Raman Sensitive Degradation and Etching Dynamics of Exfoliated Black Phosphorus

Layered black phosphorus has drawn much attention due to the existence of a band gap compared to the widely known graphene. However, environmental stability of black phosphorus is still a major issue, which hinders the realization of practical device applications. Here, we spatially Raman map exfoliated black phosphorus using confocal fast-scanning technique at different time intervals. We observe a Raman intensity modulation for , B2g, and modes. This Raman modulation is found to be caused by optical interference, which gives insights into the oxidation mechanism. Finally, we examine the fabrication compatible PMMA coating as a viable passivation layer. Our measurements indicate that PMMA passivated black phosphorus thin film flakes can stay pristine for a period of 19 days when left in a dark environment, allowing sufficient time for further nanofabrication processing. Our results shed light on black phosphorus degradation which can aid future passivation methods.

We carried out AFM measurements of the black phosphorus thin film flake in figure 2 to determine the initial thickness of the flake. We scanned the edge of the flake as shown in figure  S2. In this image, the flat part is the non-degraded part, while the lower part that exhibit high intensity shows the bubble formation that was shown previously by Island et al., where the volume of the black phosphorus flake has increased significantly after few days. The estimated flake thickness that was used in the intensity enhancement model is d = 35nm ± 5nm. c. Raman characteristics at a specific site on the flake.
The Raman characteristics of a specific site on few layers black phosphorus flake are shown in figure S7 for 1 , 2 , and 2 Raman modes. The Intensity of each mode shows a decrease, analogous to previous reports [1]. Also, we observe slight Raman upshift before the flake completely degrades. This Raman upshift reflects layers lowering, which can be interpreted as layers etching [1,2].  I. Raman intensity maps at different time intervals We exfoliated a thick black phosphorus flake as shown in figure S7. The flake intial thickness is is estimated to be as large as ~200nm. The sheet shows edge degradation as the first sign of degradation in the optical image for T=3days. Subsequently, surface degradation starts to show on the surface of the sheet. The sheet completely degrades after 57 days. We monitored the Raman maps of each of 1 , 2 , and 2 Raman modes at different time intervals in order to assess the mechanism of the degradation ( figure S8). In order to validate the intensity enhancement model due to interference, AFM measurements have been carried out on a thick flake at a specific time interval in order to compare the thickness after degradation with the intensity enhancement model, Our AFM measurements are in good agreement with our measured Raman intensity modulation, further cooperating our assessment on the intensity enhancement due to optical interference. Below is the AFM image along with the line profile showing the thickness of black phosphorus for t=23 days.

5-PMMA coated few layers black phosphorus:
In order to assess the degradation rate for PMMA coated few layers black phosphorus, we monitored the flake degradation Figure S15. Optical images and Raman Intensity maps of 1 (red color), 2 (green color), and 2 (cyan color) vibrational peaks for a few layers flake. The first signs of degradation occur on the second day. In our calculations, there are two major processes take place that contribute to this Raman intensity modulation. The first process is the absorption of the incident laser on the black phosphorus thin film as illustrated in figure 3. The final net absorption is governed by: , and 3 = 2 − 3 are the transmission coefficients. n0, n1, n2 and n3 are the refractive indices for air, BP sheet, SiO2, and Si, respectively. These values are found in ref [3,4]. d 1 and d2 are the thickness of as exfoliated BP sheet and SiO2. X is the depth in the BP sheet that the laser is being absorbed and scattered. The coefficients , 2 ,and 2 are given by = 2ᴨ 1 , 1 = 2ᴨ 1 1 , 2 = 2ᴨ 2 2 , where λex is the excitation wavelength of the laser (532nm in our case).
The second process is the Raman scattering interference, which occurs inside the BP sheet, as schematically shown in figure S10b. The scattering term (Fsc) can be calculated according to: , and for the scattering terms are given by = 2ᴨ 1 , 1 = 2ᴨ 1 1 , 2 = 2ᴨ 2 2 , where λsc is the desired Raman mode wavelength.
The net enhancement factor after taking the absorption and the Raman multiple scattered light into account is given by [5]: Where N is the normalization factor and it is introduced in the equation to remove the effects of the substrate (SiO2 and Si). Figure S17 plots the intensity enhancement factor with and without the normalization factor. In this figure, the intensity enhancement for the zigzag direction shows significant increase compared to the armchair direction.

b. Thickness and etching rate estimation.
It is possible to estimate the BP thickness using the interference model by fitting the measured intensities of each Raman mode to the calculated enhancement model. In doing so, the boundary conditions should be identified, which are the initial thickness of BP sheet and the SiO2 thickness. For the flake in figure 2, the thickness is around 35nm, confirmed by AFM measurements, and the SiO2 thickness is around 280nm. For black phosphorus sheets with thicknesses ranging between 0-40nm, the intensity enhancement profile exhibit negligible differences for zigzag and armchair directions, as illustrated in figure 4. Accordingly, the thickness can be obtained at different time intervals by fitting the normalized intensity profile to the interference model profile described above. Figure S12a, S12b, and S12c show the thickness vs. time calculated for sites A, B, and C on the flake highlighted in figure 3 for 1 , 2 , and 2 , respectively. In figure 4b of the paper, we plotted the average thickness which is obtained from the S12 figures. The estimated etching rate shown in figure 4c is taken as the derivative of the average thickness.  Figure S22. Raman intensity maps of degraded thin film black phosphorus flake (a) before annealing and (b)directly after annealing in vacuum for 1 (red), 2 (green), and 2 (cyan). The optical image of the flake shows a clean surface after annealing due to removal of liquid interfaces. After Annealing