Two-dimensional amorphous NiO as a plasmonic photocatalyst for solar H2 evolution

Amorphous materials are usually evaluated as photocatalytically inactive due to the amorphous nature-induced self-trapping of tail states, in spite of their achievements in electrochemistry. NiO crystals fail to act as an individual reactor for photocatalytic H2 evolution because of the intrinsic hole doping, regardless of their impressive cocatalytic ability for proton/electron transfer. Here we demonstrate that two-dimensional amorphous NiO nanostructure can act as an efficient and robust photocatalyst for solar H2 evolution without any cocatalysts. Further, the antenna effect of surface plasmon resonance can be introduced to construct an incorporate antenna-reactor structure by increasing the electron doping. The solar H2 evolution rate is improved by a factor of 19.4 through the surface plasmon resonance-mediated charge releasing. These findings thus open a door to applications of two-dimensional amorphous NiO as an advanced photocatalyst.


Supplementary
As shwon in Supplementary Fig. 2a, it can be clearly seen that BC is irregular with a size of several hundred nanometers. HRTEM image in Supplementary Fig. 2b presents a typical interplanar spacing of 0.2087 nm, which corresponds to the (200) crystallographic plane of cubic bunsenite NiO phase (JCPDS 47-1049). 2DC is in the shape of nanoflake with a thickness less than 10 nm, similar to 2DA and 2DPA ( Supplementary Fig. 2d) optical phonon modes (2LO), respectively. The last band at 1532.6 cm -1 can be assigned to the two-magnon scattering mode (2M). 1P modes of 2DC are slightly weaker than those of BC, suggesting 2DC is slightly disordered due to the 2D effect.
As shown in the FTIR spectra in Supplementary Fig. 2h

Supplementary Note 2
As shown in the Ni 2p 3/2 XPS spectra in Supplementary Fig. 4a  and suitable for the production of meta-stable materials.
The plasma plume was inclined to expand, however, the expansion was strongly confined by water. A shock wave was then induced by the continual supply of vaporizing species in the plasma plume. This shock wave resulted in extra pressure in the plasma plume, called plasma-induced pressure. During the expansion and condensation of the plasma plume, energy was transferred to the surrounding water.
Therefore, a thin layer of vapor appeared and evolved to a cavitation bubble around the plasma plume. This cavitation bubble produced an additional pressure over the plasma plume. It pressed the plasma back against the ablation spot. This is the so-called pressing process. Thanks to the large specific heat capacity of water, the plasma plume quenches quickly during the intermission of laser pulses. Since the quenching time was extremely short (about tens of picoseconds), the disordered distribution of Ni and O species in the plasma plume was reserved in the product released from the plasma plume. The morphology of the product was flake-like due to the pressing process. Finally, the amorphous nanoflakes aggregated due to the large surface areas of the nanoflakes. 32 It is believed that a water environment is necessary for the formation of amorphous phase by laser ablation on account of its large specific heat capacity for the rapid cooling process 1

Supplementary Note 5
An isotope labelling study is generally employed to determine the origin of the evolved gases in photocatalytic overall water splitting 4,5 . However, it may be not applicable in our case. As mentioned by Kandiel, H + is easier to be reduced than D + .
An isotope labelling will influence the reaction pathways to some extent 6 . Further, the rapid and frequent H + /D + isotopes exchange will take place in the whole photocatalytic process, making it difficult to determine the origin.

Supplementary Note 6
As shown in Supplementary Fig. 16a, in the absence of a photocatalyst, no signal could be detected (labelled as blank). When photocatalyst powders were added into the system, a four-line spectrum with the relative intensities of 1:1:1:1 formed, which is the characteristic spectrum for the BMPO-O 2 adducts 7 . Similarly, four characteristic peaks with the intensity ratios of 1:2:2:1 appeared when DMPO was used as the trapping agent, shown in Supplementary Fig. 16b. These peaks should be attributed to the DMPO-OH adducts 8 . On the contrary, no TEMP-1 O 2 adducts was detected (triplet signals), indicating the absence of 1 O 2 radicals in the process 9 . The generation of H 2 O 2 was determined by the coloration method, and the absorbance spectra are shown in Supplementary Fig. 16c 15,16,17 . Therefore, Route C may be less possible in our case.
The possible reaction route for the system containing AgNO 3 and photocatalyst (2DA or 2DPA) upon irradiation is illustrated in Supplementary Fig. 18

Supplementary Note 8
Ag + is a well-known electron scavenger that can consume the photo-generated electrons timely and retard the undesired carrier recombination 19 . At this point, ROS generation can proceed more smoothly. When the photocatalytic tests are conducted in pure water without Ag + , the consumption of electrons is weakened since H + ions are less easily reducible than Ag + ions 20 . Actually, electron consumption and hole consumption are interactional. Although some of the holes can be used for producing ROS (O 2 and OH), they cannot be consumed quickly due to the excess of electrons.
As shown in Supplementary Fig. 20, the signal to noise ratios of the ESR spectra are much smaller than those in Supplementary Fig. 16. The accumulation of electrons and holes in the photocatalysts will enhance the carrier recombination (②).

Supplementary Note 9
Generally, there are three pathways for the SPR-mediated photocatalysis based on the traditional antenna-reactor heterostructures: charge transfer, energy transfer and plasmonic heating 21 . In our case, 2DPA is a single-component photocatalyst with an incorporate antenna-reactor structure, which may obey a mechanism different from the charge transfer or energy transfer 3 . Plamonic heating is also negligible since the photocatalytic rate shows a linear correlation with the incident light intensity 22 .