Optical Signatures of Phosphorene Chemical Evolvement in Liquid Environment

Phosphorene possesses a unique combination of physical and chemical properties, including tunable direct bandgap, anisotropic electronic and optical properties. However, application of phosphorene is hampered by its fast degradation under ambient conditions. It is believed that H2O and O2 play key roles in the instability of phosphorene, but their exact contributions have not yet been completely understood in experiment. Herein, a technique of probing the mechanism of phosphorene oxidation through the evolvement of its photoluminescence spectra in a liquid suspension is introduced. With the addition of H2O2, the photoluminescence intensity of the suspension surprisingly increases, indicating a passivation effect due to the formation of phosphorene oxide, which is also verified by Raman spectroscopy. In contrast, direct addition of H2O leads to irreversible and rapid degradation as shown by the quenching of photoluminescence. Furthermore, monolayer phosphorene suspension photoluminescence at 2.17 eV is obtained by completely eliminating effects from contaminations, substrate strain, dielectric substrate, and oxidation of samples. Solvent protection in combination with the optical probing method provides an effective approach in investigating the chemistry of active materials like phosphorene. Phosphorene has great potential in the range of sensing applications including high‐resolution H2O2 and humidity sensors.

Phosphorene possesses a unique combination of physical and chemical properties, including tunable direct bandgap, anisotropic electronic and optical properties.However, application of phosphorene is hampered by its fast degradation under ambient conditions.It is believed that H 2 O and O 2 play key roles in the instability of phosphorene, but their exact contributions have not yet been completely understood in experiment.Herein, a technique of probing the mechanism of phosphorene oxidation through the evolvement of its photoluminescence spectra in a liquid suspension is introduced.With the addition of H 2 O 2 , the photoluminescence intensity of the suspension surprisingly increases, indicating a passivation effect due to the formation of phosphorene oxide, which is also verified by Raman spectroscopy.In contrast, direct addition of H 2 O leads to irreversible and rapid degradation as shown by the quenching of photoluminescence.Furthermore, monolayer phosphorene suspension photoluminescence at 2.17 eV is obtained by completely eliminating effects from contaminations, substrate strain, dielectric substrate, and oxidation of samples.Solvent protection in combination with the optical probing method provides an effective approach in investigating the chemistry of active materials like phosphorene.Phosphorene has great potential in the range of sensing applications including high-resolution H 2 O 2 and humidity sensors.
Among them, monolayer phosphorene is the most unstable, and the lack of a robust experimental technique for fast and precise monolayer identification makes its identification and characterization extremely challenging.][29] Uncertainties due to dielectric constants of surrounding media, surface defects, contamination, and oxidation of samples may affect the reliability of measured bandgap values.These contradicting studies pose great hindrance on the understanding of interesting phenomena such as bandgap tunability.
Here, we present a liquid suspension-based photoluminescence study on phosphorene after liquid-phase exfoliation.The exfoliated phosphorene is suspended in pure ethanol in a transparent cuvette with cover, which prevents phosphorene degradation or subsequent oxidation due to its isolation from oxidants (Ar purging) and also allows optical characterization of the suspension.In order to introduce oxygen and H 2 O into the system in a controllable way, we added trace amount of H 2 O 2 and H 2 O to the suspension and monitored the PL response of the system.There are multiple advantages of using H 2 O 2 here.First, we can precisely control the amount of oxygen compared with bubbling oxygen by adjusting the amount of H 2 O 2 .Second, H 2 O 2 won't introduce any new elements or compounds into the system which might make the analysis more complicated (Only H, O, and P compounds exist in the final products).In the following, we show from the PL results that phosphorene reacts with H 2 O 2 to form P x O y when H 2 O 2 is introduced and becomes resistant to further degradation, which is also verified by the Raman signatures of the oxidation process in the liquid.While on the other hand, when only H 2 O was introduced, phosphorene degrades irreversibly and rapidly as indicated by the evolvement of the PL spectrum.Because phosphorene is protected in the liquid throughout our experiments, we are able to obtain precise photoluminescence spectra of monolayer phosphorene and 2.17 eV sets a lower bound for the bandgap of monolayer BP.The revealing of oxidation mechanism of monolayer phosphorene by dynamic photoluminescence spectroscopy not only offers opportunities in phosphorene protection in further studies, but may also lead to high-sensitivity H 2 O 2 and humidity sensors.

Results and discussion
Figure 1a,b shows, respectively, the low-and high-resolution transmission electron microscopy (HRTEM) images of the samples prepared using the liquid route from bulk BP crystal powder, which are directly transferred onto the ultrathin carbon transmission electron microscope (TEM) grid.A TEM image of thin BP nanosheets is shown in Figure 1a.While the HRTEM image (Figure 1b) shows the high crystalline quality of the as-transferred BP sample onto the TEM grid. Figure 1b also shows the typical selected-area electron diffraction (SAED) pattern of the monolayer BP samples, which confirms the BP orthorhombic crystalline character.Moreover, the clear HRTEM image of the BP nanosheet obtained from the yellow rectangular area in Figure 1b shows lattice fringes of 0.214 nm (Figure 1c), which can be ascribed to the (112) plane of the BP crystal.Further evidence of the chemical nature and purity of the liquid exfoliated flakes is provided by energydispersive X-ray analysis (EDXA, Figure S1, Supporting Information).The thickness values of the phosphorene samples are further determined via atomic force microscopy (AFM).Figures 1d shows the typical AFM image and height profile, which confirm that the height of the phosphorene is %0.81 nm.The AFM procedure is done within 10 min after dropcasting sample solution on SiO 2 /Si substrate to minimize degradation.This finding indicates the monolayer nature of the BP nanosheet sample.Previous AFM measurements have suggested that single-layer phosphorene is 0.6-0.9nm thick. [24]We have also identified bilayer and trilayer BP in the sample, which are shown in Figure S2, Supporting Information and 3, respectively.
In previous studies on the degradation of few-layer BP, the role of oxygen and water in few-layer phosphorene degradation was studied under ambient conditions, [14,15,30] and it has been shown that degradation of phosphorene is initiated by contact with oxygen.Instead of getting involved in the reaction directly, water is capable of removing P x O y from the surface, exposing P 0 , and allows further oxidation to proceed. [31]Castellanos-Gomez et al. once attributed the degradation to the intrinsically hydrophilicity of few-layer BP surface with water. [25]However, there is no direct experimental result to distinguish the effects of oxygen and water for monolayer phosphorene.Here we found that our H 2 O 2 technique offers a unique opportunity to probe the degradation process of monolayer phosphorene in a controllable way.PL spectroscopy measurements are done on liquid-exfoliated phosphorene stabilized in pure ethanol (Ar purging).In particular, after adding H 2 O 2 or H 2 O in experiments, each optical spectrum was obtained within 10 s in our experiments.The samples were shielded from light exposure between two PL measurements to minimize possible degradation induced by exposure to the laser light.The PL spectra upon addition of small amount of H 2 O 2 are shown in Figure 2. Here, we identify the PL peak near 572 nm as the PL from monolayer phosphorene and perform further analysis.In order to verify this result, we first obtained the background spectrum without laser excitation and PL spectra of the blank cuvette, presented in Figure S4, Supporting Information.No discernable signal can be detected in these spectra.Optical spectra of pure ethanol and pure Nmethylpyrrolidone (NMP) are also provided in Figure S5, Supporting Information, where peaks are identified as welldocumented Raman shifts of 889, 1098, 1470, 2949, and 3365 cm À1 .This also explains that some additional peaks in PL spectra shown in Figure 2a are from the Raman shifts of ethanol. [32]Most importantly, the PL intensity of phosphorene in ethanol increases when H 2 O 2 is being added to the suspension, as shown in Figure 2b.In order to eliminate fluctuations of the measurement, the photoluminescence fluctuations are observed at about 1% at different times (we once tested photoluminescence fluctuations at different time such as 0, 5, 10, 15, 20, 25 min) for the same sample in our experiments (Figure S5c, Supporting Information).This rather counterintuitive PL behavior uncovers valuable information regarding the oxidation mechanisms of phosphorene.The phosphorene preferentially reacting with H 2 O 2 to form P x O y when H 2 O 2 is added is verified by following Raman spectroscopy.
Raman spectroscopy has been shown to provide metrics on the extent of oxidation of phosphorene. [33]Figure 2c shows the typical Raman spectrum of the phosphorene samples recorded at room temperature with an excitation wavelength of 473 nm.Three prominent peaks can be ascribed to an out-of-plane phonon mode A g 1 at 364.1 cm À1 and to two in-plane modes, namely, B 2g and A g 2 , at 441.4 and 468.9 cm À1 , respectively.It has been shown that the intensity ratio of the A g 1 /A g 2 phonons has a strong dependence on oxidation-induced sample degradation. [14]This ratio typically becomes smaller when the sample has more oxidation-induced passivation. [18,34]We therefore analyzed the A g 1 /A g 2 intensity ratio of Raman spectra when different amounts of H 2 O 2 were added into the solution.The resultant intensity ratio histogram is plotted in Figure 2d.This intensity ratio decreases with increasing amount of H 2 O 2 , indicating an increase in the extent of oxidation passivation.(We also used a 532 nm laser excitation, which is in support Figure S6, Supporting Information).Moreover, the P x O y -passivated phosphorene shows resistance to H 2 O-induced degradation.H 2 O is a product of H 2 O 2 reactions, but even with the appearance of H 2 O, the PL evolvement is still distinctive from following direct addition of H 2 O.The P x O y formed is relatively stable even though water is present (as shown in Figure 2b), as predicted in previous simulation work. [10]We postulate that the P x O y formation leads to the occupation of the lone pair electrons and hence may prevent the reaction between phosphorus and oxygen, ultimately mitigating the degradation of phosphorene.Only when large amount of H 2 O is added, P x O y may react with water and transform into phosphoric acid.The PL optical spectra of the phosphorene samples after adding large amount of H 2 O 2 or H 2 O are shown in Figure S7, Supporting Information.The PL intensity (around 572 nm) decreased monotonously with the increase in the large amount of H 2 O 2 or H 2 O. Figure S8a,c, Supporting Information, shows the resolution of the P 2p XPS spectra of the phosphorescent samples with and without the addition of large amounts of H 2 O 2 .The XPS spectrum of these BP sample with H 2 O 2 added shows a dominant maximum peak at 129 eV and a broad peak at 131-135 eV, which are attributed to P─P bonds and P─O bonds, respectively.Fitting curves further revealed three peaks at 128.5, 129.3, and 133.0 eV, which can be attributed to P 2p 3/2 , P 2p 1/2 , and P x O y , respectively.Meanwhile, the XPS spectrum of those BP sample without H 2 O 2 added shows similar peak positions; the intensity of the P x O y peak is significantly weaker than that of the P x O y peak of the BP sample after treatment with hydrogen peroxide.High-resolution O 1s XPS spectra of the phosphorescent samples with and without the addition of large amounts of H 2 O 2 (Figure S8b,d, Supporting Information) show two sharp peaks at 530.7 and 532.0 eV, which are attributed to the P═O and P─O─P bonds.We also found that P─O─P predominated in the hydrogen peroxide-treated BP samples compared to the nonhydrogen peroxide-treated BP samples.[37] The above effective H 2 O 2 technique may also allow us to precisely engineer the defects in monolayer phosphorene, triggering new applications in nanoelectronics and optoelectronic devices. [38]o gain more physical insight of the increased PL intensity, density functional theory (DFT) simulation is carried out on defect monolayer BP and O-modified BP. Figure 3a shows the local density of states (LDOS) map along armchair direction in the real space, the formation of single-P vacancy of monolayer BP decreases E F , giving rise to a P-type feature.The periodic lattice field is broken by the P vacancy, the localized states' distortion near E F can be observed at defect site.It is in accordance with the emergence of a flattened band at E F in Figure 3b, which is contributed by the P_ p z orbits.With the oxygen modifying the P vacancy, although the states' localization at E F still occurred, the flattened band is contributed by P_p y and P_p z orbits (Figure 3d and S9, Supporting Information).However, the periodic crystal lattice field below the E F is maintained, as shown in Figure 3c, which would contribute to the increased states' transition with the same momentum.Above theory is similar to the effect in TMDCs that the formation of surface passivation minimizes the defect trapping of charge carriers, leading to the enhanced PL signal. [39]n contrast, Figure 4a,b shows that the PL intensity is decreased when only H 2 O is added, degradation of phosphorene occurs immediately, which is generally believed that there is formation of phosphorus trioxide or phosphorus pentoxide or phosphates found in the literature. [40]The Raman result also indicates that phosphorene directly degrades when there is only H 2 O present.From Figure 4c,d, all of the spectra show that intensity ratio of A g 1 /A g 2 is around 0.4 when H 2 O was added into the solution, confirming that the phosphorene samples are not oxidized, but direct degradation occurs (a similar result is shown in Figure S6c,d, Supporting Information).Monolayer phosphorene is shown to directly react with water from the spectra.In other words, the degradation of monolayer phosphorene is primarily due to water rather than oxygen in our experimental results.Coleman et al. proposed the following acid-base disproportionation reaction with phosphorene. [17] BP 2VAC is a BP flake with two vacancies.The reaction energy of Equation ( 1) is evaluated %ΔE = 0.26 eV when occurring far from the edge, which is very similar to the value for the reaction started on the ideal monolayer.Considering that chemical reaction with an energy barrier less than 0.9 eV (%21 kcal mol À1 ) from DFT calculations could occur at room temperature, [11] such a process is expected to occur spontaneously at room temperature.These calculations further show that monolayer phosphorene can react with water, even in the absence of oxygen and light.All our analysis earlier is based on the PL peak near 572 nm, indicating that a bandgap no less than 2.17 eV originated from monolayer phosphorene, which is in accordance with the theory value predicted by DFT calculations (see Figure 3c,d) and the experimental value obtained using the scanning tunneling microscopy method. [25,27]It has been shown that the PL peak intensity increases dramatically when the number of layer decreases, in spite of the reduced amount of material. [41]For this reason, we target the characteristic PL peak of monolayer phosphorene in our experiment.Photoluminescence spectra of a single-monolayer phosphorene flake protected by ethanol using an excitation wavelength of 473 nm are also presented in Figure S7, Supporting Information.It shows the same peak position that is consistent with the photoluminescence spectra of large amount of phosphorene in solution using an excitation wavelength of 532 nm.These photoluminescence spectra corroborate with each other.Previously reported PL results of monolayer phosphorene often obtained by micromechanical exfoliation and transferred onto SiO 2 /Si substrates greatly varied, depending on the static dielectric constant of the surrounding medium.In addition, surface defects, contamination, and oxidation of samples may introduce further experimental uncertainties.[44] From Figure S11, Supporting Information, this is consistent with the photoluminescence spectra of the phosphorene in solution at the higher concentration of H 2 O 2 or H 2 O (Figure S7a,c, Supporting Information).These photoluminescence spectra are corroborating with each other.Our method offers a stain-free environment for BP photoluminescence.Degradation process is also observed on our monolayer phosphorene once they are exposed to ambient environment.The monolayer phosphorene degrades within around %10-20 min after evaporation of ethanol.This further demonstrates the effectiveness of our liquid suspension-based photoluminescence study.

Conclusion
In conclusion, we have demonstrated a novel method to uncover the degradation mechanism of monolayer phosphorene through observing the evolvement of the PL spectra of its liquid suspension in a controllable way.We show that phosphorene gets passivated by the formation of P x O y in the presence of H 2 O 2 , which would increase the PL intensity due to the repairing of P vacancy by O atoms.Monolayer phosphorene also degrades in the presence of water solely.Moreover, the stabilized phosphorene monolayer in solution provides us with a perfect platform to probe the bandgap of monolayer phosphorene.This work provides optical evidence of phosphorene evolvement in a protected solvent environment and paves the way toward deeper understanding of the chemistry of the degradation of phosphorene.

Experimental Section
Sample Preparation: BP crystals powder was purchased from Sigma Co., Ltd and stored in a dark Ar glovebox.N-monthylpyrrolidinone (NMP) was obtained from Sigma Co., Ltd.The concentration of H 2 O 2 was diluted into 1000 μM.In a typical procedure, 10 mg of BP powder was added into 25 mL of NMP sonicated in an ice bath for 6 h at the power of 180 W. The resultant dispersion was centrifuged for 20 min at a speed of 5000 rpm.The supernatant was further centrifuged for 20 min at a speed of 7000 rpm.
Characterization: TEM images were obtained using an aberrationcorrected TEM (Titan 80-300).AFM images were taken on a Veeco Multimode 8 in tapping mode.PL spectra with an excitation wavelength of 532 nm were measured in a customized PL system.Laser beam was focused into solution in a 1 mm-thick cuvette by a long working distance objective lens (40Â).PL signal passing through a 550 nm long-pass filter was collected by a spectrometer (Horiba iHR320) and a charge coupled device detector (Horiba Synapse).PL spectra with an excitation wavelength of 473 nm were measured by the PL function in a Raman system (LabRAM HR Evolution, HORIBA) with a 10Â objective lens (solution in a 1 mm-thick cuvette) and a 100Â objective lens (flakes).Raman spectra were recorded using a HORIBA inVia System with an excitation wavelength of 473 nm.
Theoretical Calculation: The first-principle calculations were carried on the Quantum ATK software.The exchange functional and correlation functionals were described by hybrid B3LYP while the norm conserving pseudopotential was used.The wave function cutoff was 544 eV and the k-point sampling grid for 2D crystal slabs was 3 Â 1 Â 2. The vacuum spacing layer between neighboring repeatable units of 2D slabs was set to 1 nm.All of the atomic sites of as-built 2D slabs were fully relaxed before calculations of energy dispersion curves, absorption spectra, and localized density of states.

Figure 1 .
Figure 1.Basic characterization of liquid-exfoliated phosphorene.a) TEM image of the phosphorene sample.b) HRTEM image of the area marked a red rectangle in (a).Inset shows the SAED pattern.c) HRTEM image of the area marked a yellow rectangle in (b).d) Typical AFM image of the phosphorene sample deposited on a SiO 2 /Si substrate and inset shows the corresponding AFM height profile of the sample.

Figure 2 . 1 /A g 2 from
Figure 2. Photoluminescence and Raman spectra of the phosphorene dispersed in pure ethanol.a) Photoluminescence spectra of phosphorene with different amounts of H 2 O 2 added.b) Relationship between PL intensity and H 2 O 2 volume.c) Raman spectra of phosphorene with different amounts of H 2 O 2 added.d) Relationship between the peak ratios A g 1 /A g 2 from Raman spectra and H 2 O 2 volume.

Figure 3 .
Figure 3. DFT calculations.a) Calculated local DOS of monolayer BP with single P defect.b) Orbit-resolved band structure of single P-vacancy BP models.c) Calculated local DOS of monolayer BP with O modified the single P vacancy.d) Orbit-resolved band structure of O modified the single-P vacancy BP models.

Figure 4 .
Figure 4. Photoluminescence and Raman spectra of the phosphorene dispersed in pure ethanol.a) Photoluminescence spectra of phosphorene with different amounts of H 2 O added.b) Relationship between PL intensity and H 2 O volume.c) Raman spectra of phosphorene with different amounts of H 2 O added.d) Relationship between the peak ratios A g 1 /A g 2 from Raman spectra and H 2 O volume.