Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.
Advertisement
Communications Materials volume 7, Article number: 180 (2026)
The most conventional atomic layer processing method, atomic layer deposition (ALD), delivers ultrathin blanket coatings with sub-nanometer thickness precision. Lateral confinement underpins the direct atomic layer processing (DALP®) family of deposition techniques. ALD chemistry applied to DALP® is a 3D printing method called atomic-layer additive manufacturing (ALAM). Here, we demonstrate the applicability of ALAM to the additive buildup of the crucial ZnS / Sb2S3 / V2O5 semiconductor stack which constitutes an functional inorganic solar cell. To this goal, ALAM processes are first optimized and evaluated for the individual materials vanadium(V) oxide, zinc sulfide, and antimony(III) sulfide. We establish the layer-by-layer growth mode controlled by self-limiting surface chemistry and characterize the materials’ structure and the smooth surface morphology of ALAM-coated areas. Finally, all three materials are 3D-printed in ALAM mode in combination with electrodes and the electron acceptor titania (TiO2) to form functional solar cells with a 120 nm thick Sb2S3 absorber layer. This novel fabrication of solar cells highlights the advantages of using direct patterning in the prototyping and optimization of photovoltaics in research and development.
Photovoltaics (PV) are a mature technology deployed at the terawatt scale. Yet the high energy and material consumption in production necessitates research to significantly reduce them. Beyond Si-based solar cells, which currently dominate the PV market with >100 µm thick devices1,2,3,4,5,6,7, second-generation solar cells enable thin film (<10 µm) architectures based on materials with higher absorption coefficients, such as CIGS or CdTe7,8,9,10,11,12. They allow for more flexible fabrication and applications13,14,15,16,17. However, their reliance on scarce and/or toxic materials limits their commercial breakthrough18,19,20. Therefore, third-generation solar cells emphasize sustainability as the fundamental metric rather than pure power conversion efficiency (PCE). For example, antimony(III) chalcogenides represent a promising alternative, due to their high absorption coefficients (corresponding to ideal thicknesses on the order of 0.1 µm), tunable band gap, high stability, and reasonable element abundance20,21,22,23,24,25,26. Efficiencies of up to 8% have been reported for single-junction solar cells27 using both solution-based27,28,29,30,31 and vacuum-based deposition methods. Recently, atomic layer deposition (ALD) has become increasingly established thanks to its precise control over film thickness32,33,34,35,36,37. This control becomes even more necessary the thinner each layer needs to be. However, ALD has traditionally been inefficient in precursor usage and, above all, time38,39,40,41,42,43. Furthermore, devices rely on patterned regions instead of blanket coatings, requiring subsequent etching or patterning processes, thereby increasing process complexity. These requirements associated with ALD-processed PV imply that prototyping and optimization tasks become intolerably time-consuming as large numbers of test cells must be manufactured with parameter variations and in statistically significant amounts.
Direct atomic layer processing (DALP®) enables rapid generation of patterned stacks, with the additive DALP® procedure atomic-layer additive manufacturing (ALAM) able to combine the layer-by-layer control of ALD with the flexibility of 3D printing44,45,46,47. A microfluidic gas nozzle delivers precursor gases in a laterally constrained manner (see Fig. 1a in our original publications)47,48, enabling localized deposition on a substrate carried by a motion stage. ALAM eliminates the need for large volumes of precursor vapors that is commensurate with traditional ALD chambers of macroscopic size. It minimizes the ALD cycle time and accelerates prototyping and optimization tasks by varying parameters and layer thicknesses on a single substrate.
Solar cell structure. Left, semiconductor stack based on Sb2S3 light absorber layer presented in terms of its architecture and processing. Right, sketch of its electronic band structure (band gaps quoted in eV)68,69,70,71,72,73.
Before ALAM can be applied to minimize material and energy consumption in ultrathin photovoltaics, two fundamental aspects must be established. First, ALAM processes must be developed for all semiconductors involved. Second, a geometry of the semiconductor stack and electrodes must be designed for ALAM printing. The following paragraphs establish both aspects for a PV system based on antimony sulfide (Sb2S3).
Our goal is not to deliver the world’s most efficient solar cell, and certainly not to prove economic viability of a production method. We introduce a novel prototyping method and demonstrate that a semiconductor stack can be built up in one single deposition tool—three distinct layers (ZnS, Sb2S3, V2O5) with one method and with the capability of generating several devices on one substrate. There is certainly ample room for later improvement in both the device performance and the process maturity.
The stack chosen for the present study relies on a thin film solar cell design (Fig. 1) that combines a crystalline ultrathin Sb2S3 intrinsic light absorber layer (variable, ~60–120 nm) with titania (TiO2, 50 nm) as the electron transport material and vanadium(V) oxide (V2O5, ~30 nm) as the hole transport layer49,50. The choice of V2O5 as the hole transport material differs from our previous studies in which spin-coated polythiophenes were used. It relies on the accessibility of V2O5 via atomic-layer processing in mild conditions, specifically, at a temperature below 200 °C and in the absence of an oxidizing agent such as ozone. An additional interfacial layer of zinc sulfide (ZnS, 1 nm) has been demonstrated to be crucial for both chemical and physical reasons21,32,51. This minimal stack design is well suited to demonstrate the first ALAM-printed PV device. The top gold contact is thermally evaporated through a mask, whereas commercially available transparent conducting indium tin oxide (ITO) substrates are used as the bottom contact with sputter-coated TiO2. Therefore, the ALAM processes for V2O5, ZnS, and Sb2S3 are established first.
We first investigate the surface chemistry of V2O5 ALAM on native Si(100) wafers for print and material quality. We choose the metal-organic precursor vanadyl triisopropoxide (OV(OiPr)3) to react with water at a substrate temperature of 150 °C. ALAM-patterned lines of 350 µm width appear with orange to blue hues depending on the number of passes performed (Fig. 2A). Lines featuring a distinct number of passes on specific segments allow for rapid determination of growth rates within a single sample using imaging ellipsometry and profilometry. The five individual thickness values presented in Fig. 2 have been deposited in less than 45 min. As seen in our previous work, ALAM retains the characteristic layer-by-layer growth of ALD44,45,46,47. Indeed, we find that V2O5 ALAM exhibits a linear growth rate of ~0.2 Å per pass (Fig. 2B), very similar to reports of conventional chamber-based ALD52,53. Scanning electron microscopy (SEM) and atomic force microscopy (AFM) analyses further show that the resulting V2O5 deposits are exceptionally smooth and homogenous with minimal to no impurities and a root-mean-square surface roughness of less than a nanometer. (Fig. 2D, E, G) Energy-dispersive X-ray microanalysis (EDS) mapping and grazing-incidence X-ray diffraction (GI-XRD) further demonstrate the existence of pure, crystalline shcherbinaite V2O5 before and after annealing at 250 °C for 2 min (Fig. 2C, F).
A Optical image of segmented line print of V2O5 using ALAM. B Averaged growth rates as determined by imaging ellipsometry (lavender) and profilometry (deep purple) using multiple segmented line prints show ALD-type growth. The error bars represent standard deviations derived from n = 9 individual thickness measurements on n’ = 3 distinct, nominally identical samples. C GI-XRD confirms the presence of crystalline shcherbinaite V2O5 (COD ID 9012221). D, E SEM, F EDS and G AFM show the presence of a smooth and homogenous V2O5 thin film with roughness similar to the Si(100) substrate.
For ZnS, we apply the metal-organic precursor bis(3-dimethylaminopropyl)zinc, Zn(DMP)2, with H2S as the reactant. This alternative to the classical diethylzinc54,55,56,57 features lower vapor pressure, lower reactivity, and non-pyrophoric behavior. It is therefore inherently safer with improved material economy58. Only the very sparing use of precursor vapor by ALAM renders this particular compound suitable for deposition46. The well-behaved chemistry enabled by Zn(DMP)2 allows for printing very defined 450-µm wide lines with a linear growth rate of 0.5 Å per pass, comparable to reported chamber-based ALD processes, indicating that the growth mode is controlled by surface chemistry (Fig. 3)56,57. SEM indicates the presence of smooth and homogenous ZnS thin films with additional agglomerates. AFM exhibits an increased roughness of ~2.6 nm and reveals smaller pinholes in a 50 nm thick film. Annealing the ZnS films at 300 °C under N2 for 2 min eliminates pinholes and improves the surface roughness to 1.8 nm though the agglomerates are still present.
A Optical image of a segmented line print of ZnS using ALAM. B Averaged growth rates determined by imaging ellipsometry (salmon pink) and profilometry (dark red) using multiple segmented line prints show ALD-type growth. The error bars represent standard deviations derived from n = 9 individual thickness measurements on n’ = 3 distinct, nominally identical samples. The as-printed thin film (C–E) shows a smooth base layer with discernable pinholes and agglomerates on top. F–H Upon annealing at 300 °C for 2 min the pinholes are eliminated with an improved surfaces roughness, however the thin film still features agglomerates on top.
EDS (Fig. 4A–D) characterizes a rather compositionally pure ZnS base layer, which is further supported by subsequent GI-XRD (Fig. 4F), but with high carbon and oxygen contents in the agglomerates (Fig. 4E). This is strongly indicative of unreacted Zn(DMP)2 precursor remnants, which are subsequently decomposed upon exposure to air. These agglomerates will inevitably have a negative impact on solar cell performance, and we assume that they can be avoided by running the ZnS process at a temperature higher than the current limitation of our ALAM prototype. This will involve additional thermal management between the sample heater and the motion stage.
A–D EDS mappings of the printed ZnS lines show a stoichiometric ZnS deposit with carbon and oxygen impurities. EDS quantifications (E) of the underlying ZnS film and the agglomerates confirm the purity of the ZnS thin film in contrast to the agglomerates on top of it, which are dominated by ZnO caused by precursor accumulation and decomposition. F GI-XRD measurements demonstrate the presence of crystalline ZnS (sphalerite, COD ID 9000107).
The Sb2S3 light absorber layer is printed using tris(dimethylamido)antimony(III), (Me2N)3Sb, and H2S. While in conventional chamber-based ALD (Me2N)3Sb is usually volatilized at elevated temperature, in ALAM mode it necessitates accurate cooling to –3 °C. In these conditions, ALAM yields accurate prints of thin Sb2S3 lines with slightly increased 500-µm width (Fig. 5A). Growth is linear with ALD-similar 0.7 Å per pass (Fig. 5B)33,36,59,60. When printed on Si substrates (with a native SiO2 layer) in the absence of a ZnS adhesion layer, thin Sb2S3 films exhibit discontinuities as deposited (Fig. 5C–F).
A Optical image of segmented line print of Sb2S3 using ALAM. B Averaged growth rates determined by imaging ellipsometry (pale moss) and profilometry (forest green) using multiple segmented line prints show ALD-type growth. The error bars represent standard deviations derived from n = 9 individual thickness measurements on n’ = 3 distinct, nominally identical samples. The as-printed thin films (C–F) exhibit visible pinholes.
Sb2S3 is deposited at 180 °C in its well-known orange, amorphous form. Crystallizing it to its electrically conductive, black stibnite structure involves annealing in conditions for which there is ample literature precedents. We have demonstrated in the past that annealing a-Sb2S3 as a continuous thin film on planar ZnS-coated substrates maintains the continuity and yields laterally large (micron-sized) crystals while maintaining the low roughness of the layer21,34,61,62. Upon annealing at 300 °C for 2 min, our ALAM-deposition Sb2S3 films (Fig. 6) exhibit complete dewetting on the native Si(100) wafer, as reported21. Nevertheless, EDS mapping and GI-XRD show that the solid consists of pure, crystalline Sb2S3.
A SEM measurements of the annealed Sb2S3 printed thin film shows strong dewetting on the Si(100) wafer. B GI-XRD measurements and C–F EDS mapping shows highly pure, crystalline Sb2S3 (stibnite, COD ID 9003460).
In contrast to this, depositing the same Sb2S3 onto a Si(100) wafer preliminarily coated with only 1 nm of ALAM-printed ZnS yields after annealing an exceptionally smooth (AFM roughness <1 nm), crystalline and closed thin film with discernible grain boundaries, as seen in SEM (Fig. 7A–D). This observation is fully parallel to our earlier studies using conventional chamber-based ALD21,51. Of course, the underlying ZnO-rich agglomerates are still present underneath the pristine Sb2S3 thin film.
SEM (A–C) and AFM (D) measurements of the annealed Sb2S3 thin film with a ~1 nm thin ZnS adhesion layer (on Si(100) substrate) reveal a smooth and homogenous surface morphology. E–H EDS mapping confirms the presence of ZnO-rich agglomerates underneath the pristine Sb2S3 thin film already observed in Fig. 4.
Each material of the solar cell stack printed by ALAM has sufficient thin film quality and control to allow fabrication of miniaturized functional devices that require careful consideration of the lateral geometry for the electrical contacts and the semiconductors. Inadequate fitting of the metallic contacts to the diminutive ALAM feature sizes would generate short-circuits.
Before attempting to generate large numbers of cells on one substrate for rapid prototyping, we establish a proof of principle design of the contacts that places six individual cells on one substrate (Fig. 8). We define an active window size of 0.5 mm × 3 mm. The gold connection bridge leading to the measurement pad is positioned over a preliminarily etched non-conductive glass sideline (gray surface in Fig. 8), with the solar active materials slightly overlapping. This approach necessitates sufficient accuracy in the alignment of ALAM prints and subsequent gold evaporation.
This design allows for the fabrication of six solar cells with up to six different configurations. It minimizes the risk of short-circuiting between the top gold contacts and the conductive ITO bottom electrode by judicious placement of a chemically etched ITO-free surface on either side of the substrate (gray).
The geometric precision achieved by ALAM printing must be matched during the subsequent gold contact evaporation via conventional shadow masking. Figure 9 demonstrates that laser-cut PET foil masks deliver sufficiently tight tolerances (±0.2 mm) to address each cell without causing a short-circuit. The uncertainty is caused by bending of the PET foil, which introduces a shadow effect of slightly variable magnitude (±0.1 mm on each side). Despite this limitation, EDS mapping demonstrates that the edges of all relevant fields are sufficiently sharp for solar cell fabrication.
A, B SEM micrographs of the gold contacts defined by evaporation through a mask. They are positioned near the edge of a glass substrate bearing a surface on which the ITO layer has been removed by wet chemical etching. A and B present two nominally identical but distinct individual contacts and allow for a quantification of the positioning accuracy and tolerance in the contact dimensions. Differences are due to shadowing caused by an imperfectly adhering mask. C EDS mapping highlights the precision achieved for placing the connection bridge and active window sections of the gold contact as well as the edge of the ITO etching.
All ingredients prepared so far can now be combined to prepare a fully functional device. The ITO-coated glass slides are first etched locally as described in the previous section, then sputter-coated with 50 nm TiO2 and annealed (250 °C, 30 min, in air). ALAM is used to print six individual solar cells consisting of ZnS (30 passes, 1 nm), Sb2S3 (800/1200/1600 passes, 60/90/120 nm), and V2O5 (1600 passes, 34 nm). Annealing is performed after both the Sb2S3 deposition (300 °C, 2 min, under N2) and V2O5 (250 °C, 2 min, under N2). The ALAM-coated areas consist of a rectangle of ~2 mm × 19 mm dimensions obtained by printing eleven adjacent lines spaced by 0.2 mm nominally. Each such area contains three segments of distinct Sb2S3 layer thickness, demonstrating the prototyping capability of the approach—here, we explore Sb2S3 layers of 60 nm, 90 nm, and 120 nm. We chose the light absorber layer thickness to showcase the prototyping ability of our technology, given that it is the crucial parameter exploited experimentally to adjust the balance between the contradictory requirements of maximized light absorption and minimized loss of charge carriers by recombination on their way to the carrier selective materials21. Gold contacts are finally evaporated through the mask described above.
Optical images as well as SEM micrographs (Fig. 10A, B) evidence how the appearance of the structures reflects our design. SEM micrographs further reveal the surface morphology of the ALAM deposits: the fine roughness of the V2O5 layer (Fig. 10C) and the crystalline grain boundaries of the underlying Sb2S3 thin film (Fig. 10D, E).
A Optical image of three solar cells fabricated by ALAM on one substrate. The precisely ALAM-printed features and evaporated gold contacts are visible. SEM micrographs (B–E) confirm the overall stack geometry (B) and, within it, the presence of morphological features observed previously for individual materials: homogeneous granularity of the top V2O5 layer (imaged with 3 kV acceleration voltage, C), crystalline Sb2S3/ZnS films with grain boundaries imaged underneath V2O5 (imaged under 15 kV, D) and in its absence (3 kV, E).
JV measurements (Fig. 11) establish the solar cell functionality and quantify performance. Five of the 18 stack units fabricated exhibit a rectifying behavior, and three of them clear solar cell characteristics with photocurrent and photovoltage under illumination. The occurrence of short-circuits is expected on the basis of the ZnO particles observed after ZnS deposition: they protrude vertically and increase the risk of direct contact between ZnO and V2O5. Here, the thickest Sb2S3 layer (120 nm) is needed to mitigate partly the effect of the ZnO particles by shutting down at least some of the short-circuits limiting VOC, with VOC ranging between 20 and 140 mV and JSC in the μA·cm–2 range. The very low current densities achieved here are related to the low fill factors, which are themselves due to the long lateral transport distances. In combination with the small active surface area and the low absolute values of the currents generated, the Ohmic losses are therefore significant. Additionally, the physical alignments of individual layers within the stack may not be perfect. Finally, the 120 nm thick Sb2S3 layer is necessary for avoiding short-circuits but thicker than what we have found to be optimal in a previous study21. Accordingly, the champion solar cell shows a PCE of only 0.03%. We note that this number does not fare so poorly when compared to the only example available so far in the openly accessible literature of what has been described as a “nearly all-ALD CIGS-type solar cell” (JSC ≈ 500 μA cm–2, VOC ≈ 80 mV, FF ≈ 0.5, PCE ≈ 0.02%)63.
A–C JV measurements of three functional 3D-printed solar cells with 120 nm Sb2S3 layer thickness. All curves show the resemblance of diodes in dark conditions (royal blue) and under illumination (azure blue). These solar cells exhibit positive current in short-circuit in the dark as well as negative current in reverse bias conditions, indicating a low shunt resistance and the presence of (capacitive) leakage currents due to detrimental contacts. The JV curves visibly change under illumination (azure blue) and yield measurable photocurrents.
These numbers are by no means on par with state-of-the-art photovoltaics reliant on Sb chalcogenides, and both performance and reliability must, in the future, be drastically improved. Most crucially, the ZnO particles must be eliminated using more elevated ZnS deposition temperatures, in order to avoid short-circuits. This simple proof-of-concept fabrication represents the first ever demonstration of solar cells 3D-printed with nanometer precision. We envision that this novel processing method will enable future optimization of photovoltaics that is not only faster but also significantly more efficient in both material usage and energy input. Further topics of relevance to be investigated in the future may pertain to the possibility that the imperfections of ALD chemistry yield organic residues in the films. Their presence strongly depends on the precursors chosen and the processing temperature64,65,66, and it would of course have a significant effect on the semiconductor properties.
This study demonstrated that ALAM/DALP® can be successfully employed to deposit thin films of Sb2S3, ZnS, and V2O5 in a 3D-printed manner with quality and layer thickness precision of conventional chamber-based ALD.
V2O5 ALAM relies on the metal-organic precursor (iPrO)3VO combined with water and resulting in ~350 μm-wide, continuous printed lines with a smooth, homogenous surface morphology, high compositional purity and crystalline structure after annealing.
ZnS printing is achieved by using the metal-organic precursor Zn(DMP)2 with H2S, yielding crystalline, smooth, and homogenous ZnS base layers on ~450 μm-wide defined lines. Additional ZnO agglomerates are caused by the limited substrate temperature achievable in the current prototype.
Sb2S3 thin lines are obtained from (Me2N)3Sb reacting with H2S. The as-printed ~500 μm-wide lines show high compositional purity but exhibit strong island growth and pinhole formation on metal oxide substrates, with additional dewetting of the thin films after annealing on Si(100) wafers. Introducing a ~1 nm thin ZnS adhesion layer beforehand prevents these undesirable effects. Annealed ALAM-printed ZnS / Sb2S3 structures exhibit a high crystallinity with continuous, smooth, and homogenous morphology. A limitation of the currently possible ZnS process is related to the temperatures achievable and the amount of residual water vapor present in the system. The presence of H2O has a significant, deleterious effect on the film morphology, adding large nodules to an otherwise smooth and continuous film. These structural imperfections lead to short-circuits which in the specific conditions used here have prevented us from building cells with optimal Sb2S3 layer thickness. This limitation, however, is a rather trivial one which will be addressed in a future overhaul of the ALAM tool.
A similar investigation could be performed by classical ALD. Exploiting ALAM instead offers several distinct advantages. Firstly, the amount of precursor consumed for this study is reduced by at least a factor 100046. Secondly, the rapidity with which a process can be established (GPC determination) is incommensurable. Third, optimization of devices can be carried out in one deposition run on one substrate—a feat that is not accessible using classical blanket deposition techniques available so far.
The direct patterning capability of ALAM enables the integration of these individual semiconductors and the design of flexibly configurable small-scale arrays of functional solar cells on individual ITO/glass substrates. This study delivers a glimpse of the high throughput and prototyping capabilities of ALAM (and more generally, DALP®). With the early tool prototype used for this study, the total 3000 or so passes required for depositing all three semiconductors of one cell (corresponding to about 12,000 individual line passes for the whole area) are performed within 1 h. Aligning the lines and raster patterns used for the distinct layers of the semiconductor stack has been manual on the prototype tool. It will be upgraded to an automatized system in the near future, enabling sub-micron positioning reproducibility. The recent commercialization of a DALP® reactor will ensure further time efficiency gains and improved reliability, thereby facilitating prompt acceptance of this technology.
The following materials were used: vanadium(V) oxytriisopropoxide (OV(OCH(CH3)2)3, 98%, abcr), bis(3-dimethylaminopropyl)zinc (Zn(DMP)2, synthesized)67, tris(dimethylamido)antimony(III) ([(H3C)2N]3Sb), ≥99%, Sigma Aldrich and abcr), water (Milli-Q®), hydrogen sulfide (H2S, 3% in N2, Air Liquide), Ar (N5.0, Air Liquide), N2 (N5.0, Linde Germany), Si(100) wafers (native oxide, Silicon Materials Inc.), ITO-coated glass substrates (≤10 Ω/sq., Techinstro), Hellmanex III (2% in Milli-Q® water, Hellma Analytics), isopropanol (≥98% technical, VWR), acetone (≥99% technical, VWR), hydrochloric acid (HCl, 2 M, lab grade, thermo fisher), DALP® nozzles (Atlant3D), Au (99.99% pellets, m&k GmbH), TiO2 target (99.99%), Si reference solar cell (Newport).
ALAM depositions were carried out using a DALP® reactor prototype under atmospheric pressure of constantly purged N2 gas. Precursors and reactants were stored in separate bubblers connected to the DALP® system. For vanadium(V) oxide ALAM, vanadium(V) oxytriisopropoxide was maintained at 50 °C with Ar serving as the carrier gas at a flow rate of 12 sccm. H2O was used as a reactant at room temperature (~23 °C) with N2 as the carrier gas at a flow rate of 30 sccm. Precursor and reactant vapors were transported through separate lines heated at 70 °C. The DALP® nozzle system was maintained at 90 °C with an additional Ar flow of 600 sccm to ensure separation of the precursor/reactant channels. The substrate temperature was fixed at 150 °C. For zinc(II) sulfide ALAM, bis(3-dimethylaminopropyl)zinc was maintained at 65 °C with Ar as carrier gas at a flow rate of 12 sccm. H2S was used as a reactant at room temperature (~23 °C) at a flow rate of 30 sccm. Precursor and reactant vapors were transported through separate lines heated at 90 °C. The ALAM | DALP® nozzle system was maintained at 110 °C with an additional Ar flow of 600 sccm to ensure separation of the precursor/reactant channels. The substrate temperature was fixed at 180 °C. For antimony(III) sulfide ALAM, tris(dimethylamido)antimony(III) was maintained at –3 °C with Ar as carrier gas at a flow rate of 12 sccm. H2S was used as a reactant at room temperature (~23 °C) at a flow rate of 30 sccm. Precursor and reactant vapors were transported through separate lines heated at 80 °C. The ALAM nozzle system was maintained at 130 °C with an additional Ar flow of 600 sccm to ensure separation of the precursor/reactant channels. The substrate temperature was fixed at 80 °C. Direct patterning was achieved by moving the substrate with a motorized stage at a speed of 200 mm/s and acceleration of 2000 mm/s2. The substrate was positioned visually as close as possible without contact, and the distance was kept constant throughout all prints.
ITO glass substrates were completely covered with Kapton® tape and subsequently cut using a laser cutter to obtain the ITO etching regions. ITO etching was performed with Zn powder/2 M HCl for 15 min. The Kapton® tape was removed and the ITO glass substrate was subsequently cleaned by ultrasonication in 2% Hellmanex III solution, Isopropanol, Acetone, and Milli-Q® water. As a final step, the substrates were treated with UV-ozone for 30 mins.
Subsequently, ~50 nm TiO2 from a TiO2 target (99.99%) was sputtered in RF mode on the cleaned ITO / glass substrates at a power density of 2.5 W cm–2 and Ar flow of 5 sccm with a base pressure of 1.3 × 10–4 Pa and working pressure of 4.3 Pa in a Torr International Inc. CRC 622 sputter coater. The TiO2 sputtered ITO glass substrates were subsequently annealed at 250 °C for 30 min in air.
After ZnS and Sb2S3 ALAM, the samples were annealed at 300 °C for 2 min under N2 atmosphere. After V2O5 ALAM printing, they were annealed again at 250 °C for 2 min under N2. As a last step, 80 nm thick gold back contacts were thermally evaporated using a Covap PVD instrument by Ångstrom Engineering Inc. from Au pellets (99.99%). A custom-built plastic foil mask was prepared via laser cutting, defining the active window size to ~0.5 × 3 mm2.
Imaging ellipsometry measurements were performed using an Accurion EP-4 Ellipsometer with an Omicron LED Hub providing six distinct wavelengths (384.9 nm, 447.2 nm, 522.7 nm, 595.4 nm, 657.7 nm and 840.6 nm). Angle of incidence and view were kept at 60.08°. All measurements were conducted with EP4Control 23.4.11.1 software from Accurion. Each ALAM print was scanned at all six wavelengths through automatic consecutive scans following a pre-programmed scanning pattern, with full automation of focusing and alignment at each step of the scan. The measured scans were stitched and analyzed with DataStudio 23.4.5.2 software from Accurion. Model calculations were performed using the EP4Model 23.4.4 software from Accurion. A self-calibrated Tauc-Lorentz function was used for all prints as mathematical models. To derive reliable mathematical models and obtain layer thicknesses, optical data were extracted after initial scanning with DataStudio 23.4.5.2 software. Separate datasets were generated: one for the optical properties of the underlying Si wafer (used for calibration) and others for each ALAM print segment. Clustering was performed to reduce computation time.
Optical microscopy was performed using the Imaging Ellipsometer’s built-in camera with 5-times magnification at a wavelength of 447.2 nm and image scanning/stitching.
Scanning electron microscopy was performed using a Jeol JSM-F100 at an acceleration voltage of 3 kV and a working distance of 10 mm with secondary electron detector. Energy-dispersive X-ray spectroscopy (EDS) was carried out on the same instrument with an acceleration voltage between 5 kV and 15 kV. EDS spectra or maps were analyzed with SEM Center software from Jeol.
Grazing-incidence X-ray diffraction was performed on a Bruker D8 Advance diffractometer employing Cu Kα radiation (λ = 0.15406 nm) with a slit width of 0.1 mm under an incident angle of 0.5°. The 2θ measurements were conducted with 0.02° step size and 25 s measurement duration per step. The Crystallographic Open Database (COD) was used to assign expected diffractions.
Atomic force microscopy was performed on a Bruker NanoMan VS AFM in tapping mode using an OTESPA silicon tip with a spring constant of 26 N/m. Free amplitude was 1000 mV with a setpoint of 60–80%. Surface morphology was measured over a 1 × 1 µm2 area at a resolution of 1024 × 1024 samples per line, a scan rate of 0.5 Hz and tip velocity of 3 µm/s. Data and image processing was performed with Gwyddion 2.69 software.
Profilometry was performed using a Bruker Dektak XT Profilometer with a 20 s/600 µm scan and stylus force (sic) of 3 mg. Data was analyzed with Vision64 software from Bruker. Profile height was determined by plotting the height profile with Origin2025 from OriginLabs.
Current density – voltage (JV) measurements were performed under a solar simulator employing a Xenon lamp calibrated to 1.5 AM with 100 mW cm–2 illumination using a Si reference cell. Solar cells were connected using a Gamry Reference 600 potentiostat. Measurement sweep range was –1.1 V to +1.1 V with a step of 5 mV and scanning speed of 30 mV s–1. Datasets were plotted with Origin2025 from OriginLabs.
The data (raw and analyzed) presented in the figures is available on the open access platform Zenodo under the https://doi.org/10.5281/zenodo.19473184.
Benda, V. & Černá, L. PV cells and modules – state of the art, limits and trends. Heliyon 6, e05666 (2020).
Peng, J. Q., Lu, L. & Yang, H. X. Review on life cycle assessment of energy payback and greenhouse gas emission of solar photovoltaic systems. Renew. Sustain. Energy Rev. 19, 255–274 (2013).
Article Google Scholar
Shockley, W. & Queisser, H. J. Detailed balance limit of efficiency of p-n junction solar cells. J. Appl. Phys. 32, 510–519 (1961).
Article CAS Google Scholar
Leccisi, E., Raugei, M. & Fthenakis, V. The energy and environmental performance of ground-mounted photovoltaic systems-a timely update. Energies 9, 622 (2016).
Article Google Scholar
Görig, M. & Breyer, C. Energy learning curves of PV systems. Environ. Prog. Sustain. Energy 35, 914–923 (2016).
Article Google Scholar
Ludin, N. A. et al. Prospects of life cycle assessment of renewable energy from solar photovoltaic technologies: a review. Renew. Sustain. Energy Rev. 96, 11–28 (2018).
Article Google Scholar
Razykov, T. M. et al. Solar photovoltaic electricity: current status and future prospects. Sol. Energy 85, 1580–1608 (2011).
Article CAS Google Scholar
Chirila, A. et al. Potassium-induced surface modification of Cu(In,Ga)Se2 thin films for high-efficiency solar cells. Nat. Mater. 12, 1107–1111 (2013).
Article CAS PubMed Google Scholar
Rahman, M. F. et al. A qualitative design and optimization of CIGS-based solar cells with Sn2S3 back surface field: a plan for achieving 21.83 % efficiency. Heliyon 9, e22866 (2023).
Article CAS PubMed PubMed Central Google Scholar
Kamada, R. et al. New world record Cu(In, Ga)(Se, S)2 thin film solar cell efficiency beyond 22%. In 2016 IEEE 43rd Photovoltaic Specialists Conference (PVSC), 1287–1291 (IEEE, 2016).
Birkmire, R. W. & McCandless, B. E. CdTe thin film technology: leading thin film PV into the future. Curr. Opin. Solid State Mater. Sci. 14, 139–142 (2010).
Article CAS Google Scholar
Zhao, X. G. et al. Cu-in halide perovskite solar absorbers. J. Am. Chem. Soc. 139, 6718–6725 (2017).
Article CAS PubMed Google Scholar
Wen, X. X. et al. Epitaxial CdTe thin films on mica by vapor transport deposition for flexible solar cells. ACS Appl. Energy Mater. 3, 4589–4599 (2020).
Article CAS Google Scholar
Chun, S., Han, K. S., Shin, J. H., Lee, H. & Kim, D. Fabrication and characterization of CdTe nano pattern on flexible substrates by nano imprinting and electrodeposition. Microelectron. Eng. 87, 2097–2102 (2010).
Article CAS Google Scholar
Gerthoffer, A. et al. CIGS solar cells on flexible ultra-thin glass substrates: characterization and bending test. Thin Solid Films 592, 99–104 (2015).
Article CAS Google Scholar
Moriwaki, K. et al. Monolithically integrated flexible Cu(In,Ga)Se2 solar cells and submodules using newly developed structure metal foil substrate with a dielectric layer. Sol. Energy Mater. Sol. Cells 112, 106–111 (2013).
Article CAS Google Scholar
Zheng, J. et al. Efficient flexible monolithic perovskite-CIGS tandem solar cell on conductive steel substrate. ACS Energy Lett. 9, 1545–1547 (2024).
Article CAS Google Scholar
Candelise, C., Winskel, M. & Gross, R. Implications for CdTe and CIGS technologies production costs of indium and tellurium scarcity. Prog. Photovolt. 20, 816–831 (2012).
Article Google Scholar
Tao, C. S., Jiang, J. C. & Tao, M. Natural resource limitations to terawatt-scale solar cells. Sol. Energy Mater. Sol. Cells 95, 3176–3180 (2011).
Article CAS Google Scholar
Amatya, R. & Ram, R. J. Trend for thermoelectric materials and their earth abundance. J. Electron. Mater. 41, 1011–1019 (2011).
Article Google Scholar
Büttner, P. et al. Adjusting interfacial chemistry and electronic properties of photovoltaics based on a highly pure Sb2S3 absorber by atomic layer deposition. ACS Appl. Energy Mater. 2, 8747–8756 (2019).
Article PubMed PubMed Central Google Scholar
Radzwan, A., et al. First-principles calculations of structural, electronic, and optical properties for Ni-doped Sb2S3. Comput. Condensed Matter. 24, e00477 (2020)
Nie, R., Lee, K. S., Hu, M. & Seok, S. I. Strain tuning via larger cation and anion codoping for efficient and stable antimony-based solar cells. Adv. Sci. 8, 2002391 (2020).
Article Google Scholar
Li, J., Liu, X. & Yao, J. The enhanced photovoltaic performance of Sb2S3 solar cells by thermal decomposition of antimony ethyl xanthate with thiourea doping. Energy Technol. 8, 1900841 (2019)
Lei, H. W. et al. Copper doping of Sb2S3: fabrication, properties, and photovoltaic application. J. Mater. Sci. Mater. Electron. 30, 21106–21116 (2019).
Article CAS Google Scholar
Kern, S. et al. Monolithic two-terminal tandem solar cells using Sb2S3 and solution-processed PbS quantum dots achieving an open-circuit potential beyond 1.1 V. ACS Appl. Mater. Interfaces 16, 13903–13913 (2024).
Article CAS PubMed Google Scholar
Adhikari, A. et al. Process optimization and light soaking to enhance photovoltaic performance of antimony sulfide solar cells. ACS Appl. Energy Mater. 8, 6280–6289 (2025).
Article CAS Google Scholar
Zhu, L. et al. Solution-processed compact Sb2S3 thin films by a facile one-step deposition method for efficient solar cells. Solar RRL 5, 2100666 (2021)
Han, J. et al. Solution-processed Sb2S3 planar thin film solar cells with a conversion efficiency of 6.9% at an open circuit voltage of 0.7 V achieved via surface passivation by a SbCl3 interface layer. ACS Appl. Mater. Interfaces 12, 4970–4979 (2020).
Article CAS PubMed Google Scholar
Wang, X. et al. A fast chemical approach towards Sb2S3 film with a large grain size for high-performance planar heterojunction solar cells. Nanoscale 9, 3386–3390 (2017).
Article CAS PubMed Google Scholar
Choi, Y. C. & Seok, S. I. Efficient Sb2S3-sensitized solar cells via single-step deposition of Sb2S3 using S/Sb-ratio-controlled SbCl3-thiourea complex solution. Adv. Funct. Mater. 25, 2892–2898 (2015).
Article CAS Google Scholar
Pointer, C. et al. Elucidating mechanistic details of photo-induced charge transfer in antimony sulfide-based p-i-n junctions. J. Phys. Chem. C. 125, 18429–18437 (2021).
Article CAS Google Scholar
Büttner, P. et al. Continuous, crystalline Sb2S3 ultrathin light absorber coatings in solar cells based on photonic concentric p-i-n heterojunctions Nano Energy 103, 107820 (2022)
Wu, M. et al. Correlative and in situ microscopy investigation of phase transformation, crystal growth, and degradation of antimony sulfide thin films. ACS Nano 19, 25017–25027 (2025).
Article CAS PubMed Google Scholar
Lee, S.-J. et al. Approach to transparent photovoltaics based on wide band Gap Sb2S3 absorber layers and optics-based device optimization. ACS Appl. Energy Mater. 3, 12644–12651 (2020).
Article CAS Google Scholar
Kim, D. H. et al. Highly reproducible planar Sb2S3-sensitized solar cells based on atomic layer deposition. Nanoscale 6, 14549–14554 (2014).
Article CAS PubMed Google Scholar
Büttner, P. et al. Solid state interdigitated Sb2S3 based TiO2 nanotube solar cells. RSC Adv. 10, 28225–28231 (2020).
Article PubMed PubMed Central Google Scholar
Yuan, C. Y. & Dornfeld, D. A. Integrated sustainability analysis of atomic layer deposition for microelectronics manufacturing J. Manuf. Sci. Eng. Trans. ASME, 132, 030918 (2010)
Pan, D. Q., Guan, D. S., Jen, T. C. & Yuan, C. Atomic layer deposition process modeling and experimental investigation for sustainable manufacturing of nano thin films. J. Manuf. Sci. Eng. Trans. ASME 138, 101010 (2016).
Yasmeen, S., Ryu, S. W., Lee, S. H. & Lee, H. B. R. Atomic layer deposition beyond thin film deposition technology. Adv. Mater. Technol. 8, 2200876 (2022).
Vervuurt, R. H., Sharma, A., Jiao, Y., Kessels, W. E. & Bol, A. A. Area-selective atomic layer deposition of platinum using photosensitive polyimide. Nanotechnology 27, 405302 (2016).
Article PubMed Google Scholar
Miikkulainen, V., Vehkamäki, M., Mizohata, K., Hatanpää, T. & Ritala, M. Highly material selective and self-aligned photo-assisted atomic layer deposition of copper on oxide materials. Adv. Mater. Interfaces 8, 2100014 (2021).
Chalker, P. R. Photochemical atomic layer deposition and etching. Surf. Coat. Technol. 291, 258–263 (2016).
Article CAS Google Scholar
Galván Dominguez, Y. A. et al. Atomic-layer additive manufacturing of indium(III) sulfide. Chem. Afr. 9, 43 (2026).
Article Google Scholar
Kürten, S. et al. Direct-patterning SnO2 deposition by atomic-layer additive manufacturing. Mater. Adv. 6, 3998–4002 (2025).
Article Google Scholar
Stefanovic, S. et al. Direct-patterning ZnO deposition by atomic-layer additive manufacturing using a safe and economical precursor. Small 19, e2301774 (2023).
Article PubMed Google Scholar
Kundrata, I. et al. Additive manufacturing in atomic layer processing mode. Small Methods 6, e2101546 (2022).
Article PubMed Google Scholar
Direct Link to Figure 1a of our paper “Additive Manufacturing in Atomic Layer Processing Mode” https://onlinelibrary.wiley.com/cms/asset/ec738bea-bf99-4c5b-9ea6-689f28810294/smtd202101546-fig-0001-m.png (2022).
Zhang, L. et al. V2O5 as hole transporting material for efficient all inorganic Sb2S3 solar cells. ACS Appl. Mater. Interfaces 10, 27098–27105 (2018).
Article CAS PubMed Google Scholar
Yan, W. B. et al. Hole-transporting materials in inverted planar perovskite solar cells. Adv. Energy Mater. 6, 1600474 (2016).
Article Google Scholar
Büttner, P. et al. ZnS ultrathin interfacial layers for optimizing carrier management in Sb2S3-based photovoltaics. ACS Appl. Mater. Interfaces 13, 11861–11868 (2021).
Article PubMed PubMed Central Google Scholar
Musschoot, J. et al. Comparison of thermal and plasma-enhanced ALD/CVD of vanadium pentoxide. J. Electrochem. Soc. 156, P122–P126 (2009).
Article CAS Google Scholar
Badot, J. C. et al. Atomic layer epitaxy of vanadium oxide thin films and electrochemical behavior in presence of lithium ions. Electrochem. Solid St 3, 485–488 (2000).
Article CAS Google Scholar
Sun, C. H. et al. ZnS thin films grown by atomic layer deposition on GaAs and HgCdTe substrates at very low temperature. Infrared Phys. Technol. 85, 280–286 (2017).
Article CAS Google Scholar
Bisengaliev, R. A. et al. Molecular layering of 2D films and superlattices based on II-VI compounds. Phys. Solid State 40, 754–755 (1998).
Article Google Scholar
Basit, M. A. et al. Enhanced PbS quantum dot loading on TiO2 photoanode using atomic-layer-deposited ZnS interfacial layer for quantum dot-sensitized solar cells. Mater. Chem. Phys. 220, 293–298 (2018).
Article CAS Google Scholar
Stuyven, G., De Visschere, P., Hikavyy, A. & Neyts, K. Atomic layer deposition of ZnS thin films based on diethyl zinc and hydrogen sulfide. J. Cryst. Growth 234, 690–698 (2002).
Article CAS Google Scholar
Diethylzinc fire. @ctfm-lab.bsky.social on Bluesky, https://bsky.app/profile/ctfm-lab.bsky.social/post/3ly6f3lnaic2o (2025).
Yang, R. B. et al. Atomic layer deposition of antimony oxide and antimony sulfide. Chem. Mater. 21, 2586–2588 (2009).
Article CAS Google Scholar
Wu, Y. L. et al. Antimony sulfide as a light absorber in highly ordered, coaxial nanocylindrical arrays: preparation and integration into a photovoltaic device. J. Mater. Chem. A 3, 5971–5981 (2015).
Article CAS Google Scholar
Wedemeyer, H. et al. Nanocrystalline solar cells with an antimony sulfide solid absorber by atomic layer deposition. Energy Environ. Sci. 6, 67–71 (2013).
Article CAS Google Scholar
Mahuli, N., Halder, D., Paul, A. & Sarkar, S. K. Atomic layer deposition of amorphous antimony sulfide (a-Sb2S3) as semiconductor sensitizer in extremely thin absorber solar cell. J. Vac. Sci. Technol. A 38, 032407 (2020).
Article CAS Google Scholar
Le Tulzo, H., Schneider, N., Lincot, D. & Donsanti, F. Toward an all-atomic layer deposition (ALD) process for Cu(In,Ga)(S,Se)2(CIGS)-type solar cell. Solar Energy Mater. Solar Cells, 200, 109965 (2019)
Dufond, M. et al. Quantifying the extent of ligand incorporation and the effect on properties of TiO2 thin films grown by atomic layer deposition using an alkoxide or an alkylamide. Chem. Mater. 32, 1393–1407 (2020).
Article CAS Google Scholar
Elam, J. W. et al. Atomic layer deposition of tin oxide films using tetrakis(dimethylamino) tin. J. Vac. Sci. Technol. A 26, 244–252 (2008).
Article CAS Google Scholar
Cho M., et al. Effects of carbon residue in atomic layer deposited HfO2 films on their time-dependent dielectric breakdown reliability Appl. Phys. Lett. 90, 182907 (2007).
Zanders, D. et al. Cobalt metal ALD: understanding the mechanism and role of zinc alkyl precursors as reductants for low-resistivity co thin films. Chem. Mater. 33, 5045–5057 (2021).
Article CAS Google Scholar
Liu, X. et al. Grain engineering of Sb2S3 thin films to enable efficient planar solar cells with high open-circuit voltage. Adv. Mater. 36, e2305841 (2024).
Article PubMed Google Scholar
Gassenbauer, Y. et al. Surface states, surface potentials, and segregation at surfaces of tin-doped In2O3. Phys. Rev. B 73, 245312 (2006).
Article Google Scholar
Chang, Y., Yates, J. R. & Patrick, C. E. First-principles band alignments at the Si:anatase TiO2 interface. ACS Omega 8, 20138–20147 (2023).
Article CAS PubMed PubMed Central Google Scholar
Huang, Y. H. et al. First-principles calculation of the band structure, electronic states, and optical properties of Cr-doped ZnS double-wall nanotubes. Comput. Mater. Sci. 101, 1–7 (2015).
Article CAS Google Scholar
Lee, S., Wang, Q. & Chakrapani, V. Electronic vs optical bandgap of stoichiometric V2O5: effects of surface transfer doping and electron accumulation layer. J. Appl. Phys. 136, 205101 (2024).
Article CAS Google Scholar
Michaelson, H. B. The work function of the elements and its periodicity. J. Appl. Phys. 48, 4729–4733 (1977).
Article CAS Google Scholar
Download references
We thank Profs. Marcus Halik and Dirk Guldi for graciously providing access to AFM and profilometry, respectively.
This work was financially supported by the European Research Council (ERC) under the ERC Proof-of-Concept Grant ‘ALAMS – Atomic-layer additive manufacturing for solar cells’ (grant number 101069310), by the German Ministry of Economic Affairs and Energy under the project ‘KernKat’, (FKZ: 03EI3104), and by the Deutsche Forschungsgemeinschaft (German Research Foundation, DFG) within the Collaborative Research Center ‘ChemPrint’ (project 538767711, CRC 1719), the Research Training Group ‘CorMic’ (project 537140136, GRK 3103), and the project REACTIVE (490773082). Selina Kern acknowledges financial support from the Fonds der Chemischen Industrie (Kekulé Fellowship). Open Access funding enabled and organized by Projekt DEAL.
Department of Chemistry and Pharmacy, Section Materials Chemistry, Chair Chemistry of Thin Film Materials, IZNF, Friedrich-Alexander-Universität Erlangen-Nürnberg, Erlangen, Germany
Micah Mc Naire, Sanja Pannen, Jonas Englhard, Pei-Chun Liao, Selina Kern, Ryan W. Crisp & Julien Bachmann
Inorganic Materials Chemistry, Ruhr University Bochum, Bochum, Germany
David Zanders & Anjana Devi
Leibniz Institute for Solid State and Materials Research, Dresden, Germany
Anjana Devi
Chair of Materials Chemistry, Faculty of Chemistry and Food Chemistry, TU Dresden, Dresden, Germany
Anjana Devi
ATLANT 3D Nanosystems ApS, Taastrup, Denmark
Julien Bachmann
PubMed Google Scholar
PubMed Google Scholar
PubMed Google Scholar
PubMed Google Scholar
PubMed Google Scholar
PubMed Google Scholar
PubMed Google Scholar
PubMed Google Scholar
PubMed Google Scholar
Micah Mc Naire: Investigation, data curation, writing (original draft), writing (review & editing). Sanja Pannen: Investigation. Jonas Englhard: Investigation. Pei-Chun Liao: Investigation. Selina Kern: Investigation. David Zanders: Resources. Anjana Devi: Supervision, funding acquisition. Ryan W. Crisp: Supervision, funding acquisition. Julien Bachmann: Conceptualization, writing (review & editing), supervision, funding acquisition.
Correspondence to Julien Bachmann.
J.B. declares a financial interest in the company Atlant 3D Nanosystems as a co-founder and shareholder. The authors declare no further competing interest.
Communications Materials thanks Shi-Joon Sung, and the other, anonymous, reviewers for their contribution to the peer review of this work.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
Reprints and permissions
Mc Naire, M., Pannen, S., Englhard, J. et al. Ultrathin Sb2S3 solar cells processed by atomic-layer additive manufacturing. Commun Mater 7, 180 (2026). https://doi.org/10.1038/s43246-026-01279-7
Download citation
Received:
Accepted:
Published:
Version of record:
DOI: https://doi.org/10.1038/s43246-026-01279-7
Anyone you share the following link with will be able to read this content:
Sorry, a shareable link is not currently available for this article.
Provided by the Springer Nature SharedIt content-sharing initiative
Advertisement
Communications Materials (Commun Mater)
ISSN 2662-4443 (online)
© 2026 Springer Nature Limited
Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.
