Sustainable Doping via Molecular Adsorption on Thin-Film Semiconductor Bi2O2Se

Abstract

Doping in semiconductors is commonly achieved by incorporating foreign atoms into the crystal, which is a complex process that requires precise control. In two-dimensional or thin-film semiconductors, an alternative approach involving surface modification, such as molecular adsorption, has been proposed and attempted. However, successful doping via gas adsorption has not yet been reported. In this work, we present both first-principles calculations and experimental evidence demonstrating the feasibility of this approach in thin-film semiconductor Bi₂O₂Se, which has recently gained attention for its high carrier mobility, moderate band gap, and excellent air stability. We find that p-type doping can be achieved through surface adsorption of molecules such as NO₂, which exhibits stable chemisorption and significant charge transfer. This adsorption-induced p-doping effect significantly modulates the threshold voltage in as-grown n-type samples and remains stable for more than 10 days under ambient conditions, markedly improving electrostatic control and switching behavior in Bi₂O₂Se-based devices.

Keywords: chemisorption; molecular doping; thin-film semiconductor; threshold voltage modulation.

1

Energy levels of selected molecules compared with the density of states of films of pristine Bi₂O₂Se and defective Bi₂O₂Se with 4.1% V_Se. The first group of molecules (NO₂, SO₂, NO, Cl₂) has positive electron affinity, while the second group (NH₃, NF₃, CO₂, N₂) has negative electron affinity. The HOMO is marked with an asterisk (*), and the partially occupied state is indicated by a half-filled circle (◐). Peak heights reflect orbital degeneracy. The band gap of Bi₂O₂Se is highlighted in gray, and the Fermi level of the film with 4.1% internal Se vacancies is marked by a dashed line. All energy levels are with respect to the vacuum level.

2

Optimized geometric structures and difference charge density plots. (a) Pristine Bi₂O₂Se film (side view left; top view right) with 50% Se vacancies in the Se plane on the surfaces; (b) NO₂ adsorption in the first configuration; (c) NO₂ adsorption in the second configuration; (d) NH₃ adsorption. Shown in (b)–(d) are a side view (left) with the corresponding difference charge density plot and a top view (right). Charge accumulation and depletion are indicated in red and blue, respectively, at an isosurface level of 0.0025 e/Bohr³.

3

Charge transfer and adsorption energy for NO₂ adsorption at various surface coverages on the pristine Bi₂O₂Se film. Red dashed lines serve as visual guides to highlight the overall trends. The spread in data reflects variations arising from different intermolecular interactions in the 3 × 3 and 4 × 4 surface models. Charge transfer values are obtained from Bader charge analysis..

4

(a) Projected density of states (DOS) for 6% and 19% NO₂ adsorption on a defective Bi₂O₂Se film containing 4.1% V_Se. For comparison, the DOS of an isolated NO₂ molecule and the defective film with 4.1% internal Se vacancies are shown above. The HOMO of NO₂ is marked with an asterisk (*). Black dashed lines indicate the Fermi level in each system, and the band gap regions are shaded in gray. All energies are referenced to the vacuum level, and the surface dipole created by the charge transfer causes an energy shift in the adsorption system. Projected band structures are shown for (b) a defective Bi₂O₂Se film with 4.1% V_Se and the consequent adsorption systems with a 6% molecular coverage of (c) NO₂, (d) SO₂, and (e) NO, with blue lines denoting the Fermi level. Band energies are aligned by deep O 2s core levels with energy zero set at the Fermi level of the defective Bi₂O₂Se film as in (b).

5

(a) Raman spectrum of the as-grown Bi₂O₂Se showing the A_1g peak at 159 cm ^–1. (b) Cross-sectional STEM-HAADF of a Bi₂O₂Se nanoplate. (c) Optical image of the fabricated device. (d) Step-height profile measured with the atomic force microscopy (AFM) image, confirming a thickness of 15.5 nm. (e) Transfer characteristics of Bi₂O₂Se FETs under varying NO₂ concentrations, exhibiting a positive threshold voltage shift with increasing NO₂ partial pressures. The inset shows a logarithmic increase in current variation compared to that in the undoped sample, indicating the decreasing carrier concentration and saturating doping effect with higher NO₂ exposure. (f) Temporal evolution of the transfer characteristics of the 25 ppm of NO₂-doped device, demonstrating excellent stability after 14 days. (g) Transfer characteristics of Bi₂O₂Se FETs exposed to different NH ₃ concentrations, showing negligible threshold voltage shifts. The inset confirms minimal current variation across NH₃ concentrations. These electrical measurements are carried out under a drain voltage of 0.05 V.