Adsorption Science & Technology

Review Article

A Review of the Modeling of Adsorption of Organic and Inorganic Pollutants from Water Using Artificial Neural Networks

Table 1

Summary of the ANN modeling of the adsorption of water pollutants at batch operating conditions.


Adsorbent	Adsorbate	ANN model	Experiment	Data used (Training-Testing-Validation)	Input variables	Output variable	Pollutant removal	Reference

Activated carbon	Methylene blue, Auramine-O, Crystal violet, Eosin yellow	Levenberg–Marquardt back propagation	Equilibrium and Kinetics	23-0-8	C₀, m, t	Removal percentage	99.42%	[4]
Polyaniline/carboxymethyl cellulose/TiO₂	Congo red	Multi-layered perceptron back propagation (5-1-8)	Kinetic, isotherm and RSM (CCD)	28-9-9	pH, t, C₀, T, m	Adsorption capacity	119.9 mg/g	[5]
Raw oak waste and NaOH-activated oak waste	Cd²⁺	Three–layered feed forward back propagation	Kinetic and isotherm	153-33-33	pH, t, m, C₀, T	Removal percentage	99.50%	[6]
Zeolite	Diazinon	Multi-layered feed forward (4-7-1)	Equilibrium and Kinetics	21-5-5	t, pH	Removal percentage	98.2%	[8]
Magnetic walnut shell	Methyl violet, Rhodamine 6G	Multi-layered feed forward, Levenberg-Marquardt, Bayesian regularized and scaled conjugate gradient	Equilibrium and Kinetics	75-0-25	pH, C₀, m, t	Removal percentage	Methyl violet 140.7 mg/g, R6G 78.53 mg/g	[9]
Anammox bacteria	Nitrogen	Multi-layered feed forward, Levenberg-Marquardt	Isotherms	48-16-16	Treatment time, NLR, ammonia concentration and nitrite concentration	Adsorption capacity	17.79 mg/g	[11]
Nanobiosorbent derived from waste molasses	As³⁺	Feed forward Levenberg Marquardt back propagation (4-20-1)	RSM (CCD)	21-5-4	C₀, m, pH, t	Removal percentage	878.03 mg/g	[12]
Magnetic rGO-chitosan composite	Cefixime	Random forest, Feed forward (3-60-50-1)	RSM and equilibrium	15-5-0	C₀, m, pH	Adsorption capacity	30.36 mg/g	[13]
Cu/F graphene aerogel	Perfluorooctanoic acid	Adaptive neuro-fuzzy inference system	RSM (CCD)	23-5-5	m, pH, ionic strength, T	Removal percentage	25 mg/g	[17]
Activated spent tea	Methylene blue	Three-layered feed forward back propagation (5-10-1)	Kinetics	61-10-10	C₀, m, pH, T, t	Removal percentage	1042 mg/g	[26]
SBA-15/CTAB composite	Reactive red 198	Adaptive neuro fuzzy inference system	Equilibrium and Kinetics	70-30-0	T, pH, t, m, C₀	Adsorption rate	99%	[44]
MOF-5-activated carbon	Fast green, eosin Y, quinine yellow	Multi-layered feed forward back propagation (6-11-3)	Equilibrium, Kinetics and RSM (CCD)	75-25-0	C₀, pH, m, t	Removal percentage	FG 98.1%, EY 98.1%, QY 91.9%	[45]
Orange peel	Paracetamol	Levenberg–Marquardt back propagation (3-18-1)	Kinetic	332-0-163	C₀, t, T	Pseudo second order kinetics	--	[47]
Activated carbon	Crystal violet	Three layered feed forward back propagation	Equilibrium, Kinetics and RSM (CCD)	23-8-0	C₀, pH, m, t	Removal percentage	99.82%	[185]
Sawdust	Basic red 46, Cu²⁺	Levenberg–Marquardt back propagation (5-7-1)	Kinetics	38-6-6	C₀, pH, t, m	Removal percentage	--	[186]
Polyaniline/SiO₂ nanocomposite	Amido black 10B	Multi-layered feed forward (4-8-1)	Kinetic and isotherm	10-4-4	pH, t, C₀, m	Removal percentage	99.87%	[58]
Zeolite prepared from raw fly ash	Cr⁶⁺	Multi-layered feed forward	Equilibrium and Kinetics	21-4-5	T, m, t, pH	Removal percentage	92.30%	[59]
Chemically modified sugarcane bagasse	Real textile effluent with mixture of dyes	Multilayer feed-forward	Kinetic and isotherm	60-20-20	pH, m, t, Ps,	Removal percentage	100%	[60]
Leucaena leucocephala seeds	As³⁺, As⁵⁺	Levenberg–Marquardt back propagation (4-14-1)	Kinetics	108-36-36	C₀, m, t, V	Removal percentage	>90 %	[61]
Acid-activated neutralized red mud	Phosphate	Three-layered feed forward and quasi-Newton back-propagation, scaled conjugate gradient and Levenberg–Marquardt (6-10-1)	Kinetic and equilibrium	25-8-8	C₀, m, pH, T, t	Adsorption capacity	31.34-192.62 mg/g	[74]
Polyaniline nano-adsorbent	Methyl orange	Multi-layered feed forward (5-8-1)	Kinetic and isotherm	60-20-20	pH, t, C₀, T, m	Removal percentage	32.04 mg/g	[75]
Clay mixture	Cu²⁺	Three layered Levenberg–Marquardt back propagation	Kinetics, RSM	17-5-5	pH, m, T, C₀	Adsorption capacity	23.36 mg/g	[76]
Modified bentonite	Chlorobenzene	Radial basis function feed forward, Levenberg–Marquardt back propagation	Kinetic and isotherm	--	C₀, t, m, pH	Adsorption capacity	0.9223 mmol/g	[78]
Fe₃O₄/MIL-101(Fe))	Sarafloxacin	Multi-layered feed forward (4-8-1)	Kinetic, isotherm and RSM (CCD)	17-5-6	C₀, m, t, pH	Removal percentage	93.10%	[79]
UiO-66(Zr) and NH₂-UiO-66(Zr) metal organic frameworks	Pyrene	Multi-layered perceptron feed forward back propagation	RSM (CCD)	10-10-27	T, m, C₀, p H, T	Removal percentage	UiO-66(Zr) 99.2%, NH2-UiO-66(Zr) 95.67%	[80]
Penicillium YW 01	Acid black 172, Congo red	ANN and hybrid ANN-genetic algorithm (4-(3-10)-1)	Kinetics	77-26-26	pH, T, C₀, t	Adsorption capacity	AB172 225.38 mg/g, CR 411.53 mg/g	[96]
Activated carbon from opuntia ficus indica	As³⁺	Hybridization of the traditional adsorption equations with a feed forward ANN	Kinetic and isotherm	Isotherm 21-3-3; Kinetics 44-5-5	C₀, T, pH, t (kinetics), Ce (for isotherms)	Adsorption capacity	8.32 mg/g	[106]
Activated carbon	Aliphatic and aromatic compounds	Three layered feed forward back propagation (4-3-1)	Isotherms	333-35-0	Molecular size, molecular volume, branching number, substituent length, heteroatoms and global substitution pattern	The parameter K log (qe/Ce)	--	[107]
Activated carbons	55 organic aromatic compounds	Three layered feed forward (5-3-1)	Isotherms	45-10-0	Molecular size and volume, branching number, substituent length, heteroatoms and global substitution pattern	The parameter K log (qe/Ce)	--	[108]
Hematite	Cd²⁺	Adapted neural fuzzy and back propagation (5-20-1)	Kinetics	7-4-4	C₀, v, T, pH, Ps	Ce	11.5 mg/g	[112]
Powdered activated carbon	Acid orange 7	Three layered feed forward back propagation (3-2-1)	RSM, Equilibrium and Kinetics	146-73-0	pH, C₀, t	Ce	449.77 mg/g	[113]
Antep pistachio (Pistacia Vera L.) shells	Pb²⁺	Three layered feed forward Levenberg–Marquardt back propagation (5-11-1)	Kinetics	34-16-16	pH, T, m, t, C₀	Removal percentage	98.70%	[114]
Activated carbon	Auramine O	Multi-layered feed forward back propagation (6-7-1)	Kinetics	272 data for training	C₀, pH, m, t, T, irradiation time	Adsorption capacity	700 mg/g	[115]
Spirulina sp. biomass	Methylene blue	Genetic algorithm (3-20-20-1)	Equilibrium	108-81-0	T, pH, m	Removal percentage or adsorption capacity	900 mg/g	[116]
Industrial sludge leachate with red mud	Pb²⁺	Multi-layered perceptron feed forward Levenberg–Marquardt back propagation (3-12-1)	Kinetics and RSM (CCD and BBD)	15/05/2005	t, m, pH	Removal percentage	98.51%	[117]
Bacillus sp.	Total chromium	Feed forward back propagation (3-10-1)	Equilibrium	280-36-36	C₀, pH, t	Adsorption capacity	50 mg/g	[118]
Bentonite clay	Fe³⁺	Multi-layered perceptron (3-9-1)	Equilibrium, Kinetics and RSM (CCD)	13-3-0	C₀, t, m	Removal percentage	90.12%	[119]
Fungal biomass	Reactive black 5	Feed forward Levenberg–Marquardt back propagation (9-10-1)	Equilibrium	135-45-45	C₀, pH, t, nitrogen content, carbon content, hydrogen content, BET area, pore volume and pore diameter	Adsorption capacity	179.26 mg/g	[187]
Date palm seeds	Cu²⁺	Feed forward with neuro fuzzy logic (3-3-27-1)	Equilibrium	10-10-0	pH, m, T	Adsorption capacity	4.18 mg/g	[121]
Lentil straw	Lanaset red G	Three layered feed forward quick propagation (4-24-1)	Kinetics	784-184-184	Ps, m, pH, C₀, t	Adsorption capacity	271.12 mg/g	[122]
Activated almond shell	Zn²⁺	Non-linear autoregressive model processes with exogenous input NARX (4-10-2)	Equilibrium	10-10-10	pH, m, particle size, C₀	Removal percentage, adsorption capacity	80%	[124]
Lignocellulosic Biomasses	Pb²⁺, Cd²⁺, Ni²⁺, Zn²⁺	Feed forward back propagation (6-10-1;10-10-1;5-10-1)	Equilibrium and Kinetics	169 data (kinetic) (70-15-15); 92 data (isotherm) (70-15-15)	Kinetics: C₀, t; Isotherms: Ce, BET area, content of cellulose, hemicellulose and lignin, concentration of acidic groups, MW, hydrated ionic radii, electronegativity, hydration energy	Adsorption capacity	1-7 mg/g	[125]
Gold nanoparticles loaded with activated carbon	Reactive orange 12	Hybrid three layered feed forward – Imperialist competitive (3-12-1)	Equilibrium and Kinetics	168-72-0	pH, m, T, t	Removal percentage	99%	[123]
Calcareous soil	Fe³⁺	Quantum back propagation multi-layer perceptron (6-6-1)	Equilibrium and Kinetics	40-13-13	C₀, t, m, Ps, v, T	Removal percentage	99.20%	[127]
Magnetic/talc nanocomposite	Pb²⁺, Ni²⁺, Cu²⁺	Wavelet neural network based on the incremental back propagation (3-(1-15)-3)	RSM (CCRD)	13-6-0	C₀, m, t	Removal percentage	Pb²⁺ 93.60%, Ni²⁺ 51.30%, Cu²⁺ 81%	[128]
Rice straw	Cd²⁺	Feed-forward Levenberg–Marquardt back-propagation and neural-fuzzy inference system (3-6-1)	Equilibrium	26-9-8	C₀, m, pH	Removal percentage	80%	[129]
Typha domingensis	Ni²⁺, Cd²⁺	Adaptive neuro fuzzy inference system	Equilibrium	-	pH, m, C₀, t, Ps	Removal percentage	-	[131]
Multi-walled and single-walled carbon nanotubes	Triamterene	Levenberg–Marquardt back-propagation (3-(1-25)-1)	Equilibrium and Kinetics	45-0-19	C₀, m, t	Removal percentage	95%	[132]
Coffee grounds	Pb²⁺	Three-layer feed forward and Levenberg–Marquardt back propagation (1-(4-13)-1)	Equilibrium	70-15-15	pH, C₀	Adsorption capacity	22.9 mg/g	[133]
ZnS nanoparticles-activated carbon	Brilliant green	Radial basis function	Equilibrium RSM (CCD)	36-0-18	C₀, t, m	Removal percentage	98.89%	[135]
Nanowires and activated carbon	Safranin-O, Indigo Carmine	Three layered feed forward backpropagation	Equilibrium, Kinetics and RSM (CCD)	70-15-15	C₀, m, t	Removal percentage	S-O 29.09 mg/g, IC 37.85 mg/g	[137]
Maize bran	Cr⁶⁺	Multi-layered feed forward (4-10-1)	Kinetic	81-25-18	C₀, t, pH, T	Adsorption capacity	~13.5 mg/g	[138]
Clinoptilolite	Fluoride	Hybridization of ANN and Langmuir and Pseudo-second order equations (3-4-2)	Kinetic and isotherm	69-30-30	C₀, t, T, Ce, pH, T	Adsorption capacity and parameters of the models	12.4 mg/g	[139]
Calcined Ca-Al-(NO₃)	Fluoride	Three layered feed forward Levenberg-Marquardt (5-10-1)	Kinetic	650 data	C₀, m, T, pH	Adsorption capacity	45 mg/g	[140]
Palm kernel shell activated carbon	Zn²⁺	Levenberg-Marquardt back propagation, gradient descent, resilient backpropagation and gradient descent with adaptive linear regression (5-7-1)	Kinetics, RSM (CCD)	190-40-40	pH, t, C₀, m, T	Removal percentage	90%	[141]
Centaurea stem	Crystal violet	Multi-layered feed forward network	Kinetic, isotherm, RSM (CCD)	26-0-6	pH, T, t, C₀, m	Removal percentage	476.19 mg/g	[142]
Poly-aluminium chloride	Disperse blue 79	Three layered Levenberg–Marquardt back propagation (4-4-1)	Kinetic, RSM (CCD)	27-9-9	pH, m, C₀, pH	Removal percentage	52.60%	[145]
Chitosan/nanodiopside	Crystal violet	Hybrid feed forward - genetic algorithm (4-9-1)	Kinetic and isotherm, RSM (CCD)	70-15-15	C₀, pH, m, t	Removal percentage	99.50%	[148]
Hydroxyapatite (Hap))/chitosan nanocomposite	Pb²⁺	Adaptive neuro-fuzzy inference system	Kinetic and isotherm	38-0-19	pH, t, v, T, m, C₀, Hap concentration	Removal percentage	99.20%	[150]
Activated carbon, carbon nanotubes, chitin, chitosan, corn straw, grape seeds, sugarcane bagasse, orange peel and rice husk	Indium³⁺	Levenberg–Marquardt back propagation and adaptive neuro–fuzzy inference system	Kinetic	840-180-180	Surface area, point of zero charge, t, m	Adsorption capacity	1000 mg/g	[153]
Abelmoschus esculentus seeds	Nile blue	Three–layered feed–forward back propagation and Levenberg-Marquardt	Kinetic, isotherm and RSM (CCD)	21-5-5	C₀, m, pH, t	Adsorption capacity	42.34 mg/g	[154]
Mesoporous rGO/Fe/Co nanohybrids	Methylene blue	Hybrid ANN -particle swarm, hybrid ANN-genetic algorithm	Kinetic and isotherm, RSM (CCD)	24-0-6	pH, T, t, C₀	Removal percentage	88.40%	[156]
Biochar from rice husk	Fluoride	Multi-layered back propagation (4-10-10-10-1)	Kinetics	70-15-15	m, T, pH, C₀	Removal percentage	4.45 mg/g	[158]
Magnetic clay/Graphene Oxide/Fe₃O₄	pesticide diazinon	Multi-layered back propagation (5-13-1)	RSM(CCD)	23-7-7	pH, T, t, C₀, m	Removal percentage	7.2 mg/g	[159]
Magnetic Fe₃O₄ nanoparticles loaded with cocoa pod carbon composite	Methylene blue, crystal violet	Multi-layered feed forward Levenberg-Marquardt back propagation (4-9-1)	RSM(CCD)	18-6-6	C₀, m, pH, t	Removal percentage	MB 147.43 mg/g, CV 162.25 mg/g	[160]
Mesoporous CoFe₂O₄/graphene oxide nanocomposites	Arsenide	Hybrid back propagation neural network - genetic algorithm (4-4-1)	RSM(CCD)	--	T, pH, C₀, t	Removal percentage	86.62%	[161]
Iron doped rice husk	Cr⁶⁺	Feed forward Levenberg-Marquardt back propagation (5-1-1)	Kinetics	25 data	pH, m, T, t, v	Removal percentage	11.14 mg/g	[163]
Hydrochar derived from KOH activated Crocus sativus petals	Pb²⁺	Support vector machine	Equilibrium	25-5-0	T, t, water/biomass ratio, ZnCl₂/biomass ratio	Adsorption capacity	89.52 mg/g	[165]
Water treatment plant residuals	Phosphate	Feed forward multilayer Levenberg–Marquardt back propagation (4-2-3)	Isotherms	7-1-1	pH and total phosphate in system	Adsorption capacity	400 mg/g	[188]
Activated carbon	Auramine O	Feed forward Levenberg–Marquardt back propagation (6-7-1)	Kinetic	272 data for training	C₀, t, v, T, pH, m, V	Adsorption capacity	850 mg/g	[189]
Sawdust of mangifera indica	Cu²⁺	Multi-layered partial recurrent back-propagation (4-50-40-27-1)	Isotherms	4608-256-25	C₀, pH, T, Ps	Removal percentage	6 mg/g	[190]
Activated carbon	Chlorophenol	Three layered feed forward back propagation (4-8-1)	Isotherms and Kinetics	520-140-140	pH, C₀, T, t	Removal percentage	>60 %	[191]
Alga Chara sp.	Malachite green	Three layered feed forward back propagation (5-12-1)	Kinetics	180-60-60	C₀, m, pH, T, t	Removal percentage	92.75%	[192]
Chara contraria	Lanaset red (LR) G	Three layered feed forward back propagation (4-23-1)	Kinetics	1566-369-369	Ps, m, pH, C₀, t	Adsorption capacity	372.39 mg/g	[193]
Waste acorns of Quercus ithaburensis	Radioactive gallium-67	Multi-layered perceptron feed forward Levenberg–Marquardt back propagation (3-8-5-1)	Kinetics	172-44-44	T, m, t	Adsorption capacity	91.80%	[194]
Bacillus cereus biomass	As³⁺	Feed forward back propagation	Isotherms and Kinetics	128-43-0	C₀, m, T, t	Removal percentage	85.24%	[195]
Alga Cladophora	Malachite green	Three layered feed forward back propagation (5-14-1)	Equilibrium and Kinetics	132-44-44	C₀, m, pH, t, T	Removal percentage	100%	[196]
Alga Xanthophyta sp	Malachite green	Three layered feed forward back propagation (5-12-1)	Kinetic, RSM	180-60-60	C₀, m, pH, T, t	Removal percentage	100%	[197]
Polyaniline (P)	Sodium dodecyl benzene sulfonate (SDBS)	Feed forward back propagation (5-5-5-1)	Equilibrium	69-0-69	pH, m, T, t, v	Adsorption capacity	P- ZnCl₂ 29.5 mg/g, P-CuCl₂ 32.3 mg/g	[198]
Leachate by hazelnut shell	Zn²⁺	Three layered feed forward Levenberg–Marquardt back propagation (4-8-4)	Kinetics	100 data	pH, m, T, t	Removal percentage	98%	[199]
Leucaena Leucocephala seeds	Ni²⁺	Levenberg–Marquardt back propagation	Kinetics	72-24-24	C₀, m, t, V	Removal percentage	76.23%	[200]
Black cumin	Pb²⁺	Multi-Layer perceptrón	Equilibrium, RSM (CCD)	15-11-0	T, m, pH	Adsorption capacity	8.08 mg/g	[201]
Walnut husk	Lanaset red G	Multi-layered feed forward (3-20-1)	Kinetic	278-65-65	C₀, pH, t	Adsorption capacity	187.74 mg/g	[202]
Coconut shell	Indigo	Feed forward back propagation (6-5-5-1)	Equilibrium	152-21-0	Type of adsorbent, pH, m, Ps, C₀, type of activation	Removal percentage	94.83%	[203]
Activated carbon	Direct blue 86	Levenberg–Marquardt back propagation (5-(5-12)-1)	Kinetics	100-33-33	C₀, m, t, T, pH,	Adsorption capacity	2.26 mg/g	[204]
Organoclay	Reactive red 141	Three layered feed forward scaled conjugate gradient back propagation (4-5-1)	Kinetics	85-15-0	C₀, t, m, T	Removal percentage	100%	[205]
Nanocellulose fibers	Pb²⁺	Multi-layered feed forward Levenberg–Marquardt back propagation	Equilibrium and Kinetics	108-36-36	C₀, pH, m, t, V	Removal percentage	94.21%	[206]
Alga Lemna minor L.	Acid blue 92	Three layered feed forward back propagation (5-14-1)	Kinetics	57-17-17	C₀, m, pH, T, t	Removal percentage	91%	[207]
Acinetobacter sp. bacteria	Cr⁶⁺	Three-layered feed forward back propagation (3-7-1)	Equilibrium and Kinetics	48-22-0	C₀, m, t	Removal percentage	90.91%	[208]
Eggshells	Crystal violet	Three layered feed forward Levenberg–Marquardt back propagation (4-8-1)	Equilibrium	105-52-53	pH, m, C₀, T	Removal percentage	99%	[209]
Activated carbon	Pb²⁺	Resilient back propagation (5-12-1)	Kinetics	10-5-5	pH, C₀, m, t, T	Removal percentage	--	[210]
Nanometer SiO2	Mn²⁺, Co²⁺	Hybrid ant colony optimization-ANN Levenberg–Marquardt (4-7-5-1 for Mn(II), 6-9-1 for Co(II))	Kinetics	37-10-10	pH, m, t, concentration of 1-(2-pyridylazo)-2-naphthol	Removal percentage	Mn²⁺ 99 %, Co²⁺ 98%	[211]
Zea mays Cob Powder	As³⁺, As⁵⁺	Levenberg–Marquardt back propagation (4-14-1)	Kinetics	108-36-36	C₀, m, t, V	Removal percentage	As³⁺ 81.88 %, As⁵⁺ 92.61 %	[212]
Moringa oleifera seeds	Methylene blue, Congo red	Levenberg–Marquardt back propagation (4-14-1)	Kinetics	108-36-36	C₀, m, t, V	Removal percentage	MB 90.27%, CR 98.52%	[213]
Carbon nanotubes	Methylene blue	Three layered feed forward Levenberg–Marquardt back propagation (5-10-1)	Equilibrium and Kinetics	43-21-21	C₀, m, T, pH, t	Final concentration	0.3-36.79 mg/L	[214]
Immobilized Bacillus subtilis	Cd²⁺	Feed forward Levenberg–Marquardt back propagation (5-8-1)	Equilibrium and Kinetics	63-14-13	C₀, m, t, T, pH	Adsorption capacity	251.91 mg/g	[215]
Co (III) oxide nanoparticle loaded on activated carbon	Eosin Y	Multi-layered feed forward - genetic algorithm (3-23-1)	Kinetics	176-38-38	T, m, t	Removal percentage	>90%	[216]
Lemna major biomass	Carbaryl	Multi-layered perceptron (4-6-1)	Equilibrium and Kinetics, RSM	22-0-7	C₀, pH, m, t	Removal percentage	95%	[217]
Chromolaena odorata	Trichlorophenol	Two layered feed forward Levenberg–Marquardt back propagation (2-10-1)	Kinetics	-	C₀, t	Removal percentage	99%	[218]
Poly(vinyl alcohol) PVA/chitosan nanofiber membrane	Ni²⁺, Co²⁺	Multi-layered perceptron (6-3-2-1)	Equilibrium and kinetics	70-30-0	C₀, t, pH, m, T	Removal percentage	Ni²⁺ 79.28%, Co²⁺ 77.12%	[219]
Copper sulfide nanoparticles loaded on activated carbon	Reactive orange 12	Principal component analysis-ANN Levenberg–Marquardt back propagation (3-12-1)	Equilibrium and Kinetics	182-39-39	C₀, m, t	Removal percentage	99.54%	[220]
Zn(OH)2 nanoparticles-activated carbon	Sunset yellow	Three-layered feed-forward–Imperialist competitive (3-9-1)	Equilibrium and Kinetics	224-96-0	C₀, m, t	Removal percentage	90%	[221]
Gold nanoparticles loaded on activated carbon and activated carbon Tamarisk	Methyl orange	Three layered feed forward Levenberg–Marquardt back-propagation and particle swarm (4-11-1)	Equilibrium and Kinetics	182-78-0	C₀, m, t	Removal percentage	>98%	[222]
Graphite oxide nanoparticle	Methylene blue, Brilliant green	Three layered feed forward Levenberg–Marquardt back propagation and genetic algorithm (3-23-1)	Equilibrium and Kinetics	235-101-0	C₀, m, t	Removal percentage	MB 97.4%, BG 97.3%	[223]
Gold nanoparticles-activated carbon	1,3,4-thiadiazole-2,5-dithiol	Adaptive neuro-fuzzy inference system (3-9-27-27-1)	Equilibrium and Kinetics	168-72-0	C₀, m, t	Removal percentage	>90 %	[224]
ZnO Nanoparticle	As³⁺	Feed-Forward Levenberg–Marquardt back propagation (5-20-1)	Equilibrium and Kinetics	47-24-24	C₀, t, pH, m, v	Removal percentage	98%	[225]
Turkish lignite	Methylene blue	Three-layered feed forward back propagation (6-14-1)	Equilibrium and Kinetics	60-20-20	C₀, t, m, pH, v	Adsorption capacity	41.94 mg/g	[226]
Clinoptilolite	Cu²⁺	Three layered back propagation (3-11-1)	Equilibrium and Kinetics	63-14-13	pH, T, C₀	Removal percentage	>60%	[227]
Activated carbon	Methylene blue	Hybrid multi-layered perceptron - genetic algorithm (5-7-1)	Equilibrium and Kinetics, RSM	70-27-11	pH, v, m, C₀, t	Removal percentage	99.75%	[228]
Silver nanoparticles	Molybdenum	Hybrid ANN–particle swarm (4-12-1)	Equilibrium and Kinetics	187-31-42	C₀, pH, m, t	Removal percentage	92.50%	[229]
Borasus Flabellifer Coir Powder	Cr⁶⁺	Feed forward back propagation (3-18-1)	Equilibrium	38-16-0	pH, m, C₀	Removal percentage	>95 %	[230]
De-oiled algal biomass	Methylene blue	Feed forward Levenberg-Marquardt back propagation (6-10-1)	Equilibrium and Kinetics	369-79-79	pH, m, v, T, C₀	Removal percentage	>90 %	[231]
Hydrochar	Acridine orange, Rhodamine 6G	Levenberg–Marquardt back propagation (3-(1-25)-1)	Equilibrium and Kinetics	50-25-25	T, t, pH	Adsorption capacity	AO 70.36 mg/g, R6G 60.13 mg/g	[232]
Crab shell	Pb²⁺, Zn²⁺	Multi-layered perceptron back propagation	Equilibrium	-	-	Adsorption capacity	Pb²⁺ 564 mg/g, Zn²⁺ 71.46 mg/g	[233]
Sargassum filipendula	Ni²⁺, Cr³⁺	Multi-layered feed forward (2-2-2-2)	Equilibrium and Kinetics mono-Binary	26 data for training	Ce	Adsorption capacity	Ni²⁺ 1 mg/g, Cr³⁺ 2 mg/g	[234]
Microcrystalline cellulose	Cr³⁺, Cr⁶⁺	Multi-layered feed forward Levenberg–Marquardt back propagation	Equilibrium and Kinetics	160 data	C₀, t, m	Removal percentage	Cr³⁺ 42.36%; Cr⁶⁺ 5.88%	[235]
Light expended clay aggregate	Cu²⁺	Three layered feed forward back propagation (4-4-1)	Equilibrium and Kinetics RSM (CCD)	27-8-0	pH, T, C₀, m	Adsorption capacity	113.64 mg/g	[236]
Bamboo biochar	Acid black 172	Three layered feed forward Levenberg–Marquardt back propagation	Equilibrium and Kinetics	108-36-36	pH, T, C₀, t, ionic strength	Adsorption capacity	81.88 mg/g	[237]
Intercalated tartrate-Mg–Al layered double hydroxides	Pb²⁺	Hybrid multi-layered feed forward Levenberg–Marquardt back propagation – genetic algorithm	Equilibrium	19-3-3	C₀, t, m pH	Removal percentage	99.32%	[238]
Bauxsol	Phosphorus	Three layered feed forward back propagation (3-5-1)	Equilibrium	17-5-0	T, t, HCl concentration	Adsorption capacity	55.72 mg/g	[239]
Graphene oxide nanoplatelets	Safranin	Three layered feed forward back propagation (3-10-1)	Equilibrium and Kinetics	--	pH, m, T, t	Removal percentage	97.78%	[240]
Walnut husk	Basic red 46	Three layered feed forward back propagation (5-25-1)	Kinetics	1469-345-346	Ps, m, pH, C₀, t	Adsorption capacity	66.45 mg/g	[241]
Biomass	Ni²⁺, Co²⁺	Multi-layered feed forward	Equilibrium and Kinetics	70-30-0	pH, m, t	Removal percentage	Ni²⁺ 90.25%, Co²⁺ 86.13%	[242]
Activated carbon	Sunset yellow	Three layered back propagation (3-(1-20)-1)	Equilibrium and Kinetics	218-47-47	C₀, m, t	Removal percentage	>98 %	[243]
Copper nanowires loaded on activated carbon	Malachite green	Hybrid three layered feed forward -genetic algorithm (3-(1-30)-1)	Equilibrium and Kinetics	186-62-0	C₀, m, t	Removal percentage	99%	[244]
Zinc sulfide nanoparticle loaded on activated carbon	Brilliant green	Three-layered feed forward Levenberg–Marquardt back-propagation and hybrid ANN-particle swarm optimization (3-(1-35)-1)	Equilibrium and Kinetics	176-76-0	C₀, m, t	Removal percentage	>80 %	[245]
Gold and titanium dioxide nanoparticles loaded on activated carbon	Phenol red	Hybrid three layered feed forward -genetic algorithm	Equilibrium and Kinetics	70-15-15	pH, C₀, m, t	Removal percentage	Ti-NP-AC (97.3%), Au-NP-AC (97.4%)	[246]
Spent tea leaves	Cu²⁺	Feed forward Levenberg–Marquardt back propagation (3-10-4-1)	Equilibrium and Kinetics, RSM (CCD)	12-4-4	pH, m, t	Removal percentage	96.12%	[247]
Zinc oxide nanorods loaded on activated carbon	Sunset yellow	Levenberg–Marquardt back propagation (3-6-1)	Equilibrium and Kinetics	189-81-0	C₀, m, t	Removal percentage	98.89%	[248]
Hybrid material (Ce-HAHCl)	As³⁺	Feed forward back propagation (6-7-1)	Equilibrium and Kinetics	60-40-0	C₀, m, pH, t, v, T	Removal percentage	98.80%	[249]
Cerium oxide polyaniline composite	Cr⁶⁺	Three layered feed forward back propagation (5-7-1)	Equilibrium and Kinetics	52-34-0	C₀, m, t, pH, T	Removal percentage	93.90%	[250]
Orange peel ash	Phenol	Multi-layer perceptron feed forward back propagation (6-6-1)	Equilibrium and Kinetics	--	C₀, pH, m, t, v, T	Removal percentage	97.34%	[251]
Apricot stones and corn cob	Cu²⁺	Multi-layer perceptron and Levenberg–Marquardt back propagation (3-8-6)	Equilibrium	144-48-48	C₀, m, V	Removal percentage, adsorption capacity and Ce	AS 5.9 mg/g, CC 5.3 mg/g	[252]
Sepiolite-stabilized zero-valent iron nanocomposite	Cr⁶⁺	Three-layered feed forward, train scaled conjugate gradient (4-8-1)	Kinetics	71-23-23	pH, C₀, m, chloride ion concentration.	Removal percentage	99.78%	[253]
Olive stone	Pb²⁺	Neural fuzzy inference system and multi-layered feed forward back propagation	Equilibrium	27-3-0	Chemical solution concentration for treatment, pH, C₀	Adsorption capacity	25.45mg/g	[254]
Azadirachta indica (neem) bark powder	As⁴⁺	Three-layered feed forward (7-4-2)	Kinetics	14-7-7	pH, t, C₀, m, v, T	Removal percentage and adsorption capacity	95%	[255]
Biomass of Rhizopus arrhizus	Reactive orange 13	Three-layered feed forward (4-3-1)	Equilibrium and Kinetics	25-0-5	pH, m, C₀, t, v	Adsorption capacity	132.63 mg/g	[256]
Hyacinth roots, neem leaves and coconut shells	Pb²⁺	Levenberg–Marquardt back propagation and scaled conjugate gradient	Kinetics	114 data	pH, C₀, m, t, type of adsorbent	Removal percentage	98.67%	[257]
Sargassum bevanom (S. bevanom) algae	Hg²⁺	Multi-layered feed forward, Levenberg–Marquardt back propagation (4-(4-7)-1)	Equilibrium and Kinetics	26-6-6	m, C_0, t, pH.	Removal percentage	90.24%	[258]
Valonia tannin resin	Cd²⁺	Four-layered fast artificial neural network (6-25-5-1)	Equilibrium and Kinetics	365-91-0	T, pH, C₀, v, Ps, t	Removal percentage	63.29 mg/g	[259]
ZnS nanoparticles loaded activated carbon	Brilliant green, eosin B	Multi-layered feed forward and Levenberg–Marquardt back propagation	Equilibrium, Kinetics and RSM (CCD)	41-0-13	C₀, t, m	Removal percentage	100%	[260]
Rice straw and Fe3O4 nanoparticles	Pb²⁺, Cu²⁺	Quick propagation, batch back propagation, genetic algorithm and Levenberg-Marquardt (3-9-2)	Kinetics	15-0-5	C₀, m, t	Removal percentage	Pb(II) 91.18%, Cu(II) 75.98%	[261]
CaO₂ nanoparticle	Benzeneacetic acid	Multi-layered perceptron feed forward (2-0-5-5-3)	Equilibrium	16 data	C₀, m	Ce, amount of CaO₂ and removal percentage	97.47%	[262]
Activated carbon (Bean husk)	Ibuprofen	Three-layered feed forward back propagation (3-x-1)	Equilibrium, Kinetics and RSM (CCD)	20 data	C₀, m, t, pH, v, T	Removal percentage	99%	[263]
Flax meal biomaterial	Cu²⁺	Multi-layer feed forward perceptron	Equilibrium, RSM (CCD)	528 data	C₀, m, pH	Removal percentage	40.11 mg/g	[264]
Zn(OH)₂ nanoparticles on activated carbon	Methylene blue	Multi-layered feed forward Levenberg Marquardt back propagation	Equilibrium, RSM (CCD)	21-5-5	C₀, pH, m, t	Removal percentage	98.70%	[265]
Carbon nanotube	Microcystins-LR	Three-layered Levenberg–Marquardt back propagation (4-13-1)	Kinetic	46-10-10	pH, C₀, t, m	Removal percentage	175.43 μg/g	[266]
Copper oxide nanoparticle loaded on activated carbon	Pb²⁺, malachite green	Three-layered feed forward (5-20)-2)	Isotherm, RSM (CCD)	20-6-6	C₀, pH, t, m	Removal percentage	Pb²⁺ 98.33 mg/g, MG 87.72 mg/g	[267]
Parthenium hysterophorus derived activated char	Ranitidine hydrochloride	Feed forward back-propagation	Kinetics, RSM (CCD)	--	C₀, m, t, pH, T, v	Removal percentage	99.65%	[268]
Graphene oxide	Fluoride	Three-layered feed forward, Levenberg–Marquardt back propagation	Kinetic and isotherm, RSM (CCD)	14-2-4	m, t, T	Removal percentage	28.72 mg/g	[269]
Tamarindus indica	As³⁺	Hybrid ANN-genetic algorithm (4-6-1)	Kinetic, isotherm and RSM (CCD)	24-0-7	C₀, t, pH, m, T	Adsorption capacity	91.10%	[270]
Zn-loaded pinecone biochar	As³⁺	Hybrid back propagation ANN-genetic algorithm	Kinetic and isotherm, RSM	13-4-0	C₀, pH, ratio Ethanol - adsorbate	Adsorption capacity	10.47 μg/g	[271]
Iron composite	Amido black	Three-layered feed forward (5-8-1)	Kinetics	60-20-20	pH, t, C₀, T, m	Removal percentage	88%	[272]
Iron/olivine composite	As³⁺	Multi-layered perceptron (5-(12-14)-1)	Kinetics, RSM	22-8-8	C₀, m, t, v, pH	Removal percentage	99%	[273]
Graphene-like carbon	Methylene blue	Multi-layered perceptron (4-13-1)	Kinetics	73-16-16	C₀, m, t, pH	Removal percentage	20 mg/g	[274]
Activated carbon	Ni²⁺	ANN and adaptive neuro–fuzzy inference system	Kinetic and isotherm	420-90-90	C₀, t, pH, BET area	Adsorption capacity	63.50 mg/g	[275]
Activated carbon (Medlar seed)	Cr²⁺	Multi-layered perceptron back propagation (5-1-8)	Kinetic and isotherm	41-0-18	pH, C₀, m, t	Removal percentage	200 mg/g	[276]
Sunflower seed husk	Cu²⁺	Levenberg–Marquardt back propagation (6-20-1)	Kinetic and equilibrium	119-39-39	C₀, flow rate, dosage and surfactant concentration, t, pH, T	Removal percentage	--	[277]
Chitosan	Acid orange 7	Multi-layer perceptron (4-5-1)	Kinetic and isotherm, RSM	26-4-4	pH, C₀, m, t	Removal percentage	98%	[278]
ZnO nanoparticles and copper-metal organic frameworks	Cyanide	Multi-layer perceptron and Levenberg–Marquardt back propagation (1-35-5)	Kinetic, isotherm, RSM	70-15-15	pH, t, T, m, C₀	Removal percentage	6.34 mg/g	[279]
Ho-CaWO4	Methylene blue	Multi-layer perceptron and Levenberg–Marquardt back propagation (4-8-1)	Kinetics, RSM (CCD)	15-3-3	pH, t, m, C₀	Removal percentage	71.17%	[280]
Catecholamine/starch	Methylene blue	Hybrid multi-layer feed forward - particle swarm (3-7-1)	Kinetics, RSM (CCD)	--	pH, C₀, t	Adsorption capacity	2276.5 mg/g	[281]
Neodymium modified ordered mesoporous carbon	Sunset yellow	Three-layered back propagation	Kinetic and isotherm	47-10-10	C₀, t, m	Removal percentage	--	[282]
Biochar	Methylene blue	Feed forward back propagation	Isotherm	504-108-108	Adsorbent type, Ps, C₀, pyrolysis temperature	Adsorption capacity	0.047 mmol/g	[283]
Olive pips	Pb²⁺	Three-layered feed forward back propagation	Kinetic and equilibrium	60-20-20	T, t, pH, m, C₀, v	Removal percentage	99%	[284]
Microwave-synthesized akaganeite nanoparticles	Congo red	Back-propagation	Kinetic and isotherm	27-5-5	C₀, t, m	Removal percentage	>150 mg/g	[285]

Nomenclature: adsorbent dosage (m), Box-Behnken Design (BBD), central composite design (CCD), final adsorbate concentration (Ce), initial adsorbate concentration (), particle size (Ps), response surface methodology (RSM), temperature (), time (), stirring rate (), and volume ().