├── Entity Resolution.ipynb ├── README.md ├── data ├── Keywords.xlsx ├── Property.xlsx ├── TrainSet1.0.json ├── TrainSet2.0.json └── TrainSet_bc.json └── inference_KG.py /Entity Resolution.ipynb: -------------------------------------------------------------------------------- 1 | { 2 | "cells": [ 3 | { 4 | "cell_type": "code", 5 | "execution_count": 1, 6 | "id": "90333284-3e0c-4d9a-b941-94f5312ce869", 7 | "metadata": {}, 8 | "outputs": [], 9 | "source": [ 10 | "from chemdataextractor import Document\n", 11 | "import re\n", 12 | "import json\n", 13 | "import csv\n", 14 | "import pandas as pd\n", 15 | "from gensim.models import Word2Vec\n", 16 | "from mat2vec.processing import MaterialsTextProcessor\n", 17 | "import numpy as np\n", 18 | "import copy" 19 | ] 20 | }, 21 | { 22 | "cell_type": "markdown", 23 | "id": "c8348891-0ef0-40f5-873d-7ac8c53f917c", 24 | "metadata": {}, 25 | "source": [ 26 | "# Auxiliary unction" 27 | ] 28 | }, 29 | { 30 | "cell_type": "code", 31 | "execution_count": 2, 32 | "id": "cdd60b07-17c6-4c9d-9284-f8d3bd95e507", 33 | "metadata": {}, 34 | "outputs": [], 35 | "source": [ 36 | "model_path = 'mat2vec/training/models/pretrained_embeddings'\n", 37 | "w2v_model = Word2Vec.load(model_path)\n", 38 | "\n", 39 | "def calculate_phrase_vector(model, phrase):\n", 40 | " '''\n", 41 | " Calculate the vector representation of a given phrase.\n", 42 | " For each word in the phrase, if the word exists in the model, obtain its vector and calculate the mean of all these vectors.\n", 43 | " '''\n", 44 | " words = [model.wv[word] for word in phrase.split() if word in model.wv]\n", 45 | " if words:\n", 46 | " return np.mean(words, axis=0)\n", 47 | " return None\n", 48 | "\n", 49 | "def calculate_similarity(element1, element2):\n", 50 | " '''\n", 51 | " Calculate and return the similarity between two elements.\n", 52 | " Output: The similarity score\n", 53 | " '''\n", 54 | " vector1 = calculate_phrase_vector(w2v_model, element1)\n", 55 | " vector2 = calculate_phrase_vector(w2v_model, element2)\n", 56 | " sim_score = 0\n", 57 | " if vector1 is not None and vector2 is not None:\n", 58 | " sim_score = np.dot(vector1, vector2) / (np.linalg.norm(vector1) * np.linalg.norm(vector2))\n", 59 | " else:\n", 60 | " if element1 == element2:\n", 61 | " sim_score = 1\n", 62 | "\n", 63 | " return sim_score\n", 64 | "\n", 65 | "def extract_chemical_info(text):\n", 66 | " '''\n", 67 | " Extracting NFA entities using ChemdataExtractor.\n", 68 | " '''\n", 69 | " doc = Document(text)\n", 70 | " materials = [c.text for c in doc.cems]\n", 71 | " abbreviations = {}\n", 72 | " for item in doc.abbreviation_definitions:\n", 73 | " abbr, definition = item[0], item[1] if len(item) >= 2 else ('', '')\n", 74 | " if isinstance(abbr, list):\n", 75 | " abbr = abbr[0] if abbr else ''\n", 76 | " if isinstance(definition, list):\n", 77 | " definition = ' '.join(definition)\n", 78 | " if abbr:\n", 79 | " abbreviations[abbr] = definition\n", 80 | "\n", 81 | " return list(set(materials)), abbreviations\n", 82 | "\n", 83 | "def rule_based_extension(text, materials, special_chars=['–', ':', '/', '@']):\n", 84 | " '''\n", 85 | " Rule-base enhancement of ChemdataExtractor's ability to extract NFA entities\n", 86 | " '''\n", 87 | " extended_materials = []\n", 88 | " \n", 89 | " for material in materials:\n", 90 | " positions = [m.start() for m in re.finditer(re.escape(material), text)]\n", 91 | " for pos in positions:\n", 92 | " left_index = pos - 1\n", 93 | " right_index = pos + len(material)\n", 94 | " left_char = text[left_index] if left_index >= 0 else ''\n", 95 | " right_char = text[right_index] if right_index < len(text) else ''\n", 96 | " \n", 97 | " if left_char in special_chars or right_char in special_chars:\n", 98 | " extended_left = re.search(r'\\s[^ ]*$', text[:left_index + 1])\n", 99 | " extended_right = re.search(r'^[^ ]*\\s', text[right_index:])\n", 100 | " \n", 101 | " left_boundary = extended_left.start() + 1 if extended_left else left_index\n", 102 | " right_boundary = right_index + extended_right.end() - 1 if extended_right else right_index\n", 103 | "\n", 104 | " extended_material = text[left_boundary:right_boundary]\n", 105 | " extended_materials.append(extended_material)\n", 106 | " else:\n", 107 | " extended_materials.append(material)\n", 108 | " \n", 109 | " return list(set(extended_materials))\n", 110 | "\n", 111 | "def correct_entities(entity, st_list):\n", 112 | " '''\n", 113 | " Compare the entity with all the elements in the stand list and delete the worng entity.\n", 114 | " Output: True if the entity is similar to any element in standard list else False\n", 115 | " '''\n", 116 | " sup_vocab = ['TiO2', 'CdS']\n", 117 | " for i in sup_vocab:\n", 118 | " if i in entity:\n", 119 | " return True\n", 120 | " if entity != '-':\n", 121 | " for i in range(len(st_list)):\n", 122 | " if calculate_similarity(entity, st_list[i]) >= 0.8:\n", 123 | " return True\n", 124 | "\n", 125 | " return False\n", 126 | "\n", 127 | "\n", 128 | "# Sample:\n", 129 | "# text_paragraph = \"ethylenediamine (EDA). Organic/nanostructured Si hybrid solar cells have achieved high power conversion efficiency (PCE) and shortcircuit current density (JSC), due to the excellent light harvesting ability of nanostructured Si. However, the separation of carriers is mainly affected by the low work function of PH1000-type PEDOT:PSS, resulting in poor open circuit voltage (VOC). In this work, an efficient and stable MoO3-doped copper phthalocyanine-3,4 ',4 '',4 '''- tetra-sulfonated acid tetra sodium salt (TS-CuPc) organic small molecule film was introduced between the heterojunction interfaces. The photogenerated carrier separation is promoted by the enhanced built-in potential (V-bi) owing to the high work function of TS-CuPc:MoO3, which also suppresses the carrier recombination at the surface of nanostructured Si. As a result, for PEDOT:PSS/nanostructured Si photovoltaic devices, the PCE was greatly improved from 14.15% to 16.09%. The excellent charge separation properties and interface passivation effect enable efficient devices to demonstrate the vast potential of this new type of interlayer in photovoltaic applications.\"\n", 130 | "# materials, abbreviations = extract_chemical_info(text_paragraph)\n", 131 | "# materials = rule_based_extension(text_paragraph, materials)\n", 132 | "# print(materials,'\\n', abbreviations)" 133 | ] 134 | }, 135 | { 136 | "cell_type": "markdown", 137 | "id": "6a6d5618-0fea-4df4-ad2f-e2a207637e4c", 138 | "metadata": {}, 139 | "source": [ 140 | "# NFA Label Standardization and Correct" 141 | ] 142 | }, 143 | { 144 | "cell_type": "code", 145 | "execution_count": 3, 146 | "id": "a4b92217-b361-48cf-8839-a9c41c705aec", 147 | "metadata": {}, 148 | "outputs": [ 149 | { 150 | "data": { 151 | "text/plain": [ 152 | "[[{'Name': [],\n", 153 | " 'Formula': ['BaCe1 − xNdxO3 − x 2'],\n", 154 | " 'Acronym': [],\n", 155 | " 'Descriptor': ['polymetric'],\n", 156 | " 'Structure/Phase': ['perovskites'],\n", 157 | " 'Application': [{'neodymium doping': []}],\n", 158 | " 'Property': [\"'resolution': [1000]\",\n", 159 | " \"'spontaneous strain': [0.05]\",\n", 160 | " \"'opaque': []\",\n", 161 | " \"'transparent': []\",\n", 162 | " \"'reflectance': [1000]\"],\n", 163 | " 'Synthesis': [],\n", 164 | " 'Characterization': ['neutron powder diffraction']},\n", 165 | " ''],\n", 166 | " [{'Name': ['NaFeTi2F11'],\n", 167 | " 'Formula': [],\n", 168 | " 'Acronym': [],\n", 169 | " 'Descriptor': ['cathode'],\n", 170 | " 'Structure/Phase': [],\n", 171 | " 'Application': [{'sodium-based batteries': [\"'intercalation voltage': [4.0] (property)\"]}],\n", 172 | " 'Property': [],\n", 173 | " 'Synthesis': [],\n", 174 | " 'Characterization': []},\n", 175 | " ''],\n", 176 | " [''],\n", 177 | " [''],\n", 178 | " [''],\n", 179 | " [{'Name': ['Battery manganese dioxide'],\n", 180 | " 'Formula': [],\n", 181 | " 'Acronym': [],\n", 182 | " 'Descriptor': ['electrochemically deposited', 'doped'],\n", 183 | " 'Structure/Phase': [],\n", 184 | " 'Application': [{'batteries': []}],\n", 185 | " 'Property': [\"'etymology': []\"],\n", 186 | " 'Synthesis': ['electrochemical processes'],\n", 187 | " 'Characterization': []},\n", 188 | " ''],\n", 189 | " [{'Name': [],\n", 190 | " 'Formula': ['La1-xCexCoO3'],\n", 191 | " 'Acronym': [],\n", 192 | " 'Descriptor': ['catalyst'],\n", 193 | " 'Structure/Phase': [],\n", 194 | " 'Application': [{'zinc oxide': []}],\n", 195 | " 'Property': [\"'active site': []\", \"'particle size': []\", \"'shape': []\"],\n", 196 | " 'Synthesis': ['X-ray photoelectron spectroscopy']},\n", 197 | " ''],\n", 198 | " [''],\n", 199 | " [{'Name': ['Nafion'],\n", 200 | " 'Formula': [],\n", 201 | " 'Acronym': ['SPCE'],\n", 202 | " 'Descriptor': ['emission'],\n", 203 | " 'Structure/Phase': ['film'],\n", 204 | " 'Application': [{'zinc oxide': [], 'magnetite': []}],\n", 205 | " 'Property': [\"'thickness': [0.05–0.5]\", \"'efficiency': [0.05–0.5]\"],\n", 206 | " 'Synthesis': ['electrochemiluminescence'],\n", 207 | " 'Characterization': []},\n", 208 | " ''],\n", 209 | " [''],\n", 210 | " [''],\n", 211 | " [{'Name': ['CdO'],\n", 212 | " 'Formula': [],\n", 213 | " 'Acronym': [],\n", 214 | " 'Descriptor': ['ceramics'],\n", 215 | " 'Structure/Phase': [],\n", 216 | " 'Application': [{'z-type oxide': [\"'thermoelectric figure of merit': [0.5] (property)\"]}],\n", 217 | " 'Property': [\"'electrical conductivity': []\",\n", 218 | " \"'lattice thermal conductivity': []\",\n", 219 | " \"'power factor': []\"],\n", 220 | " 'Synthesis': ['substituting'],\n", 221 | " 'Characterization': []},\n", 222 | " ''],\n", 223 | " [''],\n", 224 | " [{'Name': ['polydimethylsiloxane'],\n", 225 | " 'Formula': [],\n", 226 | " 'Acronym': ['PDMS'],\n", 227 | " 'Descriptor': ['elastomer surface functionalization'],\n", 228 | " 'Structure/Phase': [],\n", 229 | " 'Application': [{'silicone gels': []}],\n", 230 | " 'Property': [],\n", 231 | " 'Synthesis': ['thiol-ene click chemistry'],\n", 232 | " 'Characterization': ['X-ray photoelectron spectroscopy',\n", 233 | " 'transmission electron microscope',\n", 234 | " 'fluorescence labeling',\n", 235 | " 'surface cracking',\n", 236 | " 'hydrophobic recovery']},\n", 237 | " ''],\n", 238 | " [''],\n", 239 | " [''],\n", 240 | " [''],\n", 241 | " [''],\n", 242 | " [{'Name': [],\n", 243 | " 'Formula': ['Cu3Pd'],\n", 244 | " 'Acronym': [],\n", 245 | " 'Descriptor': ['polar'],\n", 246 | " 'Structure/Phase': ['L12'],\n", 247 | " 'Application': [{}],\n", 248 | " 'Property': [\"'ordering': []\"],\n", 249 | " 'Synthesis': ['annealing'],\n", 250 | " 'Characterization': ['TEM']},\n", 251 | " ''],\n", 252 | " [{'Name': [],\n", 253 | " 'Formula': ['Cu2S-SrLnCuS3'],\n", 254 | " 'Acronym': [],\n", 255 | " 'Descriptor': ['solution'],\n", 256 | " 'Structure/Phase': ['orthorhombic structure'],\n", 257 | " 'Application': [{'SrLnCuS3': [\"'melting temperature': [1050] (property)\"]}],\n", 258 | " 'Property': [\"'unit cell parameters': []\",\n", 259 | " \"'volume': []\",\n", 260 | " \"'monotony': [0.001]\",\n", 261 | " \"'periodicity': [0.002]\"],\n", 262 | " 'Synthesis': [],\n", 263 | " 'Characterization': []},\n", 264 | " ''],\n", 265 | " [''],\n", 266 | " [{'Name': [],\n", 267 | " 'Formula': ['7Li'],\n", 268 | " 'Acronym': [],\n", 269 | " 'Descriptor': ['NMR'],\n", 270 | " 'Structure/Phase': ['carbon'],\n", 271 | " 'Application': [{'lithium-ion battery': []}],\n", 272 | " 'Property': [],\n", 273 | " 'Synthesis': [],\n", 274 | " 'Characterization': ['NMR']},\n", 275 | " ''],\n", 276 | " [{'Name': [],\n", 277 | " 'Formula': ['LiCoO2'],\n", 278 | " 'Acronym': [],\n", 279 | " 'Descriptor': ['electrode'],\n", 280 | " 'Structure/Phase': [],\n", 281 | " 'Application': [{'lithium-ion battery': [\"'charge transfer kinetics': [] (property)\",\n", 282 | " \"'resistance': [] (property)\"]}],\n", 283 | " 'Property': [],\n", 284 | " 'Synthesis': [],\n", 285 | " 'Characterization': []},\n", 286 | " {'Name': [],\n", 287 | " 'Formula': ['LiFePO4'],\n", 288 | " 'Acronym': [],\n", 289 | " 'Descriptor': ['electrode'],\n", 290 | " 'Structure/Phase': [],\n", 291 | " 'Application': [{'lithium-ion battery': [\"'charge transfer kinetics': [] (property)\",\n", 292 | " \"'resistance': [] (property)\"]}],\n", 293 | " 'Property': [],\n", 294 | " 'Synthesis': [],\n", 295 | " 'Characterization': []},\n", 296 | " {'Name': [],\n", 297 | " 'Formula': ['LiMn2O4'],\n", 298 | " 'Acronym': [],\n", 299 | " 'Descriptor': ['electrode'],\n", 300 | " 'Structure/Phase': [],\n", 301 | " 'Application': [{'lithium-ion battery': [\"'charge transfer kinetics': [] (property)\",\n", 302 | " \"'resistance': [] (property)\"]}],\n", 303 | " 'Property': [],\n", 304 | " 'Synthesis': [],\n", 305 | " 'Characterization': []},\n", 306 | " ''],\n", 307 | " [{'Name': [],\n", 308 | " 'Formula': ['LiNi0.5Mn1.5O4'],\n", 309 | " 'Acronym': [],\n", 310 | " 'Descriptor': ['cathodes'],\n", 311 | " 'Structure/Phase': ['nanoparticle', 'nanorod'],\n", 312 | " 'Application': [{'lithium-ion battery': [\"'energy density': [1.6] (property)\",\n", 313 | " \"'capacity': [121] (property)\"]}],\n", 314 | " 'Property': [\"'capacity': [100] (property)\"],\n", 315 | " 'Synthesis': ['hollow'],\n", 316 | " 'Characterization': []},\n", 317 | " ''],\n", 318 | " [{'Name': ['yttrium barium copper oxide'],\n", 319 | " 'Formula': [],\n", 320 | " 'Acronym': [],\n", 321 | " 'Descriptor': ['superconducting material'],\n", 322 | " 'Structure/Phase': ['fibers'],\n", 323 | " 'Application': [{'cooling': []}],\n", 324 | " 'Property': [\"'core diameter': [100–200]\",\n", 325 | " \"'core strength': [100–200]\",\n", 326 | " \"'diameter': [300–600]\"],\n", 327 | " 'Synthesis': ['melt-draw'],\n", 328 | " 'Characterization': []},\n", 329 | " ''],\n", 330 | " [{'Name': ['copper–platinum'],\n", 331 | " 'Formula': [],\n", 332 | " 'Acronym': ['Cu–Pt'],\n", 333 | " 'Descriptor': ['transparent', 'conducting'],\n", 334 | " 'Structure/Phase': ['nanowire'],\n", 335 | " 'Application': [{'electrocatalysis': [\"'activity': [8] (property)\"]}],\n", 336 | " 'Property': [\"'transmission': [80]\"],\n", 337 | " 'Synthesis': ['electroplating'],\n", 338 | " 'Characterization': []},\n", 339 | " ''],\n", 340 | " [{'Name': [],\n", 341 | " 'Formula': ['CH3NH3PbI3'],\n", 342 | " 'Acronym': [],\n", 343 | " 'Descriptor': ['perovskites'],\n", 344 | " 'Structure/Phase': ['single-crystalline'],\n", 345 | " 'Application': [{'photovoltaics': [], 'optoelectronics': []}],\n", 346 | " 'Property': [\"'length': [80]\",\n", 347 | " \"'diameter': []\",\n", 348 | " \"'thickness': []\",\n", 349 | " \"'response time': [40]\",\n", 350 | " \"'stability': []\"],\n", 351 | " 'Synthesis': ['inverse-temperature crystallization'],\n", 352 | " 'Characterization': []},\n", 353 | " ''],\n", 354 | " [''],\n", 355 | " [''],\n", 356 | " [''],\n", 357 | " [{'Name': ['Al2O3'],\n", 358 | " 'Formula': [],\n", 359 | " 'Acronym': [],\n", 360 | " 'Descriptor': ['solar cell'],\n", 361 | " 'Structure/Phase': ['perovskite'],\n", 362 | " 'Application': [{'cell efficiency': []}],\n", 363 | " 'Property': [\"'diffusion coefficient': []\", \"'collection efficiency': []\"],\n", 364 | " 'Synthesis': ['fabrication'],\n", 365 | " 'Characterization': ['cell efficiency']},\n", 366 | " ''],\n", 367 | " [{'Name': [],\n", 368 | " 'Formula': ['Pr0.5Sr0.5Co1 − xMnxO3'],\n", 369 | " 'Acronym': [],\n", 370 | " 'Descriptor': ['perovskites'],\n", 371 | " 'Structure/Phase': [],\n", 372 | " 'Application': [{}],\n", 373 | " 'Property': [\"'magnetoresistance': [0.95]\"],\n", 374 | " 'Synthesis': [],\n", 375 | " 'Characterization': ['magnetization',\n", 376 | " 'electrical conductivity measurements',\n", 377 | " 'neighboring atoms'],\n", 378 | " 'Model': []},\n", 379 | " ''],\n", 380 | " [''],\n", 381 | " [''],\n", 382 | " [''],\n", 383 | " [''],\n", 384 | " [''],\n", 385 | " [''],\n", 386 | " [''],\n", 387 | " [''],\n", 388 | " [''],\n", 389 | " [{'Name': [],\n", 390 | " 'Formula': ['NaTi2(PO4)3'],\n", 391 | " 'Acronym': [],\n", 392 | " 'Descriptor': [],\n", 393 | " 'Structure/Phase': ['P4'],\n", 394 | " 'Application': [{}],\n", 395 | " 'Property': [\"'phase transition temperature': [355]\"],\n", 396 | " 'Synthesis': [],\n", 397 | " 'Characterization': ['X-ray diffraction']},\n", 398 | " {'Name': ['NaTi0.5Sn1.5(PO4)3'],\n", 399 | " 'Formula': [],\n", 400 | " 'Acronym': [],\n", 401 | " 'Descriptor': [],\n", 402 | " 'Structure/Phase': ['P4'],\n", 403 | " 'Application': [{}],\n", 404 | " 'Property': [\"'phase transition temperature': [26]\"],\n", 405 | " 'Synthesis': [],\n", 406 | " 'Characterization': ['X-ray diffraction']},\n", 407 | " ''],\n", 408 | " [''],\n", 409 | " [{'Name': [],\n", 410 | " 'Formula': ['AlN'],\n", 411 | " 'Acronym': [],\n", 412 | " 'Descriptor': ['thin film'],\n", 413 | " 'Structure/Phase': ['hexagonal aluminum nitride', 'cubic aluminum nitride'],\n", 414 | " 'Application': [{'silicon': [\"'magnetron sputtering': [] (property)\",\n", 415 | " \"'power impulse magnetron sputtering': [] (property)\"]}],\n", 416 | " 'Property': [],\n", 417 | " 'Synthesis': ['magnetron sputtering', 'power impulse magnetron sputtering'],\n", 418 | " 'Characterization': ['X-ray diffraction']},\n", 419 | " ''],\n", 420 | " [''],\n", 421 | " [''],\n", 422 | " [''],\n", 423 | " [''],\n", 424 | " [''],\n", 425 | " [{'Name': ['Vanadium transition metal oxide films'],\n", 426 | " 'Formula': [],\n", 427 | " 'Acronym': [],\n", 428 | " 'Descriptor': ['annealing'],\n", 429 | " 'Structure/Phase': ['film'],\n", 430 | " 'Application': [{'thin film': [\"'efficiency': [] (property)\",\n", 431 | " \"'reliability': [] (property)\"]}],\n", 432 | " 'Property': [\"'thickness': [1]\", \"'diameter': [1]\"],\n", 433 | " 'Synthesis': ['annealing'],\n", 434 | " 'Characterization': ['X-ray diffraction',\n", 435 | " 'X-ray photoelectron spectroscopy',\n", 436 | " 'scanning electron microscopy',\n", 437 | " 'optical absorption',\n", 438 | " 'near-infrared absorption']},\n", 439 | " '']]" 440 | ] 441 | }, 442 | "execution_count": 3, 443 | "metadata": {}, 444 | "output_type": "execute_result" 445 | } 446 | ], 447 | "source": [ 448 | "def NFA_SC(file_path):\n", 449 | " res = []\n", 450 | " delete_res = []\n", 451 | " with open(file_path, 'r', encoding='utf-8') as file:\n", 452 | " for line in file:\n", 453 | " entry = json.loads(line.strip())\n", 454 | " doi_part = entry['paper']\n", 455 | " abs_part = entry['instruction'].split('### Input:\\n')[1].split('\\n\\n### Response:')[0].strip()\n", 456 | " response_part = entry['response'].replace('', '').split('### Response:')[1].strip().split(' & ')\n", 457 | " temp = []\n", 458 | " for part in response_part:\n", 459 | " if part.strip():\n", 460 | " try:\n", 461 | " NERoutput = json.loads(part)\n", 462 | " \n", 463 | " #Extract chemical information and initialize the standard set for subsequent similarity comparison\n", 464 | " materials, abbreviations = extract_chemical_info(abs_part)\n", 465 | " materials = rule_based_extension(abs_part, materials)\n", 466 | " Standard_set = set(materials)\n", 467 | " Standard_set.update(abbreviations.keys())\n", 468 | " Standard_set.update(abbreviations.values())\n", 469 | " Standard_list = list(Standard_set) # Keep Chem elements single and accelerate efficiency\n", 470 | "\n", 471 | " Name = NERoutput['Name'][0] if NERoutput['Name'] else '-'\n", 472 | " Formula = NERoutput['Formula'][0] if NERoutput['Formula'] else '-'\n", 473 | " Acronym = NERoutput['Acronym'][0] if NERoutput['Acronym'] else '-'\n", 474 | "\n", 475 | " # SC1: Verify if Name, Formula, and Acronym are in the standard list. If not, set them to '-'.\n", 476 | " if Name != '-':\n", 477 | " if not correct_entities(Name, Standard_list):\n", 478 | " Name = '-'\n", 479 | " NERoutput['Name'].clear()\n", 480 | "\n", 481 | " if Formula != '-':\n", 482 | " if not correct_entities(Formula, Standard_list):\n", 483 | " Formula = '-'\n", 484 | " NERoutput['Formula'].clear()\n", 485 | "\n", 486 | " if Acronym != '-':\n", 487 | " if not correct_entities(Acronym, Standard_list):\n", 488 | " Acronym = '-'\n", 489 | " NERoutput['Acronym'].clear()\n", 490 | "\n", 491 | " # SC2: Correct Acronym tags.\n", 492 | " if Name == '-' and Formula == '-' and Acronym != '-':\n", 493 | " if Acronym.isupper() or (Acronym.endswith('s') and Acronym[:-1].isupper()):\n", 494 | " continue\n", 495 | " else:\n", 496 | " Name = Acronym\n", 497 | " Acronym = '-'\n", 498 | " NERoutput['Acronym'].clear()\n", 499 | "\n", 500 | " if Name == '-' and Formula != '-' and Acronym == '-':\n", 501 | " if Formula.isupper() or (Formula.endswith('s') and Formula[:-1].isupper()):\n", 502 | " Acronym = Formula\n", 503 | " Formula = '-'\n", 504 | " NERoutput['Formula'].clear()\n", 505 | "\n", 506 | " if Name != '-' and Formula == '-' and Acronym == '-':\n", 507 | " if Name.isupper() or (Name.endswith('s') and Name[:-1].isupper()):\n", 508 | " Acronym = Name\n", 509 | " Name = '-'\n", 510 | " NERoutput['Name'].clear()\n", 511 | "\n", 512 | " # SC3: Ensure that Name and Acronym correspond.\n", 513 | " if not abbreviations:\n", 514 | " Acronym = '-'\n", 515 | " NERoutput['Acronym'].clear()\n", 516 | " elif Name != '-':\n", 517 | " sim_Acronym = {key: 0 for key in abbreviations.keys()}\n", 518 | " sim_Name = {value: 0 for value in abbreviations.values()}\n", 519 | " for key in sim_Name.keys():\n", 520 | " sim_Name[key] = calculate_similarity(Name, key)\n", 521 | " max_sim_Name = max(sim_Name, key=sim_Name.get)\n", 522 | " max_Name = sim_Name[max_sim_Name]\n", 523 | " if max_Name > 0.85:\n", 524 | " Name = max_sim_Name\n", 525 | " if Formula != '-':\n", 526 | " sim_Formula = {key: 0 for key in abbreviations.keys()}\n", 527 | " for key in sim_Formula.keys():\n", 528 | " sim_Formula[key] = calculate_similarity(Formula, key)\n", 529 | " max_sim_Formula = max(sim_Formula, key=sim_Formula.get)\n", 530 | " max_Formula = sim_Formula[max_sim_Formula]\n", 531 | " if max_Formula > 0.85:\n", 532 | " Acronym = max_sim_Formula\n", 533 | " Formula = '-'\n", 534 | " NERoutput['Formula'].clear()\n", 535 | " for key, value in abbreviations.items():\n", 536 | " if value == max_sim_Name:\n", 537 | " Acronym = key \n", 538 | " elif Acronym != '-':\n", 539 | " sim_Acronym = {key: 0 for key in abbreviations.keys()}\n", 540 | " for key in sim_Acronym.keys():\n", 541 | " sim_Acronym[key] = calculate_similarity(Acronym, key)\n", 542 | " max_sim_Acronym = max(sim_Acronym, key=sim_Acronym.get)\n", 543 | " max_Acronym = sim_Acronym[max_sim_Acronym]\n", 544 | " if max_Acronym > 0.85:\n", 545 | " Acronym = max_sim_Acronym\n", 546 | " if Formula != '-':\n", 547 | " sim_Formula = {key: 0 for value in abbreviations.keys()}\n", 548 | " for key in sim_Formula.keys():\n", 549 | " sim_Formula[key] = calculate_similarity(Formula, key)\n", 550 | " max_sim_Formula = max(sim_Formula, key=sim_Formula.get)\n", 551 | " max_Formula = sim_Formula[max_sim_Formula]\n", 552 | " if max_Formula > 0.85:\n", 553 | " Name = max_sim_Formula\n", 554 | " Formula = '-'\n", 555 | " NERoutput['Formula'].clear()\n", 556 | " Name = abbreviations[max_sim_Acronym]\n", 557 | " else:\n", 558 | " Name = '-'\n", 559 | " NERoutput['Name'].clear()\n", 560 | " Acronym = '-'\n", 561 | " NERoutput['Acronym'].clear()\n", 562 | " elif Formula != '-':\n", 563 | " sim_val = {value: 0 for value in abbreviations.values()}\n", 564 | " for key in sim_val.keys():\n", 565 | " sim_val[key] = calculate_similarity(Formula, key)\n", 566 | " max_sim_val = max(sim_val, key=sim_val.get)\n", 567 | " max_val = sim_val[max_sim_val]\n", 568 | " if max_val > 0.85:\n", 569 | " Name = max_sim_val\n", 570 | " for key, value in abbreviations.items():\n", 571 | " if value == max_sim_val:\n", 572 | " Acronym = key\n", 573 | " Formula = '-'\n", 574 | " NERoutput['Formula'].clear()\n", 575 | " if Formula != '-':\n", 576 | " sim_key = {key: 0 for key in abbreviations.keys()}\n", 577 | " for key in sim_key.keys():\n", 578 | " sim_key[key] = calculate_similarity(Formula, key)\n", 579 | " max_sim_key = max(sim_key, key=sim_key.get)\n", 580 | " max_key = sim_key[max_sim_key]\n", 581 | " if max_key > 0.85:\n", 582 | " Acronym = max_sim_key\n", 583 | " Name = abbreviations[max_sim_key]\n", 584 | " Formula = '-'\n", 585 | " NERoutput['Formula'].clear()\n", 586 | " \n", 587 | " # SC4: Handle empty abbreviations, after SC, if all these label is empty, the output is invalid.\n", 588 | " if Name == '-' and Formula == '-' and Acronym == '-':\n", 589 | " delete_res.append([doi_part])\n", 590 | " continue\n", 591 | " elif Name != '-' and Formula != '-':\n", 592 | " if Name == Formula:\n", 593 | " Formula = '-'\n", 594 | " NERoutput['Formula'].clear()\n", 595 | " elif Name != '-' and Acronym != '-':\n", 596 | " if Name == Acronym:\n", 597 | " Acronym = '-'\n", 598 | " NERoutput['Acronym'].clear()\n", 599 | " elif Formula != '-' and Acronym != '-':\n", 600 | " if Formula == Acronym:\n", 601 | " Acronym = '-'\n", 602 | " NERoutput['Acronym'].clear()\n", 603 | "\n", 604 | " if Name != '-':\n", 605 | " NERoutput['Name'].clear()\n", 606 | " NERoutput['Name'].append(Name)\n", 607 | " if Formula != '-':\n", 608 | " NERoutput['Formula'].clear()\n", 609 | " NERoutput['Formula'].append(Formula)\n", 610 | " if Acronym != '-':\n", 611 | " NERoutput['Acronym'].clear()\n", 612 | " NERoutput['Acronym'].append(Acronym)\n", 613 | "\n", 614 | " temp.append(NERoutput)\n", 615 | " except:\n", 616 | " continue\n", 617 | " temp.append(doi_part)\n", 618 | " res.append(temp)\n", 619 | " return res\n", 620 | "\n", 621 | "NERoutput_NFASC = NFA_SC('./NC/NC_result.jsonl') # Load your inference result here\n", 622 | "NERoutput_NFASC" 623 | ] 624 | }, 625 | { 626 | "cell_type": "markdown", 627 | "id": "524b36f7-aff6-49db-8d83-499c13b2f071", 628 | "metadata": {}, 629 | "source": [ 630 | "# Other Label Standardization and Correct" 631 | ] 632 | }, 633 | { 634 | "cell_type": "code", 635 | "execution_count": 4, 636 | "id": "2d8109a7-2f9f-4bd6-b357-927bf3f8b967", 637 | "metadata": { 638 | "scrolled": true 639 | }, 640 | "outputs": [ 641 | { 642 | "data": { 643 | "text/plain": [ 644 | "[[{'Name': [],\n", 645 | " 'Formula': ['BaCe1 − xNdxO3 − x 2'],\n", 646 | " 'Acronym': [],\n", 647 | " 'Descriptor': ['polymetric'],\n", 648 | " 'Structure/Phase': ['perovskites'],\n", 649 | " 'Application': [{}],\n", 650 | " 'Property': [\"'resolution': [1000]\",\n", 651 | " \"'spontaneous strain': [0.05]\",\n", 652 | " \"'opaque': []\",\n", 653 | " \"'transparent': []\",\n", 654 | " \"'reflectance': [1000]\"],\n", 655 | " 'Synthesis': [],\n", 656 | " 'Characterization': []},\n", 657 | " ''],\n", 658 | " [{'Name': ['NaFeTi2F11'],\n", 659 | " 'Formula': [],\n", 660 | " 'Acronym': [],\n", 661 | " 'Descriptor': ['cathode'],\n", 662 | " 'Structure/Phase': [],\n", 663 | " 'Application': [{}],\n", 664 | " 'Property': [],\n", 665 | " 'Synthesis': [],\n", 666 | " 'Characterization': []},\n", 667 | " ''],\n", 668 | " [''],\n", 669 | " [''],\n", 670 | " [''],\n", 671 | " [{'Name': ['Battery manganese dioxide'],\n", 672 | " 'Formula': [],\n", 673 | " 'Acronym': [],\n", 674 | " 'Descriptor': ['electrochemically deposited', 'doped'],\n", 675 | " 'Structure/Phase': [],\n", 676 | " 'Application': [{}],\n", 677 | " 'Property': [\"'etymology': []\"],\n", 678 | " 'Synthesis': ['electrochemical deposition '],\n", 679 | " 'Characterization': []},\n", 680 | " ''],\n", 681 | " [{'Name': [],\n", 682 | " 'Formula': ['La1-xCexCoO3'],\n", 683 | " 'Acronym': [],\n", 684 | " 'Descriptor': ['catalyst'],\n", 685 | " 'Structure/Phase': [],\n", 686 | " 'Application': [{}],\n", 687 | " 'Property': [\"'active site': []\",\n", 688 | " \"'average particle size': []\",\n", 689 | " \"'shape': []\"],\n", 690 | " 'Synthesis': [],\n", 691 | " 'Characterization': ['Ultraviolet photoelectron spectroscopy']},\n", 692 | " ''],\n", 693 | " [''],\n", 694 | " [{'Name': ['Nafion'],\n", 695 | " 'Formula': [],\n", 696 | " 'Acronym': ['SPCE'],\n", 697 | " 'Descriptor': ['emission'],\n", 698 | " 'Structure/Phase': [],\n", 699 | " 'Application': [{}],\n", 700 | " 'Property': [\"'thickness': [0.05–0.5]\", \"'efficiency': [0.05–0.5]\"],\n", 701 | " 'Synthesis': [],\n", 702 | " 'Characterization': []},\n", 703 | " ''],\n", 704 | " [''],\n", 705 | " [''],\n", 706 | " [{'Name': ['CdO'],\n", 707 | " 'Formula': [],\n", 708 | " 'Acronym': [],\n", 709 | " 'Descriptor': ['ceramics'],\n", 710 | " 'Structure/Phase': [],\n", 711 | " 'Application': [{}],\n", 712 | " 'Property': [\"'electrical conductivity': []\",\n", 713 | " \"'lattice thermal conductivity': []\",\n", 714 | " \"'power factor': []\"],\n", 715 | " 'Synthesis': [],\n", 716 | " 'Characterization': []},\n", 717 | " ''],\n", 718 | " [''],\n", 719 | " [{'Name': ['polydimethylsiloxane'],\n", 720 | " 'Formula': [],\n", 721 | " 'Acronym': ['PDMS'],\n", 722 | " 'Descriptor': ['elastomer surface functionalization'],\n", 723 | " 'Structure/Phase': [],\n", 724 | " 'Application': [{'bioimaging': []}],\n", 725 | " 'Property': [],\n", 726 | " 'Synthesis': ['sol gel '],\n", 727 | " 'Characterization': ['Ultraviolet photoelectron spectroscopy',\n", 728 | " 'transmission electron microscope',\n", 729 | " 'surface cracking']},\n", 730 | " ''],\n", 731 | " [''],\n", 732 | " [''],\n", 733 | " [''],\n", 734 | " [''],\n", 735 | " [{'Name': [],\n", 736 | " 'Formula': ['Cu3Pd'],\n", 737 | " 'Acronym': [],\n", 738 | " 'Descriptor': ['polar'],\n", 739 | " 'Structure/Phase': [],\n", 740 | " 'Application': [{}],\n", 741 | " 'Property': [\"'ordering': []\"],\n", 742 | " 'Synthesis': ['annealing'],\n", 743 | " 'Characterization': ['transmission electron microscope']},\n", 744 | " ''],\n", 745 | " [{'Name': [],\n", 746 | " 'Formula': ['Cu2S-SrLnCuS3'],\n", 747 | " 'Acronym': [],\n", 748 | " 'Descriptor': ['solution'],\n", 749 | " 'Structure/Phase': ['orthorhombic'],\n", 750 | " 'Application': [{}],\n", 751 | " 'Property': [\"'unit cell parameters': []\",\n", 752 | " \"'volume': []\",\n", 753 | " \"'monotony': [0.001]\",\n", 754 | " \"'periodicity': [0.002]\"],\n", 755 | " 'Synthesis': [],\n", 756 | " 'Characterization': []},\n", 757 | " ''],\n", 758 | " [''],\n", 759 | " [{'Name': [],\n", 760 | " 'Formula': ['7Li'],\n", 761 | " 'Acronym': [],\n", 762 | " 'Descriptor': ['NMR'],\n", 763 | " 'Structure/Phase': [],\n", 764 | " 'Application': [{'lithium-ion battery': []}],\n", 765 | " 'Property': [],\n", 766 | " 'Synthesis': [],\n", 767 | " 'Characterization': []},\n", 768 | " ''],\n", 769 | " [{'Name': [],\n", 770 | " 'Formula': ['LiCoO2'],\n", 771 | " 'Acronym': [],\n", 772 | " 'Descriptor': ['electrode'],\n", 773 | " 'Structure/Phase': [],\n", 774 | " 'Application': [{'lithium-ion battery': [\"'charge transfer kinetics': [] (property)\",\n", 775 | " \"'resistance': [] (property)\"]}],\n", 776 | " 'Property': [],\n", 777 | " 'Synthesis': [],\n", 778 | " 'Characterization': []},\n", 779 | " {'Name': [],\n", 780 | " 'Formula': ['LiFePO4'],\n", 781 | " 'Acronym': [],\n", 782 | " 'Descriptor': ['electrode'],\n", 783 | " 'Structure/Phase': [],\n", 784 | " 'Application': [{'lithium-ion battery': [\"'charge transfer kinetics': [] (property)\",\n", 785 | " \"'resistance': [] (property)\"]}],\n", 786 | " 'Property': [],\n", 787 | " 'Synthesis': [],\n", 788 | " 'Characterization': []},\n", 789 | " {'Name': [],\n", 790 | " 'Formula': ['LiMn2O4'],\n", 791 | " 'Acronym': [],\n", 792 | " 'Descriptor': ['electrode'],\n", 793 | " 'Structure/Phase': [],\n", 794 | " 'Application': [{'lithium-ion battery': [\"'charge transfer kinetics': [] (property)\",\n", 795 | " \"'resistance': [] (property)\"]}],\n", 796 | " 'Property': [],\n", 797 | " 'Synthesis': [],\n", 798 | " 'Characterization': []},\n", 799 | " ''],\n", 800 | " [{'Name': [],\n", 801 | " 'Formula': ['LiNi0.5Mn1.5O4'],\n", 802 | " 'Acronym': [],\n", 803 | " 'Descriptor': ['cathodes'],\n", 804 | " 'Structure/Phase': ['nanoparticle', 'nanorod', 'hollow spheres'],\n", 805 | " 'Application': [{'lithium-ion battery': [\"'energy density': [1.6] (property)\",\n", 806 | " \"'capacity': [121] (property)\"]}],\n", 807 | " 'Property': [\"'capacity': [100] (property)\"],\n", 808 | " 'Synthesis': [],\n", 809 | " 'Characterization': []},\n", 810 | " ''],\n", 811 | " [{'Name': ['yttrium barium copper oxide'],\n", 812 | " 'Formula': [],\n", 813 | " 'Acronym': [],\n", 814 | " 'Descriptor': ['superconducting material'],\n", 815 | " 'Structure/Phase': [],\n", 816 | " 'Application': [{}],\n", 817 | " 'Property': [\"'core diameter': [100–200]\",\n", 818 | " \"'core strength': [100–200]\",\n", 819 | " \"'diameter': [300–600]\"],\n", 820 | " 'Synthesis': [],\n", 821 | " 'Characterization': []},\n", 822 | " ''],\n", 823 | " [{'Name': ['copper–platinum'],\n", 824 | " 'Formula': [],\n", 825 | " 'Acronym': ['Cu–Pt'],\n", 826 | " 'Descriptor': ['transparent', 'conducting'],\n", 827 | " 'Structure/Phase': ['nanowire'],\n", 828 | " 'Application': [{}],\n", 829 | " 'Property': [\"'transmission': [80]\"],\n", 830 | " 'Synthesis': [],\n", 831 | " 'Characterization': []},\n", 832 | " ''],\n", 833 | " [{'Name': [],\n", 834 | " 'Formula': ['CH3NH3PbI3'],\n", 835 | " 'Acronym': [],\n", 836 | " 'Descriptor': ['perovskites'],\n", 837 | " 'Structure/Phase': [],\n", 838 | " 'Application': [{}],\n", 839 | " 'Property': [\"'length': [80]\",\n", 840 | " \"'diameter': []\",\n", 841 | " \"'thickness': []\",\n", 842 | " \"'response time': [40]\",\n", 843 | " \"'stability': []\"],\n", 844 | " 'Synthesis': [],\n", 845 | " 'Characterization': []},\n", 846 | " ''],\n", 847 | " [''],\n", 848 | " [''],\n", 849 | " [''],\n", 850 | " [{'Name': ['Al2O3'],\n", 851 | " 'Formula': [],\n", 852 | " 'Acronym': [],\n", 853 | " 'Descriptor': ['solar cell'],\n", 854 | " 'Structure/Phase': ['perovskite'],\n", 855 | " 'Application': [{}],\n", 856 | " 'Property': [\"'diffusion coefficient': []\", \"'collection efficiency': []\"],\n", 857 | " 'Synthesis': [],\n", 858 | " 'Characterization': []},\n", 859 | " ''],\n", 860 | " [{'Name': [],\n", 861 | " 'Formula': ['Pr0.5Sr0.5Co1 − xMnxO3'],\n", 862 | " 'Acronym': [],\n", 863 | " 'Descriptor': ['perovskites'],\n", 864 | " 'Structure/Phase': [],\n", 865 | " 'Application': [{}],\n", 866 | " 'Property': [\"'magnetoresistance': [0.95]\"],\n", 867 | " 'Synthesis': [],\n", 868 | " 'Characterization': ['electrical conductivity measurements'],\n", 869 | " 'Model': []},\n", 870 | " ''],\n", 871 | " [''],\n", 872 | " [''],\n", 873 | " [''],\n", 874 | " [''],\n", 875 | " [''],\n", 876 | " [''],\n", 877 | " [''],\n", 878 | " [''],\n", 879 | " [''],\n", 880 | " [{'Name': [],\n", 881 | " 'Formula': ['NaTi2(PO4)3'],\n", 882 | " 'Acronym': [],\n", 883 | " 'Descriptor': [],\n", 884 | " 'Structure/Phase': [],\n", 885 | " 'Application': [{}],\n", 886 | " 'Property': [\"'phase transition temperature': [355]\"],\n", 887 | " 'Synthesis': [],\n", 888 | " 'Characterization': ['X-ray diffraction']},\n", 889 | " {'Name': ['NaTi0.5Sn1.5(PO4)3'],\n", 890 | " 'Formula': [],\n", 891 | " 'Acronym': [],\n", 892 | " 'Descriptor': [],\n", 893 | " 'Structure/Phase': [],\n", 894 | " 'Application': [{}],\n", 895 | " 'Property': [\"'phase transition temperature': [26]\"],\n", 896 | " 'Synthesis': [],\n", 897 | " 'Characterization': ['X-ray diffraction']},\n", 898 | " ''],\n", 899 | " [''],\n", 900 | " [{'Name': [],\n", 901 | " 'Formula': ['AlN'],\n", 902 | " 'Acronym': [],\n", 903 | " 'Descriptor': ['thin film'],\n", 904 | " 'Structure/Phase': ['cubic aluminum nitride'],\n", 905 | " 'Application': [{}],\n", 906 | " 'Property': [],\n", 907 | " 'Synthesis': ['magnetron sputtering', 'magnetron sputtering'],\n", 908 | " 'Characterization': ['X-ray diffraction']},\n", 909 | " ''],\n", 910 | " [''],\n", 911 | " [''],\n", 912 | " [''],\n", 913 | " [''],\n", 914 | " [''],\n", 915 | " [{'Name': ['Vanadium transition metal oxide films'],\n", 916 | " 'Formula': [],\n", 917 | " 'Acronym': [],\n", 918 | " 'Descriptor': ['annealing'],\n", 919 | " 'Structure/Phase': [],\n", 920 | " 'Application': [{}],\n", 921 | " 'Property': [\"'thickness': [1]\", \"'diameter': [1]\"],\n", 922 | " 'Synthesis': ['annealing'],\n", 923 | " 'Characterization': ['X-ray diffraction',\n", 924 | " 'Ultraviolet photoelectron spectroscopy',\n", 925 | " 'scanning electron microscopy',\n", 926 | " 'optical absorption spectra']},\n", 927 | " '']]" 928 | ] 929 | }, 930 | "execution_count": 4, 931 | "metadata": {}, 932 | "output_type": "execute_result" 933 | } 934 | ], 935 | "source": [ 936 | "NERoutput_OTHSC = copy.deepcopy(NERoutput_NFASC)\n", 937 | "\n", 938 | "def OTHER_SC(abbr_path, keyword_path, NERoutput):\n", 939 | " abbr_df = pd.read_excel(abbr_path, header=None)\n", 940 | " keywords_df = pd.read_excel(keyword_path, header=None)\n", 941 | " \n", 942 | " name2abbr = pd.Series(abbr_df[1].values, index=abbr_df[0]).to_dict()\n", 943 | " keywords = list(keywords_df.itertuples(index=False, name=None))\n", 944 | " \n", 945 | " for entries in NERoutput:\n", 946 | " for entry in entries[:-1]:\n", 947 | " if isinstance(entry, dict):\n", 948 | " for key in list(entry.keys()):\n", 949 | " if key in ['Structure/Phase', 'Synthesis', 'Characterization']:\n", 950 | " for i, entity in enumerate(entry[key]):\n", 951 | " max_similarity = 0\n", 952 | " selected_keyword = None\n", 953 | " for keyword_tuple in keywords:\n", 954 | " keyword, threshold, abbr_condition, category = keyword_tuple\n", 955 | " similarity = calculate_similarity(entity, keyword)\n", 956 | " if similarity > max_similarity:\n", 957 | " max_similarity = similarity\n", 958 | " if similarity > threshold / 100:\n", 959 | " selected_keyword = keyword_tuple\n", 960 | " if selected_keyword:\n", 961 | " keyword, _, abbr_condition, category = selected_keyword\n", 962 | " if abbr_condition == 'a':\n", 963 | " new_entity = next((k for k, v in name2abbr.items() if v == keyword), keyword)\n", 964 | " else:\n", 965 | " new_entity = keyword\n", 966 | " if category == key:\n", 967 | " entry[key][i] = new_entity # Replace the old entity with the new one\n", 968 | " else:\n", 969 | " # Remove the old entity from its current place\n", 970 | " if entity in entry[key]:\n", 971 | " entry[key].remove(entity)\n", 972 | " # Add the new entity to its correct category, handling 'Application' uniquely\n", 973 | " if category not in entry:\n", 974 | " entry[category] = [{}] if category == 'Application' else []\n", 975 | " if category == 'Application':\n", 976 | " entry[category][0][new_entity] = []\n", 977 | " else:\n", 978 | " if new_entity not in entry[category]:\n", 979 | " entry[category].append(new_entity)\n", 980 | " else:\n", 981 | " if entity in entry[key]:\n", 982 | " entry[key].remove(entity)\n", 983 | " \n", 984 | " elif key == 'Application':\n", 985 | " for app_dict in entry['Application']:\n", 986 | " app_keys_to_delete = []\n", 987 | " for app_key in list(app_dict.keys()):\n", 988 | " max_similarity = 0\n", 989 | " selected_keyword = None\n", 990 | " for keyword_tuple in keywords:\n", 991 | " keyword, threshold, abbr_condition, category = keyword_tuple\n", 992 | " similarity = calculate_similarity(app_key, keyword)\n", 993 | " if similarity > max_similarity:\n", 994 | " max_similarity = similarity\n", 995 | " if similarity > threshold / 100:\n", 996 | " selected_keyword = keyword_tuple\n", 997 | " if selected_keyword:\n", 998 | " keyword, _, abbr_condition, category = selected_keyword\n", 999 | " if abbr_condition == 'a':\n", 1000 | " new_app_key = next((k for k, v in name2abbr.items() if v == keyword), keyword)\n", 1001 | " else:\n", 1002 | " new_app_key = keyword\n", 1003 | " if category != 'Application':\n", 1004 | " app_value = app_dict.pop(app_key)\n", 1005 | " if category not in entry:\n", 1006 | " entry[category] = []\n", 1007 | " if new_app_key not in entry[category]:\n", 1008 | " entry[category].append(new_app_key)\n", 1009 | " else:\n", 1010 | " if new_app_key != app_key:\n", 1011 | " app_dict[new_app_key] = app_dict.pop(app_key)\n", 1012 | " else:\n", 1013 | " app_keys_to_delete.append(app_key)\n", 1014 | " for key_to_delete in app_keys_to_delete:\n", 1015 | " del app_dict[key_to_delete]\n", 1016 | " return NERoutput\n", 1017 | "\n", 1018 | "def process_property(file_path, NERoutput):\n", 1019 | " keywords_df = pd.read_excel(file_path, header=None)\n", 1020 | " keywords = [(row[0], row[1]) for index, row in keywords_df.iterrows()]\n", 1021 | "\n", 1022 | " for entry in NERoutput:\n", 1023 | " if entry:\n", 1024 | " for dict_entry in entry[:-1]:\n", 1025 | " if 'Application' in dict_entry:\n", 1026 | " for app_dict in dict_entry['Application']:\n", 1027 | " for key, values in app_dict.items():\n", 1028 | " new_values = []\n", 1029 | " for line in values:\n", 1030 | " property_name, new_line = extract_and_replace_property(line, keywords)\n", 1031 | " if new_line:\n", 1032 | " new_values.append(new_line)\n", 1033 | " app_dict[key] = new_values\n", 1034 | " if 'Property' in dict_entry:\n", 1035 | " new_properties = []\n", 1036 | " for line in dict_entry['Property']:\n", 1037 | " property_name, new_line = extract_and_replace_property(line, keywords)\n", 1038 | " if new_line:\n", 1039 | " new_properties.append(new_line)\n", 1040 | " dict_entry['Property'] = new_properties\n", 1041 | "\n", 1042 | "\n", 1043 | "def extract_and_replace_property(line, keywords):\n", 1044 | " try:\n", 1045 | " if line.strip().startswith(\"'\") and \"':\" in line:\n", 1046 | " property_name = line.strip().split(\"'\")[1]\n", 1047 | " max_similarity = 0\n", 1048 | " replacement_keyword = None\n", 1049 | " max_threshold = 0\n", 1050 | " \n", 1051 | " for keyword, threshold in keywords:\n", 1052 | " similarity = calculate_similarity(property_name, keyword)\n", 1053 | " if similarity > max_similarity:\n", 1054 | " max_similarity = similarity\n", 1055 | " replacement_keyword = keyword\n", 1056 | " max_threshold = threshold\n", 1057 | " \n", 1058 | " if max_threshold and max_similarity > max_threshold / 100:\n", 1059 | " new_line = line.replace(f\"'{property_name}'\", f\"'{replacement_keyword}'\")\n", 1060 | " return property_name, new_line\n", 1061 | " else:\n", 1062 | " return property_name, line\n", 1063 | " except:\n", 1064 | " print(line)\n", 1065 | " return None, None\n", 1066 | "\n", 1067 | "NERoutput_OTHSC = OTHER_SC('./dictionary/abbreviation.xlsx', './dictionary/keywords.xlsx', NERoutput_OTHSC)\n", 1068 | "process_property('./dictionary/property.xlsx', NERoutput_OTHSC)\n", 1069 | "NERoutput_OTHSC" 1070 | ] 1071 | }, 1072 | { 1073 | "cell_type": "markdown", 1074 | "id": "17fd82d9-aef5-4ea1-bcf1-5d06efaef40d", 1075 | "metadata": {}, 1076 | "source": [ 1077 | "# Binary Classification" 1078 | ] 1079 | }, 1080 | { 1081 | "cell_type": "code", 1082 | "execution_count": 23, 1083 | "id": "d31a4e7a-4c6a-4abc-b563-889b217ff358", 1084 | "metadata": {}, 1085 | "outputs": [], 1086 | "source": [ 1087 | "NERoutput_FT = copy.deepcopy(NERoutput_OTHSC)\n", 1088 | "\n", 1089 | "def NFA_FT(ner_output):\n", 1090 | " transformed_data = []\n", 1091 | " \n", 1092 | " for item in ner_output:\n", 1093 | " if isinstance(item[-1], str):\n", 1094 | " doi = item[-1]\n", 1095 | " for element in item[:-1]:\n", 1096 | " if 'Name' in element and element['Name']:\n", 1097 | " for name in element['Name']:\n", 1098 | " transformed_data.append({\n", 1099 | " \"resource\": doi,\n", 1100 | " \"instruction\": \"Below is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\\n\\n### Instruction:\\nTell me if the given material/chemical term belongs to the material/chemical Name or Formula.\\n\\n### Input:\\n\" + name + \"\\n\\n### Response:\",\n", 1101 | " })\n", 1102 | " if 'Formula' in element and element['Formula']:\n", 1103 | " for formula in element['Formula']:\n", 1104 | " transformed_data.append({\n", 1105 | " \"resource\": doi,\n", 1106 | " \"instruction\": \"Below is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\\n\\n### Instruction:\\nTell me if the given material/chemical term belongs to the material/chemical Name or Formula.\\n\\n### Input:\\n\" + formula + \"\\n\\n### Response:\",\n", 1107 | " })\n", 1108 | "\n", 1109 | " return transformed_data\n", 1110 | "\n", 1111 | "# Apply the transformation to the NER output\n", 1112 | "NFA_FT = NFA_FT(NERoutput_FT)\n", 1113 | "\n", 1114 | "# Save the transformed data to a JSON file\n", 1115 | "output_file_path = './NFA_FT/classification7.json'\n", 1116 | "with open(output_file_path, 'w') as f:\n", 1117 | " for item in NFA_FT:\n", 1118 | " f.write(json.dumps(item) + \"\\n\")" 1119 | ] 1120 | }, 1121 | { 1122 | "cell_type": "markdown", 1123 | "id": "194e14b1", 1124 | "metadata": {}, 1125 | "source": [ 1126 | "# Here you need to process the 'classification7.json' through LLMs" 1127 | ] 1128 | }, 1129 | { 1130 | "cell_type": "code", 1131 | "execution_count": 26, 1132 | "id": "587b073d-fa4b-4581-b971-04cbc1256110", 1133 | "metadata": {}, 1134 | "outputs": [], 1135 | "source": [ 1136 | "responses_file_path = './NFA_FT/processed_NFA7.json' # Load the classified result here.\n", 1137 | "responses_data = []\n", 1138 | "\n", 1139 | "with open(responses_file_path, 'r') as file:\n", 1140 | " for line in file:\n", 1141 | " response = json.loads(line)\n", 1142 | " responses_data.append(response)\n", 1143 | "\n", 1144 | "NFAFT_dict = {}\n", 1145 | " \n", 1146 | "def update_nfaft_from_response(response):\n", 1147 | " input_start = response[\"response\"].find(\"### Input:\\n\") + len(\"### Input:\\n\")\n", 1148 | " input_end = response[\"response\"].find(\"\\n\\n### Response:\")\n", 1149 | " input_term = response[\"response\"][input_start:input_end].strip()\n", 1150 | " \n", 1151 | " response_start = response[\"response\"].find(\"### Response:\") + len(\"### Response:\")\n", 1152 | " response_end = response[\"response\"].find(\"\")\n", 1153 | " extracted_response = response[\"response\"][response_start:response_end].strip()\n", 1154 | " \n", 1155 | " NFAFT_dict[input_term] = extracted_response\n", 1156 | "\n", 1157 | "for response in responses_data:\n", 1158 | " update_nfaft_from_response(response)\n", 1159 | "\n", 1160 | "def check_and_replace(ner_output, NFAFT_dict):\n", 1161 | " for item in ner_output:\n", 1162 | " for element in item[:-1]:\n", 1163 | " if isinstance(element, dict):\n", 1164 | " for key in ['Name', 'Formula']:\n", 1165 | " if key in element and element[key]:\n", 1166 | " values_to_check = element[key]\n", 1167 | " for val in list(values_to_check):\n", 1168 | " if val in NFAFT_dict:\n", 1169 | " correct_key = NFAFT_dict[val]\n", 1170 | " if correct_key != key:\n", 1171 | " opposite_key = 'Formula' if key == 'Name' else 'Name'\n", 1172 | " if not element.get(opposite_key):\n", 1173 | " element[opposite_key] = []\n", 1174 | " element[opposite_key].append(val)\n", 1175 | " element[key].remove(val)\n", 1176 | "\n", 1177 | "check_and_replace(NERoutput_FT, NFAFT_dict)" 1178 | ] 1179 | }, 1180 | { 1181 | "cell_type": "markdown", 1182 | "id": "d6621dbb-6ceb-4302-bf18-5b63e097f6ba", 1183 | "metadata": {}, 1184 | "source": [ 1185 | "# Save SCNERoutput" 1186 | ] 1187 | }, 1188 | { 1189 | "cell_type": "code", 1190 | "execution_count": 17, 1191 | "id": "a23cc237-d239-420c-97ee-4c44fbb01a62", 1192 | "metadata": {}, 1193 | "outputs": [], 1194 | "source": [ 1195 | "def save_to_json(data, file_name):\n", 1196 | " with open(file_name, 'w', encoding='utf-8') as file:\n", 1197 | " json.dump(data, file, indent=4)\n", 1198 | "\n", 1199 | "save_to_json(NERoutput_FT, 'ER_RES.json')" 1200 | ] 1201 | } 1202 | ], 1203 | "metadata": { 1204 | "kernelspec": { 1205 | "display_name": "Python 3 (ipykernel)", 1206 | "language": "python", 1207 | "name": "python3" 1208 | }, 1209 | "language_info": { 1210 | "codemirror_mode": { 1211 | "name": "ipython", 1212 | "version": 3 1213 | }, 1214 | "file_extension": ".py", 1215 | "mimetype": "text/x-python", 1216 | "name": "python", 1217 | "nbconvert_exporter": "python", 1218 | "pygments_lexer": "ipython3", 1219 | "version": "3.8.18" 1220 | } 1221 | }, 1222 | "nbformat": 4, 1223 | "nbformat_minor": 5 1224 | } 1225 | -------------------------------------------------------------------------------- /README.md: -------------------------------------------------------------------------------- 1 | # Material-Knowledge-Graph 2 | This is the directory for paper "Construction and Application of Materials Knowledge Graph in Multidisciplinary Materials Science via Large Language Mode", which presents a automatic NLP pipeline to extract information from 150,000 peer-reviewed papers and construct the material knowledge graph, which contains more than 18,000 entities and nearly 90,000 triples. 3 | 4 | The instructions of code: 5 | 6 | - data 7 | - TrainSet1.0.json: train dataset of named entity recognition and relation extraction (NERRE) task version 1. 8 | - TrainSet2.0.json: train dataset of named entity recognition and relation extraction (NERRE) task version 2. 9 | - TrainSet_bc.json: train dataset of "Name" and "Formula" label classification (BC) task. 10 | - Property.xlsx and Keywords.xlsx: expert dictionary. 11 | - [train](https://github.com/MasterAI-EAM/Darwin/blob/main/train.py): code for training LLaMA-7B (outside in main directory) 12 | - Entity Resolution.ipynb: code for Entity Resolution (ER) task. 13 | - Graph completion (non-embedding).ipynb: code for link predication. 14 | - inference_KG.py: code for LLM inference 15 | 16 | 17 | ## Data Format 18 | TrainSet{version/task}.json is the JSON file containing a list of dictionaries, and each dictionary contains the following fields: 19 | - `instruction`: `str`, describes the task the model should perform. For NERRE, we use "You're a chatbot that can extract entities and relations in a paragraph.". For BC, we use "Tell me if the given material/chemical term belongs to the material/chemical Name or Formula.". 20 | - `input`: `str`, input for the task. 21 | - `output`: `str`, the answer to the instruction. 22 | 23 | ## Getting Started 24 | 1.install the requirements in the main directory: 25 | 26 | ```bash 27 | pip install -r requirements.txt 28 | ``` 29 | 30 | Then download the checkpoints of the open-source LLaMA-7B weights from huggingface. 31 | 32 | 2.fine-tune the LLMs using NERRE dataset and inference the corpus using inference_KG.py (FT detail shows below) 33 | 34 | We use version2.0 to fine-tune the LLMs and using part of data from 1.0 for evaluation. 35 | 36 | 3.use ER code to clean the inference result and construct the KG. 37 | 38 | [Mat2vec](https://github.com/materialsintelligence/mat2vec) and [ChemDataExactor](https://github.com/CambridgeMolecularEngineering/chemdataextractor) should be installed first. 39 | 40 | 4.use graph completion code to predicte potential links (we use pykeen package for TransE predication). 41 | 42 | ## Fine-tuning 43 | To fine-tune LLaMA-7b with NERRE/BC datasets, below is a command that works on a machine with 4 A100 80G GPUs in FSDP `full_shard` mode. 44 | Replace `` with a port of your own, `` with the 45 | path to your converted checkpoint and tokenizer, and `` with where you want to store your outputs. 46 | ```bash 47 | torchrun --nproc_per_node=8 --master_port= train.py \ 48 | --model_name_or_path \ 49 | --data_path \ 50 | --bf16 True \ 51 | --output_dir \ 52 | --num_train_epochs 10 \ 53 | --per_device_train_batch_size 1 \ 54 | --per_device_eval_batch_size 1 \ 55 | --gradient_accumulation_steps 2 \ 56 | --evaluation_strategy "no" \ 57 | --save_strategy "steps" \ 58 | --save_steps 500 \ 59 | --save_total_limit 1 \ 60 | --learning_rate 2e-5 \ 61 | --weight_decay 0. \ 62 | --warmup_ratio 0.03 \ 63 | --lr_scheduler_type "cosine" \ 64 | --logging_steps 1 \ 65 | --fsdp "full_shard auto_wrap" \ 66 | --fsdp_transformer_layer_cls_to_wrap 'LlamaDecoderLayer' \ 67 | --tf32 False 68 | --model_max_length 1024 69 | ``` 70 | 71 | To run on more gpus, you may prefer to turn down `gradient_accumulation_steps` to keep a global batch size of 128. Global batch size has not been tested for optimality. 72 | 73 | 74 | ## **Authors** 75 | 76 | This project is a collaborative effort by the following: 77 | 78 | UNSW: Yanpeng Ye, Shaozhou Wang, Tong Xie, Shaozhou Wang, Imran Razzak, Wenjie Zhang 79 | 80 | CityU HK: Jie Ren, Yuwei Wan 81 | 82 | Tongji University: Haofen Wang 83 | 84 | GreenDynamics: Yixuan Liu 85 | 86 | All advised by Wenjie Zhang from UNSW Engineering 87 | 88 | ## **Citation** 89 | 90 | If you use the data or code from this repository in your work, please cite it accordingly. 91 | 92 | ## **Acknowledgements** 93 | 94 | This project has referred to the following open-source projects: 95 | 96 | - Meta LLaMA: **[LLaMA](https://github.com/facebookresearch/llama)** 97 | - Stanford Alpaca: **[Alpaca](https://github.com/tatsu-lab/stanford_alpaca)** 98 | - Mat2vec: **[mat2vec](https://github.com/materialsintelligence/mat2vec)** 99 | - ChemDataExactor: **[ChemDataExactor](https://github.com/CambridgeMolecularEngineering/chemdataextractor)** 100 | -------------------------------------------------------------------------------- /data/Keywords.xlsx: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/MasterAI-EAM/Material-Knowledge-Graph/fee563926d480b5f2a564d3cf6c57d7697af4471/data/Keywords.xlsx -------------------------------------------------------------------------------- /data/Property.xlsx: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/MasterAI-EAM/Material-Knowledge-Graph/fee563926d480b5f2a564d3cf6c57d7697af4471/data/Property.xlsx -------------------------------------------------------------------------------- /data/TrainSet2.0.json: -------------------------------------------------------------------------------- 1 | [ 2 | { 3 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 4 | "input": "Pt/TiO2 nanosheets with exposed (001) facets were fabricated by a simple hydrothermal route in a Ti(OC4H9)(4)-HF-H2O mixed solution followed by a photochemical reduction deposition of Pt nanoparticles on TiO2 nanosheets under xenon lamp irradiation. The prepared samples were characterized by transmission electron microscopy, X-ray diffraction, X-ray photoelectron spectroscopy, nitrogen adsorption-desorption isotherms, UV-vis diffuse reflectance spectroscopy, and photoluminescence (PL) spectroscopy. Production of (OH)-O-center dot radicals on the TiO2 surface was detected by the PL technique using coumarin as a probe molecule. The effects of Pt loading on the rates of photocatalytic hydrogen production of the as-prepared samples in ethanol aqueous solution were investigated and discussed. The results showed that the photocatalytic hydrogen production rates of TiO2 nanosheets from the ethanol aqueous solutions were significantly enhanced by loaded Pt on the TiO2 nanosheets, and the latter with a 2 wt % of deposited Pt exhibited the highest photocatalytic activity. All fluorinated TiO2 nanosheets exhibited much higher photocatalytic activity than Degussa P25 TiO2 and pure TiO2 nanoparticles prepared in pure water due to the synergistic effect of surface fluorination and exposed (001) facets.", 5 | "output": "{\"Name\": [], \"Formula\": [\"Pt/TiO2\"], \"Acronym\": [], \"Descriptor\": [], \"Structure/Phase\": [\"nanosheets\", \"exposed (001) facets\"], \"Application\": [{\"photocatalytic hydrogen production\": [\"'photocatalytic activity': [] (property)\"]}], \"Property\": [\"high photocatalytic activity\"], \"Synthesis\": [\"hydrothermal reaction\"], \"Characterization\": [\"transmission electron microscropy\", \"X-ray diffraction\", \"X-ray photoelectron spectroscopy\", \"nitrogen adsorption\", \"UV-vis diffuse reflectance spectroscopy\", \"photoluminescence quenching\"]}" 6 | }, 7 | { 8 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 9 | "input": "Multiple-color-emissive carbon dots (CDots) have potential applications in various fields such as bioimaging, light-emitting devices, and photocatalysis. The majority of the current CDots to date exhibit excitation-wavelength-dependent emissions with their maximum emission limited at the blue-light region. Here, a synthesis of multiple-color-emission CDots by controlled graphitization and surface function is reported. The CDots are synthesized through controlled thermal pyrolysis of citric acid and urea. By regulating the thermal-pyrolysis temperature and ratio of reactants, the maximum emission of the resulting CDots gradually shifts from blue to red light, covering the entire light spectrum. Specifically, the emission position of the CDots can be tuned from 430 to 630 nm through controlling the extent of graphitization and the amount of surface functional groups, COOH. The relative photoluminescence quantum yields of the CDots with blue, green, and red emission reach up to 52.6%, 35.1%, and 12.9%, respectively. Furthermore, it is demonstrated that the CDots can be uniformly dispersed into epoxy resins and be fabricated as transparent CDots/epoxy composites for multiple-color- and white-light-emitting devices. This research opens a door for developing low-cost CDots as alternative phosphors for light-emitting devices.", 10 | "output": "{\"Name\": [\"carbon dots\"], \"Formula\": [], \"Acronym\": [\"CDots\"], \"Descriptor\": [\"low cost\"], \"Structure/Phase\": [], \"Application\": [{\"multiple-color- and white-light-emitting devices\": [], \"bioimaging\": [], \"photocatalysis\": []}], \"Property\": [\"'emission position': [430~630]\",\"'blue emission ': [0.526]\",\"'green emission ': [0.351]\",\"'red emission ': [0.129]\"], \"Synthesis\": [\"thermal pyrolysis\"], \"Characterization\": []}" 11 | }, 12 | { 13 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 14 | "input": "Energy captured directly from sunlight provides an attractive approach towards fulfilling the need for green energy resources on the terawatt scale with minimal environmental impact. Collecting and storing solar energy into fuel through photocatalyzed water splitting to generate hydrogen in a cost-effective way is desirable. To achieve this goal, low cost and environmentally benign urea was used to synthesize the metal-free photocatalyst graphitic carbon nitride (g-C3N4). A porous structure is achieved via one-step polymerization of the single precursor. The porous structure with increased BET surface area and pore volume shows a much higher hydrogen production rate under simulated sunlight irradiation than thiourea-derived and dicyanamide-derived g-C3N4. The presence of an oxygen atom is presumed to play a key role in adjusting the textural properties. Further improvement of the photocatalytic function can be expected with after-treatment due to its rich chemistry in functionalization.", 15 | "output": "{\"Name\": [\"graphitic carbon nitride\"], \"Formula\": [\"g-C3N4\"], \"Acronym\": [], \"Descriptor\": [\"metal-free\", \"green energy\", \"cost-effective\"], \"Structure/Phase\": [\"porous structure\"], \"Application\": [{\" hydrogen production\": []}], \"Property\": [\"increased BET surface area\", \"increased pore volume\"], \"Synthesis\": [\"one-step polymerization\"], \"Characterization\": []}" 16 | }, 17 | { 18 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 19 | "input": "Visible light-active TiO2 (m-TiO2) nanoparticles were obtained by an electron beam treatment of commercial TiO2 (p-TiO2) nanoparticles. The m-TiO2 nanoparticles exhibited a distinct red-shift in the UV-visible absorption spectrum and a much narrower band gap (2.85 eV) due to defects as confirmed by diffuse reflectance spectroscopy (DRS), photoluminescence (PL), X-ray diffraction, Raman spectroscopy, electron paramagnetic resonance, transmission electron microscopy, X-ray photoelectron spectroscopy (XPS), electrochemical impedance spectroscopy (EIS) and linear scan voltammetry (LSV). The XPS revealed changes in the surface states, composition, Ti4+ to Ti3+ ratio, and oxygen deficiencies in the mTiO(2). The valence band XPS, DRS and PL results were carefully examined to understand the band gap reduction of m-TiO2. The visible light-responsive enhanced photocatalytic activity of m-TiO2 was demonstrated by degrading methylene blue and brilliant blue G. The EIS and LSV in the dark and under visible light irradiation further support the visible light-induced photocatalytic activities of the m-TiO2 due to a decrease in electron transfer resistance and an increase in photocurrent. This study confirms that m-TiO2 can be used effectively as a photocatalyst and photoelectrode material owing to its enhanced visible light-induced photocatalytic activity.", 20 | "output": "{\"Name\": [\"Visible light-active TiO2\"], \"Formula\": [\"m-TiO2\"], \"Acronym\": [], \"Descriptor\": [], \"Structure/Phase\": [\"nanoparticles\"], \"Application\": [{\"visible-light photocatalyst\": []}], \"Property\": [\"'bandgap energy': [2.85]\", \"'enhanced photocatalytic activity:' []\"], \"Synthesis\": [\"electron beam treatment\"], \"Characterization\": [\"X-ray diffraction\", \"Raman spectroscopy\", \"electron paramagnetic resonance\", \"transmission electron microscropy\", \"X-ray photoelectron spectroscopy\", \"electrochemical impedance spectroscopy\", \"linear scan voltammetry \"]}" 21 | }, 22 | { 23 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 24 | "input": "The generation of green hydrogen (H-2) energy using sunlight is of great significance to solve the worldwide energy and environmental issues. Particularly, photocatalytic H-2 production is a highly promising strategy for solar-to-H-2 conversion. Recently, various heterostructured photocatalysts with high efficiency and good stability have been fabricated. Among them, 2D/2D van der Waals (VDW) heterojunctions have received tremendous attention, since this architecture can promote the interfacial charge separation and transfer and provide massive reactive centers. On the other hand, currently, most photocatalysts are composed of metal elements with high cost, limited reserves, and hazardous environmental impact. Hence, the development of metal-free photocatalysts is desirable. Here, a novel 2D/2D VDW heterostructure of metal-free phosphorene/graphitic carbon nitride (g-C3N4) is fabricated. The phosphorene/g-C3N4 nanocomposite shows an enhanced visible-light photocatalytic H-2 production activity of 571 mu mol h(-1) g(-1) in 18 v% lactic acid aqueous solution. This improved performance arises from the intimate electronic coupling at the 2D/2D interface, corroborated by the advanced characterizations techniques, e.g., synchrotron-based X-ray absorption near-edge structure, and theoretical calculations. This work not only reports a new metal-free phosphorene/g-C3N4 photocatalyst but also sheds lights on the design and fabrication of 2D/2D VDW heterojunction for applications in catalysis, electronics, and optoelectronics.", 25 | "output": "{\"Name\": [\"graphitic carbon nitride\"], \"Formula\": [\"g-C3N4\"], \"Acronym\": [], \"Descriptor\": [\"metal-free\", \"photocatalysts\"], \"Structure/Phase\": [\"2D/2D van der Waals\"], \"Application\": [{\" hydrogen production\": [\"'H-2-production rate ': [571] (property)\"]},\"catalysis\",\"electronics\",\"optoelectronics\"], \"Property\": [], \"Synthesis\": [], \"Characterization\": [\"X-ray absorption fine structure\"]}" 26 | }, 27 | { 28 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 29 | "input": "Titanium dioxide (M-TiO2), which was doped with 13 different metal ions (i.e., silver (Ag+), rubidium (Rb+), nickel (Ni2+), cobalt (Co2+), copper (Cu2+), vanadium (V3+), ruthenium (Ru3+), iron (Fe3+), osmium (Os3+), yttrium (Y3+), lanthanum (La3+), platinum (Pt4+, Pt2+), and chromium (Cr3+, Cr6+)) at doping levels ranging from 0.1 to 1.0 at. %, was synthesized by standard sol-gel methods and characterized by X-ray diffraction, BET surface area measurement, SEM, and UV-vis diffuse reflectance spectroscopy. Doping with Pt(IV/II), Cr(III), V(III), and Fe(III) resulted ill a lower anatase to rutile phase transformation (A-R phase transformation) temperature for the resultant TiO2 particles, while doping with Ru(III) inhibited the A-R phase transformation. Metal-ion doping also resulted in a red shift of the photophysical response of TiO2 that was reflected in an extended absorption ill the visible region between 400 and 700 nm. In contrast, doping with Ag(I), Rb(I), Y(III), and La(III) did not result in a red shift of the absorption spectrum of TiO2 As confirmed by elemental composition analysis by energy dispersive X-ray spectroscopy, the latter group Of ions was unable to be substituted for Ti(IV) in the crystalline matrix due to their incompatible ionic radii. The photocatalytic activities of doped TiO2 samples were quantified in terms of the photobleaching of methylene blue, the oxidation of iodide (I-), and the oxidative degradation of phenol in aqueous solution both under visible-light irradiation (lambda > 400 nm) and Under broader-band UV-vis irradiation (lambda > 320 nm). Pt- and Cr-doped TiO2, which had relatively high percentages of rutile in the particle phase, showed significantly enhanced visible-light photocatalytic activity for all three reaction classes.", 30 | "output": "{\"Name\": [\"Titanium dioxide\"], \"Formula\": [\"M-TiO2)\"], \"Acronym\": [], \"Descriptor\": [\"polymetallic doping\"], \"Structure/Phase\": [], \"Application\": [{\"visible-light photocatalyst\": []}], \"Property\": [], \"Synthesis\": [\"sol-gel method\", \"doped\"], \"Characterization\": [\"X-ray diffraction\", \"surface area measurement\", \"scanning electron microscopy\", \"UV-vis diffuse reflectance spectroscopy\"]}" 31 | }, 32 | { 33 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 34 | "input": "Graphene oxide modified g-C3N4 (GO/g-C3N4) with efficient photocatalytic capability under visible light irradiation was fabricated by sonochemical approach. Transmission electron microscopy images demonstrated that GO was two-dimensional sheets with chiffon-like ripples and the g-C3N4 possessed a layered structure. GO was overlaid on the surface of g-C3N4 in the GO/g-C3N4 hybrids. The UV-vis diffuse reflectance spectra showed that the GO/g-C3N4 hybrid had intense optical absorption in the visible light region. Photoluminescence spectra confirmed that the separation efficiency of photogenerated charge in GO/g-C3N4 was more intensive than pristine g-C3N4, indicating the GO acts as a separation centre and electron acceptor in the GO/g-C3N4 hybrid. The effective photogenerated charge separation efficiency lead to a remarkable improvement in the visible light photocatalysis. The pseudo-first-order kinetic constants of photocatalytic degradation of rhodamine B and 2, 4-dichlorophenol under visible light irradiation with GO/g-C3N4 were 3.80 and 2.08 times as large as that with pristine g-C3N4, respectively. This work indicates that the metal-free GO/g-C3N4 hybrid photocatalyst is a promising material in waste control, and GO could be an excellent material to combine with other semiconductors to make composites.", 35 | "output": "{\"Name\": [\"Graphene oxide modified g-C3N4\"], \"Formula\": [\"GO/g-C3N4\"], \"Acronym\": [], \"Descriptor\": [], \"Structure/Phase\": [\"two-dimensional sheets\",\"layered structure\"], \"Application\": [{\"visible-light photocatalyst\": [], \"waste controlt\": []}], \"Property\": [\"'separation efficiency of photogenerated charge': []\"], \"Synthesis\": [\"sonochemical approach\"], \"Characterization\": [\"UV-vis diffuse reflectance spectroscopy\", \"photoluminescence quenching\"]}" 36 | }, 37 | { 38 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 39 | "input": "Dinitrogen reduction to ammonia using transition metal catalysts is central to both the chemical industry and the Earth's nitrogen cycle. In the Haber-Bosch process, a metallic iron catalyst and high temperatures (400 degrees C) and pressures (200 atm) are necessary to activate and cleave NN bonds, motivating the search for alternative catalysts that can transform N-2 to NH3 under far milder reaction conditions. Here, the successful hydrothermal synthesis of ultrathin TiO2 nanosheets with an abundance of oxygen vacancies and intrinsic compressive strain, achieved through a facile copper-doping strategy, is reported. These defect-rich ultrathin anatase nanosheets exhibit remarkable and stable performance for photocatalytic reduction of N-2 to NH3 in water, exhibiting photoactivity up to 700 nm. The oxygen vacancies and strain effect allow strong chemisorption and activation of molecular N-2 and water, resulting in unusually high rates of NH3 evolution under visible-light irradiation. Therefore, this study offers a promising and sustainable route for the fixation of atmospheric N-2 using solar energy.", 40 | "output": "{\"Name\": [], \"Formula\": [\"TiO2\"], \"Acronym\": [], \"Descriptor\": [\"copper-doping\", \"catalysts\"], \"Structure/Phase\": [\"nanosheets\"], \"Application\": [{\"photocatalytic reduction\": []}], \"Property\": [\"'photoactivity': [700]\"], \"Synthesis\": [\"hydrothermal reaction\"], \"Characterization\": []}" 41 | }, 42 | { 43 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 44 | "input": "Graphitic carbon nitride (g-C3N4) has been deemed a promising heterogeneous metal-free catalyst for a wide range of applications, such as solar energy utilization toward water splitting, and its photocatalytic performance is reasonably adjustable through tailoring its texture and its electronic and optical properties. Here phosphorus-doped graphitic carbon nitride nanostructured flowers of in-plane mesopores are synthesized by a co-condensation method in the absence of any templates. The interesting structures, together with the phosphorus doping, can promote light trapping, mass transfer, and charge separation, enabling it to perform as a more impressive catalyst than its pristine carbon nitride counterpart for catalytic hydrogen evolution under visible light irradiation. The catalyst has low cost, is environmentally friendly, and represents a potential candidate in photoelectrochemistry.", 45 | "output": "{\"Name\": [\"graphitic carbon nitride\"], \"Formula\": [\"g-C3N4)\"], \"Acronym\": [], \"Descriptor\": [\"phosphorus doping\",\"low cost\"], \"Structure/Phase\": [\"nanostructured flowers\"], \"Application\": [{\"hydrogen evolutio\": [], \"photoelectrochemistry\": []}], \"Property\": [], \"Synthesis\": [\" co-condensation method\"], \"Characterization\": []}" 46 | }, 47 | { 48 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 49 | "input": "beta-Graphdiyne (-GDY) is a member of 2D graphyne family with zero band gap, and is a promising material with potential applications in energy storage, organic electronics, etc. However, the synthesis of -GDY has not been realized yet, and the measurement of its intrinsic properties remains elusive. In this work, -GDY-containing thin film is successfully synthesized on copper foil using modified Glaser-Hay coupling reaction with tetraethynylethene as precursor. The as-grown carbon film has a smooth surface and is continuous and uniform. Electrical measurements reveal the conductivity of 3.47 x 10(-6) S m(-1) and the work function of 5.22 eV. TiO2@-GDY nanocomposite is then prepared and presented with an enhancement of photocatalytic ability compared to pure TiO2.", 50 | "output": "{\"Name\": [\"beta-Graphdiyne\"], \"Formula\": [], \"Acronym\": [\"GDY\"], \"Descriptor\": [\"tetraethynylethene as precursor\"], \"Structure/Phase\": [\"2D\", \"thin film\"], \"Application\": [{\"photocatalyst\": []}], \"Property\": [\"'work function': [5.22]\",\"'band gap': [0]\", \"'conductivity': [3.47 x 10(-6)]\", \"'smooth surface': []\", \"'continuous': []\", \"'uniform': []\"], \"Synthesis\": [\"Glaser-Hay coupling reaction \"], \"Characterization\": []} & {\"Name\": [], \"Formula\": [\"TiO2@-GDY\"], \"Acronym\": [], \"Descriptor\": [], \"Structure/Phase\": [], \"Application\": [{}], \"Property\": [\"'photocatalytic ability': []\"], \"Synthesis\": [], \"Characterization\": []}" 51 | }, 52 | { 53 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 54 | "input": "Artificial photosynthesis of hydrocarbon fuels by utilizing solar energy and CO2 is considered as a potential route for solving ever-increasing energy crisis and greenhouse effect. Herein, hierarchical porous O-doped graphitic carbon nitride (g-C3N4) nanotubes (OCN-Tube) are prepared via successive thermal oxidation exfoliation and curling-condensation of bulk g-C3N4. The as-prepared OCN-Tube exhibits hierarchically porous structures, which consist of interconnected multiwalled nanotubes with uniform diameters of 20-30 nm. The hierarchical OCN-Tube shows excellent photocatalytic CO2 reduction performance under visible light, with methanol evolution rate of 0.88 mu mol g(-1) h(-1), which is five times higher than bulk g-C3N4 (0.17 mu mol g(-1) h(-1)). The enhanced photocatalytic activity of OCN-Tube is ascribed to the hierarchical nanotube structure and O-doping effect. The hierarchical nanotube structure endows OCN-Tube with higher specific surface area, greater light utilization efficiency, and improved molecular diffusion kinetics, due to the more exposed active edges and multiple light reflection/scattering channels. The O-doping optimizes the band structure of g-C3N4, resulting in narrower bandgap, greater CO2 affinity, and uptake capacity as well as higher separation efficiency of photogenerated charge carriers. This work provides a novel strategy to design hierarchical g-C3N4 nanostructures, which can be used as promising photocatalyst for solar energy conversion.", 55 | "output": "{\"Name\": [\"graphitic carbon nitride\"], \"Formula\": [\"g-C3N4\"], \"Acronym\": [], \"Descriptor\": [\"O-doped\"], \"Structure/Phase\": [\"nanotubes\"], \"Application\": [{\"CO2 reduction\": []}], \"Property\": [\"'specific surface area': []\", \"'light utilization efficiency': []\", \"'molecular diffusion kinetics': []\", \"'diameter': [20-30]\", \"'methanol evolution rate': [0.88]\", \"'photocatalyst': []\", \"'bandgap': []\", \"'CO2 affinity': []\", \"'CO2 uptake capacity': []\", \"'separation efficiency of photogenerated charge': []\"], \"Synthesis\": [\"thermal oxidation exfoliation\", \"curling-condensation\"], \"Characterization\": []}" 56 | }, 57 | { 58 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 59 | "input": "CdS/g-C3N4 core/shell nanowires with different g-C3N4 contents were fabricated by a combined solvothermal and chemisorption method and characterized by X-ray powder diffraction, scanning electronic microscopy, transmission electron microscopy, and UV-vis diffuse reflection spectroscopy. The photocatalytic hydrogen-production activities of these samples were evaluated using Na2S and Na2SO3 as sacrificial reagents in water under visible-light illumination (lambda >= 420 nm). The results show that after a spontaneous adsorption process g-C3N4 is successfully coated on CdS nanowires with intimate contact and can significantly improve the photocatalytic hydrogen-production rate of CdS nanowires, which reaches an optimal value of up to 4152 mu mol h(-1) g(-1) at the g-C3N4 content of 2 wt %. More importantly, g-C3N4 coating can substantially reinforce the photostability of CdS nanowires even in a nonsacrificial system. The synergic effect between g-C3N4 and CdS, which can effectively accelerate the charge separation and transfer corrosive holes from CdS to robust C3N4, was proposed to be responsible for the enhancement of the photocatalytic activity and photostability. The possible conditions necessary for the synergic effect to work in a CdS/g-C3N4 core/shell configuration is also discussed.", 60 | "output": "{\"Name\": [], \"Formula\": [\"CdS/g-C3N4\"], \"Acronym\": [], \"Descriptor\": [], \"Structure/Phase\": [\"core/shell nanowires\"], \"Application\": [{\" hydrogen production\": [4152]}], \"Property\": [\"'photocatalytic activity': []\", \"'photostability': []\", \"'separation efficiency of photogenerated charge': []\"], \"Synthesis\": [\"hydrothermal reaction\", \"chemisorption\"], \"Characterization\": [\"X-ray diffraction\",\" scanning electronic microscopy\",\"transmission electron microscopy\",\"UV-vis diffuse reflection spectroscopy\"]}" 61 | }, 62 | { 63 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 64 | "input": "In situ sulfur-doped mesoporous g-C3N4 (mpgCNS) was synthesized from a simple organosulfur compound, thiourea, using SiO2 nanoparticles as the hard template. The resultant product has a high surface area of 128 m(2) g(-1) and mesopores in the range of 10-20 nm. Based on X-ray photoelectron spectroscopy analysis, the doped sulfur was proposed to substitute carbon in mpgCNS and a downshift of 0.25 eV was resulted in its conduction band. Optical studies indicated that mpgCNS exhibits enhanced and extended light absorbance in the visible light region and a much lower density of defects compared to the native g-C3N4. As a result, mpgCNS has been found to be 30 times more active than the native g-C3N4 for hydrogen evolution from photocatalytic water splitting. A high quantum efficiency of 5.8% at 440 nm was obtained which is among the highest for carbon nitride photocatalysts.", 65 | "output": "{\"Name\": [], \"Formula\": [\"g-C3N4\"], \"Acronym\": [], \"Descriptor\": [\"sulfur-doped\"], \"Structure/Phase\": [\"mesoporous\"], \"Application\": [{\"photocatalysts\": [], \"hydrogen evolution\": []}], \"Property\": [\"'specific surface area': [128]\", \"'diameter': [10~20]\",\"'downshift': [0.25]\",\"'quantum efficiency': [0.058]\"], \"Synthesis\": [\"template synthesis \"], \"Characterization\": [\"X-ray photoelectron spectroscopy\"]}" 66 | }, 67 | { 68 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 69 | "input": "High ionic conductivity solid polymer electrolyte (SPE) has long been desired for the next generation high energy and safe rechargeable lithium batteries. Among all of the SPEs, composite polymer electrolyte (CPE) with ceramic fillers has garnered great interest due to the enhancement of ionic conductivity. However, the high degree of polymer crystallinity, agglomeration of ceramic fillers, and weak polymer-ceramic interaction limit the further improvement of ionic conductivity. Different from the existing methods of blending preformed ceramic particles with polymers, here we introduce an in situ synthesis of ceramic filler particles in polymer electrolyte. Much stronger chemical/mechanical interactions between monodispersed 12 nm diameter SiO2 nanospheres and poly(ethylene oxide) (PEO) chains were produced by in situ hydrolysis, which significantly suppresses the crystallization of PEO and thus facilitates polymer segmental motion for ionic conduction. In addition, an improved degree of LiClO4 dissociation can also be achieved. All of these lead to good ionic conductivity (1.2 x 10(-3) cm(-1) at 60 degrees C, 4.4 x 10(-5) cm(-1) at 30 degrees C). At the same time, largely extended electrochemical stability window up to 5.5 V can be observed. We further demonstrated all-solid-state lithium batteries showing excellent rate capability as well as good cycling performance.", 70 | "output": "{\"Name\": [], \"Formula\": [\"SiO2\"], \"Acronym\": [], \"Descriptor\": [\"monodisperse\"], \"Structure/Phase\": [\"nanospheres\"], \"Application\": [{}], \"Property\": [\"'diameter': [12]\"], \"Synthesis\": [\"hydrolysis\"], \"Characterization\": []} & {\"Name\": [\"composite polymer electrolyte)\"], \"Formula\": [], \"Acronym\": [\"CPE\"], \"Descriptor\": [\"ceramic filler\", \"poly(ethylene oxide) chains\"], \"Structure/Phase\": [], \"Application\": [{\"all-solid-state lithium batteries\": [\"'rate capability': []\", \"'cycling': []\"]}], \"Property\": [\"'ionic conductivity': [1.2 x 10(-3)]\",\"'electrochemical stability window': [5.5]\"], \"Synthesis\": [\"in situ synthesis\", \"in situ hydrolysis\"], \"Characterization\": []}" 71 | }, 72 | { 73 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 74 | "input": "In this work, a microwave approach is developed to rapidly synthesize ultralong zinc pyrovanadate (Zn3V2O7(OH)(2)2H(2)O, ZVO) nanowires with a porous crystal framework. It is shown that our synthesis strategy can easily be extended to fabricate other metal pyrovanadate compounds. The zinc pyrovanadate nanowires show significantly improved electrochemical performance when used as intercalation cathode for aqueous zinc-ion battery. Specifically, the ZVO cathode delivers high capacities of 213 and 76 mA h g(-1) at current densities of 50 and 3000 mA g(-1), respectively. Furthermore, the Zn//ZVO cells show good cycling stability up to 300 cycles. The estimated energy density of this Zn cell is approximate to 214Wh kg(-1), which is much higher than commercial lead-acid batteries. Significant insight into the Zn-storage mechanism in the pyrovanadate cathodes is presented using multiple analytical methods. In addition, it is shown that our prototype device can power a 1.5 V temperature sensor for at least 24 h.", 75 | "output": "{\"Name\": [\"zinc pyrovanadate\"], \"Formula\": [\"(Zn3V2O7(OH)(2)2H(2)O\"], \"Acronym\": [\"ZVO\"], \"Descriptor\": [\"ultralong\"], \"Structure/Phase\": [\"nanowires\", \"porous\"], \"Application\": [{\"zinc-ion battery\": [\"'rate capability': [213;0.05] (property)\", \"'cycling performance': [300] (property)\", \"'energy density': [214] (property)\", \"'continuous power supply': [24] (property)\"]}], \"Property\": [], \"Synthesis\": [\"microwave approach\"], \"Characterization\": []}" 76 | }, 77 | { 78 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 79 | "input": "Novel double-shelled CoMn2O4 hollow microcubes with nanometersized building blocks have been prepared by a facile co-precipitation and annealing method. In virtue of the unique structural features, the resultant CoMn2O4 hollow structures exhibit high specific capacity (similar to 830 mA h g(-1) at 200 mA g(-1)) and good cycling performance as an anode material for lithium-ion batteries.", 80 | "output": "{\"Name\": [], \"Formula\": [\"CoMn2O4\"], \"Acronym\": [], \"Descriptor\": [\"anode\"], \"Structure/Phase\": [\"double-shelled\", \"hollow microcubes\"], \"Application\": [{\"lithium-ion batteries\": [\"'cycling performance': [] (property)\"]}], \"Property\": [\"'specific capacity':[830;200]\"], \"Synthesis\": [\"co-precipitation\", \"annealing\"], \"Characterization\": []}" 81 | }, 82 | { 83 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 84 | "input": "Heteroatom doping has always been a common modification method for lithium-ion batteries (LIBs) cathodes. The addition of guest heteroatoms to the material can adjust the microstructure of the host material, and enlarge the lattice spacing or form the vacancy defect of oxygen, which can effectively im-prove the Electrochemical performance of materials. Herein, CO2-modified VO2(B) nanorods were syn-thesized via a simple one-step hydrothermal method. Thanks to the extended layer spacing in VO2(B) that results from CO2 small molecules, the synthesized CO2-VO2 electrode exhibits a high initial capacity of 251.1 mA g-1 at 100 mA g-1, and a capacity retention of 81.8 % after 80 cycles. It is worth noting that the CO2- VO2 electrode also has a high rate capacity (170.0 mAh g-1 at 1 A g-1), and the capacity recovery rate can reach 84.5 % even under the condition of variable current. Further studies showed that the role of CO2 small molecules as a pillar in the interlayer of VO2(B) and a weak electrostatic attraction between OCO2 and Li+ existed, resulting in a robust crystal structure and convenient Li-ion diffusion conditions during the (de) intercalation of lithium ions in VO2(B). This study opens the way for further exploration of the application of small molecule doping in the field of lithium-ion batteries.(c) 2022 Elsevier B.V. All rights reserved.", 85 | "output": "{\"Name\": [], \"Formula\": [\"CO2-VO2\"], \"Acronym\": [], \"Descriptor\": [\"heteroatom doping\", \"electrode\"], \"Structure/Phase\": [\"nanorods\"], \"Application\": [{\"lithium-ion batteries\": []}], \"Property\": [\"'capacity retention':[80;0.818]\"], \"Synthesis\": [\"simple one-step hydrothermal method\"], \"Characterization\": []}" 86 | }, 87 | { 88 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 89 | "input": "Graphene-sulfur (G S) hybrid materials with sulfur nanocrystals anchored on interconnected fibrous graphene are obtained by a facile one-pot strategy using a sulfur/carbon disulfide/ alcohol mixed solution. The reduction of graphene oxide and the formation/binding of sulfur nanocrystals were integrated. The G S hybrids exhibit a highly porous network structure constructed by fibrous graphene, many electrically conducting pathways, and easily tunable sulfur content, which can be cut and pressed into pellets to be directly used as lithium sulfur battery cathodes without using a metal current-collector, binder, and conductive additive. The porous network and sulfur nanocrystals enable rapid Ion transport and short Li+ diffusion distance, the Interconnected fibrous graphene provides highly conductive electron transport pathways, and the oxygen-containing (mainly hydroxyl/epoxide) groups show strong binding with polysulfides, preventing their dissolution into the electrolyte based on first-principles calculations. As a result, the G S hybrids show a high capacity, an excellent high-rate performance, and a long life over 100 cycles. These results demonstrate the great potential of this unique hybrid structure as cathodes for high-performance lithium sulfur batteries.", 90 | "output": "{\"Name\": [\"Graphene-sulfur\"], \"Formula\": [], \"Acronym\": [\"G S\"], \"Descriptor\": [\"hybrids\"], \"Structure/Phase\": [\"porous network\"], \"Application\": [{\"lithium sulfur battery\": [\"'capacity': [] (property)\", \"'rate capability': [] (property)\", \"'cycling performance': [100] (property)\"]}], \"Property\": [], \"Synthesis\": [\"solution-processed\"], \"Characterization\": []}" 91 | }, 92 | { 93 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 94 | "input": "Lithium-sulfur batteries hold great promise for serving as next generation high energy density batteries. However, the Shuttle of pdlysulfide induces rapid capacity degradation.aticl poor cycling stability of lithium sulfur tells. Herein, we proposed a unique lithium sulfur battery configuration with an ultrathin graphene oxide (GO) membrane for high stability. The oxygen electronegative atoms modifieci60 into a polar plane, and the carboxyl groups acted as ion-hopping sites of positively charged species (Lit) and rejected the transportation of negatively charged species (S-n(2-)) due to the electrostatic interactions. Such electrostatic repulsion and physical inhibition largely decreased the transferenceof polysulfides across the GO:membrane in the lithium sulfur system. Consequently, the GO membrane with highly tunable functionalizatidn properties, high mechanical strength, low electric conductivity', and facile fabrication procedure is an effective permselective separator system in lithium sulfur batteries. By the incorporation of a=permselective do membrane, the cyclic capacity decay rate is also reduced from 0.49 to 0.23%/cycle. As the GO membrane blocks the diffusion of polystilfides through the membrane, it is also with advantages of antkelfdischarge properties.", 95 | "output": "{\"Name\": [\"graphene oxide\"], \"Formula\": [], \"Acronym\": [\"GO)\"], \"Descriptor\": [\"tunable functionalization\"], \"Structure/Phase\": [\"film\"], \"Application\": [{\"lithium sulfur battery\": [\"'antkelfdischarge': [] (property)\", \"'cyclic capacity decay rate': [0.0049;0.0023] (property)\"]}], \"Property\": [\"'stability': []\", \"'mechanical strength': []\", \"'low electric conductivity': []\"], \"Synthesis\": [], \"Characterization\": []}" 96 | }, 97 | { 98 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 99 | "input": "Replacement of noble-metal platinum catalysts with cheaper, operationally stable, and highly efficient electrocatalysts holds huge potential for large-scale implementation of clean energy devices. Metal-organic frameworks (MOFs) and metal dichalcogenides (MDs) offer rich platforms for design of highly active electrocatalysts owing to their flexibility, ultrahigh surface area, hierarchical pore structures, and high catalytic activity. Herein, an advanced electrocatalyst based on a vertically aligned MoS2 nanosheet encapsulated Mo-N/C framework with interfacial Mo-N coupling centers is reported. The hybrid structure exhibits robust multifunctional electrocatalytic activity and stability toward the hydrogen evolution reaction, oxygen evolution reaction, and oxygen reduction reaction. Interestingly, it further displays high-performance of Zn-air batteries as a cathode electrocatalyst with a high power density of approximate to 196.4 mW cm(-2) and a voltaic efficiency of approximate to 63 % at 5 mA cm(-2), as well as excellent cycling stability even after 48 h at 25 mA cm(-2). Such outstanding electrocatalytic properties stem from the synergistic effect of the distinct chemical composition, the unique three-phase active sites, and the hierarchical pore framework for fast mass transport. This work is expected to inspire the design of advanced and performance-oriented MOF/MD hybrid-based electrocatalysts for wider application in electrochemical energy devices.", 100 | "output": "{\"Name\": [], \"Formula\": [\"MoS2\"], \"Acronym\": [], \"Descriptor\": [\"Mo-N/C framework\"], \"Structure/Phase\": [\"nanosheet\"], \"Application\": [{\"electrocatalysis\": [\"'activity': [] (property)\", \"'stability': [] (property)\"],\"Zn-air batteries\": []}], \"Property\": [\"'stability': [48]\",\"'power density':[196.4]\"], \"Synthesis\": [\"encapsulating\"], \"Characterization\": []}" 101 | }, 102 | { 103 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 104 | "input": "Orientation control of anisotropic one-dimensional (1D) and two-dimensional (2D) materials in solutions is of great importance in many fields ranging from structural materials design, the thermal management, to energy storage. Achieving fine control of vertical alignment of anisotropic fillers (such as graphene, boron nitride (BN), and carbon fiber) remains challenging. This work presents a universal and scalable method for constructing vertically aligned structures of anisotropic fillers in composites assisted by the expansion flow (using 2D BN platelets as a proof-of-concept). BN platelets in the silicone gel strip are oriented in a curved shape that includes vertical alignment in the central area and horizontal alignment close to strip surfaces. Due to the vertical orientation of BN in the central area of strips, a through-plane thermal conductivity as high as 5.65 W m(-1) K-1 was obtained, which can be further improved to 6.54 W m(-1) K-1 by combining BN and pitch-based carbon fibers. The expansion-flow-assisted alignment can be extended to the manufacture of a variety of polymer composites filled with 1D and 2D materials, which can find wide applications in batteries, electronics, and energy storage devices.", 105 | "output": "{\"Name\": [\"boron nitride\"], \"Formula\": [\"BN\"], \"Acronym\": [], \"Descriptor\": [\"vertical alignment\", \"2D\"], \"Structure/Phase\": [\"platelets\"], \"Application\": [{\"batteries\": [], \"electronics\": [], \"energy storage devices\": []}], \"Property\": [\"'thermal conductivity': [5.65]\"], \"Synthesis\": [], \"Characterization\": []}" 106 | }, 107 | { 108 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 109 | "input": "In this work, nitrogen-doped graphene nanosheets serving as lithium storage materials are presented. The nitrogen-doped graphene nanosheets were prepared by heat treatment of graphite oxide under an ammonia atmosphere at 800 degrees C for 2 h. Scanning electron microscopy, transmission electron microscopy, Raman spectroscopy, and X-ray photoelectron spectroscopy were employed to characterize the prepared product as nitrogen-doped graphene nanosheets with a doping level of ca. 2% nitrogen, where the N binding configuration of the graphene includes 57.4% pyridinic, 35.0% pyrrolic and 7.6% graphitic N atoms. Galvanostatic charge/discharge experiments revealed that these nitrogen-doped graphene nanosheets exhibited a high reversible capacity (900 mA h g(-1) at 42 mA g(-1) (1/20 C)), excellent rate performance (250 mA h g(-1) at a current density of 2.1 A g(-1) (2.5 C)), and significantly enhanced cycling stability, which demonstrated nitrogen-doped graphene nanosheets to be a promising candidate for anode materials in high rate lithium-ion batteries.", 110 | "output": "{\"Name\": [\"nitrogen-doped graphene\"], \"Formula\": [], \"Acronym\": [], \"Descriptor\": [\"nitrogen-doped\", \"anode\"], \"Structure/Phase\": [\"nanosheets\"], \"Application\": [{\"lithium-ion battery\": [\"'reversible capacity': [900;-;0.042] (property)\", \"'stability': [] (property)\", \"'rate capability': [250;2.1] (property)\"]}], \"Property\": [\"'doping level': [0.02]\"], \"Synthesis\": [\"heating\"], \"Characterization\": [\"scanning electron microscopy\", \"transmission electron microscropy\", \"raman spectroscopy\", \"X-ray photoelectron spectroscopy\"]}" 111 | }, 112 | { 113 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 114 | "input": "Li-O2 batteries are mainly limited by the poor conductivity of their discharge products as well as parasitic reactions with carbon-containing electrodes and electrolytes. Here, Li-O2 cells utilizing inorganic solid-state electrolytes are investigated as a means to operate at elevated temperature, thereby increasing the conductivity of discharge products. Growth of dense, conductive LixOy products further removes the need for high-surface area support structures commonly made of carbon. Patterned Au electrodes, evaporated onto Li7La3Zr2O12 (LLZO) solid electrolyte, are used to create a triple-phase boundary for the nucleation of the discharge product, with growth outward into the cell headspace with gaseous O2. Through capacity measurements and imaging, discharge product growths are estimated to reach a critical dimension of approximately 10 mu m, far exceeding what would be possible for a conformal film based on its room temperature electronic conductivity. Raman spectroscopy and electrochemical mass spectrometry are used to characterize the discharge chemistry and reveal a mixed lithium oxide character, with evidence of trace lithium hydroxides and initial carbonate contamination. These results showcase that thermal enhancement of Li-O2 batteries could be a viable strategy to increase capacity when paired with solid electrolytes.", 115 | "output": "{\"Name\": [], \"Formula\": [\"Li7La3Zr2O12\"], \"Acronym\": [\"LLZO\"], \"Descriptor\": [\"electrolyte\",\" thermal enhancement\"], \"Structure/Phase\": [], \"Application\": [{\"lithium-ion\": []}], \"Property\": [], \"Synthesis\": [\"hydrothermal method\"], \"Characterization\": [\"Raman spectroscopy\", \"electrochemical mass spectrometry\"]}" 116 | }, 117 | { 118 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 119 | "input": "Binary metal oxides have been regarded as ideal and potential anode materials, which can ameliorate and offset the electrochemical performance of the single metal ()Aides, such as reversible capacity, structural stability and electronic conductivity. In this work, monodisperse NiCo2O4 mesoporous microspheres are fabricated by a facile solvothermal method followed by pyrolysis of the Ni0.33Co0.67CO3 precursor. The Brunauer-Emmett-Teller (BET) surface area of NiCo2O4 mesoporous microspheres is determined to be about 40.58 m(2) g(-1) with dominant pore diameter of 14.5 nm and narrow size distribution of 10-20 nm. Our as-prepared NiCo2O4 products were evaluated as the anode material for the lithium-ion-battery (LIB) application. It is demonstrated that the special structural features of the NiCo2O4 microspheres including uniformity of the surface texture, the integrity and porosity exert significant effect on the electrochemical performances. The discharge capacity of NiCo2O4 microspheres could reach 1198 mA h g(-1) after 30 discharge-charge cycles at a current density of 200 mA g(-1). More importantly, when the current density increased to 800 mk g(-1), it can render reversible capacity of 705 mA h g(-1) even after 500 cycles, indicating its potential applications for next-generation high power lithium ion batteries (LIBs). The superior battery performance is mainly attributed to the unique micro/nanostructure composed of interconnected NiCo2O4 nanocrystals, which provides good electrolyte diffusion and large electrode-electrolyte contact area, and meanwhile reduces volume change during charge/discharge process. The strategy is simple but very effective, and because of its versatility, it could be extended to other high-capacity metal oxide anode materials for LIBs.", 120 | "output": "{\"Name\": [], \"Formula\": [\"NiCo2O4\"], \"Acronym\": [], \"Descriptor\": [\"monodisperse\",\"anode\"], \"Structure/Phase\": [\"mesoporous microspheres\"], \"Application\": [{\"lithium-ion battery\": [\"'reversible capacity': [1198;30;0.2] (property)\", \"'current density ': [800] (property)\", \"'reversible capacity': [500;705] (property)\", \"electrolyte diffusion (descriptor)\", \"electrode-electrolyte contact area (descriptor)\"]}], \"Property\": [\"'specific surface': [40.58]\", \"'diameter': [14.5]\"], \"Synthesis\": [\"solvothermal method\"], \"Characterization\": [\"Brunauer-Emmett-Teller\"]}" 121 | }, 122 | { 123 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 124 | "input": "Low-temperature solution-processed photovoltaics suffer from low efficiencies because of poor exciton or electron-hole diffusion lengths (typically about 10 nanometers). Recent reports of highly efficient CH3NH3PbI3-based solar cells in a broad range of configurations raise a compelling case for understanding the fundamental photophysical mechanisms in these materials. By applying femtosecond transient optical spectroscopy to bilayers that interface this perovskite with either selective-electron or selective-hole extraction materials, we have uncovered concrete evidence of balanced long-range electron-hole diffusion lengths of at least 100 nanometers in solution-processed CH3NH3PbI3. The high photoconversion efficiencies of these systems stem from the comparable optical absorption length and charge-carrier diffusion lengths, transcending the traditional constraints of solution-processed semiconductors.", 125 | "output": "{\"Name\": [], \"Formula\": [\"CH3NH3PbI3\"], \"Acronym\": [], \"Descriptor\": [], \"Structure/Phase\": [\"bilayers\"], \"Application\": [{\"solar cells\": []}], \"Property\": [\"'diffusion length': [0.1]\"], \"Synthesis\": [\"solution-processed\"], \"Characterization\": [\"femtosecond transient absorption spectroscopy\"]}" 126 | }, 127 | { 128 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 129 | "input": "Graphene and graphitic carbon nitride (g-C3N4) composite photocatalysts were prepared by a combined impregnation-chemical reduction strategy involving polymerization of melamine in the presence of graphene oxide (precursors) and hydrazine hydrate (reducing agent), followed by thermal treatment at 550 degrees C under flowing nitrogen. The resulting, graphene/g-C3N4 composite photocatalysts were characterized by X-ray diffraction, transmission electron microscopy, UV-visible spectrophotometry, nitrogen adsorption, X-ray photoelectron spectroscopy, Fourier transform infrared spectroscopy, Raman spectroscopy, and photoluminescence spectroscopy. The transient photocurrent response was measured for several on-off cycles of intermittent irradiation. The effect of graphene content on the rate of visible-light photocatalytic hydrogen production was studied for a series of graphene-graphitic carbon nitride composite samples containing Pt as a cocatalyst in methanol aqueous solutions. This study shows that graphene sheets act as electronic conductive channels to efficiently separate the photogenerated charge carriers and, consequently, to enhance the visible-light photocatalytic H-2-production activity of g-C3N4. The optimal graphene content was determined to be similar to 1.0 wt %, and the corresponding H-2-production rate was 451 mu mol h(-1) g(-1), which exceeded that of pure g-C3N4 by more than 3.07 times. The proposed mechanism for the enhanced visible-light photocatalytic activity of g-C3N4 modified by a 5 mall amount of graphene was further confirmed by photoluminescence spectroscopy and transient photocurrent response. The metal-free graphene/g-C3N4 composites showed high visible-light photocatalytic activity, which makes them promising nanomaterials for further applications in water treatment and dye-sensitized solar cells.", 130 | "output": "{\"Name\": [\"graphene and graphitic carbon nitride composite\"], \"Formula\": [], \"Acronym\": [\"graphene/g-C3N4\"], \"Descriptor\": [\"photocatalysts\"], \"Structure/Phase\": [], \"Application\": [{\"dye-sensitized solar cells\": [],\"visible-light photocatalytic\": [\"'activity': []\"]}], \"Property\": [\"'H-2-production rate':[451]\"], \"Synthesis\": [\"impregnation chemical reduction\"], \"Characterization\": [\"X-ray diffraction\", \"transmission electron microscopy\", \"UV-visible spectrophotometry\", \"nitrogen adsorption\", \"X-ray photoelectron spectroscopy\", \"Fourier transform infrared spectroscopy\", \"Raman spectroscopy\", \"photoluminescence spectroscopy\"]}" 131 | }, 132 | { 133 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 134 | "input": "Metal halide perovskites of the general formula X(3)-where A is a monovalent cation such as caesium, methylammonium or formamidinium; B is divalent lead, tin or germanium; and X is a halide anion-have shown great potential as light harvesters for thin-film photovoltaics(1-5). Among a large number of compositions investigated, the cubic a-phase of formamidinium lead triiodide (FAPbI(3)) hasemerged as the most promising semiconductor for highly efficient and stable perovskite solar cells(6-9), and maximizing the performance of this material in such devices is of vital importance for the perovskite researchcommunity. Here we introduce an anion engineering concept that uses the pseudo-halide anion formate (HCOO-) to suppress anion-vacancy defects that are present at grain boundaries and at the surface of the perovskite films and to augment the crystallinity of the films. Theresulting solar cell devices attain a power conversion efficiency of 25.6 per cent (certified 25.2 per cent), have long-term operational stability (450 hours) and show intense electroluminescence with external quantum efficiencies of more than 10 per cent. Our findings provide a direct route to eliminate the most abundant and deleterious lattice defects present in metal halide perovskites, providing a facile access to solution-processable films with improved optoelectronic performance.", 135 | "output": "{\"Name\": [\"formamidinium lead triiodide\"], \"Formula\": [\"FAPbI(3))\"], \"Acronym\": [], \"Descriptor\": [\"Metal halide perovskites\",\"promising semiconductor\"], \"Structure/Phase\": [\"cubic a-phase\"], \"Application\": [{\"thin-film solar cell\": [], \"perovskite solar cell\": [\"'efficiency': [0.256] (property)\", \"'stability': [450] (property)\", \"'external quantum efficiency': [0.1] (property)\"]}], \"Property\": [], \"Synthesis\": [\"solution-processed\"], \"Characterization\": []}" 136 | }, 137 | { 138 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 139 | "input": "All-inorganic and lead-free cesium tin halides (CsSnX3, X=Cl, Br, I) are highly desirable for substituting the organolead halide perovskite solar cells. However, the poor stability of CsSnX3 perovskites has so far prevented the fabrication of devices that can withstand sustained operation under normal conditions. In this paper, a two-step sequential deposition method is developed to grow high-quality B-gamma-CsSnI3 thin films and their unique phase change in atmosphere is explored in detail. We find the spontaneous oxidative conversion from unstable B-gamma-CsSnI3 to air-stable Cs2SnI6 in air. Allowing the phase conversion of the CsSnI3 film to evolve in ambient air it gives the semiconducting perovskite Cs2SnI6 with a bandgap of 1.48 eV and high absorption coefficient (over 10(5) cm(-1) from 1.7 eV). More importantly, the Cs2SnI6 film, for the first time, is adopted as a light absorber layer for a lead-free perovskite solar cell and a preliminary estimate of the power conversion efficiency (PCE) about 1% with open-circuit voltage of 0.51 V and short-circuit current of 5.41 mA/cm(2) is realized by optimizing the perovskite absorber thickness. According to the bandgap and the Shockley-Queisser limit, such inorganic perovskite solar cell with higher efficiency and pronounced stability can be expected by material quality improvement and device engineering. (C) 2016 Elsevier B.V. All rights reserved.", 140 | "output": "{\"Name\": [\"inorganic and lead-free cesium tin halides\"], \"Formula\": [\"CsSnX3, X=Cl, Br, I\"], \"Acronym\": [], \"Descriptor\": [], \"Structure/Phase\": [\"perovskites\"], \"Application\": [{}], \"Property\": [], \"Synthesis\": [], \"Characterization\": []} & {\"Name\": [], \"Formula\": [\"Cs2SnI6\"], \"Acronym\": [], \"Descriptor\": [\"light absorber\"], \"Structure/Phase\": [\"B-gamma\"], \"Application\": [{\"Perovskite solar cell\": [\"'efficiency ': [0.01] (property)\", \"'open-circuit voltage ': [0.51] (property)\", \"'short-circuit current': [5.41] (property)\"]}], \"Property\": [\"'band energy ': [1.48]\", \"'absorption coefficient': [1000000]\"], \"Synthesis\": [\"two-step deposition\"], \"Characterization\": []}" 141 | }, 142 | { 143 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 144 | "input": "Metal halides perovskites, such as hybrid organic-inorganic CH3NH3PbI3, are newcomer optoelectronic materials that have attracted enormous attention as solution-deposited absorbing layers in solar cells with power conversion efficiencies reaching 20%. Herein we demonstrate a new avenue for halide perovskites by designing highly luminescent perovskite-based colloidal quantum dot materials. We have synthesized monodisperse colloidal nanocubes (4-15 nm edge lengths) of fully inorganic cesium lead halide perovskites (CsPbX3, X = Cl, Br, and I or mixed halide systems Cl/Br and Br/I) using inexpensive commercial precursors. Through compositional modulations and quantum size-effects, the bandgap energies and emission spectra are readily tunable over the entire visible spectral region of 410-700 nm. The photoluminescence of CsPbX3 nanocrystals is characterized by narrow emission line-widths of 12-42 nm, wide color gamut covering up to 140% of the NTSC color standard, high quantum yields of up to 90%, and radiative lifetimes in the range of 1-29 ns. The compelling combination of enhanced optical properties and chemical robustness makes CsPbX3 nanocrystals appealing for optoelectronic applications, particularly for blue and green spectral regions (410-530 nm), where typical metal chalcogenide-based quantum dots suffer from photodegradation.", 145 | "output": "{\"Name\": [], \"Formula\": [\"CH3NH3PbI3\"], \"Acronym\": [], \"Descriptor\": [\"optoelectronic\", \"absorber\"], \"Structure/Phase\": [\"perovskites,\"], \"Application\": [{\"solar cell\": [\"'efficiency': [0.2] (property)\"]}], \"Property\": [\"'efficiency': [0.2]\"], \"Synthesis\": [\"sol-gel\"], \"Characterization\": []} & {\"Name\": [], \"Formula\": [\"CsPbX3 X = Cl, Br, I \"], \"Acronym\": [], \"Descriptor\": [\"compositional modulation\", \"quantum size effect\"], \"Structure/Phase\": [\"perovskites\"], \"Application\": [{}], \"Property\": [[\"'nanocube edge length': [4]\", \"'nanocube edge length': [15]\"], [\"'narrow emission line width': [12]\", \"'narrow emission line width': [42]\"], \"'color gamut coverage': [1.4]\", [\"'radiative lifetime': [1]\", \"'radiative lifetime': [29]\"], \"'bandgap energy': []\", \"'emission spectrum': []\", \"'quantum yield': [0.9]\"], \"Synthesis\": [], \"Characterization\": []}" 146 | }, 147 | { 148 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 149 | "input": "We report on solid-state mesoscopic heterojunction solar cells employing nanoparticles (NPs) of methyl ammonium lead iodide (CH3NH3)PbI3 as light harvesters. The perovskite NPs were produced by reaction of methylammonium iodide with PbI2 and deposited onto a submicron-thick mesoscopic TiO2 film, whose pores were infiltrated with the hole-conductor spiro-MeOTAD. Illumination with standard AM-1.5 sunlight generated large photocurrents (J(SC)) exceeding 17 mA/cm(2), an open circuit photovoltage (V-OC) of 0.888 V and a fill factor (FF) of 0.62 yielding a power conversion efficiency (PCE) of 9.7%, the highest reported to date for such cells. Femto second laser studies combined with photo-induced absorption measurements showed charge separation to proceed via hole injection from the excited (CH3NH3)PbI3 NPs into the spiro-MeOTAD followed by electron transfer to the mesoscopic TiO2 film. The use of a solid hole conductor dramatically improved the device stability compared to (CH3NH3)PbI3 -sensitized liquid junction cells.", 150 | "output": "{\"Name\": [\"methyl ammonium lead iodide\"], \"Formula\": [\"(CH3NH3)PbI3\"], \"Acronym\": [], \"Descriptor\": [\"light harvester\", \"stability\"], \"Structure/Phase\": [\"nanoparticles\"], \"Application\": [{\"heterojunction solar cell\": [\"'short-circuit current density': [17] (property)\", \"'open-circuit voltage': [0.888] (property)\", \"'fill factor': [0.62] (property)\", \"'efficiency': [0.097] (property)\"]}], \"Property\": [], \"Synthesis\": [], \"Characterization\": [\"femtosecond laser\", \"photo-induced absorption spectroscopy\"]} & {\"Name\": [\"spiro-MeOTAD\"], \"Formula\": [], \"Acronym\": [], \"Descriptor\": [\"hole conductor\"], \"Structure/Phase\": [], \"Application\": [{}], \"Property\": [], \"Synthesis\": [], \"Characterization\": []}" 151 | }, 152 | { 153 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 154 | "input": "Transparent electrodes, indespensible in displays and solar cells, are currently dominated by indium tin oxide (ITO) Films although the high price of indium, brittleness of Films, and high vacuum deposition are limiting their applications Recently, solution-processed networks of nanostructures such as carbon nanotubes (CNTs), graphene, and silver nanowires have attracted great attention as replacements A low junction resistance between nanostructures is important for decreasing the sheet resistance However, the junction resistances between CNTs and boundry resistances between graphene nanostructures are too high The aspect ratios of silver nanowires are limited to similar to 100, and silver is relatively expensive Here, we show high-performance transparent electrodes with copper nanofiber networks by a low-cost and scalable electrospinning process Copper nanofibers have ultrahigh aspect ratios of up to 100000 and fused crossing points with ultralow junction resistances, which result in high transmitance at low sheet resistance, e g, 90% at 50 Omega/sq The copper nanofiber networks also show great flexibility and stretchabilty Organic solar cells using copper nanowire networks as transparent electrodes have a power efficiency of 30%, comparable to devices made with ITO electrodes", 155 | "output": "{\"Name\": [\"indium tin oxide\"], \"Formula\": [\"ITO\"], \"Acronym\": [], \"Descriptor\": [\"transparent electrodes\"], \"Structure/Phase\": [\"film\"], \"Application\": [{\"solar cell\": [], \"display\": []}], \"Property\": [\"'brittleness': []\", \"'vacuum deposition': []\"], \"Synthesis\": [], \"Characterization\": []} & {\"Name\": [\"copper\"], \"Formula\": [], \"Acronym\": [], \"Descriptor\": [\"electrode\"], \"Structure/Phase\": [\"nanofiber networks\"], \"Application\": [{\"solar cell\": [\"'effciency': [0.3] (property)\"]}], \"Property\": [\"'aspect ratio': [100000]\", \"'sheet resistance': [0.9]\", \"'transparent': []\", \"'flexible': []\", \"'stretchabilty': []\"], \"Synthesis\": [\"electrospinning process\"], \"Characterization\": []}" 156 | }, 157 | { 158 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 159 | "input": "Organic-inorganic perovskites have shown promise as high-performance absorbers in solar cells, first as a coating on a mesoporous metal oxide scaffold and more recently as a solid layer in planar heterojunction architectures. Here, we report transient absorption and photoluminescence-quenching measurements to determine the electron-hole diffusion lengths, diffusion constants, and lifetimes in mixed halide (CH3NH3PbI3-xClx) and triiodide (CH3NH3PbI3) perovskite absorbers. We found that the diffusion lengths are greater than 1 micrometer in the mixed halide perovskite, which is an order of magnitude greater than the absorption depth. In contrast, the triiodide absorber has electron-hole diffusion lengths of similar to 100 nanometers. These results justify the high efficiency of planar heterojunction perovskite solar cells and identify a critical parameter to optimize for future perovskite absorber development.", 160 | "output": "{\"Name\": [\"indium tin oxide\"], \"Formula\": [\"ITO\"], \"Acronym\": [], \"Descriptor\": [\"absorber\"], \"Structure/Phase\": [\"perovskites\"], \"Application\": [{\"solar cell\": [], \"heterojunction perovskite solar cell\": []}], \"Property\": [\"'diffusion length': [1]\", \"'diffusion constant': []\", \"'lifetime': []\"], \"Synthesis\": [], \"Characterization\": [\"photoluminescence quenching\"]} & {\"Name\": [], \"Formula\": [\"CH3NH3PbI3)\"], \"Acronym\": [], \"Descriptor\": [\"absorber\"], \"Structure/Phase\": [\"perovskites\"], \"Application\": [{\"solar cell\": [], \"heterojunction perovskite solar cell\": []}], \"Property\": [\"'diffusion length': [1]\", \"'diffusion constant': []\", \"'lifetime': []\"], \"Synthesis\": [], \"Characterization\": [\"photoluminescence quenching\"]}" 161 | }, 162 | { 163 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 164 | "input": "Because of their high-efficiency interfacial charge-transfer characteristics of heterojunction nanocomposites have sparked widespread interest in solar cell field. The present study reports the sol-gel electrospinning assisted Bi5FeTi3O15 (BFTO)/g-C3N4 (CN) heterojuction nanocomposites as counter electrodes (CEs) for dye-sensitized solar cells (DSSCs). XRD and TEM studies reveals that BFTO has orthorhombic crystal structure and nanotubes with diameter of 50-100 nm have been decorated on the CN nanosheets. The obvious red-shift in absorption and strong blue and green florescence has been found in the UV and PL studies. Furthermore, the BFTO/CN nanohybrid as-fabricated exhibited a synergetic effect between Bi5FeTi3O15 nanotubes and a g-C3N4 nanosheets, which could influence the electrocatalytic behavior against the triiodide reaction. The photovoltaic efficiency of the BFTO/CN CE is 11.2%, which is better than the typical Pt CE (7.6%) and bare BFTO CE (4.9%). As a result, the CE made of BFTO/CN nanohybrid material is a promising low-cost Pt-free counter electrode for DSSC.", 165 | "output": "{\"Name\": [], \"Formula\": [\"Bi5FeTi3O15\"], \"Acronym\": [\"BFTO\"], \"Descriptor\": [], \"Structure/Phase\": [\"orthorhombic crystal\", \"nanotube\"], \"Application\": [{}], \"Property\": [\"'diameter': [50;100]\"], \"Synthesis\": [], \"Characterization\": [\"X-ray diffraction\", \"transmission electron microscope\", \"photoluminescence\", \"ultraviolet–visible spectroscopy\"]} & {\"Name\": [\"Bi5FeTi3O15/g-C3N4\"], \"Formula\": [], \"Acronym\": [\"BFTO/CN\"], \"Descriptor\": [\"composite\"], \"Structure/Phase\": [\"heterojuction nanocomposites\"], \"Application\": [{\"dye-sensitized solar cell\": [\"'photovoltaic efficiency': [0.112] (property)\"], \"counter electrodes\": []}], \"Property\": [], \"Synthesis\": [\"sol-gel electrospinning\"], \"Characterization\": []}" 166 | }, 167 | { 168 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 169 | "input": "Direct methanol fuel cells (DMFCs) are an attractive portable energy technology due to their low operating temperatures, high-energy conversion efficiency, and lower pollutant production. For over a half-century, the default electrocatalyst for DMFCs has been platinum (Pt). The barriers to widespread deployment of DMFCs, however, are largely linked to the cost of this precious metal and its propensity to become poisoned in its role as the anode catalyst for the oxidation of methanol. Bimetallic platinumgold (PtAu) catalysts, however, offer superior activity for the oxidation of methanol and can operate with increased electrocatalytic stability and resistance to poisoning. Here, we demonstrate a block copolymer template strategy for the preparation of arrays of clusters of PtAu nanocatalysts with tailored composition, particle density, and electrochemical activity. A detailed characterization by XPS, TEM, EDX, and electrochemistry was used to assign a coreshell nanostructure to the 3 nm PtAu nanocatalysts that constitute the clusters. The activity Pt-rich coreshell PtAu nanocatalysts for the electrocatalytic oxidation of methanol was approximately 24-fold that of a current Pt benchmark catalyst (ETEK), only 28% less than that of the PtRu bimetallic benchmark catalyst (XC-72R), and, in comparison to these same catalysts, exhibited a 2- to 3-fold increase in its metric for tolerance to carbonaceous poisoning (I-f/I-b ratio).", 170 | "output": "{\"Name\": [\"platinumgold\"], \"Formula\": [], \"Acronym\": [\"PtAu\"], \"Descriptor\": [\"catalyst\"], \"Structure/Phase\": [\"nanocluster\", \"core shell\"], \"Application\": [{\"Direct methanol fuel cells (DMFCs)\": [\"'efficiency': [] (property)\", \"'tolerance': [2;3] (property)\", \"'activity': [] (property)\"]}], \"Property\": [\"'composition': []\", \"'density': []\", \"'activity': []\"], \"Synthesis\": [\"block copolymer template strategy\"], \"Characterization\": [\"X-ray photoelectron spectroscopy\", \"transmission electron microscope\", \"energy dispersive X-ray spectroscopy\", \"electrochemistry\"]}" 171 | }, 172 | { 173 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 174 | "input": "NiO/graphene nanocomposites are fabricated via a solvothermal method. Scanning and transmission electron microscopy results indicate that the NiO nanoplates (length, similar to 100 nm) were homogeneously distributed on the graphene sheets. The electrochemical properties of the samples as active cathode catalysts for rechargeable Li-air batteries are evaluated by constant current charge-discharge cycling. The composites exhibit a reversible capacity of 1160 mAh g(-1) after 50 cycles at a discharge current density of 50 mA g(-1); this reverse capacity is much higher than that of pure NiO nanoplates (30 mAh g(-1)). Using graphene as a conductive matrix, a homogeneous distribution of NiO nanoplates is accomplished and graphene serves as a framework for loading as produced Li2O2 during the discharge process, resulting in the excellent electrochemical performance of the composites. The mesoporous structure of the NiO nanoplates is suitable for the transfer of O-2 and deposition of Li2O2 produced by the electrochemical reaction. NiO/graphene nanocomposites are a candidate material for high-capacity, low-cost, and nontoxic cathode catalysts in rechargeable Li-air batteries. (C) 2014 Elsevier B.V. All rights reserved.", 175 | "output": "{\"Name\": [], \"Formula\": [\"NiO/graphene\"], \"Acronym\": [], \"Descriptor\": [\"cathode catalysts\"], \"Structure/Phase\": [\"nanocomposites \", \"nanoplate\"], \"Application\": [{\"Li-air batteries\": [\"'capacity': [1160;-;0.05] (property)\", \"'capacity': [1160;50;-] (property)\"]}], \"Property\": [\"'length': [100]\", \"'nontoxic': []\"], \"Synthesis\": [\"solvothermal\"], \"Characterization\": [\"scanning electron microscropy\", \"transmission electron microscropy\", \"X-ray diffraction\"]}" 176 | }, 177 | { 178 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 179 | "input": "Nickel(II) oxide (NiO) is an important wide gap p-type semiconductor used as a hole transport material for dye sensitized solar cells and as a water oxidation electrocatalyst. Here we demonstrate that nanocrystals of the material have increased p-type character and improved photocatalytic activity for hydrogen evolution from water in the presence of methanol as sacrificial electron donor. NiO nanocrystals were synthesized by hydrolysis of Ni(II) nitrate under hydrothermal conditions followed by calcination in air. The crystals have the rock salt structure type and adopt a plate-like morphology (50-90 nm x 10-15 nm). Diffuse reflectance absorbance spectra indicate a band gap of 3.45 eV, similar to bulk NiO. Photoelectrochemical measurements were performed at neutral pH with methylviologen as electron acceptor, revealing photo-onset potentials (Fermi energies) of 0.2 and 0.05 eV (NHE) for nanoscale and bulk NiO, respectively. Nano-NiO and NiO-Pt composites obtained by photodepositon of H2PtC16 catalyze hydrogen evolution from aqueous methanol at rates of 0.8 and 4.5 mu mol H-2 h(-1), respectively, compared to 0.5 and 2.1, mu mol H-2 h(-1) for bulk-NiO and NiO-Pt (20 mg of catalyst, 300 W Xe lamp). Surface photovoltage spectroscopy of NiO and NiO Pt films on Au substrates indicate a metal Pt-NiO junction with 30 mV photovoltage that promotes carrier separation. The increased photocatalytic and photoelectrochemical performance of nano-NiO is due to improved minority carrier extraction and increased p-type character, as deduced from Mott Schottky plots, optical absorbance, and X-ray photoelectron spectroscopy data.", 180 | "output": "{\"Name\": [\"Nickel(II) oxide\"], \"Formula\": [], \"Acronym\": [\"NiO\"], \"Descriptor\": [\"electrocatalyst\"], \"Structure/Phase\": [\"nanocrystal\"], \"Application\": [{\"dye sensitized solar cell\": [], \"water oxidation electrocatalyst\": []}], \"Property\": [\"'bandgap energy': [3.45]\"], \"Synthesis\": [\"hydrothermal\"], \"Characterization\": [\"Mott Schottky\", \"X-ray photoelectron spectroscopy\"]}" 181 | }, 182 | { 183 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 184 | "input": "Nitrogen-doped (N-doped) single-wall carbon nanotubes (SWCNTs) were synthesized by chemical vapor deposition using SiOx nanoparticles as a catalyst and ethylenediamine as the source of both carbon and nitrogen. The N-doped SWCNTs have a mean diameter of 1.1 nm and a narrow diameter range, with 92% of them having diameters from 0.7 to 1.4 nm. Multi-wavelength laser Raman spectra and temperature-dependent electrical resistance indicate that the SWCNT sample is enriched with metallic nanotubes. These N-doped SWCNTs showed excellent electrocatalytic activity for the oxygen reduction reaction and highly selective and sensitive sensing ability for dopamine detection.", 185 | "output": "{\"Name\": [\"single - wall carbon nanotubes\"], \"Formula\": [], \"Acronym\": [\"SWCNTs\"], \"Descriptor\": [\"N-doped\", \"catalyst\"], \"Structure/Phase\": [\"nanotube\"], \"Application\": [{\"sensor\": []}], \"Property\": [\"'diameter': [0.7;1.4]\"], \"Synthesis\": [\"chemical vapour deposition\"], \"Characterization\": [\"Raman spectra\", \"electrical resistance measurement\"]}" 186 | }, 187 | { 188 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 189 | "input": "Photoconducting properties of indium tin oxide (ITO) nanowires grown by RF sputtering associated with annealing process were studied. ITO nanowires have been grown without the use of catalysts or oblique deposition. The cubic cross-sectional nanowires length was of the order of several microns, while their diameter was similar to 150-250 nm. The photoluminescence (PL) analysis on nanowires proved their excellent photoemission characteristics. Devices based on ITO nanowires showed a substantial increase in conductance of up to three orders of magnitude upon exposure to UV light and showing reproducible UV photoresponse and remaining relatively stable. The rising speed is slightly reduced, while the decay time is prolonged. Such devices also exhibited short response times and significant shifts in the threshold gate voltage. It was found that the dynamic response of the ITO nanowires phototransistor was stable with an on/off current contrast ratio of around 10(1). It is thus found that the change in the carrier concentrations in constant gate voltage was enhanced by 1.12 x 10(17) cm(-3) for the 350 nm UV light, corresponding to threshold-voltage shifts of before to after UV-light exposition. The photoconductive gain corresponding to responsivity measured at 350 nm was 0.11 x 10(5).", 190 | "output": "{\"Name\": [\"indium tin oxide\"], \"Formula\": [], \"Acronym\": [\"ITO\"], \"Descriptor\": [], \"Structure/Phase\": [\"nanowire\"], \"Application\": [{}], \"Property\": [\"'length': []\", \"'diameter': []\", \"'carrier concentration': []\", \"'threshold gate voltage': []\", \"'dynamic response': [1000]\"], \"Synthesis\": [\"magnetron sputtering\", \"annealing\"], \"Characterization\": [\"Photoluminescence\"]}" 191 | }, 192 | { 193 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 194 | "input": "Promotion of the catalytic efficiency and reduction of the usage amount of Pd are crucial to developing effective and low-cost catalysts for catalytic reduction of organic aromatic nitro pollutant compound at present. Herein, we report a hybrid material of ultrafine Pd nanocrystals (PdNCs) grown on an excellent support of SnO2-decorated graphene nanosheets (SnO2-GNS) for a highly efficient reduction of a representative aromatic nitro compound (4-nitrophenol). The supporting material of SnO2-GNS was prepared simply by one step of homogeneous reaction with a positively charged polymer as stabilizer and active site for absorption of the PdNCs precursor. Transmittance electronic images demonstrate that the PdNCs are densely and well covered on the SnO2-GNS with a uniform size of 3.4 nm. This nanohybrid exhibits the fastest reduction time (4 minutes) compared to other controlled materials. Moreover, it shows high kinetic responses with an apparent kinetic rate constant (k(app)) of 2.03 x 10(-2) s(-1) and turnover frequency (TOF) of 1.70 s(-1). The cycle performance (10 times) experiments demonstrate that this nanohybrid also displays a good anti-poisoning capability. Thanks to the ultrafine PdNCs, this as-prepared PdNCs/SnO2-GNS nanohybrid may have broad potential in other catalytic fields, for example, organic synthesis, fuel cells and electrochemical biosensors.", 195 | "output": "{\"Name\": [], \"Formula\": [\"PdNCs/SnO2-GNS\"], \"Acronym\": [], \"Descriptor\": [\"low-cost catalysts\"], \"Structure/Phase\": [\"nanocrystal\"], \"Application\": [{\"organic aromatic nitro pollutant compound\": [\"'high kinetic responses': [] (property)\", \"'hanti-poisoning capability': [] (property)\"], \"fuel cells\": [], \"electrochemical biosensors\": []}], \"Property\": [\"'average particle size': [3.4]\"], \"Synthesis\": [\"one step\"], \"Characterization\": [\"transmission electron microscope\"]}" 196 | }, 197 | { 198 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 199 | "input": "The study reports a facile method for the green synthesis of copper oxide nanoparticles (CuO Nps) by a solution combustion method using Tinospora cordifolia water extract. The Nps were characterized by XRD, SEM, TEM and UV-visible studies. XRD data indicates the formation of pure monoclinic crystallite structures of CuO Nps. SEM images show that the particles have sponge like structure with large surface area and the average crystallite sizes were found to be similar to 6-8 nm. These observations were confirmed by TEM analysis. Photocatalytic activity studies of CuO Nps reveal that they act as very good catalyst for the effective degradation of methylene blue (MB) in the presence of IN and Sun light. Also, the degradation of MB was found to be pH dependent. The Nps found to inhibit the activity of 1,1-Diphenyl-2-picrylhydrazyl (DPPH) free radicals effectively. CuO Nps exhibit significant bactericidal activity against Klebsiella aerogenes, Pseudomonas aeruginosa, Escherichia coli and Staphylococcus aureus. The study reveals a simple, ecofriendly and robust method for the synthesis of multifunctional CuO nanoparticle employing under-utilized medicinal plants. (C) 2015 Elsevier Ltd. All rights reserved.", 200 | "output": "{\"Name\": [\"copper oxide nanoparticles\"], \"Formula\": [], \"Acronym\": [\"CuO Nps\"], \"Descriptor\": [\"catalyst\"], \"Structure/Phase\": [\"nanoparticle\", \"sponge like\", \"monoclinic\"], \"Application\": [{\"degradation\": []}], \"Property\": [\"'crystallite size': [6;8]\", \"'specific area': []\", \"'bactericidal activity': []\"], \"Synthesis\": [\"solution combustion\"], \"Characterization\": [\"scanning electron microscope\", \"X-ray diffraction\", \"UV-vis absorption spectroscopy\", \"transmission electron microscope\"]}" 201 | }, 202 | { 203 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 204 | "input": "A SiO2 coat was prepared on an HP40 steel surface via atmospheric-pressure chemical vapour deposition (APCVD). The composition and surface appearance of the coat were then investigated with the aid of energy dispersive spectrometry and scanning electron microscopy, and its anti-coking ability was analysed by virtue of anti-coking mechanism and an anti-coking test. A 5 mu m-thick SiO2 coat was prepared through APCVD at 800 degrees C, with tetraethyl orthosilicate as the source material and air as the carrier and diluent gas. The prepared SiO2 coat can effectively prevent catalyst coking (with an anti-coking rate above 70%) under pyrolysis conditions and can improve the anti-coking ability of HP40 steel.", 205 | "output": "{\"Name\": [], \"Formula\": [\"SiO2\"], \"Acronym\": [], \"Descriptor\": [\"coat\"], \"Structure/Phase\": [], \"Application\": [{\"HP40 steel\": [\"'anti-coking rate': [70] (property)\"]}], \"Property\": [\"'thickness': [5]\"], \"Synthesis\": [\"chemical vapour deposition\"], \"Characterization\": [\"scanning electron microscopy\"]}" 206 | }, 207 | { 208 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 209 | "input": "A series of mono dispersed PtxAu100-x alloy nanoparticles (NPs), with x varying from 0 to 100, were prepared by pulsed laser ablation in liquids, using a series of targets that were made by mixing pure Pt and pure Au powders. The structures of PtxAu100-x alloy NPs were assessed by transmission electron microscopy and X-ray diffraction. A face-centered solid solution is obtained over the whole composition range, and the particle size increases from 2.5 to 5.3 nm as x is increased from 0 to 100. The electrocatalytic performances of the PtxAu100-x alloy NPs towards the formic acid oxidation were investigated by cyclic voltammetry and chronoamperometry. On as-prepared PtxAu100-x alloy NPs, oxidation of formic acid occurs through dehydrogenation, while dehydration is the privileged mechanism on as-prepared mixtures of Pt and Au NPs. However, after a series of CV in 0.5 M H2SO4, both types of catalysts are able to oxidize formic acid according to the dehydrogenation pathway. After 600 s of electrolysis, the mass activities of PtxAu100-x alloy NPs is a factor of two larger than that of mixtures of pure Pt and pure Au NPs with the same surface composition, although both types of catalysts display similar activity with respect to the total electrochemically active surface area. (C) 2013 Elsevier B.V. All rights reserved.", 210 | "output":"{\"Name\": [], \"Formula\": [\"PtxAu100-x\"], \"Acronym\": [\"NPs\"], \"Descriptor\": [\"catalyst\"], \"Structure/Phase\": [\"nanoparticle\"], \"Application\": [{\"electrocatalytic\": []}], \"Property\": [\"'active surface area': []\",\"'particle size': [2.5;5.3]\"], \"Synthesis\": [\"laser ablation\"], \"Characterization\": [\"transmission electron microscopy\", \"X-ray diffraction\"]}" 211 | }, 212 | { 213 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 214 | "input": "Aligned nitrogen-doped carbon nanotubes (N-CNTs) containing 6.4-15.7 wt% of nitrogen were synthesized by pyrolysis of 3-ferrocenyl-2-(4-cyanophenyl)acrylonitrile as the catalyst in either acetonitrile or a solution of acetonitrile and ethylbenzoate. For comparison, N-CNTs were synthesized by pyrolysis of 3-ferrocenyl-2-(4-cyanophenyl)acrylonitrile in toluene. The effect of oxygen and the carbon source used during synthesis was investigated. The use of 3-ferrocenyl-2-(4-cyanophenyl)acrylonitrile in acetonitrile as a nitrogen and carbon source selectively yielded mainly N-CNTs, while use of toluene as a carbon source yielded both N-CNTs and carbon spheres. Elemental analysis of the N-CNTs synthesized using both acetonitrile and ethylbenzoate (source of oxygen) indicated that addition of oxygen enhanced the nitrogen content of N-CNTs. This was further supported by results from Raman spectroscopy, X-ray diffraction (XRD), Fourier transform infrared spectroscopy and inverse gas chromatography surface energy analysis. The higher nitrogen-containing N-CNTs were less graphitic and showed a higher base constant (K-b) compared to N-CNTs synthesized without oxygen. Analysis of transmission electron microscopy images showed that the outer diameters of the N-CNTs decreased upon increasing the oxygen composition by mass in the synthesis precursors from 1 to 4 wt% oxygen, the oxygen was derived from ethylbenzoate. In addition, the scanning electron microscopy and XRD revealed that the alignment of N-CNTs increased upon addition of oxygen. Electrical conductivity measurements of N-CNTs showed a negative relationship between the amount of oxygen in the starting materials and the conductivity of N-CNTs.", 215 | "output": "{\"Name\": [\"nitrogen - doped carbon nanotubes\"], \"Formula\": [], \"Acronym\": [\"N-CNTs\"], \"Descriptor\": [\"N drop\"], \"Structure/Phase\": [\"nanotube\"], \"Application\": [{}], \"Property\": [], \"Synthesis\": [\"pyrolyzing\"], \"Characterization\": [\"scanning electron microscope\", \"Fourier transform infrared spectrometry \", \"X-ray diffraction\", \"Raman spectrum\", \"transmission electron microscopy\", \"inverse gas chromatography surface energy analysis\"]}" 216 | }, 217 | { 218 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 219 | "input": "Amorphous agglomerates of carbon nanospheres (CNS) with a diameter range of 10-50 nm were synthesized using the solution combustion method. High-resolution transmission electron microscopy (HRTEM) revealed a densely packed high surface area of SP2-hybridized carbon; however, there were no crystalline structural components, as can be seen from the scanning electron microscopy, HRTEM, X-ray diffraction, Raman spectroscopy, and thermal gravimetric analyses. Electrochemical and thermo catalytic decomposition study results show that the material can be used as a potential electrode candidate for the fabrication of energy storage devices and also for the production of free hydrogen if such devices are used in a fluidized bed reactor loaded with the as-prepared CNS as the catalyst bed.", 220 | "output": "{\"Name\": [\"carbon nanospheres\"], \"Formula\": [], \"Acronym\": [\"CNS\"], \"Descriptor\": [], \"Structure/Phase\": [\"amorphous\", \"nanosphere\"], \"Application\": [{\"flexible energy storage\": []}], \"Property\": [\"'diameter': [10;50]\", \"'specific area': []\"], \"Synthesis\": [\"solution combustion\"], \"Characterization\": [\"thermal gravimetric analysis\", \"X-ray diffraction\", \"Raman spectroscopy\", \"Thermal gravimetric analyses\", \"HRTEM\"]}" 221 | }, 222 | { 223 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 224 | "input": "An effective solid-state reaction method is reported for synthesizing boron nitride nanotubes (BNNTs) in large scale and with high purity by annealing amorphous boron powder and ferric chloride (FeCl3) catalyst in ammonia atmosphere at elevated temperatures. FeCl3 that has rarely been utilized before is introduced not only as a catalyst but also as an efficient transforming agent which converts boron powder into boron chloride (BCl3) vapor in situ. The nanotubes are bamboo in shape and have an average diameter of about 90 nm. The effect of synthetic temperatures on nanotube morphology and yield is investigated. The photoluminescence (PL) measurement shows emission bands of the nanotubes at 354, 423, 467, and 666 nm. A combined growth mechanism of vapor-liquid-solid (VLS) and solid-liquid-solid (SLS) model is proposed for the formation of the BNNTs.", 225 | "output": "{\"Name\": [\"boron nitride nanotubes\"], \"Formula\": [], \"Acronym\": [\"BNNTs\"], \"Descriptor\": [], \"Structure/Phase\": [\"nanotube\"], \"Application\": [{}], \"Property\": [\"'diameter': [90]\"], \"Synthesis\": [\"annealing\"], \"Characterization\": [\"Photoluminescence\"]}" 226 | }, 227 | { 228 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 229 | "input": "Anodic Aluminum Oxide (AAO) template prepared in acid electrolyte possess regular and highly anisotropic porous structure with pore diameter range from five to several hundred nanometers, and with a density of pores ranging from 10(9) to 10(11) cm(-2). AAO can be used as microfilters and templates for the growth of CNTs and metal or semiconductor nanowires. Varying anodizing conditions such as temperature, electrolyte, applied voltage, anodizing and widening time, one can control the diameter, the length, and the density of pores. In this work, we deposited Al thin film by radio frequency magnetron sputtering method to fabricate AAO nano template and synthesized multi-well carbon nanotubes on a glass substrate by microwave plasma-enhanced chemical vapor deposition (MPECVD). AAO nano-porous templates with various pore sizes and depths were introduced to control the dimension and density of CNT arrays. The AAO nano template was synthesize on glass by two-step anodization technique. The average diameter and interpore distance of AAO nano template are about 65 nm and 82 nm. The pore density and AAO nano template thickness are about 2.1 x 10(10) pores/cm(2) and 1 mu m, respectively. Aligned CNTs on the AAO nano template were synthesized by MPECVD at 650 degrees C with the Ni catalyst layer. The length and diameter of CNTs were grown 2 mu m and 50 nm, respectively.", 230 | "output": "{\"Name\": [\"Anodic Aluminum Oxide\"], \"Formula\": [], \"Acronym\": [\"AAO\"], \"Descriptor\": [], \"Structure/Phase\": [\"anisotropic porous\", \"nanotubes \"], \"Application\": [{\"CNTs\": []}], \"Property\": [\"'pore diameter': [10(9);10(11)]\", \"'thickness': [2.1]\", \"'diameter': [50]\", \"'length': [2]\"], \"Synthesis\": [\"magnetron sputtering\"], \"Characterization\": []}" 231 | }, 232 | { 233 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 234 | "input": "A facile method was introduced and demonstrated to synthesize zinc oxide (ZnO) nanorods (NRs) as an electron transporting layer (ETL) for organic solar cells (OSCs). Hydrothermal synthesis of the NRs showed a constant growth rate of 5.5 nm min(-1) from germination to sub-micrometer length. The properties were characterized using scanning electron microscopy (SEM), transmission electron microscopy (TEM), absorption spectrophotometry and so on. Based on these measurements, the germinant growth mechanism and its corresponding orientation characteristics were investigated. As an ETL of the OSCs, ZnO NRs enhance the charge extraction from the active layer due to their increased interfacial surface area, but there is an optimal length because of the shunt path formation and UV absorption of long ZnO NRs. As a result, the OSC with the ZnO NRs as ETL shows power conversion efficiency (PCE) up to 6.2%. The J-V characteristics and incident photon-to-current conversion efficiency (IPCE) measurement also reveal that the efficiency enhancement is an assembly of individual results from optical, physical and electrical characteristic of the ZnO NRs. (C) 2019 Published by Elsevier Ltd on behalf of The editorial office of Journal of Materials Science & Technology.", 235 | "output": "{\"Name\": [\"zinc oxide\"], \"Formula\": [\"ZnO\"], \"Acronym\": [], \"Descriptor\": [\"electron transporting layer\"], \"Structure/Phase\": [\"nanorod\"], \"Application\": [{\"organic solar cell\": [\"'efficiency': [0.062] (property)\"], \"perovskite/hole transport layer\": []}], \"Property\": [\"'growth rate': [5.5]\"], \"Synthesis\": [\"hydrothermal\"], \"Characterization\": [\"scanning electron microscopy\", \"transmission electron microscopy\", \"atomic absorption spectrophotometry\"]}" 236 | }, 237 | { 238 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 239 | "input": "Among many candidates for photoanode materials of photoelectrochemical (PEC) cell, nanostructured tungsten trioxide (WO3) is regarded as one of the most promising materials due to its superior electrical properties and adequate bandgap (similar to 2.8 eV) and band edge position. WO3 nanoflakes (WO3 NFs), which have merits on its high surface area and crystallinity, have been actively studied for this manner but solar-to-hydrogen efficiency of WO3 NFs based photoanode is still not sufficient both in light absorption and charge separation. Plasmon-induced enhancement using Au nanoparticles is excellent approach for both the efficiency of light absorption and charge separation of WO3. However, it still needs optimization on its amount, shape, coverage, and etc. Here, we synthesized WO3 NFs by solvothermal growth and decorated gold nanoparticles on these nanoflakes by e-beam evaporation and rapid thermal annealing process in a row. By this process, a large-area AuNPs/WO3 nanocomposite structure with various size, interparticle distance, and coverage of AuNPs were fabricated. These AuNPs/WO3 NFs type photoanode achieve high light absorption both in UV and visible range and consequently higher photocurrent density. The optimized AuNPs/WO3 nanocomposite photoanode exhibits 1.01 mA cm(-2) of photocurrent density, which is increased to 19.8% compared with bare WO3 nanoflakes. Field emission-scanning electron microscope, x-ray diffraction, UV-vis spectrometer analysis were measured to analyze the morphology and crystallinity and relationship between structure and PEC performance.", 240 | "output": "{\"Name\": [\"WO3 nanoflakes\"], \"Formula\": [], \"Acronym\": [\"WO3 NFs\"], \"Descriptor\": [\"photoanode\", \"decorated gold nanoparticles\"], \"Structure/Phase\": [\"nanoflake\"], \"Application\": [{\"photoelectrochemical cell\": [\"'photocurrent density': [0.198]\"]}], \"Property\": [\"'bandgap energy': [2.8]\"], \"Synthesis\": [\"solvothermal\"], \"Characterization\": [\"field emission-scanning electron microscope\", \"X-ray diffraction\", \"UV-vis spectrometer\"]}" 241 | }, 242 | { 243 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 244 | "input": "Amorphous tin-gallium oxide (a-SGO) grown with atomic layer deposition was evaluated as a buffer layer in (Ag, Cu)(In,Ga)Se-2 thin-film solar cells in search for a new material that is compatible with a variety of absorber band gaps. Hard and soft X-ray photoelectron spectroscopy on absorber/a-SGO stacks combined with J-V characterization of solar cells that were fabricated, showed that the conduction band alignment at the absorber/a-SGO interface can be tuned by varying the cation composition and/or growth temperature. Here, the surface band gap was 1.1 eV for the absorber. However, optical band gap data for a-SGO indicate that a suitable conduction band alignment can most likely be achieved even for wider absorber band gaps relevant for tandem top cells. A best efficiency of 17.0% was achieved for (Ag,Cu)(In,Ga)Se-2/a-SGO devices, compared to eta = 18.6% for the best corresponding CdS reference. Lower fill factor and open-circuit voltage values were responsible for lower cell efficiencies. The reduced fill factor is explained by a larger series resistance, seemingly related to interface properties, which are yet to be optimized. Some layer constellations resulted in degradation in fill factor during light soaking as well. This may partly be explained by light-induced changes in the electrical properties of a-SGO, according to analysis of Al/SGO/n-Si metal-oxide-semiconductor capacitors that were fabricated and characterized with J-V and C-V. Moreover, the introduction of a 1 nm thick Ga2O3 interlayer between the absorber and a-SGO improved the open-circuit voltage, which further indicates that the absorber/a-SGO interface can be improved.", 245 | "output": "{\"Name\": [\"tin-gallium oxide\"], \"Formula\": [], \"Acronym\": [\"a-SGO\"], \"Descriptor\": [\"buffer layer\"], \"Structure/Phase\": [\"amorphous\"], \"Application\": [{\"Cu(In,Ga)Se-2 solar cell\": [\"'efficiency': [0.170] (property)\"]}], \"Property\": [\"'bandgap energy': [1.1]\"], \"Synthesis\": [\"atomic layer deposition\"], \"Characterization\": [\"X-ray photoelectron spectroscopy\", \"J-V characterization\"]}" 246 | }, 247 | { 248 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 249 | "input": "Cadmium selenide (CdSe) thin films were grown on borosilicate glass substrates using the RF magnetron sputtering method. In this study, CdSe thin film was deposited at a deposition temperature in the range of 25 degrees C to 400 degrees C. The influence of deposition or growth temperature on the structural, morphological, and opto-electrical properties of CdSe films was investigated elaborately to achieve a good-quality window layer for solar-cell applications. The crystal structure, surface morphology, and opto-electrical characteristics of sputtered CdSe films were determined using X-ray diffraction (XRD), field emission scanning electron microscopy (FESEM), UV-Vis spectrophotometry, and Hall effect measurement, respectively. The XRD results revealed the polycrystalline nature of CdSe, with a hexagonal structure having a strong preferential orientation toward the (002) plane. As evident from the FESEM images, the average grain size and surface morphology of the films were dependent on deposition temperatures. The carrier concentration was obtained as 10(14) cm(-3). The band gap in the range of 1.65-1.79 eV was found. The explored results suggested that sputtered CdSe thin film deposited at 300 degrees C has the potential to be used as a window layer in solar cells.", 250 | "output": "{\"Name\": [\"Cadmium selenide\"], \"Formula\": [\"CdSe\"], \"Acronym\": [], \"Descriptor\": [\"window layer\"], \"Structure/Phase\": [\"hexagonal\", \"nanotube\"], \"Application\": [{\"solar cell\": []}], \"Property\": [\"'bandgap energy': [1.65;1.79]\"], \"Synthesis\": [\"magnetron sputtering\"], \"Characterization\": [\"X-ray diffraction\", \"field emission scanning electron microscopy\", \"UV-Vis spectrophotometry\", \"Hall effect measurement\"]}" 251 | }, 252 | { 253 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 254 | "input": "Combustion synthesis method is an economic, one step and an effective method for the synthesis of nanomaterials. In this work, titanium dioxide nanoparticles were synthesized by combustion method. The synthesized material was characterized by XRD, SEM and BET. Dye sensitized solar cell was fabricated using the synthesized material and its performance was compared with the solar cell fabricated using commercial TiO2. In order to evaluate the photovoltaic performance of DSSCs, photocurrent density to photovoltage (J-V) characteristic and electrochemical impedance spectroscopy (EIS) measurements were carried out. The DSSC fabricated using combustion synthesized and commercial TiO2 exhibited a power conversion efficiency of 6.11% and 6.62%, respectively. Combustion synthesized TiO2 which displayed similar efficiency to that of commercial material is least studied in the literature for solar cell applications.", 255 | "output": "{\"Name\": [\"titanium dioxide\"], \"Formula\": [\"TiO2\"], \"Acronym\": [], \"Descriptor\": [\"nanomaterials\"], \"Structure/Phase\": [\"nanoparticle\"], \"Application\": [{\"dye sensitized solar cell\": [\"'efficiency': [0.0611] (property)\"]}], \"Property\": [], \"Synthesis\": [\"solution combustion\"], \"Characterization\": [\"scanning electron microscope\", \"X-ray diffraction\", \"Brunauer Emmet Teller method specific surface area\", \"J-V characterization\", \"electrochemical impedance spectroscopy\"]}" 256 | }, 257 | { 258 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 259 | "input": "Flower-like zinc oxide (ZnO) micro-/nano-structures are synthesized as thin films on glass substrates by the spray pyrolysis technique at a substrate temperature of 350 +/- 5 degrees C. The films are characterized using x-ray diffraction (XRD), scanning electron microscopy (SEM) and x-ray energy dispersion spectroscopy (EDS). Absorbance spectra (UV-Vis) and ImageJ software is used for size analysis. The XRD diffractogram revealed that the films have a hexagonal (wurtzite) phase with preferential orientation along the (002) line. SEM images showed that the films have a micro/nano flower-like structure. EDS spectra revealed that the films contain chlorine in addition to zinc and oxygen. Absorbance measurements were used to estimate the band gap energies, which were found to be larger than the bulk value. Such morphologies and band gap values are advantageous when using ZnO as fore contact and/or window material in thin-film solar cells and photoanodes in dye-sensitized solar cells, optoelectronic devices and gas sensors.", 260 | "output": "{\"Name\": [\"zinc oxide\"], \"Formula\": [\"ZnO\"], \"Acronym\": [], \"Descriptor\": [\"electrode\"], \"Structure/Phase\": [\"flower shape\", \"hexagonal\"], \"Application\": [{\"dye-sensitized solar cells\": [], \"dye-sensitized solar cells\": [], \"optoelectronic device\": [], \"gas sensor\": []}], \"Property\": [\"'bandgap energy': []\"], \"Synthesis\": [\"pyrolyzing\"], \"Characterization\": [\"scanning electron microscopy\", \"x-ray energy dispersion spectroscopy\", \"UV-Vis spectrophotometry\"]}" 261 | }, 262 | { 263 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 264 | "input": "Herein, we report the first time application of waste plastic derived 3D graphene nanosheets (GNs) for hole transport material (HTM) free perovskite solar cells (PSCs), where 3D GNs have been employed as an electrode dopant material in monolithic carbon electrode based mesoscopic PSCs. Waste plastics were upcycled into high-quality 3D GNs by using two-step pyrolysis processes, where, a nickel (99.99%) metal mesh was taken as the catalytic and degradation template to get an acid free route for the synthesis of 3D GNs. Raman spectroscopy, HRTEM analysis and XRD analysis show the presence of 1-2 graphene layers within the 3D GNs. Further, the optical band gap study has also been performed to analyze the applicability of 3D GNs for PSCs. The optimized device with 3D GNs shows a power conversion efficiency (PCE) of 12.40%, whereas the carbon-based control device shows a PCE of 11.04%. Further, all other device parameters such as short circuit current (J(sc)), open circuit voltage (V-oc) and fill factor (FF) have been improved with the addition of 3D GNs. The performance enhancement in 3D GN doped HTM free PSC solar cells is attributed to the enhancement in conductivity and reduced recombination within the device. Further, the photocurrent study shows that the 3D GN device shows better performance as compared to the reference device due to the larger diffusion current. Thus, the upcycling of waste plastics into 3D GNs and their exploitation for application in energy conversion show an effective and potential way to convert waste into energy.", 265 | "output": "{\"Name\": [\"graphene nanosheets\"], \"Formula\": [], \"Acronym\": [\"GNs\"], \"Descriptor\": [\"electrocatalyst\"], \"Structure/Phase\": [\"3D\", \"nanosheet\", \"monolithic\"], \"Application\": [{\"hole transport material\": [], \"perovskite solar cells\": [\"'power conversion efficiency': [0.124] (property)\"]}], \"Property\": [\"'bandgap energy': []\", \"'conductivity': []\"], \"Synthesis\": [\"pyrolysis\"], \"Characterization\": [\"raman spectroscopy\", \"high-resolution transmission electron microscropy\", \"X-ray diffraction\"]}" 266 | }, 267 | { 268 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 269 | "input": "In order to further optimize the solvothermal method for preparing Ti-doped CuGaS2 (CGS(2):Ti) film, we have developed a simpler and rapider process of chemical synthesis for 30 min followed by annealing for 30 min. XRD, FESEM, EDAX, XPS, Raman spectra, UV-Vis-NIR absorption spectra and photoelectricity properties have been characterized. As a result, by partially replacing Ga atoms with Ti in the CGS(2) chalcopyrite crystal, a narrow partially filled intermediate band (IB) was introduced, and two additional absorption bands were directly observed at 1.61 eV and 0.84 eV spectrum. A photoelectrochemical study of CGS(2):Ti confirmed the broad-spectrum solar response produced by the IB effect. The present method is facile, does not need sulfuration process and large energy consumption, avoids the use of toxic substances and vacuum equipment, and may be a promising method for intermediate-band solar cells (IBSC).", 270 | "output": "{\"Name\": [\"Ti-doped CuGaS2\"], \"Formula\": [], \"Acronym\": [\"CGS(2):Ti\"], \"Descriptor\": [\"Ti-doped\", \"film\"], \"Structure/Phase\": [\"chalcopyrite\"], \"Application\": [{\"intermediate-band solar cells\": []}], \"Property\": [\"'bandgap energy': [1.61;0.84]\"], \"Synthesis\": [\"solvothermal\"], \"Characterization\": [\"X-ray diffraction\", \"field emission scanning electron microscope\", \"UV-vis-nIR absorption spectroscopy\", \"energy-dispersive X-ray spectroscopy\", \"Raman spectra\", \"X-ray photoelectron spectroscopy\"]}" 271 | }, 272 | { 273 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 274 | "input": "In this paper, embedded amorphous-silicon (a-Si) and titanium (Ti) ultrathin-films forming a multilayer structure is proposed as a new efficient absorber material for thin-film solar cells (TFSCs). Promising design strategy based on combining FDTD (finite difference time domain) with particle swarm optimization (PSO) was adopted to identify the a-Si/Ti multilayer structure offering the highest total absorbance efficiency (TAE). It is found that the proposed multilayer structure can serve as an effective absorber, yielding superb TAE exceeding 80%. The a-Si/Ti multilayer was then elaborated by successive growth of a-Si and Ti ultrathin layers using RF magnetron sputtering technique. The sputtered a-Si/Ti thin-film was characterized by scanning electron microscopy (SEM), X-ray diffraction (XRD), and UV-visible absorption spectroscopy. Measurements showed a unique optical behavior, promoting broadband absorbance over the visible and even NIR spectrum ranges. In particular, the prepared aSi/Ti absorber exhibits an optical band-gap of 1.36 eV, which is suitable for photovoltaic applications. A performance assessment of the elaborated absorber was investigated by extracting I-V characteristics and electrical parameters under dark and 1-sun illumination. It is revealed that the proposed absorber demonstrates outstanding electrical and sensing performances. Therefore, promoting enhanced resistive behavior and lightscattering effects, this innovative concept of optimized a-Si/Ti multilayer provides a sound pathway for designing promising alternative absorbers for the future development of a-Si-based TFSCs.", 275 | "output": "{\"Name\": [\"amorphous-silicon and titanium ultrathin-films\"], \"Formula\": [], \"Acronym\": [\"a-Si/Ti\"], \"Descriptor\": [\"absorbent\"], \"Structure/Phase\": [\"multilayer\"], \"Application\": [{\"absorber\": [\"'efficiency': [0.8] (property)\"], \"thin-film solar cells\": []}], \"Property\": [\"'bandgap energy': [1.36]\"], \"Synthesis\": [\"magnetron sputtering\"], \"Characterization\": [\"scanning electron microscopy\", \"X-ray diffraction\", \"UV-vis absorption spectroscopy\"]}" 276 | }, 277 | { 278 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 279 | "input": "In this study, we successfully prepared oxidatively functionalized graphene quantum dots (GQDs) to engineer the absorption band of a photoanode. By employing a facile hydrothermal process, we demonstrate the degree of intrinsic oxidation at the polar carbene edges of the GQDs, using Raman spectroscopy, X-ray photoelectron spectroscopy, and Fourier-transform infrared spectroscopy. The cyclic voltammetry results show that oxygen functional groups on the GQDs can change the highest occupied molecular orbital and lowest unoccupied molecular orbital levels. Dye-sensitized solar cells, with the GQDs dissolved in the dye, were evaluated in terms of their I V characteristics and conversion efficiencies. The GQDs dissolved in the dye were easily attachable to TiO2 porous particles, and also contributed to the photoelectric conversion efficiency.", 280 | "output": "{\"Name\": [\"graphene quantum dots\"], \"Formula\": [], \"Acronym\": [\"GQDs\"], \"Descriptor\": [\"oxidatively functionalized\", \"photoanode\"], \"Structure/Phase\": [], \"Application\": [{\"dye-sensitized solar cells\": [\"'I V characteristics': [1000] (property)\", \"'power conversion efficiency': [] (property)\"]}], \"Property\": [\"'molecular orbital level': []\"], \"Synthesis\": [\"hydrothermal\"], \"Characterization\": [\"Raman spectroscopy\", \"X-ray photoelectron spectroscopy\", \"Fourier-transform infrared spectroscopy\"]}" 281 | }, 282 | { 283 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 284 | "input": "In this work correlations between thin film crystallinity of plasma ion assisted electron beam evaporated vanadium oxide (VOx) and fluctuations of the deposition parameters during the growth process could be observed by in situ monitoring deposition conditions and electron microscopy studies. In the presented case, unintentional fluctuations in the gas flow at the plasma source caused by inhomogeneous melting of the target material lead to an increase in discharge current and therefore a decrease of the oxygen flow in the plasma source, resulting in the formation of highly crystalline bands due to a temporary increase in energy flux. The major part of the VOx thin film consists of a large number of nanocrystals embedded in an amorphous phase. In-depth structural analysis confirms a mixture of V2O5, in different modifications, VO2, as well as the mixed-valence oxides V4O9 and V6O13, for nanocrystalline parts and crystalline bands. These differ mainly in the degree of crystallinity being influenced by variations in discharge current, and partly in the amount of higher oxidized vanadium oxides. In future, precisely controlled variation of plasma source conditions will open up pathways to control and tailor crystallinity of electron beam evaporated thin films, allowing for production methods for patterned thin films or layers with graduated crystallinity. This may give rise to a new class of coatings of nanohybrids combining amorphous VOx with low electrical conductivity and crystalline domains providing a higher electrical conductivity which is useful for electrochromic displays, smart windows, and solar cells.", 285 | "output": "{\"Name\": [\"Vanadium oxide\"], \"Formula\": [\"VOx\"], \"Acronym\": [], \"Descriptor\": [\"thin film\"], \"Structure/Phase\": [\"nanocrystal\"], \"Application\": [{\"electrocrystalline display\": [], \"smart window\": [], \"solar cell\": []}], \"Property\": [\"'crystallinity': []\"], \"Synthesis\": [\"electron beam evaporation\"], \"Characterization\": [\"in situ monitor\", \"electron microscopy\"]}" 286 | }, 287 | { 288 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 289 | "input": "In this work, coral-like micro/nano zinc oxide (ZnO) thin films are synthesized onto glass substrates using the low-cost spray pyrolysis (SP) method without using a metal catalyst or surfactant. The films are prepared at substrate temperature T-s = 350 +/- 3 degrees C, using oxygen as the carrier gas. The structure, surface morphology, composition, and optical properties of the as-deposited thin films are characterized using x-ray diffraction (XRD), scanning electron microscopy with energy dispersive x-ray spectroscopy (SEM-EDS), and UV-VIS spectroscopy, respectively. ImageJ software is used to produce the plot profiles and surface occupancy plots. The XRD diffractogram showed that the films are polycrystalline with hexagonal (wurtzite) structure. SEM observations revealed the formation of coral-like products of different shapes with a wide range of dimensions. The EDS spectrum showed that the films contain Cl in addition to Zn and O. Absorbance is used to infer the transmittance and to estimate the band gap energy, which is found to be E-g = 3.37 eV. The fourth derivative of the absorbance revealed the three free exciton peaks A, B, and C. In addition, a strong broad absorption band with several sharp lines is detected in the energy range 1.77-2.50 eV and a weaker absorption band in the energy range 2.48-2.80 eV. This non-catalyst production of coral-like ZnO micro/nano thin films on glass substrates may have potential applications in the industry of nanoelectronic and nanooptical devices, optoelectronics, sensors, nanoscale mechanics research, and solar cells.", 290 | "output": "{\"Name\": [\"zinc oxide\"], \"Formula\": [\"ZnO\"], \"Acronym\": [], \"Descriptor\": [\"thin film\"], \"Structure/Phase\": [\"coral-like\", \"hexagonal\"], \"Application\": [{\"nanoelectronic and nanooptical devices\": [], \"optoelectronics\": [], \"sensor\": [], \"nanoscale mechanics research\": [], \"solar cells\": []}], \"Property\": [\"'bandgap energy': [3.37]\", \"'enhanced absorption in the visible region': [2.48;2.80]\"], \"Synthesis\": [\"pyrolyzing\"], \"Characterization\": [\"scanning electron microscopy\", \"UV-vis absorption spectroscopy\", \"X-ray diffraction\", \"energy dispersive x-ray spectroscopy\", \"UV-VIS spectroscopy\"]}" 291 | }, 292 | { 293 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 294 | "input": ")Silicon nanocrystals in a host matrix of silicon carbide SiC are important to high band gap dielectrics for solar cells applications. Forming Si-NC is difficult considering that SiC nano-clusters are easily formed in the host matrix. This analysis focuses on the influence of deposition temperature (T-d) and annealing ambience to synthesise SiC film for forming Si-NC in the host matrix. Here Si-NC is obtained by co-sputtering Si and C targets at 350 degrees C and 500 degrees C. Annealing is done at various temperatures using vacuum (VA) or conventional thermal annealing (CTA). Its structural and optical properties are investigated using spectroscopic ellipsometry, Fourier-transform spectroscopy (FTIR) and X-ray spectroscopy (XPS). The findings are its refractive index is between 2.69-4.08, showing Si abundance in the film. FTIR spectra shows change in intensity and position of Si-C bond variations of 620-700 cm(-1)with change in T(d )and XPS spectra confirms the presence of Si-C- and Si-Si-related bonds, influenced with increase in T-d and VA. CTA samples show increase in Si-O related bonds which indicates increase in oxygen presence. Hence a higher T-d=500(0)Cwith VA appears to be a vital to synthesize Si-NC predominantly for optoelectronic applications.", 295 | "output": "{\"Name\": [\"Silicon nanocrystals\"], \"Formula\": [\"Si-NC\"], \"Acronym\": [], \"Descriptor\": [\"high band gap dielectrics\"], \"Structure/Phase\": [\"nanocrystal\"], \"Application\": [{\"solar cell\": []}], \"Property\": [], \"Synthesis\": [\"magnetron sputtering\"], \"Characterization\": [\"spectroscopic ellipsometry\", \"Fourier-transform spectroscopy\", \"X-ray spectroscopy\"]}" 296 | }, 297 | { 298 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 299 | "input": "20 nm thick anatase niobium-doped titanium dioxide (TNO) nanofilm with resistivity of 2.9 x 10(-3) omega cm, visible-light transmittance of over 92% and work function of 4.43 eV was obtained by radio-frequency magnetron sputtering method and a diluted H-2 atmosphere annealing. A rapid annealing in a forming gas atmosphere of 3%H-2 and 97%N-2 at 500 degrees C reduced the resistivity by 3 orders of magnitude. Furthermore, the utilization of a \"glass cover\" (GC) in the annealing process decreased the resistivity by approximately 3 times and increased the refractive index by approximately 3.6% for the anatase TNO nanofilm. The secondary ion mass spectrometry measurement confirmed that the GC restrained the hydrogen diffusion from the surface into TNO occurred during the annealing. Employed as the antireflection layer in hydrogenated amorphous Si solar cells, the TNO nanofilm annealed with GC enabled an open-circuit voltage of 0.91 V, a gain of 6.5% in short-circuit current density and a reduction of 30.3% in series resistance, resulting in a gain of approximately 12.8% in photoelectric conversion efficiency. These results suggest that moderate H incorporation is beneficial for improving the electrical and optical characteristics of TNO thin films.", 300 | "output": "{\"Name\": [\"titanium dioxide\"], \"Formula\": [], \"Acronym\": [\"TNO\"], \"Descriptor\": [\"antireflection layer\"], \"Structure/Phase\": [\"nanofilm\"], \"Application\": [{\"silicon solar cell\": [\"'current density': [1.4] (property)\", \"'series resistance': [0.303] (property)\", \"'open-circuit voltage': [0.91] (property)\", \"'photoelectric conversion efficiency': [0.128] (property)\"]}], \"Property\": [\"'resistivity': [2.9]\", \"'refractive index': [3.6]\", \"'transmittance': [0.92]\"], \"Synthesis\": [\"annealing\"], \"Characterization\": [\"time-of-flight secondary ion mass spectrometry \"]}" 301 | }, 302 | { 303 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 304 | "input": "A facile approach for the synthesis of Au- and Pt-decorated CuInS2 nanocrystals (CIS NCs) as sensitizer materials on the top of MoS2 bilayers is demonstrated. A single surfactant (oleylamine) is used to prepare such heterostructured noble metal decorated CIS NCs from the pristine CIS. Such a feasible way to synthesize heterostructured noble metal decorated CIS NCs from the single surfactant can stimulate the development of the functionalized heterostructured NCs in large scale for practical applications such as solar cells and photodetectors. Photodetectors based on MoS2 bilayers with the synthesized nanocrystals display enhanced photocurrent, almost 20-40 times higher responsivity and the On/Off ratio is enlarged one order of magnitude compared with the pristine MoS2 bilayers-based photodetectors. Remarkably, by using Pt- or Au-decorated CIS NCs, the photocurrent enhancement of MoS2 photodetectors can be tuned between blue (405 nm) to green (532 nm). The strategy described here acts as a perspective to significantly improve the performance of MoS2-based photodetectors with the controllable absorption wavelengths in the visible light range, showing the feasibility of the possible color detection.", 305 | "output": "{\"Name\": [\"CuInS2\"], \"Formula\": [], \"Acronym\": [\"CIS\"], \"Descriptor\": [\"sensitizer\"], \"Structure/Phase\": [\"nanocrystal\"], \"Application\": [{\"photodetectors\": [], \"solar cells\": []}], \"Property\": [\"'photoresponse': [200-400]\"], \"Synthesis\": [], \"Characterization\": []}" 306 | }, 307 | { 308 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 309 | "input": "A facile hydrothermal method to synthesize water-soluble copper indium sulfide (CIS) nanocrystals (NCs) at 150 degrees C is presented. The obtained samples exhibited three distinct photoluminescence peaks in the red, green and blue spectral regions, corresponding to three size fractions, which could be separated by means of size-selective precipitation. While the red and green emitting fractions consist of 4.5 and 2.5 nm CIS NCs, the blue fraction was identified as in situ formed carbon nanodots showing excitation wavelength dependent emission. When used as light absorbers in quantum dot sensitized solar cells, the individual green and red fractions yielded power conversion efficiencies of 2.9% and 2.6%, respectively. With the unfractionated samples, the efficiency values approaching 5% were obtained. This improvement was mainly due to a significantly enhanced photocurrent arising from complementary panchromatic absorption.", 310 | "output": "{\"Name\": [\"copper indium sulfide\"], \"Formula\": [], \"Acronym\": [\"CIS\"], \"Descriptor\": [], \"Structure/Phase\": [\"nanocrystal\"], \"Application\": [{\"quantum dot-sensitized solar cell\": [\"'efficiency': [0.029;0.026] (property)\"]}], \"Property\": [\"'water-soluble': []\"], \"Synthesis\": [\"hydrothermal\"], \"Characterization\": []}" 311 | }, 312 | { 313 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 314 | "input": "A highly conductive carbon-doped cadmium oxide (CdO) has been fabricated by a catalyst free, compatible and low-cost one-step technique using an annealing method. In the process, the cadmium acetate was used directly as a precursor, and no further assistance from chemicals, multiple steps or instruments was needed. The morphology and phase identification were confirmed by x-ray diffraction, x-ray photoelectron and Raman spectroscopy analyses. From the analyses, it was confirmed that the carbons are inserted at the vacant oxygen sites of the prepared CdO, and the Cd-C linking bridges play a significant role in increasing the electrical conductivity. The thin film prepared by the fabricated CdO showed an excellent electrical conductivity of 858.51 S/cm, suggesting that a highly conductive CdO material fabricated by a simple annealing method can serve as a good electrical conductor material. The thin film fabricated from the carbon-doped CdO material can be used for electrical applications such as solar cells, photo transistors, photo diodes and super capacitors.", 315 | "output": "{\"Name\": [\"cadmium oxide\"], \"Formula\": [], \"Acronym\": [\"CdO\"], \"Descriptor\": [\"electrical conductor\"], \"Structure/Phase\": [], \"Application\": [{\"solar cells\": [], \"super capacitors\": [], \"photo diodes\": [], \"photo transistors\": []}], \"Property\": [\"'electrical conductivity': [858.51]\"], \"Synthesis\": [\"annealing\"], \"Characterization\": [\"X-ray diffraction\", \"X-ray photoelectron\", \"Raman spectroscopy\"]}" 316 | }, 317 | { 318 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 319 | "input": "A kesterite Cu2ZnGeS4 (CZGS) thin film was prepared on an Mo-coated glass substrate by spray deposition of an aqueous solution containing constituent elements followed by heat treatment in a sulfur atmosphere. The results of analysis of the ionization potential of the CZGS thin film by photoelectron spectroscopy (PES) and the external quantum efficiency (EQE) spectrum of a solar cell device based on the CZGS thin film showed that this compound had ca. 0.7 eV more negative conduction band minimum (CBM) than that of the well-studied kesterite Cu2ZnSnS4 (CZTS) but had almost the same energy level of the valence band maximum (VBM) as that of CZTS. Owing to the relatively shallow CBM, the CZGS film worked as a photocathode for photoelectrochemical (PEC) CO2 reduction in a neutral aqueous solution (0.1 M KHCO3) saturated with CO2 by applying a bias potential lower than that for inducing the reaction under a dark condition, whereas the CZTS film could not work under the same condition. Enhancement of PEC CO2 reduction activity by surface coverage of the CZGS thin film with a ZnS layer is also discussed on the basis of the electronic structure of the thus-formed CZGS-ZnS interface.", 320 | "output": "{\"Name\": [], \"Formula\": [\"Cu2ZnGeS4\"], \"Acronym\": [\"CZGS\"], \"Descriptor\": [\"photocathode\"], \"Structure/Phase\": [\"kesterite\"], \"Application\": [{\"solar cell device\": [], \"CO2 reduction\": []}], \"Property\": [\"'conduction band minimum': [0.7]\", \"'valence band maximum': [1]\"], \"Synthesis\": [], \"Characterization\": [\"photoelectron spectroscopy\", \"external quantum efficiency spectrum\"]}" 321 | }, 322 | { 323 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 324 | "input": "A novel set of hole-transporting materials (HTMs) based on pi-extended diindolotriazatruxene (DIT) core structure with electron-rich methoxy-engineered functional groups were designed and synthesized via a facile two-step procedure. All compounds were afforded from inexpensive precursors without a complex purification process. Cyclic voltammograms indicate that the resulting HTMs exhibit suitable highest occupied molecular orbital (HOMO) energy levels, which facilitate efficient hole injection from the valence band of perovskites into the HOMO of DIT-based HTMs as confirmed by time-resolved photoluminescence. Notable power conversion efficiency of the planar perovskite solar cells with low-temperature device fabrication achieved 18.21% utilizing D2, which is competitive with the corresponding devices based on the common Spiro-OMeTAD-based HTMs. The results manifest that DIT-based compounds are promising HTMs for constructing high-efficiency planar perovskite solar cells with low-cost solution processing procedures.", 325 | "output": "{\"Name\": [\"diindolotriazatruxene\"], \"Formula\": [], \"Acronym\": [\"DIT\"], \"Descriptor\": [\"hole-transporting materials\"], \"Structure/Phase\": [\"core shell\"], \"Application\": [{\"perovskite solar cells\": [\"'power conversion efficiency': [0.1821] (property)\"]}], \"Property\": [\"'HOMO energy': []\"], \"Synthesis\": [], \"Characterization\": [\"cyclic voltammetry\"]}" 326 | }, 327 | { 328 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 329 | "input": "A polyaniline (PANI)-based TiO2-sensitized polymeric active layer was synthesized via an in situ polymerization technique in the presence of ammonium persulphate (APS as the oxidizer) for its use in hybrid solar cell application.TiO2-sensitized tubular PANI nanocomposites with an average diameter of approximate to 50 nm proved to be very efficient for optimizing the optical band gap so as to absorb the maximum solar spectrum. The synergistic combination of PANI-TiO2 (4%) featured the maximum current for varying concentrations of TiO2 towards current-voltage (I-V) and cyclic voltammetry measurements. Further, the thermal stability was markedly enhanced by incorporating TiO2 nanoparticles into the polymer matrix as compared to pure PANI.", 330 | "output": "{\"Name\": [], \"Formula\": [\"TiO2\"], \"Acronym\": [], \"Descriptor\": [\"polyaniline (PANI)-based\"], \"Structure/Phase\": [\"tubular\", \"nanocomposites\"], \"Application\": [{\"hybrid solar cell\": []}], \"Property\": [\"'bandgap energy': []\"], \"Synthesis\": [], \"Characterization\": [\"current-voltage\", \"cyclic voltammetry\"]}" 331 | }, 332 | { 333 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 334 | "input": "A safe and environmentally benign solvent is a requisite for mass production of thin film solar cells via solution methods. Here, a highly industry suitable solvent N-methyl-pyrrolidone (NMP) is used for the first time to make a precursor solution with simple compounds of CuCl, InCl3 center dot 4H(2)O and thiourea and fabricate CuIn(S,Se)(2) (CISSe) solar cells. A power conversion efficiency of 10.23% has been achieved from the NMP-based solution when the precursor film was processed in air, which is only 8.55% for film processed in glove box. Characterizations using XRD, Raman, SEM, EDX, and FTIR show air annealing favors decomposition of organic species in the precursor film, which results in high quality absorber materials. Further improvement in device efficiency can be expected by gallium alloying and optimization of device fabrication conditions. The results demonstrate highly efficient thin film solar cells can be fabricated from an industry suitable NMP precursor solution in air, which is promising for simplifying film processing and reducing manufacture cost.", 335 | "output": "{\"Name\": [\"N-methyl-pyrrolidone\"], \"Formula\": [], \"Acronym\": [\"NMP\"], \"Descriptor\": [\"solvent\"], \"Structure/Phase\": [], \"Application\": [{\"CuIn(S,Se)(2) (CISSe) solar cells\": [\"'efficiency': [0.1023] (property)\"]}], \"Property\": [], \"Synthesis\": [\"annealing\"], \"Characterization\": [\"X-ray diffraction\", \"raman spectroscopy\", \"scanning electron microscopy\", \"fourier transform infrared spectroscopy\", \"energy dispersive spectrometer\"]}" 336 | }, 337 | { 338 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 339 | "input": "A solution combustion synthesized hole transport layer (HTL) of spinel nickel cobaltite (NiCo2O4) incorporating 3% Cu-2% Li was fabricated using the doctor-blading technique for planar inverted perovskite solar cells (PVSCs). PVSCs incorporating 3% Cu-2% Li-doped NiCo2O4 showed an increase in Jsc and Voc device performance parameters compared to unmodified NiCo2O4, leading to power conversion efficiency (PCE) of 16.5%. X-ray photoelectron spectroscopy measurements revealed the tendency of Cu cations to replace preferably the surface Ni atoms by changing the surface stoichiometry of NiCo2O4, inducing a cathodic polarization. Ultraviolet photoelectron spectroscopy measurements unveiled the increase in the ionization potential by 0.1 eV for a co-doped NiCo2O4 film compared to unmodified NiCo2O4-based HTL. We attribute the enhanced PCE of the inverted PVSCs presented to the improved hole extraction properties of 3% Cu-2% Li NiCO2O4 HTL. (C) 2019 Author(s).", 340 | "output": "{\"Name\": [], \"Formula\": [\"NiCo2O4\"], \"Acronym\": [], \"Descriptor\": [\"hole transport layer\"], \"Structure/Phase\": [], \"Application\": [{\"perovskite solar cells\": [\"'power conversion efficiency': [0.165] (property)\"]}], \"Property\": [\"'ionization potential': [0.1]\"], \"Synthesis\": [\"solution combustion\"], \"Characterization\": [\"X-ray photoelectron spectroscopy\", \"Ultraviolet photoelectron spectroscopy\"]}" 341 | }, 342 | { 343 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 344 | "input": "Although SnO2 is widely used in perovskite solar cells (PSCs) as the electron transport layer, there is still a need for a detailed investigation regarding the impact of physisorption on the performance of PSCs. Here, we report the role of physisorption of oxygen species from the ambient atmosphere in the oxidation states and defective oxides at the surface of SnO2 nanoparticles for the performance of PSCs. The SnO2 nanoparticle's thin film shows a higher oxidation state (Sn-O) and lower charge trap sites at an annealing temperature of 100 degrees C, and the SnO(2 )film based PSCs yield an efficiency of 18.04%. The annealing process results in the gradual desorption of physically adsorbed oxygen species. Therefore, the annealing process of SnO2 nanoparticle thin films in atmosphere is the key factor to control the oxidation states and defective oxides at the surface of SnO2 nanoparticle thin films.", 345 | "output": "{\"Name\": [], \"Formula\": [\"SnO2\"], \"Acronym\": [], \"Descriptor\": [\"electron transport layer\"], \"Structure/Phase\": [\"nanoparticle\"], \"Application\": [{\"perovskite solar cell\": [\"'efficiency': [0.1804] (property)\"]}], \"Property\": [\"'charge trap sites': [0.1804]\"], \"Synthesis\": [\"annealing\"], \"Characterization\": []}" 346 | }, 347 | { 348 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 349 | "input": "Although carbon based hole-conductor-free perovskite solar cells have acquired interest in the scientific community in the past few years, some issues with regard to relatively low efficiency and toxicity need to be addressed before their commercialization. For the first time, the environmentally friendly ammonium acetate CH3COONH4 (NH4Ac) and Zn(CH3COO)(2) (ZnAc2) were adopted as additives for perovskite preparation. By introducing a very small amount (0.08%) of volatile NH4Ac through a simple one-step spin coating method, high-quality perovskite layer with lower defect density, larger grain size and pure perovskite composition without alien atom was obtained, resulting in a power conversion efficiency (PCE) of 13.9%, which is similar to 25% enhancement compared to that of the device without additive. Moreover, stability was also investigated, 6% degradation appeared in PCE of NH4Ac sample after being stored for 1900h in dark. The promoted interface contact of carbon/perovskite results in higher efficiency and longer stability of NH4Ac added perovskite solar cell (PSC). On the other hand, addition of non-volatile ZnAc2 only lead to a slight enhancement of the efficiency. The factors affecting the properties will be discussed in detail. (C) 2019 Elsevier B.V. All rights reserved.", 350 | "output": "{\"Name\": [\"ammonium acetate CH3COONH4\"], \"Formula\": [\"CH3COONH4\"], \"Acronym\": [\"NH4Ac\"], \"Descriptor\": [\"additive\"], \"Structure/Phase\": [], \"Application\": [{\"perovskite solar cells\": [\"'power conversion efficiency': [0.139] (property)\"]}], \"Property\": [\"'degradation': [0.06]\", \"'stability': [0.06]\"], \"Synthesis\": [], \"Characterization\": []} & {\"Name\": [], \"Formula\": [\"ZnAc2\"], \"Acronym\": [], \"Descriptor\": [\"additive\"], \"Structure/Phase\": [], \"Application\": [{\"perovskite solar cells\": [\"'power conversion efficiency': [0.139] (property)\"]}], \"Property\": [], \"Synthesis\": [], \"Characterization\": []}" 351 | }, 352 | { 353 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 354 | "input": "Although halide perovskite solar cell (PSC) technology reaches, in few years, efficiencies greater than 25%, the cost-ceffective perspective is achievable only if scalable processes in real manufacturing conditions (i.e., pilot line and/or plant factory) are designed and optimized for the full device stack. Herein, a full semiautomatic scalable process based on the blade-coating technique is demonstrated to fabricate perovskite solar modules in ambient conditions. An efficient and stable triple-cation cesium methylammonium formamidinium (CsMAFA) perovskite is deposited in ambient air with a two-step process assisted by air and green anti-solvent quenching. The developed industry compatible coating process enables the fabrication of several highly reproducible small-area cells on module size substrate with an efficiency exceeding 17% and with high reproducibility. Corresponding reproducible modules (less than 2% variability) with a 90% geometrical fill factor achieve an efficiency larger than 16% and T-80 = 750 h in light-soaking condition at maximum power point and room temperature/ambient after encapsulation. Film deposition properties are assessed by different characterization techniques, namely, scanning electron microscopy, profilometry, UV-vis and photoluminescence (PL) spectroscopy, and PL and electroluminescence imaging. The techniques confirm less defects and local coating variations of the ambient air/bladed devices with respect to the nitrogen air/spinned devices.", 355 | "output": "{\"Name\": [\"cesium methylammonium formamidinium\"], \"Formula\": [], \"Acronym\": [\"CsMAFA\"], \"Descriptor\": [\"coating\"], \"Structure/Phase\": [\"perovskite\"], \"Application\": [{\"halide perovskite solar cell\": [\"'efficiency': [00.17] (property)\"]}], \"Property\": [], \"Synthesis\": [\"two-step process\"], \"Characterization\": [\"scanning electron microscopy\", \"UV-vis absorption spectroscopy\", \"profilometry\", \"photoluminescence\"]}" 356 | }, 357 | { 358 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 359 | "input": "An Nb-doped TiO2 (Nb-TiO2) film comprising a double structure stacked with a bottom compact layer and top mesoporous layers was synthesized by treating a Ti precursor-coated substrate using a one-step low-temperature steam-annealing (SA) method. The SA-based Nb-TiO2 films possess high crystallinity and conductivity, and that allows better control over the conduction band (CB) of TiO2 for the electron transport layer (ETL) of the perovskite solar cells by the Nb doping level. Optimization of power conversion efficiency (PCE) for the Nb-TiO2-based ETL was combined with the CB level tuning of the mixed-halide perovskite by changing the Br/I ratio. This band offset management enabled to establish the most suitable energy levels between the ETL and the perovskites. This method was applied to reduce the band gap of perovskites to enhance the photocurrent density while maintaining a high open-circuit voltage. As a result, the optimal combination of 5 mol % Nb-TiO2 ETL and 10 mol % Br in the mixed-halide perovskite exhibited high photovoltaic performance for low-temperature device fabrication, achieving a high-yield PCE of 21.3%.", 360 | "output": "{\"Name\": [], \"Formula\": [\"Nb-TiO2\"], \"Acronym\": [], \"Descriptor\": [\"Nb-doped\", \"electron transport layer\"], \"Structure/Phase\": [\"bottom compact layer\", \"top mesoporous layers\"], \"Application\": [{\"perovskite solar cell\": [\"'efficiency': [0.213] (property)\"]}], \"Property\": [\"'bandgap energy': []\", \"'photocurrent density': []\", \"'high photovoltaic performance': []\"], \"Synthesis\": [\"steam-annealing\"], \"Characterization\": []}" 361 | }, 362 | { 363 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 364 | "input": "Anomalous photovoltaic effect and switchable photovoltaic output have propelled much recent research and development of ferroelectric photovoltaic (FE-PV) materials in solar cell, photodetection and optical-read nonvolatile memory fields. However, extremely small output photocurrent density in the order of nA/cm2 or mu A/cm2 is the primary problem to limit applications of FE-PV materials. Herein, a large short-circuit current density is achieved in polycrystalline hexagonal YMnO3 (h-YMO) thin film prepared by sol-gel method, which reaches 3.92 mA/cm2 under 1 sun illumination (100 mW/cm2), much larger than that of other narrow-bandgap hexagonal rare-earth manganates or ferrites epitaxial thin film materials. A systematic nanoscale analysis suggests that the grain with ferroelectric polarization and grain boundary with high conductivity are the active structures that cause the photovoltaic effect. Especially, grain boundary induces a downward band bending due to potential difference at the grain/grain boundary interface, which forms a new build-in electric field to promote carrier separation. Moreover, grain boundary as the conduction path achieves effective transport of carriers. The work demonstrates the positive effect of grain boundaries in FE-PV materials, and opens up a promising route to the development of high-performance photovoltaic materials via the combined effect of ferroelectric polarization and grain boundary.", 365 | "output": "{\"Name\": [], \"Formula\": [\"YMnO3\"], \"Acronym\": [\"h-YMO\"], \"Descriptor\": [\"thin film\"], \"Structure/Phase\": [\"hexagonal\"], \"Application\": [{\"photodetection\": [], \"solar cell\": [], \"optical-read nonvolatile memor\": []}], \"Property\": [\"'bandgap energy': []\", \"'conductivity': []\", \"'short-circuit current density': [3.92]\"], \"Synthesis\": [\"sol-gel\"], \"Characterization\": []}" 366 | }, 367 | { 368 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 369 | "input": "Cadmium Sulphide is the conventional n-type buffer layer used in thin film Cadmium Telluride solar cells. It is well known that Cadmium Sulphide causes optical losses and sulphur diffuses into the absorber during high temperature activation. Sputter-deposited Mg-doped ZnO (MZO) has been shown to be an attractive buffer layer for Cadmium Telluride solar cells due to its transparency and tuneable band gap. It is also stable to high temperature processing and avoids diffusion of elements into the cadmium telluride absorber during the cadmium chloride activation treatment. However, degradation is observed in solar cells incorporating MZO buffer layers. Analysis of the MZO film surface potential has revealed significant fluctuations in the thin film work function once the layer is exposed to the atmosphere following deposition. These fluctuations are attributed to the high reactivity to water vapour of the MgO contained in the MZO films. This has been analysed using X-ray Photoelectron Spectroscopy to determine corresponding changes in the surface chemistry. The Zinc Oxide component is relatively stable, but the analysis shows that MgO forms a Mg(OH)(2) layer on the MZO surface which forms a secondary barrier at the MZO/CdTe interface and/or at the interface between MZO and the Fluorine-doped SnO2. This affects the Fill Factor and as a consequence it degrades the conversion efficiency.", 370 | "output": "{\"Name\": [\"Mg - doped ZnO\"], \"Formula\": [], \"Acronym\": [\"MZO\"], \"Descriptor\": [\"buffer layer\"], \"Structure/Phase\": [], \"Application\": [{\"thin film Cadmium Telluride solar cell\": []}], \"Property\": [\"'bandgap energy': []\"], \"Synthesis\": [\"sputtering\"], \"Characterization\": [\"X-ray photoelectron spectroscopy\"]}" 371 | }, 372 | { 373 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 374 | "input": "Carbon-based mesoscopic perovskite solar cells (PSCs) are promising for printable next-generation photovoltaic applications, but the optical properties of their carbon layer limit their light harvesting efficiency. We developed solution-processable indium-tin-oxide (ITO) nanoparticles to replace carbon electrodes for mesoscopic lead-free tin-based PSCs. The ITO electrodes were fabricated via screen printing with 1, 2, 4, 6 and 8 layers corresponding to the thicknesses of 2.1-14.3 mu m; tin perovskite (GA(0.2)FA(0.8)SnI(3)) with SnF2 (20%) and EDAI(2) (15%) (GA represents guanidinium; EDAI(2) represents ethylenediammonium diiodide) was drop-casted on the device to produce a mesoporous structure of FTO/TiO2/perovskite/Al2O3/ITO. The best device achieved a power conversion efficiency of 5.4% with great stability. These solution-processed ITO electrodes are a landmark to shed light on new paths for the commercialization of lead-free PSCs.", 375 | "output": "{\"Name\": [\"indium tin oxide\"], \"Formula\": [], \"Acronym\": [\"ITO\"], \"Descriptor\": [\"electrode\"], \"Structure/Phase\": [\"nanoparticle\"], \"Application\": [{\"perovskite solar cell\": [\"'power conversion efficiency': [0.054] (property)\"]}], \"Property\": [\"'thickness': [2.1;14.3]\"], \"Synthesis\": [\"ininkjet printing\"], \"Characterization\": []}" 376 | }, 377 | { 378 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 379 | "input": "CdS and different concentrations of Mn+2 doped CdS quantum dots are synthesized via simple chemical method on poly-vinyl alcohol matrix. The doped and undoped CdS quantum dots are characterized by using UV-Vis absorption spectroscopy, X-ray diffraction, and energy dispersive X-ray spectroscopy to revel the changes in optical and structural properties due to Mn doping. The synthesized quantum dots were used as sensitizing layer on a ZnO thin film solar cell and tested under artificial illumination condition to obtain the current densityvoltage characteristics. A significantly higher efficiency of about 2.09% was obtained for Mn doped CdS quantum dots compared to that of 1% efficiency in pure CdS quantum dot sensitized solar cell.", 380 | "output": "{\"Name\": [], \"Formula\": [\"CdS\"], \"Acronym\": [], \"Descriptor\": [\"quantum dots\"], \"Structure/Phase\": [\"nanoparticle\"], \"Application\": [{\"solar cell\": [\"'efficiency': [0.209] (property)\"]}], \"Property\": [], \"Synthesis\": [\"solution-processed\"], \"Characterization\": [\"UV-vis absorption spectroscopy\", \"X-ray diffraction\", \"energy dispersive X-ray spectroscopy\"]}" 381 | }, 382 | { 383 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 384 | "input": "Ce-doped ZnO hollow spheres (Ce@ZHSp) with various Ce-contents of 0, 0.25, 0.50 and 1.0 mol% were synthesized via a template-free hydrothermal method. The calcined samples were characterized using XRD, FESEM, TEM, BET-BJH and UV-Vis spectroscopic techniques. The analytical results showed that all the samples possessed a hexagonal wurtzite structure. The crystallite size of Ce@ZHSp decreased with increasing Ce-content. Calcined samples had a hollow sphere morphology with an average diameter range of 4.63-5.16 mu m. Diffuse reflectance was increased by adding an appropriate Ce-dopant of 0.25 mol% into the ZnO hollow spheres. The energy-conversion efficiency of a dye-sensitized solar cell was examined under the A.M. 1.5 direct spectrum. These results showed that the highest efficiency was derived from a bilayer photoanode using 0.25 mol% Ce-doped ZnO hollow spheres as a light scattering layer. This finding was related to an increase in the amount of dye adsorption and a photonic reflection effect originating from the relatively larger particle sizes with higher porosity compared to the other samples.", 385 | "output": "{\"Name\": [\"e-doped ZnO\"], \"Formula\": [\"Ce@ZHSp\"], \"Acronym\": [], \"Descriptor\": [\"light scattering layer\"], \"Structure/Phase\": [\"hollow spheres\"], \"Application\": [{\"dye-sensitized solar cell\": [\"'efficiency': [] (property)\"]}], \"Property\": [\"'diameter': [4.63;5.16]\"], \"Synthesis\": [\"hydrothermal\"], \"Characterization\": [\"fourier transform infrared spectroscopy\", \"X-ray diffraction\", \"transmission electron microscopy\", \"fourier transform infrared spectroscopy\", \"UV-Vis spectroscopic\", \"Brunauer Emmet Teller method\"]}" 386 | }, 387 | { 388 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 389 | "input": "Indium tin oxide (ITO) film is the most widely used as front electrodes in solar cells with a copper metallization scheme. No work has focused on the barrier properties of the ITO layer on the textured silicon for solar cells. In this work, a thin indium tin oxide barrier layer and copper metal layer were deposited on textured (001) silicon by a sputtering. The stacks present Cu/ITO/Si. The structures of Cu/ITO/Si characterized by scanning transmission electron microscope, energy-dispersive X-ray spectrometer, and powder X-ray diffraction. The results show that the stacks of Cu/ITO/Si can be preserved up to 600 degrees C. The 35-nm thickness ITO layer was found to be a diffusion barrier against Cu up to 600 degrees C. The copper thin films were agglomerated and formed the particle at a temperature of 700 degrees C. The failure of Cu/ITO/Si can be attributed to agglomerate copper thin films and breakdown of ITO thin films at a temperature of 700 degrees C.", 390 | "output": "{\"Name\": [], \"Formula\": [\"Cu/ITO/Si\"], \"Acronym\": [], \"Descriptor\": [\"electrode\"], \"Structure/Phase\": [\"(001)\"], \"Application\": [{\"solar cell\": []}], \"Property\": [\"'thickness': [35]\", \"'diffusion barrier': []\"], \"Synthesis\": [\"magnetron sputtering\"], \"Characterization\": [\"scanning electron microscope\", \"transmission electron microscope\", \"X-ray diffraction\", \"energy-dispersive X-ray spectrometer\"]}" 391 | }, 392 | { 393 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 394 | "input": "Inorganic lead halide perovskite CsPbIBr2 possesses good stability with a suitable band gap for tandem solar cells. Decreasing the defect concentration and improving the film quality is crucial to further increase the power conversion efficiency of CsPbIBr2 solar cells. Here, the crystallization dynamics of CsPbIBr2 films is regulated by introducing the volatile organic salt, formamidinium acetate (FAAc) into the precursor solution. It is found that FAAc slows the crystallization process of CsPbIBr2 films and pinhole-free films with large grains and smooth surfaces are obtained. The defect concentration of the films is decreased and the nonradiative recombination is significantly inhibited. By improving the film quality, the FAAc remarkably enhances the efficiency of CsPbIBr2 solar cells. The champion device delivers a power conversion efficiency of 9.44% and exhibits higher stability than the reference device. This finding provides an effective strategy for reducing defects, suppressing the recombination, and improving the performance of CsPbIBr2 solar cells.", 395 | "output": "{\"Name\": [\"formamidinium acetate\"], \"Formula\": [], \"Acronym\": [\"FAAc\"], \"Descriptor\": [\"organic salt\"], \"Structure/Phase\": [], \"Application\": [{\"solar cell\": [\"'efficiency': [0.944] (property)\"]}], \"Property\": [\"'volatile': []\"], \"Synthesis\": [\"solution-processed\"], \"Characterization\": []}" 396 | }, 397 | { 398 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 399 | "input": "Interfacial engineering plays an important role to improve the photovoltaic performance of organic solar cells (OSCs). Herein, CdSe/ZnS quantum dots (QDs) are used as a cathode interlayer (CIL) modifier. By using this strategy, an enhanced power conversion efficiency (PCE) from 13.0% to 14.6% is achieved, mainly due to the increase in open-circuit voltage (V-oc) and short-circuit current density (J(sc)). A single QD layer of a proper size can reduce the defects on the surface of the active layer and smoothen the interface between the active layer and cathode. Furthermore, the low work function of the QDs with dipole moment facilitates charge transport and suppresses charge recombination at the interface by strengthening the built-in field, thus contributing to the enhancement of PCE. The excitons generated by the QDs can also be dissociated at the IT-4F/QD interface, which boosts the photon harvesting capability of the device. As a result, a high PCE of 14.6% is achieved for QD-modified OSCs.", 400 | "output": "{\"Name\": [\"CdSe/ZnS quantum dots\"], \"Formula\": [\"CdSe/ZnS QDs\"], \"Acronym\": [], \"Descriptor\": [\"modifier\"], \"Structure/Phase\": [\"quantum dot\"], \"Application\": [{\"organic solar cell\": [\"'PCE': [0.146] (property)\"]}], \"Property\": [\"'work function': []\"], \"Synthesis\": [], \"Characterization\": []}" 401 | }, 402 | { 403 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 404 | "input": "Lead (Pb) in conventional perovskite solar cells (PSCs) is toxic and has to be replaced. Situated in one group of the periodic table of elements, tin (Sn) has the same valence electrons' configuration as Pb (ns(2)np(2)), promising analogous chemical properties. Hence, Sn is considered a suitable replacement to Pb. However, because of the lack of lanthanide shrinkage, Sn behaves differently: Pb is stable in Pb2+ form, an 2 oxidation state needed for perovskite structure, while Sn tends to lose all its valence electrons forming Sn4+. As a result, PSCs based on Sn are not efficient. Traces of oxygen have been conventionally discussed as a source of Sn oxidation. But recent findings point to the oxidation of Sn-based perovskites even in the absence of oxygen. This perspective summarizes recently-discovered unconventional oxidation pathways of Sn perovskites, including reaction with solvent molecules and disproportionation. We explain these phenomena by a Frost-Ebsworth diagram and argue that a deeper understanding of this diagram is a key toward stable and efficient Pb-free Sn-based PSCs.", 405 | "output": "{\"Name\": [\"tin\"], \"Formula\": [\"Sn\"], \"Acronym\": [], \"Descriptor\": [\"lanthanide\"], \"Structure/Phase\": [], \"Application\": [{\"perovskite solar cell\": []}], \"Property\": [], \"Synthesis\": [\"solution-processed\"], \"Characterization\": [\"Frost-Ebsworth\"]}" 406 | }, 407 | { 408 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 409 | "input": "Lead halide perovskite solar cells afford high power conversion efficiencies, even though the photoactive layer is formed in a solution process. At the same time, solution processing may impose some severe dewetting issues, especially if organic, hydrophobic charge transport layers are considered. Ultimately, very narrow processing windows with a relatively large spread in device performance and a considerable lab-to-lab variation result. Here, we unambiguously identify dimethylsulfoxide (DMSO), which is commonly used as a co-solvent and complexing agent, to be the main reason for dewetting of the precursor solution on hydrophobic hole transport layers, such as polytriarylamine, in a gas-quenching-assisted deposition process. In striking contrast, we will show that N-methyl-2-pyrrolidon (NMP), which has a lower hydrophilic lipophilic-balance, can be favorably used instead of DMSO to strongly mitigate these dewetting issues. The resulting high-quality perovskite layers are extremely tolerant with respect to the mixing ratio (NMP/dimethylformamide) and other process parameters. Thus, our findings afford an outstandingly robust, easy to use, and fail-safe deposition technique, yielding single (MAPbI(3)) and double (FA(0.94)Cs(0.06)PbI(3)) cation perovskite solar cells with high efficiencies (similar to 18.5%). Most notably, the statistical variation of the devices is significantly reduced, even if the deposition process is performed by different persons. We foresee that our results will further the reliable preparation of perovskite thin films and mitigate process-to-process variations that still hinder the prospects of upscaling perovskite solar technology.", 410 | "output": "{\"Name\": [\"N-methyl-2-pyrrolidon\"], \"Formula\": [], \"Acronym\": [\"NMP\"], \"Descriptor\": [\"photoactive layer\"], \"Structure/Phase\": [], \"Application\": [{\"perovskite solar cell\": [\"'efficiency': [0.185] (property)\"]}], \"Property\": [\"'lower hydrophilic lipophilic-balance': []\"], \"Synthesis\": [\"deposition\"], \"Characterization\": []}" 411 | }, 412 | { 413 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 414 | "input": "Tantalum (Ta)-doped titanium oxide (TiO2) thin films are grown by plasma enhanced atomic layer deposition (PEALD), and used as both an electron transport layer and hole blocking compact layer of perovskite solar cells. The metal precursors of tantalum ethoxide and titanium isopropoxide are simultaneously injected into the deposition chamber. The Ta content is controlled by the temperature of the metal precursors. The experimental results show that the Ta incorporation introduces oxygen vacancies defects, accompanied by the reduced crystallinity and optical band gap. The PEALD Ta-doped films show a resistivity three orders of magnitude lower than undoped TiO2, even at a low Ta content (0.8-0.95 at.%). The ultraviolet photoelectron spectroscopy spectra reveal that Ta incorporation leads to a down shift of valance band and conduction positions, and this is helpful for the applications involving band alignment engineering. Finally, the perovskite solar cell with Ta-doped TiO2 electron transport layer demonstrates significantly improved fill factor and conversion efficiency as compared to that with the undoped TiO2 layer.", 415 | "output": "{\"Name\": [\"Tantalum-doped titanium oxide\"], \"Formula\": [], \"Acronym\": [\"Ta-doped TiO2\"], \"Descriptor\": [\"electron transport layer\", \"hole blocking compact layer\"], \"Structure/Phase\": [], \"Application\": [{\"perovskite solar cell\": [\"'improved fill factor': [] (property)\", \"'efficiency': [] (property)\"]}], \"Property\": [\"'resistivity': []\", \"'crystallinity': []\", \"'bandgap energy': []\"], \"Synthesis\": [\"atomic layer deposition\"], \"Characterization\": [\"Ultraviolet photoelectron spectroscopy\"]}" 416 | }, 417 | { 418 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 419 | "input": "Technological implementation of organolead trihalide perovskite (OTP) photovoltaics requires suppression of the surface ionic defects and grain boundaries of OTP films. These surface ionic defects have a detrimental effect on the power conversion efficiency (PCE), are notorious for introducing hysteresis into the current density-voltage (J-V) characteristics, and decrease the stability of perovskite solar cells (PSCs). Here, we report the use of core/shell quantum dots (QDs) as passivation layers on the OTP surfaces to decrease the trap density and, simultaneously, stabilize the OTP chemical structure and extend the charge carrier lifetime. Density functional theory (DFT) calculations indicated that the OTP surface defects and grain boundaries were effectively suppressed by the presence of the CdSe/ZnS QDs. We attribute the lower trap density of the OTP to the Se2- anions from the CdSe/ZnS passivation layer, inducing van der Waals interactions between the organic and inorganic components of the framework. For PSCs featuring CdSe/ZnS QD passivation, the PCE reached close to 20% with diminishing hysteresis of the J-V characteristics and fill factor (FF) of 81.44%. Moreover, the PSCs incorporating the CdSe/ZnS QD passivation layer exhibited long-term stability, retaining 75 and 80% of their initial performance after 2400 and 720 h, respectively. This facile interfacial strategy appears highly applicable for preparing high-performance durable OTP-based high optoelectronic devices.", 420 | "output": "{\"Name\": [\"quantum dots\"], \"Formula\": [], \"Acronym\": [\"QDs\"], \"Descriptor\": [\"passivation layer\"], \"Structure/Phase\": [\"core shell\"], \"Application\": [{\"OTP photovoltaics\": [\"'efficiency': [0.2] (property)\", \"'fill factor': [0.8144] (property)\"]}], \"Property\": [], \"Synthesis\": [], \"Characterization\": [\"density functional theory\"]}" 421 | }, 422 | { 423 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 424 | "input": "The cadmium sulfide (CdS) is a promising electron transport layer (ETL) material for perovskite solar cells (PSCs) due to its low photocatalytic activity toward perovskite materials under UV light. The critical problem responsible for the moderate performance of CdS-based PSCs is the parasitic light absorption of CdS, which drives researchers to deposit ultrathin ETLs. However, the ultrathin ETL often involves the undesirable shunt current leakage because of the direct contact between conducting substrate and perovskite layer. Herein, a fully low-temperature solution-processed colloidal SnO2-assisted CdS (S-CdS) ETL for planar CH3NH3PbI3 PSCs is constructed. The detailed characterizations of morphological, optical, and energy levels confirm that the assistance of colloidal SnO2 provides the ameliorated continuity, reduces surface roughness and superior wettability of ETLs for high-quality perovskite formation as well as the favorable cascade band structure for efficient charge transfer. The study of charge transfer mechanisms reveals that the S-CdS ETL effectively inhibits the shunt leakage, promotes the electron extraction and suppresses the charge recombination at the ETL/perovskite interface. Consequently, the S-CdS ETL-based PSCs deliver an appreciable efficiency of 16.26%, doubling that of conventional CdS-based devices. To the best of our knowledge, this value is the champion efficiency reported for CdS-based CH3NH3PbI3 PSCs.", 425 | "output": "{\"Name\": [\"SnO2-assisted cadmium sulfide\"], \"Formula\": [\"S-CdS\"], \"Acronym\": [], \"Descriptor\": [\"electron transport layer\"], \"Structure/Phase\": [\"cascade band\"], \"Application\": [{\"perovskite solar cell\": [\"'effiency': [0.1626] (property)\"]}], \"Property\": [], \"Synthesis\": [\"solution-processed\"], \"Characterization\": []}" 426 | }, 427 | { 428 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 429 | "input": "The copper based semiconductors have now proved to the most promising solar cell material nowadays among all the available p type semiconducting materials. Present work deals with fabrication of thin film layer of CuO and CuS material on to a glass substrate by physical vapour deposition method. The structural, surface morphological, compositional, optical and electrical characterization of the fabricated thin films is to be carried out with the help of X-ray diffraction (XRD), scanning electron microscopy, energy dispersive X-ray, UV-visible spectrophotometer and I-V characterization respectively confirms the formation of CuO and CuS film on glass substrate.The structural property of CuO and CuS are obtained from XRD pattern. Both CuO and CuS material obtained are of crystalline nature with monolithic and hexagonal structure respectively. The surface morphology of CuO films shows that as the temperature increases, the fabricated film shows prominent granular texture and CuS film surface shows flower like texture. The optical band gap of CuO varies from 1.56 to 1.78 eV as the temperature decrease from 450 degrees C to 350 degrees C. The band gap of CuS lies within the range of 2.50 eV. The contact resistivity and bulk resistivity of CuO sample (S1) is 7.3 x 10(-4) ohm m(2) and 2.4 x 10(-4) ohm m respectively. The contact resistivity and bulk resistivity of CuS sample (S4) is 5.1 x 10(-4) ohm m(2) and 1.27 x 10(-4) ohm m respectively. The I-V characteristics curve of CuO and CuS obtained shows the semiconducting behaviour of the thin film formed; both in the forward bias and reverse bias region.", 430 | "output": "{\"Name\": [], \"Formula\": [\"CuO\"], \"Acronym\": [], \"Descriptor\": [\"semiconductor\"], \"Structure/Phase\": [\"monolithic\"], \"Application\": [{\"solar cell\": []}], \"Property\": [\"'bandgap energy': [1.56;1.78]\", \"'resistivity': [7.3;2.4]\"], \"Synthesis\": [\"physical vapor deposition\"], \"Characterization\": [\"X-ray diffraction\", \"scanning electron microscopy\", \"UV-visible spectrophotomete\", \"energy dispersive spectrometry\", \"I-V characterization\"]} & {\"Name\": [], \"Formula\": [\"CuS\"], \"Acronym\": [], \"Descriptor\": [\"semiconductor\"], \"Structure/Phase\": [\"hexagonal\"], \"Application\": [{\"solar cell\": []}], \"Property\": [\"'bandgap energy': [2.5]\", \"'resistivity': [5.1;1.27]\"], \"Synthesis\": [\"physical vapor deposition\"], \"Characterization\": [\"X-ray diffraction\", \"scanning electron microscopy\", \"UV-visible spectrophotomete\", \"energy dispersive spectrometry\", \"I-V characterization\"]}" 431 | }, 432 | { 433 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 434 | "input": "The cubic and spherical shaped iron pyrite (FeS2) nanocrystals were synthesized in a pure phase form by an efficient hot injection method. These FeS2 nanocrystals were used as a counter electrode (CE) alternative to the conventional Pt CE in dye-sensitized solar cells (DSSCs) owing to its tremendous optical properties and low-cost. The obtained FeS2 nanocrystalline materials with excellent shape and phase purity were confirmed through XRD and Raman spectroscopy data. From Tafel, and impedance spectroscopy studies, the catalytic activity FeS2 CEs are found to be comparable with that of Pt CE. Along with the I-3(-)/I- electrolyte, photo-conversion efficiency is found to be 6.9% (spherical), 6.2% (cubic) for the FeS2 CE, and 7% for Pt CE. The excellent performance of the FeS2 CE in DSSCs makes it a distinctive choice among the various CE materials studied including low-cost photovoltaics.", 435 | "output": "{\"Name\": [\"iron pyrite\"], \"Formula\": [\"FeS2\"], \"Acronym\": [], \"Descriptor\": [\"counter electrode\"], \"Structure/Phase\": [\"nanocrystal\"], \"Application\": [{\"dye-sensitized solar cell\": [\"'effiency': [0.062] (property)\"]}], \"Property\": [], \"Synthesis\": [\"hot injection\"], \"Characterization\": [\"X-ray diffraction\", \"raman spectroscopy\", \"Tafel slope\", \"electrochemical impedance spectroscopy\"]}" 436 | }, 437 | { 438 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 439 | "input": "The development of metal oxide-based electron transport layers in perovskite solar cells (PSCs) has received intensive research interest for achieving high-efficiency PSCs. Herein, TiO(2)nanorods (TiO(2)NRs) are grown onTiO(2)seed layers coated on fluorine-doped tin oxide (FTO) glass substrate by using a hydrothermal method and then are utilized as the electronic transport layer in PSCs. The main concern, after hydrothermal growth of TiO(2)NRs, is that their crystallinity can be improved by a sequential high-temperature treatment at 450 degrees C. In addition to high-temperature annealing, a low-temperature treatment with boiling water, which is expected to clean the surface of the TiO(2)NRs, is developed. In this contribution, the champion PSCs are those based on TiO(2)NRs where boiling water treatment achieves a maximum power conversion efficiency (PCE) of 15.50%, whereas a PCE of 12.91% is obtained from PSCs based on TiO(2)NRs with high-temperature annealing. The remarkable ease of using a water-assisted process offers an efficient approach to the removal of residuals adsorbed on the surface and circumvents the disadvantage of a thermal annealing method resulting in high-production costs. This low-temperature treatment can be used to improve TiO(2)films in flexible PSCs.", 440 | "output": "{\"Name\": [\"TiO(2)nanorods\"], \"Formula\": [], \"Acronym\": [\"TiO(2)NRs\"], \"Descriptor\": [\"electron transport layer\"], \"Structure/Phase\": [\"nanorod\"], \"Application\": [{\"perovskite solar cell\": [\"'power conversion efficiency': [0.155] (property)\"]}], \"Property\": [], \"Synthesis\": [\"hydrothermal\"], \"Characterization\": []}" 441 | }, 442 | { 443 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 444 | "input": "The development of polysulfide reduction electrocatalysts with abundant active sites, high conductivity and good stability is vital to the fabrication of quantum dot sensitized solar cells (QDSCs) but is challenging. Herein, we report a convenient approach for the synthesis of Cu nanoparticle@carbon nanorod (Cu@CNR) composites by the direct pyrolysis of HKUST-1 precursors at temperatures above 1000 degrees C. The Cu@CNR composite whose carbon nanorods (CNRs) are encased with Cu nanoparticles (Cu NPs) exhibits superior catalytic activity towards polysulfide reduction in QDSCs, which might be ascribed to the abundant CuxS active sites derived from Cu NPs and the excellent charge transfer capability of graphitized CNRs. Benefiting from the encasement in CNRs, the dissociation of CuxS from counter electrodes (CEs) is effectively suppressed. When Cu@CNR composites deposited on F-doped tin oxide glass were used as CEs, a champion power conversion efficiency of 9.50% (J(sc) = 26.50 mA cm(-2), V-oc = 0.628 V, and FF = 0.573) under one sun illumination was observed with Zn-Cu-In-Se QD sensitizers and polysulfide electrolyte.", 445 | "output": "{\"Name\": [\"Cu nanoparticle@carbon nanorod\"], \"Formula\": [], \"Acronym\": [\"Cu@CNR\"], \"Descriptor\": [\"electrocatalyst\", \"composites\"], \"Structure/Phase\": [\"nanoparticle\"], \"Application\": [{\"solar cell\": [\"'power conversion efficiency': [0.095] (property)\"]}], \"Property\": [\"'active sites': []\", \"'charge transfer capability': []\"], \"Synthesis\": [\"pyrolysis\"], \"Characterization\": []}" 446 | }, 447 | { 448 | "instruction": "You're a chatbot that can extract entities and relations in a paragraph.", 449 | "input": "S, N-co-doped porous carbon (SNPC) materials are good candidates for the cathodic oxygen-reduction reaction (ORR) and lithium-ion batteries (LIBs). However, SNPC gives low yield and is expensive. Herein, we report a new and efficient method for synthesizing a S, N-co-doped mesoporous carbon material through the carbonization of S, N-containing precursors in molten ZnCl2, where ZnCl2 served as the ionic solvent and Lewis acid catalyst. The resultant SNPC-800 showed a mesoporous structure with a specific surface area of 1235 m(2) g(-1) and a mesopore-size range of 10-45 nm, which were considerably larger than those obtained through the carbonization of ionic liquids and fabrication of graphene oxides. Furthermore, ORR measurements indicated good catalytic activity, comparable to the commercial Pt/C catalyst. Also the SNPC-800 material exhibited excellent catalytic stability, and high methanol tolerance compared to the commercial Pt/C catalyst. Density functional theory calculation results revealed that the catalytic properties originated from the synergistic effect of the S/N dopant and that the main catalytic reaction path followed an associative mechanism. LIB tests further showed high reversible capacity, as well as excellent cycling stability and rate performance.", 450 | "output": "{\"Name\": [\"S, N-co-doped porous carbon\"], \"Formula\": [], \"Acronym\": [\"SNPC\"], \"Descriptor\": [], \"Structure/Phase\": [\"mesoporous\"], \"Application\": [{\"flexible lithium-ion battery\": [\"'capacity': [] (property)\", \"'cycling stability': [] (property)\", \"'rate performance': [] (property)\"], \"oxygen-reduction reaction\": [\"'stability': [] (property)\", \"'tolerance': [] (property)\"]}], \"Property\": [\"'specific area': [1235]\", \"'mesopore size': [10;45]\"], \"Synthesis\": [\"pyrolyzing\"], \"Characterization\": [\"density functional theory\"]}" 451 | } 452 | ] -------------------------------------------------------------------------------- /inference_KG.py: -------------------------------------------------------------------------------- 1 | import time 2 | import torch 3 | from transformers import LlamaTokenizer, LlamaForCausalLM 4 | import sys 5 | import json 6 | from tqdm import tqdm 7 | import signal 8 | import multiprocessing 9 | import argparse 10 | import os 11 | from peft import PeftModel 12 | 13 | # clear cache 14 | torch.cuda.empty_cache() 15 | num_gpus = torch.cuda.device_count() 16 | print(num_gpus) 17 | os.environ['CUDA_LAUNCH_BLOCKING'] = "1" 18 | 19 | ############################## data part ########################################## 20 | IO_chunk_size = 50 21 | 22 | def load_txt_files(file_path): 23 | data_items = [] 24 | with open(file_path, 'r', encoding='utf-8') as file: 25 | for line in file: 26 | json_data = json.loads(line.strip()) 27 | data_items.append((json_data['resource'], json_data['instruction'])) 28 | return data_items 29 | 30 | 31 | def save_and_exit(signum, frame): 32 | print('terminating by signal') 33 | results = list(shared_results) 34 | with open(filename, "w") as file: 35 | for i in tqdm(range(0, len(results), IO_chunk_size)): 36 | chunk = results[i:i + IO_chunk_size] 37 | file.writelines([json.dumps(item) + "\n" for item in chunk]) 38 | file.flush() 39 | sys.exit(0) # Exit gracefully 40 | 41 | 42 | # Set the signal handler 43 | signal.signal(signal.SIGTERM, save_and_exit) 44 | # Set the signal handler for SIGINT 45 | signal.signal(signal.SIGINT, save_and_exit) 46 | 47 | ########################## model part ###################################### 48 | # inference 49 | def inference(sent, tokenizer, model): 50 | input_ids = tokenizer(sent, return_tensors="pt").input_ids.to("cuda") 51 | generated_ids = model.generate( 52 | input_ids=input_ids, 53 | # max_length=2048, 54 | max_new_tokens=1024, 55 | do_sample=True, 56 | repetition_penalty=1.0, 57 | temperature=0.8, 58 | top_p=0.75, 59 | top_k=40 60 | ) 61 | res = tokenizer.decode(generated_ids[0]) 62 | 63 | del input_ids 64 | del generated_ids 65 | return res 66 | 67 | 68 | def process_on_gpu(gpu_id, data_subset, shared_results, model_path, basemodel_path, use_PEFT_model, counts, output_path): 69 | try: 70 | torch.cuda.set_device(gpu_id) 71 | 72 | s = time.time() 73 | tokenizer = LlamaTokenizer.from_pretrained(model_path, maxlength=2048) 74 | if not use_PEFT_model: 75 | model = LlamaForCausalLM.from_pretrained( 76 | model_path, 77 | load_in_8bit=False, 78 | torch_dtype=torch.float16 79 | ).cuda(gpu_id) 80 | else: 81 | base_model = LlamaForCausalLM.from_pretrained(basemodel_path, load_in_8bit=False, 82 | torch_dtype=torch.float16, use_auth_token='your_use_auth_token') 83 | peft_model = PeftModel.from_pretrained(base_model, model_path, load_in_8bit=False, 84 | torch_dtype=torch.float16,) 85 | model = peft_model.cuda(gpu_id) 86 | 87 | e = time.time() 88 | print(f'\nModel load in time on {gpu_id}: {e - s}\n') 89 | 90 | s = time.time() 91 | count = 0 92 | results_in_file = [] 93 | for key, data in data_subset: 94 | res = inference(data, tokenizer, model) 95 | shared_results.append({'paper': key, 'instruction': data, 'response': res}) 96 | results_in_file.append({'paper': key, 'instruction': data, 'response': res}) 97 | del res 98 | torch.cuda.empty_cache() 99 | 100 | count += 1 101 | if count % 10 == 0: 102 | with open(f'{output_path}/inference{gpu_id}.json', "a") as file: 103 | file.writelines([json.dumps(item) + "\n" for item in results_in_file]) 104 | results_in_file = [] 105 | e = time.time() 106 | print(f'\ngpu {gpu_id}: {count} files has already been inferenced, the inference time is {e - s}\n') 107 | s = time.time() 108 | 109 | counts.value += 1 110 | 111 | except Exception as e: 112 | print(f"Error on GPU {gpu_id}: {e}") 113 | 114 | return 0 115 | 116 | 117 | if __name__ == '__main__': 118 | # Set up argument parser 119 | parser = argparse.ArgumentParser(description='Process paths.') 120 | parser.add_argument('--data_path', type=str, help='Path to the data file') 121 | parser.add_argument('--output_path', type=str, help='Path to save the output') 122 | parser.add_argument('--model_path', type=str, help='Path to the model') 123 | parser.add_argument('--timeout', type=int, default=20000, help='the forceable time to stop program') 124 | parser.add_argument('--basemodel_path', type=str, default='none', help='basemodel for PEFT model') 125 | parser.add_argument('--use_PEFT', type=bool, default=False, help='if uses PEFT model. If True, basemodel_path must be added') 126 | 127 | # Parse arguments 128 | args = parser.parse_args() 129 | 130 | multiprocessing.set_start_method('spawn', force=True) 131 | 132 | data_path = args.data_path 133 | output_path = args.output_path 134 | model_path = args.model_path 135 | filename = os.path.join(output_path, 'inference.json') 136 | timeout = args.timeout #seconds 137 | basemodel_path = args.basemodel_path 138 | use_PEFT_model = args.use_PEFT 139 | 140 | print("loading model, path:", model_path) 141 | 142 | # Check if output_path exists, if not create it 143 | if not os.path.exists(output_path): 144 | os.makedirs(output_path) 145 | 146 | for i in range(num_gpus): 147 | with open(f'{output_path}/inference{i}.json', "w") as file: 148 | pass 149 | 150 | # Load data from txt files 151 | data_items = load_txt_files(data_path) 152 | # data_items = list(data_dic.items()) 153 | num_data = len(data_items) 154 | chunk_size = len(data_items) // num_gpus + 1 155 | data_chunks = [data_items[i:i + chunk_size] for i in range(0, len(data_items), chunk_size)] 156 | print(f'{num_data} data, {num_gpus} gpus, {chunk_size} chunks, chunk len = {len(data_chunks)}') 157 | 158 | 159 | # Create a shared list using multiprocessing.Manager 160 | manager = multiprocessing.Manager() 161 | shared_results = manager.list() 162 | counts = manager.Value('i', 0) 163 | 164 | # Use multiprocessing to process each chunk on a separate GPU with a timeout 165 | pool = multiprocessing.Pool(processes=num_gpus) 166 | for args in enumerate(data_chunks): 167 | pool.apply_async(process_on_gpu, args=(args[0], args[1], shared_results, model_path, basemodel_path, use_PEFT_model, counts, output_path)) 168 | pool.close() 169 | 170 | start_time = time.time() 171 | while True: 172 | if counts.value >= num_data: 173 | print('All processes success') 174 | pool.terminate() 175 | break 176 | if time.time() - start_time > timeout: 177 | print('terminal because of timeout') 178 | pool.terminate() 179 | break 180 | time.sleep(60) 181 | 182 | results = list(shared_results) 183 | 184 | # Flatten the results and write to file 185 | print('\nWriting into file!!!\n') 186 | with open(filename, "w") as file: 187 | try: 188 | for i in tqdm(range(0, len(results), IO_chunk_size)): 189 | chunk = results[i:i + IO_chunk_size] 190 | file.writelines([json.dumps(item) + "\n" for item in chunk]) 191 | file.flush() 192 | except Exception as e: 193 | file.flush() 194 | print(f"An error occurred: {e}") 195 | 196 | --------------------------------------------------------------------------------