From prehistoric symbolic murals carved on caves by stone, to the confidential information written on letter paper by invisible ink, the continuous improvement of writing system has witnessed human progresses. A special type of writing, steganography, which is usually taken as a top secret, is evolving with time. The formulas of secret ink used in the First World War were ordered to remain hidden from the public in 20021 , even though these secret writing methods merely stay at the primary level of chromogenic reaction. At the turn of this century, it was reported that the steganographic technique could be advanced to the molecular level by concealing secret messages in DNA-based microdot2,3 and macromolecules4. By thinning down paper and ink to the atomic scale, one may reach the limit of writing system and take steganography a step further, even enabling the storage of data in an atomic-thin memory. The path to store information at the atomic limit can be carried out by scanning probe microscopes (SPM), e.g. scanning tunneling microscope (STM) through directly manipulating lattice locations5,6, charge states7,8 or magnetization states9–11, or atomic force microscope (AFM) methods like dip-pen nanolithography (DPN)12–14 that can deliver a wide range of substances onto various substrates13,15,16. The SPM methods require low temperature, clean surface, ultrahigh vacuum environment and/or complicated operations17,18, and their information readout typically relies on delicate SPM or electronic microscopes. These are undesirable features for steganography of which users are usually under intense pressure with limited conditions and have limited time to work.
Herein, using a monolayer of water as ink, an atomic layer of MoS2 as paper, and a laser as pen, concealing information at the atomic-thin level is achieved under ambient conditions in a rewritable manner. The optical image of an n-doped MoS2 monolayer is displayed in Fig. 1a. Its PL is enhanced upon water absorption, and the desorption of water by laser irradiation suppresses the PL enhancement, as shown in Fig. 1b. Raman spectroscopic measurements suggest that the laser irradiation does little damage to the MoS2 monolayer. As displayed in Fig. 1c, absorbing H2O molecules blueshifts and intensifies the out-of-plane vibrational mode A1’ of MoS2, whereas the in-plane vibrational mode E’ stays unchanged (blue line). After the laser irradiation treatment, the Raman spectrum (red line) changes back to be almost identical to the pristine one (black line), suggesting that there is little structural change within the monolayer of MoS2. To investigate the mechanism of PL enhancement by absorbing H2O molecules, a field-effect transistor (FET) device is fabricated to explore the variations in carrier concentration of MoS2. Fig. 1d&e display the optical images of the FET device before and after adsorbing H2O molecules, respectively. Their appearances look the same, but their electronic properties vary significantly. The source-drain current (ISD) versus back gate voltage (VG) curves in Fig. 1f indicate that the pristine device (black curve) shows an n-type characteristic. After absorbing H2O molecules on the surface, the electron concentration drops sharply (blue curve), suggesting that the physically adsorbed H2O molecules have a p-type doping to MoS2. A p-type doping favors the formation and recombination of neutral excitons over negative trions. According to literatures19–21, neutral excitons have a faster radiative recombination rate than the negative trions and thus have a higher PL efficiency, leading to the PL enhancement. Once the absorbed H2O molecules are erased by annealing in vacuum (200 oC for 3 h), the n-type characteristic (green curve) is restored, and in fact is even slightly enhanced compared to the pristine sample, because fewer water molecules are on the annealed sample.
The laser irradiation removing the surface H2O molecules is further verified with AFM mapping. A square region by laser irradiation on the MoS2 sample (the blue rectangle area denoted in Fig. 1a) is not distinguishable by optical microscope, whereas it is clearly visible under AFM mapping (Fig. 1g). The line profile in Fig. 1h indicates that the thickness of the monolayer MoS2 before laser treatment is 1.43 nm, which is higher than the theoretical value of monolayered MoS2 owing to the adsorbed water. The average thickness of the laser-treated square region (Fig. 1i) is measured to be 0.75 nm, consistent with the reported monolayer thickness of MoS222. The thickness reduction caused by laser treatment is about 0.68 nm, which agrees well with the thickness of a single layer of water molecules (~0.4 nm) plus the adsorption equilibrium distance (~0.2 nm)23. The measured thickness of MoS2 monolayer with absorbed water is consistent with previous theoretical studies24,25.
By removing surface water in designed patterns with lasers and the re-addition of water by humidification, rewritable data-encoding and readout through simple PL mapping can be achieved. The encoding scheme is illustrated in Fig. 2a. The sample is moisturized before laser writing, and then patterns are written directly onto the MoS2 monolayer with a laser by removing the surface water. A home-built laser writing system is used to conduct the information-writing process (Supplementary Fig. 1). The laser-written information is read out by PL imaging, and is erased by exposing the sample to a humid environment. Fig. 2b displays a Chinese character “仁” with the meaning of benevolence that is written onto the MoS2 monolayer. The writing of “仁” (Fig. 2c) and the erasing of it with water absorption (Fig. 2d) don’t produce observable changes on the sample’s optical images. Through PL mapping, the written pattern is clearly visible (Fig. 2e). Under ambient conditions (humidity < 30%), the written pattern can last for at least one week. The pattern is successfully erased after the sample is placed in a humid environment (relative humidity ~85% for about 10 minutes) (Fig. 2f). The PL spectra of the sample before any treatments (black curve), after laser writing (red curve), and after pattern-erasing by adsorbing H2O molecules (blue curve) are displayed in Fig. 2g. They are very similar to those in Fig. 1b, of which the PL intensity is dependent on the surface water content as discussed above. During writing, the laser power density is carefully controlled at about 3.0 × 105 w/cm2 to ensure that the thermal energy generated by MoS2 photo-absorption is sufficient for the removal of water but below the MoS2 damage threshold. Overwriting with too much laser power can cause permanent damages to the MoS2 monolayer and the written pattern cannot be erased by adsorbing H2O molecules (Supplementary Fig. 2).
The MoS2 monolayer is rewritable for many times. Fig. 3 displays the PL images of “仁” written in the solid-filled form (upper panels) and erased (bottom panels) at the 1st, 10th, 20th, 50th cycle. The upper panels are taken after each writing. “仁” is still visible after 50 writing/erasing cycles. The PL images in the bottom panels are taken after each humidification treatment, showing that “仁” is completely removed by the erasing process. To test how long the written pattern can last, we keep the sample in a container at 30% relative humidity (RH) for a week right after the 20th writing. Water in the air has not been able to erase the written pattern for such a long period of time, and “仁” is still clearly distinguishable, as shown in the upper 20th panel in Fig. 3. The corresponding optical images of the sample after each writing/erasing cycle also show no noticeable changes on the MoS2 monolayer during the processes (Supplementary Fig. 3).
Not only can a simple word or symbol be written with the method, but also a sentence and even a painting. The phrase “Nothing is Impossible” (Fig. 4c) and the famous painting Mona Lisa (Fig. 4f) are successfully written on continuous MoS2 monolayered films. The line drawing of Mona Lisa (Fig. 4f) is probably the thinnest portrait by far, since the “paper” is a monolayer of MoS2 and the “pigment” is one layer of H2O molecules.
The writing/erasing process is reasonably fast, dependent on the complexity of the writen pattern. A more complexed pattern takes a longer time. With the same condition, writing “仁” and “π” of a similar size takes about 60 s and 15 s, respectively. It takes shorter than 10 mins to obtain their PL images by mapping. The erasing can be extremely fast. In a destructive way, if the sample is dipped into water or blown by breath, the pattern disappears immediately, but, in most of our tries the MoS2 monolayer is broken after the erasing process. In a nondestructive manner, the sample is placed in an environment with a RH of 85%, the written pattern disappears within 10 minutes. In total, it takes minutes for the entire writing/erasing/readout process. The time can be significantly shortened by optimizing the operation conditions and tools, e.g. using 2D PL imaging rather than point mapping used in this work can cut the readout time from minutes to seconds, and using a pulsed laser that can deliver enough heat to desorb water molecules before the thermal energy diffuses out of the focus spot would increase both the writing speed and the spatial resolution.
In summary, atomic-thin erasable and rewritable information recorded on a MoS2 monolayer under ambient conditions is achieved by modulating and imaging the PL of a MoS2 monolayer with surface water reversible absorption/desorption. The messages encoded in this way can be at least under triple protections: information is encrypted, the writing appearance is invisible, and the designated MoS2 triangle is camouflaged among a jumble of MoS2 triangles; and they can be easily and immediately erased by breath or even saliva. The miniaturization of the equipment needed for the writing/readout processes is feasible and their operations can be foolproof. These are desirable and promising features for many applications, e.g. steganography, anti-counterfeiting, sensors, and transient information storage.