Satu-Pia Reinikainen

Interactions of soil components and their effects on speciation of chromium in soils

For evaluation of real risks of soil contaminated with chromium it is important to identify and understand the reactions of Cr species with main soils components. In this research reactions of Cr(III) and Cr(VI) with Fe(OH)3, MnO2, CaCO3, kaolinite and natural organic matter (NOM) were examined in batch experiments. Since the batches were compositions of soil component mixtures the experimental design was carried out by applying Simplex algorithm and the results of the experimental domain were interpreted by partial least squares (PLS). The research confirmed the significance of studying all the soil components simultaneously instead of studying only the effects of one component at a time. The ability of pure components to change the speciation of Cr(III) and Cr(VI) in liquid phase was different to that in mixtures, which indicated that the soil components interacted with each other. The order of pure components to convert Cr(III) by adsorption, precipitation or oxidation was Fe(OH)3=kaolinite=CaCO3>NOM>MnO2, while in mixtures the order was: Fe(OH)3>CaCO3>kaolinite>MnO2>NOM. This means that in mixtures with the highest amount of Fe(OH)3 the conversion of Cr(III) was highest. In mixtures with the highest amount of NOM the changes in speciation of dissolved Cr(III) were the lowest. The order for the pure components to change the speciation of Cr(VI) was: Fe(OH)3>NOM>kaolinite>CaCO3>MnO2 and in mixtures the order was: NOM>Fe(OH)3>kaolinite>MnO2>CaCO3. The reactions between different soil components also affected the oxidation/reduction ability of soils. As separate component MnO2 oxidized Cr(III) to Cr(VI). However, in mixtures with NOM or Fe(OH)3 the oxidation was hindered. The effect of pH on speciation of Cr in mixtures was also examined and is discussed.

Analysis of protein filtration data by PLS regression

Six different globular proteins were filtered with hydrophilic regenerated cellulose ultrafiltration membranes at their isoelectric points. The PLS regression was used to analyse the filtration data. The modelling of the relative flux failed, but that of the transmission of the proteins succeeded relatively well. As a conclusion of the data analysis the most important explanatory variables affecting the transmission of proteins are the permeate flux and the membrane cut-off value divided by the molar mass of the protein. They both have a positive linear relationship with the transmission.

Raman spectroscopic imaging of indomethacin loaded in porous silica

Loading of a poorly soluble drug such as indomethacin (IMC) into porous silica particles enhances its dissolution upon administration. The distribution of the different solid forms in which IMC may appear was studied using Raman spectroscopy. Raman mapping of the samples was performed with a Raman microscope equipped with an automated xy-stage. The spectral data were extracted in the range 1500–1750 cm−1, which represents the stretching of the CO bond in the IMC molecule. To alleviate the problem of overlapping peaks in the Raman spectra of the different IMC forms, the spectral data were analyzed using partial least squares (PLS) and principal component analysis (PCA). Despite the problems caused by fluorescence, the method gave valuable information about the occurrence and distribution of the solid forms of IMC. The same approach was utilized for analysis of the heterogeneity of recrystallized IMC samples, and PCA was shown to be capable of revealing the presence of solvates or polymorphs not included in the model.

Measurement uncertainty of lactase-containing tablets analyzed with FTIR

Uncertainty is one of the most critical aspects in determination of measurement reliability. In order to ensure accurate measurements, results need to be traceable and uncertainty measurable. In this study, homogeneity of FTIR samples is determined with a combination of variographic and multivariate approach. An approach for estimation of uncertainty within individual sample, as well as, within repeated samples is introduced. FTIR samples containing two commercial pharmaceutical lactase products (LactaNON and Lactrase) are applied as an example of the procedure. The results showed that the approach is suitable for the purpose. The sample pellets were quite homogeneous, since the total uncertainty of each pellet varied between 1.5% and 2.5%. The heterogeneity within a tablet strip was found to be dominant, as 15-20 tablets has to be analyzed in order to achieve <5.0% expanded uncertainty level. Uncertainty arising from the FTIR instrument was <1.0%. The uncertainty estimates are computed directly from FTIR spectra without any concentration information of the analyte.

Detection of Novel Carbohydrate-Related Compounds in Aqueous Samples Using a Capillary Electrophoretic Profiling Method

The aim of this research was to extend an existing capillary electrophoresis (CE) method, originally developed for the determination of mono- and disaccharides, to the determination of alternative carbohydrate compounds, namely furfural and polydatin. Empirical validation confirms that this novel method can be applied for the determination of analyte concentrations from complex matrices, and the evaluation of their carbohydrate composition. It is concluded that the approach has validity as an analytical procedure and has the ability to determine industrially important analytes from a heterogeneous biological sample matrix, and thus, the method adds value to the development of large scale separation processes. However, some additional optimization is required before online applications.

Evaluation of variation in dynamic processes via online spectrometers

Theory of sampling offers powerful tools for process optimization. An adequate sampling interval can be determined for spectral measurements when utilizing a multivariate extension of variography by applying score vectors as independent sources of uncertainty. The traditional way is to apply variographic analyses into single process variables independently. In the multivariate extension, those process variables are replaced with score vectors of principal component analysis. The combined uncertainty found this way depends not only on the variance in the spectra, but also, for example, on the number of utilized score vectors and the preprocessing method. This approach is illustrated with a crystallization process continuously followed with an attenuated total reflectance Fourier transform infrared instrument. The results show that the approach is highly applicable but should only be utilized as an indicative tool. Copyright

On the effectiveness of cross-fitting in multi-block PLS (CF-MBPLS)

Multi‐block PLS is an extension of partial least squares or projection to latent structures (PLS), where the descriptor matrix is divided into meaningful blocks based on either process units or type of data. A typical application is using process variables as one block and spectral data on another block. It has been utilized in obtaining more information of processes and the effect of different types of variables. In comparison with priority or hierarchical PLS, in multi‐block PLS, there is no need to prioritize blocks in advance because they are iteratively calculated at the same time. With multi‐block PLS, however, it is easy to overfit data resulting in a poor predictive ability. A recent development called cross‐fitting has been reported to alleviate the problem of overfitting in PLS. This approach was adjusted to multi‐block PLS and is tested on two different data sets, where overfitting and sensitivity to outliers are issues. Copyright

Visual tool for real-time monitoring of membrane fouling via Raman spectroscopy and process model based on principal component analysis

Membrane fouling, i.e. accumulation of unwanted material on the surface of the membrane is a significant problem in filtration processes since it commonly degrades membrane performance and increases operating costs. Therefore, the advantages of early stage monitoring and control of fouling are widely recognized. In this work, the potential of using Raman spectroscopy coupled to chemometrics in order to quantify degree of membrane fouling in real-time was investigated. The Raman data set collected from adsorption experiments with varying pHs and concentrations of model compound vanillin was used to develop a predictive model based on principal component analysis (PCA) for the quantification of the vanillin adsorbed on the membrane. The correspondence between the predicted concentrations based on the PCA model and actual measured concentrations of adsorbed vanillin was moderately good. The model developed was successful in monitoring both adsorption and desorption processes. Furthermore, the model was able to detect abnormally proceeding experiment based on differentiating PCA score and loading values. The results indicated that the presented approach of using Raman spectroscopy combined with a PCA model has potential for use in monitoring and control of fouling and cleaning in membrane processes.

Detection of current inefficiencies in copper electrowinning with multivariate data analysis

To further advance existing laboratory studies, the influence of different process parameters onto current efficiency was evaluated based on real industrial process history data obtained from conventional electrowinning circuit. Multivariate calibration model under partial least squares algorithm was applied to predict current efficiency in the process. The basic model was developed using values of electrolyte cupric and ferric concentrations, and total current applied. Pairwise interaction of parameters and moving average technique were applied to improve the prediction ability of the calibration. However, model construction based on the entire data set appeared to be unreliable due to high unexplained variance in the target variable, as sensor data were daily averaged. According to cluster analysis and further Monte-Carlo simulation, the phenomena of current inefficiency causing variation in the prediction of current efficiency appeared to be of random nature, i.e. daily averaging brought random variation to the multivariate model. For this reason, the data set was analyzed with multivariate process control charts to reveal the most important samples for predictive control. Multivariate calibration model was obtained using 58 samples, while the original data set contained 214 observations. Using the model, current efficiency values can be predicted on-line based on process sensor data. Multivariate process control tool was proposed in order to effectively monitor electrowinning process and detect current inefficiencies based on direct comparison of predicted and measured values of current efficiency.

Detection of abnormal process behavior in copper solvent extraction by Hotelling T 2 and squared prediction error control chart

Once a multivariate model is developed, it can be combined with tools and techniques from univariate statistical process control to form multivariate statistical process control tools. It allows development of advanced process monitoring strategies. In the current study, copper plant history data with multiple variables was successfully treated by principal component analysis to detect abnormal process behavior, particularly, in copper solvent extraction. The multivariate model was based on the concentration levels of main process metals recorded by the industrial on-stream x-ray fluorescence analyzer. Normal operating conditions were defined through control limits that were assigned to Hotelling T2 values on x-axis and to squared prediction error values on y-axis. Samples that were beyond the limits were classified as either systematic or random errors, or outliers. Model testing showed successful application of control limits to detect abnormal behavior of copper solvent extraction process as early warnings. Compared to the conventional univariate techniques of analyzing one variable at a time, the proposed model allows to detect on-line a process failure summarizing information from all process variables simultaneously. The proposed methodology was combined with on-line quality monitoring tool developed by VTT, Technical Research Center of Finland, to visualize the results. Thus, the proposed approach has a potential in on-line industrial instrumentation providing fast, robust and cheap application with automation abilities.