.
Our interpretation of those images, supported by molecular dynamics simulations, was that SP isomer were not immobilized on the surface, even though the alkyl chains anchored them on the substrate. In the present case, the situation at the solid/liquid interface is more dynamic, as molecules are more mobile; hence, we are not capable of imaging even the alkyl chains. The desorption is proportional to the number of MC molecules on the substrate Nsub times the desorption rate Kdes. Thus, one can write: In which the first term describes the adsorption of MC molecules from solution, and the second term the desorption of the MC molecules from the substrate. It is worth noting that during photoisomerization, the number of MC molecules Nsol is not constant, but it varies with time. In particular, during UV irradiation the MC concentration in solution increases, following the exponential trend which can be monitored by the absorbance: Where Nsta is the number of MC molecules at the stationary state, reached when the SP → MC photoisomerization is balanced by the MC → SP thermal conversion, and τhv,O is the time constant extracted from the optical characterization, as defined in the main text. Instead, after UV irradiation, the number of MC molecules in solution decays as: Where τΔ,O is the time constant extracted from the optical characterization for the thermal decay.
Substituting the concentration dependence into (1), one obtains a non-homogeneous first order differential equation which does not have any trivial exponential solution.
However, one can understand the exponential dependence of the current during UV irradiation by considering, as a rather brute approximation, that the number of MC molecules does not varies with time, but it is rather constant at a value Nsol ~ C. In this case, one can rewrite Eq. (2) as: which has solutions in the form: High initial SP concentration C0 will generate a high number of MC C, so that C will be higher for higher C0, in agreement with the electrical measurements. We point out that the approximation of constant Nsol ~ C fails immediately after switching on the UV light. Indeed, at the very beginning of the irradiation, the MC concentration is zero, and it requires a certain time before reaching a finite concentration. During such time, the relative variation in MC concentration is maximal. Instead, after a relatively short time (a few s) a critical MC concentration is reached, and further increase in Nsol become less relevant. In this regard, by fitting the IDS(t) curves after the initial time required for a significant MC concentration to build up (see Figs. S4-S6), we capture the dynamics of assembly formation with the simple exponential fit.
In the light of this discussion, another interesting observation can be extracted by comparing the decay in I(t) related to the desorption of the self-assembled MC adlayer and the decay in absorbance, as reported in Fig. 3e. During approximately one minute after switching off the UV light, the decay in I(t) is faster than that in Abs(t). On the contrary, according to equation 1, one S21 would expect that the desorption of the self-assembled adlayer were slower than the thermal recovery in solutionwhich would reflect a physical situation in which desorption takes place after the density of MC in solution has decreased significantly. This indicates that there is a physical phenomenon which favors the desorption of the self-assembled adlayer at a timescale faster than the thermal reconversion in solution, which is not contained in equation (2). While our experiment does not provide direct information on such phenomenon, we postulate that it is the MC → SP conversion occurring on the surface rather than in the solution, which leads to the dissolution of the MC assembly following other (faster) dynamics.

Ensemble dynamics
Ensemble molecular processes involving a large number of molecules, such as photoisomerization in solution or the formation of self-assembly, typically evolve according to relatively slow time constants, ranging from the ms to the s. While (ultra)fast molecular events take place at the single molecule level, the dynamics of systems composed of a large number of molecules are typically dominated by ensemble quantities describing population-averaged stochastic processes 5 , which lead to orders-of-magnitude slower dynamics. For example, while at the single molecule level light-induced isomerization of photochromes takes place at an ultra-fast timescale (fs-ns) 6,7 , at the ensemble level the dynamics are determined by absorption and photoisomerization quantum yield 8,9 , leading to ensemble dynamics in the few-seconds range 10 .
Analogously, the phenomena governing the motion of single molecules on surfaces (vibrations and hopping rates) occur at the ps timescale at room temperature, 11 while self-organized supramolecular assemblies composed of a large number of molecules evolve on slower (fewseconds) timescales, which are ideal to be captured by our device method using a 10 -100 Hz sampling rate easily achievable with conventional electronics.