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 size: 322,193 Bytes
4b6a819 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 | {"file_name":"images/test_atomic-layer-deposition_experimental-usecase_10_fig_13.jpg","caption":"FIG. 13. (Color online) Results from Monte Carlo simulations investigating the influence of surface recombination of radicals during plasma-assisted ALD (Ref. 304). (a) Equivalent thickness profile in a trench of aspect ratio 10 for different deposition regimes, obtained for various combinations of values for the sticking probability, $s$ , and surface recombination probability, $r$ . The positions within the trench labeled 0 and $100\\%$ correspond to the trench opening and trench bottom, respectively. Note that both recombination-limited cases show almost perfect overlap. (b) The dose required to reach saturation in trenches with aspect ratios of 10 and 30 for nonzero values of $r$ . This dose is normalized to the dose required to reach saturation in these trenches when $r = 0$ . For the simulations $s = 0.01$ was assumed. From H.C.M. Knoops et al., J. Electrochem. Soc. 157, G241 (2010). Reproduced with permission of ECS—The Electrochemical Society.","id":"test/atomic-layer-deposition/experimental-usecase/10/fig_13","sample_id":"atomic-layer-deposition/experimental-usecase/10/fig_13","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"},{"panel_id":"b","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"Thickness profiles across a trench reveal how different growth regimes—reaction-limited, diffusion-limited, and recombination-limited—affect uniformity, with recombination-limited cases showing more variation along the depth.\"},{\"panel_id\":\"b\",\"text\":\"The dose necessary for saturation increases with higher surface recombination probability, and this effect becomes more pronounced at higher aspect ratios (AR = 30).\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Distance into Trench (%) | Reaction-limited (s = 0.01; r = 0) | Diffusion-limited (s = 1; r = 0) | Recombination-limited (s = 0.01; r = 0.9) | Recombination-limited (s = 1; r = 0.9) |\\n|--------------------------|------------------------------------|----------------------------------|-------------------------------------------|----------------------------------------|\\n| 0 | 100 | 100 | 100 | 100 | \\n| 20 | 98 | 100 | 95 | 97 |\\n| 40 | 95 | 99 | 90 | 93 |\\n| 60 | 92 | 98 | 88 | 91 |\\n| 80 | 90 | 98 | 86 | 90 |\\n| 100 | 92 | 99 | 88 | 92 |\"},{\"panel_id\":\"b\",\"text\":\"| Surface Recombination Probability (r) | Dose Necessary for Saturation (Aspect Ratio = 10) | Dose Necessary for Saturation (Aspect Ratio = 30) |\\n|---------------------------------------|---------------------------------------------------|---------------------------------------------------|\\n| 0.00 | 1 | 1 |\\n| 0.05 | 3 | 10 |\\n| 0.10 | 6 | 25 |\\n| 0.15 | 10 | 50 |\\n| 0.20 | 20 | 70 |\\n| 0.25 | 40 | 85 | \\n| 0.30 | 60 | 100 |\"}]","vqa":"[{\"panel_id\":\"b\",\"items\":[{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"List\",\"answer\":\"Surface recombination probability (r), Aspect ratio of the structure\"}]},{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Thickness remains nearly uniform for diffusion-limited and reaction-limited processes but decreases notably deeper into the trench for recombination-limited conditions, indicating non-uniform film growth.\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Film thickness (%) as a function of trench depth.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Yes/No\",\"answer\":\"Yes\"}]}]","bbox":[{"panel_id":"a","x":17,"y":14,"width":670,"height":508},{"panel_id":"b","x":5,"y":556,"width":682,"height":472}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/10/images/fig_13.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/10/images/fig_13.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/10/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/10/H. B. Profijt et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"experimental-usecase","paper_id":"10","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:img_caption"},"width":689,"height":1033,"image_format":"jpeg","image_sha256":"b6751d4682ee73e5c954d57032cb22cc25c9ad21eeb17b4363db8c865e9080ac","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/test_atomic-layer-deposition_experimental-usecase_10_fig_14.jpg","caption":"FIG. 14. (Color online) Experiments proving that VUV radiation from the plasma affects the surface passivation of crystalline Si by $\\mathrm{Al}_2\\mathrm{O}_3$ when deposited by plasma-assisted ALD (Refs. 43, 303). After annealing, the wafers were exposed to an $\\mathrm{O_2}$ plasma for various exposure times. (a) The effective charge carrier lifetimes degraded for increasing exposure times at a rate which increased with increasing VUV radiation present in the plasma (higher intensity for higher power and/or lower pressure). (b) Results are also given for the situation in which the substrate is covered by quartz and $\\mathrm{MgF_2}$ windows that respectively block and do not block the VUV photons of $9.5\\mathrm{eV}$ . From H.B. Profijt et al., ECS Trans. 33, 61 (2010). Reproduced with permission of ECS—The Electrochemical Society.","id":"test/atomic-layer-deposition/experimental-usecase/10/fig_14","sample_id":"atomic-layer-deposition/experimental-usecase/10/fig_14","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"},{"panel_id":"b","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"Increasing plasma power accelerates carrier lifetime reduction, with 500 W exposure causing rapid degradation compared to 100 W.\"},{\"panel_id\":\"b\",\"text\":\"The use of different chamber windows (no window, quartz, MgF₂) alters the plasma interaction with the surface, showing that optical transparency and window material strongly influence degradation rates.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| O₂ Plasma Exposure Time (s) | 100 W, 7.5 mTorr | 500 W, 7.5 mTorr | 100 W, 7.5 mTorr (quartz window) |\\n|----------------------------|------------------|------------------|----------------------------------|\\n| 0 | 10.0 | 9.8 | 10.0 |\\n| 25 | 6.5 | 4.5 | 9.5 |\\n| 50 | 2.0 | 1.0 | 9.0 |\\n| 75 | 0.9 | 0.6 | 8.5 |\\n| 100 | 0.5 | 0.3 | 8.0 |\\n| 125 | 0.3 | 0.2 | 7.5 |\\n| 150 | 0.2 | 0.1 | 7.0 |\"}]","vqa":"[{\"panel_id\":\"b\",\"items\":[{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Effective carrier lifetime (ms).\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Yes/No\",\"answer\":\"Yes\"}]},{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The effective carrier lifetime decreases exponentially with increasing O₂ plasma exposure time, with higher plasma power or direct exposure (no window) causing faster degradation.\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"List\",\"answer\":\"Plasma power (100 W and 500 W), Pressure (7.5 mTorr), Window material (none, quartz, MgF₂)\"}]}]","bbox":[{"panel_id":"a","x":42,"y":11,"width":632,"height":462},{"panel_id":"b","x":40,"y":511,"width":636,"height":466}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/10/images/fig_14.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/10/images/fig_14.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/10/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/10/H. B. Profijt et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"experimental-usecase","paper_id":"10","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:img_caption"},"width":678,"height":978,"image_format":"jpeg","image_sha256":"eae3238a3e3999c1efb9ceab9a2c05ea03521258070446004a7458a8396d7b2d","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/test_atomic-layer-deposition_experimental-usecase_10_fig_16.jpg","caption":"FIG. 16. Barrier failure temperatures for TaN and Ta films deposited by PVD and plasma-assisted ALD (Refs. 204, 219). After deposition the TaN and Ta layers were capped by $200 \\mathrm{nm}$ thick Cu films using the PVD technique. Barrier failure temperatures were determined by monitoring the disappearance of the Cu (111) peak by X-ray diffraction measurements. The ALD films were deposited using an $\\mathrm{H}_2 / \\mathrm{N}_2$ plasma film in TaN and an $\\mathrm{H}_2$ plasma for Ta. From H. Kim et al., J. Appl. Phys. 95, 5848 (2004). Reprinted with permission. Copyright 2004, American Institute of Physics.","id":"test/atomic-layer-deposition/experimental-usecase/10/fig_16","sample_id":"atomic-layer-deposition/experimental-usecase/10/fig_16","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The figure shows how the failure temperature increases with film thickness for PE-ALD and PVD TaN and Ta films. Across all materials, thicker films withstand higher temperatures, with ALD-based films generally exhibiting higher thermal stability than their PVD counterparts.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Film thickness (nm) | PE-ALD TaN | PVD TaN | PE-ALD Ta | PVD Ta |\\n|---|---|---|---|---|\\n| 0 | 580 | 590 | 600 | 610 |\\n| 5 | 670 | 680 | 690 | 700 |\\n| 10 | 760 | 770 | 780 | 790 |\\n| 15 | 850 | 860 | 870 | 880 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Failure temperature increases with film thickness for all materials, with PE-ALD films exhibiting higher stability compared to PVD films across the measured range.\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Failure temperature (°C).\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"List\",\"answer\":\"(PE-ALD) TaN, (PVD), TaN\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Yes/No\",\"answer\":\"Yes\"}]}]","bbox":[{"panel_id":"a","x":0,"y":0,"width":605,"height":447}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/10/images/fig_16.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/10/images/fig_16.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/10/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/10/H. B. Profijt et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"experimental-usecase","paper_id":"10","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:img_caption"},"width":605,"height":447,"image_format":"jpeg","image_sha256":"7f5f55ebe1f0d3ac33f9b16ea3ace68c7acdad2c5c0a37f2a80fbe87e5cb4d3f","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/test_atomic-layer-deposition_experimental-usecase_10_fig_4.jpg","caption":"Fig. 4. (Color online) Ion energy distribution as measured by a retarding field energy analyzer (RFEA) in $\\mathrm{O}_2$ , $\\mathrm{H}_2$ and $\\mathrm{N}_2$ plasmas (operating pressure: 8 mTorr; plasma power: 100 W) used for remote plasma-assisted ALD. The RFEA was positioned at the substrate stage. Measurements were performed in the home-built ALD-I reactor installed at Eindhoven University of Technology. Due to non-ideal effects such as capacitive coupling, the ion energies measured are higher than those measured in the Oxford Instruments FlexAL reactor, which are reported elsewhere (Ref. 303).","id":"test/atomic-layer-deposition/experimental-usecase/10/fig_4","sample_id":"atomic-layer-deposition/experimental-usecase/10/fig_4","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The figure presents ion energy distribution curves for O₂, H₂, and N₂ plasmas. Each plasma type exhibits a peak in ion energy distribution near 30–35 eV, with the H₂ plasma showing the highest intensity, O₂ plasma a moderate peak, and N₂ plasma the lowest distribution. This comparison highlights differences in ion energetics between the three plasma environments.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Ion Energy (eV) | O₂ Plasma (a.u.) | H₂ Plasma (a.u.) | N₂ Plasma (a.u.) |\\n|-----------------|------------------|------------------|------------------|\\n| 0 | 0.00 | 0.00 | 0.00 |\\n| 10 | 0.05 | 0.10 | 0.00 |\\n| 20 | 0.30 | 0.40 | 0.10 |\\n| 25 | 0.60 | 0.80 | 0.20 |\\n| 30 | 0.90 | 1.00 | 0.30 |\\n| 35 | 0.80 | 0.70 | 0.20 |\\n| 40 | 0.20 | 0.20 | 0.10 |\\n| 50 | 0.00 | 0.00 | 0.00 |\\n| 60 | 0.00 | 0.00 | 0.00 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The H₂ plasma shows the highest ion energy distribution peak around 30–35 eV, followed by O₂ plasma with a moderate peak and N₂ plasma with the lowest intensity.\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Ion energy distribution (arbitrary units.)\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"List\",\"answer\":\"H₂ plasma high energy, O₂ moderate, N₂ low energy and ion impact\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Yes/No\",\"answer\":\"Yes\"}]}]","bbox":[{"panel_id":"a","x":0,"y":0,"width":606,"height":472}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/10/images/fig_4.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/10/images/fig_4.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/10/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/10/H. B. Profijt et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"experimental-usecase","paper_id":"10","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:img_caption"},"width":606,"height":472,"image_format":"jpeg","image_sha256":"33337cd6ce6b8f9ae4de2c78e45af55d0edd55f8d5cca3d0514b4dc0fd8d464a","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/test_atomic-layer-deposition_experimental-usecase_10_fig_8.jpg","caption":"FiG. 8. (Color online) Growth per cycle of $\\mathrm{Al}_2\\mathrm{O}_3$ films as a function of the substrate temperature. The films were deposited by plasma-assisted ALD ( $\\mathrm{O}_2$ plasma) and thermal ALD ( $\\mathrm{H}_2\\mathrm{O}$ ). Two different ALD reactors were used; one operating at $15\\mathrm{mTorr}$ and the other at $170\\mathrm{mTorr}$ (Refs. 43, 323). From S.E. Potts et al., J. Electrochem. Soc. 157, P66 (2010). Reproduced by permission of ECS—The Electrochemical Society.","id":"test/atomic-layer-deposition/experimental-usecase/10/fig_8","sample_id":"atomic-layer-deposition/experimental-usecase/10/fig_8","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The multiple line chart compares growth per cycle as a function of deposition temperature for both O₂ plasma and thermal (H₂O) ALD processes under two pressure conditions (15 mTorr and 170 mTorr). Growth per cycle decreases with temperature for plasma processes, while remaining nearly constant for thermal ALD. Higher pressure generally results in slightly higher growth rates.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Deposition Temperature (°C) | O₂ Plasma (170 mTorr) | O₂ Plasma (15 mTorr) | Thermal (H₂O, 170 mTorr) | Thermal (H₂O, 15 mTorr) |\\n|-----------------------------|----------------------|----------------------|---------------------------|--------------------------|\\n| 50 | 1.7 | 1.6 | 0.8 | 0.7 |\\n| 100 | 1.6 | 1.5 | 0.8 | 0.8 |\\n| 150 | 1.4 | 1.3 | 0.9 | 0.9 |\\n| 200 | 1.2 | 1.1 | 0.9 | 0.9 |\\n| 250 | 1.1 | 1.0 | 0.9 | 0.9 |\\n| 300 | 1.0 | 0.9 | 0.9 | 0.9 |\\n| 350 | 0.9 | 0.8 | 0.9 | 0.9 |\\n| 400 | 0.8 | 0.7 | 0.9 | 0.9 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The growth per cycle decreases with temperature for O₂ plasma ALD but remains almost constant for thermal (H₂O) ALD, indicating that plasma-based processes are more temperature-sensitive.\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"List\",\"answer\":\"O₂ plasma 170 mTorr, O₂ plasma 15 mTorr, H₂O thermal 170 mTorr, H₂O thermal 15 mTorr\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Yes/No\",\"answer\":\"Yes\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Growth per cycle (Å/cycle).\"}]}]","bbox":[{"panel_id":"a","x":0,"y":0,"width":675,"height":520}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/10/images/fig_8.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/10/images/fig_8.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/10/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/10/H. B. Profijt et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"experimental-usecase","paper_id":"10","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:img_caption"},"width":675,"height":520,"image_format":"jpeg","image_sha256":"7fa62c00a7f742804a3143a2cd7c472f3775129593bd67d3726da42f1e5b382f","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/test_atomic-layer-deposition_experimental-usecase_10_fig_9.jpg","caption":"FiG. 9. (Color online) Growth per cycle of $\\mathrm{TiO}_2$ films as a function of the substrate temperature. Plasma-assisted ALD was carried out using $\\mathrm{Ti(O^iPr)_4}$ , $\\mathrm{Ti(Cp^Me)(O^iPr)_3}$ , $\\mathrm{Ti(Cp^*)OMe)_3}$ , and $\\mathrm{Ti(Cp^Me)(NMe_2)_3}$ as precursors in combination with an $\\mathrm{O}_2$ plasma (Refs. 56, 248). Data for thermal ALD with $\\mathrm{H}_2\\mathrm{O}$ (Ref. 325) and $\\mathrm{O}_3$ (Ref. 324) using the widely employed $\\mathrm{Ti(O^iPr)_4}$ precursor are given for comparison. From S.E. Potts et al., J. Electrochem. Soc. 157, P66 (2010). Reproduced by permission of ECS—The Electrochemical Society.","id":"test/atomic-layer-deposition/experimental-usecase/10/fig_9","sample_id":"atomic-layer-deposition/experimental-usecase/10/fig_9","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The multiple line chart compares growth per cycle for several titanium precursors under both plasma and thermal ALD conditions. Plasma ALD processes, using precursors such as Ti(OPr)₄ and Ti(Cp*) derivatives, maintain higher and more stable growth rates across all temperatures, while thermal ALD processes show minimal growth and stronger temperature dependence.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Deposition Temperature (°C) | Ti(OPr)₄ (Plasma) | Ti(Cp*)Y(OPr)₄ (Plasma) | Ti(Cp*)Y(OMe)₄ (Plasma) | Ti(Cp*)X(NMe₂)₃ (Plasma) | Ti(OPr)₄ + H₂O (Thermal) | Ti(OPr)₄ + O₃ (Thermal) |\\n|-----------------------------|-------------------|--------------------------|--------------------------|---------------------------|---------------------------|--------------------------|\\n| 50 | 0.65 | 0.72 | 0.68 | 1.10 | 0.05 | 0.12 |\\n| 100 | 0.64 | 0.70 | 0.67 | 1.08 | 0.04 | 0.11 |\\n| 150 | 0.63 | 0.68 | 0.65 | 1.07 | 0.03 | 0.10 |\\n| 200 | 0.62 | 0.67 | 0.64 | 1.05 | 0.03 | 0.09 |\\n| 250 | 0.62 | 0.67 | 0.64 | 1.04 | 0.03 | 0.09 |\\n| 300 | 0.61 | 0.66 | 0.63 | 1.04 | 0.02 | 0.08 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Plasma ALD shows consistently higher and more stable growth per cycle across all deposition temperatures, while thermal ALD yields lower growth rates that decline further with temperature.\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Deposition temperature.\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"List\",\"answer\":\"Ti(OPr)₄ plasma, Ti(Cp*)Y(OPr)₄ plasma, Ti(Cp*)Y(OMe)₄ plasma, Ti(Cp*)X(NMe₂)₃ plasma, Ti(OPr)₄ + H₂O thermal, Ti(OPr)₄ + O₃ thermal.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Yes/No\",\"answer\":\"Yes\"}]}]","bbox":[{"panel_id":"a","x":0,"y":0,"width":681,"height":525}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/10/images/fig_9.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/10/images/fig_9.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/10/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/10/H. B. Profijt et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"experimental-usecase","paper_id":"10","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:img_caption"},"width":681,"height":525,"image_format":"jpeg","image_sha256":"89831d0049dc299f6c32ec14efd020630b262814d55c708174cde431e3c5350a","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/test_atomic-layer-deposition_experimental-usecase_15_fig_2.jpg","caption":"FIG. 2. Normalized capacitance of BST films grown on $\\mathrm{MgO}$ $\\mathrm{TiO_2 / Si}$ and $\\mathrm{Ta}_2\\mathrm{O}_5 / \\mathrm{Si}$ substrates.","id":"test/atomic-layer-deposition/experimental-usecase/15/fig_2","sample_id":"atomic-layer-deposition/experimental-usecase/15/fig_2","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The chart shows the normalized capacitance of BST films with different dielectric layers (BST/TiO2/Si, BST/Ta2O5/Si, BST/MgO) as a function of applied voltage. The tunability of capacitance is indicated at specific voltages.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Applied Voltage (V) | Normalized Capacitance |\\n|---------------------|------------------------|\\n| 0 | 1.0 |\\n| 5 | 0.7 |\\n| 10 | 0.4 |\\n| 35 | 0.6 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"At an applied voltage of 10 V, the tunability values of BST films grown on Ta2O5/Si and TiO2 /Si were 53.1% and 72.9%, respectively and 20.7% for films on MgO single crystal substrate.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Improved tunability and, thus, higher frequency tuning agility, is possible with TiO2 or Ta2O5 buffer layers which increase the overall average dielectric constant and thereby minimize electric field attenuation\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Yes/No\",\"answer\":\"No\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"BST films grown on MgO single crystal substrates showed highly (100) preferred orientation and reached a tunability of only 41.6% with a much higher applied voltage of 35 V. The measured\\ncapacitance of BST grown on TiO2/Si,Ta2O5 / Si, and MgO substrates at zero bias voltage was 125.8, 30.8, and 9.95 pF, respectively. In comparison to parallel-plate capacitors, coplanar designs generally require higher control voltages and\\noffer lower tunability, potential disadvantages of the IDC structure. The drive voltage needed to obtain appropriate tunability is very high in the BST/MgO structure given the low dielectric constant of the MgO single crystal substrate\"}]}]","bbox":[{"panel_id":"a","x":1,"y":2,"width":572,"height":566}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/15/images/fig_2.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/15/images/fig_2.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/15/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/15/Il-Doo Kim et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"experimental-usecase","paper_id":"15","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:image_caption"},"width":572,"height":567,"image_format":"jpeg","image_sha256":"1d59fc6a4a86573f6a7b9735ad1cdc0fe0f248b22080cd2ed01684bf61358273","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/test_atomic-layer-deposition_experimental-usecase_19_fig1.jpg","caption":"Fig.1 Deposition rate of $\\mathrm{ZrO_2}$ thin films from the $(\\mathrm{CpMe})_2\\mathrm{Zr(OMe)Me}$ and $(\\mathrm{CpMe})_2\\mathrm{ZrMe}_2$ precursors and ozone as function of the deposition temperature. Pulsing times were $1.0$ and $2.0~\\mathrm{s}$ for the $\\mathrm{Zr}$ precursor and $\\mathrm{O}_3$ respectively. Lines are guides to the eye.","id":"test/atomic-layer-deposition/experimental-usecase/19/fig1","sample_id":"atomic-layer-deposition/experimental-usecase/19/fig1","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The multiple line chart shows the deposition rate (Å/cycle) of Zirconia (ZrO₂ ) thin films from two different zirconium precursors—(CpMe)₂Zr(OMe)Me and (CpMe)₂ZrMe₂—using O₃ as the oxidant, plotted against deposition temperature (250–500 °C). The growth rate seems to increase with increase in temperature and the behaviours are slightly different for the two precursors.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Deposition Temperature (°C) | Growth Rate (Å/cycle) | Process |\\n|-----------------------------|------------------------|-----------------------------|\\n| 250 | 0.24 | (CpMe)₂Zr(OMe)Me / O₃ |\\n| 250 | 0.21 | (CpMe)₂ZrMe₂ / O₃ |\\n| 300 | 0.56 | (CpMe)₂Zr(OMe)Me / O₃ |\\n| 300 | 0.52 | (CpMe)₂ZrMe₂ / O₃ |\\n| 350 | 0.64 | (CpMe)₂Zr(OMe)Me / O₃ |\\n| 350 | 0.56 | (CpMe)₂ZrMe₂ / O₃ |\\n| 400 | 0.75 | (CpMe)₂Zr(OMe)Me / O₃ |\\n| 400 | 0.58 | (CpMe)₂ZrMe₂ / O₃ |\\n| 450 | 0.88 | (CpMe)₂Zr(OMe)Me / O₃ |\\n| 450 | 0.71 | (CpMe)₂ZrMe₂ / O₃ |\\n| 500 | 1.10 | (CpMe)₂Zr(OMe)Me / O₃ |\\n| 500 | — | (CpMe)₂ZrMe₂ / O₃ |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"1. Precursors Used:\\n\\n(CpMe)₂Zr(OMe)Me\\n\\n(CpMe)₂ZrMe₂\\n\\n2. Experimental Conditions for One ALD Cycle:\\n\\nZr Precursor Pulse Time: 1.0 s\\n\\nO₃ (Oxidant) Pulse Time: 2.0 s\\n\\nDeposition Temperature Range: 250–500 °C\\n\\nOxidant: O₃ (Ozone)\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"(CpMe)₂Zr(OMe)Me provides a stable, low deposition rate (~0.5 Å/cycle) from 250–400 °C, while (CpMe)₂ZrMe₂ only achieves a stable rate above 350 °C. The balance is between low-temperature capability and growth rate. Choosing the (OMe)Me precursor enables processing below 350 °C, but at the cost of a roughly 2x slower growth rate compared to the Me₂ precursor at its optimal temperature.\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"It reduces the deposition rate by approximately half (from ~1.0 Å/cycle to ~0.5 Å/cycle).\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The (CpMe)₂Zr(OMe)Me precursor must be used, as it shows a stable rate at 300 °C. The consequence is a slower growth rate, which will increase the total deposition time per wafer to achieve the target film thickness. This reduces tool throughput and increases the cost of the manufacturing step.\"}]}]","bbox":[{"panel_id":"a","x":1,"y":2,"width":419,"height":346}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/19/images/fig1.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/19/images/fig1.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/19/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/19/Jaakko Niinisto et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"experimental-usecase","paper_id":"19","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:image_caption"},"width":422,"height":350,"image_format":"jpeg","image_sha256":"19570d36026756cad3bafec249f9b2bbdf172a421a9a183404c5416b86bd203f","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/test_atomic-layer-deposition_experimental-usecase_19_fig2.jpg","caption":"Fig.2 Deposition rate of $\\mathrm{ZrO_2}$ thin films grown from $(\\mathrm{CpMe})_2\\mathrm{Zr(OMe)Me}$ and $(\\mathrm{CpMe})_2\\mathrm{ZrMe}_2$ as function of the $\\mathrm{Zr}$ precursor pulse length at the substrate temperatures indicated. Pulsing time was $2.0~\\mathrm{s}$ for $\\mathrm{O}_3$ . Lines are guides to the eye.","id":"test/atomic-layer-deposition/experimental-usecase/19/fig2","sample_id":"atomic-layer-deposition/experimental-usecase/19/fig2","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The multiple chart shows how the growth rate of ZrO₂ thin films changes with zirconium precursor pulse length for two precursors—(CpMe)₂Zr(OMe)Me and (CpMe)₂ZrMe₂—using O₃ as the oxidant. Measurements were taken at 400 °C and 350 °C. As the pulse length increases, the growth rate approaches saturation, but the two precursors show slightly different saturation behaviors. The O₃ pulse time was fixed at 2.0 s.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Zr Precursor Pulse Length (s) | Growth Rate (Å/cycle) | Process |\\n|-------------------------------|------------------------|-----------------------------------------------|\\n| 0.5 | — | (CpMe)₂Zr(OMe)Me / O₃ at 400°C |\\n| 0.5 | — | (CpMe)₂Zr(OMe)Me / O₃ at 350°C |\\n| 0.5 | — | (CpMe)₂ZrMe₂ / O₃ at 350°C |\\n| 1.0 | 0.72 | (CpMe)₂Zr(OMe)Me / O₃ at 400°C |\\n| 1.0 | 0.60 | (CpMe)₂Zr(OMe)Me / O₃ at 350°C |\\n| 1.0 | 0.52 | (CpMe)₂ZrMe₂ / O₃ at 350°C |\\n| 1.5 | 0.72 | (CpMe)₂Zr(OMe)Me / O₃ at 400°C |\\n| 1.5 | 0.62 | (CpMe)₂Zr(OMe)Me / O₃ at 350°C |\\n| 1.5 | — | (CpMe)₂ZrMe₂ / O₃ at 350°C |\\n| 2.0 | 0.75 | (CpMe)₂Zr(OMe)Me / O₃ at 400°C |\\n| 2.0 | 0.65 | (CpMe)₂Zr(OMe)Me / O₃ at 350°C |\\n| 2.0 | 0.56 | (CpMe)₂ZrMe₂ / O₃ at 350°C |\\n| 2.5 | 0.75 | (CpMe)₂Zr(OMe)Me / O₃ at 400°C |\\n| 2.5 | 0.65 | (CpMe)₂Zr(OMe)Me / O₃ at 350°C |\\n| 2.5 | — | (CpMe)₂ZrMe₂ / O₃ at 350°C |\\n| 3.0 | — | (CpMe)₂Zr(OMe)Me / O₃ at 400°C |\\n| 3.0 | 0.66 | (CpMe)₂Zr(OMe)Me / O₃ at 350°C |\\n| 3.0 | 0.55 | (CpMe)₂ZrMe₂ / O₃ at 350°C |\\n| 3.5 | — | (CpMe)₂Zr(OMe)Me / O₃ at 400°C |\\n| 3.5 | — | (CpMe)₂Zr(OMe)Me / O₃ at 350°C |\\n| 3.5 | — | (CpMe)₂ZrMe₂ / O₃ at 350°C |\\n| 30 | — | (CpMe)₂Zr(OMe)Me / O₃ at 400°C |\\n| 30 | 0.62 | (CpMe)₂Zr(OMe)Me / O₃ at 350°C |\\n| 30 | 0.52 | (CpMe)₂ZrMe₂ / O₃ at 350°C |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"List\",\"answer\":\"Variable Zr precursor pulse (x s)\\n\\n, Purge\\n\\n, Fixed O₃ pulse (2.0 s)\\n\\n, Purge\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The (CpMe)₂Zr(OMe)Me precursor at 400°C reaches saturation with a shorter pulse length (~1.5 s) than the (CpMe)₂ZrMe₂ precursor at 350°C, which requires a longer pulse (~2.5 s) to saturate. This indicates the (OMe)Me precursor reacts and saturates the surface more rapidly at a higher temperature, while the Me₂ precursor has slower surface reaction kinetics at 350°C.\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Replacing one -OMe ligand with a second -Me ligand.\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"(CpMe)₂Zr(OMe)Me at 350°C is preferable. It shows a clear, achievable saturation plateau at this temperature, ensuring self-limiting growth for uniformity. Using the Me₂ precursor at 350°C would require a longer, harder-to-control pulse to reach saturation, risking incomplete reactions and less uniform films.\"}]}]","bbox":[{"panel_id":"a","x":0,"y":2,"width":475,"height":352}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/19/images/fig2.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/19/images/fig2.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/19/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/19/Jaakko Niinisto et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"experimental-usecase","paper_id":"19","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:image_caption"},"width":475,"height":358,"image_format":"jpeg","image_sha256":"7523d44a3e1cd01d01e4aa3cb1bce322ab89f98d891775ff2c91745cb128db25","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/test_atomic-layer-deposition_experimental-usecase_19_fig_7.jpg","caption":"Fig. 7 Current density-voltage (a) and Poole-Frenkel (b) plots of the current-voltage dependences measured on some representative $\\mathrm{ZrO_2}$ films deposited from precursors indicated by labels. Layer thicknesses are also given in the labels.","id":"test/atomic-layer-deposition/experimental-usecase/19/fig_7","sample_id":"atomic-layer-deposition/experimental-usecase/19/fig_7","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"},{"panel_id":"b","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The muliple line chart shows Current Density-Voltage (J-V) plots for ZrO₂ films of varying thickness (3.6–13.1 nm) deposited from two different precursors. The analysis identifies dominant leakage current mechanisms: Schottky emission (interface-limited) at lower fields.\"},{\"panel_id\":\"b\",\"text\":\"The muliple line chart shows Poole-Frenkel plots for ZrO₂ films of varying thickness (3.6–13.1 nm) deposited from two different precursors. The analysis identifies dominant leakage current mechanisms: Poole-Frenkel emission (bulk trap-limited) at higher fields.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Voltage, V | Current density, A/cm² |\\n|---|---|\\n| 0 | 10^-9 |\\n| 1 | 10^-8 |\\n| 2 | 10^-7 |\\n| 3 | 10^-6 |\\n| 4 | 10^-5 |\\n| 5 | 10^-4 |\\n| 6 | 10^-3 |\\n| 7 | 10^-2 |\\n| 8 | 10^-1 |\"},{\"panel_id\":\"b\",\"text\":\"| E^(1/2), (MV/cm)^(1/2) | ln(J/E) |\\n|---|---|\\n| 0.0 | -20 |\\n| 0.5 | -15 |\\n| 1.0 | -10 |\\n| 1.5 | -5 |\\n| 2.0 | 0 |\\n| 2.5 | 5 |\\n| 3.0 | 10 |\\n| 3.5 | 15 |\\n| 4.0 | 20 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Deposit a top electrode (like Al/Ag/Au) to form a capacitor.\\n\\nApply a series of increasing voltage biases to the capacitor.\\n\\nMeasure the resulting current flowing through the film at each applied voltage.\\nThe final critical step is measuring the current, which directly gives the leakage density (J) for that specific voltage (V), allowing the trace to be plotted.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"(CpMe)₂Zr(OMe)Me (3.9 nm): Pro: Lower leakage current (~10⁻⁷ A/cm² at 2V).\\n\\n(CpMe)₂ZrMe₂ (3.6 nm): Con: Higher leakage current (~10⁻⁴ A/cm² at 2V).\\n\\n(CpMe)₂Zr(OMe)Me precursor is better for low leakage.\"}]},{\"panel_id\":\"b\",\"items\":[{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The slope is proportional to the square root of the trap's dynamic dielectric constant and its energy depth. A steeper slope (like for the (CpMe)₂ZrMe₂ film) suggests a lower dynamic dielectric constant or different trap energetics. The OMe ligand in the (CpMe)₂Zr(OMe)Me precursor may incorporate oxygen differently, potentially passivating defects or creating a distinct trap distribution with a shallower effective energy level, leading to its different slope.\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"This implies that in thin films (3.9 nm), the dominant leakage mechanism shifts from bulk trap-assisted (Poole-Frenkel) to a different mechanism, most likely direct tunneling. Tunneling current is thickness-sensitive and can be high and unpredictable.For a capacitor requiring long-term charge retention, this makes very thin ZrO₂ films unreliable, as charge will leak rapidly via tunneling, regardless of improvements to bulk film quality.\"}]}]","bbox":[{"panel_id":"a","x":1,"y":0,"width":515,"height":382},{"panel_id":"b","x":5,"y":395,"width":505,"height":382}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/19/images/fig_7.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/19/images/fig_7.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/19/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/19/Jaakko Niinisto et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"experimental-usecase","paper_id":"19","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:image_caption"},"width":514,"height":781,"image_format":"jpeg","image_sha256":"1a79f653c114dec6ff9881b9d460a0c98ca30e5621b8c616767536e2457dfcfc","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/test_atomic-layer-deposition_experimental-usecase_20_fig_1.jpg","caption":"Fig. 1 Composition of deposited films versus percentage of titanium oxide sub-cycles. Experimental data are compared with a model based on surface utilization10 and the growth rates of the individual binary oxides.","id":"test/atomic-layer-deposition/experimental-usecase/20/fig_1","sample_id":"atomic-layer-deposition/experimental-usecase/20/fig_1","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The chart shows the titanium concentration (at.% relative to total metal content) as a function of TiO₂ subcycles (%). It compares experimental data with theoretical values from a surface utilization model with calculated values from individual growth rates. The surface utilization model better predicts experimental values.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| TiO₂ subcycles (%) | Titanium concentration (at.%): Experimental data | Titanium concentration (at.%): from surface utilization model | Titanium concentration (at.%): calculated from individual growth rates |\\n|---|---|---|---|\\n| 0.0 | 0.0 | 0.0 | 0.0 |\\n| 5.8 | 47.2 | 25.3 | 13.5 |\\n| 10.9 | 58.1 | 40.3 | 23.3 |\\n| 13.8 | 62.5 | 47.4 | 29.0 |\\n| 17.0 | 62.4 | 52.8 | 32.9 |\\n| 19.8 | 63.0 | 57.7 | 37.7 |\\n| 25.0 | 68.7 | 64.6 | 45.3 |\\n| 29.8 | 77.5 | 70.3 | 51.0 |\\n| 32.8 | 76.5 | 73.2 | 54.5 |\\n| 50.2 | 84.6 | 84.5 | 71.2 |\\n| 67.2 | 93.5 | 92.0 | 83.3 |\\n| 75.2 | 94.2 | 94.5 | 88.3 |\\n| 100 | 100 | 100 | 100 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Surface utilization model.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"The calculations generally underestimate the experimental values.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Approx. 5-6%.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"All three datasets intersect at 100% TiO₂ subcycles, which corresponds to 100 at.% Ti. It is physically necessary for all methods to reach this value because at 100% subcycles, the deposition consists exclusively of titanium precursor pulses. Therefore, the resulting film is pure titanium oxide, physically forcing the cation composition to be 100% titanium.\"}]}]","bbox":[{"panel_id":"a","x":0,"y":0,"width":539,"height":420}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/20/images/fig_1.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/20/images/fig_1.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/20/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/20/Titta Aaltonen et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"experimental-usecase","paper_id":"20","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:image_caption"},"width":539,"height":420,"image_format":"jpeg","image_sha256":"bc15b6f488357c14e4e67e6c1a9c3da40655699508782b8d7f65b558bee56eb9","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/test_atomic-layer-deposition_experimental-usecase_20_fig_3.jpg","caption":"Fig. 3 Growth per cycle as a function of the percentage of $\\mathrm{TiO_2}$ subcycles.","id":"test/atomic-layer-deposition/experimental-usecase/20/fig_3","sample_id":"atomic-layer-deposition/experimental-usecase/20/fig_3","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The chart shows the growth per cycle (Å/cycle) as a function of the percentage of TiO₂ subcycles. It compares theoretical values derived from the individual growth rates (open circles) with experimental values calculated from film thicknesses measured by X-ray Reflectivity (solid squares). Experimental growth rates are consistently lower than the prediction.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| TiO₂ subcycles (%) | Growth per cycle (Å/cycle): calculated from individual growth rates | Growth per cycle (Å/cycle): calculated from XRR thickness |\\n|---|---|---|\\n| 0 | 0.280 | 0.280 |\\n| 6 | 0.293 | 0.250 |\\n| 12 | 0.306 | 0.178 |\\n| 14 | 0.313 | 0.180 |\\n| 18 | 0.320 | 0.239 |\\n| 20 | 0.326 | 0.260 |\\n| 25 | 0.338 | 0.278 |\\n| 33 | 0.357 | 0.239 |\\n| 50 | 0.399 | 0.300 |\\n| 67 | 0.440 | 0.359 |\\n| 75 | 0.459 | 0.379 |\\n| 100 | 0.519 | 0.519 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Approx. 0.519 Å/cycle (100% TiO₂ subcycles).\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Approx. 0.28 Å/cycle (0% TiO₂ subcycles).\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Theoretical values overestimate the experimental ones (derived from XRR measurement).\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Theoretical values calculated from individual growth rates.\"}]}]","bbox":[{"panel_id":"a","x":0,"y":0,"width":514,"height":375}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/20/images/fig_3.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/20/images/fig_3.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/20/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/20/Titta Aaltonen et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"experimental-usecase","paper_id":"20","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:image_caption"},"width":514,"height":375,"image_format":"jpeg","image_sha256":"26e61de4d3234cb771152ae745529ac154387120e1198ab187b56d5b1eb461d1","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/test_atomic-layer-deposition_experimental-usecase_20_fig_6.jpg","caption":"Fig. 6 SIMS depth profiles of LLT films grown with $\\mathrm{LiO}^{\\prime}\\mathrm{Bu}$ pulse times of 2, 8, and $12\\mathrm{s}$ .","id":"test/atomic-layer-deposition/experimental-usecase/20/fig_6","sample_id":"atomic-layer-deposition/experimental-usecase/20/fig_6","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The chart displays SIMS depth profiles showing the distribution of Ti-48, Li-6 and La-138 throughout the thickness of the film in nanometers for films grown with different precursor pulse times (12, 8, 2 s).\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Species | Pulse time (s) | Depth (nm) | Intensity (counts/second) |\\n|---|---|---|---|\\n| La-138 | 12 | 2.85 | 240 |\\n| La-138 | 12 | 20.4| 369 |\\n| La-138 | 12 | 95.1| 376 |\\n| La-138 | 12 | 111.45| 779 |\\n| La-138 | 12 | 128.7 | 87 |\\n| La-138 | 8 | 2.85 | 241 | \\n| La-138 | 8 | 18.9 | 362 | \\n| La-138 | 8 | 83.7 |376 | \\n| La-138 | 8 | 96.3 | 867 | \\n| La-138 | 8 | 115.05 | 72 | \\n| La-138 | 2 | 2.85 | 241 | \\n| La-138 | 2 | 18.9 | 362 | \\n| La-138 | 2 | 71.85 | 389 | \\n| La-138 | 2 | 79.8 | 499 | \\n| La-138 | 2 | 95.85 | 71 | \\n| Li-6 | 12 | 0.89 | 3869268 | \\n| Li-6 | 12 | 5.7 | 448809 | \\n| Li-6 | 12 | 91.35 | 490593 | \\n| Li-6 | 12 | 94.95 | 851946 | \\n| Li-6 | 12 | 102.3 | 349796 | \\n| Li-6 | 12 | 108.45 | 481936 | \\n| Li-6 | 12 | 127.2 | 42045 | \\n| Li-6 | 8 | 3.15 | 481936 | \\n| Li-6 | 8 | 79.35 | 481936 | \\n| Li-6 | 8 | 84.75 | 882828 | \\n| Li-6 | 8 | 89.4 | 282508 | \\n| Li-6 | 8 | 95.4 | 481936 | \\n| Li-6 | 8 | 123.15 | 7748 | \\n| Li-6 | 2 | 0.75 | 1402514 | \\n| Li-6 | 2 | 8.25 | 228164 | \\n| Li-6 | 2 | 60.75 | 382362 | \\n| Li-6 | 2 | 68.4 | 675923 | \\n| Li-6 | 2 | 75 | 362475 | \\n| Li-6 | 2 | 123.159 | 7748 | \\n| Li-6 | 2 | 79.2 | 490593 | \\n| Li-6 | 2 | 103 | 14448 | \\n| Ti-48 | 12 | 2.1 | 7215184 |\\n| Ti-48 | 12 | 5.25 | 8933691 |\\n| Ti-48 | 12 | 104.1 | 5931928 |\\n| Ti-48 | 12 | 128.7 | 75659.94 |\\n| Ti-48 | 8 | 2.1 | 8621186 |\\n| Ti-48 | 8 | 4.95 | 9940831 |\\n| Ti-48 | 8 | 90.6 | 4876905 |\\n| Ti-48 | 8 | 123.6 | 8029 |\\n| Ti-48 | 2 | 2.1 |8621186 |\\n| Ti-48 | 2 | 4.95 | 9940831 |\\n| Ti-48 | 2 | 74.55 | 5524183 |\\n| Ti-48 | 2 | 103.95 | 14971 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"The film grown with 12 s pulse.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"No. The lithium signal curve after 2 s is distinctly lower in intensity compared to the curves at 8 s and 12 s. Since the 2 s curve is below the maximum level, it indicates that the maximum possible amount of lithium has not been incorporated yet.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"List\",\"answer\":\"Ti-48, Li-6, La-138\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"No. The signal intensities for all three isotopes drop by orders of magnitude once the depth passes the film thickness (e.g. for >120 nm for the 12s sample). This indicates the substrate is a different material.\"}]}]","bbox":[{"panel_id":"a","x":0,"y":0,"width":641,"height":413}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/20/images/fig_6.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/20/images/fig_6.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/20/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/20/Titta Aaltonen et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"experimental-usecase","paper_id":"20","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:image_caption"},"width":641,"height":413,"image_format":"jpeg","image_sha256":"e08cf451755efb2bfd1030b25fed1395e6deb95270248a1f74b0faf79c080304","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/test_atomic-layer-deposition_experimental-usecase_20_fig_7.jpg","caption":"Fig. 7 SIMS depth profiles of lithium, titanium, and lanthanum in the LLT films with one and three $\\mathrm{LiO}^{\\prime}\\mathrm{Bu}+$ water sub-cycles.","id":"test/atomic-layer-deposition/experimental-usecase/20/fig_7","sample_id":"atomic-layer-deposition/experimental-usecase/20/fig_7","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The chart displays SIMS depth profiles comparing the intensity at various depths for isotopes of titanium (Ti-48), lithium (Li-6), and lanthanum (La-138) for films grown using one lithium precursor sub-cycle and three lithium sub-cycles.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Species | Number of lithium sub-cycles | Depth (nm) | Intensity (counts/second) |\\n|---|---|---|---|\\n| La-138 | 3 Li sub-cycles | 3 | 272 |\\n| La-138 | 3 Li sub-cycles | 128 | 372 |\\n| La-138 | 3 Li sub-cycles | 134 | 469 |\\n| La-138 | 3 Li sub-cycles | 159 | 165 |\\n| La-138 | 1 Li sub-cycle | 3 | 273 |\\n| La-138 | 1 Li sub-cycle | 84 | 379 |\\n| La-138 | 1 Li sub-cycle | 97 | 925 |\\n| La-138 | 1 Li sub-cycle | 115 | 83 |\\n| Li-6 | 3 Li sub-cycles | 0.6 | 2173095 |\\n| Li-6 | 3 Li sub-cycles | 3.6 | 478382 |\\n| Li-6 | 3 Li sub-cycles | 55 | 527121 |\\n| Li-6 | 3 Li sub-cycles | 78 | 1474142 |\\n| Li-6 | 3 Li sub-cycles | 158 | 158284 |\\n| Li-6 | 1 Li sub-cycle | 0.6 | 2173095 |\\n| Li-6 | 1 Li sub-cycle | 7.4 | 357578 |\\n| Li-6 | 1 Li sub-cycle | 78 | 497312 |\\n| Li-6 | 1 Li sub-cycle | 84 | 823627 |\\n| Li-6 | 1 Li sub-cycle | 90 | 294511 |\\n| Li-6 | 1 Li sub-cycle | 95 | 478382 |\\n| Li-6 | 1 Li sub-cycle | 124 | 7237 |\\n| Ti - 48 | 3 Li sub-cycles | 2.2 | 7097797 |\\n| Ti - 48 | 3 Li sub-cycles | 4.4 | 8130408 |\\n| Ti - 48 | 3 Li sub-cycles | 115 | 5733594 |\\n| Ti - 48 | 3 Li sub-cycles | 151 | 442657 |\\n| Ti - 48 | 3 Li sub-cycles | 1597 | 330874 |\\n| Ti - 48 | 1 Li sub-cycle | 2.2 | 7097797 |\\n| Ti - 48 | 1 Li sub-cycle | 5.2 | 9495724 |\\n| Ti - 48 | 1 Li sub-cycle | 88 | 4909206 |\\n| Ti - 48 | 1 Li sub-cycle | 123 | 7379 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"The one grown with 3 lithium sub-cycles.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"No. Between 60 nm and 120 nm there is a broad surge in intensity. While the upper region (10–60 nm) overlaps with the 1-sub-cycle reference, suggesting normal growth, the massive surge in intensity near the bottom might indicate that excess lithium from the additional pulses migrated and segregated at the interface.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"The film grown with one lithium precursor pulse.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"At the interface, as indicated by the surge in intensity at approx. 80 nm.\"}]}]","bbox":[{"panel_id":"a","x":0,"y":0,"width":564,"height":458}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/20/images/fig_7.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/20/images/fig_7.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/20/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/20/Titta Aaltonen et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"experimental-usecase","paper_id":"20","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:image_caption"},"width":564,"height":458,"image_format":"jpeg","image_sha256":"baadd747ed96f2fd566753604a91666f857d9c944da07bac651816d9dd4a714c","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/test_atomic-layer-deposition_experimental-usecase_21_fig_3.jpg","caption":"Fig. 3 (a) In situ thickness measurements by SE for plasma-assisted ALD of $\\mathrm{Li}_{2}\\mathrm{CO}_{3}$ at 50, 200, 275 and $300^{\\circ}\\mathrm{C}$ and thermal ALD of $\\mathrm{Li}_{2}\\mathrm{CO}_{3}$ at $150^{\\circ}\\mathrm{C}$ . The film growth between 50 and $200^{\\circ}\\mathrm{C}$ with plasma ALD was very similar (see ESI†). (b) Growth per cycle as a function of process table temperature for both plasma-assisted (squares) and thermal (circles) ALD processes.","id":"test/atomic-layer-deposition/experimental-usecase/21/fig_3","sample_id":"atomic-layer-deposition/experimental-usecase/21/fig_3","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"},{"panel_id":"b","label":"multi-axis chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The chart plots the thickness of a material as a function of the number of cycles for different thermal and plasma treatments at various temperatures.\"},{\"panel_id\":\"b\",\"text\":\"The chart plots the GPC SE and GPC EBS values against temperature for both thermal and plasma treatments.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Number of Cycles | Thermal 150 °C | Plasma 50 °C | Plasma 200 °C | Plasma 275 °C | Plasma 300 °C |\\n|-------------------|-----------------|--------------|---------------|---------------|---------------|\\n| 0 | 0 | 0 | 0 | 0 | 0 |\\n| 100 | 8 | 10 | 15 | 18 | 20 |\\n| 200 | 15 | 18 | 25 | 28 | 30 |\\n| 300 | 22 | 25 | 32 | 36 | 38 |\\n| 400 | 28 | 32 | 38 | 42 | 45 |\\n| 500 | 35 | 38 | 44 | 48 | 50 |\\n| 600 | 42 | 45 | 50 | 54 | 58 |\"},{\"panel_id\":\"b\",\"text\":\"| Temperature (°C) | GPC SE Thermal (Å) | GPC SE Plasma (Å) | GPC EBS Thermal (Li.at.nm⁻²) | GPC EBS Plasma (Li.at.nm⁻²) |\\n|-------------------|----------------------|---------------------|--------------------------------|--------------------------------|\\n| 50 | 0.3 | 0.8 | - | - |\\n| 100 | 0.6 | 0.8 | - | - |\\n| 150 | 0.5 | 0.8 | 3 | 4 |\\n| 200 | 0.5 | 0.8 | 3 | 4 |\\n| 275 | 0.5 | 0.8 | 3 | 4 |\\n| 300 | 0.5 | 1.6 | 3 | 7 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"For plasma-assisted ALD and Tsub </=200 °C, the thickness of the films develops linearly with the number of cycles. For temperatures higher than 250 °C, the film thickness develops according to two slopes, which correlates with a transition in film chemical composition\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Yes/No\",\"answer\":\"Yes\"}]},{\"panel_id\":\"b\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"It can be noticed that the growth per cycle of the thermal ALD process strongly decreases at 50 °C, indicating the lower limit of the temperature window. The ALD process window for the plasma-assisted process is extended towards lower temperatures, as it has a similar growth per cycle at 50 °C as at higher temperatures\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The overall growth per cycle is higher for the plasma-assisted process as compared to the thermal process. This is more ofen observed for plasma processes due to the relatively high reactivity of the plasma species which could create a higher density of reactive surface sites\"}]}]","bbox":[{"panel_id":"a","x":11,"y":5,"width":582,"height":453},{"panel_id":"b","x":9,"y":461,"width":594,"height":455}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/21/images/fig_3.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/21/images/fig_3.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/21/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/21/N. Hornsveld et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"experimental-usecase","paper_id":"21","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:image_caption"},"width":603,"height":911,"image_format":"jpeg","image_sha256":"34429e72a439b3254541679fcbbc1b41dc6a7a908fb65564d8648ad2b1f926f5","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/test_atomic-layer-deposition_experimental-usecase_39_fig_2.jpg","caption":"FIG. 2. (Color online) Thickness of $\\mathrm{Al}_2\\mathrm{O}_3$ as a function of the number of cycles as measured by SE for deposition temperatures of 70 and $200^{\\circ}\\mathrm{C}$ . The data shown are for films deposited on plasma oxidized $c$ -Si substrates.","id":"test/atomic-layer-deposition/experimental-usecase/39/fig_2","sample_id":"atomic-layer-deposition/experimental-usecase/39/fig_2","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The line chart shows the relationship between the number of cycles and the SE thickness at two different temperatures, 70 °C and 200 °C.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Number of cycles | SE: Thickness (Å) 70 °C | SE: Thickness (Å) 200 °C |\\n|---|---|---|\\n| 0 | 0 | 0 |\\n| 50 | 60 | 60 |\\n| 100 | 160 | 120 |\\n| 150 | 250 | 180 |\\n| 200 | 340 | 250 |\\n| 250 | 420 | 320 |\\n| 300 | 500 | 380 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Plasma oxidized Si.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"70 and 200 °C.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"1.66 Å/cycle.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"1.27 Å/cycle.\"}]}]","bbox":[{"panel_id":"a","x":0,"y":0,"width":669,"height":536}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/39/images/fig_2.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/39/images/fig_2.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/39/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/39/Heil et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"experimental-usecase","paper_id":"39","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:image_caption"},"width":669,"height":536,"image_format":"jpeg","image_sha256":"684b9e9235d82be94f1cc202177356b26b26e9ca34da13ea915e1af5ade928b2","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/test_atomic-layer-deposition_experimental-usecase_39_fig_9.jpg","caption":"FIG. 9. (Color online) Time-resolved OES measurements at the wavelengths of 519, 656, and $777~\\mathrm{nm}$ corresponding to $\\mathrm{CO^{*}}$ $\\mathrm{H^{*}}$ and $\\mathrm{O^{*}}$ respectively. (a) Emission for an $\\mathrm{O}_2$ plasma without preceding $\\mathrm{Al(CH_3)_3}$ dosing and (b) the emission for a plasma exposure step during an ALD cycle.","id":"test/atomic-layer-deposition/experimental-usecase/39/fig_9","sample_id":"atomic-layer-deposition/experimental-usecase/39/fig_9","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"},{"panel_id":"b","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The line chart displays the intensity of OES (Optical Emission Spectroscopy) during O2 plasma exposure with preceding Al(CH3)3 dosing.\"},{\"panel_id\":\"b\",\"text\":\"The line chart shows the intensity of OES during O2 plasma exposure during an ALD cycle step.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Time (s) | Intensity (a.u.) 777 nm - O* | Intensity (a.u.) 656 nm - H* | Intensity (a.u.) 519 nm - CO* |\\n|---|---|---|---|\\n| 0 | 0 | 0 | 0 |\\n| 1 | 60 | 0 | 0 |\\n| 2 | 60 | 0 | 0 |\"},{\"panel_id\":\"b\",\"text\":\"| Time (s) | Intensity (a.u.) 777 nm - O* | Intensity (a.u.) 656 nm - H* | Intensity (a.u.) 519 nm - CO* |\\n|---|---|---|---|\\n| 0 | 0 | 0 | 0 |\\n| 0.5 | 40 | 80 | 140 |\\n| 1 | 60 | 10 | 20 |\\n| 1.5 | 60 | 0 | 10 |\\n| 2 | 60 | 0 | 10 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"List\",\"answer\":\"519 nm (CO*), 656 nm (H*), 777 nm (O*)\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"In a pure O₂ plasma strike, a stable O* emission at 777 nm appears immediately, while no significant emission is observed at 519 or 656 nm. During an ALD-cycle plasma strike, strong initial emissions at 519 and 656 nm are observed, reflecting the formation of reaction products such as CO, CO₂, and H₂O.\"}]},{\"panel_id\":\"b\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"The surface reactions occur on the time scale of 1 second.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"When the plasma is ignited after an aluminium precursor dose, transient increases in CO* and H* emission are observed, indicating the formation of gas-phase reaction products such as CO, CO₂, and H₂O. Simultaneously, the O* emission is initially suppressed, reflecting the consumption of oxygen radicals in surface oxidation and by-product formation. As the surface reactions approach saturation, the emissions associated with reaction products decay due to pumping and reduced reactant availability, while the O* emission recovers to the level seen in a pure O₂ plasma. The short decay time of these signals shows that reaction completion occurs well within one second, consistent with independent spectroscopic ellipsometry results demonstrating saturated growth for plasma exposure times of half a second or longer.\"}]}]","bbox":[{"panel_id":"a","x":3,"y":6,"width":377,"height":487},{"panel_id":"b","x":364,"y":4,"width":309,"height":489}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/39/images/fig_9.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/39/images/fig_9.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/39/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/39/Heil et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"experimental-usecase","paper_id":"39","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:image_caption"},"width":675,"height":492,"image_format":"jpeg","image_sha256":"62690cbd737002aa04e0a85eb981ccb451b8229bb28715fd6eee26c8ea809524","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/test_atomic-layer-deposition_experimental-usecase_48_fig_2.jpg","caption":"FIG. 2. (Color online) (a) Film thickness as a function of number of ALD cycles with temperature ranging between 150 and $400^{\\circ}\\mathrm{C}$ . The thickness values have been obtained from in situ SE. (b) GPC values corresponding to the studied substrate stage temperatures and the average number of Hf atoms deposited per ALD cycle as determined from RBS measurements.","id":"test/atomic-layer-deposition/experimental-usecase/48/fig_2","sample_id":"atomic-layer-deposition/experimental-usecase/48/fig_2","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"},{"panel_id":"b","label":"multi-axis chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The figure shows the thickness of a material as a function of ALD cycles at various temperatures (150–400 °C). For all temperatures, the thickness increases linearly with cycle count.\"},{\"panel_id\":\"b\",\"text\":\"The figure compares the growth per cycle (GPC) and the number of Hf atoms per cycle at different temperatures (150–400 °C). The Hf atoms per cycle show a small increase at higher temperatures; GPC stays relatively constant.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| ALD cycles | 150°C : Thinkness (nm)| 200°C: Thinkness (nm) | 250°C: Thinkness (nm) | 300°C: Thinkness (nm) | 350°C: Thinkness (nm) | 400°C: Thinkness (nm) |\\n|---|---|---|---|---|---|---|\\n| 0 | 0 | 0 | 0 | 0 | 0 | 0 |\\n| 100 | 11 | 10| 10 | 9.9 | 9.76 | 9.3 |\\n| 200 | 22 | 22 | 21 | 20.9 | 20.5 | 20.4 |\\n| 300 | 34 | 33 | 32 | 31.8 | 32 | 31.8 |\"},{\"panel_id\":\"b\",\"text\":\"| Table temperature (°C) | GPC (nm) | Hf atoms per cycle (nm²) |\\n|---|---|---|\\n| 150 | 0.113 ± 0.008 | - |\\n| 200 | 0.111 ± 0.008 | 2.71 ± 0.01 |\\n| 250 | 0.108 ± 0.008 | - |\\n| 300 | 0.106 ± 0.008 | 2.77 ± 0.01 |\\n| 350 | 0.107 ± 0.008 | - |\\n| 400 | 0.109 ± 0.008 | 2.91 ± 0.01 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"At lower temeperatures.\"}]},{\"panel_id\":\"b\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The GPC stays nearly constant (~0.11 nm/cycle) across temperature, but the Hf atoms incorporated per cycle are lower at the lower temperatures and rise as temperature increases. If the film grows by about the same thickness per cycle but contains fewer Hf atoms per cycle, the added volume must include a larger fraction of non-Hf species. This mismatch supports the inference of greater impurity incorporation (or reduced film density) at lower temperatures.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"They correlate positively.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"From about 150 to 400 °C.\"}]}]","bbox":[{"panel_id":"a","x":4,"y":7,"width":496,"height":348},{"panel_id":"b","x":491,"y":4,"width":510,"height":350}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/48/images/fig_2.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/48/images/fig_2.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/48/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/48/Sharma et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"experimental-usecase","paper_id":"48","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:image_caption"},"width":1003,"height":355,"image_format":"jpeg","image_sha256":"413aac1fdf71a3b4381ecf6a3444ff9d9d3146551ab1b73384f810cb98913cbc","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/test_atomic-layer-deposition_experimental-usecase_51_fig_10.jpg","caption":"Fig. 10 Bright IV curves of SHJ cells with $10\\mathrm{nmMoO}_x$","id":"test/atomic-layer-deposition/experimental-usecase/51/fig_10","sample_id":"atomic-layer-deposition/experimental-usecase/51/fig_10","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The chart displays the relationship between voltage and current density for three different plasma times and/or compositions.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| voltage (V) | 10.6% 2 second O2 | 10.4% 2 second O2 & Ar | 7.3% 1 second O2 |\\n|---|---|---|---|\\n|-0.2 | -30 | -30 | -30 |\\n| 0 | -30 | -30 | -30 |\\n| 0.2 | -30 | -30 | -30 |\\n| 0.4 | -25 | -25 | -5 |\\n| 0.6 | 0 | 0 | 0 |\\n| 0.8 | 15 | 15 | 5 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Yes/No\",\"answer\":\"No\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Around 0.65 V.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"10 nm.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"About 10 mA/cm^2.\"}]}]","bbox":[{"panel_id":"a","x":0,"y":0,"width":650,"height":539}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/51/images/fig_10.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/51/images/fig_10.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/51/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/51/Ziegler et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"experimental-usecase","paper_id":"51","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:image_caption"},"width":650,"height":539,"image_format":"jpeg","image_sha256":"5dba508c321c47f75c0b0c3345980c59cc8dcdbcf7420f1706ffb7dd808650f4","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/test_atomic-layer-deposition_experimental-usecase_51_fig_11.jpg","caption":"Fig. 11 Bright IV curves of SHJ cells with $20 \\mathrm{nm} \\mathrm{MoO}_x$","id":"test/atomic-layer-deposition/experimental-usecase/51/fig_11","sample_id":"atomic-layer-deposition/experimental-usecase/51/fig_11","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"IV curve of three samples, all with different plasma exposure times and/or compositions. The tested solar cells had a MnOx thickness of 20 nm.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| voltage (V) | 9.0% 2 second O2 | 10.4% 2 second O2 & Ar | 5.8% 1 second O2 |\\n|---|---|---|---|\\n|-0.2 | -30 | -30 | -30 |\\n| 0 | -30 | -30 | -30 |\\n| 0.2 | -30 | -30 | -28 |\\n| 0.4 | -20 | -24 | -2 |\\n| 0.6 | 0 | 0 | 0 |\\n| 0.8 | 18 | 6 | 8 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"20 nm.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"The current decreases in all produced solar cells when the oxide layer thickness is doubled.\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Because their lower initial oxygen content can lead to a reduced MoOx work function after ITO sputtering and annealing, increasing the hole transport barrier and worsening VOC and fill factor.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"A pronounced s-shape in the illuminated IV curves limits the fill factor and efficiency.\"}]}]","bbox":[{"panel_id":"a","x":0,"y":0,"width":653,"height":539}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/51/images/fig_11.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/51/images/fig_11.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/51/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/51/Ziegler et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"experimental-usecase","paper_id":"51","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:image_caption"},"width":653,"height":539,"image_format":"jpeg","image_sha256":"620f4aa90a2da12facc76476c9f38d6afb0facad5dcc89eb8c592291a731cf5a","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/test_atomic-layer-deposition_experimental-usecase_56_figure_2.jpg","caption":"Figure 2. (a) Amount of hydrogen and carbon impurities in $40\\mathrm{nm}\\mathrm{Li}_3\\mathrm{PO}_4$ layers deposited at different processing temperatures as determined by ERD. A clear increase of the amount of incorporated hydrogen is seen on reducing the temperature. (b) Shows the dependency of the Li-ion conductivity, as determined by impedance spectroscopy, on the processing temperature. A strong rise in conductivity is seen when reducing the processing temperature, attributed to the reduced crystallinity of the layers.","id":"test/atomic-layer-deposition/experimental-usecase/56/figure_2","sample_id":"atomic-layer-deposition/experimental-usecase/56/figure_2","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"},{"panel_id":"b","label":"line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The line chart shows the atomic percentage of hydrogen (H Content) and carbon (C Content) as a function of processing temperature. The atomic percentage of H decreases as the processing temperature increases, while the atomic percentage of C remains almost constant.\"},{\"panel_id\":\"b\",\"text\":\"The line chart illustrates the change in conductivity (Cond) with respect to processing temperature. The Li-ion conductivity decreases as the processing temperature increases.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Processing T (°C) | H content | C content |\\n| --- | --- | --- |\\n| 200 | 14 | 2 |\\n| 250 | 12 | 2.2 |\\n| 278 | 10 | 2 |\"},{\"panel_id\":\"b\",\"text\":\"| Processing T (°C) | Cond (nS/cm) |\\n|---|---|\\n| 200 | 0.14 |\\n| 226 | 0.07 |\\n| 250 | 0.05 |\\n| 280 | 0.05 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Li3PO4 films with a thickness of 40 nm.\"}]},{\"panel_id\":\"b\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"The Li-ion conductivity decreases when the processing temperature increases.\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"The li-ion conductivity is highest at 200 °C.\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"The Li-ion conductivity decreases at higher temperatures due to increased crystallinity of the layers.\"}]}]","bbox":[{"panel_id":"a","x":1,"y":3,"width":577,"height":523},{"panel_id":"b","x":581,"y":1,"width":589,"height":521}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/56/images/figure_2.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/56/images/figure_2.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/56/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/56/Put et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"experimental-usecase","paper_id":"56","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:image_caption"},"width":1172,"height":522,"image_format":"jpeg","image_sha256":"24781a05c797f3bf4cce2d54d0a6026d100ebdbc0537d0caaf937f151f99dd94","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/test_atomic-layer-deposition_experimental-usecase_65_fig_1.jpg","caption":"FIG. 1. (a) Optical emission signals during a 100 mTorr $\\mathbf{N}_2$ plasma exposure step following the precursor exposure in the cycle. A constant signal is observed for plasma feedstock species (here illustrated for $\\mathbf{N}_2^*$ emission at $337\\mathrm{nm}$ $(\\mathrm{C}^3\\prod_u\\rightarrow \\mathrm{B}^3\\prod_g))$ , and a time-dependent signal is observed for dissociated reaction products (here illustrated for $\\mathbf{CN}^*$ emission at $388\\mathrm{nm}$ $(\\mathrm{B}^2\\Sigma \\rightarrow X^2\\Sigma))$ .14 Note that the emission fluctuation during the start-up of the plasma exposure can be assigned due to plasma matching effects which indicate the large influence of the released reaction products on the plasma composition. (b) Schematic illustration of the redeposition process. During the plasma exposure step, precursor ligands are liberated from the surface and enter the gas phase. These reaction products can dissociate in the plasma (e.g., the electron impact) leading to reactive species which can redeposit on the surface. The extent of the redeposition will depend on the gas residence time $\\tau$","id":"test/atomic-layer-deposition/experimental-usecase/65/fig_1","sample_id":"atomic-layer-deposition/experimental-usecase/65/fig_1","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"},{"panel_id":"b","label":"apparatus diagram"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The line chart the optical emission signal over time, for two distinct peaks at 337 nm and 388 nm related to plasma feedstock species and reaction products.\"},{\"panel_id\":\"b\",\"text\":\"The apparatus diagram depicts the flow of feedstock gas through a plasma reactor, showing the movement of feedstock species, reaction products, and redepositing species.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Time (s) | Optical emission signal (a.u.) plasma feedstock|Optical emission signal (a.u.) reaction products|\\n|---|---|---|\\n| 1 | 0|0|\\n|1.5|3|1|\\n| 2 | 3|0.5|\\n| 3 | 3|0|\\n| 4 |3|0|\\n| 5 | 3|0|\"},{\"panel_id\":\"b\",\"text\":\"\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Due to the incorporation of other species in the plasma there is an effect of the matching of the plasma in the first second.\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Carbon.\"}]},{\"panel_id\":\"b\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The reactor is a downstream plasma reactor, the gases come in from the top and are pump away from underneath the substrate.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The gas residence time can be reduced by increasing the flow into the reactor or by pumping faster.\"}]}]","bbox":[{"panel_id":"a","x":0,"y":0,"width":603,"height":513},{"panel_id":"b","x":0,"y":513,"width":588,"height":561}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/65/images/fig_1.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/65/images/fig_1.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/65/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/65/Knoops et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"experimental-usecase","paper_id":"65","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:image_caption"},"width":608,"height":1067,"image_format":"jpeg","image_sha256":"272a7d1211301bbbbe53ec980dbb7533a0bb776ffeb3aa9a92b9b434f740adbc","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/test_atomic-layer-deposition_experimental-usecase_65_fig_3.jpg","caption":"FIG. 3. (a) Growth per cycle (GPC) as a function of $\\mathbb{N}_2$ plasma exposure time for $\\mathrm{SiN}_x$ films deposited at $200^{\\circ}\\mathrm{C}$ and three different plasma gas residence times $\\tau = 0.44\\mathrm{s}$ for $100\\mathrm{sccm}$ $\\mathbf{N}_2$ at $40\\mathrm{mTorr}$ $\\tau = 0.15\\mathrm{s}$ for $100\\mathrm{sccm}$ $\\mathrm{N}_2 + 200$ sccm Ar at $40\\mathrm{mTorr}$ and $\\tau = 0.09\\mathrm{s}$ for $100\\mathrm{sccm}$ $\\mathrm{N}_2 + 200$ sccm Ar at $24\\mathrm{mTorr}$ . The lines are fits to the expression given in (a). (b) The values of the time constants $t_A$ and $t_B$ extracted from the fits in (a) as a function of residence time $\\tau$ . The solid lines are linear fits to the data points.","id":"test/atomic-layer-deposition/experimental-usecase/65/fig_3","sample_id":"atomic-layer-deposition/experimental-usecase/65/fig_3","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"},{"panel_id":"b","label":"multiple scatter plot"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The chart shows the growth per cycle (Å) over plasma exposure time (s) for different plasma gas residence times. The growth rate initially increases rapidly, reaches a peak and then gradually decreases for all conditions.\"},{\"panel_id\":\"b\",\"text\":\"The chart displays the relationship between plasma gas residence time (s) and time constant (s). There is a linear increase in time constant B with increasing plasma gas residence time, whilst time constan A remains constant.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Plasma exposure time (s) | Growth per cycle (Å) 0.44 s|Growth per cycle (Å) 0.15 s|Growth per cycle (Å) 0.09 s|\\n|---|---|---|---|\\n| 0 | 0.0 |0.0|0.0|\\n| 2 | 0.46 |0.31|0.27|\\n| 4 | 0.44 |0.29|0.26|\\n| 6 | 0.4 |0.26|0.25|\\n| 8 | 0.35 |0.25|0.25|\\n| 12 | 0.3 |0.24|0.24|\\n| 16 | 0.27 |0.24|0.24|\"},{\"panel_id\":\"b\",\"text\":\"| Plasma gas residence time (s) | Time constant tB (s) | Time constant tA (s) |\\n|---|---|---|\\n| 0.0 | 0.0 |0.8|\\n| 0.1 | 2 ± 0.7 |0.8|\\n| 0.15 | 2.5 ± 0.3 |0.8|\\n| 0.45 | 8 ± 2 |0.8|\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Usually a saturation curve is fitted with a single exponential like: GPC = A(1-e^(-t/tA)).\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The height of the peak is related to the amount of redeposition of ligand species that come in to the plasma.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Longer plasma exposure times give lower GPC values indicating that the species resulting of redeposition are removed from the surface.\"}]},{\"panel_id\":\"b\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Only time constant B.\"}]}]","bbox":[{"panel_id":"a","x":1,"y":1,"width":599,"height":476},{"panel_id":"b","x":0,"y":483,"width":601,"height":491}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/65/images/fig_3.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/65/images/fig_3.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/65/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/experimental-usecase/65/Knoops et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"experimental-usecase","paper_id":"65","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:image_caption"},"width":603,"height":972,"image_format":"jpeg","image_sha256":"6d134e9326358abe7531cacb282b56b25f115cf8d1d367a9fa7f2e1f795c0a69","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/test_atomic-layer-deposition_simulation-usecase_1_figure_5.jpg","caption":"Figure 5. Arrhenius plot of $\\Gamma (T)$ , $k^{\\mathrm{TST}}(T)$ , and $k^{\\mathrm{SC - TST}}(T)$ for the $(\\mathrm{SiH}_3\\mathrm{O})_3\\mathrm{Si - OH}^* + \\mathrm{Ti}_4$ reaction.","id":"test/atomic-layer-deposition/simulation-usecase/1/figure_5","sample_id":"atomic-layer-deposition/simulation-usecase/1/figure_5","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"Multi-line Arrhenius plot for the (SiH₃O)₃Si–OH* + TiI₄ reaction, showing temperature dependence of rate constants for the tunneling correction factor (¡), classical transition-state theory (kTST), and semiclassical tunneling-corrected rate constant (kSC-TST). At low temperatures, quantum tunneling dominates, with kSC-TST higher than kTST, while at high temperatures classical Arrhenius behavior is recovered. The crossover temperature (Tx ≈ 151 K) indicates where tunneling effects become negligible.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"|Temperature (K)|¡|kTST (s⁻¹)|kSC-TST (s⁻¹)|\\n|---------------|--|----------|-------------|\\n|60.00|6.77|5.57 × 10⁻³⁶|3.77 × 10⁻⁶|\\n|90.00|1.99|1.55 × 10⁻¹⁹|3.10 × 10⁻⁶|\\n|120.00|2.58|1.86 × 10⁻¹¹|4.79 × 10⁻⁶|\\n|150.00|3.54|1.92 × 10⁻⁶|6.78 × 10⁻⁵|\\n|180.00|4.93|4.37 × 10⁻³|2.15 × 10⁻²|\\n|210.00|2.78|1.12 × 10⁰|3.11 × 10⁰|\\n|250.00|1.95|2.06 × 10²|4.01 × 10²|\\n|300.00|1.55|1.87 × 10⁴|2.91 × 10⁴|\\n|350.00|1.37|4.81 × 10⁵|6.59 × 10⁵|\\n|423.15|1.23|1.63 × 10⁷|2.01 × 10⁷|\\n|473.15|1.18|8.18 × 10⁷|9.67 × 10⁷|\\n|523.15|1.15|3.69 × 10⁸|4.23 × 10⁸|\\n|573.15|1.12|1.08 × 10⁹|1.21 × 10⁹|\\n|623.15|1.10|2.92 × 10⁹|3.22 × 10⁹|\\n|673.15|1.08|6.85 × 10⁹|7.42 × 10⁹|\\n|723.15|1.07|1.43 × 10¹⁰|1.54 × 10¹⁰|\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Yes/No\",\"answer\":\"Yes\"}]},{\"panel_id\":\"b\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"At high temperatures, kSC-TST closely matches kTST, indicating tunneling effects are negligible in this regime.\"}]},{\"panel_id\":\"d\",\"items\":[{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The Arrhenius plot shows a clear crossover temperature (Tx ≈ 151 K) below which tunneling significantly increases the reaction rate. By analyzing this plot, one can determine the temperature range where classical TST underestimates the rate and tunneling must be considered. For practical ALD applications, this allows process engineers to predict whether tunneling effects can be ignored at typical operating temperatures, ensuring accurate rate estimates and proper design of deposition cycles.\"}]},{\"panel_id\":\"c\",\"items\":[{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"List\",\"answer\":\"¡ (tunneling correction factor) – Indicates the contribution of quantum tunneling to the rate constant, kTST – Classical rate constant from transition-state theory without tunneling correction, kSC-TST – Semiclassical rate constant including tunneling effects\"}]}]","bbox":[{"panel_id":"a","x":0,"y":0,"width":641,"height":586}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/1/images/figure_5.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/1/images/figure_5.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/1/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/1/Zheng Hu et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"simulation-usecase","paper_id":"1","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:img_caption"},"width":641,"height":586,"image_format":"jpeg","image_sha256":"1902163e1d59805790c59ca01304f170aaf6b3a8270afa952d7c82f19b7e220b","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/test_atomic-layer-deposition_simulation-usecase_1_figure_8.jpg","caption":"Figure 8. Arrhenius plot of the rate constants of the $\\mathrm{(SiH_3O)_3Si - OTi - }$ $\\mathrm{I}_3^* +\\mathrm{H}_2\\mathrm{O}$ reaction calculated with different treatments of the rotation of the absorbed $\\mathrm{H}_2\\mathrm{O}$","id":"test/atomic-layer-deposition/simulation-usecase/1/figure_8","sample_id":"atomic-layer-deposition/simulation-usecase/1/figure_8","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"Figure 8 presents an Arrhenius plot of the rate constants for the (SiH₃O)₃Si−O−TiI₃* + H₂O reaction, calculated under different treatments of the rotational motion of the adsorbed H₂O molecule. Multiple curves compare how ignoring rotation, treating rotation as restricted, or allowing free rotation affects the temperature dependence of the rate constant. The plot highlights that including rotational degrees of freedom lowers the rate constants and slightly modifies the Arrhenius activation energy, reflecting the impact of rotational entropy on the reaction kinetics.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"|T (K)|k_TST (s⁻¹)|k_SC-TST (s⁻¹)|\\n|-----:|-----------:|--------------:|\\n|423.15|1.43 × 10⁰|1.45 × 10⁰|\\n|473.15|1.56 × 10¹|1.57 × 10¹|\\n|523.15|1.08 × 10²|1.08 × 10²|\\n|573.15|5.82 × 10²|5.84 × 10²|\\n|623.15|2.04 × 10³|2.04 × 10³|\\n|673.15|6.92 × 10³|6.94 × 10³|\\n|723.15|1.72 × 10⁴|1.73 × 10⁴|\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Yes, including rotational motion decreases the rate constant compared to ignoring rotation.\"}]},{\"panel_id\":\"b\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"The restricted rotation model produces a slightly lower activation energy (4.8 kcal/mol) compared to ignoring rotation (5.1 kcal/mol).\"}]},{\"panel_id\":\"d\",\"items\":[{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The treatment of H₂O rotation directly affects the free energy of the adsorbed complex, which in turn changes the activation barrier and the magnitude of the rate constant. Ignoring rotation overestimates the rate constant by underestimating the free energy barrier, while including full rotational freedom can underestimate it due to overestimation of rotational entropy. This highlights the importance of accurately modeling molecular rotations when predicting surface reaction kinetics in ALD processes, as small differences can significantly impact predicted reaction rates and process design.\"}]},{\"panel_id\":\"c\",\"items\":[{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"List\",\"answer\":\"No rotation (rotation ignored, qrot = 1), Restricted rotation (partial rotational freedom), Free rotation (V₀ = 0, σ = 1)\"}]}]","bbox":[{"panel_id":"a","x":0,"y":0,"width":659,"height":606}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/1/images/figure_8.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/1/images/figure_8.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/1/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/1/Zheng Hu et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"simulation-usecase","paper_id":"1","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:img_caption"},"width":659,"height":606,"image_format":"jpeg","image_sha256":"f746f886be76658375c2e7df013e882146b66e7b774b48b13a8f17cae56ac735","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/test_atomic-layer-deposition_simulation-usecase_14_fig_3.jpg","caption":"Fig. 3. Coverage of the substrate surface during the first 10 ALD cycles. Simulations with and without densification mechanisms are compared to experimental results. The starting surface used was a chemical $\\mathrm{SiO_2}$ oxide surface prepared with DDC3 method of SMS machine (representing an experimental OH surface density of $50\\%$ ). The ALD reactor used was PulsarTM 2000.","id":"test/atomic-layer-deposition/simulation-usecase/14/fig_3","sample_id":"atomic-layer-deposition/simulation-usecase/14/fig_3","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The line chart illustrates the surface coverage % over number of ALD cycles, comparing experiment with two KMC simulations (with and without densification). Coverage increases steadily with more cycles for all cases, 'With Densification' achievies the coverage closest to experiment.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| ALD Cycle | Experimental | Without Densification | With Densification |\\n|------------|--------------|----------------------|--------------------|\\n| 0 | 0 | 0 | 0 |\\n| 1 | 39 | 39 | 39 |\\n| 2 | 44 | 40 | 49 |\\n| 3 | 53 | 41 | 62 |\\n| 4 | 63 | 42 | 69 |\\n| 5 | 71 | 43 | 77 |\\n| 6 | 74 | 43 | 81 |\\n| 7 | 78 | 44 | 86 |\\n| 8 | 84 | 45 | 89 |\\n| 9 | 94 | 47 | 92 |\\n| 10 | 100 | 47 | 94 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"List\",\"answer\":\"Experimental, With densification, Without densification\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The plot shows that including densification prevents early stalling and allows coverage to keep increasing with cycle number. Without densification, the simulated coverage quickly plateaus around ~40–50%. The “with densification” curve continues rising and closely follows the experimental trend toward near-complete coverage by 10 cycles. This indicates densification is necessary in the model to reproduce the experiment more closely.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"It generally underestimates the coverage compared to the experimental data.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"47 %\"}]}]","bbox":[{"panel_id":"a","x":0,"y":0,"width":514,"height":370}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/14/images/fig_3.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/14/images/fig_3.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/14/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/14/C. Mastail et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"simulation-usecase","paper_id":"14","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:img_caption"},"width":514,"height":370,"image_format":"jpeg","image_sha256":"39979be1ea6178c1f8183c09b64c7d7176f62c6fd5ddbc6bba95c0a50762a861","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/test_atomic-layer-deposition_simulation-usecase_18_fig_6.jpg","caption":"Fig. 6. Comparison of numerically calculated growth rate with experimental data of Chen et al. [28].","id":"test/atomic-layer-deposition/simulation-usecase/18/fig_6","sample_id":"atomic-layer-deposition/simulation-usecase/18/fig_6","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"This plot represents the reactor-scale model by comparing predicted GaN growth rates with experimental data from Chen et al. over 700–1300 K. The model correctly reproduces the shift from kinetic control at low temperatures to a diffusion-limited plateau near 900 K and the decline at very high temperatures. The close agreement demonstrates that the mechanism reliably captures GaN growth behavior across key temperature regimes. Graph has inverse temperature (1000/T).\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| 1000/T (1/K) | Growth Rate (µm/hr) | Work |\\n|-------------|---------------------|--------------------|\\n| 0.73 | 1.45 | This work |\\n| 0.75 | 1.80 | Experimental study |\\n| 0.77 | 1.90 | Experimental study |\\n| 0.80 | 2.00 | Experimental study |\\n| 0.83 | 2.10 | This work |\\n| 0.87 | 2.05 | This work |\\n| 0.92 | 2.00 | This work |\\n| 0.97 | 1.95 | This work |\\n| 1.02 | 1.90 | Experimental study |\\n| 1.10 | 1.70 | Experimental study |\\n| 1.27 | 0.50 | Experimental study |\\n| 1.30 | 0.25 | This work |\\n| 1.32 | 0.05 | Experimental study |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"List\",\"answer\":\"1. Desorption/Thermodynamic limitation (High T).\\n\\n, 2. Mass Transport (Diffusion) limitation (Mid T).\\n\\n, 3. Surface Kinetic limitation (Low T).\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The comprehensive multi-regime model is computationally expensive and impractical for real-time control. The risk of using the simple Arrhenius model is severe error if the process operates outside its fitted kinetic zone. It would fail to predict the growth rate plateau (leading to overestimation) and the high-T decline (leading to unsafe temperature choices), potentially causing poor film quality and reactor problems.\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Low-T (High 1000/T): Kinetic control – Growth rate increases sharply with temperature; limited by the thermal activation of surface chemical reactions.\\n\\nMid-T (Plateau): Diffusion control – Rate is constant; limited by the transport of reactants through the boundary layer to the surface.\\n\\nHigh-T (Low 1000/T): Declining regime – Rate decreases; limited by premature gas-phase reactions or depletion of reactants before they reach the surface, or thermodynamic reversal.\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Yes.The model's proven accuracy across regimes allows it to identify the diffusion-limited plateau (near 900 K) as the optimal window. In this regime, the growth rate is high and, most importantly, insensitive to minor temperature fluctuations, which is the key to achieving excellent thickness uniformity across a production wafer. The model reliably defines this window, avoiding the inefficient kinetic and problematic high-T regimes.\"}]}]","bbox":[{"panel_id":"a","x":1,"y":3,"width":588,"height":417}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/18/images/fig_6.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/18/images/fig_6.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/18/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/18/Debasis Sengupta et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"simulation-usecase","paper_id":"18","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:img_caption"},"width":592,"height":419,"image_format":"jpeg","image_sha256":"1e0847ed6df2eb28468f284ae98abcf3699369af34de3a2ad14bd17749f4932a","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/test_atomic-layer-deposition_simulation-usecase_20_fig_4.jpg","caption":"Fig. 4. Comparison of relative intermixing ratios for position B (water side) and C (TMA side) with different gap sizes, and the relative intermixing ratios is defined as the concentration ratio of the foreign gas to the home gas.","id":"test/atomic-layer-deposition/simulation-usecase/20/fig_4","sample_id":"atomic-layer-deposition/simulation-usecase/20/fig_4","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The line chart shows the relationship between gap size and relative intermixing ratio at two positions, B and C\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Gap size (mm) | Position B (Ratio) | Position C (Ratio) |\\n|---|---|---|\\n| 1.0 | 0.005 | 0.250 |\\n| 1.5 | 0.001 | 0.060 |\\n| 2.0 | 0.000 | 0.030 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Position C (TMA side) experiences higher intermixing because water has a higher inlet vapor pressure than TMA, resulting in stronger diffusive transport of water molecules toward the TMA zone. At 1 mm gap, the relative intermixing ratio at position C is approximately 0.25 (1:4 water to TMA), while position B shows only about 0.01. This asymmetry reflects the different volatilities of the two precursors.\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"It is defined as the concentration ratio of the foreign gas to the home gas at each monitoring position.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"The relative intermixing ratio decreases dramatically from approximately 0.25 to 0.03, representing roughly an 8-fold reduction.\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"No. At 2 mm gap size, both positions show low intermixing ratios (below 0.03), indicating effective precursor separation with minimal risk of unwanted CVD-type reactions.\"}]}]","bbox":[{"panel_id":"a","x":4,"y":2,"width":679,"height":577}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/20/images/fig_4.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/20/images/fig_4.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/20/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/20/Dongqing Pan et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"simulation-usecase","paper_id":"20","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:img_caption"},"width":684,"height":580,"image_format":"jpeg","image_sha256":"9ef981acb8d839efc6d133aa3acf0af5a37c173a838bbb27e7ca035deff09a6b","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/test_atomic-layer-deposition_simulation-usecase_20_fig_6.jpg","caption":"Fig. 6. The effects of wafer temperature $150^{\\circ} \\mathrm{C}$ , $200^{\\circ} \\mathrm{C}$ and $250^{\\circ} \\mathrm{C}$ on the precursor intermixing with gap size and pumping pressure fixed at $1.5 \\mathrm{~mm}$ and 0.22 torr: (a) TMA and water concentration in $\\mathrm{mol} / \\mathrm{m}^3$ at position B (water side), and (b) at position C (TMA side). The simulation time is $0.04 \\mathrm{~s}$ with $1.5 \\mathrm{~mm}$ gap size. Fig. 9. Growth per cycle of bulk $\\mathrm{Al}_2\\mathrm{O}_3$ thin films with different relative pumping pressure (to the base pressure, 0.22 torr) 0, 0.05 and 0.1 torr with $1.5 \\mathrm{~mm}$ gap size and $200^{\\circ} \\mathrm{C}$ wafer temperature.","id":"test/atomic-layer-deposition/simulation-usecase/20/fig_6","sample_id":"atomic-layer-deposition/simulation-usecase/20/fig_6","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"Growth per cycle (GPC) of Al₂O₃ thin films as a function of dosing time for three relative pumping pressures (0, 0.05, and 0.1 torr above the base pressure of 0.22 torr) at 1.5 mm gap size and 200°C wafer temperature. All three curves show nearly identical behavior, with rapid growth during the unsaturated period (0 to 0.015 s) followed by saturation at approximately 1.2 Å/cycle.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Time (s) | Al₂O₃-0 torr (Å/cycle) | Al₂O₃-0.05 torr (Å/cycle) | Al₂O₃-0.1 torr (Å/cycle) |\\n|---|---|---|---|\\n| 0.000 | 0.00 | 0.00 | 0.00 |\\n| 0.002 | 0.20 | 0.18 | 0.15 |\\n| 0.005 | 0.68 | 0.62 | 0.58 |\\n| 0.008 | 0.95 | 0.90 | 0.85 |\\n| 0.010 | 1.12 | 1.08 | 1.02 |\\n| 0.012 | 1.18 | 1.16 | 1.14 |\\n| 0.015 | 1.21 | 1.21 | 1.20 |\\n| 0.020 | 1.23 | 1.23 | 1.23 |\\n| 0.030 | 1.23 | 1.23 | 1.23 |\\n| 0.040 | 1.23 | 1.23 | 1.23 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"The saturated GPC remains essentially constant at approximately 1.2 Å/cycle regardless of pumping pressure, indicating that pumping pressure has very weak influence on the chemical deposition process.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The ALD process is self-limiting, meaning deposition stops once all reactive surface sites are consumed. While different pumping pressures affect precursor intermixing levels, the dominant precursor concentrations at the wafer surface remain comparable across all conditions. Since surface reaction kinetics depend primarily on temperature and precursor availability at saturating conditions, the final GPC is determined by the surface site density rather than pumping efficiency.\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Approximately 0.015 s, after which the growth per cycle plateaus.\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"It represents the thickness of Al₂O₃ film deposited during one complete ALD cycle, measured in angstroms.\"}]}]","bbox":[{"panel_id":"a","x":3,"y":3,"width":641,"height":544}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/20/images/fig_6.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/20/images/fig_6.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/20/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/20/Dongqing Pan et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"simulation-usecase","paper_id":"20","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:img_caption"},"width":647,"height":548,"image_format":"jpeg","image_sha256":"b1fff92a2ff9f6fd0b46f152a1b31c653fde2583de21f169dbf81b863de6c68c","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/test_atomic-layer-deposition_simulation-usecase_20_fig_7.jpg","caption":"Fig. 7. Growth per cycle of bulk $\\mathrm{Al}_2\\mathrm{O}_3$ thin films with different deposition temperatures, $150^{\\circ} \\mathrm{C}$ , $200^{\\circ} \\mathrm{C}$ and $250^{\\circ} \\mathrm{C}$ with $1.5 \\mathrm{~mm}$ gap size and the ideal pumping condition (0.22 torr).","id":"test/atomic-layer-deposition/simulation-usecase/20/fig_7","sample_id":"atomic-layer-deposition/simulation-usecase/20/fig_7","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"Growth per cycle (GPC) of Al₂O₃ thin films as a function of dosing time at three deposition temperatures (150°C, 200°C, 250°C) with 1.5 mm gap size and 0.22 torr pumping pressure. Higher temperatures yield higher saturated GPC values, with 250°C reaching approximately 1.23 Å/cycle and 150°C reaching approximately 1.18 Å/cycle. All curves saturate around 0.015 s.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Time (s) | Al₂O₃-250°C (Å/cycle) | Al₂O₃-200°C (Å/cycle) | Al₂O₃-150°C (Å/cycle) |\\n|---|---|---|---|\\n| 0.000 | 0.00 | 0.00 | 0.00 |\\n| 0.002 | 0.35 | 0.30 | 0.25 |\\n| 0.005 | 0.80 | 0.75 | 0.70 |\\n| 0.008 | 1.10 | 1.05 | 1.00 |\\n| 0.010 | 1.19 | 1.16 | 1.12 |\\n| 0.012 | 1.22 | 1.19 | 1.16 |\\n| 0.015 | 1.24 | 1.21 | 1.18 |\\n| 0.020 | 1.25 | 1.22 | 1.20 |\\n| 0.030 | 1.25 | 1.22 | 1.20 |\\n| 0.040 | 1.25 | 1.22 | 1.20 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Temperature has a noticeably stronger effect on saturated GPC than pumping pressure. While all pumping pressure curves in Figure 9 converge to essentially the same value, the temperature curves show clear separation at saturation, with 250°C achieving approximately 0.05 Å/cycle higher GPC than 150°C. This reflects the temperature dependence of surface reaction kinetics described by the Arrhenius equation, where higher temperatures increase reaction rates and surface reactivity.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Higher temperature increases the reactivity of surface species and molecular kinetic energy, leading to higher collision probabilities and faster reaction rates, which ultimately produces a higher saturated film thickness per cycle.\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"No. The three curves are nearly overlapping during the early unsaturated growth period, with differences becoming apparent only as saturation is approached.\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Approximately 4 nm/s, calculated from the saturated GPC of about 1.2 Å/cycle divided by a total cycle time of 0.03 s (two half-reactions of 0.015 s each).\"}]}]","bbox":[{"panel_id":"a","x":5,"y":6,"width":552,"height":461}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/20/images/fig_7.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/20/images/fig_7.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/20/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/20/Dongqing Pan et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"simulation-usecase","paper_id":"20","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:img_caption"},"width":558,"height":467,"image_format":"jpeg","image_sha256":"21cc855505281085ba525a55c8258e79348fa8353c144d44f0bfd7748fb0dc75","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/test_atomic-layer-deposition_simulation-usecase_20_fig_8.jpg","caption":"Fig. 8. Growth per cycle of bulk $\\mathrm{Al}_2\\mathrm{O}_3$ thin films with different gap sizes 1, 1.5 and $2 \\mathrm{~mm}$ with ideal pumping condition (0.22 torr) and $200^{\\circ} \\mathrm{C}$ wafer temperature.","id":"test/atomic-layer-deposition/simulation-usecase/20/fig_8","sample_id":"atomic-layer-deposition/simulation-usecase/20/fig_8","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"Growth per cycle (GPC) of Al₂O₃ thin films as a function of dosing time for three gap sizes (1, 1.5, and 2 mm) at 200°C and 0.22 torr pumping pressure. Larger gaps reach saturation faster (2 mm saturates at approximately 0.01 s) while smaller gaps require longer dosing times (1 mm saturates at approximately 0.025 s). The 1 mm and 1.5 mm gaps achieve slightly higher saturated GPC than 2 mm.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Time (s) | Al₂O₃-1mm (Å/cycle) | Al₂O₃-1.5mm (Å/cycle) | Al₂O₃-2mm (Å/cycle) |\\n|---|---|---|---|\\n| 0.000 | 0.00 | 0.00 | 0.00 |\\n| 0.002 | 0.02 | 0.15 | 0.35 |\\n| 0.004 | 0.18 | 0.50 | 0.75 |\\n| 0.006 | 0.38 | 0.82 | 1.00 |\\n| 0.008 | 0.60 | 1.02 | 1.12 |\\n| 0.010 | 0.78 | 1.15 | 1.17 |\\n| 0.012 | 0.92 | 1.19 | 1.19 |\\n| 0.015 | 1.08 | 1.22 | 1.20 |\\n| 0.020 | 1.19 | 1.23 | 1.20 |\\n| 0.025 | 1.22 | 1.23 | 1.20 |\\n| 0.030 | 1.23 | 1.23 | 1.20 |\\n| 0.040 | 1.23 | 1.23 | 1.20 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Smaller gap sizes significantly increase the critical dosing time. The 1 mm gap requires approximately 0.025 s to reach saturation, while 1.5 mm needs about 0.012 s and 2 mm only 0.01 s. This occurs because smaller gaps hinder gas flow and reduce convective mass transport, making diffusion the dominant transport mechanism. Since diffusion is slower than convection, precursor delivery to the surface takes longer with narrower gaps.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"The narrow 1 mm gap restricts gas flow, causing mass transport to be dominated by slower diffusive processes rather than faster convective transport, which delays precursor delivery to the wafer surface.\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"The 1 mm gap achieves the highest saturated GPC (approximately 1.23 Å/cycle), slightly higher than 1.5 mm and 2 mm gaps.\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"The 1.5 mm gap offers the best compromise, achieving a saturated GPC comparable to 1 mm while requiring a much shorter critical dosing time (0.012 s versus 0.025 s), and also showing less precursor intermixing than 1 mm.\"}]}]","bbox":[{"panel_id":"a","x":5,"y":6,"width":635,"height":537}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/20/images/fig_8.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/20/images/fig_8.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/20/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/20/Dongqing Pan et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"simulation-usecase","paper_id":"20","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:img_caption"},"width":639,"height":545,"image_format":"jpeg","image_sha256":"6b1f49df4af45f4bf425c6cdcc03c4a402fb7c7ce8b0737123e03a0d68e45e79","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/test_atomic-layer-deposition_simulation-usecase_3_fig_3.jpg","caption":"FIG. 3. (Color online) $\\mathrm{Al}_2\\mathrm{O}_3$ film growth rate in $\\mathring{\\mathrm{A}}/$ cycle for: (a) 8 s purging experiments; (b) $20\\mathrm{s}$ purging experiments. Sample A is located in the inlet area and sample C in the outlet area.","id":"test/atomic-layer-deposition/simulation-usecase/3/fig_3","sample_id":"atomic-layer-deposition/simulation-usecase/3/fig_3","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"},{"panel_id":"b","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"Al₂O₃ growth per cycle (GPC) versus temperature for five wafer positions with 8 s purging time. GPC increases from 100°C to 200°C then decreases at 250°C, demonstrating the ALD temperature window. Sample A (inlet) shows highest GPC (~1.18 Å/cycle at 200°C) while sample C (outlet) shows lowest values throughout.\"},{\"panel_id\":\"b\",\"text\":\"Same measurement with 20 s purging time. The longer purge slightly increases peak GPC for most samples. Notably, the turning temperature shifts from 200°C to 150°C for samples D and E, indicating earlier surface saturation due to extended precursor contact time.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Temperature (°C) | A | B | C | D | E |\\n|---|---|---|---|---|---|\\n| 100 | 1.05 | 1.02 | 0.98 | 1.00 | 1.00 |\\n| 150 | 1.10 | 1.07 | 1.00 | 1.05 | 1.05 |\\n| 200 | 1.15 | 1.12 | 1.10 | 1.15 | 1.15 |\\n| 250 | 1.10 | 1.08 | 0.95 | 1.10 | 1.05 |\"},{\"panel_id\":\"b\",\"text\":\"| Temperature (°C) | A | B | C | D | E |\\n|---|---|---|---|---|---|\\n| 100 | 1.10 | 1.05 | 0.98 | 1.00 | 1.00 |\\n| 150 | 1.15 | 1.10 | 1.00 | 1.05 | 1.05 |\\n| 200 | 1.20 | 1.15 | 1.10 | 1.15 | 1.15 |\\n| 250 | 1.10 | 1.08 | 0.95 | 1.10 | 1.05 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Sample A (inlet position) consistently shows the highest GPC across all temperatures. This occurs because precursor concentration remains higher for longer at the inlet, increasing collision probability with reactive surface sites. At the outlet (sample C), precursors are diluted by byproduct CH₄ and partially consumed by reactions upstream.\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"The five wafer sample positions inside the chamber: A at inlet, C at outlet, B and D at upper and lower positions, and E at center.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Enhanced desorption of formed surface species at higher temperatures reduces net film deposition.\"}]},{\"panel_id\":\"b\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"(1) Peak GPC slightly increases for most samples due to longer precursor contact time, (2) the turning temperature shifts to 150°C for samples D and E as surface saturation is reached earlier, and (3) outlet sample C remains at similar GPC since it was already saturated in 8 s purging.\"}]}]","bbox":[{"panel_id":"a","x":0,"y":0,"width":601,"height":505},{"panel_id":"b","x":0,"y":507,"width":601,"height":502}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/3/images/fig_3.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/3/images/fig_3.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/3/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/3/Dongqing Pan et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"simulation-usecase","paper_id":"3","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:img_caption"},"width":606,"height":1017,"image_format":"jpeg","image_sha256":"38ca78a452d18a5e838d3d76457cf82cc58c5d86139b599c73f11aa8a59ebc4b","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/test_atomic-layer-deposition_simulation-usecase_3_fig_4.jpg","caption":"FIG. 4. (Color online) Comparisons of surface-averaged $\\mathrm{Al}_2\\mathrm{O}_3$ film growth rates in $\\mathring{\\mathrm{A}}/$ cycle by experimental and numerical investigations.","id":"test/atomic-layer-deposition/simulation-usecase/3/fig_4","sample_id":"atomic-layer-deposition/simulation-usecase/3/fig_4","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The line chart illustrates the growth per cycle at different temperatures for two experimental conditions and one numerical simulation\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Temperature (°C) | Experimental-8s purging (Å/Cycle) | Experimental-20s purging (Å/Cycle) | Numerical simulation (Å/Cycle) |\\n|---|---|---|---|\\n| 100 | 1.01 | 1.04 | 1.09 |\\n| 150 | 1.06 | 1.14 | 1.19 |\\n| 200 | 1.14 | 1.16 | 1.23 |\\n| 250 | 1.10 | 1.07 | 1.21 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The numerical GPC closely follows the experimental trends, with values approximately 0.05 Å/cycle higher on average. This slight overestimation is attributed to model simplifications such as assuming 100% hydroxyl coverage at the initial state and neglecting intermediate reaction products. The overall agreement validates the numerical model for studying ALD process details.\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Black squares: experimental data with 8 s purging; red circles: experimental data with 20 s purging; green triangles: numerical simulation results.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"200°C, with approximately 1.22 Å/cycle.\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"(1) The model assumes perfect 100% hydroxyl coverage on wafer surfaces initially, (2) complex intermediate products from DFT studies are neglected for computational feasibility, and (3) local variations in precursor concentration and surface conditions are simplified.\"}]}]","bbox":[{"panel_id":"a","x":1,"y":0,"width":595,"height":511}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/3/images/fig_4.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/3/images/fig_4.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/3/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/3/Dongqing Pan et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"simulation-usecase","paper_id":"3","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:img_caption"},"width":598,"height":511,"image_format":"jpeg","image_sha256":"e9ddf95008041c5e8588e553ff448dda549f5bc6d398a3c0e820428daebd43a1","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/test_atomic-layer-deposition_simulation-usecase_3_fig_8.jpg","caption":"FIG. 8. (Color online) Surface deposition rate of bulk $\\mathrm{Al}_2\\mathrm{O}_3\\langle \\mathrm{s}\\rangle$ at five different positions in TMA pulsing step for 100, 150, 200, and $250^{\\circ}\\mathrm{C}$ , respectively. The bulk $\\mathrm{Al}_2\\mathrm{O}_3$ growth rate is obtained from summation of the resulted bulk species $\\mathrm{O}\\langle \\mathrm{s}\\rangle$ and $\\mathrm{Al}\\langle \\mathrm{s}\\rangle$ in the two half reactions described in Eq. (17). Point A is located in the inlet area and point C in the outlet area.","id":"test/atomic-layer-deposition/simulation-usecase/3/fig_8","sample_id":"atomic-layer-deposition/simulation-usecase/3/fig_8","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"},{"panel_id":"b","label":"multiple line chart"},{"panel_id":"c","label":"multiple line chart"},{"panel_id":"d","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"Bulk Al₂O₃ deposition rate at five chamber positions during TMA pulse at 100°C. Inlet sample A shows the fastest and highest response (~1.6×10⁻⁴ kg/m²s), peaking around 3 ms. Other positions (B, C, D, E) show delayed and lower peak rates.\"},{\"panel_id\":\"b\",\"text\":\"Same at 150°C. Sample A responds similarly; other samples show slightly earlier peaks compared to 100°C.\"},{\"panel_id\":\"c\",\"text\":\"Same at 200°C. Peak rates and timing patterns remain consistent with lower temperatures.\"},{\"panel_id\":\"d\",\"text\":\"Same at 250°C. Overall behavior is similar, confirming that positional differences in deposition rate are determined primarily by precursor transport rather than temperature.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Flow Time (ms) | A ( kg/m^2s ) | B ( kg/m^2s ) | C ( kg/m^2s ) | D ( kg/m^2s ) | E ( kg/m^2s ) |\\n|---|---|---|---|---|---|\\n| 1.5 | 0.80e-4 | 0.00 | 0.00 | 0.00 | 0.10e-4 |\\n| 2.0 | 1.65e-4 | 0.00 | 0.00 | 0.00 | 0.40e-4 |\\n| 2.5 | 1.20e-4 | 0.00 | 0.00 | 0.00 | 0.70e-4 |\\n| 3.0 | 0.40e-4 | 0.00 | 0.00 | 0.00 | 0.95e-4 |\\n| 3.5 | 0.10e-4 | 0.00 | 0.00 | 0.00 | 1.00e-4 |\\n| 4.0 | 0.00 | 0.00 | 0.05e-4 | 0.00 | 0.80e-4 |\\n| 5.0 | 0.00 | 0.05e-4 | 0.30e-4 | 0.05e-4 | 0.50e-4 |\\n| 6.0 | 0.00 | 0.20e-4 | 0.60e-4 | 0.20e-4 | 0.20e-4 |\\n| 7.0 | 0.00 | 0.50e-4 | 0.80e-4 | 0.50e-4 | 0.05e-4 |\\n| 8.0 | 0.00 | 0.75e-4 | 0.70e-4 | 0.75e-4 | 0.00 |\\n| 9.0 | 0.00 | 0.75e-4 | 0.40e-4 | 0.75e-4 | 0.00 |\\n| 10.0 | 0.00 | 0.50e-4 | 0.15e-4 | 0.50e-4 | 0.00 |\\n| 12.0 | 0.00 | 0.10e-4 | 0.00 | 0.10e-4 | 0.00 |\"},{\"panel_id\":\"b\",\"text\":\"| Flow Time (ms) | A ( kg/m^2s ) | B ( kg/m^2s ) | C ( kg/m^2s ) | D ( kg/m^2s ) | E ( kg/m^2s ) |\\n|---|---|---|---|---|---|\\n| 2.0 | 1.68e-4 | 0.00 | 0.00 | 0.00 | 0.60e-4 |\\n| 3.0 | 0.10e-4 | 0.00 | 0.00 | 0.00 | 1.05e-4 |\\n| 4.0 | 0.00 | 0.00 | 0.05e-4 | 0.00 | 0.50e-4 |\\n| 5.0 | 0.00 | 0.05e-4 | 0.35e-4 | 0.05e-4 | 0.10e-4 |\\n| 6.0 | 0.00 | 0.30e-4 | 0.70e-4 | 0.30e-4 | 0.00 |\\n| 7.0 | 0.00 | 0.60e-4 | 0.90e-4 | 0.60e-4 | 0.00 |\\n| 8.0 | 0.00 | 0.85e-4 | 0.60e-4 | 0.85e-4 | 0.00 |\\n| 9.0 | 0.00 | 0.70e-4 | 0.20e-4 | 0.70e-4 | 0.00 |\\n| 10.0 | 0.00 | 0.40e-4 | 0.00 | 0.40e-4 | 0.00 |\"},{\"panel_id\":\"c\",\"text\":\"| Flow Time (ms) | A ( kg/m^2s ) | B ( kg/m^2s ) | C ( kg/m^2s ) | D ( kg/m^2s ) | E ( kg/m^2s ) |\\n|---|---|---|---|---|---|\\n| 2.0 | 1.70e-4 | 0.00 | 0.00 | 0.00 | 0.80e-4 |\\n| 3.0 | 0.05e-4 | 0.00 | 0.00 | 0.00 | 1.25e-4 |\\n| 4.0 | 0.00 | 0.00 | 0.05e-4 | 0.00 | 0.40e-4 |\\n| 5.0 | 0.00 | 0.05e-4 | 0.35e-4 | 0.05e-4 | 0.05e-4 |\\n| 6.0 | 0.00 | 0.35e-4 | 0.75e-4 | 0.35e-4 | 0.00 |\\n| 7.0 | 0.00 | 0.70e-4 | 0.90e-4 | 0.70e-4 | 0.00 |\\n| 8.0 | 0.00 | 0.90e-4 | 0.50e-4 | 0.90e-4 | 0.00 |\\n| 9.0 | 0.00 | 0.60e-4 | 0.10e-4 | 0.60e-4 | 0.00 |\\n| 10.0 | 0.00 | 0.30e-4 | 0.00 | 0.30e-4 | 0.00 |\"},{\"panel_id\":\"d\",\"text\":\"| Flow Time (ms) | A ( kg/m^2s ) | B ( kg/m^2s ) | C ( kg/m^2s ) | D ( kg/m^2s ) | E ( kg/m^2s ) |\\n|---|---|---|---|---|---|\\n| 2.0 | 1.72e-4 | 0.00 | 0.00 | 0.00 | 0.90e-4 |\\n| 2.5 | 0.40e-4 | 0.00 | 0.00 | 0.00 | 1.25e-4 |\\n| 3.0 | 0.05e-4 | 0.00 | 0.00 | 0.00 | 0.90e-4 |\\n| 4.0 | 0.00 | 0.00 | 0.05e-4 | 0.00 | 0.20e-4 |\\n| 5.0 | 0.00 | 0.10e-4 | 0.30e-4 | 0.10e-4 | 0.00 |\\n| 6.0 | 0.00 | 0.40e-4 | 0.80e-4 | 0.40e-4 | 0.00 |\\n| 7.0 | 0.00 | 0.80e-4 | 0.85e-4 | 0.80e-4 | 0.00 |\\n| 8.0 | 0.00 | 0.92e-4 | 0.40e-4 | 0.92e-4 | 0.00 |\\n| 9.0 | 0.00 | 0.50e-4 | 0.10e-4 | 0.50e-4 | 0.00 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Sample A (inlet) consistently shows the highest and fastest deposition rate across all temperatures. This correlates with flow simulation results showing that A experiences higher precursor concentrations for longer periods. The sample closest to precursor inlets reacts first and most intensely before precursors diffuse to other regions.\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Approximately 1.6–1.8×10⁻⁴ kg/m²s across all temperatures.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"By summing the deposition rates of bulk species O⟨s⟩ and Al⟨s⟩ from the two half reactions (TMA and water pulses).\"}]},{\"panel_id\":\"d\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"No. The relative ordering (A highest, then E, then B/D, then C lowest) remains consistent across all four temperatures, indicating that flow and precursor distribution dominate over temperature effects for spatial variations.\"}]}]","bbox":[{"panel_id":"a","x":0,"y":0,"width":660,"height":522},{"panel_id":"b","x":664,"y":3,"width":647,"height":518},{"panel_id":"c","x":3,"y":513,"width":684,"height":520},{"panel_id":"d","x":646,"y":509,"width":665,"height":531}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/3/images/fig_8.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/3/images/fig_8.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/3/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/3/Dongqing Pan et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"simulation-usecase","paper_id":"3","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:img_caption"},"width":1317,"height":1042,"image_format":"jpeg","image_sha256":"c8fea5536d529924ce2b20d7848fe7ecb09bd2c5695dc462b24cc3d0f7523cc9","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/test_atomic-layer-deposition_simulation-usecase_41_fig_6.jpg","caption":"Fig. 6 Comparison of the dissociation of $\\mathrm{Cu(acac)}_2$ into $\\mathrm{Cu(acac)}$ and acac on $\\mathrm{Cu(110)}$ , $\\mathrm{Cu}_2\\mathrm{O(111)}$ , and $\\mathrm{CuO(111)}$ surfaces at $600\\mathrm{K}$ .","id":"test/atomic-layer-deposition/simulation-usecase/41/fig_6","sample_id":"atomic-layer-deposition/simulation-usecase/41/fig_6","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The line chart shows the decrease in the number of Cu(acac)_2 over time for three different surfaces: Cu(110), Cu_2O(111), and CuO(111).\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Time (ps) | Cu(110) | Cu₂O(111) | CuO(111) |\\n|---|---|---|---|\\n| 0.0 | 24 | 24 | 24 |\\n| 5.0 | 24 | 24 | 24 |\\n| 10.0 | 24 | 22 | 24 |\\n| 12.5 | 21 | 20 | 23 |\\n| 15.0 | 17 | 18 | 22 |\\n| 17.5 | 17 | 18 | 22 |\\n| 20.0 | 16 | 16 | 21 |\\n| 22.5 | 13 | 12 | 19 |\\n| 25.0 | 9 | 12 | 19 |\\n| 27.5 | 5 | 12 | 18 |\\n| 30.0 | 5 | 9 | 17 |\\n| 32.5 | 4 | 9 | 16 |\\n| 35.0 | 4 | 4 | 13 |\\n| 37.5 | 2 | 3 | 11 |\\n| 40.0 | 2 | 3 | 7 |\\n| 42.5 | 2 | 2 | 4 |\\n| 45.0 | 1 | 2 | 3 |\\n| 47.5 | 1 | 1 | 3 |\\n| 50.0 | 0 | 0 | 2 |\\n| 52.5 | 0 | 0 | 2 |\\n| 55.0 | 0 | 0 | 1 |\\n| 60.0 | 0 | 0 | 1 |\\n| 70.0 | 0 | 0 | 1 |\\n| 72.5 | 0 | 0 | 0 |\\n| 80.0 | 0 | 0 | 0 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"List\",\"answer\":\"Black squares represent Cu(110).\\n\\n, Red circles represent Cu2O(111).\\n\\n, Blue triangles represent CuO(111).\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"CuO(111) is slowest, because its curve drops later and reaches zero at a later time than the other two.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"About 70 ps.\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The paper concludes that copper rich surfaces react faster than CuO(111) in decomposing Cu(acac)2. This trend suggests that surface Cu atoms are the more reactive sites toward the acac ligand. In an ALD context, this implies that the substrate oxidation state can strongly influence precursor dissociation kinetics\"}]}]","bbox":[{"panel_id":"a","x":10,"y":9,"width":488,"height":407}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/41/images/fig_6.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/41/images/fig_6.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/41/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/41/Xiao Hu et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"simulation-usecase","paper_id":"41","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:img_caption"},"width":511,"height":419,"image_format":"jpeg","image_sha256":"31b4ddbbfa32321d6ed3f78dfd6c7657e485b2f50ac942377c52a8c127d7438c","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/test_atomic-layer-deposition_simulation-usecase_41_fig_8.jpg","caption":"Fig. 8 Product evolution for the surface reactions between $\\mathrm{Cu(acac)}_2$ and atomic H (a and b), water (c and d), and ozone (e and f), respectively.","id":"test/atomic-layer-deposition/simulation-usecase/41/fig_8","sample_id":"atomic-layer-deposition/simulation-usecase/41/fig_8","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"},{"panel_id":"b","label":"multiple line chart"},{"panel_id":"c","label":"multiple line chart"},{"panel_id":"d","label":"multiple line chart"},{"panel_id":"e","label":"multiple line chart"},{"panel_id":"f","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"Gas phase product evolution versus time for the surface reaction between Cu(acac)2 and atomic H.\"},{\"panel_id\":\"b\",\"text\":\"Surface elemental inventory (C, H, O) versus time during the surface reaction between Cu(acac)2 and atomic H.\"},{\"panel_id\":\"c\",\"text\":\"Gas phase product evolution versus time for the surface reaction between Cu(acac)2 and H2O.\"},{\"panel_id\":\"d\",\"text\":\"Surface elemental inventory (C, H, O) versus time during the surface reaction between Cu(acac)2 and H2O.\"},{\"panel_id\":\"e\",\"text\":\"Gas phase product evolution versus time for the surface reaction between Cu(acac)2 and ozone, showing COx and HxOy as major byproducts.\"},{\"panel_id\":\"f\",\"text\":\"Surface elemental inventory (C, H, O) versus time during the surface reaction between Cu(acac)2 and ozone, showing strong surface oxidation and carbon loss.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Time (ps) | H (×0.05) | CxHy | H₂O | H₂ (×0.05) | Others |\\n|---|---|---|---|---|---|\\n| 0 | 100 | 0 | 0 | 0 | 0 |\\n| 20 | 80 | 10 | 35 | 10 | 15 |\\n| 40 | 45 | 40 | 60 | 22 | 10 |\\n| 60 | 20 | 65 | 68 | 32 | 5 |\\n| 80 | 5 | 85 | 68 | 37 | 2 |\\n| 100 | 0 | 90 | 65 | 39 | 1 |\\n| 150 | 0 | 91 | 55 | 39 | 0 |\\n| 200 | 0 | 93 | 50 | 39 | 0 |\"},{\"panel_id\":\"b\",\"text\":\"| Time (ps) | C (Number) | H (Number) | O (Number) |\\n|---|---|---|---|\\n| 0 | 240 | 350 | 100 |\\n| 25 | 180 | 300 | 60 |\\n| 50 | 40 | 100 | 30 |\\n| 75 | 10 | 50 | 20 |\\n| 100 | 10 | 50 | 35 |\\n| 150 | 10 | 70 | 40 |\\n| 200 | 10 | 70 | 45 |\"},{\"panel_id\":\"c\",\"text\":\"| Time (ps) | H₂O (×0.01) | H(acac) | acac (approx) | H₂(acac) |\\n|---|---|---|---|---|\\n| 0 | 20.0 | 0 | 0 | 0 |\\n| 100 | 19.8 | 6 | 1 | 0 |\\n| 200 | 19.5 | 8 | 1 | 0 |\\n| 400 | 19.0 | 14 | 2 | 0 |\\n| 600 | 18.8 | 14 | 2 | 2 |\\n| 800 | 18.5 | 15 | 2 | 2 |\"},{\"panel_id\":\"d\",\"text\":\"| Time (ps) | H (Number) | O (Number) | C (Number) |\\n|---|---|---|---|\\n| 0 | 350 | 150 | 240 |\\n| 100 | 500 | 190 | 210 |\\n| 200 | 530 | 210 | 200 |\\n| 400 | 530 | 230 | 170 |\\n| 600 | 530 | 240 | 150 |\\n| 800 | 540 | 240 | 150 |\"},{\"panel_id\":\"e\",\"text\":\"| Time (ps) | O₃ (×0.1) | O | O₂ (×0.1) | CO+CO₂ | OH+H₂O+HO₂ | Others |\\n|---|---|---|---|---|---|---|\\n| 0 | 100 | 0 | 0 | 0 | 0 | 0 |\\n| 50 | 60 | 60 | 50 | 10 | 60 | 25 |\\n| 100 | 20 | 120 | 85 | 40 | 90 | 30 |\\n| 150 | 5 | 130 | 98 | 55 | 90 | 30 |\\n| 200 | 0 | 100 | 102 | 70 | 80 | 25 |\\n| 300 | 0 | 50 | 108 | 90 | 55 | 22 |\\n| 450 | 0 | 15 | 110 | 100 | 35 | 22 |\\n| 600 | 0 | 0 | 110 | 110 | 20 | 22 |\"},{\"panel_id\":\"f\",\"text\":\"| Time (ps) | O (Number) | H (Number) | C (Number) |\\n|---|---|---|---|\\n| 0 | 100 | 330 | 240 |\\n| 50 | 300 | 220 | 180 |\\n| 150 | 480 | 200 | 130 |\\n| 300 | 550 | 240 | 100 |\\n| 450 | 580 | 270 | 80 |\\n| 600 | 600 | 290 | 70 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Reactivity order is atomic H, then ozone, then H2O. \\n\\nAtomic H shows rapid formation of hydrogen containing products and fast changes in surface composition in (a,b).\\n\\nOzone shows substantial formation of HxOy and COx products in (e) and strong surface oxygen buildup in (f). \\n\\nWater shows slower product formation and more gradual surface evolution in (c,d).\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The paper reports that HxOy evolution starts about 10 ps earlier than COx evolution. This implies that ozone reacts first with hydrogen in the Cu(acac)2 derived ligands before extensive carbon oxidation occurs. That sequencing is consistent with an initial hydrogen abstraction step followed by deeper oxidation to CO and CO2.\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"They argue that if HxOy re adsorb, then protons become available on the surface. Those protons can participate in ligand exchange during the next Cu(acac)2 pulse, similar to the water based ligand exchange mechanism. In practice, this links ozone exposure to creation of reactive, protonated surface species that can help sustain self limiting cycling.\"}]},{\"panel_id\":\"b\",\"items\":[{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"“Gas phase” panels plot counts of gaseous products, while “Surface” panels plot how many C, H, and O atoms remain associated with the surface region during the simulation.\"}]}]","bbox":[{"panel_id":"a","x":1,"y":0,"width":442,"height":369},{"panel_id":"b","x":443,"y":1,"width":428,"height":365},{"panel_id":"c","x":0,"y":367,"width":450,"height":323},{"panel_id":"d","x":441,"y":360,"width":431,"height":330},{"panel_id":"e","x":1,"y":690,"width":438,"height":320},{"panel_id":"f","x":445,"y":688,"width":431,"height":321}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/41/images/fig_8.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/41/images/fig_8.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/41/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/41/Xiao Hu et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"simulation-usecase","paper_id":"41","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:img_caption"},"width":877,"height":1011,"image_format":"jpeg","image_sha256":"dcba7016ede83af2d48f420c08f97fa58216676fcf5ccb6857b8260c55d98811","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/test_atomic-layer-deposition_simulation-usecase_47_figure_4.jpg","caption":"Figure 4. ALD of $\\mathrm{SiO_2}$ and $\\mathrm{Si}_3\\mathrm{N}_4$ . Energetics for the Bound Reactants (BR), Transition State (TS), Bound Products (BP) and Unbound Products (UP) relative to the Unbound Reactants (UR) (described in Figure 3) for the reaction of $\\mathrm{SiH}_2(\\mathrm{NH}_2)_2$ (triangles, dashed lines), $\\mathrm{SiH}_2(\\mathrm{N}(\\mathrm{CH}_3)_2)_2$ (squares, solid lines) with the cluster models of an aminated nitride surface (red) and a hydroxylated oxide surface (green). $\\Delta E$ values are given in $\\mathrm{kJ} \\mathrm{mol}^{-1}$ and were determined from BP86/SV(P) DFT calculations. Activation energies $E_{\\mathrm{act}}$ are listed in the inset. Reproduced with permission.[69] Copyright 2014, American Chemical Society.","id":"test/atomic-layer-deposition/simulation-usecase/47/figure_4","sample_id":"atomic-layer-deposition/simulation-usecase/47/figure_4","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"}],"summarization":"[]","data_extraction":"[]","vqa":"[]","bbox":[{"panel_id":"a","x":0,"y":0,"width":869,"height":659}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/47/images/figure_4.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/47/images/figure_4.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/47/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-deposition/simulation-usecase/47/Simon D. Elliott et al.pdf"],"main_category":"atomic-layer-deposition","sub_category":"simulation-usecase","paper_id":"47","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:image_caption"},"width":869,"height":659,"image_format":"jpeg","image_sha256":"4cb259afded5b538f88cf71372b5417f1f3b8434c8741d0f3935c5b2d67ea247","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/test_atomic-layer-etching_experimental-usecase_1_4a058218120a769b3ad915dc1273c5d615aa9628e0975eb1b8445f403f7d782c.jpg","caption":"","id":"test/atomic-layer-etching/experimental-usecase/1/4a058218120a769b3ad915dc1273c5d615aa9628e0975eb1b8445f403f7d782c","sample_id":"atomic-layer-etching/experimental-usecase/1/4a058218120a769b3ad915dc1273c5d615aa9628e0975eb1b8445f403f7d782c","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The figure shows two overlapping software screenshots displaying line plots of electrochemical measurements. The left panel appears to show a voltage profile (likely charge-discharge cycles over time), while the right panel displays a capacity (or similar cumulative metric) plotted against cycle number or time. Different colored curves in the right plot suggest comparison across multiple channels or cell conditions (e.g., different materials or electrolytes). The consistent curvature upward indicates increasing cumulative response over repeated measurements\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| x | y1 | y2 | y3 |\\n|---|---|---|---|\\n| 0 | 0.178 | 0.178 | 0.178 |\\n| 1 | 0.256 | 0.256 | 0.256 |\\n| 2 | 0.334 | 0.334 | 0.334 |\\n| 3 | 0.412 | 0.412 | 0.412 |\\n| 4 | 0.490 | 0.490 | 0.490 |\\n| 5 | 0.568 | 0.568 | 0.568 |\\n| 6 | 0.646 | 0.646 | 0.646 |\\n| 7 | 0.724 | 0.724 | 0.724 |\\n| 8 | 0.802 | 0.802 | 0.802 |\\n| 9 | 0.880 | 0.880 | 0.880 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The curvature in the right plot suggests that the system exhibits a non-linear increase in response over time or cycles. This is commonly seen in battery or capacitor systems, where the cumulative capacity or state of charge grows more slowly as the system approaches saturation or stabilises after initial activation. The separation between the curves may indicate differences in performance across different test conditions, such as varying materials, electrolytes, or cycle numbers. This type of data helps evaluate long-term stability and efficiency.\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Three.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Yes/No\",\"answer\":\"No\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"List\",\"answer\":\"Charge - discharge efficiency, Cycle-to-cycle capacity retention\"}]}]","bbox":[{"panel_id":"a","x":0,"y":0,"width":311,"height":189}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/1/images/4a058218120a769b3ad915dc1273c5d615aa9628e0975eb1b8445f403f7d782c.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/1/images/4a058218120a769b3ad915dc1273c5d615aa9628e0975eb1b8445f403f7d782c.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/1/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/1/Atomic Layer Etching An Industry Perspective.pdf"],"main_category":"atomic-layer-etching","sub_category":"experimental-usecase","paper_id":"1","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"not_found"},"width":311,"height":189,"image_format":"jpeg","image_sha256":"594aff5e9a144e67a0c5b09b262733efe80670cff01766689e2c0966407f276c","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/test_atomic-layer-etching_experimental-usecase_12_fig_4.jpg","caption":"FIG. 4. Comparison of literature values of the etch per cycle (EPC) for thermal HF/TMA processes and the $\\mathrm{SF_6}$ plasma/TMA process outlined here.<sup>10,14,15,18</sup> Temperatures reported from this work are measured using in situ spectroscopic ellipsometry, while the literature values are the reported set point temperatures. Substrate temperature values from the literature may deviate depending on process conditions and reactors used, as discussed in more detail in the supplementary material. Additional data for thermal ALE processes utilizing alternate reactants are presented in Fig. S3 in the supplementary material.","id":"test/atomic-layer-etching/experimental-usecase/12/fig_4","sample_id":"atomic-layer-etching/experimental-usecase/12/fig_4","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"Etch per cycle increases with substrate temperature for all reported thermal ALE chemistries, with the SF₆ plasma/TMA process showing the strongest activation. This work reaches ~1.8 Å/cycle at 200 °C and increases to ~3.5 Å/cycle at 300 °C, exceeding all HF/TMA literature values. The highest reported HF/TMA result (Cano et al.) reaches ~2.5 Å/cycle at 300 °C, while other studies remain at or below ~1 Å/cycle. The comparison highlights a clear performance advantage of plasma-assisted fluorination.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Substrate Temperature (°C) | This work (SF₆/TMA) | Canon HF/TMA | Cheng HF/TMA | Lee HF/TMA | Zyvox HF/TMA |\\n|---|---|---|---|---|---|\\n| 150 | ~0.2 | ~0.0 | ~0.0 | ~0.0 | ~0.0 |\\n| 200 | ~1.2 | ~0.6 | ~0.4 | ~0.3 | ~0.2 |\\n| 250 | ~2.7 | ~1.5 | ~0.9 | ~0.7 | ~0.6 |\\n| 300 | ~3.5 | ~2.2 | ~1.3 | ~1.1 | ~1.0 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Thermal HF/TMA process and SF6 plasma/TMA process.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"The current work with SF6/TMA.\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Approximately 275 °C.\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"The etch per cycle increases as the substrate temperature gets higher.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"For increasing temperature, regardless of the chemistry, the EPC increases.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"It could be that different types of AlOx are etched. It is known that deposition method of the substrate can have an impact on etch rate.\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Zywotko has shown that there is deposition of 0.5 angstrom/cycle at 250 degrees celcius.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"at 250 degrees celcius an etch rate of 3 angstrom/cycle is achieved.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The SF₆ plasma/TMA process shows much stronger temperature activation than HF/TMA chemistries. At 250 °C it already matches or exceeds the highest HF/TMA values reported at 300 °C, and at 300 °C it reaches ~3.5 Å/cycle. This indicates that plasma-assisted fluorination produces a more reactive surface for ligand exchange with TMA.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"SF₆ plasma/TMA (this work), at ~3.5 Å/cycle.\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"SF₆ plasma produces a higher density of reactive fluorinated sites than HF gas exposure. The presence of energetic fluorine species increases fluorine coverage and accessibility at the surface, which enhances the efficiency of TMA-driven ligand exchange. This results in greater material removal per cycle at elevated temperature.\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Yes/No\",\"answer\":\"Yes\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The SF₆ plasma ALE process (green squares) exhibits a significantly higher EPC compared to all the thermal HF/TMA processes (dashed lines) across the measured temperature range. For example, at 250 °C, the plasma process achieves an EPC of nearly 3.0 Å/cycle, whereas the thermal processes range from approximately 0.1 to 2.3 Å/cycle. Additionally, the plasma process maintains viable etch rates at lower temperatures where some thermal processes are inactive or show negligible rates.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The etch rate acts as a strong function of the substrate temperature, increasing steeply as the temperature rises. The EPC grows from approximately 0.2 Å/cycle at 155 °C to 3.4 Å/cycle at 285 °C. The data point at 140 °C lies slightly below zero, indicating a transition from etching to deposition (ALD mode) at very low temperatures.\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"A key advantage is the ability to achieve significant etching at lower processing temperatures. The figure shows that the plasma process delivers an etch rate of 0.8 Å/cycle at 185 °C, a temperature where thermal kinetics typically limit the reaction rate for HF-based processes. This makes the plasma variant suitable for substrates that cannot withstand the higher thermal budgets required for efficient thermal ALE.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"The transition occurs around 140 °C, as indicated by the data point at 140 °C showing a slightly negative value (net deposition) and the point at 155 °C showing a positive etch rate.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"List\",\"answer\":\"SF₆ plasma / TMA: The chemistry utilized in \\\"This work\\\" (Green square trace).\\n\\nThermal HF / TMA: The standard chemistry cited from literature sources (Cano, Hennessy, Lee, Zywotko).\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The SF₆/TMA process exhibits a significantly higher etch rate compared to the HF/TMA processes. While the HF/TMA data (excluding Cano) generally remains below 1.5 Å/cycle, the SF₆/TMA process rapidly climbs from ~2.0 Å/cycle to >3.0 Å/cycle in this temperature window.\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"It implies the etch rate is limited by a plasma-driven surface reaction with low activation energy, not by a thermally activated step like precursor decomposition or ligand desorption.\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Modern memory chips contain temperature-sensitive materials (e.g., metals, organics) that cannot withstand the >275°C required for thermal HF/TMA etching. The SF₆ plasma/TMA process achieves the necessary etch rate at ~200°C, compatible with these thermal constraints. Furthermore, the wide temperature window (200-300°C) offers robust process control across a wafer, which is critical for the high-aspect-ratio structures in 3D memory.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Yes, depending on the reactor and process conditions used.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"In previous work, ALD of AlF3 was observed to proceed in the wide temperature range of 50–300 °C when using a short 10 ms TMA dose.This suggests\\nthat not only the temperature but also exposure to the precursor determines whether deposition or etching occurs.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"In comparison to thermal ALE, the plasma ALE process exhibits a higher etch per cycle (EPC) over the entire temperature range investigated\\nhere.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"A significant etch rate of 0.8A˚ /cycle was already observed at a relatively low temperature of 185 °C, highlighting one of the benefits of using a plasma process. It is thought that the fluorine radicals provided by the plasma allow for higher etch rates as a thicker fluorinated surface region is formed as compared to when using HF.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"3.4 Å.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"140 °C.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"The plasma ALE process shows a higher EPC at all temperatures.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"The TMA exposure dose, which resulted in AlF3 ALD for exposure times of 10 ms.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The plot shows that SF₆ plasma/TMA gives the highest EPC at a given substrate temperature compared with the HF/TMA studies. That means you can achieve the same total etch depth at a lower temperature than with HF/TMA. Based on this trend alone, SF₆ plasma/TMA is the better choice for temperature-sensitive substrates.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"At approximately 150 °C (EPC crossing 0).\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Plasma etching with SF6/TMA.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"List\",\"answer\":\"Cano (HF/TMA): 2.5 Å/cycle, Hennessy (HF/TMA): 1.2 Å/cycle, Lee (HF/TMA): 0.45 Å/cycle\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"EPC increases with substrate temperature for all reported processes. The SF₆ plasma/TMA process shows a steeper increase and higher EPC values compared to thermal HF/TMA processes reported in the literature, particularly above 200 °C\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"The SF₆ plasma/TMA process reported in this work.\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Temperatures in this work are measured in situ, while literature values are reported set points that may vary with reactor design and process conditions.\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"List\",\"answer\":\"Higher EPC, Stronger T-dependence, Better efficiency vs. thermal ALE\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Yes. The plasma-assisted ALE process consistently shows higher EPC than reported thermal HF/TMA processes for the entire temperature range studied.\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"At low substrate temperatures (~140 °C), the EPC is negative, indicating a transition from etching to deposition, which reflects ALD-like behavior.\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"List\",\"answer\":\"Achieves higher etch rates even at relatively low temperatures, Ensures isotropic etching suitable for high aspect ratio 3D trenches, Allows precise control of film removal per cycle through temperature and pulse adjustments\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Increasing the substrate temperature increases the EPC. At low temperatures (~140 °C), the EPC is slightly negative, indicating deposition-like behavior typical of ALD. As the temperature rises, the EPC increases steadily, reaching ~3.4 Å/cycle at 285 °C, showing that higher substrate temperatures enhance etching efficiency in the plasma-assisted ALE process.\"}]}]","bbox":[{"panel_id":"a","x":0,"y":0,"width":598,"height":450}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/12/images/fig_4.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/12/images/fig_4.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/12/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/12/Isotropic plasma atomic layer etching of Al2O3 using a fluorine containing plasma and Al(CH3)3.pdf"],"main_category":"atomic-layer-etching","sub_category":"experimental-usecase","paper_id":"12","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:img_caption"},"width":598,"height":450,"image_format":"jpeg","image_sha256":"8c6d933d6117c56fdd962da8a41bce7235d360674e1d8aae675b600767265eb2","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/test_atomic-layer-etching_experimental-usecase_20_figure_12.jpg","caption":"Figure 12. Ion signal intensities versus temperature during $\\mathrm{BCl}_3$ exposure on $\\mathrm{AlF}_3$ powder with a heating rate of $3^{\\circ}\\mathrm{C / min}$ . (a) Boron ligand-exchange species $(\\mathrm{BCl}_2\\mathrm{F}^+$ , $\\mathrm{BClF}_2^+$ and $\\mathrm{BF}_3^+$ ) and $\\mathrm{BCl}_3^+$ reactant. (b) $\\mathrm{AlCl}_3^+$ and $\\mathrm{AlCl}_2^+$ etch products.","id":"test/atomic-layer-etching/experimental-usecase/20/figure_12","sample_id":"atomic-layer-etching/experimental-usecase/20/figure_12","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"},{"panel_id":"b","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The line chart shows the intensity of various ions (BCl₃⁺, BCIF₂⁺, BCl₂F⁺, BF₃⁺) as a function of temperature. The intensity decreases for BCl₃⁺ while it increases for BCIF₂⁺, BCl₂F⁺ and BF₃⁺, the latter showing a particularly strong increase.\"},{\"panel_id\":\"b\",\"text\":\"The line chart displays the intensity of AlCl₂⁺ and AlCl₃⁺ as a function of temperature. Both intensities increase with temperature, with AlCl₂⁺ showing a steeper rise than AlCl₃⁺.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Temperature (°C) | BCl₃⁺ | BCIF₂⁺ | BCl₂F⁺ | BF₃⁺ |\\n|----------------|-------|--------|---------|-------|\\n| 0 | 200 | 40 | 0 | 0 |\\n| 100 | 160 | 40 | 20 | 10 |\\n| 200 | 150 | 40 | 20 | 20 |\\n| 300 | 80 | 100 | 80 | 100 |\\n| 400 | - | - | - | - |\"},{\"panel_id\":\"b\",\"text\":\"| Temperature (°C) | AlCl₂⁺ | AlCl₃⁺ |\\n|----------------|--------|--------|\\n| 0 | 0 | 0 |\\n| 100 | 0 | 0 |\\n| 200 | 0 | 0 |\\n| 300 | 200 | 80 |\\n| 400 | - | - |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"The boron species originate from ligand-exchange reactions.\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"This is a ligand-exchange reaction that leads to F/Cl transfer and the formation of Cl in the AlF3 powder.\"}]},{\"panel_id\":\"b\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"The intensity of both etch products remains constant at 0 mV until 200 °C. After this point, the peak intensity increases for both products.\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"As can be seen in the subfigure b, the ACl2+ shows a higher peak intensity.\"}]}]","bbox":[{"panel_id":"a","x":3,"y":6,"width":657,"height":244},{"panel_id":"b","x":2,"y":254,"width":656,"height":297}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/20/images/figure_12.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/20/images/figure_12.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/20/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/20/Thermal Atomic Layer Etching of Aluminum Nitride.pdf"],"main_category":"atomic-layer-etching","sub_category":"experimental-usecase","paper_id":"20","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:img_caption"},"width":661,"height":553,"image_format":"jpeg","image_sha256":"dc461b367b9358d7ebebd8e0bfa9d3461adc3500f80497c4e4f07ab68b8aa8d2","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/test_atomic-layer-etching_experimental-usecase_21_fig_4.jpg","caption":"Fig. 4. Self-limiting behavior for CoO thermal ALE for long Hacac and $\\mathrm{O_3}$ exposures at $250^{\\circ}C$ Hacac exposure time was varied with $\\mathrm{O_3}$ exposure fixed at $1.0~\\mathrm{s}$ $\\mathrm{O_3}$ exposure time was varied with Hacac exposure fixed at $1.5~\\mathrm{s}$","id":"test/atomic-layer-etching/experimental-usecase/21/fig_4","sample_id":"atomic-layer-etching/experimental-usecase/21/fig_4","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The figure shows the etch rate of CoOₓ (Å/cycle) as a function of exposure time to Hacac or O₃ (s) at 250 °C with a 40 s purge. Both curves show an increase in the etch ratio as exposure time increases.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Hacac or O₃ Exposure Times (s) | Hacac | O3 |\\n| --- | --- | --- |\\n| 0.0 | 0.0 | 0.0 |\\n| 0.5 | - | - |\\n| 1.0 | 0.36 | 0.44 |\\n| 1.5 | 0.44 | 0.46 |\\n| 2.0 | 0.46 | 0.48 |\\n| 2.5 | 0.48 | 0.52 |\\n| 3.0 | - | - |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"The experiments were performed under 40 s of purge at 250 °C.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"The CoOx etch rate increases for both cases as the exposure time increases.\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"After 2.5 s of exposure to Hacac or O3.\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Yes/No\",\"answer\":\"Yes\"}]}]","bbox":[{"panel_id":"a","x":0,"y":0,"width":660,"height":550}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/21/images/fig_4.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/21/images/fig_4.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/21/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/21/Thermal atomic layer etching of CoO using acetylacetone and ozone.pdf"],"main_category":"atomic-layer-etching","sub_category":"experimental-usecase","paper_id":"21","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:img_caption"},"width":664,"height":553,"image_format":"jpeg","image_sha256":"c4af285810dca305c19cdb1e96517046f81d53e4c3336de198787145a30533a6","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/test_atomic-layer-etching_experimental-usecase_27_fig_4.jpg","caption":"Fig. 4. (a) The atomic percentage of the etched $\\mathrm{Al}_2\\mathrm{O}_3$ surface and (b) the $\\mathrm{Al / O}$ ratio of the etched surface of $\\mathrm{Al}_2\\mathrm{O}_3$ measured by X-ray photoelectron spectroscopy (XPS) for the reference, the $\\mathrm{Al}_2\\mathrm{O}_3$ etched by partial layer etching $(\\Theta_{\\mathrm{Al - Cl / BCl - O}}< 1)$ , and the $\\mathrm{Al}_2\\mathrm{O}_3$ etched by ALET $(\\Theta_{\\mathrm{Al - Cl / BCl - O}} = 1)$ .","id":"test/atomic-layer-etching/experimental-usecase/27/fig_4","sample_id":"atomic-layer-etching/experimental-usecase/27/fig_4","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"},{"panel_id":"b","label":"bar chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The line chart illustrates the atomic percentages of Al and O of the three samples.\"},{\"panel_id\":\"b\",\"text\":\"The bar chart compares the Al/O ratio across three samples: Ref., Partial Layer Etching, and ALET.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Sample | Atomic Percentage Al | Atomic Percentage O |\\n|---|---|---|\\n| Ref. | 35 | 65 |\\n| Partial Layer Etching | 30 | 70 |\\n| ALET | 35 | 65 |\"},{\"panel_id\":\"b\",\"text\":\"| Sample | Al/O Ratio |\\n|---|---|\\n| Ref. | 0.55 |\\n| Partial Layer Etching | 0.4 |\\n| ALET | 0.55 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Approximately 1 nm.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"The ALD Al2O3 reference has an Al/O ratio lower than 0.67, whilst sapphire is exactly 0.67.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Under partial layer etching, the Al/O ratio decreases below the reference value, indicating non-stoichiometric surface modification.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Figure 4 shows that the surface stoichiometry of Al2O3 responds very differently to partial layer etching compared to ALET. Under partial layer etching conditions, the atomic percentages shift and the Al/O ratio becomes lower than that of the ALD reference film. This suggests that the process removes oxygen and aluminum unevenly, likely due to preferential etching of surface-exposed sites. In contrast, when ALET is performed at a reactant ratio of unity, the surface stoichiometry remains unchanged. Both the elemental percentages and the Al/O ratio match the reference, showing that a complete monolayer is removed in each cycle without perturbing the underlying material.\"}]}]","bbox":[{"panel_id":"a","x":6,"y":6,"width":629,"height":213},{"panel_id":"b","x":5,"y":228,"width":629,"height":223}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/27/images/fig_4.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/27/images/fig_4.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/27/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/27/Atomic layer etching of Al2O3 using BCl3Ar.pdf"],"main_category":"atomic-layer-etching","sub_category":"experimental-usecase","paper_id":"27","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:image_caption"},"width":636,"height":450,"image_format":"jpeg","image_sha256":"d108946ba1ce254c906748d60e65e888e4b96903cb029988d56819570b342691","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/test_atomic-layer-etching_experimental-usecase_35_fig_3.jpg","caption":"FIG. 3. Thickness change of $\\mathrm{TiO_2}$ and $\\mathrm{ZrO_2}$ thin films when exposed to $\\mathrm{WF_6}$ [(a) and (e)], $\\mathrm{BCl_3}$ [(b) and (f)], $\\mathrm{SOCl_2}$ [(c) and (g)], or $\\mathrm{TiCl_4}$ [(d) and (h)] doses at various temperatures as measured by in situ ellipsometry.","id":"test/atomic-layer-etching/experimental-usecase/35/fig_3","sample_id":"atomic-layer-etching/experimental-usecase/35/fig_3","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"},{"panel_id":"b","label":"multiple line chart"},{"panel_id":"c","label":"multiple line chart"},{"panel_id":"d","label":"multiple line chart"},{"panel_id":"e","label":"multiple line chart"},{"panel_id":"f","label":"multiple line chart"},{"panel_id":"g","label":"multiple line chart"},{"panel_id":"h","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"For the TiO₂ + WF₆ system, thickness change becomes increasingly negative with higher WF₆ doses and temperature, indicating progressive material removal dominated by fluorination reactions, with etching strongly enhanced above 200 °C.\"},{\"panel_id\":\"b\",\"text\":\"In the TiO₂ + BCl₃ system, thickness change remains close to zero across all BCl₃ doses and temperatures, showing that BCl₃ does not significantly etch or grow TiO₂ under these conditions and mainly results in surface-limited reactions.\"},{\"panel_id\":\"c\",\"text\":\"For TiO₂ + SOCl₂, higher SOCl₂ doses and elevated temperatures lead to gradual negative thickness changes, demonstrating temperature-activated chlorination and etching, with the strongest material removal observed at 240 °C.\"},{\"panel_id\":\"d\",\"text\":\"In the TiO₂ + TiCl₄ system, thickness changes are small and slightly positive at most temperatures, indicating limited surface modification or redeposition rather than net etching, consistent with reversible ligand exchange behavior.\"},{\"panel_id\":\"e\",\"text\":\"For the ZrO₂ + WF₆ system, thickness decreases significantly with increasing WF₆ dose, particularly at lower temperatures, confirming efficient fluorination-driven etching of ZrO₂, with stronger material loss at 250–300 °C.\"},{\"panel_id\":\"f\",\"text\":\"In the ZrO₂ + BCl₃ system, thickness changes are slightly positive or near zero depending on temperature, suggesting weak surface reactions and minimal etching, with higher temperatures favoring marginal film growth or surface stabilization.\"},{\"panel_id\":\"g\",\"text\":\"For ZrO₂ + SOCl₂, thickness variations remain small and fluctuate around zero, indicating limited and unstable chlorination reactions, with only minor etching occurring at specific temperatures and doses.\"},{\"panel_id\":\"h\",\"text\":\"In the ZrO₂ + TiCl₄ system, thickness increases steadily with TiCl₄ dose at higher temperatures, showing net film growth or surface chlorination, while lower temperatures exhibit minimal thickness change.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Chemistry|Substrate|Temperatures (°C)|Trend Summary|\\n|----------|---------|-----------------|-------------|\\n|WF6|TiO2|160, 180, 200, 220, 240|Initial swelling → etching at high T|\"},{\"panel_id\":\"b\",\"text\":\"| Chemistry|Substrate|Temperatures (°C)|Trend Summary|\\n|----------|---------|-----------------|-------------|\\n|BCl3|TiO2|200–280|Mild etch; self-limiting behavior|\"},{\"panel_id\":\"c\",\"text\":\"| Chemistry|Substrate|Temperatures (°C)|Trend Summary|\\n|----------|---------|-----------------|-------------|\\n|SOCl2|TiO2|160–240|Clean etching; strong self-limiting|\"},{\"panel_id\":\"d\",\"text\":\"| Chemistry|Substrate|Temperatures (°C)|Trend Summary|\\n|----------|---------|-----------------|-------------|\\n|TiCl4|TiO2|160–240|Weak reaction; minimal change|\"},{\"panel_id\":\"e\",\"text\":\"| Chemistry|Substrate|Temperatures (°C)|Trend Summary|\\n|----------|---------|-----------------|-------------|\\n|WF6|ZrO2|250, 275, 300, 325|Immediate shrinkage → stronger at high T|\"},{\"panel_id\":\"f\",\"text\":\"| Chemistry|Substrate|Temperatures (°C)|Trend Summary|\\n|----------|---------|-----------------|-------------|\\n|BCl3|ZrO2|250–325|Best etching at high T|\"},{\"panel_id\":\"g\",\"text\":\"| Chemistry|Substrate|Temperatures (°C)|Trend Summary|\\n|----------|---------|-----------------|-------------|\\n|SOCl2|ZrO2|250–325|Poor etching; sulfate formation|\"},{\"panel_id\":\"h\",\"text\":\"| Chemistry|Substrate|Temperatures (°C)|Trend Summary|\\n|----------|---------|-----------------|-------------|\\n|TiCl4|ZrO2|250–325|Slight etching at high T|\"}]","vqa":"[{\"panel_id\":\"d\",\"items\":[{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The curves reveal how each co-reactant affects thickness per dose and how strongly temperature influences reactivity. For TiO₂, SOCl₂ causes clean, stable etching with clear saturation, identifying it as the most effective co-reactant. For ZrO₂, BCl₃ shows the strongest negative thickness change at elevated temperatures, matching the expected formation of volatile ZrCl₄ and making it the best choice for ZrO₂ ALE. Meanwhile, SOCl₂ and TiCl₄ perform poorly on ZrO₂, confirming thermodynamic predictions.\"}]},{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Yes, several chemistries (especially WF₆ + SOCl₂ for TiO₂ and WF₆ + BCl₃ for ZrO₂) show clear saturation behavior after a few doses, confirming self-limiting reactions consistent with ALE.\"}]},{\"panel_id\":\"c\",\"items\":[{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"List\",\"answer\":\"TiO₂ + WF₆. TiO₂ + SOCl₂, ZrO₂ + WF₆, ZrO₂ + BCl₃\"}]},{\"panel_id\":\"b\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"SOCl₂ produces strong, clean etching on TiO₂ at all temperatures but yields minimal or even positive thickness change (surface swelling from sulfate formation) on ZrO₂, demonstrating poor ALE performance for ZrO₂.\"}]}]","bbox":[{"panel_id":"a","x":12,"y":20,"width":510,"height":376},{"panel_id":"b","x":8,"y":392,"width":499,"height":384},{"panel_id":"c","x":5,"y":776,"width":504,"height":384},{"panel_id":"d","x":3,"y":1159,"width":512,"height":384},{"panel_id":"e","x":530,"y":20,"width":492,"height":345},{"panel_id":"f","x":528,"y":379,"width":494,"height":368},{"panel_id":"g","x":528,"y":762,"width":494,"height":369},{"panel_id":"h","x":528,"y":1144,"width":494,"height":368}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/35/images/fig_3.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/35/images/fig_3.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/35/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/35/Comparison of BCl3, TiCl4, and SOCl2 chlorinating agents for.pdf"],"main_category":"atomic-layer-etching","sub_category":"experimental-usecase","paper_id":"35","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:image_caption"},"width":1022,"height":1548,"image_format":"jpeg","image_sha256":"f87a582668358ffb3e7c0410d84540d31c24cc02d4de1cf1bbba0ee7ffd92fc7","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/test_atomic-layer-etching_experimental-usecase_35_fig_6.jpg","caption":"FIG. 6. Influence of substrate temperature on the etch rates during atomic layer etching of (a) $\\mathrm{TiO_2}$ and (b) $\\mathrm{ZrO_2}$ thin films.","id":"test/atomic-layer-etching/experimental-usecase/35/fig_6","sample_id":"atomic-layer-etching/experimental-usecase/35/fig_6","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"},{"panel_id":"b","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"shows TiO₂ etch rate rising linearly with substrate temperature for WF₆ paired with SOCl₂, BCl₃ and TiCl₄ — with SOCl₂ giving the highest rates and TiCl₄ the lowest.\"},{\"panel_id\":\"b\",\"text\":\"shows ZrO₂ etch rate vs temperature: BCl₃ produces the largest and most temperature-sensitive etch rates, TiCl₄ is moderate, and SOCl₂ is essentially inactive. The trends confirm that etching is thermally activated and that optimal co-reactant choice differs by oxide (SOCl₂ best for TiO₂; BCl₃ best for ZrO₂).\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Substrate|Co-reactants Compared|Temperature Range (°C)|Trend Summary|\\n|----------|---------------------|----------------------|-------------|\\n|TiO₂|BCl₃, SOCl₂, TiCl₄|160–190|Etch rate increases with T; SOCl₂ highest, TiCl₄ lowest|\"},{\"panel_id\":\"b\",\"text\":\"| Substrate|Co-reactants Compared|Temperature Range (°C)|Trend Summary|\\n|----------|---------------------|----------------------|-------------|\\n|ZrO₂|BCl₃, SOCl₂, TiCl₄|250–325|BCl₃ highest and strongly T-dependent; SOCl₂ nearly flat|\"}]","vqa":"[{\"panel_id\":\"d\",\"items\":[{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The plots clearly show that SOCl₂ provides the highest etch rates for TiO₂ across the tested temperatures, identifying it as the most efficient co-reactant for TiO₂ ALE. For ZrO₂, BCl₃ exhibits by far the strongest temperature-driven increase in etch rate, making it the preferred co-reactant for high-performance etching. TiCl₄ remains intermediate for both oxides, while SOCl₂ performs poorly on ZrO₂. These relationships help determine suitable chemistries and processing windows for selective, optimized ALE.\"}]},{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Yes, both TiO₂ and ZrO₂ show higher etch rates at elevated temperatures, demonstrating thermally activated reaction kinetics.\"}]},{\"panel_id\":\"c\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Volatility of the reaction by-products (e.g., ZrCl₄, TiCl₄, WOCl₄), Stability of intermediate fluorides or sulfates, Activation energy of the ligand-exchange step, Ability of each co-reactant to remove fluorinated surface species\"}]},{\"panel_id\":\"b\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"BCl₃ shows the steepest increase in etch rate with temperature for ZrO₂, significantly outperforming SOCl₂ and TiCl₄.\"}]}]","bbox":[{"panel_id":"a","x":12,"y":0,"width":527,"height":433},{"panel_id":"b","x":552,"y":0,"width":529,"height":435}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/35/images/fig_6.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/35/images/fig_6.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/35/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/35/Comparison of BCl3, TiCl4, and SOCl2 chlorinating agents for.pdf"],"main_category":"atomic-layer-etching","sub_category":"experimental-usecase","paper_id":"35","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:image_caption"},"width":1081,"height":439,"image_format":"jpeg","image_sha256":"03381831d7ebb13107724f64f9f7bcc2cc77857d121bb69af5017e0ccaea9912","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/test_atomic-layer-etching_experimental-usecase_35_fig_7.jpg","caption":"FIG. 7. (a) In situ ellipsometry and [(b)-(d)] Auger electron spectroscopy measurements of $\\mathrm{TiO_2}$ thermal ALE using $\\mathrm{WF_6}$ and $\\mathrm{BCl_3}$ , $\\mathrm{SOCl_2}$ , or $\\mathrm{TiCl_4}$ at $170^{\\circ}\\mathrm{C}$ . The ellipsometry data were collected after every ALE cycle, while AES measurements were done before etching and after 5 and 25 ALE cycles.","id":"test/atomic-layer-etching/experimental-usecase/35/fig_7","sample_id":"atomic-layer-etching/experimental-usecase/35/fig_7","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"},{"panel_id":"b","label":"multiple line chart"},{"panel_id":"c","label":"multiple line chart"},{"panel_id":"d","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"shows TiO₂ thickness measured by in-situ ellipsometry during repeated ALE cycles at 170 °C for three chemistries. All curves fall to zero (film removed), but WF₆/SOCl₂ removes TiO₂ fastest, WF₆/TiCl₄ is intermediate, and WF₆/BCl₃ is slowest — each trace shows the number of cycles to full removal and the characteristic self-limiting stepwise decrease.\"},{\"panel_id\":\"b\",\"text\":\"shows AES time-series measured before etching and after 5 and 25 cycles for each chemistry. They track surface composition (O, Ti, Si, W, B, C, Cl etc.). AES corroborates ellipsometry: the Si substrate signal rises as TiO₂ is removed, while different chemistries leave distinct residues — e.g. measurable B for BCl₃, little residue for SOCl₂, and higher Cl/W signals for TiCl₄ — indicating differences in by-product volatility and surface cleanliness.\"},{\"panel_id\":\"c\",\"text\":\"shows AES time-series measured before etching and after 5 and 25 cycles for each chemistry. They track surface composition (O, Ti, Si, W, B, C, Cl etc.). AES corroborates ellipsometry: the Si substrate signal rises as TiO₂ is removed, while different chemistries leave distinct residues — e.g. measurable B for BCl₃, little residue for SOCl₂, and higher Cl/W signals for TiCl₄ — indicating differences in by-product volatility and surface cleanliness.\"},{\"panel_id\":\"d\",\"text\":\"shows AES time-series measured before etching and after 5 and 25 cycles for each chemistry. They track surface composition (O, Ti, Si, W, B, C, Cl etc.). AES corroborates ellipsometry: the Si substrate signal rises as TiO₂ is removed, while different chemistries leave distinct residues — e.g. measurable B for BCl₃, little residue for SOCl₂, and higher Cl/W signals for TiCl₄ — indicating differences in by-product volatility and surface cleanliness.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| ALE Cycles | WF6/SOCl2 Thickness (nm) | WF6/TiCl4 Thickness (nm) | WF6/BCl3 Thickness (nm) |\\n|------------|-------------------------|--------------------------|-------------------------|\\n| 0 | 1.9 | 1.7 | 1.8 |\\n| 2 | 1.5 | 1.4 | 1.3 |\\n| 4 | 1.2 | 1.1 | 0.9 |\\n| 6 | 0.9 | 0.8 | 0.5 |\\n| 8 | 0.5 | 0.4 | 0.2 |\\n| 10 | 0.1 | 0.1 | 0.0 |\\n| ≥12 | 0.0 | 0.0 | 0.0 |\"},{\"panel_id\":\"b\",\"text\":\"| ALE Cycles | Si (%) | O (%) | Ti (%) | B (%) | Cl (%) | C (%) |\\n|------------|--------|-------|--------|-------|--------|-------|\\n| 0 | 5 | 55 | 20 | 15 | 3 | 2 |\\n| 5 | 60 | 25 | 8 | 5 | 1 | 1 |\\n| 10 | 65 | 25 | 5 | 3 | 1 | 1 |\\n| 25 | 70 | 25 | 3 | 1 | 1 | 2 |\"},{\"panel_id\":\"c\",\"text\":\"| ALE Cycles | Si (%) | O (%) | Ti (%) | W (%) | Cl (%) | C (%) |\\n|------------|--------|-------|--------|-------|--------|-------|\\n| 0 | 5 | 55 | 20 | 10 | 5 | 5 |\\n| 5 | 45 | 35 | 10 | 6 | 3 | 1 |\\n| 10 | 55 | 38 | 5 | 2 | 0 | 0 |\\n| 25 | 65 | 40 | 2 | 1 | 0 | 0 |\"},{\"panel_id\":\"d\",\"text\":\"| ALE Cycles | Si (%) | O (%) | Ti (%) | Cl (%) | C (%) |\\n|------------|--------|-------|--------|--------|-------|\\n| 0 | 5 | 55 | 20 | 10 | 10 |\\n| 5 | 50 | 28 | 12 | 6 | 4 |\\n| 10 | 60 | 30 | 6 | 3 | 1 |\\n| 25 | 70 | 30 | 3 | 1 | 0 |\"}]","vqa":"[{\"panel_id\":\"d\",\"items\":[{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The AES results show that SOCl₂ leaves minimal surface contamination and enables rapid, complete removal of TiO₂, making it the most effective co-reactant. WF₆/BCl₃ leaves boron residues, and WF₆/TiCl₄ leaves significant chlorine and tungsten species, suggesting poorer by-product volatility. Therefore, SOCl₂ provides the cleanest and most efficient ALE chemistry for TiO₂ at 170 °C.\"}]},{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Yes, ellipsometry shows a stepwise, self-limiting decrease in TiO₂ thickness across ALE cycles for all chemistries, confirming self-limiting etching behavior.\"}]},{\"panel_id\":\"c\",\"items\":[{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"List\",\"answer\":\"Chlorine residues (especially for WF₆/TiCl₄), Boron residues (for WF₆/BCl₃), Detected tungsten signal (from WF₆ fluorination step), Carbon traces in some conditions\"}]},{\"panel_id\":\"b\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"SOCl₂ removes TiO₂ the fastest, TiCl₄ is intermediate, and BCl₃ is the slowest.\"}]}]","bbox":[{"panel_id":"a","x":205,"y":4,"width":723,"height":441},{"panel_id":"b","x":13,"y":449,"width":381,"height":325},{"panel_id":"c","x":393,"y":449,"width":343,"height":325},{"panel_id":"d","x":735,"y":449,"width":346,"height":325}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/35/images/fig_7.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/35/images/fig_7.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/35/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/35/Comparison of BCl3, TiCl4, and SOCl2 chlorinating agents for.pdf"],"main_category":"atomic-layer-etching","sub_category":"experimental-usecase","paper_id":"35","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:image_caption"},"width":1083,"height":775,"image_format":"jpeg","image_sha256":"a8f0edbe7b043bafc30cbedfc51762de6e15a1a3a7fa3b0b36f7feee9e3ba5b0","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/test_atomic-layer-etching_experimental-usecase_35_fig_8.jpg","caption":"FIG. 8. (a) In situ ellipsometry and [(b)-(d)] Auger electron spectroscopy measurements of $\\mathrm{ZrO_2}$ thermal ALE using $\\mathrm{WF_6}$ and $\\mathrm{BCl_3}$ , $\\mathrm{SOCl_2}$ , or $\\mathrm{TiCl_4}$ at $325^{\\circ}\\mathrm{C}$ . The ellipsometry data were collected after every ALE cycle, while AES measurements were done before etching and after 5 and 25 ALE cycles.","id":"test/atomic-layer-etching/experimental-usecase/35/fig_8","sample_id":"atomic-layer-etching/experimental-usecase/35/fig_8","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"},{"panel_id":"b","label":"multiple line chart"},{"panel_id":"c","label":"multiple line chart"},{"panel_id":"d","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"shows ZrO₂ thickness measured during repeated ALE cycles at 325 °C. WF₆/BCl₃ removes ZrO₂ fastest (film gone in ~3 cycles), WF₆/TiCl₄ is slightly slower (~4–5 cycles), and WF₆/SOCl₂ is ineffective (film persists after 25 cycles). The curves demonstrate self-limiting removal for the effective chemistries and a clear chemistry-dependent rate.\"},{\"panel_id\":\"b\",\"text\":\"present AES composition measured before etching and after 5 and 25 cycles for each chemistry. AES corroborates ellipsometry: for BCl₃, Zr and O drop rapidly and B residues appear; for SOCl₂, S signal is detectable and Zr/O remain high (sulfate formation and poor removal); for TiCl₄, Zr decreases moderately while Cl/W signatures indicate incomplete volatile conversion. Together the panels link kinetics (how quickly ZrO₂ is removed) with surface chemistry (what residues form), explaining why BCl₃ is optimal for ZrO₂ ALE and SOCl₂ fails.\"},{\"panel_id\":\"c\",\"text\":\"present AES composition measured before etching and after 5 and 25 cycles for each chemistry. AES corroborates ellipsometry: for BCl₃, Zr and O drop rapidly and B residues appear; for SOCl₂, S signal is detectable and Zr/O remain high (sulfate formation and poor removal); for TiCl₄, Zr decreases moderately while Cl/W signatures indicate incomplete volatile conversion. Together the panels link kinetics (how quickly ZrO₂ is removed) with surface chemistry (what residues form), explaining why BCl₃ is optimal for ZrO₂ ALE and SOCl₂ fails.\"},{\"panel_id\":\"d\",\"text\":\"present AES composition measured before etching and after 5 and 25 cycles for each chemistry. AES corroborates ellipsometry: for BCl₃, Zr and O drop rapidly and B residues appear; for SOCl₂, S signal is detectable and Zr/O remain high (sulfate formation and poor removal); for TiCl₄, Zr decreases moderately while Cl/W signatures indicate incomplete volatile conversion. Together the panels link kinetics (how quickly ZrO₂ is removed) with surface chemistry (what residues form), explaining why BCl₃ is optimal for ZrO₂ ALE and SOCl₂ fails.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| ALE Cycles | WF6/SOCl2 Thickness (nm) | WF6/TiCl4 Thickness (nm) | WF6/BCl3 Thickness (nm) |\\n|------------|-------------------------|--------------------------|-------------------------|\\n| 0 | 3.1 | 3.5 | 3.7 |\\n| 2 | 2.7 | 2.0 | 1.6 |\\n| 4 | 2.5 | 0.6 | 0.8 |\\n| 6 | 2.3 | 0.1 | 0.1 |\\n| 8 | 2.1 | 0.0 | 0.0 |\\n| 12 | 1.8 | 0.0 | 0.0 |\\n| 16 | 1.4 | 0.0 | 0.0 |\\n| 20 | 1.0 | 0.0 | 0.0 |\\n| 25 | 0.5 | 0.0 | 0.0 |\"},{\"panel_id\":\"b\",\"text\":\"| ALE Cycles | Al (%) | O (%) | C (%) | Zr (%) | B (%) |\\n|------------|--------|-------|-------|--------|-------|\\n| 0 | 0 | 55 | 20 | 25 | 0 |\\n| 5 | 15 | 30 | 25 | 20 | 10 |\\n| 10 | 25 | 35 | 20 | 15 | 8 |\\n| 25 | 45 | 40 | 8 | 5 | 2 |\"},{\"panel_id\":\"c\",\"text\":\"| ALE Cycles | Al (%) | O (%) | C (%) | Zr (%) | W (%) | S (%) |\\n|------------|--------|-------|-------|--------|-------|-------|\\n| 0 | 0 | 55 | 20 | 25 | 0 | 0 |\\n| 5 | 5 | 35 | 30 | 25 | 3 | 2 |\\n| 10 | 15 | 33 | 20 | 20 | 2 | 2 |\\n| 25 | 45 | 28 | 10 | 15 | 1 | 1 |\"},{\"panel_id\":\"d\",\"text\":\"| ALE Cycles | Al (%) | O (%) | C (%) | Zr (%) | W (%) | Ti (%) |\\n|------------|--------|-------|-------|--------|-------|--------|\\n| 0 | 0 | 55 | 20 | 25 | 0 | 0 |\\n| 5 | 10 | 35 | 25 | 25 | 2 | 3 |\\n| 10 | 20 | 35 | 20 | 20 | 2 | 3 |\\n| 25 | 55 | 32 | 10 | 12 | 1 | 2 |\"}]","vqa":"[{\"panel_id\":\"d\",\"items\":[{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The ellipsometry data show that BCl₃ and TiCl₄ enable rapid, self-limiting ZrO₂ removal, whereas SOCl₂ fails due to slow kinetics. AES confirms this by revealing B residues for BCl₃ and Cl/W residues for TiCl₄, but heavy S residues and persistent Zr/O for SOCl₂. Therefore, WF₆/BCl₃ is the most efficient chemistry for ZrO₂ ALE, while SOCl₂ should be avoided because it forms nonvolatile sulfates.\"}]},{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Yes. WF₆/BCl₃ and WF₆/TiCl₄ rapidly remove ZrO₂ within a few cycles, while WF₆/SOCl₂ barely etches the film even after 25 cycles.\"}]},{\"panel_id\":\"c\",\"items\":[{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"List\",\"answer\":\"Sulfur-containing species (S), Persistent Zr and O (indicating unremoved oxide), Minimal exposure of the Si substrate, These suggest sulfate formation, which is nonvolatile and blocks etching.\"}]},{\"panel_id\":\"b\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"WF₆/BCl₃ yields the fastest ZrO₂ removal, with the film almost completely etched in about 3 cycles.\"}]}]","bbox":[{"panel_id":"a","x":211,"y":7,"width":712,"height":436},{"panel_id":"b","x":5,"y":449,"width":385,"height":326},{"panel_id":"c","x":392,"y":451,"width":343,"height":323},{"panel_id":"d","x":735,"y":450,"width":342,"height":324}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/35/images/fig_8.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/35/images/fig_8.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/35/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/35/Comparison of BCl3, TiCl4, and SOCl2 chlorinating agents for.pdf"],"main_category":"atomic-layer-etching","sub_category":"experimental-usecase","paper_id":"35","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:image_caption"},"width":1080,"height":775,"image_format":"jpeg","image_sha256":"294403cc9d8f5fdf52853e928ab47e068f2c6ecc549cd1a21c0d3417c98e0054","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/test_atomic-layer-etching_experimental-usecase_35_fig_9.jpg","caption":"FIG. 9. RMS surface roughness evolution as a function of ALE cycles as measured by atomic force microscopy. The ALE of $\\mathrm{TiO_2}$ films was performed at $170^{\\circ}\\mathrm{C}$ and the $\\mathrm{ZrO_2}$ films at $325^{\\circ}\\mathrm{C}$ .","id":"test/atomic-layer-etching/experimental-usecase/35/fig_9","sample_id":"atomic-layer-etching/experimental-usecase/35/fig_9","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The plot shows RMS surface roughness (pm) measured by AFM as a function of ALE cycles for TiO₂ (170 °C) and ZrO₂ (325 °C) using WF₆ paired with BCl₃, SOCl₂, or TiCl₄. Both materials start with higher roughness and show a rapid drop after a few cycles; ZrO₂ begins much rougher (~180–190 pm) and smooths more dramatically than TiO₂. Effective chemistries (WF₆/BCl₃ and WF₆/TiCl₄ for ZrO₂; WF₆/SOCl₂ and WF₆/TiCl₄ for TiO₂) produce greater smoothing, while SOCl₂ is least effective at smoothing ZrO₂ — matching its poor etch performance. Overall, ALE leads to net surface smoothing, and the degree of smoothing correlates with the chemistry’s etch effectiveness.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Material|Initial Roughness (0 cycles)|Roughness after 5 cycles|Roughness after 25 cycles|Trend Summary|\\n|---------|----------------------------|------------------------|-------------------------|-------------|\\n|TiO₂|~90–100 pm|Decreases for all chemistries|Further decreases; becomes smoother overall|ALE smooths TiO₂ surfaces|\\n|ZrO₂|~180–190 pm|Sharp drop for all chemistries|Chemistry-dependent smoothing; SOCl₂ least effective|ZrO₂ starts rougher but smooths strongly with effective chemistries|\"}]","vqa":"[{\"panel_id\":\"d\",\"items\":[{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Roughness evolution provides insight into how uniformly each chemistry removes material. Effective chemistries produce greater smoothing due to consistent layer removal. For TiO₂, SOCl₂ and TiCl₄ lead to smoother surfaces, while for ZrO₂, BCl₃ and TiCl₄ yield the strongest smoothing effects. In contrast, SOCl₂ produces limited smoothing on ZrO₂, reflecting its poor etch performance. Thus, AFM roughness trends help confirm which co-reactant chemistries produce uniform, high-quality ALE.\"}]},{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Yes. Both materials exhibit noticeable smoothing after ALE cycles, although the degree of smoothing depends on the co-reactant chemistry.\"}]},{\"panel_id\":\"c\",\"items\":[{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"List\",\"answer\":\"Initial surface morphology (TiO₂ smoother; ZrO₂ rougher), Reaction efficiency of each co-reactant, Volatility of reaction by-products, Uniformity of layer-by-layer removal\"}]},{\"panel_id\":\"b\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"ZrO₂ shows the largest reduction in roughness, dropping from ~180–190 pm initially to much lower values after 25 cycles.\"}]}]","bbox":[{"panel_id":"a","x":0,"y":0,"width":544,"height":450}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/35/images/fig_9.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/35/images/fig_9.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/35/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/35/Comparison of BCl3, TiCl4, and SOCl2 chlorinating agents for.pdf"],"main_category":"atomic-layer-etching","sub_category":"experimental-usecase","paper_id":"35","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:image_caption"},"width":544,"height":450,"image_format":"jpeg","image_sha256":"eb057e3c6e483479fb992f76b63458f72fbca7e249306e615bad90b7d7385db1","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/test_atomic-layer-etching_experimental-usecase_37_fig_2.jpg","caption":"FIG. 2. In situ ellipsometry data collected during deposition and etching steps on the Si-OH surface. (a) and (b) show the raw delta $(\\Delta)$ and psi $(\\Psi^{\\prime})$ parameters, respectively, which are fitted by a Cauchy model to (c) measure the $\\mathrm{TiO_2}$ layer thickness.","id":"test/atomic-layer-etching/experimental-usecase/37/fig_2","sample_id":"atomic-layer-etching/experimental-usecase/37/fig_2","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"},{"panel_id":"b","label":"multiple line chart"},{"panel_id":"c","label":"line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The line chart shows the trend of Δ values over cycles for four different wavelengths (465 nm, 525 nm, 580 nm, and 635 nm). The Δ values decrease with increasing cycles until around cycle 240, after which they start to increase again.\"},{\"panel_id\":\"b\",\"text\":\"The line chart displays the constant values of Ψ for four different wavelengths (465 nm, 525 nm, 580 nm, and 635 nm) over cycles. The values remain stable throughout the cycles.\"},{\"panel_id\":\"c\",\"text\":\"The line chart illustrates the thickness of a material over cycles. The thickness increases steadily until cycle 240, after which it decreases slightly.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Cycles | 465 nm | 525 nm | 580 nm | 635 nm |\\n|---|---|---|---|---|\\n| 210 | 164 | 166 | 168 | 168 |\\n| 215 | 163 | 165 | 167 | 167 |\\n| 220 | 162 | 164 | 166 | 166 |\\n| 225 | 161 | 163 | 165 | 165 |\\n| 230 | 160 | 162 | 164 | 164 |\\n| 235 | 159 | 161 | 163 | 163 |\\n| 240 | 158 | 160 | 162 | 162 |\\n| 245 | 161 | 163 | 165 | 166 |\"},{\"panel_id\":\"b\",\"text\":\"| Cycles | 465 nm | 525 nm | 580 nm | 635 nm |\\n|---|---|---|---|---|\\n| 210 | 21 | 19 | 17 | 17 |\\n| 215 | 21 | 19 | 17 | 17 |\\n| 220 | 21 | 19 | 17 | 17 |\\n| 225 | 21 | 19 | 17 | 17 |\\n| 230 | 21 | 19 | 17 | 17 |\\n| 235 | 21 | 19 | 17 | 17 |\\n| 240 | 21 | 19 | 17 | 17 |\\n| 245 | 21 | 19 | 17 | 17 |\"},{\"panel_id\":\"c\",\"text\":\"| Cycles | Thickness (nm) |\\n|---|---|\\n| 210 | 4.8 |\\n| 215 | 5.0 |\\n| 220 | 5.2 |\\n| 225 | 5.4 |\\n| 230 | 5.6 |\\n| 235 | 5.8 |\\n| 240 | 6.0 |\\n| 245 | 5.8 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Δ decreases linearly during deposition and increases linearly during etching.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"At 170 °C.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"In figure 2, the raw ellipsometry parameters Ψ and Δ, measured at four wavelengths, are shown alongside the thickness extracted using a temperature-dependent Cauchy model. While Ψ remains essentially constant over the small thickness range involved, Δ exhibits a clear correlation with the TiO₂ film thickness. This demonstrates that Δ is the primary ellipsometric parameter for detecting incremental growth and removal during ALD and ALE.\"}]},{\"panel_id\":\"b\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Ψ does not change significantly during both deposition and etching.\"}]}]","bbox":[{"panel_id":"a","x":2,"y":8,"width":611,"height":498},{"panel_id":"b","x":5,"y":496,"width":611,"height":508},{"panel_id":"c","x":0,"y":1002,"width":616,"height":521}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/37/images/fig_2.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/37/images/fig_2.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/37/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/37/Effect of reactant dosing on selectivity during area-selective.pdf"],"main_category":"atomic-layer-etching","sub_category":"experimental-usecase","paper_id":"37","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:image_caption"},"width":619,"height":1528,"image_format":"jpeg","image_sha256":"8b69b59580729319c832ee6fcf7fca6d5c1715a0743dd66d14bbfe3d0b0edb11","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/test_atomic-layer-etching_experimental-usecase_37_fig_8.jpg","caption":"FIG. 8. $\\mathrm{TiO_2}$ film thickness on $\\mathrm{Si - H}$ and $\\mathrm{Si - OH}$ surfaces during ALE/ALE super-cycles with $90~\\mathrm{ms}$ dosing conditions for $\\mathrm{TiCl_4 / H_2O}$ . The measurement was done by in situ ellipsometry, with the data point taken after every five cycles for ALD and every cycle for ALE.","id":"test/atomic-layer-etching/experimental-usecase/37/fig_8","sample_id":"atomic-layer-etching/experimental-usecase/37/fig_8","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The line chart shows the thickness of Si-OH and Si-H as a function of super-cycles. Both thicknesses increase linearly with increasing super-cycles.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Super-Cycles | Thickness (nm) Si-H | Thickness (nm) Si-OH |\\n|---|---|---|\\n| 0 | 0 | 0 |\\n| 5 | 0 | 4 |\\n| 10 | 0 | 8 |\\n| 15 | 0 | 13 |\\n| 20 | 0 | 17.5 |\\n| 25 | 0.5 | 22.5 |\\n| 30 | 0.5 | 25 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Si-OH and Si-H.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"30.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"90 ms for both.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"170 °C.\"}]}]","bbox":[{"panel_id":"a","x":0,"y":0,"width":672,"height":492}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/37/images/fig_8.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/37/images/fig_8.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/37/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/37/Effect of reactant dosing on selectivity during area-selective.pdf"],"main_category":"atomic-layer-etching","sub_category":"experimental-usecase","paper_id":"37","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:image_caption"},"width":672,"height":492,"image_format":"jpeg","image_sha256":"504269b9a3d7a1f4bc42d1c9686a34847d9b839b84d61184f9f51951fef3e456","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/test_atomic-layer-etching_experimental-usecase_44_fig_5.jpg","caption":"FIG. 5. (Color online) Film thickness vs number of AlN ALE reaction cycles at $275^{\\circ}\\mathrm{C}$ in pure AlN region of AlN film showing results for thermal AlN ALE and plasma-enhanced ALE. (a) $\\mathrm{H}_{2}$ plasma increases AlN etch rate from $0.36$ to $1.96\\mathrm{\\AA}$ /cycle. (b) Ar plasma increases AlN etch rate from $0.36$ to $0.66\\mathrm{\\AA}$ /cycle.","id":"test/atomic-layer-etching/experimental-usecase/44/fig_5","sample_id":"atomic-layer-etching/experimental-usecase/44/fig_5","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"},{"panel_id":"b","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The line chart shows the thickness (Å) of AlN as a function of ALE reaction cycle number at 275 °C. Two ALE processes are compared: thermal ALE from 0 to 14 cycles and H₂ plasma-enhanced ALE from 15 to 26 cycles. The introduction of H₂ plasma increases the AlN etch rate from 0.36 to 1.96 Å/cycle.\"},{\"panel_id\":\"b\",\"text\":\"The line chart shows the thickness (Å) of AlN as a function of ALE reaction cycle number at 275 °C. Two ALE processes are compared: thermal ALE from 0 to 14 cycles and Ar plasma-enhanced ALE from 15 to 26 cycles. The use of Ar plasma increases the AlN etch rate from 0.36 to 0.66 Å/cycle. This image clearly indicates that Ar Plasma enhnaced ALE shows a higher etch rate.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Cycle Number | Thickness (Å) |\\n|---|---|\\n| 0 | 450 |\\n| 4 | 449 |\\n| 8 | 447 |\\n| 12 | 446 |\\n| 16 | 441 |\\n| 20 | 433 |\\n| 24 | 438 |\\n| 28 | 425 |\"},{\"panel_id\":\"b\",\"text\":\"| Cycle Number | Thickness (Å) |\\n|---|---|\\n| 0 | 438 |\\n| 4 | 436 |\\n| 8 | 435 |\\n| 12 | 433 |\\n| 16 | 430 |\\n| 20 | 428 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"List\",\"answer\":\"HF Pulse & Purge.\\n\\n, Sn(acac)₂ Pulse & Purge.\\n\\n, H₂ Plasma Exposure (new step).\\n\\n, Purge.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"H₂ plasma provides a ~5x enhancement (chemical), while Ar plasma gives only a ~2x enhancement (physical). This reveals the trade-off: H₂ plasma offers higher efficiency via chemical radical reactions (removing blocking ligands), but may introduce more hydrogen incorporation. Ar plasma offers a cleaner, purely physical enhancement (ion/radiation) but with significantly lower etch rate gains, requiring a balance between speed and potential plasma-induced damage or impurity introduction.\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Residual acac surface species (ligands) that block reactive sites.\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Ar plasma may be better suited for high-aspect-ratio features because it avoids the radical depletion and non-uniformity issues that H₂ radicals can cause in deep trenches. While slower, its more isotropic ion-driven enhancement can improve etch conformity and reduce loading effects in complex geometries.\"}]}]","bbox":[{"panel_id":"a","x":6,"y":5,"width":592,"height":530},{"panel_id":"b","x":0,"y":550,"width":601,"height":530}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/44/images/fig_5.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/44/images/fig_5.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/44/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/44/Thermal atomic layer etching of crystalline aluminum nitride.pdf"],"main_category":"atomic-layer-etching","sub_category":"experimental-usecase","paper_id":"44","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:image_caption"},"width":605,"height":1083,"image_format":"jpeg","image_sha256":"4442b55d732128cfd726cfc195a60c74235cf28ae1301645f198389631099f3a","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/test_atomic-layer-etching_experimental-usecase_47_FIG1.jpg","caption":"FIG. 1. Values of Gibbs free energy change for oxidation reactions of molybdenum to $\\mathrm{MoO}_2$ and $\\mathrm{MoO}_3$ with $\\mathrm{H}_2\\mathrm{O}$ , $\\mathrm{N}_2\\mathrm{O}$ , $\\mathrm{O}_2$ , and $\\mathrm{O}_3$ at $0 - 500^{\\circ}\\mathrm{C}$ .","id":"test/atomic-layer-etching/experimental-usecase/47/FIG1","sample_id":"atomic-layer-etching/experimental-usecase/47/FIG1","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The chart displays the Gibbs free energy change (ΔG) for various reactions of molybdenum (Mo) with different oxidants at different temperatures.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Reaction | ΔG Range (kJ/mol) | Trend with Temperature |\\n|----------------------------------------------|--------------------|-------------------------|\\n| Mo + 3 H₂O(g) → MoO₃ + 3 H₂(g) | ~0 to +100 | Slightly increases |\\n| Mo + H₂O(g) → MoO₂ + H₂(g) | ~-200 to -100 | Increases |\\n| Mo + O₂(g) → MoO₂ | ~-600 to -500 | Increases |\\n| Mo + 1.5 O₂(g) → MoO₃ | ~-600 to -500 | Increases |\\n| Mo + 2 N₂O(g) → MoO₂ + 2 N₂(g) | ~-800 to -700 | Increases |\\n| Mo + 2 O₃(g) → MoO₂ + O₂(g) | ~-900 to -800 | Increases |\\n| Mo + 3 N₂O(g) → MoO₃ + 3 N₂(g) | ~-1000 to -900 | Increases |\\n| Mo + 3 O₃(g) → MoO₃ + O₂(g) | ~-1100 to -1000 | Increases |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Gibbs free energy changes of the oxidation reactions of molybdenum using H2O, N2O, O2, and O3 as oxidizing agents were calculated to understand the feasibility of ALE\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"MoO2 and MoO3 were considered, as they can form volatile chlorides and oxychlorides via deoxychlorination reactions with NbCl5. Oxides with lower oxidation states were also not available in the thermodynamic database.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Yes/No\",\"answer\":\"No\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The energy changes are negative over the whole temperature range, except for the reaction with H2O, which has a positive Gibbs free energy change in this temperature range.\"}]}]","bbox":[{"panel_id":"a","x":3,"y":6,"width":664,"height":533}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/47/images/FIG1.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/47/images/FIG1.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/47/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/47/Atomic layer etching of molybdenum with O2 and NbCl5.pdf"],"main_category":"atomic-layer-etching","sub_category":"experimental-usecase","paper_id":"47","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:image_caption"},"width":669,"height":542,"image_format":"jpeg","image_sha256":"af6ab55fe4fc356703b246e5ba4baa6f56512cb113634fb9e0f4996466331fc9","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/test_atomic-layer-etching_experimental-usecase_47_FIG2.jpg","caption":"FIG. 2. Values of Gibbs free energy changes of etching reactions of $\\mathrm{MoO}_2$ and $\\mathrm{MoO}_3$ with $\\mathrm{NbCl}_5$ at $0 - 500^{\\circ}\\mathrm{C}$ .","id":"test/atomic-layer-etching/experimental-usecase/47/FIG2","sample_id":"atomic-layer-etching/experimental-usecase/47/FIG2","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The chart shows the gibbs free energy changes as a function of temperature for various chemical reactions.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Reaction | ΔG Range (kJ/mol) | Trend with Temperature |\\n|--------------------------------------------------------------|--------------------|-------------------------|\\n| MoO₂ + NbCl₅(g) → MoOCl₂(g) + NbOCl₃(g) | ~150 to ~50 | Decreases |\\n| MoO₃ + 3 NbCl₅(g) → MoCl₆(g) + 3 NbOCl₃(g) | ~120 to ~20 | Decreases |\\n| MoO₂ + 2 NbCl₅(g) → MoOCl₄(g) + 2 NbOCl₃(g) | ~50 to ~-50 | Decreases |\\n| MoO₃ + 2 NbCl₅(g) → MoOCl₄(g) + 2 NbOCl₃(g) | ~0 to ~-100 | Decreases |\\n| MoO₃ + NbCl₅(g) → MoO₂Cl₂(g) + NbOCl₃(g) | ~-50 to ~-150 | Decreases |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Yes/No\",\"answer\":\"Yes\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The etching of oxides occurs by deoxychlorination, producing NbOCl3 and molybdenum chlorides or oxychlorides as reaction products. MoCl4 is the only thermodynamically favorable etch product for MoO2 and is able to form above 200 °C. Conversion of MoO2 to a solid MoOCl2 is also favorable, but etching via this reaction pathway is prevented by the low volatility of MoOCl2\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"For MoO3, full chlorination is not thermodynamically favorable, but both MoO2Cl2 and MoOCl4 are possible etch products above about 20 °C.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The etching of metallic molybdenum by NbCl5 alone was also considered. Possible reactions leading to continuous etching would be oxidation of the metallic Mo to either MoCl4 or MoCl6 by NbCl5, which itself would be reduced to NbCl4. Neither reaction is thermodynamically favorable as both have highly positive Gibbs free energy changes at 0–500 °C. This indicates that continuous etching of Mo by NbCl5 is not an issue for this process, but does not mean that NbCl5 is completely inert toward Mo\"}]}]","bbox":[{"panel_id":"a","x":4,"y":6,"width":664,"height":534}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/47/images/FIG2.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/47/images/FIG2.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/47/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/47/Atomic layer etching of molybdenum with O2 and NbCl5.pdf"],"main_category":"atomic-layer-etching","sub_category":"experimental-usecase","paper_id":"47","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:image_caption"},"width":669,"height":542,"image_format":"jpeg","image_sha256":"279faf211e5037756e795bcacfe2e1aa2346ed3086f9699677333f502ef3fbd9","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/test_atomic-layer-etching_experimental-usecase_47_FIG3.jpg","caption":"FIG. 3. Equilibrium composition for an input of $0.1\\mathrm{kmol}$ $\\mathrm{NbCl}_5$ and $0.02\\mathrm{kmol}$ Mo at 10 mbar pressure. $5\\mathrm{kmol}$ of $\\mathbb{N}_2$ was used as background gas to approximate the partial pressure of $\\mathrm{NbCl}_5$ in the reactor setup used in this paper.","id":"test/atomic-layer-etching/experimental-usecase/47/FIG3","sample_id":"atomic-layer-etching/experimental-usecase/47/FIG3","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The chart shows the amount of various chlorides (NbCl₄(s), NbCl₅(s), MoCl₃(s), NbCl₃(s), NbCl₂.67(s), NbCl₄(g), MoCl₂(s), MoCl₄(g)) in kmol as a function of temperature in °C.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Temperature (°C) | NbCl₅(g) | NbCl₅(s) | NbCl₄(s) | NbCl₄(g) | NbCl₃(s) | NbCl₂.₆₇(s) | MoCl₃(s) | MoCl₂(s) | MoCl₄(g) |\\n|-------------------|----------|----------|----------|----------|----------|-------------|----------|----------|----------|\\n| 0 | 0 | 0.04 | 0.06 | 0 | 0 | 0 | 0.02 | 0 | 0 |\\n| 50 | 0.01 | 0.04 | 0.05 | 0 | 0 | 0 | 0.02 | 0 | 0 |\\n| 100 | 0.05 | 0.03 | 0.02 | 0.005 | 0 | 0 | 0.02 | 0 | 0 |\\n| 150 | 0.07 | 0.01 | 0.005 | 0.01 | 0.015 | 0.005 | 0.02 | 0 | 0 |\\n| 200 | 0.07 | 0 | 0 | 0.015 | 0.02 | 0.01 | 0.02 | 0 | 0 |\\n| 250 | 0.07 | 0 | 0 | 0.02 | 0.015 | 0.015 | 0.02 | 0.005 | 0.005 |\\n| 300 | 0.07 | 0 | 0 | 0.025 | 0.01 | 0.015 | 0.02 | 0.01 | 0.01 |\\n| 350 | 0.07 | 0 | 0 | 0.03 | 0.005 | 0.015 | 0.02 | 0.015 | 0.015 |\\n| 400 | 0.07 | 0 | 0 | 0.035 | 0 | 0.015 | 0.02 | 0.02 | 0.02 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Equilibrium composition calculated by minimizing the Gibbs\\nfree energy of the system when Mo and NbCl5(g) are given as the\\nstarting materials indicates that they can react to form nonvolatile\\nMoCl3 and various lower niobium chlorides (Fig. 3). Below 150 °C,\\nthe dominant Nb-containing species is solid NbCl4, and above this\\ntemperature up to 400 °C, the reacted NbCl5 is mostly expected to\\nform NbCl3 and Nb3Cl8. Once the oxidized surface layer has been\\netched, these lower chlorides could form a layer that prevents\\nfurther reactions.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Equilibrium composition calculated by minimizing the Gibbs free energy of the system when Mo and NbCl5(g) are given as the starting materials indicates that they can react to form nonvolatile MoCl3 and various lower niobium chlorides.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Below 150 °C, the dominant Nb-containing species is solid NbCl4, and above this temperature up to 400 °C, the reacted NbCl5 is mostly expected to\\nform NbCl3 and Nb3Cl8. Once the oxidized surface layer has been etched, these lower chlorides could form a layer that prevents further reactions.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Equilibrium compositions of both NbCl3 and MoCl3 were calculated in the presence of excess O2 to evaluate their effect on the etching process. It was\\nfound that for NbCl3 the most thermodynamically favorable product is NbO2Cl below about 240 °C and Nb2O5 above it. For MoCl3, the product is MoO3 below 350 °C and gaseous MoO2Cl2 above 350 °C. This indicates that if the lower chlorides are formed, they are readily oxidized and then etched in the subsequent ALE cycle, as NbCl5 has been shown to also etch Nb2O5\"}]}]","bbox":[{"panel_id":"a","x":5,"y":3,"width":665,"height":520}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/47/images/FIG3.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/47/images/FIG3.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/47/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/experimental-usecase/47/Atomic layer etching of molybdenum with O2 and NbCl5.pdf"],"main_category":"atomic-layer-etching","sub_category":"experimental-usecase","paper_id":"47","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:image_caption"},"width":673,"height":528,"image_format":"jpeg","image_sha256":"89930675ddc469508084cf51e8d9e32177e4451374dc00d4b8d2f80eca645281","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/test_atomic-layer-etching_simulation-usecase_11_fig_4.jpg","caption":"Fig. 4 - Reaction progression of half-cycles for both steps over time with an operating temperature of 573 K. The full progression is achieved at 1.38 s for Step A and 2.38 s for Step B, respectively.","id":"test/atomic-layer-etching/simulation-usecase/11/fig_4","sample_id":"atomic-layer-etching/simulation-usecase/11/fig_4","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"Kinetic Monte Carlo simulation results showing the progression of half-cycle completion (%) versus time for both ALE steps at 573 K. Step A (HF fluorination, blue) reaches 100% surface coverage at 1.38 s, while Step B (TMA etching, orange) completes at 2.38 s. Both curves show rapid initial progress followed by a gradual approach to saturation, characteristic of self-limiting ALE surface reactions.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Time (s) | Step A (%) | Step B (%) |\\n|---|---|---|\\n| 0.00 | 0 | 0 |\\n| 0.05 | 16 | 0 |\\n| 0.10 | 41 | 10 |\\n| 0.15 | 50 | 25 |\\n| 0.20 | 64 | 38 |\\n| 0.25 | 72 | 48 |\\n| 0.30 | 77 | 56 |\\n| 0.40 | 84 | 68 |\\n| 0.50 | 91 | 78 |\\n| 0.60 | 94 | 82 |\\n| 0.80 | 98 | 91 |\\n| 1.00 | 99 | 95 |\\n| 1.25 | 99.5 | 96 |\\n| 1.50 | 100 | 98 |\\n| 2.00 | 100 | 99.5 |\\n| 2.50 | 100 | 100 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Step B (TMA etching) involves a more complex reaction mechanism with multiple sequential steps including TMA adsorption, ligand exchange, and DMAF desorption at two different reaction sites (l-AlF2 and r-AlF3). Additionally, TMA has a lower sticking coefficient (0.02) compared to HF (0.15), which slows the adsorption kinetics. The requirement for multiple TMA molecules to complete the etching at each unit cell further extends the process time.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"The stepwise jumps reflect the discrete nature of the kMC simulation, where individual reaction events occur stochastically at specific time intervals rather than continuously.\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"For Step A, 100% means complete fluorination of the Al2O3 surface as AlF3; for Step B, 100% means complete removal of the fluorinated layer through DMAF volatilization.\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The simulated times agree well with experiments. For Step A, experiments showed self-limiting behavior after 1.0 s of HF exposure with no significant change after 1.5 s, consistent with the simulated 1.38 s. For Step B, experiments indicated completion after approximately 2.25 s of TMA exposure, closely matching the simulated 2.38 s. This agreement validates the kMC model.\"}]}]","bbox":[{"panel_id":"a","x":2,"y":4,"width":607,"height":445}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/11/images/fig_4.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/11/images/fig_4.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/11/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/11/Microscopic and data-driven modeling andoperation.pdf"],"main_category":"atomic-layer-etching","sub_category":"simulation-usecase","paper_id":"11","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:img_caption"},"width":609,"height":450,"image_format":"jpeg","image_sha256":"dc452a2b98d9de89fd573d1a6a4878896833900a6fedab44bb145db6a7503f5a","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/test_atomic-layer-etching_simulation-usecase_14_fig_1.jpg","caption":"Fig. 1. Changes in the heights of the material surfaces during $300\\mathrm{eV}$ $\\mathrm{Ar^{+}}$ ion irradiation, obtained from MD simulations. With HFC, an about $2\\mathrm{nm}$ thick HFC layer was deposited on the flat model SiN surface. Without HFC, the flat model SiN surface was etched without HFC deposition. The horizontal axis represents the ion dose. Two vertical broken lines delineate different desorption regimes I, II, and, III, discussed in the main text. For the atomic representations of the surfaces, the reader is referred to section II of the Supplementary Material.","id":"test/atomic-layer-etching/simulation-usecase/14/fig_1","sample_id":"atomic-layer-etching/simulation-usecase/14/fig_1","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The line chart presents depth (in nm) versus ion dose (in ×10¹⁶ ions/cm²) for Ar⁺ bombardment under two experimental conditions: with and without hydrofluorocarbon (HFC). The sample treated with HFC initially shows greater surface material removal (positive depth), followed by gradual etching until reaching a net depth loss similar to the untreated case. Three regimes (I–III) are marked with dashed lines, indicating transitions in material removal behavior. The untreated sample displays relatively consistent shallow etching across the dose range.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Ion Dose (×10¹⁶ ions cm⁻²) | Depth [nm] with HFC | Depth [nm] without HFC |\\n|----------------------------|---------------------|-------------------------|\\n| 0.0 | 2.0 | 0.1 |\\n| 0.2 | 1.5 | 0.0 |\\n| 0.5 | 1.0 | -0.1 |\\n| 0.8 | 0.4 | -0.2 |\\n| 1.0 | 0.0 | -0.3 |\\n| 1.2 | -0.2 | -0.4 |\\n| 1.5 | -0.4 | -0.5 |\\n| 2.0 | -0.6 | -0.6 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Depth (nm) change of the sample surface.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Yes/No\",\"answer\":\"Yes\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The segmentation into regions I, II, and III indicates distinct phases of the etching process. In Region I, there is rapid removal of material in the HFC-treated sample, likely due to a reactive surface layer. Region II shows a slower reduction, suggesting the surface becomes more resistant to ion bombardment. Region III stabilises, implying that a steady-state etch rate is reached. This behavior contrasts with the untreated sample, which exhibits a relatively linear and shallow etch throughout.\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"List\",\"answer\":\"Catalytic effect of HFC, Surface conditioning with Ar⁺, Ion dose optimization, Reaction regime identification\"}]}]","bbox":[{"panel_id":"a","x":0,"y":0,"width":667,"height":619}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/14/images/fig_1.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/14/images/fig_1.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/14/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/14/Molecular dynamics simulations of silicon nitride.pdf"],"main_category":"atomic-layer-etching","sub_category":"simulation-usecase","paper_id":"14","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:img_caption"},"width":667,"height":619,"image_format":"jpeg","image_sha256":"8114977bbf93a24caf2b3de6184bafa11395fe2e27caec1c6c83bd653a01aab7","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/test_atomic-layer-etching_simulation-usecase_14_fig_2.jpg","caption":"Fig. 2. Changes in the heights of the material surface during three-step SiN PE-ALE with (a) $\\mathrm{Ar^{+}}$ b $\\mathrm{Kr^{+}}$ and (c) $\\mathrm{Xe^{+}}$ ion irradiations over the first five cycles, obtained from MD simulations. Each FE-ALE cycle consists of the adsorption step (denoted by $①$ , desorption step $(2)$ , and oxidation step $(3)$ . The lower curves denoted as physical suttering indicate the changes in the SIN material surface positions by physical suttering with the corresponding ions as functions of the ion dose. For the curves representing the PE-ALE processes, the height changes are also represented as functions of the ion dose during the desorption step and a unit length of the horizontal axis represents the same ion dose for both PE-ALE and physical sputtering curves. The PE-ALE simulation conditions are summarized in Table I.","id":"test/atomic-layer-etching/simulation-usecase/14/fig_2","sample_id":"atomic-layer-etching/simulation-usecase/14/fig_2","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"},{"panel_id":"b","label":"multiple line chart"},{"panel_id":"c","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The figure compares atomic layer etching (ALE) and physical sputtering depths over five cycles for three different noble gas ions: Ar⁺, Kr⁺, and Xe⁺. In each subplot, the ALE condition (black line) shows periodic, steep decreases in surface depth, indicating controlled etching per cycle. In contrast, the red traces for physical sputtering display minimal and more linear depth reduction, suggesting that physical sputtering alone removes significantly less material per cycle. The graphs are labeled with Roman numerals (I–III) that denote distinct etch regions, possibly representing phases of material response or tool behavior during processing.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Cycle | ALE Depth [nm] | Sputtering Depth [nm] |\\n|-------|----------------|------------------------|\\n| 0 | 0 | 0 |\\n| 1 | 1.9 | -0.2 |\\n| 2 | 3.5 | -0.4 |\\n| 3 | 5.2 | -0.6 |\\n| 4 | 7.0 | -0.8 |\\n| 5 | 8.5 | -1.0 |\"},{\"panel_id\":\"b\",\"text\":\"| Cycle | ALE Depth [nm] | Sputtering Depth [nm] |\\n|-------|----------------|------------------------|\\n| 0 | 0 | 0 |\\n| 1 | 2.1 | -0.3 |\\n| 2 | 4.3 | -0.6 |\\n| 3 | 6.2 | -0.8 |\\n| 4 | 8.1 | -1.0 |\\n| 5 | 9.8 | -1.2 |\"},{\"panel_id\":\"c\",\"text\":\"| Cycle | ALE Depth [nm] | Sputtering Depth [nm] |\\n|-------|----------------|------------------------|\\n| 0 | 0 | 0 |\\n| 1 | 2.3 | -0.4 |\\n| 2 | 4.7 | -0.7 |\\n| 3 | 7.0 | -1.0 |\\n| 4 | 9.3 | -1.3 |\\n| 5 | 11.5 | -1.5 |\"}]","vqa":"[{\"panel_id\":\"b\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Yes/No\",\"answer\":\"Yes\"}]},{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"To evaluate the effectiveness and control of etching using noble gas ions (Ar⁺, Kr⁺, Xe⁺) under ALE and sputtering conditions\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"List\",\"answer\":\"Xe⁺, Kr⁺, Ar⁺\"}]},{\"panel_id\":\"c\",\"items\":[{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The labeled regions likely denote distinct etching regimes: Region I shows the initial rapid etching response, possibly due to surface activation or reactive layer removal. Region II marks a more stable, consistent etch rate per cycle, typical of self-limiting ALE behavior. Region III may indicate a saturation or termination phase, where etch depth stabilises or material is fully removed.\"}]}]","bbox":[{"panel_id":"a","x":9,"y":0,"width":655,"height":507},{"panel_id":"b","x":740,"y":0,"width":652,"height":508},{"panel_id":"c","x":375,"y":555,"width":650,"height":498}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/14/images/fig_2.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/14/images/fig_2.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/14/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/14/Molecular dynamics simulations of silicon nitride.pdf"],"main_category":"atomic-layer-etching","sub_category":"simulation-usecase","paper_id":"14","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:img_caption"},"width":1392,"height":1053,"image_format":"jpeg","image_sha256":"dbea9967a3cfd22ea9fe0541034b9b839f10415c303a0217d23bd56629ab172a","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/test_atomic-layer-etching_simulation-usecase_14_fig_4.jpg","caption":"Fig. 4. Depth profiles of atomic concentrations at the end of the oxidation step of the first SiN PE-ALE cycle with (a) $\\mathrm{Ar^{+}}$ , (b) $\\mathrm{Kr^{+}}$ , and (c) $\\mathrm{Xe^{+}}$ ion irradiations, obtained from MD simulations.","id":"test/atomic-layer-etching/simulation-usecase/14/fig_4","sample_id":"atomic-layer-etching/simulation-usecase/14/fig_4","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"},{"panel_id":"b","label":"multiple line chart"},{"panel_id":"c","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The figure compares atomic density profiles across depth for six or seven different elements under three experimental conditions: Ar (a), Kr (b), and Xe (c) ion exposure. In each case, the profiles illustrate how specific elements (e.g., H, O, C, Si, N) are distributed within and near the surface layers. The presence and penetration depth of each element vary depending on the ion species used, with notable differences in how deeply Ar, Kr, and Xe penetrate the material and affect the distribution of native and implanted species.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Depth [nm] | Atomic density [x 10<sup>23</sup>/cm<sup>3</sup>] |\\n|---|---|\\n|-6 | 0 |\\n|-5 | 0 |\\n|-4 | 0 |\\n|-3 | 0 |\\n|-2 | 0 |\\n|-1 | 0 |\\n| 0 | 0 |\\n| 1 | 0 |\\n| 2 | 0 |\\n| 3 | 0 |\\n| 4 | 0 |\\n| 5 | 0 |\\n| 6 | 0 |\"},{\"panel_id\":\"b\",\"text\":\"| Depth [nm] | Atomic density [x 10<sup>23</sup>/cm<sup>3</sup>] |\\n|---|---|\\n|-6 | 0 |\\n|-5 | 0 |\\n|-4 | 0 |\\n|-3 | 0 |\\n|-2 | 0 |\\n|-1 | 0 |\\n| 0 | 0 |\\n| 1 | 0 |\\n| 2 | 0 |\\n| 3 | 0 |\\n| 4 | 0 |\\n| 5 | 0 |\\n| 6 | 0 |\"},{\"panel_id\":\"c\",\"text\":\"| Depth [nm] | Atomic density [x 10<sup>23</sup>/cm<sup>3</sup>] |\\n|---|---|\\n|-6 | 0 |\\n|-5 | 0 |\\n|-4 | 0 |\\n|-3 | 0 |\\n|-2 | 0 |\\n|-1 | 0 |\\n| 0 | 0 |\\n| 1 | 0 |\\n| 2 | 0 |\\n| 3 | 0 |\\n| 4 | 0 |\\n| 5 | 0 |\\n| 6 | 0 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"The ion species used for exposure: Ar in (a), Kr in (b), and Xe in (c).\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"List\",\"answer\":\"Ar: shows deeper penetration, Kr: intermediate penetration , Xe: least penetration\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Yes/No\",\"answer\":\"No\"}]},{\"panel_id\":\"c\",\"items\":[{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The depth-dependent profiles indicate that heavier ions like Xe tend to stay near the surface, likely due to lower penetration power and stronger localised interaction, while lighter ions like Ar penetrate deeper into the material. This trend supports the hypothesis that ion mass influences the energy transfer efficiency and resulting implantation depth, which could be critical in tailoring surface modifications or etch profiles in atomic-scale processing\"}]}]","bbox":[{"panel_id":"a","x":11,"y":10,"width":439,"height":734},{"panel_id":"b","x":483,"y":10,"width":438,"height":734},{"panel_id":"c","x":953,"y":9,"width":430,"height":735}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/14/images/fig_4.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/14/images/fig_4.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/14/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/14/Molecular dynamics simulations of silicon nitride.pdf"],"main_category":"atomic-layer-etching","sub_category":"simulation-usecase","paper_id":"14","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:img_caption"},"width":1383,"height":744,"image_format":"jpeg","image_sha256":"5063d7e67b4d6983cdc2546da3ce3c0a96f9ccca31953e5e83f1e4713f2a57e3","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/test_atomic-layer-etching_simulation-usecase_18_fig_3.jpg","caption":"FIG. 3. Free energy profile for the SE1 (blue) and SL1 (orange) reactions of $\\mathsf{aHfO_2}$ from 0 to $1000\\mathrm{K}$ at the pressures given in the text. T1 is where the SL and SE reactions cross over for the 80/16F model, and T2 is where spontaneous etching is preferred.","id":"test/atomic-layer-etching/simulation-usecase/18/fig_3","sample_id":"atomic-layer-etching/simulation-usecase/18/fig_3","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The chart shows the change in Gibbs free energy (ΔG) per HfO₂ unit as a function of temperature. \\\"Preferred self-limiting\\\", \\\"Preferred etching\\\" and \\\"Purely etching\\\" regions are highlighted as well as T1 and T2 temperatures.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Temperature [K] | SE1 | SL1 |\\n| --- | --- | --- |\\n| 0 | -1.3 | -3.8 |\\n| 200 | -1.3 | -3.2 |\\n| 400 | -1.3 | -2.4 |\\n| 600 | -1.2 | -1.8 |\\n| 800 | -1.1 | -0.8 |\\n| 1000 | -1.0 | 0.0 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"T2 is the point at which spontaneous etching is favoured.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"For the SL1 reaction (orange line), the change in the Gibbs free energy values increases from -3.8 eV/HfO2 to nearly 0, as the temperature rises from 0 K to 1000 K.\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"As can be seen in the figure, self-limiting behavior is preferred from 0 to 718 K.\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"The temperature at which different self-limited conversion or spontaneous etching of aHfO2 occur can be observed. Therefore, it is possible to control the regime by adjusting the operating temperature.\"}]}]","bbox":[{"panel_id":"a","x":4,"y":6,"width":528,"height":478}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/18/images/fig_3.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/18/images/fig_3.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/18/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/18/Origin of enhanced thermal atomic layer etching of amorphous HfO2.pdf"],"main_category":"atomic-layer-etching","sub_category":"simulation-usecase","paper_id":"18","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:img_caption"},"width":534,"height":486,"image_format":"jpeg","image_sha256":"47f12f35c3e5c7f1ea123d1f249b05e41ed81fdaa4b780f57858d2a603ecb620","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/test_atomic-layer-etching_simulation-usecase_18_fig_4.jpg","caption":"FIG. 4. Free energy profiles of the spontaneous etching and self-limiting reactions for $\\mathrm{aHfO_2}$ . T1, T2, and T3 are the temperatures where the self-limiting and spontaneous etch reactions cross over for 80/16F, 100/20F, and 16O/32F models, respectively.","id":"test/atomic-layer-etching/simulation-usecase/18/fig_4","sample_id":"atomic-layer-etching/simulation-usecase/18/fig_4","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The chart shows the Gibbs free energy (ΔG) change per HfO₂ molecule as a function of temperature. Three lines represent different hydrogen fluoride (HF) concentrations (16 HF, 20 HF) and (32 HF). Three different temperatures are highlighted T1=718 K, T2=770 K and T3=1302 K.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Temperature [K] | 16HF | 20HF | 32HF |\\n| --- | --- | --- | --- |\\n| 0 | -3.8 | -3.8 | -5.2 |\\n| 400 | -2.2 | -2.2 | -4.0 |\\n| 800 | -0.8 | -1.0 | -2.6 |\\n| 1200 | 1.2 | 0.6 | -1.2 |\\n| 1600 | 3.0 | 2.2 | 0.4 |\\n| 2000 | 4.4 | 3.2 | 1.6 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Three different HF concentrations were studied: 16 HF, 20 HF and 32 HF.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"The Gibbs free energy increases with the temperature for all of the concentrations. All of them show the same behavior with the temperature.\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"As can be seen in the figure, the concentration of 16 HF shows higher Gibbs free energu values as a function of the temperature.\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"T1, T2 and T3 represent the temperatures where the self-limiting and spontaneous etch reaction cross over for different models with different concentrations of O to F.\"}]}]","bbox":[{"panel_id":"a","x":3,"y":6,"width":534,"height":464}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/18/images/fig_4.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/18/images/fig_4.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/18/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/18/Origin of enhanced thermal atomic layer etching of amorphous HfO2.pdf"],"main_category":"atomic-layer-etching","sub_category":"simulation-usecase","paper_id":"18","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:img_caption"},"width":539,"height":470,"image_format":"jpeg","image_sha256":"7b99d3219ba4ed527a1ef55adb7cd5bb37dacc502946fbb8caf7ecded6b1ff9f","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/test_atomic-layer-etching_simulation-usecase_25_figure_4.jpg","caption":"Figure 4. Contributions to the reaction FEP of the SE1 reaction of $\\mathrm{HfO}_2$ shown in Table 1. $RT\\ln (Q)$ in gold accounts for the partial pressures of the reactants and products, enthalpic contribution $(\\Delta W)$ is given in dark blue, entropy term $(T\\Delta S)$ in gray, reaction energy $(\\Delta E)$ in light blue, the sum of reaction energy and zero-point energy change $(\\Delta E + \\Delta \\mathrm{ZPE})$ in brown, and reaction free energy $(\\Delta G)$ in green.","id":"test/atomic-layer-etching/simulation-usecase/25/figure_4","sample_id":"atomic-layer-etching/simulation-usecase/25/figure_4","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"Multiple-line chart showing the temperature dependence of individual contributions to the reaction free energy for the SE1 etching reaction of HfO₂ with HF. The plot separates constant energetic terms (ΔE, ΔE+ΔZPE) from temperature-dependent contributions (TΔS, ΔW, RT ln Q), illustrating how entropy and pressure effects progressively increase ΔG with temperature while the reaction remains exergonic across the range shown.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Temperature [K] | ΔE [eV] | ΔE+ΔZPE [eV] | ΔG [eV] | TΔS [eV] | ΔW [eV] |\\n|---|---|---|---|---|---|\\n| 0 | ~ -0.9 | ~ -0.76 | ~ -0.76 | 0.0 | 0.0 |\\n| 200 | ~ -0.9 | ~ -0.76 | ~ -0.75 | ~ -0.05 | ~ -0.02 |\\n| 400 | ~ -0.9 | ~ -0.76 | ~ -0.68 | ~ -0.15 | ~ -0.06 |\\n| 600 | ~ -0.9 | ~ -0.76 | ~ -0.60 | ~ -0.25 | ~ -0.09 |\\n| 800 | ~ -0.9 | ~ -0.76 | ~ -0.52 | ~ -0.40 | ~ -0.11 |\\n| 1000 | ~ -0.9 | ~ -0.76 | ~ -0.44 | ~ -0.50 | ~ -0.13 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"The entropy term, TΔS.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Yes/No\",\"answer\":\"No\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"ΔE remains constant across the entire temperature range, reflecting its purely electronic nature. In contrast, TΔS becomes increasingly negative with temperature, contributing significantly to the upward shift in ΔG. This contrast highlights why entropy dominates the temperature dependence of reaction free energy.\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"ΔG remains negative throughout the plotted temperature range, indicating that the SE1 reaction is thermodynamically favorable under the conditions considered. Although the driving force decreases with temperature, the reaction does not become endergonic. This supports the feasibility of spontaneous etching at elevated temperatures.\"}]}]","bbox":[{"panel_id":"a","x":3,"y":9,"width":438,"height":413}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/25/images/figure_4.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/25/images/figure_4.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/25/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/25/Self-Limiting Temperature Window for Thermal Atomic Layer Etching.pdf"],"main_category":"atomic-layer-etching","sub_category":"simulation-usecase","paper_id":"25","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:img_caption"},"width":441,"height":423,"image_format":"jpeg","image_sha256":"19196ecdb540ed33a84b152021daddf91de4d291e62f1fcfdfb2d4c11ce0609e","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/test_atomic-layer-etching_simulation-usecase_25_figure_5.jpg","caption":"Figure 5. Free energy profiles for the SE1 (blue) and SL1 (orange) reactions of (a) $\\mathrm{HfO}_2$ and (b) $\\mathrm{ZrO}_2$ from 0 to $1000\\mathrm{K}$ at the pressures given in the text. $T_{1}$ is where the self-limiting and spontaneous etch reactions cross over for the 8O/16F model, and $T_{2}$ is where spontaneous etching is preferred.","id":"test/atomic-layer-etching/simulation-usecase/25/figure_5","sample_id":"atomic-layer-etching/simulation-usecase/25/figure_5","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"},{"panel_id":"b","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"Free-energy profiles for the SE1 (blue) and SL1 (orange) reactions of HfO₂ as functions of temperature. Below T₁ = 657 K, the self-limiting pathway is thermodynamically favored. Between T₁ and T₂ = 813 K, spontaneous etching becomes preferred. Above T₂, the self-limiting reaction is endergonic, and the system operates in a purely etching regime.\"},{\"panel_id\":\"b\",\"text\":\"Free-energy profiles for the SE1 and SL1 reactions of ZrO₂ showing analogous temperature-dependent behavior, with transitions at lower temperatures (T₁ = 534 K and T₂ = 770 K). Compared to HfO₂, ZrO₂ enters the preferred and purely etching regimes at lower temperatures, indicating reduced thermal stability of the self-limiting state.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Temperature [K] | ΔG (SE1 - Blue) [eV] | ΔG (SL1 - Orange) [eV] | Condition |\\n|---|---|---|---|\\n| 0 | ~ -0.75 | ~ -3.0 | Start |\\n| 657 | ~ -0.6 | ~ -0.6 | T1 (Crossover) |\\n| 813 | ~ -0.55 | 0.0 | T2 (Zero crossing) |\\n| 1000 | ~ -0.45 | ~ +0.7 | End |\"},{\"panel_id\":\"b\",\"text\":\"| Temperature [K] | ΔG (SE1 - Blue) [eV] | ΔG (SL1 - Orange) [eV] | Condition | \\n|---|---|---|---| \\n| 0 | ~ -1.0 | ~ -2.8 | Start | \\n| 534 | ~ -0.9 | ~ -0.9 | T1 (Crossover) | \\n| 770 | ~ -0.8 | 0.0 | T2 (Zero crossing) | \\n| 1000 | ~ -0.7 | ~ +0.8 | End |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Above 813 K, the free energy change (ΔG) for the self-limiting reaction becomes positive. This means the reaction becomes endergonic and is no longer thermodynamically spontaneous, indicating that a stable self-limiting layer cannot form under these conditions.\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"The self-limiting reaction (SL1).\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Temperatures below approximately 657 K are optimal, as the self-limiting reaction is thermodynamically favored while spontaneous etching is suppressed, allowing precise atomic layer removal.\"}]},{\"panel_id\":\"b\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Negative.\"}]}]","bbox":[{"panel_id":"a","x":0,"y":2,"width":438,"height":392},{"panel_id":"b","x":0,"y":407,"width":442,"height":389}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/25/images/figure_5.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/25/images/figure_5.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/25/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/25/Self-Limiting Temperature Window for Thermal Atomic Layer Etching.pdf"],"main_category":"atomic-layer-etching","sub_category":"simulation-usecase","paper_id":"25","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:img_caption"},"width":444,"height":798,"image_format":"jpeg","image_sha256":"6657ae071b8f3d67ad2f7121e41a3b8bb102159851c509db579070ec16ad1403","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/test_atomic-layer-etching_simulation-usecase_25_figure_6.jpg","caption":"Figure 6. Free energy profiles of the continuous etching and self-limiting reactions for $\\mathrm{HfO}_2$ . $T_{1}$ , $T_{2}$ , and $T_{3}$ are where the self-limiting and spontaneous etch reactions cross over for 16O/32F, 12O/24F, and 8O/16F models, respectively.","id":"test/atomic-layer-etching/simulation-usecase/25/figure_6","sample_id":"atomic-layer-etching/simulation-usecase/25/figure_6","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"Line chart showing free-energy Delta G profiles for continuous etching and self-limiting reactions of HfO2 under three HF surface coverages (16 HF, 24 HF, and 32 HF) as functions of temperature. The figure identifies three crossover temperatures T1 = 429 K (16O/32F), T2 = 618 K (12O/24F), and T3 = 657K (8O/16F)—at which the self-limiting and spontaneous etch pathways become thermodynamically equivalent.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Temperature [K] | Ref (Blue) [eV] | 32 HF (Yellow) [eV] | 24 HF (Gray) [eV] | 16 HF (Orange) [eV] | Note |\\n|---|---|---|---|---|---|\\n| 0 | ~ -0.8 | ~ -2.2 | ~ -2.8 | ~ -3.0 | Start |\\n| 429 | ~ -0.7 | ~ -0.7 | ~ -1.3 | ~ -1.5 | T1 (Crossover) |\\n| 618 | ~ -0.65 | 0.0 | ~ -0.6 | ~ -0.8 | T2 (Yellow Zero) |\\n| 657 | ~ -0.6 | ~ +0.2 | 0.0 | 0.0 | T3 (Gray/Orange Zero) |\\n| 1000 | ~ -0.5 | ~ +1.5 | ~ +0.8 | ~ +0.7 | End |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"The 32 HF surface model.\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Increasing HF coverage raises the free energy of the self-limiting pathway, causing it to intersect the continuous etching pathway at lower temperatures. This indicates that densely fluorinated surfaces destabilize the self-limiting configuration, narrowing the temperature window where controlled ALE can occur.\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The reference curve represents continuous etching, which is dominated by reaction energetics that vary weakly with surface coverage. In contrast, the self-limiting reactions depend strongly on surface fluorination, making their free energy more sensitive to temperature and coverage-dependent entropy contributions.\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"The 16 HF surface model is most suitable, as it maintains self-limiting behavior up to the highest temperature (T3 = 657K), offering a wider operational window before uncontrolled etching dominates.\"}]}]","bbox":[{"panel_id":"a","x":4,"y":6,"width":439,"height":384}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/25/images/figure_6.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/25/images/figure_6.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/25/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/25/Self-Limiting Temperature Window for Thermal Atomic Layer Etching.pdf"],"main_category":"atomic-layer-etching","sub_category":"simulation-usecase","paper_id":"25","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:img_caption"},"width":444,"height":392,"image_format":"jpeg","image_sha256":"4c22d1599c095a7b7888312679e84c87ecb1ef081c073260c5566ac3f64d93d5","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/test_atomic-layer-etching_simulation-usecase_25_figure_7.jpg","caption":"Figure 7. Free energy profiles of the continuous etching and self-limiting reactions for $\\mathrm{ZrO}_2$ . $T_{1}$ , $T_{2}$ , and $T_{3}$ are where the self-limiting and spontaneous etch reactions cross over for 16O/32F, 12O/24F, and 8O/16F models, respectively.","id":"test/atomic-layer-etching/simulation-usecase/25/figure_7","sample_id":"atomic-layer-etching/simulation-usecase/25/figure_7","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"Line chart showing free-energy (ΔG) profiles for continuous etching and self-limiting reactions of ZrO2 for three HF surface models (16 HF, 24 HF, and 32 HF) as functions of temperature. The figure marks three crossover temperatures—T1 = 273 K, T2 = 502 K, and T3 = 534 K—corresponding to the 16O/32F, 12O/24F, and 8O/16F models, respectively, where self-limiting and spontaneous etch pathways become thermodynamically equivalent.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Temperature [K] | Ref (Blue) [eV] | 32 HF (Yellow) [eV] | 24 HF (Gray) [eV] | 16 HF (Orange) [eV] | Note |\\n|---|---|---|---|---|---|\\n| 0 | ~ -1.0 | ~ -1.9 | ~ -2.7 | ~ -2.8 | Start |\\n| 273 | ~ -0.95 | ~ -0.95 | ~ -1.8 | ~ -1.9 | T1 (Crossover) |\\n| 502 | ~ -0.9 | 0.0 | ~ -1.0 | ~ -1.1 | T2 (Yellow Zero) |\\n| 534 | ~ -0.9 | ~ +0.2 | 0.0 | 0.0 | T3 (Gray/Orange Zero) |\\n| 1000 | ~ -0.7 | ~ +1.8 | ~ +0.9 | ~ +0.8 | End |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"The 32 HF surface model.\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Higher HF surface coverage raises the free energy of the self-limiting pathway, causing it to intersect the continuous etching pathway at lower temperatures. This indicates that heavily fluorinated surfaces destabilize the self-limiting configuration more rapidly as temperature increases.\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"1) 32 HF (smallest range)\\n2) 24 HF\\n3) 16 HF (largest range)\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The 16 HF surface model is most suitable, as it maintains a negative ΔG for self-limiting reactions up to the highest temperature (T3 = 534 K), providing the widest operational window merging stability with temperature tolerance.\"}]}]","bbox":[{"panel_id":"a","x":3,"y":3,"width":401,"height":353}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/25/images/figure_7.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/25/images/figure_7.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/25/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/25/Self-Limiting Temperature Window for Thermal Atomic Layer Etching.pdf"],"main_category":"atomic-layer-etching","sub_category":"simulation-usecase","paper_id":"25","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:img_caption"},"width":406,"height":358,"image_format":"jpeg","image_sha256":"01cbfd099c62b9657cf85c7bd0f25b2f38e18b8d4e11eb00211e883917afce90","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/test_atomic-layer-etching_simulation-usecase_28_fig12.jpg","caption":"FIG.12. G free energy profiles of CVE1, CVE2, and SL reactions vs temperature at a constant HF reactant pressure of $0.2$ Torr and a product pressure of $0.01$ Torr. SL reaction becomes endergonic at $110\\mathrm{K}$ .","id":"test/atomic-layer-etching/simulation-usecase/28/fig12","sample_id":"atomic-layer-etching/simulation-usecase/28/fig12","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"Thermodynamic profile comparing the Gibbs free energy (ΔG) of three etching mechanisms (CVE1, CVE2, SL) versus temperature. CVE1 and CVE2 remain exergonic (ΔG < 0) across the entire temperature range. The SL reaction transitions from exergonic to endergonic at 110 K, becoming thermodynamically unfavorable at higher temperatures.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Temperature (K) | ΔG (eV) - SL(Green) | ΔG (eV) - CVE2 (Blue) | ΔG (eV)- CVE1 (Red) |\\n|-----------------|---------------------|------------------------|-------------------------|\\n| 0 | -0.5 | -1.0 | -1.3 |\\n| 110 | 0.0 | -1.1 | -1.5 |\\n| 200 | 0.5 | -1.1 | -1.6 |\\n| 400 | 1.8 | -1.1 | -1.7 |\\n| 600 | 3.0 | -1.1 | -1.8 |\\n| 800 | 4.2 | -1.2 | -1.9 |\\n| 1000 | 5.2 | -1.2 | -2.0 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Above 110 K, the two mechanisms diverge significantly. The SL reaction becomes endergonic (ΔG > 0), meaning it is thermodynamically unfavorable and unlikely to proceed spontaneously. In contrast, the CVE1 reaction remains strongly exergonic (ΔG becoming more negative), indicating it is the thermodynamically preferred and dominant etching pathway at higher temperatures.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"110 K\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"List\",\"answer\":\"CVE1, CVE2, SL\"},{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"No, it remains exergonic (ΔG < 0) throughout the entire range.\"}]}]","bbox":[{"panel_id":"a","x":2,"y":0,"width":662,"height":606}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/28/images/fig12.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/28/images/fig12.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/28/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/28/Spontaneous etching of B2O3 by HF gas studied using infrared spectroscopy.pdf"],"main_category":"atomic-layer-etching","sub_category":"simulation-usecase","paper_id":"28","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:image_caption"},"width":666,"height":609,"image_format":"jpeg","image_sha256":"63b26bbe6337354a2ddff1d5b8ac610dcaade4fadd6c0b1939b57806d9cf6278","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/test_atomic-layer-etching_simulation-usecase_34_fig_12.jpg","caption":"Fig. 12. Graphical illustrations depicting the influence of various R2R controller algorithms on an environment that is introduced to a mild kinetic shift disturbance of 0.8 on the input parameters including the (a) HF flow rate, (b) TMA flow rate, (c) substrate velocity, which are computed from the output variable, (d) etching per cycle (EPC).","id":"test/atomic-layer-etching/simulation-usecase/34/fig_12","sample_id":"atomic-layer-etching/simulation-usecase/34/fig_12","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"},{"panel_id":"b","label":"multiple line chart"},{"panel_id":"c","label":"multiple line chart"},{"panel_id":"d","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The multiple line chart shows HF flow rate control over batch number. The EWMA and ANN predictions initially deviate slightly from the target but converge toward it as batches progress.\"},{\"panel_id\":\"b\",\"text\":\"The multiple line chart shows TMA flow rate adjustment across batches. Early deviations are reduced over time, with ANN tracking the target more closely than the EWMA curves.\"},{\"panel_id\":\"c\",\"text\":\"The multiple line chart shows substrate velocity over batch number. Initial oscillations decrease with batch number, and all methods converge toward the target velocity.\"},{\"panel_id\":\"d\",\"text\":\"The multiple line chart shows EPC over batch number. Both EWMA and ANN predictions converge toward the target EPC, with ANN exhibiting faster stabilization.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Batch number | Target | EWMA (λ=0.7) | EWMA (λ=0.3) | ANN |\\n|--------------|--------|--------------|--------------|-------|\\n| 1 | ~20 | ~20.2 | ~19.8 | ~20.1 |\\n| 10 | ~20 | ~20.1 | ~19.9 | ~20.0 |\\n| 20 | ~20 | ~20.0 | ~20.0 | ~20.0 |\"},{\"panel_id\":\"b\",\"text\":\"| Batch number | Target | EWMA (λ=0.7) | EWMA (λ=0.3) | ANN |\\n|--------------|--------|--------------|--------------|-------|\\n| 1 | ~45 | ~47 | ~43 | ~45 |\\n| 10 | ~45 | ~46 | ~44 | ~45 |\\n| 20 | ~45 | ~45 | ~45 | ~45 |\"},{\"panel_id\":\"c\",\"text\":\"| Batch number | Target | EWMA (λ=0.7) | EWMA (λ=0.3) | ANN |\\n|--------------|--------|--------------|--------------|-------|\\n| 1 | ~70 | ~75 | ~65 | ~70 |\\n| 10 | ~70 | ~72 | ~68 | ~70 |\\n| 20 | ~70 | ~70 | ~70 | ~70 |\"},{\"panel_id\":\"d\",\"text\":\"| Batch number | Target | EWMA (λ=0.7) | EWMA (λ=0.3) | ANN |\\n|--------------|--------|--------------|--------------|-------|\\n| 1 | ~0.04 | ~0.06 | ~0.03 | ~0.04 |\\n| 10 | ~0.04 | ~0.05 | ~0.04 | ~0.04 |\\n| 20 | ~0.04 | ~0.04 | ~0.04 | ~0.04 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Both methods show larger deviations from the target. At early batches methods progressively converge as batch number increases. The ANN curves generally stabilize faster and exhibit smaller oscillations compared to the EWMA curves.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"List\",\"answer\":\"HF flow rate, TMA flow rate, Substrate velocity, Etching per cycle (EPC\"}]},{\"panel_id\":\"d\",\"items\":[{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Yes/No\",\"answer\":\"Yes\"}]},{\"panel_id\":\"c\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Batch number.\"}]}]","bbox":[{"panel_id":"a","x":185,"y":0,"width":650,"height":40},{"panel_id":"b","x":514,"y":55,"width":497,"height":380},{"panel_id":"c","x":9,"y":457,"width":497,"height":384},{"panel_id":"d","x":516,"y":462,"width":495,"height":379}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/34/images/fig_12.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/34/images/fig_12.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/34/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/34/Machine learning-based run-to-run control of a spatial thermal.pdf"],"main_category":"atomic-layer-etching","sub_category":"simulation-usecase","paper_id":"34","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:image_caption"},"width":1011,"height":841,"image_format":"jpeg","image_sha256":"a49a4d8333313c0029c0c343945861c33aa58ae743486a5394094766f7404b88","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/test_atomic-layer-etching_simulation-usecase_34_fig_13.jpg","caption":"Fig. 13. Graphical illustrations depicting the influence of various R2R controller algorithms on an environment that is introduced to a severe kinetic shift disturbance of 0.6 on the input parameters including the (a) HF flow rate, (b) TMA flow rate, and (c) substrate velocity, which are computed from the output variable (d) etching per cycle (EPC).","id":"test/atomic-layer-etching/simulation-usecase/34/fig_13","sample_id":"atomic-layer-etching/simulation-usecase/34/fig_13","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"},{"panel_id":"b","label":"multiple line chart"},{"panel_id":"c","label":"multiple line chart"},{"panel_id":"d","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The multiple line chart shows HF flow rate control over batch number. EWMA initially deviates from the target, while ANN tracks the target more closely and stabilizes faster.\"},{\"panel_id\":\"b\",\"text\":\"The multiple line chart shows TMA flow rate over batch number. Early fluctuations decrease with batch number, and the ANN method maintains values closest to the target.\"},{\"panel_id\":\"c\",\"text\":\"The multiple line chart shows substrate velocity over batch number. Initial oscillations are damped over time, with ANN converging to the target more quickly than EWMA.\"},{\"panel_id\":\"d\",\"text\":\"The multiple line chart shows EPC over batch number. All methods converge toward the target EPC, with ANN exhibiting the smallest deviation during stabilization.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Batch number | Target | EWMA (λ=0.7) | EWMA (λ=0.3) | ANN |\\n|--------------|--------|--------------|--------------|-------|\\n| 1 | ~20 | ~21.0 | ~20.5 | ~20.0 |\\n| 10 | ~20 | ~20.3 | ~20.1 | ~20.0 |\\n| 20 | ~20 | ~20.0 | ~20.0 | ~20.0 |\"},{\"panel_id\":\"b\",\"text\":\"| Batch number | Target | EWMA (λ=0.7) | EWMA (λ=0.3) | ANN |\\n|--------------|--------|--------------|--------------|-------|\\n| 1 | ~45 | ~47 | ~44 | ~45 |\\n| 10 | ~45 | ~46 | ~44.5 | ~45 |\\n| 20 | ~45 | ~45 | ~45 | ~45 |\"},{\"panel_id\":\"c\",\"text\":\"| Batch number | Target | EWMA (λ=0.7) | EWMA (λ=0.3) | ANN |\\n|--------------|--------|--------------|--------------|-------|\\n| 1 | ~70 | ~75 | ~65 | ~70 |\\n| 10 | ~70 | ~72 | ~68 | ~70 |\\n| 20 | ~70 | ~70 | ~70 | ~70 |\"},{\"panel_id\":\"d\",\"text\":\"| Batch number | Target | EWMA (λ=0.7) | EWMA (λ=0.3) | ANN |\\n|--------------|--------|--------------|--------------|-------|\\n| 1 | ~0.35 | ~0.38 | ~0.33 | ~0.35 |\\n| 10 | ~0.35 | ~0.36 | ~0.34 | ~0.35 |\\n| 20 | ~0.35 | ~0.35 | ~0.35 | ~0.35 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"Both methods reduce deviations over batch number, but ANN consistently tracks the target more closely and stabilizes faster than EWMA, which shows larger initial oscillations.\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Yes/No\",\"answer\":\"Yes\"}]},{\"panel_id\":\"c\",\"items\":[{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Batch Number.\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"List\",\"answer\":\"HF flow rate, TMA flow rate, Substrate velocity, Etching per cycle\"}]}]","bbox":[{"panel_id":"a","x":190,"y":0,"width":648,"height":39},{"panel_id":"b","x":518,"y":54,"width":502,"height":380},{"panel_id":"c","x":10,"y":457,"width":499,"height":385},{"panel_id":"d","x":519,"y":461,"width":501,"height":381}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/34/images/fig_13.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/34/images/fig_13.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/34/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/34/Machine learning-based run-to-run control of a spatial thermal.pdf"],"main_category":"atomic-layer-etching","sub_category":"simulation-usecase","paper_id":"34","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:image_caption"},"width":1020,"height":842,"image_format":"jpeg","image_sha256":"54221088f5425cb59ddcf8de94ecf56065b92d41f3dea8655b745059b70e90bf","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/test_atomic-layer-etching_simulation-usecase_34_fig_15.jpg","caption":"Fig. 15. Graphical illustrations depicting the influence of various R2R controller algorithms on an environment that is introduced to a kinetic drift disturbance on the input parameters including the (a) HF flow rate, (b) TMA flow rate, and (c) substrate velocity, which are computed from the output variable (d) etching per cycle (EPC).","id":"test/atomic-layer-etching/simulation-usecase/34/fig_15","sample_id":"atomic-layer-etching/simulation-usecase/34/fig_15","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"},{"panel_id":"b","label":"multiple line chart"},{"panel_id":"c","label":"multiple line chart"},{"panel_id":"d","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The multiple line chart shows HF flow rate control over batch number. Without control, the flow rate drifts upward, while EWMA partially corrects the drift and ANN maintains the flow rate close to the target.\"},{\"panel_id\":\"b\",\"text\":\"The multiple line chart shows TMA flow rate over batch number. The no-control case increases steadily, whereas the ANN tracks the target closely and EWMA reduces but does not eliminate the drift.\"},{\"panel_id\":\"c\",\"text\":\"The multiple line chart shows substrate velocity over batch number. Velocity decreases significantly without control, while EWMA mitigates the drop and ANN maintains values near the target.\"},{\"panel_id\":\"d\",\"text\":\"The multiple line chart shows EPC over batch number. EPC declines strongly in the no-control case, while both EWMA and ANN stabilize EPC near the target, with ANN showing the smallest deviation.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Batch number | Target | EWMA | ANN | No control |\\n|--------------|--------|------|------|------------|\\n| 1 | ~20 | ~21 | ~20 | ~22 |\\n| 10 | ~20 | ~22 | ~20 | ~24 |\\n| 20 | ~20 | ~23 | ~20 | ~25 |\"},{\"panel_id\":\"b\",\"text\":\"| Batch number | Target | EWMA | ANN | No control |\\n|--------------|--------|------|------|------------|\\n| 1 | ~45 | ~46 | ~45 | ~48 |\\n| 10 | ~45 | ~47 | ~45 | ~50 |\\n| 20 | ~45 | ~48 | ~45 | ~52 |\"},{\"panel_id\":\"c\",\"text\":\"| Batch number | Target | EWMA | ANN | No control |\\n|--------------|--------|------|------|------------|\\n| 1 | ~70 | ~72 | ~70 | ~75 |\\n| 10 | ~70 | ~68 | ~70 | ~65 |\\n| 20 | ~70 | ~65 | ~70 | ~55 |\"},{\"panel_id\":\"d\",\"text\":\"| Batch number | Target | EWMA | ANN | No control |\\n|--------------|--------|------|------|------------|\\n| 1 | ~0.35 | ~0.36| ~0.35| ~0.34 |\\n| 10 | ~0.35 | ~0.33| ~0.35| ~0.28 |\\n| 20 | ~0.35 | ~0.32| ~0.35| ~0.20 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"In all panels, the no-control case shows significant drift away from the target over batch number. EWMA reduces the drift but still deviates over time. The ANN consistently maintains the variables closest to their target values.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Batch number.\"}]},{\"panel_id\":\"d\",\"items\":[{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"List\",\"answer\":\"HF flow rate, TMA flow rate, Substrate velocity, Etching per cycle (EPC)\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Yes/No\",\"answer\":\"Yes\"}]}]","bbox":[{"panel_id":"a","x":88,"y":0,"width":847,"height":39},{"panel_id":"b","x":516,"y":54,"width":495,"height":379},{"panel_id":"c","x":9,"y":456,"width":498,"height":383},{"panel_id":"d","x":516,"y":460,"width":495,"height":379}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/34/images/fig_15.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/34/images/fig_15.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/34/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/34/Machine learning-based run-to-run control of a spatial thermal.pdf"],"main_category":"atomic-layer-etching","sub_category":"simulation-usecase","paper_id":"34","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:image_caption"},"width":1011,"height":839,"image_format":"jpeg","image_sha256":"1ffcd13765d1dcfa5fe4f3dc9e3401a76450bcc7875873d329ce424c6ba1a6d6","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/test_atomic-layer-etching_simulation-usecase_7_fig_4.jpg","caption":"FIG. 4. Free energy profiles of the overall etch reactions given in Table I. Solid lines correspond to the etch of an adatom and dotted lines correspond to the etch of a surface atom.","id":"test/atomic-layer-etching/simulation-usecase/7/fig_4","sample_id":"atomic-layer-etching/simulation-usecase/7/fig_4","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"The figure presents a line chart that plots the Gibbs free energy change (ΔG) per cobalt atom as a function of temperature (T) for eight surface reactions involving different precursors: propene, butyne, silane, and trimethylsilane (TMS). Each precursor is represented under two surface states, adsorbed (-a) and saturated (-s). The chart reveals that ΔG increases linearly with temperature for all cases, with the magnitude of change varying among the precursors. Propene and TMS show the most favorable ΔG values (more negative) across the full temperature range, while butyne shows the least favorable thermodynamics. The comparison between adsorbed and saturated cases demonstrates consistent trends across species.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| T [K] | Propene-a | Butyne-a | Silane-a | TMS-a | Propene-s | Butyne-s | Silane-s | TMS-s |\\n|---|---|---|---|---|---|---|---|---|\\n| 0 | -9.5 | -2.5 | -6.5 | -8.5 | -9.5 | -2.5 | -6.5 | -8.5 |\\n| 250 | -7.5 | -1.5 | -4.5 | -6.5 | -7.5 | -1.5 | -4.5 | -6.5 |\\n| 500 | -5.5 | -0.5 | -2.5 | -4.5 | -5.5 | -0.5 | -2.5 | -4.5 |\\n| 750 | -3.5 | 0.5 | -0.5 | -2.5 | -3.5 | 0.5 | -0.5 | -2.5 |\\n| 1000 | -1.5 | 1.5 | 1.5 | 0.5 | -1.5 | 1.5 | 1.5 | 0.5 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The ΔG values for both propene and butyne increase with temperature, indicating that the thermodynamic favorability of surface binding or reaction decreases at higher temperatures. Propene shows significantly more negative ΔG values than butyne across the entire temperature range, suggesting it is more thermodynamically favorable for surface reactions. The difference between the adsorbed (a) and saturated (s) states for both molecules is minimal, implying that the surface termination state has a limited effect on the general thermodynamic trend. This suggests that propene may be a more suitable candidate for temperature-sensitive surface chemistry applications like selective etching or deposition.\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Yes/No\",\"answer\":\"Yes\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"a -2.5 eV/Co\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"List\",\"answer\":\"Propene-a, Propene-s, TMS-a, TMS-s\"}]}]","bbox":[{"panel_id":"a","x":0,"y":0,"width":653,"height":605}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/7/images/fig_4.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/7/images/fig_4.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/7/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/7/In silico design of a thermal atomic layer etch process of cobalt.pdf"],"main_category":"atomic-layer-etching","sub_category":"simulation-usecase","paper_id":"7","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:img_caption"},"width":653,"height":605,"image_format":"jpeg","image_sha256":"d1f971c226560860424bb46b213f29d205db5f0dec0d2cb78d582e2ca5c976d6","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/test_atomic-layer-etching_simulation-usecase_7_fig_6.jpg","caption":"FIG. 6. (a) Free energy profiles of the individual reaction steps for the ALE process with H desorption after propene adsorption (in dotted lines) or no desorption of H after propene adsorption (in solid lines) given in Fig. 5. (b) The pressure-temperature process window of the complete ALE cycle (H desorbed) computed by considering different $\\ln (Q)$ values for each temperature. The red line indicated the CO decomposition temperature on the Co surface. The region marked by the black rectangle gives the Co ALE process window when propene and CO are used as reactants.","id":"test/atomic-layer-etching/simulation-usecase/7/fig_6","sample_id":"atomic-layer-etching/simulation-usecase/7/fig_6","subset":"multiple-line-chart","split":"test","classification":[{"panel_id":"a","label":"multiple line chart"},{"panel_id":"b","label":"contour heatmap"}],"summarization":"[{\"panel_id\":\"a\",\"text\":\"Figure (a) presents a multi-line chart showing the Gibbs free energy (ΔG) as a function of temperature for different surface reaction steps involved in propene chemistry on cobalt. These steps include propene adsorption, its dissociation, and byproduct formation such as H₂ and CO₂. The slope and relative positioning of the lines indicate which reactions become more or less thermodynamically favorable with increasing temperature.\"},{\"panel_id\":\"b\",\"text\":\"Figure (b) is a heatmap that maps the equilibrium constant (lnQ) over a temperature-pressure space. The diagram includes a shaded process window indicating practical temperature and pressure ranges where the targeted surface reactions may be viable, providing a framework for selecting operating conditions.\"}]","data_extraction":"[{\"panel_id\":\"a\",\"text\":\"| Temperature [K] | ΔG [eV / Co] |\\n|---|---|\\n| 0 | -6 |\\n| 200 | -5 |\\n| 400 | -4 |\\n| 600 | -3 |\\n| 800 | -2 |\\n| 1000 | -1 |\"},{\"panel_id\":\"b\",\"text\":\"| Temperature [K] | Equilibrium constant: lnQ |\\n|---|---|\\n| 0 | -75 |\\n| 200 | -50 |\\n| 400 | -25 |\\n| 600 | 0 |\\n| 800 | 25 |\\n| 1000 | 50 |\"}]","vqa":"[{\"panel_id\":\"a\",\"items\":[{\"question_type\":\"Comparative/Trend\",\"questions\":\"\",\"answer_type\":\"Paragraph\",\"answer\":\"The ΔG trends in figure (a) show which reaction steps become more or less favorable as temperature increases. For instance, propene adsorption has a negative ΔG at low temperatures, indicating thermodynamic favourability, while certain desorption reactions (e.g., CO₂ release) become favorable only at higher temperatures. Figure (b) complements this by showing the process window where lnQ indicates strong driving forces for reactions (lnQ ≪ 0) or equilibrium conditions (lnQ ≈ 0). By aligning these trends, one can identify temperature-pressure zones where specific reaction pathways are both thermodynamically favorable and kinetically accessible. This approach enables the rational design of conditions that favor desired surface terminations or product release while suppressing unwanted side reactions\"},{\"question_type\":\"Process-Oriented\",\"questions\":\"\",\"answer_type\":\"Yes/No\",\"answer\":\"Yes\"},{\"question_type\":\"Structure-Property\",\"questions\":\"\",\"answer_type\":\"Factoid\",\"answer\":\"Around 750 K.\"},{\"question_type\":\"Application/Performance\",\"questions\":\"\",\"answer_type\":\"List\",\"answer\":\"Propene adsorption, dissociative H₂ adsorption, CO₂ adsorption, surface reaction of volatiles.\"}]}]","bbox":[{"panel_id":"a","x":6,"y":36,"width":602,"height":580},{"panel_id":"b","x":640,"y":17,"width":602,"height":582}],"source":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/7/images/fig_6.jpg","provenance":{"source_annotation":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/7/images/fig_6.json","source_content":"icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/7/content.json","source_pdf":["icdar2026-competition-data/test/gold_standard_test_set/atomic-layer-etching/simulation-usecase/7/In silico design of a thermal atomic layer etch process of cobalt.pdf"],"main_category":"atomic-layer-etching","sub_category":"simulation-usecase","paper_id":"7","first_classification_panel_id":"a","first_classification_label":"multiple line chart","caption_source":"content.json:img_caption"},"width":1356,"height":620,"image_format":"jpeg","image_sha256":"de52a3b4c8c82f64e86a3a1d1df5ed72daad26c85f904e0f8db546f2cd79baab","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"}
|