Data processing: measuring – calibrating – or testing – Measurement system in a specific environment – Chemical analysis
Reexamination Certificate
2002-10-16
2004-11-16
Hoff, Marc S. (Department: 2857)
Data processing: measuring, calibrating, or testing
Measurement system in a specific environment
Chemical analysis
C702S023000, C702S027000, C702S028000, C702S030000, C702S031000, C702S032000, C702S172000
Reexamination Certificate
active
06820013
ABSTRACT:
The invention relates to a method for the on-line analysis of liquid substance mixtures by means of near infrared spectroscopy (NIRS), and to an apparatus for carrying out this method and to the use of this apparatus in regulated substance distribution devices.
In chemical production, waste in liquid form is produced daily. This is collected in the form of large batches in areas which have to be classified as explosion-risk zones. Owing to recent legislation regarding the handling of waste (German Circular Economy and Waste Act), solvent mixtures can be subjected to recovery, material recycling, energy recovery or disposal, depending on their composition, with the higher-value process having priority over disposal. Mixtures of spent solvents can be re-used directly or their components can be recovered if their composition meets certain criteria. In general, mixtures of this type should not have more than three components and should at most be weakly coloured or have a low level of contamination with coarse particles. Precise limits cannot be set owing to the wide variety of mixtures produced and the user- or recycler-specific acceptance criteria. If solvent mixtures are to be subjected to material recycling or returned to production, a detailed analytical result of the composition is a vital prerequisite. Analysis of the composition is also essential as goods-in or declaration inspection on delivery of solvent mixtures. The analysis must provide information on the qualitative and also the quantitative composition of the mixtures, otherwise only energy recovery and disposal remain as options.
For efficient substance stream management, the analysis of liquid substance mixtures should therefore generally meet the following requirements:
The measurement must be sufficiently quick that it does not delay the time taken to combine the individual containers to form large batches
The measurement must be sufficiently quick that it does not delay the time taken to combine the individual containers to form large batches (i e. the measurement including evaluation should require less than 30 seconds).
The measurement must take place in the explosion-protection area.
Since the water content is a crucial parameter for the application or recycling method, quantitative determination of water must be possible.
Measurement in the gas phase at first appears appropriate for analysis of solvent mixtures since the interactions between mixture components are so weak there that they do not have to be taken into account in the measurement and in particular in the evaluation. Owing to the different vapour pressures of the components, the composition of a solvent mixture in the gas phase differs from that in the liquid phase. This is disadvantageous and requires subsequent correction calculations. A further disadvantage of measurement in the gas phase consists in the determination of the water content. In the gas phase, this can only be analysed under conditions which would require complex safety measures in the explosion-risk zone.
These disadvantages do not arise in the case of measurements in the liquid phase. Here, however, increased interactions between the components have to be taken into account.
One possibility for fast, qualitative or quantitative on-line analysis of solvent mixtures is near infrared spectroscopy (NIRS), which offers the following advantages.
Measurements are possible within seconds.
The use of quartz fibre optics which are transparent in the near infrared spectral region enables the spectrometer to be positioned outside the explosion-risk zone.
NIRS enables water to be determined quantitatively.
On use of NIRS for analytical problems, it is necessary firstly for spectra of “known” samples to be measured and filed in a reference database for calibration. Based on this database, spectra of “unknown” samples are then measured in routine operation, compared with the reference spectra in the database and accordingly evaluated qualitatively or quantitatively. In order to avoid incorrect results in this evaluation, the spectra of the samples that have been included in the reference database must cover at least the same variation latitude as the samples which are to be measured in routine operation and evaluated on the basis of the reference spectra.
Although NIRS has been used for on-line analysis in goods-in and production, it has not hitherto been employed for waste analysis since it has not been possible to use it for the simultaneous analysis of the type and proportion of the individual components in complex solvent mixtures of variable composition, without taking into account the declaration and without considerable calibration effort.
Mixtures of a plurality of solvents exhibit solvatochromic effects in the liquid phase, the extent of which are determined both by the type and the proportion of the individual components. These effects have been investigated for solutions of chromophoric substances as a function of various solvents. Data for mixtures of different solvents are rare. However, available data show that the solvatochromic interactions are very complex and can only be described or calculated systematically with difficulty. The creation of a database with reference spectra is therefore essential for calibration.
The test for the benefit of higher-value recycling methods as part of waste analysis comprises the recognition of, for example, at least 20 different solvents in mixtures, which may consist of 1, 2, 3 or 4 components. A database would have to include all possible combinations, i.e. count, and k is the number of components to be taken into account in a mixture.
In order to determine the quantitative composition and in order to take into account any solvatochromic effects, the percentage of components in the mixture, i.e. their quantitative graduation, must also be taken into account. This is calculated from
&AutoLeftMatch;
(
s
k
)
where s is the number of quantitative graduations of k components. In accordance with the basic procedure described above for building up a reference database, the number of solvent mixtures to be calibrated for analysis of up to 4 of 20 components with a 10% quantitative graduation works out in accordance with these two formulae as 449 730. This number relates merely to the possible combinations of the pure solvents which are to be expected in routine operation. Additional variations in the solvent mixtures to be taken into account, for example in the form of particulate or coloured impurities, would further increase this number significantly. The recording of such a number of reference spectra is clearly not sensible.
The use of factorial experiment plans for calculating the minimum numbers of calibration and reference samples required is known, but these cannot be used sensibly for systems which are defined by interactions and which are to be characterised qualitatively and quantitatively at the same time.
In addition, a further problem arises in the evaluation of the spectra. A known method is principal component analysis (PCA), which, although based on all the information in a spectrum, reduces this to a few principal components, known as factors, which represent the most important information for the analytical differentiation desired. The remainder of the information is collected in a so-called residue. This makes it possible to make “yes
o” statements, but not to make quantitative statements and to assess “outliers”. Quantitative evaluation by means of principal component regression (PCR) or partial least squares regression (PLSR) requires knowledge of the identity of the substances to be investigated.
The only method hitherto that uses all the spectral information without data reduction is direct spectral matching. The disadvantage of this spectra comparison method is the high demand for computer resources for repeated comparative calculation of all data points of the measured and reference spectra.
The object is thus to reduce the calibration effort and simplify the evaluation in order that NIRS can also be used for waste ana
Born Joachim
Frickel Hans
Itzel Hanshelmut
Hoff Marc S.
Merck Patent GmbH
Millen White Zelano & Branigan P.C.
Tsai Carol S.
LandOfFree
Method and device for the online analysis of solvent mixtures does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and device for the online analysis of solvent mixtures, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and device for the online analysis of solvent mixtures will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3359271