Datasets:
Formats:
imagefolder
Languages:
English
Size:
1K - 10K
ArXiv:
Tags:
scientific-images
materials-science
scientific-figures
atomic-layer-deposition
atomic-layer-etching
experimental-usecase
License:
| {"file_name":"images/train_atomic-layer-deposition_experimental-usecase_13_fig_1.jpg","caption":"FIG. 1. Dependence of the Si content in the $\\mathrm{Ti - Si - N}$ films and deposition thickness per cycle on the $\\mathrm{SiH_4}$ partial pressure for the films grown on $\\mathrm{SiO_2}$ at the substrate temperature of $180^{\\circ}\\mathrm{C}$ . $\\mathrm{Ti - Si - N}$ films were grown by sequential supply of TDMAT for $5 \\mathrm{s}$ , $\\mathrm{SiH_4}$ for $10 \\mathrm{s}$ , and $\\mathrm{NH_3}$ for $10 \\mathrm{s}$ .","id":"train/atomic-layer-deposition/experimental-usecase/13/fig_1","sample_id":"atomic-layer-deposition/experimental-usecase/13/fig_1","subset":"multi-axis-chart","split":"train","classification":[{"panel_id":"a","label":"multi-axis chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The chart shows the relationship between the Si content and the film thickness, and the partial pressure of SiH4.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Partial pressure of SiH₄ [Pa] | Si contents [at.%] | Film thickness per cycle [nm/cycle] |\\n|---|---|---|\\n| 0.1 | 17 | 0.25 |\\n| 1 | 15 | 0.23 |\\n| 10 | 18 | 0.21 |\\n| 100 | 19 | 0.19 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Yes/No\",\"answer\":\"No\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"As the partial pressure increases, so does the Si content, albeit slightly.\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"At that pressure, the Si content is closer to 15 at% and the film thickness is around 0.3 nm/cycle.\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"In this case, the Si deposited through MOALD is not comparable with that deposited through MOCVD, because the latter increases logarithmically with the partial pressure of SiH4, which differs with the data shown in the figure.\"}]}]","bbox":[{"panel_id":"a","x":2,"y":2,"width":601,"height":513}],"source":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/13/images/fig_1.jpg","provenance":{"source_annotation":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/13/images/fig_1.json","source_content":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/13/content.json","source_pdf":["icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/13/Jae-Sik Min et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"experimental-usecase","paper_id":"13","first_classification_panel_id":"a","first_classification_label":"multi-axis chart","caption_source":"content.json:image_caption"},"width":605,"height":519,"image_format":"jpeg","image_sha256":"b35a81d8e759d4a2f739ef56a1e1ba1b6d9e872f0da3939fbc7c8773623d3143","metadata_license":"CC BY 4.0","image_license":"source_publisher_rights_reserved","image_reuse_status":"non_commercial_research_use_only","schema_version":"1.0.0"} | |
| {"file_name":"images/train_atomic-layer-deposition_experimental-usecase_13_fig_2.jpg","caption":"FIG. 2. Dependence of the Si content in the $\\mathrm{Ti - Si - N}$ films and deposition thickness per cycle on the $\\mathrm{SiH_4 / NH_3}$ ratio for the films grown on $\\mathrm{SiO_2}$ at the substrate temperature of $180^{\\circ}\\mathrm{C}$ . $\\mathrm{SiH_4}$ and $\\mathrm{NH_3}$ were simultaneously supplied in the sequence of TDMAT for $5\\mathrm{s}$ and $\\mathrm{SiH_4 / NH_3}$ for $10\\mathrm{s}$ . On the contrary, when TDMAT and $\\mathrm{NH_3}$ were alternately supplied without $\\mathrm{SiH_4}$ , the saturated TiN deposition thickness per cycle was $0.44~\\mathrm{nm / cycle}$ .","id":"train/atomic-layer-deposition/experimental-usecase/13/fig_2","sample_id":"atomic-layer-deposition/experimental-usecase/13/fig_2","subset":"multi-axis-chart","split":"train","classification":[{"panel_id":"a","label":"multi-axis chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The chart shows the relation between SiH₄/NH₃ ratio and Si content and film thickness per cycle. As the ratio increases, the Si content increases until it reaches a plateau, while the film thickness per cycle decreases.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| SiH₄/NH₃ ratio | Si Contents [at.%] | Film thickness per cycle [nm/cycle] |\\n|---|---|---|\\n| 10⁻² | 1 | 0.38 |\\n| 10⁻¹ | 18 | 0.28 |\\n| 10⁰ | 23 | 0.19 |\\n| 10¹ | 23 | 0.01 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Yes/No\",\"answer\":\"Yes\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"There is an increase in the Si content until SiH4/NH3 ratio of 1 is reached, where a plateau of 23 at% is achieved. Conversely, the film thickness decreases with increasing Si content and SiH4/NH3 ratio.\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Saturation is reached at a value of 23 at% when the SiH4/NH3 ratio is 1.\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"The film thickness decreases as the Si content increase, due to a reduction in the number of molecules caused by a SiH4 blocking effect, or NH3 deficiency.\"}]}]","bbox":[{"panel_id":"a","x":2,"y":2,"width":602,"height":535}],"source":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/13/images/fig_2.jpg","provenance":{"source_annotation":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/13/images/fig_2.json","source_content":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/13/content.json","source_pdf":["icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/13/Jae-Sik Min et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"experimental-usecase","paper_id":"13","first_classification_panel_id":"a","first_classification_label":"multi-axis chart","caption_source":"content.json:image_caption"},"width":608,"height":539,"image_format":"jpeg","image_sha256":"e2b30c1f3b37d695bfdf01d35ae2eebad0fd7e36df3bd9015f8ff20bf025a52b","metadata_license":"CC BY 4.0","image_license":"source_publisher_rights_reserved","image_reuse_status":"non_commercial_research_use_only","schema_version":"1.0.0"} | |
| {"file_name":"images/train_atomic-layer-deposition_experimental-usecase_18_fig_2.jpg","caption":"FIG. 2. (a) SE parameters $(\\Psi ,\\Delta)$ measured on blank Si substrate (solid line) and after 10 cycles $\\mathrm{ZrN}$ deposition (dotted line). (b) The dynamic SE parameter $\\Delta$ measured at $4.0\\mathrm{eV}$ for 10 cycles $\\mathrm{ZrN}$ deposition. Systematic variation in $\\Delta$ with precursor pulse $(\\mathrm{t}_1)$ , plasma exposure $(\\mathrm{t}_3)$ , and inert gas purges $(\\mathrm{t}_2$ and $\\mathrm{t}_4)$ are highlighted.","id":"train/atomic-layer-deposition/experimental-usecase/18/fig_2","sample_id":"atomic-layer-deposition/experimental-usecase/18/fig_2","subset":"multi-axis-chart","split":"train","classification":[{"panel_id":"a","label":"multi-axis chart"},{"panel_id":"b","label":"line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"Spectroscopic ellipsometry parameters Ψ and Δ versus photon energy for a blank Si substrate and after 10 ZrN PEALD cycles. Spectral shifts after deposition confirm ultrathin ZrN film formation and demonstrate SE sensitivity to early-stage growth.\"},{\"panel_id\":\"b\",\"text\":\"Dynamic evolution of ellipsometric parameter Δ at 4.0 eV during 10 PEALD cycles. Periodic, step-like changes in Δ aligned with precursor, plasma, and purge steps provide in-situ evidence of self-limiting, cycle-by-cycle ZrN growth.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Photon Energy (eV) | Δ (°) | Ψ (°) |\\n|---|---|---|\\n| 1.0 | ~175 | ~35 |\\n| 2.0 | ~170 | ~33 |\\n| 3.0 | ~165 | ~32 |\\n| 4.0 | ~150 | ~30 |\\n| 5.0 | ~165 | ~31 |\\n| 6.0 | ~170 | ~32 |\"},{\"panel_id\":\"b\",\"text\":\"| Time (s) | Δ (°) |\\n|---|---|\\n| 100 | ~144 |\\n| 200 | ~143 |\\n| 300 | ~141 |\\n| 400 | ~140 |\\n| 500 | ~139 |\\n| 550 | ~138 |\"}]","vqa":"[{\"panel_id\":\"d\",\"items\":[{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The in-situ ellipsometry measurements enable real-time monitoring of ZrN film growth, allowing direct observation of cycle-resolved thickness changes and confirmation of self-limiting behavior. This capability supports precise control of growth per cycle, rapid identification of saturation conditions, and reliable optimization of PEALD parameters for uniform, ultrathin ZrN films.\"}]},{\"panel_id\":\"b\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Both Ψ and Δ exhibit systematic spectral shifts after 10 cycles, indicating the formation of an ultrathin ZrN film.\"}]},{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Yes/No\",\"answer\":\"Yes\"}]},{\"panel_id\":\"c\",\"items\":[{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"List\",\"answer\":\"Spectral shift in Ψ and Δ after deposition, Periodic step-like changes in Δ during each ALD cycle, Reproducibility of Δ response across multiple cycles, Distinct Δ responses during precursor and plasma steps\"}]}]","bbox":[{"panel_id":"a","x":13,"y":9,"width":494,"height":357},{"panel_id":"b","x":524,"y":11,"width":479,"height":355}],"source":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/18/images/fig_2.jpg","provenance":{"source_annotation":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/18/images/fig_2.json","source_content":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/18/content.json","source_pdf":["icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/18/Triratna Muneshwar et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"experimental-usecase","paper_id":"18","first_classification_panel_id":"a","first_classification_label":"multi-axis chart","caption_source":"content.json:image_caption"},"width":1003,"height":366,"image_format":"jpeg","image_sha256":"f4210e67b9670ff5a81442730500152efe3b346a34cf84ffeec5430209fa7b72","metadata_license":"CC BY 4.0","image_license":"source_publisher_rights_reserved","image_reuse_status":"non_commercial_research_use_only","schema_version":"1.0.0"} | |
| {"file_name":"images/train_atomic-layer-deposition_experimental-usecase_27_fig_1.jpg","caption":"Fig. 1. Growth rate and resistivity of TiN films deposited by remote PEALD technique using TDMAT precursor and nitrogen plasma as a function of deposition temperature. Inset of this figure shows the film thickness as a function of process cycles.","id":"train/atomic-layer-deposition/experimental-usecase/27/fig_1","sample_id":"atomic-layer-deposition/experimental-usecase/27/fig_1","subset":"multi-axis-chart","split":"train","classification":[{"panel_id":"a","label":"multi-axis chart"},{"panel_id":"b","label":"line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The chart shows the relationship between deposition temperature and thickness and resistivity of a material.\"},{\"panel_id\":\"b\",\"text\":\"The line chart displays the thickness of a material as a function of the number of cycles.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Deposition Temperature (°C) | Thickness (nm) | Resistivity (Ω·cm) |\\n|---|---|---|\\n| 175 | 15 | 500 |\\n| 200 | 19 | 450 |\\n| 250 | 19 | 350 |\\n| 300 | 19 | 400 |\\n| 350 | 25 | 500 |\"},{\"panel_id\":\"b\",\"text\":\"| Number of Cycles | Thickness (nm) |\\n|---|---|\\n| 20 | 8 |\\n| 40 | 15 |\\n| 60 | 22 |\\n| 80 | 30 |\"}]","vqa":"[{\"panel_id\":\"b\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"The GPC is 20/50 = 0.4 nm/cycle.\"}]},{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"A temperature window is a range of temperatures where the GPC is independent of temperature. The graph shows this between 200 and 300 degrees celcius.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Within the range of 200 to 300 degrees celcius there is a change in resistivity, indicating that the temperature window does not apply for the resistivity.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"A possible explaination for more growth at higher temperatures is parasitic CVD due to decomposition of precursor molecules.\"}]}]","bbox":[{"panel_id":"b","x":182,"y":21,"width":257,"height":184},{"panel_id":"a","x":3,"y":5,"width":659,"height":463}],"source":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/27/images/fig_1.jpg","provenance":{"source_annotation":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/27/images/fig_1.json","source_content":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/27/content.json","source_pdf":["icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/27/Ju Youn KIM et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"experimental-usecase","paper_id":"27","first_classification_panel_id":"a","first_classification_label":"multi-axis chart","caption_source":"content.json:image_caption"},"width":664,"height":469,"image_format":"jpeg","image_sha256":"279593c94d95a41565944df4a37503d9c29cf9d9253f76b169df26f72a6afa9a","metadata_license":"CC BY 4.0","image_license":"source_publisher_rights_reserved","image_reuse_status":"non_commercial_research_use_only","schema_version":"1.0.0"} | |
| {"file_name":"images/train_atomic-layer-deposition_experimental-usecase_28_figure_11.jpg","caption":"Figure 11. Time-resolved TS behavior of the GeSe selector devices through fabrication using (a) the conventional ALD process and (b) the ALD-DFM process. Magnified view of the transition region in (b) to the (c) ON-state and (d) OFF-state.","id":"train/atomic-layer-deposition/experimental-usecase/28/figure_11","sample_id":"atomic-layer-deposition/experimental-usecase/28/figure_11","subset":"multi-axis-chart","split":"train","classification":[{"panel_id":"a","label":"multi-axis chart"},{"panel_id":"b","label":"multi-axis chart"},{"panel_id":"c","label":"multi-axis chart"},{"panel_id":"d","label":"multi-axis chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The figure shows a multi-axis chart with voltage and current plotted against time. The voltage and current traces are shown for two different channels, V_Ch1 and V_DUT, and I_DUT respectively where there is a peak after 2 micro seconds.\"},{\"panel_id\":\"b\",\"text\":\"The figure shows a multi-axis chart with voltage and current plotted against time. The voltage and current traces are shown for two different channels, V_Ch1 and V_DUT, and I_DUT respectively where there is a peak after 2 micro seconds.\"},{\"panel_id\":\"c\",\"text\":\"The figure shows a multi-axis chart with voltage and current plotted against time. The voltage and current traces are shown for two different channels, V_Ch1 and V_DUT, and I_DUT respectively in the build up phase for the switch.\"},{\"panel_id\":\"d\",\"text\":\"The figure shows a multi-axis chart with voltage and current plotted against time. The voltage and current traces are shown for two different channels, V_Ch1 and V_DUT, and I_DUT respectively for the settling after the switch.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Time (µs) | Voltage (V) | Current (mA) |\\n|---|---|---|\\n| 0 | 0 | 0 |\\n| 0.5 | 0.5 | 0 |\\n| 1 | 1 | 0 |\\n| 1.5 | 2 | 3 |\\n| 2 | 3 | 5 |\\n| 2.5 | 2 | 3 |\\n| 3 | 1 | 0 |\\n| 3.5 | 0.5 | 0 |\\n| 4 | 0 | 0 |\"},{\"panel_id\":\"b\",\"text\":\"| Time (µs) | Voltage (V) | Current (mA) |\\n|---|---|---|\\n| 0 | 0 | 0 |\\n| 0.5 | 1 | 0 |\\n| 1 | 1.8 | 0 |\\n| 1.5 | 2 | 12 |\\n| 2 | 2 | 20 |\\n| 2.5 | 1.5 | 10 |\\n| 3 | 1.2 | 5 |\\n| 3.5 | 1 | 0 |\\n| 4 | 0 | 0 |\"},{\"panel_id\":\"c\",\"text\":\"| Time (µs) | Voltage (V) | Current (mA) |\\n|---|---|---|\\n| 0.9 | 1.9 | 0 |\\n| 1 | 2 | 0 |\\n| 1.1 | 1.5 | 8 |\\n| 1.2 | 1.9 | 8 |\"},{\"panel_id\":\"d\",\"text\":\"| Time (µs) | Voltage (V) | Current (mA) |\\n|---|---|---|\\n| 3.0 | 1.2 | 5 |\\n| 3.1 | 1.2 | 3 |\\n| 3.2 | 1.2 | 0 |\\n| 3.3 | 1.0 | 0 |\\n| 3.4 | 0.8 | 0 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"2 microseconds\"}]},{\"panel_id\":\"c\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"2.1 V.\"}]},{\"panel_id\":\"d\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Yes/No\",\"answer\":\"No\"}]},{\"panel_id\":\"b\",\"items\":[{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The moment at which the GeSe transitions from a resistive phase to a low resistance mode.\"}]}]","bbox":[{"panel_id":"a","x":15,"y":38,"width":312,"height":270},{"panel_id":"b","x":363,"y":38,"width":312,"height":270},{"panel_id":"c","x":15,"y":354,"width":312,"height":268},{"panel_id":"d","x":363,"y":354,"width":312,"height":268}],"source":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/28/images/figure_11.jpg","provenance":{"source_annotation":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/28/images/figure_11.json","source_content":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/28/content.json","source_pdf":["icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/28/Woohyun Kim et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"experimental-usecase","paper_id":"28","first_classification_panel_id":"a","first_classification_label":"multi-axis chart","caption_source":"content.json:image_caption"},"width":675,"height":622,"image_format":"jpeg","image_sha256":"82e317b84a292fc23be9c22f3a10bd652a321e05a85cecf8b315ab67a4703572","metadata_license":"CC BY 4.0","image_license":"source_publisher_rights_reserved","image_reuse_status":"non_commercial_research_use_only","schema_version":"1.0.0"} | |
| {"file_name":"images/train_atomic-layer-deposition_experimental-usecase_28_figure_4.jpg","caption":"Figure 4. (a) Layer densities versus cycle number to confirm the saturation growth rate without an incubation cycle. (b) Effect of the substrate temperature on the growth rate.","id":"train/atomic-layer-deposition/experimental-usecase/28/figure_4","sample_id":"atomic-layer-deposition/experimental-usecase/28/figure_4","subset":"multi-axis-chart","split":"train","classification":[{"panel_id":"a","label":"multi-axis chart"},{"panel_id":"b","label":"multi-axis chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The chart shows the layer density and composition of GeSe<sub>(1-x)</sub> as a function of cycle number for SiO<sub>2</sub>, TiN, and SiO<sub>2</sub>, DFM.\"},{\"panel_id\":\"b\",\"text\":\"The chart illustrates the growth rate and composition of GeSe<sub>(1-x)</sub> as a function of substrate temperature for SiO<sub>2</sub> and TiN.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Cycle Number | Layer Density (µg/cm²) | x in GeSe<sub>(1-x)</sub> |\\n|---|---|---|\\n| 0 | 0 | 0 |\\n| 100 | 5 | 0.5 |\\n| 200 | 9 | 0.5 |\\n| 400 | 17 | 0.5 |\"},{\"panel_id\":\"b\",\"text\":\"| Substrate Temperature (°C) | Growth rate (ngcm<sup>-2</sup>cy<sup>-1</sup>) | x in GeSe<sub>(1-x)</sub> |\\n|---|---|---|\\n| 70 | 57 | 0.5 |\\n| 90 | 26 | 0.5 |\\n| 110 | 19 | 0.5 |\\n| 130 | 10 | 0.6 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"For perfect ALD, every cycle, the amount of deposited material is the same, meaning there should be a linear relation between the cycle number and the layer density.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"For all substrates and dosing methods the composition is the same, i.e. Ge:Se is 1:1. This means that differently dosing the precursors will not result in different composition.\"}]},{\"panel_id\":\"b\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"This can be caused by desorption of adsorbed species leading to less film growth.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"With increasing temperature the Ge contribution increases in the film, meaning there is less Se. This means that the Se precursor desorbs more easily.\"}]}]","bbox":[{"panel_id":"a","x":2,"y":46,"width":508,"height":408},{"panel_id":"b","x":522,"y":44,"width":521,"height":411}],"source":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/28/images/figure_4.jpg","provenance":{"source_annotation":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/28/images/figure_4.json","source_content":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/28/content.json","source_pdf":["icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/28/Woohyun Kim et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"experimental-usecase","paper_id":"28","first_classification_panel_id":"a","first_classification_label":"multi-axis chart","caption_source":"content.json:image_caption"},"width":1048,"height":458,"image_format":"jpeg","image_sha256":"fb327650cc20cc06c97109a92e542c76603d522aa76e401420908901195e9697","metadata_license":"CC BY 4.0","image_license":"source_publisher_rights_reserved","image_reuse_status":"non_commercial_research_use_only","schema_version":"1.0.0"} | |
| {"file_name":"images/train_atomic-layer-deposition_experimental-usecase_31_figure_3.jpg","caption":"Figure 3. Growth rate and resistivity vs precursor exposure time of PE-ALD Co using (a) $\\mathrm{CoCp(CO)_2}$ and (b) $\\mathrm{CoCp_2}$ as a Co precursor, respectively.","id":"train/atomic-layer-deposition/experimental-usecase/31/figure_3","sample_id":"atomic-layer-deposition/experimental-usecase/31/figure_3","subset":"multi-axis-chart","split":"train","classification":[{"panel_id":"a","label":"multi-axis chart"},{"panel_id":"b","label":"multi-axis chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"For PE-ALD Co on SiO₂ using CoCp(CO)₂ at 300 °C, the growth rate increases with substrate exposure time and reaches a maximum at longer exposure, while the resistivity initially decreases and then increases at high exposure times, indicating an optimal time window for achieving low-resistivity cobalt films.\"},{\"panel_id\":\"b\",\"text\":\"In the PE-ALD Co process using CoCp₂ at 300 °C, the growth rate increases rapidly at short exposure times and then saturates, while the resistivity remains relatively low and stable with increasing exposure time, suggesting a more robust and self-limiting growth behavior.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| t<sub>S</sub> (s)|Growth Rate (Å/cycle)|Resistivity (µΩcm)|\\n|------------------|---------------------|------------------|\\n|0|0|0|\\n|1|2|250|\\n|2|1|100|\\n|3|0.5|50|\\n|4|0|25|\"},{\"panel_id\":\"b\",\"text\":\"| t<sub>S</sub> (s)|Growth Rate (Å/cycle)|Resistivity (µΩcm)|\\n|------------------|---------------------|------------------|\\n|0|0|0|\\n|2|0.5|250|\\n|4|0.5|150|\\n|6|0.5|100|\\n|8|0.5|50|\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Yes/No\",\"answer\":\"NO\"}]},{\"panel_id\":\"b\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"CoCp₂.\"}]},{\"panel_id\":\"d\",\"items\":[{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Growth saturation confirms true ALD behavior, ensuring precise thickness control and excellent conformality on complex device structures. The consistently low resistivity obtained with CoCp₂ indicates high-purity metallic cobalt, which is essential for interconnect and contact applications in advanced CMOS technology. Together, these properties make CoCp₂ a suitable precursor for scalable and reliable cobalt deposition.\"}]},{\"panel_id\":\"c\",\"items\":[{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"List\",\"answer\":\"Lack of growth-rate saturation, Strong increase in resistivity at longer exposure times, Unstable electrical properties, Evidence of CVD-like or decomposition-driven growth\"}]}]","bbox":[{"panel_id":"a","x":0,"y":2,"width":542,"height":370},{"panel_id":"b","x":3,"y":386,"width":536,"height":355}],"source":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/31/images/figure_3.jpg","provenance":{"source_annotation":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/31/images/figure_3.json","source_content":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/31/content.json","source_pdf":["icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/31/Han-Bo-Ram Lee et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"experimental-usecase","paper_id":"31","first_classification_panel_id":"a","first_classification_label":"multi-axis chart","caption_source":"content.json:image_caption"},"width":544,"height":744,"image_format":"jpeg","image_sha256":"05a494ef100bcaf930a71c2c186e8a732fa89c0bcfa2edbb38b0510fef0ea07b","metadata_license":"CC BY 4.0","image_license":"source_publisher_rights_reserved","image_reuse_status":"non_commercial_research_use_only","schema_version":"1.0.0"} | |
| {"file_name":"images/train_atomic-layer-deposition_experimental-usecase_36_fig_10.jpg","caption":"Fig. 10 a Transfer characteristics and b log $I_{\\mathrm{DS}}$ vs $V_{\\mathrm{GS}}$ plots of the conventional HEMT and the $\\mathrm{Ga_2O_3}$ MOS-HEMT","id":"train/atomic-layer-deposition/experimental-usecase/36/fig_10","sample_id":"atomic-layer-deposition/experimental-usecase/36/fig_10","subset":"multi-axis-chart","split":"train","classification":[{"panel_id":"a","label":"multi-axis chart"},{"panel_id":"b","label":"multiple scatter plot"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The chart shows the drain current (I<sub>DS</sub>) and transconductance (g<sub>m</sub>) for both Ni/Au-HEMT and Ga<sub>2</sub>O<sub>3</sub> MOS-HEMT as a function of gate-source voltage (V<sub>GS</sub>) with a VDS value of 8V.\"},{\"panel_id\":\"b\",\"text\":\"The chart displays the drain current (I<sub>DS</sub>) of Ni/Au-HEMT and Ga<sub>2</sub>O<sub>3</sub> MOS-HEMT as a function of gate-source voltage (V<sub>GS</sub>) on a logarithmic scale with a VDS value of 8V.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| I<sub>DS</sub> (mA/mm) | Ni/Au-HEMT | Ga2O3 MOS-HEMT |\\n| --- | --- | --- |\\n|-6| 0 | 0 |\\n|-4| 0 | 0 |\\n|-2| 300 | 400 |\\n|0| 500 | 800 |\\n|2| 800 | 1000 |\"},{\"panel_id\":\"b\",\"text\":\"| I<sub>DS</sub> (mA/mm) | Ni/Au-HEMT | Ga2O3 MOS-HEMT |\\n| --- | --- | --- |\\n| -6 | 10^-3 | 10^-4 |\\n| -4 | 10^-1 | 10^2 |\\n| -2 | 10^2 | 10^2 |\\n| 0 | 10^3 | 10^3 |\\n| 2 | 10^3 | 10^3 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Ni/Au-HEMT and Ga2O3 MOS-HEMT were both performed with a V_DS value of 8 V.\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"As can be seen in the subfigure a, the Ga2O3 MOS-HEMT has higher transcondutance values.\"}]},{\"panel_id\":\"b\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"The samples remain constant until approximately -4 V, at which point they both increases and reach a plateau.\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"The subthreshold swing parameter is higher for the Ni/Au-HEMT sample with a value of 188 mV/dec.\"}]}]","bbox":[{"panel_id":"a","x":8,"y":26,"width":427,"height":427},{"panel_id":"b","x":453,"y":33,"width":344,"height":422}],"source":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/36/images/fig_10.jpg","provenance":{"source_annotation":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/36/images/fig_10.json","source_content":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/36/content.json","source_pdf":["icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/36/Huan-Yu Shih et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"experimental-usecase","paper_id":"36","first_classification_panel_id":"a","first_classification_label":"multi-axis chart","caption_source":"content.json:image_caption"},"width":797,"height":455,"image_format":"jpeg","image_sha256":"b224e365c23d3b47e197db4d9638e0287052928f96bd46a32104e259abddafa1","metadata_license":"CC BY 4.0","image_license":"source_publisher_rights_reserved","image_reuse_status":"non_commercial_research_use_only","schema_version":"1.0.0"} | |
| {"file_name":"images/train_atomic-layer-deposition_experimental-usecase_36_fig_13.jpg","caption":"Fig. 13 Microwave power characteristics of the conventional HEMT and the $\\mathrm{Ga_2O_3}$ MOS-HEMT","id":"train/atomic-layer-deposition/experimental-usecase/36/fig_13","sample_id":"atomic-layer-deposition/experimental-usecase/36/fig_13","subset":"multi-axis-chart","split":"train","classification":[{"panel_id":"a","label":"multi-axis chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The chart shows the output power and power drain, and power added efficiency as a function of the input power at a frequency of 2.4G and drain source voltage of 16V, for Ni/Au-HEMT and Ga2O3 MOS-HEMT.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Input Power (dBm) | G_p Ga2O3 | G_p Ni/Au | P_out Ga2O3 | P_out Ni/Au | P.A.E. Ni/Au | P.A.E. Ga2O3 |\\n| --- | --- | --- | --- | --- | --- | --- |\\n| -25 | 25 | 24 | 0 | 0 | 0 | 0 |\\n| -20 | 25 | 24 | 4 | 3 | 0 | 0 |\\n| -15 | 25 | 24 | 9 | 8 | 0 | 0 |\\n| -10 | 25 | 24 | 15 | 14 | 2 | 2 |\\n| -5 | 25 | 24 | 19 | 18 | 8 | 8 |\\n| 0 | 25 | 24 | 23 | 22 | 25 | 24 |\\n| 5 | 20 | 19 | 20 | 19 | 45 | 35 |\\n| 10 | 15 | 14 | 17 | 16 | 50 | 40 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"They were performed at 2.4 G and 16 V.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"The output power of both the Ni/Au and G2O3 samples increases with the input power.\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"At approximately 0 dBm.\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Yes/No\",\"answer\":\"Yes\"}]}]","bbox":[{"panel_id":"a","x":3,"y":10,"width":621,"height":465}],"source":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/36/images/fig_13.jpg","provenance":{"source_annotation":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/36/images/fig_13.json","source_content":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/36/content.json","source_pdf":["icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/36/Huan-Yu Shih et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"experimental-usecase","paper_id":"36","first_classification_panel_id":"a","first_classification_label":"multi-axis chart","caption_source":"content.json:image_caption"},"width":625,"height":475,"image_format":"jpeg","image_sha256":"86f1078552c1840ea35b64569c2e45185c9fda76b588236b05b1970d7ed1e736","metadata_license":"CC BY 4.0","image_license":"source_publisher_rights_reserved","image_reuse_status":"non_commercial_research_use_only","schema_version":"1.0.0"} | |
| {"file_name":"images/train_atomic-layer-deposition_experimental-usecase_36_fig_5.jpg","caption":"Fig. 5 Refractive index and extinction coefficient of the $\\mathrm{Ga}_{2}\\mathrm{O}_{3}$ thin film plotted with respect to the wavelength","id":"train/atomic-layer-deposition/experimental-usecase/36/fig_5","sample_id":"atomic-layer-deposition/experimental-usecase/36/fig_5","subset":"multi-axis-chart","split":"train","classification":[{"panel_id":"a","label":"multi-axis chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The chart displays the extinction coefficient and refractive index of a material as a function of wavelength. Both of them decreases as the wavelength value increases.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Wavelength (nm) | Extinction coefficient | Refractive index |\\n|----------------|------------------------|------------------|\\n| 300 | 0 | 2.2 |\\n| 400 | 0 | 2.0 |\\n| 500 | 0 | 1.9 |\\n| 600 | 0 | 1.9 |\\n| 700 | 0 | 1.9 |\\n| 800 | 0 | 1.9 |\\n| 900 | 0 | 1.9 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"The refractive index (n) and extinction coefficient (k) of the Ga2O3 thin film.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"At lower wavelength values, the extinction coefficient is 0.15. However, after 300 nm, it decreases rapidly and significantly to 0.\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"The refractive index exhibit higher values than the extinction coefficient.\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Both can be obtained using different models and spectroscopic ellipsometry measurements.\"}]}]","bbox":[{"panel_id":"a","x":1,"y":12,"width":621,"height":393}],"source":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/36/images/fig_5.jpg","provenance":{"source_annotation":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/36/images/fig_5.json","source_content":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/36/content.json","source_pdf":["icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/36/Huan-Yu Shih et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"experimental-usecase","paper_id":"36","first_classification_panel_id":"a","first_classification_label":"multi-axis chart","caption_source":"content.json:image_caption"},"width":625,"height":406,"image_format":"jpeg","image_sha256":"36b4445ad8594fa37183fef95eb78420a0c85256a27be6ff107fef6572166c1e","metadata_license":"CC BY 4.0","image_license":"source_publisher_rights_reserved","image_reuse_status":"non_commercial_research_use_only","schema_version":"1.0.0"} | |
| {"file_name":"images/train_atomic-layer-deposition_experimental-usecase_4_fig_1.jpg","caption":"Fig. 1. Growth rate and refractive index of $\\mathrm{HfO_2}$ films as a function of hafnium precursor pulse length. The water pulse and purge times were $0.5\\mathrm{s}$ . The error bars express the variation of the refractive index in each sample.","id":"train/atomic-layer-deposition/experimental-usecase/4/fig_1","sample_id":"atomic-layer-deposition/experimental-usecase/4/fig_1","subset":"multi-axis-chart","split":"train","classification":[{"panel_id":"a","label":"multi-axis chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The chart shows the growth rate and refractive index at 580 nm of HfO₂ deposited on borosilicate glass as a function of Hf[N(CH₃)(C₂H₅)₄] pulse length at a growth temperature of 250 °C.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Hf[N(CH3)(C2H5)]4 PULSE LENGTH, s | GROWTH RATE, nm/cycle | REFRACTIVE INDEX AT 580 nm |\\n|---|---|---|\\n| 0.2 | 0.08 | 2.08 |\\n| 0.4 | 0.09 | 2.09 |\\n| 0.6 | 0.09 | 2.09|\\n| 0.8 | 0.09 | 2.09 |\\n| 1.0 | 0.09 | 2.08 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"It first increases and then decreases after having reached a maximal refractive index of about 2.10 after 0.4 – 0.6 seconds. This can arise from the thermal decomposition of Hf[N(CH3)(C2H5)]4. This thermal decomposition, which may become more significant during longer exposure times, may slightly reduce the film density or change its composition. Although the changes in refractive index are measurable, they are however very small.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Borosilicate glass.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"250 °C.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"After 0.4 seconds, the growth rate no longer is affected by the pulse length, indicating that the surface reactions are self-limiting.\"}]}]","bbox":[{"panel_id":"a","x":0,"y":3,"width":641,"height":399}],"source":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/4/images/fig_1.jpg","provenance":{"source_annotation":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/4/images/fig_1.json","source_content":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/4/content.json","source_pdf":["icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/4/Kukli et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"experimental-usecase","paper_id":"4","first_classification_panel_id":"a","first_classification_label":"multi-axis chart","caption_source":"content.json:img_caption"},"width":642,"height":403,"image_format":"jpeg","image_sha256":"dc6e84ee27a5c559e30e7fc0891ee55b534353bdc7283062a6d82015c1184669","metadata_license":"CC BY 4.0","image_license":"source_publisher_rights_reserved","image_reuse_status":"non_commercial_research_use_only","schema_version":"1.0.0"} | |
| {"file_name":"images/train_atomic-layer-deposition_experimental-usecase_4_fig_3.jpg","caption":"Fig. 3. Growth rate and refractive index of $\\mathrm{HfO_2}$ films as a function of growth temperature.","id":"train/atomic-layer-deposition/experimental-usecase/4/fig_3","sample_id":"atomic-layer-deposition/experimental-usecase/4/fig_3","subset":"multi-axis-chart","split":"train","classification":[{"panel_id":"a","label":"multi-axis chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The chart shows the growth rate of HfO₂ on borosilicate glass at various temperatures, along with the corresponding refractive index at 580 nm.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| GROWTH TEMPERATURE, °C | GROWTH RATE, nm/cycle | REFRACTIVE INDEX AT 580 nm |\\n|---|---|---|\\n| 150 | 0.10 | 2.08 |\\n| 175 | 0.09 | 2.08 |\\n| 200 | 0.09 | 2.10 |\\n| 225 | 0.09 | 2.09 |\\n| 250 | 0.09 | 2.09 |\\n| 275 | 0.10 | 2.08 |\\n| 300 | 0.12 | 2.06 |\\n| 325 | 0.15 | 2.05 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"150 to 325 °C.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"0.4 s Hf[N(CH3)(C2H5)]4 exposure, 0.5 s purge, 0.5 s H2O exposure, 0.5 s purge.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"The growth rate as a function of temperature is stable/constant from 200 to 250 °C.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"It varied between 2.05 and 2.10.\"}]}]","bbox":[{"panel_id":"a","x":2,"y":5,"width":642,"height":399}],"source":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/4/images/fig_3.jpg","provenance":{"source_annotation":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/4/images/fig_3.json","source_content":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/4/content.json","source_pdf":["icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/4/Kukli et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"experimental-usecase","paper_id":"4","first_classification_panel_id":"a","first_classification_label":"multi-axis chart","caption_source":"content.json:img_caption"},"width":644,"height":406,"image_format":"jpeg","image_sha256":"9fda5b98f5039166a81521e8b225fe63a60bd6ade0d2baa6de95e4b553233980","metadata_license":"CC BY 4.0","image_license":"source_publisher_rights_reserved","image_reuse_status":"non_commercial_research_use_only","schema_version":"1.0.0"} | |
| {"file_name":"images/train_atomic-layer-deposition_experimental-usecase_4_fig_8.jpg","caption":"Fig. 8. Selected $CV$ and dissipation curves measured for $\\mathrm{Al / HfO_2 / Si}$ capacitor structures at $AC$ signal frequency of $500\\mathrm{kHz}$ . Labels indicate $\\mathrm{HfO_2}$ growth temperatures. The thickness of the films grown at $200^{\\circ}\\mathrm{C}$ and $300^{\\circ}\\mathrm{C}$ are $133\\mathrm{nm}$ and $178\\mathrm{nm}$ , respectively.","id":"train/atomic-layer-deposition/experimental-usecase/4/fig_8","sample_id":"atomic-layer-deposition/experimental-usecase/4/fig_8","subset":"multi-axis-chart","split":"train","classification":[{"panel_id":"a","label":"multi-axis chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The chart displays capacitance and dissipation against bias voltage for two different growth temperatures of HfO₂.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| BIAS VOLTAGE, V | 200 °C, CAPACITANCE, nF | 300 °C, CAPACITANCE, nF | 200 °C, DISSIPATION, D | 300 °C, DISSIPATION, D |\\n|---|---|---|---|---|\\n| -4 | 0.16 | | 0.05 | |\\n| -3 | 0.16 | 0.15 | 0.05 | 0.1 |\\n| -2 | 0.15 | 0.14 | 0.05 | 0.1 |\\n| -1 | 0.12 | 0.12 | 0.1 | 0.5 |\\n| 0 | 0.04 | 0.04 | 0.05 | 0.025 |\\n| 1 | 0.03 | 0.04 | 0.025 | 0.025 |\\n| 2 | 0.03 | | 0.025 | |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"It is 0.16 nF at -4 V and 0.03 nF at 2.5 V.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"It is 0.15 nF at -4 V and 0.04 nF at 1.5 V.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"At about 0.5 and -0.5 V.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"At about -0.25 and -0.5 V.\"}]}]","bbox":[{"panel_id":"a","x":1,"y":2,"width":555,"height":354}],"source":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/4/images/fig_8.jpg","provenance":{"source_annotation":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/4/images/fig_8.json","source_content":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/4/content.json","source_pdf":["icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/4/Kukli et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"experimental-usecase","paper_id":"4","first_classification_panel_id":"a","first_classification_label":"multi-axis chart","caption_source":"content.json:img_caption"},"width":558,"height":358,"image_format":"jpeg","image_sha256":"4603bedd3a706fba1141ba86b271103093fde2767f9d671d9647f84f37511ec4","metadata_license":"CC BY 4.0","image_license":"source_publisher_rights_reserved","image_reuse_status":"non_commercial_research_use_only","schema_version":"1.0.0"} | |
| {"file_name":"images/train_atomic-layer-deposition_experimental-usecase_40_fig_2.jpg","caption":"FIG. 2. Dependence of the growth rate and silicon content in the $\\mathrm{Ti - Si - N}$ films on the $\\mathrm{SiH_4}$ pulse time. The films were grown on $\\mathrm{SiO_2}$ at $350^{\\circ}\\mathrm{C}$ with the sequential supply of $\\mathrm{TiCl_4}$ for $1\\mathrm{s}$ , $\\mathrm{SiH_4}$ for $0\\mathrm{s}\\sim 15\\mathrm{s}$ , and $\\mathrm{N}_2 / \\mathrm{H}_2 / \\mathrm{Ar}$ plasma for $10\\mathrm{s}$ . The partial pressure of $\\mathrm{SiH_4}$ was fixed at 0.1 torr.","id":"train/atomic-layer-deposition/experimental-usecase/40/fig_2","sample_id":"atomic-layer-deposition/experimental-usecase/40/fig_2","subset":"multi-axis-chart","split":"train","classification":[{"panel_id":"a","label":"multi-axis chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"Multi-axis chart showing the dependence of growth per cycle (left axis) and silicon atomic content (right axis) on SiH₄ pulse time during plasma-enhanced ALD of Ti–Si–N at 350 °C. Both quantities increase rapidly at short pulse times and saturate for SiH₄ exposures of ~8 s or longer, reaching ~0.7 Å/cycle growth and ~9–15 at.% Si, indicating self-limiting surface reactions.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| SiH₄ pulse time (s) | Thickness per cycle (Å/cycle, left axis) | Si content (at.%, right axis) |\\n|--------------------|-------------------------------------------|-------------------------------|\\n| 0 | ~0.45 | ~0.0 |\\n| 1 | ~0.50 | ~4.0 |\\n| 2 | ~0.55 | ~5.0 |\\n| 4 | ~0.60 | ~7.5 |\\n| 6 | ~0.65 | ~8.0 |\\n| 8 | ~0.70 | ~9.0 |\\n| 12 | ~0.70 | ~9.0 |\\n| 16 | ~0.70 | ~9.0 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Both the growth per cycle and the silicon content increase rapidly as the SiH₄ pulse time is extended from 0 to about 8 s. This reflects increased surface reaction completion during the SiH₄ exposure. Beyond this range, further increases in pulse time do not significantly change either quantity.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Yes/No\",\"answer\":\"No\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Once the surface reactive sites available for Si incorporation are fully occupied during the SiH₄ half-cycle, additional precursor exposure cannot introduce more silicon. This reflects the self-limiting surface chemistry characteristic of ALD, where adsorption and reaction terminate once saturation is reached.\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Approximately 8 s.\"}]}]","bbox":[{"panel_id":"a","x":0,"y":3,"width":634,"height":555}],"source":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/40/images/fig_2.jpg","provenance":{"source_annotation":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/40/images/fig_2.json","source_content":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/40/content.json","source_pdf":["icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/40/Park et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"experimental-usecase","paper_id":"40","first_classification_panel_id":"a","first_classification_label":"multi-axis chart","caption_source":"content.json:image_caption"},"width":637,"height":561,"image_format":"jpeg","image_sha256":"4a1f3fd55c02d66873772763b31497ca784c5cc9dec3e25da07ed5573e0e1f72","metadata_license":"CC BY 4.0","image_license":"source_publisher_rights_reserved","image_reuse_status":"non_commercial_research_use_only","schema_version":"1.0.0"} | |
| {"file_name":"images/train_atomic-layer-deposition_experimental-usecase_40_fig_3.jpg","caption":"FiG. 3. Dependence of the growth rate and silicon content in the $\\mathrm{Ti - Si - N}$ films on the $\\mathrm{SiH_4}$ pulse time by different reactant exposure sequence. The reactants were supplied in the sequence of $\\mathrm{SiH_4}$ for $0\\mathrm{s} - 15\\mathrm{s}$ , $\\mathrm{TiCl_4}$ for $1\\mathrm{s}$ , and $\\mathrm{N}_2 / \\mathrm{H}_2 / \\mathrm{Ar}$ plasma for $10\\mathrm{s}$ .","id":"train/atomic-layer-deposition/experimental-usecase/40/fig_3","sample_id":"atomic-layer-deposition/experimental-usecase/40/fig_3","subset":"multi-axis-chart","split":"train","classification":[{"panel_id":"a","label":"multi-axis chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"Multi-axis chart showing growth per cycle (left axis) and silicon atomic content (right axis) in plasma-enhanced ALD Ti–Si–N films deposited using a reversed reactant sequence (SiH₄ → TiCl₄ → N₂/H₂/Ar plasma). Both quantities increase with SiH₄ pulse time at short exposures and saturate at approximately 0.60–0.62 Å/cycle and 11–12 at.% Si for pulse times of ~8–10 s, indicating self-limiting surface reactions under this sequence.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"|SiH4 Pulse Time (s) |Thickness (S-T-N) [Å/cycle]|Si Content (S-T-N) [at.%]|\\n|--------------------|-------------------------------|-------------------|\\n| 0 | ~0.46 | – |\\n| 2 | ~0.50 | ~7.5 |\\n| 4 | ~0.55 | ~8.0 |\\n| 6 | ~0.60 | ~10.0 |\\n| 8 | ~0.60 | ~11.0 |\\n| 10 | ~0.62 | ~12.0 |\\n| 12 | ~0.60 | ~12.0 |\\n| 14 | ~0.60 | ~12.0 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Approximately 8–10 seconds.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The silicon content increases rapidly at short SiH₄ pulse times, rising from below 8 at.% at 2 s to about 10–11 at.% by 6–8 s. Beyond this range, the silicon content plateaus near 11–12 at.% despite further increases in pulse time, indicating saturation.\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"When SiH₄ is pulsed first, its interaction with the surface is limited by the initial surface chemistry. The subsequent TiCl₄ pulse can occupy or replace reactive sites before nitrogen incorporation, reducing the number of sites available for silicon retention. This leads to lower final silicon incorporation compared with sequences where TiCl₄ primes the surface first.\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Yes/No\",\"answer\":\"No\"}]}]","bbox":[{"panel_id":"a","x":0,"y":2,"width":433,"height":382}],"source":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/40/images/fig_3.jpg","provenance":{"source_annotation":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/40/images/fig_3.json","source_content":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/40/content.json","source_pdf":["icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/40/Park et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"experimental-usecase","paper_id":"40","first_classification_panel_id":"a","first_classification_label":"multi-axis chart","caption_source":"content.json:image_caption"},"width":436,"height":386,"image_format":"jpeg","image_sha256":"1cd211bc080a849c0c81f889ff8b11d69473d749969bc26d60567ccf3632b57a","metadata_license":"CC BY 4.0","image_license":"source_publisher_rights_reserved","image_reuse_status":"non_commercial_research_use_only","schema_version":"1.0.0"} | |
| {"file_name":"images/train_atomic-layer-deposition_experimental-usecase_5_fig_10.jpg","caption":"Fig. 10. Densities of TiN ALD films measured using QCM and surface profilometry measurements for long, medium and short reactant pulse sequences at deposition temperatures of $60 - 240^{\\circ}\\mathrm{C}$ .","id":"train/atomic-layer-deposition/experimental-usecase/5/fig_10","sample_id":"atomic-layer-deposition/experimental-usecase/5/fig_10","subset":"multi-axis-chart","split":"train","classification":[{"panel_id":"a","label":"multi-axis chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The figure shows TiN ALD film density and porosity as functions of deposition temperature for short, medium, and long reactant pulse sequences. Film density increases with temperature for all pulse lengths, while porosity decreases correspondingly, indicating densification. Short pulse sequences exhibit the strongest temperature-driven densification, reaching ~3.0 g/cm³ at 240 °C with porosity near ~40%. Medium pulses show moderate densification, whereas long pulses yield comparatively low-density, high-porosity films at intermediate temperatures. Overall, the data indicate a strong coupling between pulse timing and thermal activation of film densification.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Temperature (°C) | Pulse Length | Density (g/cm³) | Porosity (%) |\\n|---|---|---|---|\\n| 60 | Short | 1.8 | 65 |\\n| 60 | Medium | 1.5 | 70 |\\n| 120 | Short | 2.1 | 60 |\\n| 120 | Medium | 1.6 | 70 |\\n| 180 | Short | 2.6 | 50 |\\n| 180 | Medium | 2.0 | 60 |\\n| 180 | Long | 1.2 | 80 |\\n| 240 | Short | 3.0 | 42 |\\n| 240 | Medium | 2.2 | 60 |\\n| 240 | Long | 2.6 | 50 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Increasing temperature leads to higher TiN film density for all pulse lengths. The trend is strongest for short pulses, where density rises sharply with temperature. Medium pulses show a more gradual increase, while long pulses remain relatively low in density at intermediate temperatures. This indicates temperature-activated densification that depends on pulse timing.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Short pulse sequences.\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"-Higher density corresponds to lower porosity\\n-Lower density corresponds to higher porosity\\n-The inverse relationship holds for all pulse lengths\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Yes/No\",\"answer\":\"No\"}]}]","bbox":[{"panel_id":"a","x":1,"y":2,"width":669,"height":620}],"source":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/5/images/fig_10.jpg","provenance":{"source_annotation":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/5/images/fig_10.json","source_content":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/5/content.json","source_pdf":["icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/5/Elam et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"experimental-usecase","paper_id":"5","first_classification_panel_id":"a","first_classification_label":"multi-axis chart","caption_source":"content.json:img_caption"},"width":672,"height":620,"image_format":"jpeg","image_sha256":"e31cb2de14954c0e6c7371dd16a6fe0b582d09233f93f580db3cc762e7a3f2d1","metadata_license":"CC BY 4.0","image_license":"source_publisher_rights_reserved","image_reuse_status":"non_commercial_research_use_only","schema_version":"1.0.0"} | |
| {"file_name":"images/train_atomic-layer-deposition_experimental-usecase_5_fig_4.jpg","caption":"Fig. 4. Mass deposited vs. number of AB cycles measured using the in situ QCM at $120^{\\circ}\\mathrm{C}$ for the long reactant pulses with a timing sequence of (100-10-20-30).","id":"train/atomic-layer-deposition/experimental-usecase/5/fig_4","sample_id":"atomic-layer-deposition/experimental-usecase/5/fig_4","subset":"multi-axis-chart","split":"train","classification":[{"panel_id":"a","label":"multi-axis chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"This figure shows in situ QCM measurements of mass deposited during TiN ALD at 120 °C using long reactant pulses (100–10–20–30). The mass increases approximately linearly with the number of AB cycles, while the raw trace exhibits step-like features corresponding to individual precursor exposures. The equivalent TiN thickness, derived from the mass uptake, also increases linearly with cycle number. Together, these trends indicate stable, cycle-by-cycle growth consistent with self-limiting ALD behavior under long-pulse conditions.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| AB Cycles | Mass Deposited (ng/cm²) | TiN Thickness (Å) |\\n|---|---|---|\\n| 0 | ~0 | ~0 |\\n| 1 | ~120 | ~1.5 |\\n| 3 | ~280 | ~4.5 |\\n| 5 | ~420 | ~7.0 |\\n| 7 | ~560 | ~10.0 |\\n| 10 | ~800 | ~15.0 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Each step corresponds to an individual precursor exposure within an AB cycle, producing a discrete mass increase. This staircase behavior is characteristic of self-limiting surface reactions rather than continuous CVD-like growth.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"It increases approximately linearly.\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Linear mass accumulation indicates that each cycle deposits a nearly constant amount of material. This reflects uniform surface saturation and confirms that the process is operating in a true ALD regime.\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Yes/No\",\"answer\":\"Yes\"}]}]","bbox":[{"panel_id":"a","x":1,"y":0,"width":656,"height":655}],"source":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/5/images/fig_4.jpg","provenance":{"source_annotation":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/5/images/fig_4.json","source_content":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/5/content.json","source_pdf":["icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/5/Elam et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"experimental-usecase","paper_id":"5","first_classification_panel_id":"a","first_classification_label":"multi-axis chart","caption_source":"content.json:img_caption"},"width":664,"height":661,"image_format":"jpeg","image_sha256":"f168a87bfc15c9d7c7eda3a85de7f03dc5a49b3dfbb847b6c872d80086044f88","metadata_license":"CC BY 4.0","image_license":"source_publisher_rights_reserved","image_reuse_status":"non_commercial_research_use_only","schema_version":"1.0.0"} | |
| {"file_name":"images/train_atomic-layer-deposition_experimental-usecase_5_fig_5.jpg","caption":"Fig. 5. (a) Mass deposition rate vs. TDMAT exposure time at $120^{\\circ}\\mathrm{C}$ on an initial TiN surface terminated with $\\mathrm{NH}_x^*$ species. (b) Mass deposition rate vs. $\\mathrm{NH}_3$ exposure time at $120^{\\circ}\\mathrm{C}$ on an initial TiN surface terminated with $\\mathrm{T}_2(\\mathrm{N}(\\mathrm{CH}_3)_2)_x^*$ species.","id":"train/atomic-layer-deposition/experimental-usecase/5/fig_5","sample_id":"atomic-layer-deposition/experimental-usecase/5/fig_5","subset":"multi-axis-chart","split":"train","classification":[{"panel_id":"a","label":"multi-axis chart"},{"panel_id":"b","label":"multi-axis chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"This figure shows ALD saturation behavior for TiN growth at 120 °C by varying the exposure time of each reactant while monitoring mass deposition rate and equivalent thickness growth per cycle. In panel (a), increasing TDMAT exposure time produces a rapid rise in deposition rate at short times followed by a slower approach to a near-saturation regime at longer exposures, consistent with progressively more complete surface reaction of the metal precursor on an NHx*-terminated TiN surface.\"},{\"panel_id\":\"b\",\"text\":\"In panel (b), increasing NH₃ exposure time similarly increases deposition rate and then reaches a clear plateau around ~20 s, indicating saturation of the nitridation/ligand-exchange step on a surface terminated with precursor fragments. The plateau regions demonstrate self-limiting ALD chemistry, while the “short/medium/long” regimes highlight the exposure windows that transition from incomplete reaction to near-saturated growth.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| TDMAT Exposure Time (s) | Mass Deposition Rate (ng/cm²/cycle) | TiN Growth Rate (Å/cycle) |\\n|---|---|---|\\n| 0 | ~0 | ~0.0 |\\n| 5 | ~39 | ~0.8 |\\n| 20 | ~46 | ~0.9 |\\n| 50 | ~52 | ~1.0 |\\n| 100 | ~55 | ~1.05 |\\n| 200 | ~60 | ~1.2 |\"},{\"panel_id\":\"b\",\"text\":\"| NH₃ Exposure Time (s) | Mass Deposition Rate (ng/cm²/cycle) | TiN Growth Rate (Å/cycle) |\\n|---|---|---|\\n| 0 | ~5 | ~0.1 |\\n| 2 | ~41 | ~0.8 |\\n| 5 | ~49 | ~0.9 |\\n| 10 | ~55 | ~1.05 |\\n| 20 | ~63 | ~1.2 |\\n| 40 | ~63 | ~1.2 |\"}]","vqa":"[{\"panel_id\":\"b\",\"items\":[{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The plateau indicates that ligand removal and nitridation reactions have reached completion across the surface. All available precursor-derived fragments have been converted into a stable Ti–N bonding environment. This confirms that the NH₃ half-reaction is self-limiting.\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Knowing the saturation exposure allows selection of the shortest NH₃ pulse that still achieves complete surface reaction. This minimizes cycle time and reagent usage while maintaining consistent growth. Such optimization is critical for scalable and cost-efficient ALD processing.\"}]},{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"At short exposures, reactant delivery is insufficient to fully react with all available surface sites, so growth is limited by incomplete surface coverage. As exposure time increases, more molecules reach the surface until reactive sites become saturated. Once saturation is achieved, additional exposure does not significantly increase growth, producing the characteristic ALD plateau.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"The TiN growth rate increases rapidly at short exposure times and then gradually levels off at longer exposures.\"}]}]","bbox":[{"panel_id":"b","x":0,"y":598,"width":652,"height":601},{"panel_id":"a","x":0,"y":1,"width":652,"height":587}],"source":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/5/images/fig_5.jpg","provenance":{"source_annotation":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/5/images/fig_5.json","source_content":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/5/content.json","source_pdf":["icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/5/Elam et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"experimental-usecase","paper_id":"5","first_classification_panel_id":"a","first_classification_label":"multi-axis chart","caption_source":"content.json:img_caption"},"width":656,"height":1203,"image_format":"jpeg","image_sha256":"b7b5907bb513dbb65b54d4b3302bff4d08bda9343f83135405c2c5ffa24f8625","metadata_license":"CC BY 4.0","image_license":"source_publisher_rights_reserved","image_reuse_status":"non_commercial_research_use_only","schema_version":"1.0.0"} | |
| {"file_name":"images/train_atomic-layer-deposition_experimental-usecase_5_fig_6.jpg","caption":"Fig. 6. Arrhenius plot of TiN ALD mass deposition rates for long, medium and short reactant pulse sequences at deposition temperatures of $60 - 240^{\\circ}\\mathrm{C}$ .","id":"train/atomic-layer-deposition/experimental-usecase/5/fig_6","sample_id":"atomic-layer-deposition/experimental-usecase/5/fig_6","subset":"multi-axis-chart","split":"train","classification":[{"panel_id":"a","label":"multi-axis chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"This figure presents an Arrhenius plot of TiN ALD mass deposition rates for long, medium, and short reactant pulse sequences over a deposition temperature range of 60–240 °C. The x-axis is expressed as 1000/T, and distinct linear segments are annotated with apparent activation energies. The long pulse sequence shows the steepest slope at high temperature, while the short pulse sequence exhibits a weaker temperature dependence across the same range. These trends suggest that pulse duration strongly influences the temperature sensitivity of the ALD growth process.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Pulse sequence | 1000/T (K⁻¹) | Deposition Temperature (°C) | Mass deposition rate (ng/cm²/cycle) |\\n|---|---|---|---|\\n| Long (100-10-20-30) | ~1.95 | 240 | ~900 |\\n| Long (100-10-20-30) | ~2.05 | 210 | ~350 |\\n| Long (100-10-20-30) | ~2.20 | 180 | ~160 |\\n| Long (100-10-20-30) | ~2.40 | 150 | ~110 |\\n| Long (100-10-20-30) | ~2.60 | 120 | ~75 |\\n| Long (100-10-20-30) | ~2.80 | 90 | ~55 |\\n| Long (100-10-20-30) | ~3.00 | 60 | ~40 |\\n| Medium (20-10-6-10) | ~1.95 | 240 | ~300 |\\n| Medium (20-10-6-10) | ~2.05 | 210 | ~180 |\\n| Medium (20-10-6-10) | ~2.20 | 180 | ~120 |\\n| Medium (20-10-6-10) | ~2.40 | 150 | ~85 |\\n| Medium (20-10-6-10) | ~2.60 | 120 | ~60 |\\n| Medium (20-10-6-10) | ~2.80 | 90 | ~45 |\\n| Medium (20-10-6-10) | ~3.00 | 60 | ~30 |\\n| Short (1-0.5-1-5) | ~1.95 | 240 | ~70 |\\n| Short (1-0.5-1-5) | ~2.05 | 210 | ~55 |\\n| Short (1-0.5-1-5) | ~2.20 | 180 | ~45 |\\n| Short (1-0.5-1-5) | ~2.40 | 150 | ~35 |\\n| Short (1-0.5-1-5) | ~2.60 | 120 | ~28 |\\n| Short (1-0.5-1-5) | ~2.80 | 90 | ~22 |\\n| Short (1-0.5-1-5) | ~3.00 | 60 | ~16 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Labeling activation energies highlights the temperature sensitivity of each linear segment without requiring separate fitting plots. It visually emphasizes that different regimes exhibit different slopes, helping the reader identify transitions in growth behavior directly from the figure.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Yes/No\",\"answer\":\"Yes\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"-Long pulses: highest apparent activation energy at high temperature.\\n-Medium pulses: intermediate activation energy.\\n-Short pulses: lowest apparent activation energy.\\n-Shorter pulses appear less temperature sensitive overall.\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"The long pulse sequence.\"}]}]","bbox":[{"panel_id":"a","x":1,"y":5,"width":668,"height":655}],"source":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/5/images/fig_6.jpg","provenance":{"source_annotation":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/5/images/fig_6.json","source_content":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/5/content.json","source_pdf":["icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/5/Elam et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"experimental-usecase","paper_id":"5","first_classification_panel_id":"a","first_classification_label":"multi-axis chart","caption_source":"content.json:img_caption"},"width":670,"height":664,"image_format":"jpeg","image_sha256":"347c179c65e6b4cc65edd0e6f8a53ade6bac2cd682cc1a93ea4966c6d684d4bd","metadata_license":"CC BY 4.0","image_license":"source_publisher_rights_reserved","image_reuse_status":"non_commercial_research_use_only","schema_version":"1.0.0"} | |
| {"file_name":"images/train_atomic-layer-deposition_experimental-usecase_5_fig_7.jpg","caption":"Fig. 7. Mass deposited during one reactant cycle using the long pulse sequence at $120^{\\circ}\\mathrm{C}$ . The quantity $m_0$ is the mass change resulting from TDMAT exposure and purge. The quantity $m_0 + m_1$ is the net mass change resulting from one complete TiN ALD cycle.","id":"train/atomic-layer-deposition/experimental-usecase/5/fig_7","sample_id":"atomic-layer-deposition/experimental-usecase/5/fig_7","subset":"multi-axis-chart","split":"train","classification":[{"panel_id":"a","label":"multi-axis chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"This figure shows in situ QCM mass changes during a single TiN ALD cycle at 120 °C using long reactant pulses. During TDMAT exposure and the first purge, the mass increases by m₀ due to precursor adsorption. Subsequent NH₃ exposure produces an additional mass change, and the final mass after the second purge corresponds to m₀+m₁, the net per-cycle gain. The stepwise features confirm that mass uptake occurs in discrete, self-limiting half-reactions.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Time (s) | Mass Deposited (ng/cm²) | TiN Thickness (Å) | Stage |\\n|---|---|---|---|\\n| 0 | ~0 | ~0.0 | Start |\\n| 100 | ~95 | ~1.6 | End TDMAT exposure |\\n| 110 | ~90 | ~1.5 | End first purge (m₀) |\\n| 130 | ~165 | ~3.0 | End NH₃ exposure |\\n| 160 | ~75 | ~1.3 | End second purge (m₀ + m₁) |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"m₀ represents the mass gained during the TDMAT exposure and subsequent purge. It reflects adsorption of TDMAT-derived species on the surface prior to nitridation.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"During the NH₃ exposure.\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Separating these quantities isolates the contributions of precursor adsorption and nitridation/ligand removal. This helps identify which half-reaction controls mass uptake and confirms self-limiting behavior within the cycle.\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Yes/No\",\"answer\":\"Yes\"}]}]","bbox":[{"panel_id":"a","x":0,"y":3,"width":653,"height":598}],"source":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/5/images/fig_7.jpg","provenance":{"source_annotation":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/5/images/fig_7.json","source_content":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/5/content.json","source_pdf":["icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/5/Elam et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"experimental-usecase","paper_id":"5","first_classification_panel_id":"a","first_classification_label":"multi-axis chart","caption_source":"content.json:img_caption"},"width":655,"height":603,"image_format":"jpeg","image_sha256":"8e28948e43eba96d95d2639b5e7a3bd21dada87252f21da6656b1c68454a7b51","metadata_license":"CC BY 4.0","image_license":"source_publisher_rights_reserved","image_reuse_status":"non_commercial_research_use_only","schema_version":"1.0.0"} | |
| {"file_name":"images/train_atomic-layer-deposition_experimental-usecase_52_figure_6.jpg","caption":"Figure 6. Number of atoms deposited per $\\mathrm{nm}^2$ per super-cycle and the Co/Li ratio in the film as a function of the Co/Li dosing ratio $(x)$ as determined by Nuclear reaction analysis (NRA) and Elastic backscattering spectroscopy (EBS). For the Co/Li dosing ratio the amount of Li cycles per super-cycle is kept constant at 1.","id":"train/atomic-layer-deposition/experimental-usecase/52/figure_6","sample_id":"atomic-layer-deposition/experimental-usecase/52/figure_6","subset":"multi-axis-chart","split":"train","classification":[{"panel_id":"a","label":"multi-axis chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The chart displays the number of atoms per nm² supercycle for Li NRA, Li RBS, and Co RBS measurements, along with the Co/Li atomic ratio, against the Co/Li dosing ratio\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Co/Li dosing ratio (x) | Li NRA (atoms/nm² supercycle) | Li RBS (atoms/nm² supercycle) | Co RBS (atoms/nm² supercycle) | Co/Li atomic ratio |\\n|---|---|---|---|---|\\n| 2 | 5.0 | 4.9 | 1.6 | 0.35 |\\n| 3 | 5.0 | 4.7 | 2.7 | 0.50 |\\n| 4 | 4.8 | 4.5 | 4.0 | 0.80 |\\n| 5 | 5.0 | 6.0 | 4.9 | 0.85 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"The amount of Li cycles per super cycle is kept constant at 1\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The Co amount per super cycle increases as x increases, and the paper notes it continues to increase roughly linearly with x. The Li amount per super cycle is much less sensitive to x, but the text notes a slight increase in the Li sub cycle deposition rate at higher Co to Li dosing ratios. Together these trends drive the Co to Li atomic ratio upward with x until it approaches a plateau.\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"The red squares and red line labeled Co to Li ratio correspond to the Co to Li atomic ratio, and they are read from the right y axis.\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The paper states that knowing the overall film composition is essential to evaluate ALD deposited LiCoO2 thin films and to properly interpret the electrochemical results. Changing the Co to Li dosing ratio changes the atomic ratios in the film, so composition must be tracked rather than assumed from the process recipe. This figure provides that link between the ALD dosing ratio and the resulting film composition.\"}]}]","bbox":[{"panel_id":"a","x":0,"y":2,"width":670,"height":441}],"source":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/52/images/figure_6.jpg","provenance":{"source_annotation":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/52/images/figure_6.json","source_content":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/52/content.json","source_pdf":["icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/52/Donders et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"experimental-usecase","paper_id":"52","first_classification_panel_id":"a","first_classification_label":"multi-axis chart","caption_source":"content.json:image_caption"},"width":675,"height":448,"image_format":"jpeg","image_sha256":"62773e8f76a25b0d5ba81402dfcc1d6dae4a8b8e01afa7ff1d01a83bb0725692","metadata_license":"CC BY 4.0","image_license":"source_publisher_rights_reserved","image_reuse_status":"non_commercial_research_use_only","schema_version":"1.0.0"} | |
| {"file_name":"images/train_atomic-layer-deposition_experimental-usecase_54_figure_8.jpg","caption":"Figure 8. (Color online) Peak intensity and crystallite size (derived using Scherrer's equation) of the $\\mathrm{Cu(111)}$ peak in the XRD spectra for a $\\mathrm{Cu - }$ TaN-Si (Fig. 6) and b) $\\mathrm{Cu - }$ TiN-Si (Fig. 7) stacks annealed at different temperatures.","id":"train/atomic-layer-deposition/experimental-usecase/54/figure_8","sample_id":"atomic-layer-deposition/experimental-usecase/54/figure_8","subset":"multi-axis-chart","split":"train","classification":[{"panel_id":"a","label":"multi-axis chart"},{"panel_id":"b","label":"multi-axis chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The chart shows the intensity of Cu(111) and crystallite size at different anneal temperatures for a Cu-TaN-Si stack.\"},{\"panel_id\":\"b\",\"text\":\"The chart shows the intensity of Cu(111) and crystallite size at different anneal temperatures for a Cu-TiN-Si stack.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Anneal temperature (°C) | Intensity Cu(111) (counts/s) | Crystallite size (nm) |\\n|---|---|---|\\n| 200 | 100 | 30 |\\n| 400 | 135 | 30 |\\n| 600 | 145 | 30 |\\n| 650 | 100 | 30 |\\n| 700 | 70 | 30 |\"},{\"panel_id\":\"b\",\"text\":\"| Anneal temperature (°C) | Intensity Cu(111) (counts/s) | Crystallite size (nm) |\\n|---|---|---|\\n| 200 | 50 | 30 |\\n| 300 | 70 | 30 |\\n| 400 | 70 | 30 |\\n| 500 | 80 | 30 |\\n| 600 | 90 | 30 |\\n| 700 | 150 | 40 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"200, 400, 600, 650 and 700 °C.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Around 145 counts/s.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Initially, as the temperature increases, so too does the intensity of the Cu(111) peak. This can be related to the growth of the copper crystalline phase and also means that the resistance of the sheet decreases. Above 600 °C, the intensity of the Cu(111) peak rapidly decreases. At this point, copper diffuses through the film into the Si layer and Cu3Si is formed. This also causes a significant increase in film resistance.\"}]},{\"panel_id\":\"b\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"200, 300, 400, 500, 600 and 700 °C.\"}]}]","bbox":[{"panel_id":"a","x":0,"y":4,"width":672,"height":470},{"panel_id":"b","x":1,"y":490,"width":667,"height":445}],"source":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/54/images/figure_8.jpg","provenance":{"source_annotation":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/54/images/figure_8.json","source_content":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/54/content.json","source_pdf":["icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/54/Knoops et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"experimental-usecase","paper_id":"54","first_classification_panel_id":"a","first_classification_label":"multi-axis chart","caption_source":"content.json:image_caption"},"width":675,"height":953,"image_format":"jpeg","image_sha256":"df39c239ee40c7b8487eb6cfda56611592692e1868ea43a2a7d9d17dbefa8b11","metadata_license":"CC BY 4.0","image_license":"source_publisher_rights_reserved","image_reuse_status":"non_commercial_research_use_only","schema_version":"1.0.0"} | |
| {"file_name":"images/train_atomic-layer-deposition_experimental-usecase_63_fig_5.jpg","caption":"FIG. 5. (Color online) Optical constants of as-deposited $(250^{\\circ}\\mathrm{C})$ and annealed $\\sim 26 \\mathrm{nm}$ thick $\\mathrm{Ga}_{2}\\mathrm{O}_{3}$ thin films.","id":"train/atomic-layer-deposition/experimental-usecase/63/fig_5","sample_id":"atomic-layer-deposition/experimental-usecase/63/fig_5","subset":"multi-axis-chart","split":"train","classification":[{"panel_id":"a","label":"multi-axis chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The chart displays the refractive index and extinction coefficient of a material as a function of wavelength, comparing 'as-deposited' and 'annealed' conditions.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Wavelength, nm | Refractive index n (as-deposited) | Refractive index n (annealed) | Extinction coefficient k (as-deposited) | Extinction coefficient k (annealed) |\\n|---|---|---|---|---|\\n| 300 | 2.08 | 2.09 | 0.025 | 0.000 |\\n| 350 | 1.98 | 2.02 | 0.005 | 0.000 |\\n| 400 | 1.93 | 1.97 | 0.000 | 0.000 |\\n| 450 | 1.91 | 1.95 | 0.000 | 0.000 |\\n| 500 | 1.89 | 1.93 | 0.000 | 0.000 |\\n| 600 | 1.88 | 1.92 | 0.000 | 0.000 |\\n| 700 | 1.87 | 1.92 | 0.000 | 0.000 |\\n| 800 | 1.86 | 1.92 | 0.000 | 0.000 |\\n| 900 | 1.86 | 1.92 | 0.000 | 0.000 |\\n| 1000 | 1.85 | 1.92 | 0.000 | 0.000 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"List\",\"answer\":\"n is plotted in black using the left y axis.\\n\\nk is plotted in blue using the right y axis.\\n\\nSolid lines indicate as deposited, dashed lines indicate annealed.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"The annealed film has a higher refractive index than the as deposited film across the full 300 to 1000 nm range.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"For the as deposited film, k is highest at the shortest wavelengths near 300 nm and drops quickly toward near zero as wavelength increases. For the annealed film, k stays close to zero across the plotted range. Overall, absorption is concentrated near the short wavelength edge and is smaller after annealing.\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"After annealing, the refractive index increases relative to the as deposited film, which the authors attribute to structural enhancement after crystallization. A higher n is consistent with a denser and more ordered film compared with an amorphous state. In this dataset, annealing changes n without a large increase in k across most of the visible range, so the film can become optically denser while remaining largely transparent at longer wavelengths.\"}]}]","bbox":[{"panel_id":"a","x":3,"y":5,"width":642,"height":423}],"source":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/63/images/fig_5.jpg","provenance":{"source_annotation":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/63/images/fig_5.json","source_content":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/63/content.json","source_pdf":["icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/63/Donmez et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"experimental-usecase","paper_id":"63","first_classification_panel_id":"a","first_classification_label":"multi-axis chart","caption_source":"content.json:image_caption"},"width":644,"height":431,"image_format":"jpeg","image_sha256":"67b8d98be77b97e91c7d96b9213eef2de77600cfee6132455bbe743269674892","metadata_license":"CC BY 4.0","image_license":"source_publisher_rights_reserved","image_reuse_status":"non_commercial_research_use_only","schema_version":"1.0.0"} | |
| {"file_name":"images/train_atomic-layer-deposition_experimental-usecase_69_FIG2_a.jpg","caption":"","id":"train/atomic-layer-deposition/experimental-usecase/69/FIG2_a","sample_id":"atomic-layer-deposition/experimental-usecase/69/FIG2_a","subset":"multi-axis-chart","split":"train","classification":[{"panel_id":"a","label":"multi-axis chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The chart displays the relationship between the atomic ratio of P/(P + O + Al) and GPC (nm/cycle) as a function of TDMAP exposure time (in seconds).\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| O - T<sub>Al</sub> - O | GPC (nm/cycle) | P/(P + O + Al) atomic ratio |\\n|---|---|---|\\n| 0 | 0.12 | 0.00 |\\n| 5 | 0.10 | |\\n| 10 | 0.10 | 0.14 |\\n| 15 | 0.12 | |\\n| 20 | | 0.15 |\\n| 30 | 0.13 | 0.15 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"List\",\"answer\":\"15 s TDMAP; 10 s first O2 plasma; 15s TMA; 10 s second O2 plasma.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"0.15.\"}]},{\"panel_id\":\"d\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"After about 10 seconds of TMA exposure, the GPC plateaus at a value of 0.12 nm/cycle.\"}]},{\"panel_id\":\"b\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"After about 2 seconds of O2 plasma exposure, the GPC plateaus at a value of 12 nm/cycle.\"}]}]","bbox":[{"panel_id":"a","x":4,"y":0,"width":755,"height":318}],"source":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/69/images/FIG2_a.jpg","provenance":{"source_annotation":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/69/images/FIG2_a.json","source_content":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/69/content.json","source_pdf":["icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/69/Ruben Blomme et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"experimental-usecase","paper_id":"69","first_classification_panel_id":"a","first_classification_label":"multi-axis chart","caption_source":"not_found"},"width":761,"height":320,"image_format":"jpeg","image_sha256":"6187c65a593c1bae5aed559873e272998f132325207360ddda0c2ff37d7eaf17","metadata_license":"CC BY 4.0","image_license":"source_publisher_rights_reserved","image_reuse_status":"non_commercial_research_use_only","schema_version":"1.0.0"} | |
| {"file_name":"images/train_atomic-layer-deposition_experimental-usecase_69_FIG3_a.jpg","caption":"","id":"train/atomic-layer-deposition/experimental-usecase/69/FIG3_a","sample_id":"atomic-layer-deposition/experimental-usecase/69/FIG3_a","subset":"multi-axis-chart","split":"train","classification":[{"panel_id":"a","label":"multi-axis chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The chart displays the relationship between average GPC (nanometers per cycle) and the P/(P+O+Al) atomic fraction as a function of temperature (°C).\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Temperature (°C) | Average GPC (nm/cycle) | P/(P+O+Al) atomic fraction |\\n|---|---|---|\\n| 125 | 0.20 | 0.17 |\\n| 150 | 0.18 | |\\n| 175 | 0.17 | 0.15 |\\n| 200 | 0.17 | |\\n| 225 | 0.13 | 0.14 |\\n| 250 | 0.13 | 0.15 |\\n| 275 | 0.11 | 0.14 |\\n| 300 | 0.12 | 0.13 |\\n| 325 | 0.10 | 0.15 |\\n| 350 | 0.07 | 0.13 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"25.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"It stays relatively constant around a value of 0.15.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Figure 3(a) demonstrates that increasing substrate temperature leads to a decrease in GPC and in the thickness contribution from each individual precursor pulse. The GPC decreases from approximately 0.20 nm per cycle at 125 °C to about 0.08 nm per cycle at 350 °C, whilst XPS analysis indicates that the phosphorus atomic fraction remains essentially independent of temperature. Figure 3(b) further shows that the average thickness addition per pulse for TDMAP, TMA, and the O₂* steps all decline with increasing temperature, consistent with the overall GPC trend. The marginal increase in GPC observed during extended TDMAP exposure at low temperature confirms that precursor condensation is not responsible for the higher growth rate at 125 °C. Overall, the figure indicates that reduced surface reactivity at elevated temperatures, likely due to desorption of reactive surface groups, governs the observed decrease in growth rate.\"}]},{\"panel_id\":\"b\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Around 0.2 nm.\"}]}]","bbox":[{"panel_id":"a","x":0,"y":3,"width":671,"height":442}],"source":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/69/images/FIG3_a.jpg","provenance":{"source_annotation":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/69/images/FIG3_a.json","source_content":"icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/69/content.json","source_pdf":["icdar2026-competition-data/train/atomic-layer-deposition/experimental-usecase/69/Ruben Blomme et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"experimental-usecase","paper_id":"69","first_classification_panel_id":"a","first_classification_label":"multi-axis chart","caption_source":"not_found"},"width":675,"height":447,"image_format":"jpeg","image_sha256":"186beff4cfbde0673e3e4afe2f3d9c27e28e62b801e7672e33fe042a29962f5a","metadata_license":"CC BY 4.0","image_license":"source_publisher_rights_reserved","image_reuse_status":"non_commercial_research_use_only","schema_version":"1.0.0"} | |
| {"file_name":"images/train_atomic-layer-deposition_simulation-usecase_26_figure_12.jpg","caption":"Figure 12. (A) Proton population in film and number of events for 20 cycles. Pulse and purge time have been both fixed at $0.1~\\mathrm{ms}$ and temperature is $500~\\mathrm{K}$ Proton population and number of events increase constantly due to the accumulation of protons in the subsurface layers. (B) (a) Start of metal pulse (cycle), (b) start of purge, (c) start of oxygen pulse, (d) start of purge, and (e) end of cycle. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]","id":"train/atomic-layer-deposition/simulation-usecase/26/figure_12","sample_id":"atomic-layer-deposition/simulation-usecase/26/figure_12","subset":"multi-axis-chart","split":"train","classification":[{"panel_id":"a","label":"multi-axis chart"},{"panel_id":"b","label":"line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The chart shows two time-series: events/μs (red, left axis) and total protons in film/cell (green, right axis). Both signals oscillate in a repeating cycle pattern and generally increase with time, with higher event rates occurring alongside higher proton counts.\"},{\"panel_id\":\"b\",\"text\":\"The chart zooms into one cycle (0.0024–0.0028 s) and marks step transitions (metal pulse, purge, oxygen pulse, purge, end). The events/μs drops to a minimum after the metal pulse and then rises to a peak after the oxygen pulse.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Time (s) | Number of events/μs | Total protons in film/cell |\\n|---|---|---|\\n| 0.0000 | 275 | 27 |\\n| 0.0003 | 984 | 1239 |\\n| 0.0005 | 300 | 183 |\\n| 0.0007 | 1406 | 1842 |\\n| 0.0010 | 471 | 543 |\\n| 0.0011 | 1547 | 1799 |\\n| 0.0014 | 383 | 419 |\\n| 0.0015 | 1714 | 1979 |\\n| 0.0018 | 449 | 481 |\\n| 0.0019 | 1884 | 2122 |\\n| 0.0022 | 471 | 568 |\\n| 0.0023 | 2235 | 2234 |\\n| 0.0025 | 748 | 1047 |\\n| 0.0027 | 2441 | 2414 |\\n| 0.0029 | 772 | 947 |\\n| 0.0031 | 2403 | 2476 |\\n| 0.0033 | 863 | 1090 |\\n| 0.0035 | 2809 | 2954 |\\n| 0.0037 | 1057 | 1464 |\\n| 0.0039 | 2982 | 3111 |\\n| 0.0041 | 1107 | 1451 |\\n| 0.0043 | 3435 | 3681 |\\n| 0.0045 | 1595 | 2146 |\\n| 0.0047 | 3591 | 3644 |\\n| 0.0049 | 1736 | 2084 |\\n| 0.0051 | 3653 | 3483 |\\n| 0.0053 | 1827 | 2146 |\\n| 0.0055 | 3774 | 3799 |\\n| 0.0057 | 1930 | 2488 |\\n| 0.0059 | 3823 | 3893 |\\n| 0.0061 | 1975 | 2264 |\\n| 0.0063 | 4057 | 4315 |\\n| 0.0065 | 2406 | 2905 |\\n| 0.0068 | 4232 | 4533 |\\n| 0.0069 | 2554 | 3190 |\\n| 0.0071 | 4361 | 4496 |\\n| 0.0073 | 2709 | 3296 |\\n| 0.0075 | 4517 | 4719 |\\n| 0.0077 | 2850 | 3209 |\\n| 0.0079 | 4907 | 4968 |\"},{\"panel_id\":\"b\",\"text\":\"| Step | Time (s) | Number of events/μs |\\n|---|---|---|\\n| Start of metal pulse (cycle) | 0.0024 | 2018 |\\n| Start of purge | 0.0025 | 789 |\\n| Start of oxygen pulse | 0.0026 | 768 |\\n| Start of purge | 0.0027 | 2453 |\\n| End of cycle | 0.0028 | 2249 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Yes/No\",\"answer\":\"Yes\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"20 cycles.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The plot indicates net proton accumulation in the film over time. The proton trace shows a repeating up–down pattern each cycle, but its overall level shifts upward across the time window. That upward drift means any proton losses during parts of the cycle are smaller than the gains, so the net balance is positive.\"}]},{\"panel_id\":\"b\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The event rate is high at the start of the metal pulse (~2018 events/μs) and then drops sharply by the start of the first purge to ~789 events/μs. After the oxygen pulse, the event rate jumps to a peak (~2453 events/μs). By the end of the cycle it stays high but drops slightly (~2249 events/μs).\"}]}]","bbox":[{"panel_id":"a","x":2,"y":0,"width":653,"height":459},{"panel_id":"b","x":152,"y":472,"width":362,"height":528}],"source":"icdar2026-competition-data/train/atomic-layer-deposition/simulation-usecase/26/images/figure_12.jpg","provenance":{"source_annotation":"icdar2026-competition-data/train/atomic-layer-deposition/simulation-usecase/26/images/figure_12.json","source_content":"icdar2026-competition-data/train/atomic-layer-deposition/simulation-usecase/26/content.json","source_pdf":["icdar2026-competition-data/train/atomic-layer-deposition/simulation-usecase/26/Mahdi Shirazi et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"simulation-usecase","paper_id":"26","first_classification_panel_id":"a","first_classification_label":"multi-axis chart","caption_source":"content.json:img_caption"},"width":655,"height":1003,"image_format":"jpeg","image_sha256":"c0b1646d4591e2737a29189047d818cac19847ebeefe20b2cbd9bb29d031cb6a","metadata_license":"CC BY 4.0","image_license":"source_publisher_rights_reserved","image_reuse_status":"non_commercial_research_use_only","schema_version":"1.0.0"} | |
| {"file_name":"images/train_atomic-layer-deposition_simulation-usecase_28_fig_3.jpg","caption":"Fig. 3. Temperature dependence of the film growth rate. Squares and circles are experimental thickness and mass increments per ALD cycle in $\\mathrm{ZrO_2}$ film growth [10]. Curves are the hydroxylation degree calculated by Eq. (6) for $E_{\\mathrm{ad}} = 25$ kcal/mol (dashed-dotted line), $E_{\\mathrm{ad}} = 35$ kcal/mol (dashed line) and $45\\mathrm{kcal / mol}$ (dashed-dotted line).","id":"train/atomic-layer-deposition/simulation-usecase/28/fig_3","sample_id":"atomic-layer-deposition/simulation-usecase/28/fig_3","subset":"multi-axis-chart","split":"train","classification":[{"panel_id":"a","label":"multi-axis chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The chart shows the relationship between temperature (T., °C) and mass/thickness increment per cycle (a.u.) and hydroxylation degree (a.u.). The mass/thickness increment decreases with increasing temperature, while the hydroxylation degree shows a profile starting at one, gradually decreasing to zero after a threshold.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| T., °C | Mass, thickness increment per cycle, a.u. | Hydroxylation degree, a.u. Left |Hydroxylation degree, a.u. Middle |Hydroxylation degree, a.u. Right |\\n|---|---|---|---|---|\\n| 100 | 1.00 | 1.00 | 1.00 | 1.00 |\\n| 200 | 0.79 | 0.17 | 1.00 | 1.00 |\\n| 300 | 0.67 | 0.00 | 1.00 | 1.00 |\\n| 400 | 0.60 | 0.00 | 0.17 | 1.00 |\\n| 500 | 0.57 | 0.00 | 0.00 | 0.83 |\\n| 600 | 0.54 | 0.00 | 0.00 | 0.17 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The reason for this threshold is that there is a certain activation energy that has been set for the calculations. Only after this energy barrier has been met by the temperature, there will be a drop in hydroxynation level.\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Less Cl is removed from the surface. Cl is heavier than oxygen so it is likely that the film will be slightly havier if there is, for the same thickness, more Cl.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"0.61\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"As long as the increase in temperature only influences the process and not the sample itself it is likely to saturate the mass increment per cycle.\"}]}]","bbox":[{"panel_id":"a","x":3,"y":4,"width":590,"height":442}],"source":"icdar2026-competition-data/train/atomic-layer-deposition/simulation-usecase/28/images/fig_3.jpg","provenance":{"source_annotation":"icdar2026-competition-data/train/atomic-layer-deposition/simulation-usecase/28/images/fig_3.json","source_content":"icdar2026-competition-data/train/atomic-layer-deposition/simulation-usecase/28/content.json","source_pdf":["icdar2026-competition-data/train/atomic-layer-deposition/simulation-usecase/28/Maxim Deminsky et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"simulation-usecase","paper_id":"28","first_classification_panel_id":"a","first_classification_label":"multi-axis chart","caption_source":"content.json:img_caption"},"width":594,"height":444,"image_format":"jpeg","image_sha256":"c66b29bc7913760ea3f96bcf1897e8cba924ca9c106f041617b163490f4c48ca","metadata_license":"CC BY 4.0","image_license":"source_publisher_rights_reserved","image_reuse_status":"non_commercial_research_use_only","schema_version":"1.0.0"} | |
| {"file_name":"images/train_atomic-layer-deposition_simulation-usecase_55_figure_7.jpg","caption":"Figure 7.Surface $-\\mathrm{OH}$ concentrations and roughness of the $\\mathrm{Al}_2\\mathrm{O}_3$ film at $300^{\\circ}\\mathrm{C}$ as a function of ALD cycles.","id":"train/atomic-layer-deposition/simulation-usecase/55/figure_7","sample_id":"atomic-layer-deposition/simulation-usecase/55/figure_7","subset":"multi-axis-chart","split":"train","classification":[{"panel_id":"a","label":"multi-axis chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The figure presents the evolution of surface –OH concentration and surface roughness of an Al2O3 film during ALD at 300 °C as a function of cycle number. The –OH concentration increases progressively with ALD cycles, with a pronounced acceleration at later stages. Surface roughness also increases with cycle number, but shows a slower rise and partial saturation compared to the –OH concentration. This indicates that chemical surface activation and morphological roughening evolve at different rates during film growth.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| ALD cycles | –OH concentration (nm⁻²) | Roughness (Å) |\\n|---|---|---|\\n| 1 | 0.20 | 0.95 |\\n| 2 | 0.25 | 1.00 |\\n| 3 | 0.40 | 1.10 |\\n| 4 | 0.60 | 1.70 |\\n| 5 | 0.90 | 2.15 |\\n| 6 | 1.10 | 2.30 |\\n| 7 | 1.50 | 2.60 |\\n| 8 | 1.90 | 3.00 |\\n| 9 | 2.00 | 2.85 |\\n| 10 | 2.50 | 3.10 |\\n| 11 | 2.80 | 3.10 |\\n| 12 | 3.40 | 3.20 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"300 °C.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Approximately a factor of 17 (from ~0.2 to ~3.35 nm⁻²).\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Yes/No\",\"answer\":\"No\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The figure suggests that ALD can achieve high surface reactivity without inducing excessive roughness if growth is stopped before roughness saturation. This allows optimization of cycle count to maximize –OH availability for subsequent reactions while maintaining acceptable surface smoothness. Such insight is valuable for applications requiring both reactive and morphologically controlled oxide films.\"}]}]","bbox":[{"panel_id":"a","x":3,"y":2,"width":634,"height":512}],"source":"icdar2026-competition-data/train/atomic-layer-deposition/simulation-usecase/55/images/figure_7.jpg","provenance":{"source_annotation":"icdar2026-competition-data/train/atomic-layer-deposition/simulation-usecase/55/images/figure_7.json","source_content":"icdar2026-competition-data/train/atomic-layer-deposition/simulation-usecase/55/content.json","source_pdf":["icdar2026-competition-data/train/atomic-layer-deposition/simulation-usecase/55/Zheng Hu et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"simulation-usecase","paper_id":"55","first_classification_panel_id":"a","first_classification_label":"multi-axis chart","caption_source":"content.json:image_caption"},"width":642,"height":517,"image_format":"jpeg","image_sha256":"516353b11abc4b8ac124dae5a4cf2db77dd196e454ef8662035d5ca626d435e6","metadata_license":"CC BY 4.0","image_license":"source_publisher_rights_reserved","image_reuse_status":"non_commercial_research_use_only","schema_version":"1.0.0"} | |
| {"file_name":"images/train_atomic-layer-deposition_simulation-usecase_55_figure_8.jpg","caption":"Figure 8. Growth per cycle in terms of mass (red) and average height (blue) as a function of ALD cycles during the $\\mathrm{Al}_2\\mathrm{O}_3$ ALD at $300^{\\circ}\\mathrm{C}$ .","id":"train/atomic-layer-deposition/simulation-usecase/55/figure_8","sample_id":"atomic-layer-deposition/simulation-usecase/55/figure_8","subset":"multi-axis-chart","split":"train","classification":[{"panel_id":"a","label":"multi-axis chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The figure presents the growth per ALD cycle of an Al2O3 film at 300 °C expressed in terms of deposited mass and average height. Both quantities increase overall with ALD cycles, but exhibit noticeable fluctuations. The height-based growth shows stronger oscillations compared to mass-based growth, indicating differences between local thickness variation and overall material uptake during deposition.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| ALD cycles | Growth per cycle (10⁻⁹ g/cm²)-Mass | Growth per cycle (Å)-Height |\\n|---|---|---|\\n| 1 | 1.2 | 0.07 |\\n| 2 | 1.0 | 0.05 |\\n| 3 | 1.9 | 0.10 |\\n| 4 | 2.0 | 0.12 |\\n| 5 | 2.4 | 0.14 |\\n| 6 | 3.0 | 0.20 |\\n| 7 | 4.0 | 0.18 |\\n| 8 | 5.2 | 0.30 |\\n| 9 | 6.0 | 0.33 |\\n| 10 | 7.0 | 0.38 |\\n| 11 | 6.0 | 0.35 |\\n| 12 | 7.4 | 0.40 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Yes/No\",\"answer\":\"No\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Height-based growth.\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"List\",\"answer\":\"- Non-uniform surface coverage\\n- Local roughness evolution\\n- Island growth or restructuring\\n- Differences between average thickness and total material uptake\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Using both metrics provides complementary insight into ALD growth behavior. Mass-based growth reflects the total amount of material deposited per cycle, while height-based growth captures how that material translates into film thickness and roughness. Together, they help distinguish between uniform layer-by-layer growth and morphological effects such as roughening or density changes.\"}]}]","bbox":[{"panel_id":"a","x":2,"y":0,"width":632,"height":500}],"source":"icdar2026-competition-data/train/atomic-layer-deposition/simulation-usecase/55/images/figure_8.jpg","provenance":{"source_annotation":"icdar2026-competition-data/train/atomic-layer-deposition/simulation-usecase/55/images/figure_8.json","source_content":"icdar2026-competition-data/train/atomic-layer-deposition/simulation-usecase/55/content.json","source_pdf":["icdar2026-competition-data/train/atomic-layer-deposition/simulation-usecase/55/Zheng Hu et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"simulation-usecase","paper_id":"55","first_classification_panel_id":"a","first_classification_label":"multi-axis chart","caption_source":"content.json:image_caption"},"width":636,"height":500,"image_format":"jpeg","image_sha256":"939a359145aeba59801a368883a7a26de2e43fc1904579c77608922a0ae96fa7","metadata_license":"CC BY 4.0","image_license":"source_publisher_rights_reserved","image_reuse_status":"non_commercial_research_use_only","schema_version":"1.0.0"} | |
| {"file_name":"images/train_atomic-layer-etching_experimental-usecase_14_figure_12.jpg","caption":"Figure 12. Si ALE using $\\mathrm{O}_2$ , $\\mathrm{HF}$ , and $\\mathrm{Al(CH_3)_3}$ as reactants at $290^{\\circ}\\mathrm{C}$ . Si etching occurs with an etch rate of $0.4 \\mathrm{\\AA}$ cycle, while the top $\\mathrm{SiO}_2$ layer remains constant at $\\sim 11 \\mathrm{\\AA}$ . Reproduced with permission from ref 37. Copyright 2018 American Chemical Society.","id":"train/atomic-layer-etching/experimental-usecase/14/figure_12","sample_id":"atomic-layer-etching/experimental-usecase/14/figure_12","subset":"multi-axis-chart","split":"train","classification":[{"panel_id":"a","label":"multi-axis chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The chart shows a multi-axis chart with the change in silicon thickness and the SiO2 thickness over cycles of Si ALE at 290°C with O₂, 250 Torr, HF, 1 Torr, TMA, 1 Torr, and O₂-HF-TMA.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Si ALE cycles | Si Thickness Change (Å) | SiO2 Thickness (Å) |\\n| --- | --- | --- |\\n| 0 | 0 | 15 |\\n| 20 | -8 | 10 |\\n| 40 | -16 | 10 |\\n| 60 | -24 | 10 |\\n| 80 | -30 | 10 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"The experiment was perfomed at 290 °C, with O₂, 250 Torr, HF, 1 Torr, TMA, 1 Torr, and O₂-HF-TMA cycles.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"The thickness of the Si decreases over the Si ALE cycles. The thickness of the SiO2 decreases but reaches a plateau at approximately 20 ALE cycles.\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Approximately 10 Å, after 20 ALE cycles.\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"As can be seen in the figure, this ALE process reaches a etch per cycle of 0.4 Å/cycle.\"}]}]","bbox":[{"panel_id":"a","x":5,"y":3,"width":657,"height":501}],"source":"icdar2026-competition-data/train/atomic-layer-etching/experimental-usecase/14/images/figure_12.jpg","provenance":{"source_annotation":"icdar2026-competition-data/train/atomic-layer-etching/experimental-usecase/14/images/figure_12.json","source_content":"icdar2026-competition-data/train/atomic-layer-etching/experimental-usecase/14/content.json","source_pdf":["icdar2026-competition-data/train/atomic-layer-etching/experimental-usecase/14/Mechanisms of Thermal Atomic Layer Etching.pdf"],"main_category":"atomic-layer-etching","sub_category":"experimental-usecase","paper_id":"14","first_classification_panel_id":"a","first_classification_label":"multi-axis chart","caption_source":"content.json:img_caption"},"width":664,"height":506,"image_format":"jpeg","image_sha256":"3bda732fe981d5b8db9bd9185e0f53753f42b9d1ffc4323c376abb336e2748f7","metadata_license":"CC BY 4.0","image_license":"source_publisher_rights_reserved","image_reuse_status":"non_commercial_research_use_only","schema_version":"1.0.0"} | |
| {"file_name":"images/train_atomic-layer-etching_experimental-usecase_23_figure_10.jpg","caption":"Figure 10. Time-resolved QMS results during the fourth TMA exposure in Figure 9. (a) Ion intensities for the $^{68}\\mathrm{Zn(CH_3)_2}$ conversion product at $m / z$ 83 and 98 appear rapidly in advance of the ion intensity for the TMA reactant at $m / z$ 72. (b) Intensity for the $\\mathrm{Al}_2(\\mathrm{CH}_3)_4\\mathrm{F}^+$ ion signal at $m / z$ 133 appears rapidly in advance of the ion intensity for the TMA reactant at $m / z$ 72.","id":"train/atomic-layer-etching/experimental-usecase/23/figure_10","sample_id":"atomic-layer-etching/experimental-usecase/23/figure_10","subset":"multi-axis-chart","split":"train","classification":[{"panel_id":"a","label":"multi-axis chart"},{"panel_id":"b","label":"multi-axis chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The multi axis chart shows time-resolved QMS ion intensities during the TMA pulse. Volatile zinc byproducts, ⁶⁸Zn(CH₃)⁺ (m/z 83) and ⁶⁸Zn(CH₃)₂⁺ (m/z 98), appear immediately at the start of the pulse and decay rapidly, well before the TMA signal (m/z 72) saturates. This indicates a fast, self-limiting surface reaction that releases dimethylzinc.\"},{\"panel_id\":\"b\",\"text\":\"The multi axis bottom chart shows the aluminum-fluoride dimer fragment Al₂(CH₃)₄F⁺ (m/z 133) relative to the TMA signal (m/z 72). The product signal spikes instantly at the beginning of the TMA exposure, confirming rapid removal of surface fluorine through volatile dimethylaluminum fluoride formation, occurring concurrently with zinc removal.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Time (min) | TMA Intensity (mV) | Product Intensity (mV) | Material |\\n|------------|---------------------|--------------------------|------------------------------|\\n| 38 | 0 | | TMA, m/z 72 |\\n| 38 | | 5.84 | ⁶⁸Zn(CH₃)⁺, m/z 83 |\\n| 38 | | 1.62 | ⁶⁸Zn(CH₃)₂⁺, m/z 98 |\\n| 39 | 52.87 | | TMA, m/z 72 |\\n| 39 | | 0.54 | ⁶⁸Zn(CH₃)⁺, m/z 83 |\\n| 39 | | 0.34 | ⁶⁸Zn(CH₃)₂⁺, m/z 98 |\\n| 39.5 | 54.16 | | TMA, m/z 72 |\\n| 39.5 | | 0.44 | ⁶⁸Zn(CH₃)⁺, m/z 83 |\\n| 39.5 | | 0.33 | ⁶⁸Zn(CH₃)₂⁺, m/z 98 |\\n| 40 | 55.09 | | TMA, m/z 72 |\\n| 40 | | 0.56 | ⁶⁸Zn(CH₃)⁺, m/z 83 |\\n| 40 | | 0.27 | ⁶⁸Zn(CH₃)₂⁺, m/z 98 |\\n| 40.5 | 55.65 | | TMA, m/z 72 |\\n| 40.5 | | 0.69 | ⁶⁸Zn(CH₃)⁺, m/z 83 |\\n| 40.5 | | 0.29 | ⁶⁸Zn(CH₃)₂⁺, m/z 98 |\\n| 41 | 0 | | TMA, m/z 72 |\\n| 41 | | 0.23 | ⁶⁸Zn(CH₃)⁺, m/z 83 |\\n| 41 | | 0.33 | ⁶⁸Zn(CH₃)₂⁺, m/z 98 |\"},{\"panel_id\":\"b\",\"text\":\"| Time (min) | TMA Intensity (mV) | Product Intensity (mV) | Material |\\n|------------|---------------------|--------------------------|------------------------------|\\n| 38.5 | 0 | | TMA, m/z 72 |\\n| 38.5 | | 9.377 | Al₂(CH₃)₄F⁺, m/z 133 |\\n| 39 | 51.04 | | TMA, m/z 72 |\\n| 39 | | 8.77 | Al₂(CH₃)₄F⁺, m/z 133 |\\n| 39.5 | 52.01 | | TMA, m/z 72 |\\n| 39.5 | | 8.64 | Al₂(CH₃)₄F⁺, m/z 133 |\\n| 40 | 54.87 | | TMA, m/z 72 |\\n| 40 | | 9.12 | Al₂(CH₃)₄F⁺, m/z 133 |\\n| 40.5 | 54.3 | | TMA, m/z 72 |\\n| 40.5 | | 9.24 | Al₂(CH₃)₄F⁺, m/z 133 |\\n| 41 | 0 | | TMA, m/z 72 |\\n| 41 | | 0 | Al₂(CH₃)₄F⁺, m/z 133 |\"}]","vqa":"[{\"panel_id\":\"b\",\"items\":[{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"This allows you to minimize the TMA exposure time per cycle. Since the reaction completes instantly, the pulse time can be very short, significantly reducing the total cycle time and increasing etch throughput without losing etch per cycle.\"}]},{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"1. Set the mass spectrometer to monitor specific m/z values (72, 83, 98). \\n2. Initiate a short, controlled pulse of TMA into the reactor. \\n3. Record ion intensities versus time. \\nThe timed valve creates a sharp TMA pulse to resolve the rapid surface reaction kinetics from the slower gas-phase filling.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The surface-limited reaction is better for efficiency, as all TMA is consumed at the surface with no waste. A potential downside is extreme sensitivity to surface defects, which could lead to non-uniform etching if the starting surface is inconsistent.\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"This indicates TMA is consumed instantly upon reaching the surface, confirming a highly reactive, fluorinated Zn-F surface* from the prior HF step. The surface reaction is faster than the gas-phase residence time, making product detection the first signature of the TMA pulse.\"}]}]","bbox":[{"panel_id":"b","x":22,"y":518,"width":645,"height":465},{"panel_id":"a","x":6,"y":37,"width":658,"height":464}],"source":"icdar2026-competition-data/train/atomic-layer-etching/experimental-usecase/23/images/figure_10.jpg","provenance":{"source_annotation":"icdar2026-competition-data/train/atomic-layer-etching/experimental-usecase/23/images/figure_10.json","source_content":"icdar2026-competition-data/train/atomic-layer-etching/experimental-usecase/23/content.json","source_pdf":["icdar2026-competition-data/train/atomic-layer-etching/experimental-usecase/23/Thermal Atomic Layer Etching of Zinc Sulfide.pdf"],"main_category":"atomic-layer-etching","sub_category":"experimental-usecase","paper_id":"23","first_classification_panel_id":"a","first_classification_label":"multi-axis chart","caption_source":"content.json:img_caption"},"width":669,"height":986,"image_format":"jpeg","image_sha256":"5da3c6a5da166d94cf956ab8504418e7ccda5b653c5392cda74b6e15525475b8","metadata_license":"CC BY 4.0","image_license":"source_publisher_rights_reserved","image_reuse_status":"non_commercial_research_use_only","schema_version":"1.0.0"} | |
| {"file_name":"images/train_atomic-layer-etching_experimental-usecase_32_figure_2.jpg","caption":"Figure 2. Relative atomic percentages among In, Ga, and As in InGaAs, and the atomic percentage of Cl on the InGaAs surface measured by XPS as a function of Cl adsorption time from 2.5 to $20\\mathrm{s}$ . The power to the ICP ion source was $200\\mathrm{W}$ at $\\mathrm{Cl}_2$ $1.0\\mathrm{mTorr}$ of process chamber pressure. The grid voltages to the ICP ion gun were maintained at $-10\\mathrm{V}$ for the first grid and $-20\\mathrm{V}$ for the second grid.","id":"train/atomic-layer-etching/experimental-usecase/32/figure_2","sample_id":"atomic-layer-etching/experimental-usecase/32/figure_2","subset":"multi-axis-chart","split":"train","classification":[{"panel_id":"a","label":"multi-axis chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"Multi-axis chart showing the atomic percentages of In, Ga, and As (left axis) and Cl (right axis) as a function of Cl adsorption time measured by XPS. The substrate element percentages remain approximately constant across the time range, while the Cl signal increases at short times and saturates at about 16% after ~10 s, as indicated by the annotation in the figure.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Time (s) | In (%) | Ga (%) | As (%) | Cl (%) |\\n|---|---|---|---|---|\\n| Pristine (0) | ~45 | ~36 | ~15 | ~1 |\\n| 2.5 | ~48 | ~38 | ~14 | ~13 |\\n| 5.0 | ~49 | ~36 | ~14 | ~20 |\\n| 10.0 | ~49 | ~38 | ~13 | ~26 |\\n| 20.0 | ~48 | ~38 | ~15 | ~26 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"About 16%.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"No, they remain relatively stable across the measured time range.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The Chlorine (Cl) atomic percentage increases rapidly during the initial adsorption period and then plateaus at approximately 16%, indicating a saturation behavior. In contrast, the atomic percentages of In, Ga, and As remain relatively constant throughout the adsorption time range. This suggests that Cl adsorption is self-limiting under these conditions and does not significantly perturb the near-surface elemental composition of the InGaAs substrate.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"1.0 mTorr.\"}]}]","bbox":[{"panel_id":"a","x":1,"y":4,"width":653,"height":457}],"source":"icdar2026-competition-data/train/atomic-layer-etching/experimental-usecase/32/images/figure_2.jpg","provenance":{"source_annotation":"icdar2026-competition-data/train/atomic-layer-etching/experimental-usecase/32/images/figure_2.json","source_content":"icdar2026-competition-data/train/atomic-layer-etching/experimental-usecase/32/content.json","source_pdf":["icdar2026-competition-data/train/atomic-layer-etching/experimental-usecase/32/Atomic layer etching of InGaAs.pdf"],"main_category":"atomic-layer-etching","sub_category":"experimental-usecase","paper_id":"32","first_classification_panel_id":"a","first_classification_label":"multi-axis chart","caption_source":"content.json:image_caption"},"width":655,"height":461,"image_format":"jpeg","image_sha256":"ec2a399861ca1451ecaa64275283926f3dead4015e42a08d73fd14a4b4fd7bf9","metadata_license":"CC BY 4.0","image_license":"source_publisher_rights_reserved","image_reuse_status":"non_commercial_research_use_only","schema_version":"1.0.0"} | |
| {"file_name":"images/train_atomic-layer-etching_experimental-usecase_32_figure_5.jpg","caption":"Figure 5. Etch depth (Å) and etch rate $(\\mathrm{\\AA} / \\mathrm{cycle})$ of InGaAs, Si, $\\mathrm{SiO}_2$ , $\\mathrm{HfO}_2$ , PR, and ACL measured as a function of $\\mathrm{Ar^{+}}$ ion desorption time (s/cycle) with the optimized ALE condition in figures 3 and 4.","id":"train/atomic-layer-etching/experimental-usecase/32/figure_5","sample_id":"atomic-layer-etching/experimental-usecase/32/figure_5","subset":"multi-axis-chart","split":"train","classification":[{"panel_id":"a","label":"multi-axis chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"Multi-axis chart showing etch depth (left axis) and etch rate (right axis) as a function of Ar⁺ ion exposure time for multiple materials under optimized ALE conditions. InGaAs exhibits a clear increase in etch depth that saturates near ~110 Å, while Si, SiO₂, HfO₂, PR, and ACL show no measurable etching across the entire time range.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Time (s/cycle) | InGaAs Etch Depth (Å) | InGaAs Etch Rate (Å/Cycle) | Etch Depth of Si, SiO₂, HfO₂, PR, ACL (Å) | Etch Rate of Si, SiO₂, HfO₂, PR, ACL (Å/Cycle) |\\n|---|---|---|---|---|\\n| 0 | 0 | 0.0 | 0 | 0 |\\n| 20 | ~45 | ~0.4 | 0 | 0 |\\n| 30 | ~60 | ~0.6 | 0 | 0 |\\n| 50 | ~110 | ~1.1 | 0 | 0 |\\n| 80 | ~115 | ~1.2 | 0 | 0 |\\n| 100 | ~110 | ~1.1 | 0 | 0 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"List\",\"answer\":\"1. InGaAs\\n, 2. Si\\n, 3. SiO2\\n, 4. HfO2\\n, 5. PR (Photoresist)\\n, 6. ACL (Amorphous Carbon Layer)\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Yes/No\",\"answer\":\"No\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The figure demonstrates extremely high selectivity for InGaAs. While the InGaAs sample (black squares) shows a significant etch depth that increases to saturation around 110 Å, all other tested materials (Si, SiO2, HfO2, PR, and ACL) exhibit zero etch depth. This indicates that the developed ALE process is highly specific to InGaAs and does not damage or remove common mask or substrate materials.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Around 50 seconds/cycle.\"}]}]","bbox":[{"panel_id":"a","x":3,"y":1,"width":649,"height":440}],"source":"icdar2026-competition-data/train/atomic-layer-etching/experimental-usecase/32/images/figure_5.jpg","provenance":{"source_annotation":"icdar2026-competition-data/train/atomic-layer-etching/experimental-usecase/32/images/figure_5.json","source_content":"icdar2026-competition-data/train/atomic-layer-etching/experimental-usecase/32/content.json","source_pdf":["icdar2026-competition-data/train/atomic-layer-etching/experimental-usecase/32/Atomic layer etching of InGaAs.pdf"],"main_category":"atomic-layer-etching","sub_category":"experimental-usecase","paper_id":"32","first_classification_panel_id":"a","first_classification_label":"multi-axis chart","caption_source":"content.json:image_caption"},"width":653,"height":444,"image_format":"jpeg","image_sha256":"965e6fddb0022eb82ebbe0a3b65d0f8de71e8855f0066e243349c4384beacdf8","metadata_license":"CC BY 4.0","image_license":"source_publisher_rights_reserved","image_reuse_status":"non_commercial_research_use_only","schema_version":"1.0.0"} | |
| {"file_name":"images/train_atomic-layer-etching_experimental-usecase_32_figure_6.jpg","caption":"Figure 6. Etch depth $(\\mathring{\\mathrm{A}})$ and etch rate (A/cycle) of InGaAs measured as a function of the number of ALE cycles.","id":"train/atomic-layer-etching/experimental-usecase/32/figure_6","sample_id":"atomic-layer-etching/experimental-usecase/32/figure_6","subset":"multi-axis-chart","split":"train","classification":[{"panel_id":"a","label":"multi-axis chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"Multi-axis chart showing InGaAs etch depth (left axis, black squares) and etch rate per cycle (right axis, red circles) as a function of the number of ALE cycles. Etch depth increases approximately linearly from 100 to 400 cycles, while the etch rate remains nearly constant at ~1.1 Å/cycle across the range.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| No. of ALE Cycles | Etch Depth (Å) [Left Axis] | Etch Rate (Å/Cycle) [Right Axis] |\\n|---|---|---|\\n| 100 | ~105 | ~1.10 |\\n| 200 | ~220 | ~1.10 |\\n| 300 | ~335 | ~1.13 |\\n| 400 | ~460 | ~1.15 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Approximately 1.1 Å/Cycle.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Yes/No\",\"answer\":\"No\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The etch depth exhibits a perfectly linear increase as the number of cycles increases from 100 to 400. The black squares fall along a straight line, rising from approximately 100 Å to roughly 460 Å. This linearity, combined with the constant etch rate, indicates that the etching process is highly stable and allows for precise control of the etch depth at the angstrom level.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"No. of ALE Cycles.\"}]}]","bbox":[{"panel_id":"a","x":4,"y":3,"width":665,"height":446}],"source":"icdar2026-competition-data/train/atomic-layer-etching/experimental-usecase/32/images/figure_6.jpg","provenance":{"source_annotation":"icdar2026-competition-data/train/atomic-layer-etching/experimental-usecase/32/images/figure_6.json","source_content":"icdar2026-competition-data/train/atomic-layer-etching/experimental-usecase/32/content.json","source_pdf":["icdar2026-competition-data/train/atomic-layer-etching/experimental-usecase/32/Atomic layer etching of InGaAs.pdf"],"main_category":"atomic-layer-etching","sub_category":"experimental-usecase","paper_id":"32","first_classification_panel_id":"a","first_classification_label":"multi-axis chart","caption_source":"content.json:image_caption"},"width":669,"height":450,"image_format":"jpeg","image_sha256":"b0d1b89df148cef029667379e1942de01f583ff63652cfb3e76e082cc00f5c6c","metadata_license":"CC BY 4.0","image_license":"source_publisher_rights_reserved","image_reuse_status":"non_commercial_research_use_only","schema_version":"1.0.0"} | |
| {"file_name":"images/train_atomic-layer-etching_experimental-usecase_32_figure_7.jpg","caption":"Figure 7. Surface roughness and surface composition after the 100 cycles of ALE, investigated using AFM and XPS, respectively. As references, the surface roughness and surface composition of as-received un-etched InGaAs and the InGaAs etched by conventional RIE were included. For the InGaAs RIE, InGaAs was etched for 1 min using an ICP etcher operated at $13.56\\mathrm{MHz}200\\mathrm{W}$ $-50\\mathrm{V}$ of bias voltage, and $\\mathrm{Cl}_2$ gas pressure of $50\\mathrm{mTorr}$","id":"train/atomic-layer-etching/experimental-usecase/32/figure_7","sample_id":"atomic-layer-etching/experimental-usecase/32/figure_7","subset":"multi-axis-chart","split":"train","classification":[{"panel_id":"a","label":"multi-axis chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"Multi-axis chart comparing surface composition (atomic %) and RMS roughness for Reference (Ref.), ALE, and RIE treated InGaAs. The RIE sample shows a measurable chlorine signal (7.8%) and the highest roughness on the right axis, while both Ref. and ALE show no detectable chlorine and lower roughness levels. The composition of In, Ga, and As remains broadly comparable between Ref. and ALE, indicating limited surface modification under ALE conditions.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Metric | Ref. | ALE | RIE |\\n|---------------|------|------|-------|\\n| In (Atomic %) | 29.4 | 29.5 | 38.0 |\\n| Ga (Atomic %) | 20.6 | 21.8 | 18.4 |\\n| As (Atomic %) | 6.6 | 5.3 | 2.4 |\\n| Cl (Atomic %) | ~0 | ~0 | 7.8 |\\n| C (Atomic %) | 43.4 | 43.4 | ~33.4 |\\n| Rq (nm) | ~60 | ~68 | ~78 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Chlorine (Cl).\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"No, the roughness values are very similar (4.6 nm for ALE vs. 4.3 nm for Reference).\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The ALE sample exhibits lower roughness and no detectable chlorine signal compared with the RIE sample. In contrast, the RIE condition shows a clear chlorine residue (7.8%) and the highest roughness level on the right axis. This indicates that ALE produces a smoother and chemically cleaner surface than RIE under the conditions shown.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Reactive Ion Etching (RIE).\"}]}]","bbox":[{"panel_id":"a","x":2,"y":4,"width":670,"height":420}],"source":"icdar2026-competition-data/train/atomic-layer-etching/experimental-usecase/32/images/figure_7.jpg","provenance":{"source_annotation":"icdar2026-competition-data/train/atomic-layer-etching/experimental-usecase/32/images/figure_7.json","source_content":"icdar2026-competition-data/train/atomic-layer-etching/experimental-usecase/32/content.json","source_pdf":["icdar2026-competition-data/train/atomic-layer-etching/experimental-usecase/32/Atomic layer etching of InGaAs.pdf"],"main_category":"atomic-layer-etching","sub_category":"experimental-usecase","paper_id":"32","first_classification_panel_id":"a","first_classification_label":"multi-axis chart","caption_source":"content.json:image_caption"},"width":672,"height":425,"image_format":"jpeg","image_sha256":"f312a63bb472ffd0042d7ac3987ec0d9b1cfcf8dcb2f93200ab1f33fb9814758","metadata_license":"CC BY 4.0","image_license":"source_publisher_rights_reserved","image_reuse_status":"non_commercial_research_use_only","schema_version":"1.0.0"} | |
| {"file_name":"images/train_atomic-layer-etching_experimental-usecase_33_fig_5.jpg","caption":"FIG. 5. IR annealing time dependence of the EPC of TiN films with error bars representing the standard deviation $\\sigma$ . Temperature (solid line), shown on the right axis, was measured using a thermocouple attached to the carrier wafer during processing.","id":"train/atomic-layer-etching/experimental-usecase/33/fig_5","sample_id":"atomic-layer-etching/experimental-usecase/33/fig_5","subset":"multi-axis-chart","split":"train","classification":[{"panel_id":"a","label":"multi-axis chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The chart shows the relationship between IR annealing time and EPC (nanometers per cycle) along with the corresponding temperature in degrees Celsius.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| IR Annealing Time (s) | EPC (nm/cycle) | Temperature (degrees celcius)|\\n|---|---|---|\\n| 0 | 0 |-10|\\n| 1 | 0.1 |-10|\\n| 2 | 0.2 |-10|\\n| 3 | 0.3 |-10|\\n| 4 | 0.3 |0|\\n| 5 | 0.5 |0|\\n| 6 | 0.4 |10|\\n| 7 | 1.1 |15|\\n| 8 | 1.3 |20|\\n| 9 | 1.2 |25|\\n| 10 | 1.0 |30|\\n| 15 | 2.6 |60|\\n| 20 | 2.8 |100|\\n| 25 | 2.6 |140|\\n| 30 | 2.7 |170|\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The EPC seems to show self-limiting behaviour, however there is also increase in table temperature. This means that the actual influence can not be seen easily.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Both the annealing duration and the temperature at which the experiment takes place have been changed. Usually in experiment designs the goal is to change only one parameter at a time.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"2.7 nm/cycle\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"For industrial applications the shortest exposures with still good controllablity should be chosen, this would be 15 s.\"}]}]","bbox":[{"panel_id":"a","x":3,"y":4,"width":671,"height":530}],"source":"icdar2026-competition-data/train/atomic-layer-etching/experimental-usecase/33/images/fig_5.jpg","provenance":{"source_annotation":"icdar2026-competition-data/train/atomic-layer-etching/experimental-usecase/33/images/fig_5.json","source_content":"icdar2026-competition-data/train/atomic-layer-etching/experimental-usecase/33/content.json","source_pdf":["icdar2026-competition-data/train/atomic-layer-etching/experimental-usecase/33/Atomic layer etching of titanium nitride.pdf"],"main_category":"atomic-layer-etching","sub_category":"experimental-usecase","paper_id":"33","first_classification_panel_id":"a","first_classification_label":"multi-axis chart","caption_source":"content.json:image_caption"},"width":677,"height":537,"image_format":"jpeg","image_sha256":"ef9029fce514a240f5534d7277a9324f226bf79b21a6ca01d69047cf14c48281","metadata_license":"CC BY 4.0","image_license":"source_publisher_rights_reserved","image_reuse_status":"non_commercial_research_use_only","schema_version":"1.0.0"} | |
| {"file_name":"images/train_atomic-layer-etching_experimental-usecase_34_fig_8.jpg","caption":"FIG. 8. Roughness and thickness change of $60 \\text{nm} \\text{Al}_2\\text{O}_3$ after 5 min of $50 \\text{W}$ dry etching at 0.06 Torr, 50-cycle pALE at $400^{\\circ}\\text{C}$ , and 100-cycle pALE at $400^{\\circ}\\text{C}$ .","id":"train/atomic-layer-etching/experimental-usecase/34/fig_8","sample_id":"atomic-layer-etching/experimental-usecase/34/fig_8","subset":"multi-axis-chart","split":"train","classification":[{"panel_id":"a","label":"multi-axis chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The figure compares thickness and RMS roughness of Al₂O₃ across four processing steps—initial deposition, dry etching, and two pALE cycle conditions. Thickness continuously decreases from ~60 nm to ~10 nm, while roughness first increases after dry etching and then decreases again with pALE, as confirmed by both plotted values and AFM images.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Process Condition|Thickness (nm)|RMS Roughness (nm)|\\n|------------------|--------------|------------------|\\n|as-deposited|55–60|0.8–1.0|\\n|50 W – 5 min|~40|1.3–1.4|\\n|50 cycles pALE|~25|~1.0|\\n|100 cycles pALE|~10–12|0.7–0.8|\"}]","vqa":"[{\"panel_id\":\"d\",\"items\":[{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The results show that while dry etching quickly reduces thickness, it also damages the surface by increasing roughness. In contrast, pALE offers both thinning and smoothing, restoring the surface to near-original roughness levels. This indicates that pALE is better suited for applications requiring precise surface control, smoother interfaces, and minimal plasma damage. The combined data allow researchers to balance thinning rate with surface quality for optimal device fabrication.\"}]},{\"panel_id\":\"b\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Dry etching leaves the film around ~40 nm thick with high roughness (~1.3–1.4 nm), while 50 and 100 pALE cycles further thin the film to ~25 nm and ~10–12 nm and reduce roughness to ~1.0 and ~0.7–0.8 nm.\"}]},{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Yes/No\",\"answer\":\"Yes\"}]},{\"panel_id\":\"c\",\"items\":[{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"List\",\"answer\":\"Self-limiting etch reactions improve smoothness, Surface becomes more uniform with controlled etching, Damage from aggressive dry etching is gradually repaired, Lower RMS roughness reflects higher surface quality\"}]}]","bbox":[{"panel_id":"a","x":0,"y":0,"width":866,"height":617}],"source":"icdar2026-competition-data/train/atomic-layer-etching/experimental-usecase/34/images/fig_8.jpg","provenance":{"source_annotation":"icdar2026-competition-data/train/atomic-layer-etching/experimental-usecase/34/images/fig_8.json","source_content":"icdar2026-competition-data/train/atomic-layer-etching/experimental-usecase/34/content.json","source_pdf":["icdar2026-competition-data/train/atomic-layer-etching/experimental-usecase/34/CF4 plasma-based atomic layer etching of Al2O3.pdf"],"main_category":"atomic-layer-etching","sub_category":"experimental-usecase","paper_id":"34","first_classification_panel_id":"a","first_classification_label":"multi-axis chart","caption_source":"content.json:image_caption"},"width":866,"height":617,"image_format":"jpeg","image_sha256":"2411de5bd8ec2bd8266e094994406e77899a006bc968c239022b913704ffc8d7","metadata_license":"CC BY 4.0","image_license":"source_publisher_rights_reserved","image_reuse_status":"non_commercial_research_use_only","schema_version":"1.0.0"} | |
| {"file_name":"images/train_atomic-layer-etching_experimental-usecase_46_figure_5.jpg","caption":"Figure 5. Peak intensity of high resolution XPS scans of (a) Ti 2p $(459.1~\\mathrm{eV})$ and Si 2p $(99.3\\mathrm{eV})$ and (b) W 4f and $(38\\mathrm{eV})$ and F 1s $(685.2\\mathrm{eV})$ for $5.5\\mathrm{nm}$ $\\mathrm{TiO_2}$ films exposed to 0, 10, 25, 50, and $100\\mathrm{WF}_6$ doses at $220^{\\circ}\\mathrm{C}$ . Lines are a guide for the eye. The decrease in the Ti 2p signal confirms film etching.","id":"train/atomic-layer-etching/experimental-usecase/46/figure_5","sample_id":"atomic-layer-etching/experimental-usecase/46/figure_5","subset":"multi-axis-chart","split":"train","classification":[{"panel_id":"a","label":"multi-axis chart"},{"panel_id":"b","label":"multi-axis chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"XPS Ti 2p and Si 2p spectra show a linear decrease in the Ti signal and a corresponding increase in the Si substrate signal with increasing WF₆ exposure, confirming progressive thinning and etching of the 5.5 nm TiO₂ film at 220 °C.\"},{\"panel_id\":\"b\",\"text\":\"XPS W 4f and F 1s spectra increase rapidly and then saturate with WF₆ dosing, indicating the formation of a stable tungsten–fluoride–rich surface layer and associated chemical modification during the process.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| WF₆ Doses | Intensity (a.u.) | Material |\\n|----------|------------------|----------|\\n| 0 | 28600 | Ti 2p Intensity |\\n| 0 | 1600 | Si 2p Intensity |\\n| 10 | 26200 | Ti 2p Intensity |\\n| 10 | 1667 | Si 2p Intensity |\\n| 25 | 25300 | Ti 2p Intensity |\\n| 25 | 1800 | Si 2p Intensity |\\n| 50 | 22500 | Ti 2p Intensity |\\n| 50 | 2490 | Si 2p Intensity |\\n| 75 | 21100 | Ti 2p Intensity |\\n| 75 | 3000 | Si 2p Intensity |\\n| 100 | 19300 | Ti 2p Intensity |\\n| 100 | 3645 | Si 2p Intensity |\"},{\"panel_id\":\"b\",\"text\":\"| WF₆ Doses | Intensity (a.u.) | Material |\\n|----------|------------------|----------|\\n| 0 | 2000 | W 4f Intensity |\\n| 0 | 5500 | F 1s Intensity |\\n| 10 | 4300 | W 4f Intensity |\\n| 10 | 6550 | F 1s Intensity |\\n| 25 | 4300 | W 4f Intensity |\\n| 25 | 6600 | F 1s Intensity |\\n| 50 | 4500 | W 4f Intensity |\\n| 50 | 7000 | F 1s Intensity |\\n| 75 | 4334 | W 4f Intensity |\\n| 75 | 7100 | F 1s Intensity |\\n| 100 | 4000 | W 4f Intensity |\\n| 100 | 7200 | F 1s Intensity |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"At 100 doses, the sample is in a near-breakthrough state. The Si 2p signal (substrate) has more than doubled in intensity compared to 0 doses, indicating the TiO₂ overburden is significantly thinner, allowing photoelectrons from the underlying silicon wafer to escape and be detected.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The WF₆ chemical method is preferable for leaving a passivated surface, as it forms a stable W-F termination. The trade-off is that it leaves a foreign material residue (W) on the surface. Physical sputtering would avoid residue but would create a damaged, reactive surface with dangling bonds requiring subsequent passivation.\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The composition changes from pure TiO₂ to a mixed tungsten oxyfluoride layer. The inverse trends represent a fluorination and exchange reaction: WF₆ fluorinates the TiO₂ network, forming volatile TiFₓ species that etch away (decreasing Ti), while non-volatile tungsten fluoride (WFₓ) products remain on the surface (increasing W and F).\"}]},{\"panel_id\":\"b\",\"items\":[{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Yes. The consistent, dose-dependent changes show the reaction is predictable and controllable. By calibrating the number of WF₆ doses to the decrease in Ti signal (thickness), engineers can reliably etch TiO₂ to a target thickness with minimal variation, which is essential for device performance.\"}]}]","bbox":[{"panel_id":"a","x":0,"y":0,"width":516,"height":371},{"panel_id":"b","x":526,"y":0,"width":520,"height":371}],"source":"icdar2026-competition-data/train/atomic-layer-etching/experimental-usecase/46/images/figure_5.jpg","provenance":{"source_annotation":"icdar2026-competition-data/train/atomic-layer-etching/experimental-usecase/46/images/figure_5.json","source_content":"icdar2026-competition-data/train/atomic-layer-etching/experimental-usecase/46/content.json","source_pdf":["icdar2026-competition-data/train/atomic-layer-etching/experimental-usecase/46/Thermal Selective Vapor Etching of TiO2.pdf"],"main_category":"atomic-layer-etching","sub_category":"experimental-usecase","paper_id":"46","first_classification_panel_id":"a","first_classification_label":"multi-axis chart","caption_source":"content.json:image_caption"},"width":1047,"height":372,"image_format":"jpeg","image_sha256":"8a44c5d879912d6d9fcf87fdb299abf719c01d55dd3e59c6ddff36640d8bb7af","metadata_license":"CC BY 4.0","image_license":"source_publisher_rights_reserved","image_reuse_status":"non_commercial_research_use_only","schema_version":"1.0.0"} | |
| {"file_name":"images/train_atomic-layer-etching_simulation-usecase_13_fig_3.jpg","caption":"FIG. 3. A typical plot of Ar kinetic energy (filled circles) and Ar projected position along the vertical axis (empty circles) as a function of time. The dashed horizontal line indicates the position at which an atom moving away from the lattice is removed from the simulation cell.","id":"train/atomic-layer-etching/simulation-usecase/13/fig_3","sample_id":"atomic-layer-etching/simulation-usecase/13/fig_3","subset":"multi-axis-chart","split":"train","classification":[{"panel_id":"a","label":"multi-axis chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The chart displays the kinetic energy of Ar atoms (in eV) and their y-position (in Angstroms) over time (in ps). The kinetic energy decreases rapidly at the beginning and then plateaus. The y-position also decreases rapidly initially, but then increases linearly after 0.2 ps.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Time (ps) | Ar Kinetic Energy (eV) | y-position (Angstrom) |\\n|---|---|---|\\n| 0.0 | 50.0 | 5.0 |\\n| 0.1 | 50.0 | -3.0 |\\n| 0.2 | 0.2 | -3.0 |\\n| 0.3 | 1.0 | -2.0 |\\n| 0.4 | 2.0 | 0.0 |\\n| 0.5 | 2.0 | 1.25 |\\n| 0.6 | 2.0 | 5.0 |\\n| 0.7 | 2.0 | 8.75 |\\n| 0.8 | 2.0 | 12.5 |\\n| 0.9 | 2.0 | 16.25 |\\n| 1.0 | 2.0 | 20.0 |\\n| 1.1 | 2.0 | |\\n| 1.2 | 2.0 | |\\n| 1.3 | 2.0 | |\\n| 1.4 | 2.0 | |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"0.0 to 1.4 ps.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"20.0 Å.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Roughly 2.0 eV.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Figure 3 shows how a 50 eV Ar⁺ ion interacts with the surface over the first few tenths of a picosecond, highlighting both the ion’s kinetic energy and its y-coordinate as functions of time. When the ion is introduced at \\nt = t50 at y = 8.5 Å, it rapidly penetrates the lattice and collides with atoms in the top layers. These collisions cause a fast drop in kinetic energy. As its energy continues to decrease, the Ar ion reaches its maximum penetration depth at y = -2 Å around 0.17 ps with a minimum kinetic energy below 0.2 eV. Subsequent collisions eventually impart energy back to the ion, reversing its motion and causing it to travel upward through the lattice. The figure shows the ion exiting the material at around 0.4 ps with a kinetic energy of roughly 2 eV. Once the ion crosses the y = 4.5 Å threshold, its kinetic energy becomes constant, indicating that it has left the “influence region” of the lattice and is no longer interacting with surface atoms.\"}]}]","bbox":[{"panel_id":"a","x":1,"y":5,"width":668,"height":432}],"source":"icdar2026-competition-data/train/atomic-layer-etching/simulation-usecase/13/images/fig_3.jpg","provenance":{"source_annotation":"icdar2026-competition-data/train/atomic-layer-etching/simulation-usecase/13/images/fig_3.json","source_content":"icdar2026-competition-data/train/atomic-layer-etching/simulation-usecase/13/content.json","source_pdf":["icdar2026-competition-data/train/atomic-layer-etching/simulation-usecase/13/Molecular dynamics simulation of atomic layer etching of silicon.pdf"],"main_category":"atomic-layer-etching","sub_category":"simulation-usecase","paper_id":"13","first_classification_panel_id":"a","first_classification_label":"multi-axis chart","caption_source":"content.json:img_caption"},"width":670,"height":439,"image_format":"jpeg","image_sha256":"a6e2e46004b022a71166766e8e224c8dcd37231bd11842d09981ffbae09276e1","metadata_license":"CC BY 4.0","image_license":"source_publisher_rights_reserved","image_reuse_status":"non_commercial_research_use_only","schema_version":"1.0.0"} | |
| {"file_name":"images/train_atomic-layer-etching_simulation-usecase_13_fig_4.jpg","caption":"FIG. 4. The hot-spot temperature (triangles) and the number of atoms in the hot spot (circles) as a function of time. The figure shows an average over 300 simulation runs.","id":"train/atomic-layer-etching/simulation-usecase/13/fig_4","sample_id":"atomic-layer-etching/simulation-usecase/13/fig_4","subset":"multi-axis-chart","split":"train","classification":[{"panel_id":"a","label":"multi-axis chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The chart shows the average hot spot temperature and the average number of atoms in the hot spot over time.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Time (ps) | Average hot spot temperature (K) | Average number of atoms in the hot spot |\\n|---|---|---|\\n| 0.0 | 300 | 0 |\\n| 0.2 | 1000 | 70 |\\n| 0.4 | 800 | 80 |\\n| 0.6 | 700 | 60 |\\n| 0.8 | 600 | 50 |\\n| 1.0 | 500 | 40 |\\n| 1.2 | 400 | 25 |\\n| 1.4 | 400 | 20 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"300.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Around 400 K.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Around 30.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Figure 4 shows how a localized hot spot develops immediately after the energetic Ar ion collides with the surface and transfers energy into the lattice. The average crystal temperature rises sharply and reaches a maximum of about 1070 K around the moment when most of the ion’s kinetic energy has been dissipated, roughly 0.15 ps after impact. When focusing only on the atoms directly affected by the collision, the hot spot reaches an even higher peak temperature of more than 3000 K at around 0.10 ps. At this very early stage, only a small group of atoms carry most of the deposited energy. As these atoms transfer energy into their surroundings, the hot spot spreads, which increases the number of atoms involved but lowers the average temperature of this region. This spreading continues until about 0.35 ps, when the hot spot reaches its maximum size of about 75 atoms. Thereafter both the local temperature and the number of affected atoms decrease as energy continues to dissipate through the lattice and into the surrounding temperature-controlled region. The authors emphasize that the hot-spot temperature is only a qualitative measure because temperature is not rigorously defined when only a small number of atoms are involved.\"}]}]","bbox":[{"panel_id":"a","x":6,"y":6,"width":664,"height":421}],"source":"icdar2026-competition-data/train/atomic-layer-etching/simulation-usecase/13/images/fig_4.jpg","provenance":{"source_annotation":"icdar2026-competition-data/train/atomic-layer-etching/simulation-usecase/13/images/fig_4.json","source_content":"icdar2026-competition-data/train/atomic-layer-etching/simulation-usecase/13/content.json","source_pdf":["icdar2026-competition-data/train/atomic-layer-etching/simulation-usecase/13/Molecular dynamics simulation of atomic layer etching of silicon.pdf"],"main_category":"atomic-layer-etching","sub_category":"simulation-usecase","paper_id":"13","first_classification_panel_id":"a","first_classification_label":"multi-axis chart","caption_source":"content.json:img_caption"},"width":669,"height":427,"image_format":"jpeg","image_sha256":"5371dee74b91cea750c6e6aa30be93e73d12a5103f1fad0e03600a41a89a78e2","metadata_license":"CC BY 4.0","image_license":"source_publisher_rights_reserved","image_reuse_status":"non_commercial_research_use_only","schema_version":"1.0.0"} | |
| {"file_name":"images/train_atomic-layer-etching_simulation-usecase_17_fig1.jpg","caption":"Fig.1. t time of G0 and G3 for both steps, respectively. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)","id":"train/atomic-layer-etching/simulation-usecase/17/fig1","sample_id":"atomic-layer-etching/simulation-usecase/17/fig1","subset":"multi-axis-chart","split":"train","classification":[{"panel_id":"a","label":"multi-axis chart"},{"panel_id":"b","label":"multi-axis chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The figure shows the consumption of HF at different feed flow rates, along with the half-cycle time for G0 and G3 processes for Step A.\"},{\"panel_id\":\"b\",\"text\":\"The figure shows the consumption of TMA at different feed flow rates, along with the half-cycle time for G0 and G3 processes for Step B.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| HF Feed Flow Rate (sccm) | G0: HF Consumed (×10⁵ std cm³/year) | G3: HF Consumed (×10⁵ std cm³/year) | G0: Half-Cycle Time (s) | G3: Half-Cycle Time (s) |\\n|---------------------------|--------------------------------------|--------------------------------------|---------------------------|---------------------------|\\n| 50 | 0.5 | 0.5 | 1.6 | 1.6 |\\n| 150 | 1.2 | 1.2 | 1.4 | 1.4 |\\n| 300 | 2.5 | 2.5 | 1.4 | 1.4 |\\n| 600 | 4.8 | 4.2 | 1.3 | 1.2 |\\n| 600 | 5.8 | 5.3 | 1.1 | 1.0 |\"},{\"panel_id\":\"b\",\"text\":\"| TMA Feed Flow Rate (sccm) | G0: TMA Consumed (×10⁵ std cm³/year) | G3: TMA Consumed (×10⁵ std cm³/year) | G0: Half-Cycle Time (s) | G3: Half-Cycle Time (s) |\\n|----------------------------|---------------------------------------|---------------------------------------|---------------------------|---------------------------|\\n| 70 | 1.0 | 1.0 | 2.5 | 2.5 | \\n| 150 | 2.2 | 2.2 | 2.5 | 2.5 |\\n| 300 | 4.2 | 4.2 | 2.4 | 2.4 |\\n| 600 | 7.8 | 7.0 | 2.1 | 1.8 |\\n| 900 | 10 |9.8 | 1.8 | 1.8 |\"}]","vqa":"[{\"panel_id\":\"b\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Yes/No\",\"answer\":\"Yes\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"For both reactors, the HF cycle time and TMA cycle time remain constant until 500 sccm and change respectively when\\n1. HF flow rate is 600 and 900 sccm\\n2. TMA flow rate is 600 and 900 sccm.\"}]},{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The annual feed consumption is calculated by assuming that 96 cycles of etching are conducted daily and that the half-cycle times for each feed flow rate remain constant with each cycle in a single wafer system. The estimates for the precursor consumption are calculated for a single wafer.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The results for Step A a indicate that the half-cycle time for Step A for G3 is consistently faster compared to that of G0. Thus, the amount of precursor needed to ensure complete coverage is less than that of G0. By utilizing G3, at least 1.3 ×10 3 std cm 3 and at most 5.5 ×10 4 std cm 3 of HF can be saved for the range of the simulated flow rates with 600 sccm flow rate achieving the greatest amount of precursor that could be saved\"}]}]","bbox":[{"panel_id":"a","x":0,"y":9,"width":613,"height":433},{"panel_id":"b","x":644,"y":11,"width":632,"height":431}],"source":"icdar2026-competition-data/train/atomic-layer-etching/simulation-usecase/17/images/fig1.jpg","provenance":{"source_annotation":"icdar2026-competition-data/train/atomic-layer-etching/simulation-usecase/17/images/fig1.json","source_content":"icdar2026-competition-data/train/atomic-layer-etching/simulation-usecase/17/content.json","source_pdf":["icdar2026-competition-data/train/atomic-layer-etching/simulation-usecase/17/Multiscale computational fluid dynamics modeling of thermal atomic layer etching.pdf"],"main_category":"atomic-layer-etching","sub_category":"simulation-usecase","paper_id":"17","first_classification_panel_id":"a","first_classification_label":"multi-axis chart","caption_source":"content.json:img_caption"},"width":1286,"height":445,"image_format":"jpeg","image_sha256":"9508d408c4b0a5b92669c875e17333bb5e949c066b703791ffca664c98a95b41","metadata_license":"CC BY 4.0","image_license":"source_publisher_rights_reserved","image_reuse_status":"non_commercial_research_use_only","schema_version":"1.0.0"} | |
| {"file_name":"images/train_atomic-layer-etching_simulation-usecase_17_fig_10.jpg","caption":"Fig. 10. Complete cycle of G3 displaying the pressure of HF and TMA and coverage of $\\mathrm{AlF_3}$ for all HF and TMA feed flow rate of $150~\\mathrm{scm / h}$ and $70~\\mathrm{scm / h}$ , respectively. The blue solid line and the orange dashed line indicate the pressure of HF and TMA over time, respectively. The yellow solid line shows the coverage of $\\mathrm{AlF_3}$ . The $\\mathrm{AlF_3}$ is formed in Step A and etched in Step B. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)","id":"train/atomic-layer-etching/simulation-usecase/17/fig_10","sample_id":"atomic-layer-etching/simulation-usecase/17/fig_10","subset":"multi-axis-chart","split":"train","classification":[{"panel_id":"a","label":"multi-axis chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The chart displays the pressure and coverage of AIF<sub>3</sub> over time, with two separate axes for each metric for reactor model G3.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Time (s) | HF Pressure (Pa) | TMA Pressure (Pa) | Coverage of AlF₃ (%) |\\n|----------|-------------------|--------------------|------------------------|\\n| 0 | 0 | 0 | 0 |\\n| 0.5 | 30 | 0 | 40 |\\n| 1.0 | 65 | 0 | 80 |\\n| 1.5 | 68 | 0 | 95 |\\n| 2.0 | 65 | 0 | 100 |\\n| 3.0 | 0 | 0 | 100 |\\n| 6.0 | 0 | 0 | 100 |\\n| 7.0 | 0 | 30 | 100 |\\n| 8.0 | 0 | 60 | 100 |\\n| 9.0 | 0 | 65 | 100 |\\n| 10.0 | 0 | 60 | 100 |\\n| 11.0 | 0 | 0 | 100 |\\n| 15.0 | 0 | 0 | 100 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"The HF and TMA feed flow rate of 150 sccm and 70 sccm, respectively.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"The inclined plate reactor (G3) shows the best performance in terms of the film uniformity and etching speed.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"It means a complete etching cycle consisting of \\n1. HF dose of 2 s\\n2. N 2 purge of 5 s\\n3. A TMA dose of 3 s\\n4. N 2 purge of 5 s\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"HF Pressure starts increasing at 0 s, peaks around 65–70 Pa between 1–2 s, then drops sharply to near zero by 3 s, indicating that the HF pulse occurs early in the cycle (roughly 0–2.5 s). TMA Pressure (Red dashed line), remains at zero until about 7 s, then rises sharply to about 60–65 Pa between 8–10 s, and falls back to zero by 11 s. This shows the TMA pulse happens later in the cycle (roughly 7–10.5 s). Coverage of AlF₃ (Yellow dotted line): Starts at 0%, rises rapidly during the HF pulse, reaching 100% coverage by ~2 s, and stays constant for the rest of the cycle. This means the surface becomes fully covered with AlF₃ after the HF exposure and does not change during the TMA pulse. Thus the process is sequential as it represents complete cycle with TMA following HF.\"}]}]","bbox":[{"panel_id":"a","x":5,"y":5,"width":664,"height":326}],"source":"icdar2026-competition-data/train/atomic-layer-etching/simulation-usecase/17/images/fig_10.jpg","provenance":{"source_annotation":"icdar2026-competition-data/train/atomic-layer-etching/simulation-usecase/17/images/fig_10.json","source_content":"icdar2026-competition-data/train/atomic-layer-etching/simulation-usecase/17/content.json","source_pdf":["icdar2026-competition-data/train/atomic-layer-etching/simulation-usecase/17/Multiscale computational fluid dynamics modeling of thermal atomic layer etching.pdf"],"main_category":"atomic-layer-etching","sub_category":"simulation-usecase","paper_id":"17","first_classification_panel_id":"a","first_classification_label":"multi-axis chart","caption_source":"content.json:img_caption"},"width":670,"height":333,"image_format":"jpeg","image_sha256":"2d501670d73fb8ab08d958236c895294c75fabf8289dfff2eb8c3c35a30e5d77","metadata_license":"CC BY 4.0","image_license":"source_publisher_rights_reserved","image_reuse_status":"non_commercial_research_use_only","schema_version":"1.0.0"} | |
| {"file_name":"images/train_atomic-layer-etching_simulation-usecase_26_fig_4.jpg","caption":"FIG. 4. Changes in the etched depth (black) and surface concentration of Cl atoms (blue) as functions of the $\\mathsf{Ar}^+$ ion dose in the desorption step. Here, the etched depth is defined as the position of the top surface measured from its position at the end of the previous desorption step. The ALE conditions for (a)–(d) are the same as those for (a)–(d) of Fig. 3, respectively. The results shown here are the averages of those over C3–C6 cycles.","id":"train/atomic-layer-etching/simulation-usecase/26/fig_4","sample_id":"atomic-layer-etching/simulation-usecase/26/fig_4","subset":"multi-axis-chart","split":"train","classification":[{"panel_id":"a","label":"multi-axis chart"},{"panel_id":"b","label":"multi-axis chart"},{"panel_id":"c","label":"multi-axis chart"},{"panel_id":"d","label":"multi-axis chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The chart shows the depth change and concentration of Cl atoms on Si as a function of ion dose for 20 eV Ar<sup>+</sup> ions for thin Cl layer adsorption.\"},{\"panel_id\":\"b\",\"text\":\"The chart shows the depth change and concentration of Cl atoms on Si as a function of ion dose for 50 eV Ar<sup>+</sup> ions for thin Cl layer adsorption.\"},{\"panel_id\":\"c\",\"text\":\"The chart shows the depth change and concentration of Cl atoms on Si as a function of ion dose for 20 eV Ar<sup>+</sup> ions for thick Cl layer adsorption.\"},{\"panel_id\":\"d\",\"text\":\"The chart shows the depth change and concentration of Cl atoms on Si as a function of ion dose for 50 eV Ar<sup>+</sup> ions for thick Cl layer adsorption.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Ion Dose (x10<sup>16</sup> ions/cm<sup>2</sup>) | Depth Change (Å) | Cl Atoms on Si (10<sup>15</sup>/cm<sup>2</sup>) |\\n|---|---|---|\\n| 0.0 | 0.0 | 1.5 |\\n| 5.0 | 0.5 | 1.0 |\\n| 10.0 | 1.0 | 0.8 |\\n| 15.0 | 1.5 | 0.75 |\"},{\"panel_id\":\"b\",\"text\":\"| Ion Dose (x10<sup>16</sup> ions/cm<sup>2</sup>) | Depth Change (Å) | Cl Atoms on Si (10<sup>15</sup>/cm<sup>2</sup>) |\\n|---|---|---|\\n| 0.0 | 0.0 | 6.2 |\\n| 5.0 | 2.2 | 2.2 |\\n| 10.0 | 4.4 | 1.2 |\\n| 15.0 | 6.6 | 0.4 |\"},{\"panel_id\":\"c\",\"text\":\"| Ion Dose (x10<sup>16</sup> ions/cm<sup>2</sup>) | Depth Change (Å) | Cl Atoms on Si (10<sup>15</sup>/cm<sup>2</sup>) |\\n|---|---|---|\\n| 0.0 | 0.0 | 4.2 |\\n| 5.0 | 1.5 | 1.2 |\\n| 10.0 | 3.0 | 1 |\\n| 15.0 | 4.5 | 0.4 |\"},{\"panel_id\":\"d\",\"text\":\"| Ion Dose (x10<sup>16</sup> ions/cm<sup>2</sup>) | Depth Change (Å) | Cl Atoms on Si (10<sup>15</sup>/cm<sup>2</sup>) |\\n|---|---|---|\\n| 0.0 | 0.0 | 12.5 |\\n| 5.0 | 2.5 | 1 |\\n| 10.0 | 5.0 | 0.6 |\\n| 15.0 | 7.5 | 0.4 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Here, the etched depth is defined as the position of the top surface measured from its position at the end of the previous desorption step.\"}]},{\"panel_id\":\"c\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"At low Ar+ ion incident energy, it is seen that, with the increasing Ar+ ion dose, the etched depth in (a) increased slowly with a weakly decreasing slope, whereas the etched depth in (c) increased rapidly with a clear sign of saturation, i.e., the self-limit. The surface concentrations of Cl atoms also\\ndecreased slowly and nearly one-half of the initial Cl atoms remained on the surface at the end of the ALE cycle in (a), whereas the surface concentrations of Cl atoms in (c) decreased sharply with the Ar+ ion dose. It should be noted, however, that, even in the case of (c), a small amount of Cl, with a surface concentration similar to that in (a), still remained on the Si surface at the end of\\nthe ALE cycle.\"}]},{\"panel_id\":\"d\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"At high Ar+ ion incident energy, it is seen in (b) that, with a small amount of Cl deposited in the adsorption step, the etched depth monotonically increased all throughout the Ar+ ion injection,\\nindicating that physical sputtering dominates the etching process. In (d), where a larger amount of Cl was deposited in the adsorption step, the etched depth increased drastically until about\\n5 * 10^16/cm2 and then increased monotonically in the same way as in (b). In both cases, the density of Cl atoms at the end of the ALE cycle is almost negligible.\"}]},{\"panel_id\":\"b\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Yes/No\",\"answer\":\"No\"}]}]","bbox":[{"panel_id":"a","x":8,"y":49,"width":652,"height":508},{"panel_id":"b","x":687,"y":53,"width":654,"height":503},{"panel_id":"c","x":9,"y":583,"width":654,"height":512},{"panel_id":"d","x":687,"y":583,"width":655,"height":512}],"source":"icdar2026-competition-data/train/atomic-layer-etching/simulation-usecase/26/images/fig_4.jpg","provenance":{"source_annotation":"icdar2026-competition-data/train/atomic-layer-etching/simulation-usecase/26/images/fig_4.json","source_content":"icdar2026-competition-data/train/atomic-layer-etching/simulation-usecase/26/content.json","source_pdf":["icdar2026-competition-data/train/atomic-layer-etching/simulation-usecase/26/Surface damage formation during atomic layer etching of silicon with chlorine adsorption.pdf"],"main_category":"atomic-layer-etching","sub_category":"simulation-usecase","paper_id":"26","first_classification_panel_id":"a","first_classification_label":"multi-axis chart","caption_source":"content.json:img_caption"},"width":1350,"height":1095,"image_format":"jpeg","image_sha256":"5b5d209be74bf51302baabf3c02eec128c59ee453025b5fa27f0f872a8bd57f3","metadata_license":"CC BY 4.0","image_license":"source_publisher_rights_reserved","image_reuse_status":"non_commercial_research_use_only","schema_version":"1.0.0"} | |
| {"file_name":"images/train_atomic-layer-etching_simulation-usecase_27_fig_11.jpg","caption":"FIG. 11. Al2p XPS peak area and peak binding energy after 30 ALE cycles as a function of table temperature. The binding energies of $\\mathrm{AlF}_3$ and $\\mathrm{Al}_2\\mathrm{O}_3$ have been included as dashed lines to show the transition between the two binding environments. The insets show the fitted Al2p XPS spectra for the films etched at 100 and $200^{\\circ}\\mathrm{C}$ .","id":"train/atomic-layer-etching/simulation-usecase/27/fig_11","sample_id":"atomic-layer-etching/simulation-usecase/27/fig_11","subset":"multi-axis-chart","split":"train","classification":[{"panel_id":"a","label":"multi-axis chart"},{"panel_id":"b","label":"spectra chart"},{"panel_id":"c","label":"spectra chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"XPS analysis showing Al2p peak area (black) and binding energy (blue) as a function of process temperature after 30 ALE cycles. The Al2p signal is high at low temperatures (~100°C) with binding energy near AlF3 (76.5 eV), indicating AlF3 deposition dominates. Above 200°C, the peak area decreases and stabilizes while binding energy shifts toward Al2O3 (74.5 eV), indicating successful ALE with minimal Al contamination.\"},{\"panel_id\":\"b\",\"text\":\"Fitted Al2p XPS spectrum at 100°C showing a large peak consistent with AlF3 formation.\"},{\"panel_id\":\"c\",\"text\":\"Fitted Al2p XPS spectrum at 200°C showing a smaller peak shifted toward Al2O3 binding energy.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Temperature (°C) | Al2p Peak Area (x10³ CPS eV) | Peak Binding Energy (eV) |\\n|---|---|---|\\n| 100 | 17.8 | 76.25 |\\n| 125 | 16.2 | 76.45 |\\n| 150 | 14.5 | 76.35 |\\n| 175 | 5.2 | 75.80 |\\n| 200 | 1.2 | 74.85 |\\n| 225 | 0.9 | 74.50 |\\n| 250 | 0.8 | 74.85 |\\n| 275 | 0.9 | 74.65 |\\n| 300 | 1.0 | 74.75 |\"},{\"panel_id\":\"b\",\"text\":\"| Binding Energy (eV) | Intensity 100°C (x10³ CPS) | Intensity 200°C (x10³ CPS) |\\n|---|---|---|\\n| 80.0 | 0.1 | 0.0 |\\n| 78.0 | 1.5 | 0.0 |\\n| 77.0 | 5.5 | 0.1 |\\n| 76.2 | 7.3 | 0.2 |\\n| 75.5 | 5.0 | 0.4 |\\n| 74.8 | 1.2 | 0.8 |\\n| 74.0 | 0.2 | 0.3 |\\n| 72.0 | 0.0 | 0.0 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"At low temperatures (100°C), significant Al contamination occurs due to dominant AlF3 ALD reactions, resulting in a large Al2p peak area. Between 125 and 175°C, the Al content decreases as the AlF3 ALD reaction is suppressed while the ALE ligand-exchange reaction becomes more effective. Above 200°C, the Al2p signal stabilizes at a minimal level (<1 at.%), indicating that Al atoms deposited during the TMA step are efficiently removed.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"At low temperatures, AlF3 deposition from the SF6 plasma/TMA chemistry dominates, resulting in Al-F bonding. At higher temperatures, the AlF3 ALD reaction is suppressed and any residual Al on the surface comes from unreacted TMA in the final half-cycle. When samples are transferred through ambient air to the XPS, this residual TMA oxidizes, producing Al2O3-like binding environments rather than AlF3.\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"200°C.\"}]},{\"panel_id\":\"b\",\"items\":[{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"76.5 eV corresponds to Al-F bonds (AlF3), while 74.5 eV corresponds to Al-O bonds (Al2O3).\"}]}]","bbox":[{"panel_id":"a","x":1,"y":0,"width":670,"height":519},{"panel_id":"b","x":245,"y":121,"width":165,"height":161},{"panel_id":"c","x":407,"y":120,"width":166,"height":164}],"source":"icdar2026-competition-data/train/atomic-layer-etching/simulation-usecase/27/images/fig_11.jpg","provenance":{"source_annotation":"icdar2026-competition-data/train/atomic-layer-etching/simulation-usecase/27/images/fig_11.json","source_content":"icdar2026-competition-data/train/atomic-layer-etching/simulation-usecase/27/content.json","source_pdf":["icdar2026-competition-data/train/atomic-layer-etching/simulation-usecase/27/Isotropic atomic layer etching of GaN using SF6 plasma and Al(CH3)3.pdf"],"main_category":"atomic-layer-etching","sub_category":"simulation-usecase","paper_id":"27","first_classification_panel_id":"a","first_classification_label":"multi-axis chart","caption_source":"content.json:image_caption"},"width":675,"height":523,"image_format":"jpeg","image_sha256":"63218294d8562bd5b4928ee54197b794d5e5ad6a6b60642f93d478db8a50da59","metadata_license":"CC BY 4.0","image_license":"source_publisher_rights_reserved","image_reuse_status":"non_commercial_research_use_only","schema_version":"1.0.0"} | |
| {"file_name":"images/train_atomic-layer-etching_simulation-usecase_3_fig_2.jpg","caption":"FIG. 2. (Color online) Etch front characteristics during ideal ALE etching. (a) Etch depth and surface roughness as a function of time for the second through fifth ALE pulses. The different subcycles $\\mathrm{Cl} =$ passivation, $\\mathrm{Ar}^+$ $=$ ion bombardment) are shown at the top for reference, with passivation phases being highlighted by gray bands in the figure. Four pulse periods are shown in total. (b) Surface coverage of Si, $\\mathrm{SiCl}$ $\\mathrm{SiCl}_2$ $\\mathrm{SiCl}_3$ and average chlorine per site at the etch front as a function of time for the second pulse.","id":"train/atomic-layer-etching/simulation-usecase/3/fig_2","sample_id":"atomic-layer-etching/simulation-usecase/3/fig_2","subset":"multi-axis-chart","split":"train","classification":[{"panel_id":"a","label":"multi-axis chart"},{"panel_id":"b","label":"multi-axis chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The chart shows the etch depth and roughness over time during a process involving alternating cycles of Ar+ and Cl treatments.\"},{\"panel_id\":\"b\",\"text\":\"The chart illustrates the surface coverage of Si, SiCl, SiCl2, and SiCl3 along with the average chlorine site over time during a process involving Cl and Ar+ treatments.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Time (s) | Etch Depth (ML) | Roughness |\\n|----------|-----------------|-----------|\\n| 3.0 | 0 | 0.00 |\\n| 6.5 | -1.5 | 1.00 |\\n| 10.0 | -3.0 | 1.00 |\\n| 13.5 | -4.5 | 1.00 |\\n| 17.0 | -5.0 | 0.00 |\"},{\"panel_id\":\"b\",\"text\":\"| Time (s) | Si Coverage | <Cl> | SiCl3 | SiCl2 | SiCl |\\n|----------|-------------|------|-------|-------|------|\\n| 3.0 | 1.00 | 0.00 | 0.00 | 0.00 | 0.00 |\\n| 3.3 | 0.00 | 2.80 | 0.80 | 0.50 | 0.20 |\\n| 4.8 | 0.90 | 0.50 | 0.10 | 0.05 | 0.00 |\\n| 6.8 | 1.00 | 2.90 | 0.90 | 0.60 | 0.30 |\"}]","vqa":"[{\"panel_id\":\"b\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The surface coverages of SiClx plotted as a function of time for one ALE cycle indicate that there is a short (50 ms) transient period at the beginning\\nof the Cl passivation step where SiClx x<3 species dominate.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"After the initial transient the surface quickly establishes a steady state coverage of \\n1. 1% SiCl\\n2. 21% SiCl2\\n3. 78% SiCl3\\nmaking an average chlorination per surface site, <Cl>, of 2.77.\"}]},{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The etch depth as a function of time for the AR2 feature shows exactly one ML of material being removed in each ALE cycle from the second cycle after starting the etch, to the fifth cycle. The roughness increases at the beginning of each ion bombardment step as the passivated layer is partially and statistically eroded.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Yes after etching\"}]}]","bbox":[{"panel_id":"a","x":11,"y":8,"width":561,"height":448},{"panel_id":"b","x":0,"y":501,"width":572,"height":427}],"source":"icdar2026-competition-data/train/atomic-layer-etching/simulation-usecase/3/images/fig_2.jpg","provenance":{"source_annotation":"icdar2026-competition-data/train/atomic-layer-etching/simulation-usecase/3/images/fig_2.json","source_content":"icdar2026-competition-data/train/atomic-layer-etching/simulation-usecase/3/content.json","source_pdf":["icdar2026-competition-data/train/atomic-layer-etching/simulation-usecase/3/Atomic layer etching of 3D structures in silicon Self-limiting.pdf"],"main_category":"atomic-layer-etching","sub_category":"simulation-usecase","paper_id":"3","first_classification_panel_id":"a","first_classification_label":"multi-axis chart","caption_source":"content.json:img_caption"},"width":572,"height":928,"image_format":"jpeg","image_sha256":"2edddb76f1104fe8853353a63d29261d61d1b3f7ee86eb918cafb4d2536d76a7","metadata_license":"CC BY 4.0","image_license":"source_publisher_rights_reserved","image_reuse_status":"non_commercial_research_use_only","schema_version":"1.0.0"} | |
| {"file_name":"images/train_atomic-layer-etching_simulation-usecase_3_fig_3.jpg","caption":"FIG. 3. (Color online) Surface coverage of Si and average chlorine per site at the etch front for ideal ALE as a function of time for the second ALE pulse for $\\mathrm{AR} = 2$ and 10 trenches.","id":"train/atomic-layer-etching/simulation-usecase/3/fig_3","sample_id":"atomic-layer-etching/simulation-usecase/3/fig_3","subset":"multi-axis-chart","split":"train","classification":[{"panel_id":"a","label":"multi-axis chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The chart shows the Si surface coverage and average chlorine per site over time, with two curves labeled 'Ar = 2' and 'Ar = 10' as a function of time.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Time (s) | Si Coverage (Ar=2) | Si Coverage (Ar=10) | Cl Coverage (Ar=2) | Cl Coverage (Ar=10) |\\n|----------|---------------------|----------------------|---------------------|----------------------|\\n| 2.8 | 1.00 | 1.00 | 0.00 | 0.00 |\\n| 3.0 | 0.00 | 0.00 | 2.80 | 2.50 |\\n| 3.2 | 0.00 | 0.00 | 2.90 | 2.70 |\\n| 3.4 | 0.10 | 0.05 | 2.80 | 2.60 |\\n| 3.6 | 0.50 | 0.40 | 2.00 | 1.80 |\\n| 3.8 | 0.90 | 0.80 | 1.00 | 0.80 |\\n| 4.0 | 1.00 | 0.95 | 0.20 | 0.10 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The only significant difference between the ideal ALE of features having AR 2 and AR 10 is the time required to achieve the steady state value of <Cl>\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Yes/No\",\"answer\":\"Yes\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"A longer time is required to fully passivate the AR 10 feature compared to AR 2. As the AR increases, the neutral conductance through the feature decreases, and a larger portion of the incoming neutral flux is reflected back into the plasma by collisions with the sidewalls before the flux can passivate the etch front\\nat the bottom of the feature. The lower conductance of the AR 10 feature requires a larger fluence of Cl radicals, and therefore a longer passivation time, to achieve the same <Cl > as the AR 2 feature.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"From the image, it can be observed thattThe Cl coverage for AR 2 is always higher than AR 10\"}]}]","bbox":[{"panel_id":"a","x":4,"y":5,"width":572,"height":432}],"source":"icdar2026-competition-data/train/atomic-layer-etching/simulation-usecase/3/images/fig_3.jpg","provenance":{"source_annotation":"icdar2026-competition-data/train/atomic-layer-etching/simulation-usecase/3/images/fig_3.json","source_content":"icdar2026-competition-data/train/atomic-layer-etching/simulation-usecase/3/content.json","source_pdf":["icdar2026-competition-data/train/atomic-layer-etching/simulation-usecase/3/Atomic layer etching of 3D structures in silicon Self-limiting.pdf"],"main_category":"atomic-layer-etching","sub_category":"simulation-usecase","paper_id":"3","first_classification_panel_id":"a","first_classification_label":"multi-axis chart","caption_source":"content.json:img_caption"},"width":578,"height":439,"image_format":"jpeg","image_sha256":"8ec8c1817093ee2f9a398eaa9ed66a79719c911811d92da25c5cd6247b9d9b75","metadata_license":"CC BY 4.0","image_license":"source_publisher_rights_reserved","image_reuse_status":"non_commercial_research_use_only","schema_version":"1.0.0"} | |
| {"file_name":"images/train_atomic-layer-etching_simulation-usecase_31_figure_5.jpg","caption":"Figure 5. Removal step energies by terminations on the $\\mathrm{Ni / N}$ , $\\mathrm{Cu / N}$ , $\\mathrm{Ni / O}$ , and $\\mathrm{Cu / O}$ systems as functions of coverage, evaluated at $80^{\\circ}\\mathrm{C}$ . Two axes are shown. The x-axis on the left corresponds to the formamidine chemistry. The y-axis on the right corresponds to the formic acid chemistry. The bulk model results on the pristine and activated substrates are, respectively, marked with colored and black horizontal lines for comparison. The values are in eV per-metal atom etched.","id":"train/atomic-layer-etching/simulation-usecase/31/figure_5","sample_id":"atomic-layer-etching/simulation-usecase/31/figure_5","subset":"multi-axis-chart","split":"train","classification":[{"panel_id":"a","label":"multi-axis chart"},{"panel_id":"b","label":"multi-axis chart"},{"panel_id":"c","label":"multi-axis chart"},{"panel_id":"d","label":"multi-axis chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"Multi-axis chart showing the Gibbs free energy of removal (ΔG_rm at 80 °C) as a function of nitrogen coverage for the Ni/N system. Dual y-axes compare formamidine and formic acid chemistries across multiple crystallographic facets, with bulk reference energies indicated for context.\"},{\"panel_id\":\"b\",\"text\":\"Multi-axis chart showing removal energies versus nitrogen coverage for the Cu/N system. The data compare facet-dependent trends and reveal generally less favorable removal energetics than the corresponding Ni/N system.\"},{\"panel_id\":\"c\",\"text\":\"Multi-axis chart showing removal energies versus oxygen coverage for the Ni/O system. Several surface reconstructions are explicitly labeled, including (100) p(2×2) and (110) missing-row structures, enabling comparison of reconstruction-dependent stability.\"},{\"panel_id\":\"d\",\"text\":\"Multi-axis chart showing removal energies versus oxygen coverage for the Cu/O system. The figure highlights reconstruction-dependent trends, including (110) (2×1)-PR, and compares surface removal energetics against bulk reference states.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Coverage θ_N (Å⁻²) | Blue Solid | Green Dash | Red Dash-Dot | Blue Dash | Purple Dash Dot | Yellow Solid |Teal Dash |\\n|---|---|---|---|---|---|---|---|\\n| 0.01 | ~ 0.80 | ~ 0.40 | ~ 0.3 | ~ 0.5 | ~ 0.6 | ~ 0.5 | ~ 0.4 |\\n| 0.05 | ~ 0.75 | ~ 0.45 | ~ 0.3 | ~ 0.4 | ~ 0.5 | ~ 0.4 | ~ 0.4 |\\n| 0.10 | ~ 0.30 | ~ 0.25 | ~ -0.2| ~ 0.2 | ~ 0.4 | ~ 0.2 | ~ 0.1 |\\n| 0.15 | - | ~ -0.25| ~ -0.6| - | ~ -0.1| ~ -0.3 | ~ -0.3 |\"},{\"panel_id\":\"b\",\"text\":\"| Coverage θ_N (Å⁻²) | Blue Solid | Green Dash | Red Dash-Dot | Blue Dash | Purple Dash Dot | Yellow Solid |Teal Dash |\\n|---|---|---|---|---|---|---|---|\\n| 0.01 | ~ 0.0 | ~ -0.20 | ~ -0.25 | ~ -0.15 | ~ -0.15 | ~ -0.10 | ~ -0.10 |\\n| 0.05 | ~ 0.5 | ~ -0.15 | ~ -0.20 | ~ -0.20 | ~ -0.05 | ~ -0.10 | ~ -0.10 |\\n| 0.10 | ~ -0.2 | ~ -0.30 | ~ 0.40 | - | - | ~ -0.25 | ~ -0.20 | \\n| 0.15 | ~ -0.5 | - | - | - | - | - | - |\"},{\"panel_id\":\"c\",\"text\":\"| Coverage θ_N (Å⁻²) | Blue Solid | Green Dash | Red Dash-Dot | Blue Dash | Purple Dash Dot | Yellow Solid |Teal Dash |\\n|---|---|---|---|---|---|---|---|\\n| 0.01 | ~ 0.95 | ~ 0.40 | ~ 0.60 | ~ 0.60 | ~ 0.75 | ~ 0.75 | ~ 0.75 |\\n| 0.05 | ~ 0.80 | ~ 0.40 | ~ 0.50 | ~ 0.50 | ~ 0.75 | ~ 0.70 | ~ -0.10 |\\n| 0.10 | ~ 0.25 | ~ 0.20 | ~ 0.00 | ~ 0.25 | ~ 0.25 | ~ 0.20 | ~ -0.20 | \\n| 0.15 | ~ -0.10 | ~ -0.10| ~ -0.50| ~ 0.00 | ~ -0.10| ~ -0.10 | ~ -0.10 |\"},{\"panel_id\":\"d\",\"text\":\"| Coverage θ_N (Å⁻²) | Blue Solid | Green Dash | Red Dash-Dot | Blue Dash | Purple Dash Dot | Yellow Solid |Teal Dash |\\n|---|---|---|---|---|---|---|---|\\n| 0.01 | ~ 0.30 | ~ 0.00 | ~ 0.10 | ~ 0.10 | ~ 0.15 | ~ 0.15 | ~ 0.35 |\\n| 0.05 | ~ 0.20 | ~ 0.10 | ~ 0.00 | ~ 0.10 | ~ 0.20 | ~ 0.05 | ~ 0.15 |\\n| 0.10 | ~ 0.00 | ~ 0.00 | ~ -0.25| ~ -0.05| ~ 0.00 | ~ -0.05 | ~ 0.00 | \\n| 0.15 | ~ -0.40 | ~ -0.25| ~ -0.80| ~ -0.20| ~ -0.20| ~ -0.35 | ~ -0.20 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"It reflects how the removal energy changes with nitrogen coverage, indicating the sensitivity of surface stability to adsorbate density.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"No, the energy generally increases (becomes more positive and less favorable) as coverage increases.\"}]},{\"panel_id\":\"c\",\"items\":[{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Subfigure (c) labels p(2×2) and c(2×2) reconstructions on the (100) surface and (2×1)-MR and (3×1)-MR missing-row reconstructions on the (110) surface. At low oxygen coverage, the (110) (2×1)-MR reconstruction shows the lowest ΔG_rm values, indicating the most favorable removal energetics among the labeled structures.\"}]},{\"panel_id\":\"d\",\"items\":[{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"It suggests that both chemistries follow comparable surface reaction energetics. Either pathway could drive controlled removal if conditions are tuned for the appropriate surface state, giving flexibility in precursor choice for ALE design.\"}]}]","bbox":[{"panel_id":"a","x":9,"y":0,"width":485,"height":365},{"panel_id":"b","x":577,"y":0,"width":484,"height":367},{"panel_id":"c","x":11,"y":414,"width":480,"height":362},{"panel_id":"d","x":578,"y":415,"width":486,"height":359}],"source":"icdar2026-competition-data/train/atomic-layer-etching/simulation-usecase/31/images/figure_5.jpg","provenance":{"source_annotation":"icdar2026-competition-data/train/atomic-layer-etching/simulation-usecase/31/images/figure_5.json","source_content":"icdar2026-competition-data/train/atomic-layer-etching/simulation-usecase/31/content.json","source_pdf":["icdar2026-competition-data/train/atomic-layer-etching/simulation-usecase/31/Thermodynamics of Atomic Layer Etching Chemistry on Copper and Nickel Surfaces from First Principles.pdf"],"main_category":"atomic-layer-etching","sub_category":"simulation-usecase","paper_id":"31","first_classification_panel_id":"a","first_classification_label":"multi-axis chart","caption_source":"content.json:image_caption"},"width":1091,"height":861,"image_format":"jpeg","image_sha256":"e9b1108735d904f7a68b3e3d58430af46b4b8cce645f3411dd96308d9a9175fd","metadata_license":"CC BY 4.0","image_license":"source_publisher_rights_reserved","image_reuse_status":"non_commercial_research_use_only","schema_version":"1.0.0"} | |
| {"file_name":"images/train_atomic-layer-etching_simulation-usecase_31_figure_6.jpg","caption":"Figure 6. Wulff construction-averaged removal step energies as functions of coverage (dashed lines) for (a) nickel and (b) copper systems. The solid lines correspond to the results from the bulk model for the pristine (black) and modified (colored) systems. The values are in eV per-metal atom etched.","id":"train/atomic-layer-etching/simulation-usecase/31/figure_6","sample_id":"atomic-layer-etching/simulation-usecase/31/figure_6","subset":"multi-axis-chart","split":"train","classification":[{"panel_id":"a","label":"multi-axis chart"},{"panel_id":"b","label":"multi-axis chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"Multi-axis line chart showing Wulff-averaged Gibbs free energies of removal (ΔG_rm at 80 °C) for nickel as a function of surface coverage. The left y-axis corresponds to formamidine (HFamd) chemistry and the right y-axis to formic acid (HFA). Horizontal lines indicate bulk reference energies for pristine Ni and modified Ni₃N.\"},{\"panel_id\":\"b\",\"text\":\"Multi-axis line chart showing Wulff-averaged removal energies for copper as a function of coverage. Dual y-axes distinguish formamidine and formic acid chemistries, with bulk reference lines for Cu and Cu₃N included for comparison.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Coverage θ (Å⁻²) | Del_G (HFAmd) (Ni/N)| Del_G (HFA) (Ni/N) | Del_G (HFAmd) (Ni/O)| Del_G (HFA) (Ni/O) |\\n|---|---|---|---|---|\\n| 0.01 | ~ 0.60 | ~ 0.85 | ~ 0.75 | ~ 0.95 |\\n| 0.05 | ~ 0.50 | ~ 0.75 | ~ 0.60 | ~ 0.85 |\\n| 0.10 | ~ 0.35 | ~ 0.60 | ~ 0.20 | ~ 0.50 |\\n| 0.15 | ~ -0.20| ~ 0.00 | ~ -0.10| ~ 0.20 |\\n| 0.20 | - | - | ~ -0.40| ~ -0.20 |\"},{\"panel_id\":\"b\",\"text\":\"| Coverage θ (Å⁻²) | Del_G (HFAmd) (Cu/N)| Del_G (HFA) (Cu/N) | Del_G (HFAmd) (Cu/O)| Del_G (HFA) (Cu/O) |\\n|---|---|---|---|---|\\n| 0.01 | ~ 0.20 | ~ 0.20 | ~ 0.30 | ~ 0.30 |\\n| 0.05 | ~ 0.00 | ~ 0.05 | ~ 0.20 | ~ 0.20 |\\n| 0.10 | ~ -0.20| ~ 0.15 | ~ 0.00 | ~ 0.00 |\\n| 0.15 | - | - | ~ -0.20| ~ -0.10|\\n| 0.20 | - | - | - | - |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"List\",\"answer\":\"-Formamidine (HFamd) — left y-axis, -Formic acid (HFA) — right y-axis\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Bulk reference lines provide a baseline to evaluate if surface removal is energetically accessible. Comparing these helps determine if etching remains surface-limited rather than bulk-like, which is essential for achieving atomic-layer control. It's also useful for identifying when modified surface layers become easier to remove than the pristine bulk material.\"}]},{\"panel_id\":\"b\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Yes/No\",\"answer\":\"No\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The trend suggests that once sufficient surface coverage is achieved, removal becomes thermodynamically favorable without requiring continued activation. This behavior supports self-limiting etching, where removal naturally saturates after one atomic layer is processed.\"}]}]","bbox":[{"panel_id":"a","x":1,"y":17,"width":667,"height":433},{"panel_id":"b","x":3,"y":473,"width":664,"height":426}],"source":"icdar2026-competition-data/train/atomic-layer-etching/simulation-usecase/31/images/figure_6.jpg","provenance":{"source_annotation":"icdar2026-competition-data/train/atomic-layer-etching/simulation-usecase/31/images/figure_6.json","source_content":"icdar2026-competition-data/train/atomic-layer-etching/simulation-usecase/31/content.json","source_pdf":["icdar2026-competition-data/train/atomic-layer-etching/simulation-usecase/31/Thermodynamics of Atomic Layer Etching Chemistry on Copper and Nickel Surfaces from First Principles.pdf"],"main_category":"atomic-layer-etching","sub_category":"simulation-usecase","paper_id":"31","first_classification_panel_id":"a","first_classification_label":"multi-axis chart","caption_source":"content.json:image_caption"},"width":670,"height":900,"image_format":"jpeg","image_sha256":"8bfd0231d7ff76934762f62cdbde0ba2cec378c2b943bcd98a255b4a3fb799db","metadata_license":"CC BY 4.0","image_license":"source_publisher_rights_reserved","image_reuse_status":"non_commercial_research_use_only","schema_version":"1.0.0"} | |
| {"file_name":"images/train_atomic-layer-etching_simulation-usecase_6_fig_6.jpg","caption":"FIG. 6. Amount of Si etched (in units of $\\mathrm{A}$ ) (blue lines) and Cl uptake (red lines) as a function of cycle number for $\\mathrm{Si - Cl_2 - Ar}$ ALE simulation using $100\\mathrm{eV}$ $\\mathrm{Ar^+}$ . One cycle consists of $28.2\\mathrm{ML}$ of $\\mathrm{Cl}_2$ molecule impacts and $28.2\\mathrm{ML}$ of $\\mathrm{Ar^+}$ impacts. The surface is nearly saturated with Cl before ion bombardment starts each cycle. Ion bombardment reduces the Cl content of the layer (shown as an equivalent surface coverage of $\\mathrm{Cl / cm^2}$ ).","id":"train/atomic-layer-etching/simulation-usecase/6/fig_6","sample_id":"atomic-layer-etching/simulation-usecase/6/fig_6","subset":"multi-axis-chart","split":"train","classification":[{"panel_id":"a","label":"multi-axis chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"This multi-axis chart tracks the progression of Si etching and Cl uptake across five repeated cycles under 100 eV Ar⁺ ion bombardment. The red curve (left y-axis) indicates the amount of Si etched per cycle, which increases sharply during each ion bombardment step and then returns to baseline, showing a reproducible, step-like etching profile. Meanwhile, the blue line (right y-axis) demonstrates a continuous, linear increase in chlorine uptake with each cycle, indicating cumulative surface adsorption. Together, these trends reveal a consistent etch–adsorb dynamic where physical sputtering removes Si while chemical Cl incorporation steadily rises, likely due to residual Cl interactions at the surface\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Cycle Number | Si Etched (Å) | Cl Uptake (Cl/cm² × 10¹⁵) |\\n|--------------|----------------|----------------------------|\\n| 0 | 0 | 0.10 |\\n| 1 | 32 | 0.22 |\\n| 2 | 32 | 0.34 |\\n| 3 | 32 | 0.46 |\\n| 4 | 32 | 0.58 |\\n| 5 | 32 | 0.70 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The Si etching profile exhibits a periodic, stepwise increase and reset pattern with each cycle, reflecting ion-driven material removal during plasma exposure. In contrast, the chlorine uptake increases steadily without any resetting, indicating cumulative surface incorporation. This divergence highlights the physical versus chemical nature of each process: sputtering is transient and self-limiting per cycle, whereas adsorption builds progressively with repeated Cl exposure.\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Cl uptake (Cl/cm² × 10¹⁵)\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"List\",\"answer\":\"Cl₂ exposure, Pump and purge, Ar⁺ bombardment, Pump and purge\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Yes/No\",\"answer\":\"Yes\"}]}]","bbox":[{"panel_id":"a","x":0,"y":0,"width":675,"height":428}],"source":"icdar2026-competition-data/train/atomic-layer-etching/simulation-usecase/6/images/fig_6.jpg","provenance":{"source_annotation":"icdar2026-competition-data/train/atomic-layer-etching/simulation-usecase/6/images/fig_6.json","source_content":"icdar2026-competition-data/train/atomic-layer-etching/simulation-usecase/6/content.json","source_pdf":["icdar2026-competition-data/train/atomic-layer-etching/simulation-usecase/6/Dynamics of plasma atomic layer etching Molecular.pdf"],"main_category":"atomic-layer-etching","sub_category":"simulation-usecase","paper_id":"6","first_classification_panel_id":"a","first_classification_label":"multi-axis chart","caption_source":"content.json:img_caption"},"width":675,"height":428,"image_format":"jpeg","image_sha256":"b9ff54633a63f0101dcad3d42c903b3704c574e242610dec7e33238ca40e69d1","metadata_license":"CC BY 4.0","image_license":"source_publisher_rights_reserved","image_reuse_status":"non_commercial_research_use_only","schema_version":"1.0.0"} | |
| {"file_name":"images/train_atomic-layer-etching_simulation-usecase_6_fig_7.jpg","caption":"FIG. 7. Amount of Si etched in units of $\\dot{A}$ (blue lines) and Cl uptake (red lines) during the final cycle for the $\\mathrm{Si - Cl_2 - Ar}$ ALE simulation using $100\\mathrm{eVAr^{+}}$ The final portion of the ion bombardment step (purple dashed-dotted line) is used to calculate the etch yield shown as the purple dashed-dotted line. The purple dashed-dotted line is extrapolated to the estimated experimental $\\mathsf{Ar}^+$ fluence.","id":"train/atomic-layer-etching/simulation-usecase/6/fig_7","sample_id":"atomic-layer-etching/simulation-usecase/6/fig_7","subset":"multi-axis-chart","split":"train","classification":[{"panel_id":"a","label":"multi-axis chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"This multi-axis chart illustrates the relationship between species dosage (Cl₂ or Ar⁺/cm²) and two parameters: Si etching depth and chlorine uptake. The red curve (left y-axis) represents silicon etching at 100 eV Ar⁺, which begins only after Cl₂ exposure has saturated the surface. Etching increases steadily with Ar⁺ dosage and plateaus at ~9.8 Å, indicating the removal of one atomic layer of Si. The blue curve (right y-axis) shows Cl uptake, which rises quickly during Cl₂ exposure and saturates around 1.35 × 10¹⁵ Cl/cm². The dashed lines and arrows illustrate the sequential process: first Cl uptake during Cl₂ exposure, followed by Si etching upon Ar⁺ bombardment, revealing a synergistic etch mechanism.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Species Dosage (Cl₂ or Ar⁺/cm²) × 10¹⁵ | Si Etched (Å) | Cl Uptake (Cl/cm²) × 10¹⁵ |\\n|----------------------------------------|---------------|----------------------------|\\n| 0 | 0 | 0 |\\n| 10 | 0 | 0.75 |\\n| 20 | 0 | 1.25 |\\n| 25 | 0 | 1.35 |\\n| 30 | 1.5 | 1.35 |\\n| 50 | 4 | 1.35 |\\n| 75 | 6.5 | 1.35 |\\n| 100 | 9.0 | 1.35 |\\n| 125 | 9.8 | 1.35 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Chlorine uptake increases rapidly during initial Cl₂ exposure and saturates near 1.35 × 10¹⁵ Cl/cm², showing a self-limiting surface reaction. In contrast, silicon etching does not begin during Cl₂ exposure but increases progressively during the Ar⁺ dosage phase. This distinction highlights that Cl₂ alone does not etch Si, but pre-functionalises the surface, enabling etching when energy is provided by Ar⁺ bombardment. The etching depth reaches a maximum around 9.8 Å, corresponding to the removal of a single Si layer.\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"1.35 × 10¹⁵ Cl/cm².\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"List\",\"answer\":\"Cl adsorption, Surface saturation, Ar⁺ bombardment, Si–Cl bond breaking, Silicon etching\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Yes/No\",\"answer\":\"No\"}]}]","bbox":[{"panel_id":"a","x":0,"y":0,"width":669,"height":425}],"source":"icdar2026-competition-data/train/atomic-layer-etching/simulation-usecase/6/images/fig_7.jpg","provenance":{"source_annotation":"icdar2026-competition-data/train/atomic-layer-etching/simulation-usecase/6/images/fig_7.json","source_content":"icdar2026-competition-data/train/atomic-layer-etching/simulation-usecase/6/content.json","source_pdf":["icdar2026-competition-data/train/atomic-layer-etching/simulation-usecase/6/Dynamics of plasma atomic layer etching Molecular.pdf"],"main_category":"atomic-layer-etching","sub_category":"simulation-usecase","paper_id":"6","first_classification_panel_id":"a","first_classification_label":"multi-axis chart","caption_source":"content.json:img_caption"},"width":669,"height":425,"image_format":"jpeg","image_sha256":"622fa2e49e26029251a18f19be61e72c00740aa2a49bbc5484896e13bdc66f42","metadata_license":"CC BY 4.0","image_license":"source_publisher_rights_reserved","image_reuse_status":"non_commercial_research_use_only","schema_version":"1.0.0"} | |
| {"file_name":"images/train_atomic-layer-etching_simulation-usecase_8_fig_11.jpg","caption":"FIG. 11. Al2p XPS peak area and peak binding energy after 30 ALE cycles as a function of table temperature. The binding energies of $\\mathrm{AlF}_3$ and $\\mathrm{Al}_2\\mathrm{O}_3$ have been included as dashed lines to show the transition between the two binding environments. The insets show the fitted Al2p XPS spectra for the films etched at 100 and $200^{\\circ}\\mathrm{C}$ .","id":"train/atomic-layer-etching/simulation-usecase/8/fig_11","sample_id":"atomic-layer-etching/simulation-usecase/8/fig_11","subset":"multi-axis-chart","split":"train","classification":[{"panel_id":"a","label":"multi-axis chart"},{"panel_id":"b","label":"spectra chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The chart shows the relationship between the table temperature and the Al2p peak area and peak binding energy of AlF3 and Al2O3.\"},{\"panel_id\":\"b\",\"text\":\"The inset spectra chart displays the intensity of Al2p peaks at different binding energies.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Table temperature°C | Al2p peak area (CPS eV) | Peak binding energy (eV) |\\n|---|---|---|\\n| 100 | 17000 | 76.0 |\\n| 125 | 17000 | 76.0 |\\n| 150 | 15000 | 76.0 |\\n| 175 | 5000 | 75.7|\\n| 200 | 2000 | 74.7 |\\n| 225 | 2000 | 74.5 |\\n| 250 | 2000 | 74.7 |\\n| 275 | 2000 | 74.5 |\\n| 300 | 2000 | 74.6 |\"},{\"panel_id\":\"b\",\"text\":\"| Binding energy (eV) | Intensity (CPS) |\\n|---|---|\\n| 72 | 0 |\\n| 74 | 0 |\\n| 76 | Peak |\\n| 78 | 0 |\\n| 80 | 0 |\"}]","vqa":"[{\"panel_id\":\"b\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"It seems that upon etching there is a shift in binding energy.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"All data points are from films after 30 cycles of etching. There is way less al2p peak intensity for 200 degrees celcius indicating a higher EPC.\"}]},{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"A reason could be that fluorine species get more volatile at higher temperatures, meaning they will leave the surface, resulting in more Al2O3.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Temperatures above 200 degrees celcius show almost no fluorine.\"}]}]","bbox":[{"panel_id":"b","x":246,"y":123,"width":162,"height":160},{"panel_id":"a","x":0,"y":2,"width":669,"height":515}],"source":"icdar2026-competition-data/train/atomic-layer-etching/simulation-usecase/8/images/fig_11.jpg","provenance":{"source_annotation":"icdar2026-competition-data/train/atomic-layer-etching/simulation-usecase/8/images/fig_11.json","source_content":"icdar2026-competition-data/train/atomic-layer-etching/simulation-usecase/8/content.json","source_pdf":["icdar2026-competition-data/train/atomic-layer-etching/simulation-usecase/8/Isotropic atomic layer etching of GaN using SF6 plasma and AI(CH3)3.pdf"],"main_category":"atomic-layer-etching","sub_category":"simulation-usecase","paper_id":"8","first_classification_panel_id":"a","first_classification_label":"multi-axis chart","caption_source":"content.json:img_caption"},"width":675,"height":523,"image_format":"jpeg","image_sha256":"63218294d8562bd5b4928ee54197b794d5e5ad6a6b60642f93d478db8a50da59","metadata_license":"CC BY 4.0","image_license":"source_publisher_rights_reserved","image_reuse_status":"non_commercial_research_use_only","schema_version":"1.0.0"} | |