A novel approach based on recurrent neural networks applied to nonlinear systems optimization

This paper presents an efficient approach based on recurrent neural network for solving nonlinear optimization. More specifically, a modified Hopfield network is developed and its internal parameters are computed using the valid subspace technique. These parameters guarantee the convergence of the network to the equilibrium points that represent an optimal feasible solution. The main advantage of the developed network is that it treats optimization and constraint terms in different stages with no interference with each other. Moreover, the proposed approach does not require specification of penalty and weighting parameters for its initialization. A study of the modified Hopfield model is also developed to analyze its stability and convergence. Simulation results are provided to demonstrate the performance of the proposed neural network.     

Use of modified rice husks as a natural solid adsorbent of trace metals: characterisation and development of an on-line preconcentration system for cadmium and lead determination by FAAS

The use of rice husks as an alternative adsorbent in an on-line preconcentration system for Cd (II) and Pb (II) determination by flame atomic absorption spectrometry (FAAS) is described. The potential of rice husks as a natural adsorbent was evaluated as a material modified with 0.75 mol l⁻¹ NaOH solution and in the unmodified form. For this task, several techniques such as spectroscopy and thermogravimetry were used for elucidation of possible functional groups responsible for the uptake of Cd (II) and Pb (II). Furthermore, based on adsorption studies and adsorption isotherms applied to the Langmüir model, it was possible to verify that modified rice husks present a higher adsorption capacity for both metals. After establishing this material as a promising natural adsorbent, it was used for on-line preconcentration of Cd (II) and Pb (II) metals. The multivariate optimisation of chemical and flow variables was performed by using a full factorial design (2⁴) including the following factors: preconcentration time, preconcentration flow rate, concentration and volume of eluent. The optimum pH values used for on-line preconcentration were taken from prior univariate experiments. Under optimised conditions for Cd (II) determination (4 min of preconcentration at a 6 ml min⁻¹ preconcentration flow rate, in which comprises 24 ml of preconcentration volume, 200 μl elution volume and 1.0 mol l⁻¹ HNO₃ solution as eluent), the system achieved a detection limit of 1.14 μg l⁻¹ and an enrichment factor of 72.4. Similar conditions were used for Pb (II) determination (4 min of preconcentration, 6 ml min⁻¹ preconcentration flow rate, 300 μl elution volume and 1.0 mol l⁻¹ HNO₃ solution as eluent) from which a detection limit of 14.1 μg l⁻¹ and enrichment factor of 46.0 were achieved. Also, rice husks have been shown to be a homogeneous and stable adsorbent in which more than 100 preconcentration/elution cycles provide a relative standard deviation (RSD) of less than 6.0% on the analytical signal. The satisfactory accuracy of the method developed was obtained by using spiked water samples (mineral water and lake water) and spiked red wine samples. These values were confirmed by electrothermal atomic absorption spectrometry (ETAAS). The certified reference material [pig kidney (CRM 186)] and the reference material [beech leaves (CRM 100)] were also used.     

Trends in Preconcentration Procedures for Metal Determination Using Atomic Spectrometry Techniques

 Methods for metal preconcentration are often described in the literature. However, purposes are often different, depending on whether the methods are applied in environmental, clinical or technological fields. The respective method needs to be efficient, give high sensitivity, and ideally also is selective which is useful when used in combination with atomic spectroscopy. This review presents the actual tendencies in metal preconcentration using techniques such as Flame Atomic Absorption Spectrometry (FAAS), Electrothermal Atomic Absorption Spectrometry (ETAAS), Hydride Generation Atomic Absorption Spectrometry (HGAAS), Inductively Coupled Plasma Optical Emission Spectrometry (ICP OES) and Inductively Coupled Plasma Mass Spectrometry (ICP-MS). Procedures based on related to electrochemical, coprecipitation/precipitation, liquid–liquid and solid–liquid extraction and atom trapping mechanisms are presented. Correspondence: Department of Analytical Chemistry, Universidade Estadual de Campinas – UNICAMP, P.O. Box 6154, 13083-970 Campinas, Brazil. e-mail: zezzi@iqm.unicamp.br Received December 20, 2001; accepted October 11, 2002     

Metal furnace heated by flame as a hydride atomizer for atomic absorption spectrometry: Sb determination in environmental and pharmaceutical samples

The present work describes a metallic hydride atomizer for atomic absorption spectrometry, by evaluating the performance of the Inconel 600^® tube. For this purpose, stibine was used as the model volatile compound and antimony determination in river and lake sediments and in pharmaceutical samples was carried out to assess the metal furnace performance. Some parameters are evaluated such as those referring to the generation and transport of the hydride (such as KBH₄ and acid concentrations, carrier gas flow rate, injected volume, etc.), as well as those referring to the metal furnace (such as tube hole area, flame composition, long-term stability, etc.). The method presents linear Sb concentration from 2 to 80 μg L⁻¹ range (r > 0.998; n = 3) and the analytical frequency of ca. 140 h⁻¹. The limit of detection (LOD) is 0.23 μg L⁻¹ and the precision, expressed as R.S.D., is less than 5% (40 μg L⁻¹; n = 10). The accuracy is evaluated through the reference materials, and the results are similar at 95% confidence level according to the t-test.     

On-line electrolytic dissolution of alloys in flow-injection analysis : Part 1. Principles and application in the determination of soluble aluminium in steels

A fast system for steel analysis is described. Immediately after collection, the solidified sample is polished and placed on a small electrochemical dissolution unit. The dissolution step is accomplished in a few seconds, and the dissolved material is passed directly “on line” to a flow-injection manifold. The feasibility of the approach is demonstrated by determining soluble aluminium in steels (0.01–0.13% w/w) spectrophotometrically. Dissolution conditions such as current density, electrolysis duration, composition and flow rate of the electrolyte carrier stream are described. The proposed system is very stable and the consumption of reagents is low; 20–40 samples per hour can be handled. The results compare well with the values for reference standard steels.     

Sample preparation for metalloprotein analysis: A case study using horse chestnuts

In the present work, 11 different procedures for protein and metalloprotein extraction from horse chestnuts (Aescullus hippocastanum L.) in natura were tested. After each extraction, total protein was determined and, after protein separation through sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE), those metals belonging to the protein structure were mapped by synchrotron radiation X-ray fluorescence (SRXRF). After mapping the elements (Cr, Fe and Mn) in the protein bands (ca. 33 and 23.7 kDa), their concentrations were determined using atomic absorption spectrometry (ET AAS).

Good results were obtained for protein extraction using a combination of grinding and sonication. However, this strategy was not suitable to preserve metal ions in the protein structure. In fact, there was 42% decrease on Mn concentration using this procedure, compared to that performed with sample agitation in water (taken as reference). On the other hand, when grinding and agitation with an extracting buffer was used, there was a 530% increase of Mn concentration, when compared to the reference procedure.

These results indicate agreement between metal identification and determination in proteins as well as the great influence of the extraction procedure (i.e., the sample preparation step) for preserving metals in the protein structures.     

Evaluation of a synergetic effect between Rh as permanent chemical modifier and acetylacetone as complexing agent in Sc determination in sediment slurry samples by ETAAS

In the present work, scandium was determined in sediment slurry samples (from three different rivers) by electrothermal atomic absorption spectrometry (ETAAS). Slurries were prepared by weighting 100 mg of dry sediment samples (≤53 μm particle sizes) and adding 6 ml of HCl:HNO₃:HF (3:1:2, v/v). Accurate results were only possible due to the synergetic effect between Rh as permanent chemical modifier and acetylacetone (Acac) as complexing agent. The same platform was used for 400 heating cycles. The performance of the chemical modification was evaluated by using scanning electron microscopy (SEM), synchrotron radiation X-ray fluorescence (SRXRF) and some figures of merit (precision and detectability). The best analytical conditions were attained using 1500 and 2550 °C as pyrolysis and atomization temperatures. The scandium content in the liquid phase of the slurries ranged from 61 to 73%, thus indicating, in this study, that both liquid and solid phases play an important role in slurry analyses. An amount of 5.0–20.0 μg l⁻¹ Sc linear range as well as LOD and LOQ of 0.19 and 0.62 μg l⁻¹, respectively, were obtained under these conditions. The accuracy was checked by using microwave-assisted decomposition, and the results compared to those obtained with the proposed methodology (slurry analysis). By checking both sets of the results, there is no statistical difference at the 95% confidence levels.     

Evaluation of soybean seed protein extraction focusing on metalloprotein analysis

Two methods of protein extraction for soybean seeds were evaluated in terms of preservation of the metal ions bound to proteins after the extraction and separation procedures. The proteins were firstly separated according to their molar masses by polyacrylamide gel electrophoresis. Then, the protein bands were mapped by synchrotron radiation X-ray fluorescence in order to establish which metal ions were present in each one. Finally, some mapped protein bands were decomposed by microwave-assisted combustion and Ca, Cu, K, Mg, Mn, and Zn were quantified by inductively coupled plasma mass spectrometry or inductively coupled plasma optical emission spectrometry. The extraction methods studied were Method A (based on the treatment of ground soybean seeds with hexane and their extraction with Tris–HCl and β-mercaptoethanol) and Method B (based on the treatment of ground soybean seeds with petroleum ether and their extraction with Tris–HCl, dithiothreitol, phenylmethanesulfonyl fluoride, sodium dodecyl sulfate and potassium chloride). The best method was Method B, in which a 78% higher extraction efficiency was obtained when compared to Method A. Additionally, the metal-protein interactions were more appropriately preserved when Method B was applied, where the most affected ions were those that are bound weakly to proteins, such as Ca, K, and Mg.     

Electrochemical DNA biosensor for human papillomavirus 16 detection in real samples

An electrochemical DNA biosensor for human papillomavirus (HPV) 16 detection has been developed. For this proposed biosensor, l-cysteine was first electrodeposited on the gold electrode surface to form l-cysteine film (CYSFILM). Subsequently, HPV16-specific probe was immobilized on the electrode surface with CYSFILM. Electrochemistry measurement was studied by differential pulse voltammetry method (DPV). The measurement was based on the reduction signals of methylene blue (MB) before and after hybridization either between probe and synthetic target or extracted DNA from clinical samples. The effect of probe concentration was analyzed and the best results were seen at 1000 nM. The hybridization detection presented high sensitivity and broad linear response to the synthetic-target concentration comprised between 18.75 nM and 250 nM as well as to a detection limit of 18.13 nM. The performance of this biosensor was also investigated by checking probe-modified electrode hybridization with extracted DNA from samples. The results showed that the biosensor was successfully developed and exhibited high sensitivity and satisfactory selectivity to HPV16. These results allow for the possibility of developing a new portable detection system for HPVs and for providing help in making an effective diagnosis in the early stages of infection.     

Contribution of CuxO distribution,shape and ratio on TiO2 nanotubes to improve methanol production from CO2 photoelectroreduction

Many studies are focused on the development of materials for converting carbon dioxide into multicarbon oxygenates such as methanol and ethanol, because of their higher energy density and wider applicability. In this work, TiO₂ nanotubes (NT/TiO₂) were modified with Cu_xO nanoparticles in order to investigate the contribution of different ratio of Cu₂O/CuO and its distribution over NT/TiO₂ for CO₂ photoelectro-conversion to methanol. The photoelectrodes were built by anodization process to obtain NT/TiO₂ layer, and the decoration with Cu_xO hybrid system was carried out by electrodeposition process, using Na₂SO₄ or acid lactic as electrolyte, followed by annealing at different temperatures. X-ray photoelectron spectroscopy analysis revealed the predominance of Cu⁺¹ and Cu⁺² at 150 °C and 300 °C, respectively. X-ray diffraction and scanning electron microscopy indicated that under lactic acid solution, the oxide nanoparticles exhibited small size, cubic shape, and uniform distribution on the nanotube wall. While under Na₂SO₄ electrolyte, large nanoparticles with two different morphologies, octahedral and cubic shapes, were deposited on the top of the nanotubes. All modified electrodes converted CO₂ in methanol in different quantities, identified by gas chromatograph. However, the NT/TiO₂ modified with CuO/Cu₂O (80:20) nanoparticles using lactic acid as electrolyte showed better performance in the CO₂ reduction to methanol (0.11 mmol L⁻¹) in relation to the other electrodes. In all cases, a blend among the structures and nanoparticle morphologies were achieved and essential to create new site of reactions what improved the use of light irradiation, minimization of charge recombination rate and promoted high selectivity of products.