NIR spectrometry in pharmaceutical analysis.  Raman and NIR spectroscopy Problems of NIR spectrometry and how to solve them

ANIMAL AND VETERINARY

UDC 636.087.72:546.6.018.42 APPLICATION OF NIR SPECTROSCOPY TO DETERMINE THE AMOUNT OF INORGANIC AND ORGANIC COMPOUNDS IN FEED

S.I. Nikolaev, Doctor of Agricultural Sciences I.O. Kulago, Candidate of Chemical Sciences S.N. Rodionov, Candidate of Agricultural Sciences

Volgograd State Agrarian University

This work examines the possibilities of the express method of NIR spectroscopy for determining the amount of inorganic and organic compounds in feed. As a result of the research, the performance of the constructed calibrations was tested using a model mixture of “grain - bischofite” to quantify the mineral composition of biological samples. The results show that calibration data can be used to assess the mineral composition of feed mixtures.

Key words: NIR method, calibration model, bischofite.

The NIR method is based on measuring the reflection or transmission spectra of samples in the spectral range of manifestation of the component frequencies and overtones of the fundamental vibration frequencies of molecules of water, protein, fat, fiber, starch and other important components of the samples under study, followed by calculation of the value of the indicator using the calibration model built into the analyzer. The NIR spectral region covers the wavelength range 750-2500 nm (0.75-2.5 µm) or the wavenumber range 14000-4000 cm -1. Radiation in this spectral region has great penetrating power and at the same time is completely safe for biological objects. Thanks to this, it is possible to analyze whole grains of various crops without any damage to the sample. The main advantages of NIR analyzers are: rapid measurement, lack of sample preparation and reagents. The analysis process itself takes 2-3 minutes.

One of the new areas of application of the NIR method in the study of biological objects is the study of the composition of aqueous solutions.

It is known from the literature that salt solutions are directly inactive in the NIR region and signal registration is based on changes in hydrogen bonds between salts.

A typical example of measuring the “non-spectral properties” of a substance using near-infrared spectroscopy is the determination of the salt composition of seawater. In this regard, the concept of an IR shift agent becomes significant. Sodium chloride changes the structure of water by modifying hydrogen bonds, which is reflected in the spectra in the near-IR region.

In scientific developments in recent years, an important place has been devoted to the study of the effects of various macro- and microelements in mineral additives on the metabolic processes of the body of animals and poultry and the influence of these additives on the qualitative and quantitative indicators of manufactured products.

As indicated by Ba11oi'^ deficiency of feed in amino acids and energy

usually only leads to a decrease in weight gain and a deterioration in feed payments, while

how a deficiency in minerals and vitamins can cause various diseases and even the death of farm animals.

The main source of minerals for farm animals is plant feed (with some exceptions), which is introduced into the diet as mineral supplements (lick salt for animals, chalk, shells for poultry, etc.). The mineral composition of feed varies depending on their quality, plant growth conditions, the level of their agricultural technology and a number of other factors, including the so-called biogeochemical province.

Since animals receive elements of mineral nutrition with food and partly with water, in this work, studies were carried out on aqueous solutions of salts (sodium chloride and magnesium chloride) and some organic compounds (sugar, amino acid) using modern spectral methods with registration of signals in NIR ( near-IR) - areas.

To measure the concentrations of aqueous solutions of bischofite using the NIR method, a calibration model was built:

1) measurements were carried out at 4 points (cuvette positions);

2) each point was scanned twenty-four times;

3) measurements started with the lowest concentration of bischofite (1%);

4) each sample was measured three times, the first two times with the same filling of the cuvette, the third time the cuvette was filled anew;

5) the samples were selected in such a way as to characterize three concentration ranges.

As a result, a calibration model was obtained to determine the concentration of bischofite in water with a correlation coefficient of 0.99 (Figure 1).

SEC J SECV I SEV ] MD | Samples with poor chemical analysis | Accounts | Spectrum, loads | Chem. load | Total spectra: 99

Predicted value

;-N" rk- RP. u.

Reference value

Emission control Criteria: 12"00001

Exclude selected spectra

Cancel all changes

SEC R2sec indicator

Quantity 0.432567 0.999078

Spicy trend y = 0.0175+0.9991 x

Figure 1 - Calibration model of bischofite

Figure 1 shows a calibration model of bischofite built on the basis of bischofite solutions with concentrations from 1% to 10%, from 18% to 28%, from 32% to 42%.

Calibration model Quantitative

SEC SECV | SEV J MD | Samples with poor chemistry Total spectra: 48

analysis) Accounts | Spectrum, loads | Chem. i

Predicted value

I. . 0 5 . . , . . . . 1 . . . . , . 10 15 20

Reference value

Index:

|Quantity

Display data as: | Schedule

Emission control

Criterion: I 2-0000< *SECV Обновить |

Exclude selected spectra

Cancel all changes

SECV indicator R2secv F Trend line

Quantity 0.092000 0.999799 72877.753658 y = -0.0027+ 0.9996 X

Figure 2 - Calibration model of sodium chloride

In the same sequence, a calibration model for sodium chloride was built for comparative assessment. The correlation coefficient of the model was 0.99.

Figure 2 shows a calibration model of a sodium chloride solution with concentrations from 1% to 10%, from 18% to 20%.

To determine the concentration of sugar dissolved in distilled water, a calibration model was built in the above sequence. The correlation coefficient of the model was 0.99 (Figure 3).

Calibration model Quantity

BES 5ES\/ | BEU) MO | Samples with poor chemical AI Total spectra: 107

m | Accounts ] Spectrum, loads | Chem. load |

Predicted value 60-

Reference value

Quantity

Display data as: | Schedule

Emission control

Criteria: | 2-0000(“BESU Update |

Exclude selected spectra

Cancel all changes

BESU indicator (ggees/ P Direct trend

Quantity 0.218130 0.999851 230092.131072 y =0.0114 + 0.9996 x

Figure 3 - Calibration model of sugar

Figure 3 shows a calibration model of a sugar solution with concentrations from 1% to 10%, from 18% to 28%, from 40% to 45%.

Calibration model Qualitative

Figure 4 - Distribution of calibration models: 1) P-alanine, 2) sugar,

3) bischofite, 4) sodium chloride in a single coordinate system To evaluate the obtained models in the coordinates of two principal components, a qualitative comparison of the distribution points of the calibration models was carried out: 1) P-alanine, 2) sugar, 3) bischofite, 4) sodium chloride.

Using these calibrations, the following studies were carried out. Solutions of bischofite were prepared with a mass fraction of dissolved substance of 2%, 4%, 10%, with which grain (wheat, barley, oats) was moistened. When measuring the concentration of a bischofite solution using the NIR method, which was used to wet grain (wheat, barley, oats), the following data were obtained (Table 1).

Table 1 - Concentration of bischofite

Concentration of bischofite solution before wetting the grain (wheat, barley, oats) Concentration of bischofite solution after wetting the grain (wheat, barley, oats)

wheat barley oats

10 % 10,1 10,2 10,3

When wetting grain (wheat, barley, oats) with a bischofite solution with different concentrations (2%, 4%, 10%), the color of the bischofite solution changed.

In each case, the bischofite solution with which the grain was wetted was colored, possibly by organic matter (pigments) of the grain, and visually the solution had a more saturated color at a bischofite concentration of 2%; with an increase in the concentration of the bischofite solution, the color intensity of the solution with which the grain was wetted decreased.

From the analysis of the results in Table 1, it can be seen that the concentration of the bischofite solution (2%, 4%, 10%) with which the grain (wheat, barley, oats) was moistened practically did not change. The grain absorbed some volume of liquid. After this, the unused solution was drained and its volume was measured. It can be assumed that the amount of salt remaining on the grain (wheat, barley, oats) was dissolved in the consumed volume of bischofite.

Calculations have shown that when a wheat grain weighing 1000 g is wetted with a bischofite solution with concentrations (2%, 4%, 10%) the amount of magnesium and chlorine indicated in Table 2 should remain on the grain (wheat, barley, oats).

Table 2 - Calculated content of magnesium cations and chlorine anions on grain _______ (wheat, barley, oats), after treatment with bischofite solution_______

Amount of magnesium g remaining on a grain weighing 1000 g when wetted with bischofite Amount of chlorine g remaining on a grain weighing 1000 g when wetted with bischofite

2 % 4 % 10 % 2 % 4 % 10 %

Wheat grain 2.4 5.0 11.2 7.1 14.8 33.2

Barley grain 2.0 4.2 10.6 6.1 12.6 31.6

Oat grain 4.8 9.8 21.2 14.2 29.2 62.8

To determine the amount of magnesium cations and chlorine anions of grain (wheat, barley, oats) treated with bischofite solution (2%, 4%, 10%), the capillary electrophoresis (CEP) method was used. The studies were carried out on a Kapel 105 analyzer, using the method for determining cations in feed M 04-65-2010 developed (LUMEX LLC), the method for determining anions in feed M 04-73-2011 developer (LUMEX LLC). We studied grain (wheat, barley, oats) moistened with a bischofite solution (2%, 4%, 10%). The research results are displayed in Table 3.

Table 3 - Content of cations and anions in grain (wheat, barley, oats).

Amount of magnesium, g Amount of chlorine, g

in 1000 g of grain in 1000 g of grain

Without bischofite Bischofite 2% o4 4 t and & o sh and B Bischofite 10% Without bischofite o4 2 t and & o sh and B o4 4 t and & o sh and B Bischofite 10%

Wheat grain 2.8 4.5 6.7 11.4 3.3 8.5 12.G 22.7

Barley grain 2.4 3.9 5.6 16.G 4.5 5.6 1G.4 26.G

Oat grain 2.3 6.2 11.6 36.G 4.1 1G.G 26.G 44.G

1. Traditionally, it is customary in assessing the quality of water and feed to consider the presence of the amount of a particular mineral in water and feed; in this case, we came into contact with the quality of the mineral’s influence on the physical and chemical properties of water and possibly on the feed mixture.

2. A comparison of two calibration models (solutions of sodium chloride and magnesium chloride) showed that the calibration model of sodium chloride is based on the spectral range from 10400 to 10900 cm-1, and for bischofite (magnesium chloride) from 10100 to 10600 cm-1. It is known from the literature that salt solutions are directly inactive in the NIR region and signal registration is based on changes in hydrogen bonds between salts.

Therefore, the effect of sodium chloride on hydrogen bonds in a salt-water system is different from the effect of magnesium chloride on hydrogen bonds in the same system.

3. In a single coordinate system, organic and inorganic components were distributed in a certain sequence, without mixing.

4. The calculated amount of magnesium that should have remained on the grain (wheat, barley, oats) almost completely coincides with the actual amount of magnesium determined using the Capel-105 capillary electrophoresis system.

The amount of chlorine is significantly less than calculated.

5. Analysis of Table 3 shows that the data obtained using calibrations of the NIR method are confirmed by CEF studies.

6. As a result of the research, the performance of the constructed calibrations was tested using a model mixture of “grain - bischofite” to quantitatively assess the mineral composition of biological samples. The results show that calibration data can be used to assess the mineral composition of feed mixtures.

Bibliography

1. Georgievsky, V.I. The influence of magnesium level in the diet on the growth and development of broiler chickens [Text] / V.I. Georgievsky, A.K. Osmanyan, I. Tsitskiev // Chemistry in agriculture. - 1973. - No. 10. - P. 68-71.

2. Sheptun, V.L. Introduction to the method of spectroscopy in the near-infrared region [Text]: methodological manual / V.L. Whisperer. - Kyiv: Center for Infrared Spectroscopy Methods LLC "Analit-Standard", 2005. - 85 p.

3. Schmidt, V. Optical spectroscopy for chemists and biologists [Text] /V. Schmidt. -M.: Tekhnosphere, 2007. - 368 p.

WHAT IS NEAR IR?

The near infrared (NIR) region of the electromagnetic spectrum extends from 800 nm to 2500 nm (12500 to 4000 cm-1 ) and lies between the mid-IR region with longer wavelengths and the visible region with shorter wavelengths. The mid and near ranges can be used for vibrational spectroscopy. While mid-IR spectra record mainly atomic vibrations in the individual chemical bonds of most molecules, the corresponding NIR spectra show so-called overtones and Raman bands.

On the wave number scale (see-1 ) these overtones appear as something less than the constituent frequencies of the fundamental vibrations. For example, the main vibration of the C-H bond (n) of the trichloromethane molecule (CHCl3) occurs at 3040 cm-1 , the first three overtones (2n, 3n and 4n) are observed at 5907 cm-1, 8666cm -1 and 11338cm -1 respectively.

At the same time, the absorption capacity decreases with increasing overtone number, for example, a series of these values ​​for CHCl3 is 25000, 1620, 48,

1.7 cm-1 /mol respectively.

Due to the sharp decrease in intensity of higher overtones, NIR spectra are typically suppressed by overlapping overtones and Raman bands of structurally lighter groups (e.g., C-H, N-H, and O-H). Within these NIR spectra there is significant information about the molecular structure of the sample under study, and this information can be extracted by modern data processing methods.

Benefits of NIR spectroscopy

Speed ​​(usually 5 – 10s)

No sample preparation required

Easy to measure

High accuracy and reproducibility of analysis

No pollution

Process control

Possibility of taking measurements via glass and plastic packaging

Automation of measurements

Transferring a method from one device to another

Analysis of physical and chemical properties

Compared to liquid-based chemical analysis methods, NIR spectroscopy analysis is faster, simpler and more accurate. Measurements can be carried out very quickly, usually the analysis time is only 5-10 seconds. No preliminary sample preparation or special training of personnel is required. These spectra may contain information about the physical properties of the material, such as particle size, thermal and mechanical pretreatment, viscosity, density, etc.

COMPARISON OF IR SPECTROSCOPY

near and mid range

Reducing sample preparation time is one of the main advantages of near-IR compared to mid-IR. This is primarily due to the relatively low absorption coefficient of most materials in the NIR range. Mid-range measurements of powdered samples are traditionally performed either by diffuse reflectance or by compressing samples into tablets and measuring spectra in transmission mode. In both cases, the samples must first be ground into a fine powder and then mixed with a non-absorbent substance such as KBr. The powders, crushed and mixed with KBr, are placed in a mold and pressed into tablets at high pressure using a hydraulic or manual press. For diffuse reflection measurements, the crushed sample mixed with KBr is placed directly into the sample cup, the surface of the sample is leveled, and then introduced into the diffuse reflection attachment for measurements. These sample preparation methods are widely and successfully used, but have disadvantages such as longer sample preparation times, higher potential for sample contamination, possible decreased sample-to-sample and user-to-user reproducibility due to variations encountered during sample preparation, and additional cost of KBr diluent.

In addition, the advantage of NIR spectroscopy is that it uses fairly inexpensive optical fiber to measure solid and liquid samples. Comparable mid-IR accessories are either limited by their physical reach or by being fragile and difficult to handle. All this makes NIR spectroscopy much more attractive for use in the production process.

COMPARISON OF BIR spectroscopy

and dispersing devices

Fourier spectrometers in the near-IR range differ significantly from dispersive spectrometers in the near-IR range in the method of obtaining the spectrum. Dispersive devices use a narrow slit and a dispersing element, such as a grating, to convert light into a spectrum. This spectrum is projected onto a sensor or array of sensors, where the intensity of light at each wavelength is determined. The spectral resolution of dispersive devices is determined by a fixed slit width, usually 6-10 nm (from 15 cm-1 to 25cm -1 , at 2000nm). Resolution cannot be selected in software, and increasing resolution requires a narrower slit and attenuation of the resulting signal. Thus, for all dispersive devices there is a problem of choosing between resolution and signal-to-noise ratio.

A Fourier transform spectrometer, in contrast, uses an interferometer to scan combinations of wavelengths of light emerging from a broad swath of a near-infrared source and sends these combinations to a single detector.

In each interferometer scan, data is collected in the form of an interferogram, in which the signal intensity is correlated with the displacement of the moving part of the interferometer. This interferometer offset is directly related to wavelength, and a mathematical transformation (Fourier transform) is applied to plot the signal intensity as a function of wavelength, from which a measure of spectral absorption or spectral transmittance is calculated.

At the same time, a HeNe laser beam passes through the interferometer and is directed to its own detector. The displacement of the interferometer results in signal maxima and minima at this laser detector, which occur at precisely defined intervals that are multiples of the laser wavelength. Where this signal passes through zero is used as collection points for digitization of the NIR detector signal. Thus, due to the control of digital conversion, the FTIR spectrometer has significantly higher wavelength accuracy than any other dispersive instrument. This length accuracy has a direct impact on the stability conditions of calibration models developed on Fourier systems, as well as on the ability to transfer the calibration model to other Fourier instruments, which will be described below.

The spectral resolution for Fourier spectrometers is determined by the degree of mobility of the interferometer, which is controlled by software, which makes it possible to significantly increase the resolution compared to a dispersive spectrometer, and, with the help of software, to select the resolution during research. In addition, the FTIR's broadband near-infrared beam is directed through large circular apertures instead of the narrow rectangular slit used in the dispersive document, illuminating a larger area of ​​the sample and increasing the light intensity at the detector. This performance advantage results in a higher signal-to-noise ratio for FTIR spectrometers compared to dispersive instruments. A better signal-to-noise ratio leads to a significant reduction in detection time and, as a consequence, to obtaining spectra of higher quality on the Fourier instrument at any spectral resolution.

FOURIER - NEAR IR SPECTROSCOPY for qualitative and quantitative analysis

Today, many manufacturers strive not only to deliver the highest quality end product, but also to improve production efficiency through laboratory analysis and use of the resulting results in production. By gaining tighter control over technology, it is possible to optimize the use of substances by adding or eliminating them to produce specified products, thereby minimizing distribution or processing costs.

NIR is a spectroscopic technique ideal for measurement processing due to its ability to quickly perform remote measurements through high-performance quartz optical fiber. The signal attenuation within such fibers is very low (eg, 0.1 dB/km), and NIR fiber optic cables and sensors are rugged, relatively inexpensive, and widely available. Processing sensors can be located hundreds of meters from the spectrometer, and multiple sensors can be connected to a single spectrometer.

NIR MEASUREMENT METHODS

NIR sampling methods for solids are based on either diffuse reflectance or simple transmission measurements. Diffuse reflectance measurements are mainly made using a fiber optic sensor or an integrating sphere.

In Fig. Figure 2 shows a Fourier transform NIR spectrometer MPA (manufactured by Bruker Optik GmbH, Germany), which has 2 fiber optic sensor ports and a separate sample compartment, which allows the use of the direct transmission method.

This photo shows a common reflectance sensor used to analyze powder samples in test tubes.

Samples are analyzed by contacting the sensor with the sample material. The completion of the analysis is indicated by illuminated LEDs.

The integrating sphere (Fig. 3) allows you to collect spectral data from inhomogeneous substances, for example, mixed powders, grains, polymer granules, etc. The resulting spectra represent the spatial averaging of all material located in a circular measurement window (diameter 25 mm).

For better averaging, a rotating cup and automatic samplers can be used.

BIC REVOLUTION

IN PHARMACEUTICAL

INDUSTRY.

QUALITY ASSESSMENT ISSUES

The pharmaceutical industry is known as one of the most heavily regulated industries in the world, and Bruker manufactures quality testing instruments for pharmaceutical consumers that enable consumers to check whether their drugs meet required requirements. The OPUS software package controls all functions of the spectrometer. This software package includes a comprehensive test of the software and hardware suite. OPUS will fully check the correct functioning by pressing a key. This includes testing the internal test device built into the spectrometer.

The software can be run in password-protected "GLP" mode, with full administrator control over the user's access to menus, settings and custom macro programs. The data block provides complete and automatic control of all actions performed with the spectra. An icon-based programming language is built into the software to automate complex procedures. As a result, repeatability increases and potential errors are reduced.

Bruker is an ISO9000 company and all software and hardware are subject to strict quality control, multiple stages of final testing and verification before delivery to the customer. Installation of the device at the customer's site is carried out by our experienced technical engineers, who provide the customer with a working device upon delivery and then continuously throughout the life of the device.

RAW MATERIAL IDENTIFICATION

One of the first steps in the production of any pharmaceutical product is to identify and verify that the various incoming raw materials meet the necessary requirements. NIR spectroscopy through fiber optic sensors is quickly becoming the standard method for performing this compliance check, providing unprecedented speed in the identification of both solids and liquids.

To perform this type of analysis, a calibration model must be created that involves the substances of interest. First, it is necessary to obtain several spectra for each raw material, taking into account any possible variations that may occur. This usually includes types of raw materials obtained from different sellers, from different places, etc. Once the spectra are measured, an average spectrum of each material is generated, and a library of all such average spectra is created, along with statistically determined acceptable criteria (or thresholds) for all substances in the library.

The library then confirms that all materials are uniquely identified. The library can now be used to identify new unknown substances by comparing their spectra with those of the library, and determining the quality of the hit for each substance in the library. If this hit quality is less than the threshold for one substance and greater than the threshold for all other substances, the unknown substance is identified.

The liquids to be identified can be measured either by transmission measurement in the sample compartment (as shown in Figure 1) or by using a fiber immersion probe. In any case, the lower absorption coefficients of NIR (compared to mid-IR) allow the use of much longer sample path lengths (i.e. 1 - 10mm). Due to this difference in path length, measurements in the sample compartment become more advantageous, since it allows the use of standard inexpensive glass tubes instead of precision cells, reducing the cost and duration of measurements.

QUANTITATIVE ANALYSIS OF ACTIVE INGREDIENTS

Another important part of qualitative/quantitative analysis in the pharmaceutical industry is the quantitative analysis of concentrated active ingredients. This type of analysis often requires extensive laboratory testing of test prints of samples that are destroyed during testing. In contrast, FTIR provides a time-saving and non-destructive way to perform quantitative analysis of concentrates in mixtures of powder or liquid substances, as well as in already manufactured pharmaceutical tablets and capsules.

EFFECTIVE SAMPLING

A key factor in the success of FTIR for quantitative analysis is the choice of sampling method, often a combination of automated and manual sampling. Bruker manufactures sampling accessories specifically for the pharmaceutical industry. For example, an automatic sampler (Fig. 5) can be installed in the sample compartment of any Bruker FTIR spectrometer.

This accessory features a customizable sample disk that can hold up to 30 samples. The user processes the tablet slots and the movement of the disk by OPUS software or a user-definable macro command and/or communication with a centralized distributed control system within the manufacturing plant.

EXAMPLES OF ACTIVE INGREDIENT ANALYSIS

An example of the quantitative analysis of an active ingredient concentrate in a finished pharmaceutical product by Fourier Transform Infrared (NIR) is the determination of the concentration of acetylsalicylic acid (ASA) in aspirin tablets. To conduct this analysis, the least squares method (OLS) was used to process the spectra obtained from aspirin tablets with a known concentration of ASA. The concentration of ASA in the samples ranged from 85% to 90%. In addition to ASA, the tablets contained two types of starch in the range of 0%-10%.

To install the OLS model for this multicomponent system, with a resolution of only 8 cm-1 44 spectra were recorded. The optimal range for ASA was determined using the OPUS-Quant/2 software package (mutual validation). The root mean square error was 0.35%, and the discrepancy R 2 - 93.8%. This error was within the limits specified by the customer. A plot of true and calculated concentrations is shown in Figure 6.

SAMPLING THROUGH PACKAGING

In addition, the determination of the concentration of the active ingredient of aspirin tablets through the plastic materials of the clear packaging using a fiber optic diffuse reflectance sensor was demonstrated, as shown in Figure 7. The resulting spectra showed convex ranges from the polymer material of the clear packaging, but two distinct regions (6070-5900 cm-1 and 4730-4580cm -1 ) containing peaks from aspirin are still visible and were used to create the calibration model.

A graph of true and found concentrations is shown in Figure 8). The root mean square error was 0.46%, and the discrepancy R 2 - 91.30%, these values ​​are again within the limits specified by the customer. The spectra obtained in this example are shown in Figure 9.

BENEFITS OF INCREASING RESOLUTION

IN SPECTRAL ANALYSIS

Until recently, most published results in NIR spectroscopy were obtained using low-resolution dispersive instruments, with spectral resolution ranging between 6 and 10 nm (from 15 cm-1 to 25 cm -1 , at 2000 nm). The advent of FT-NIR spectrometers has led to significant advances in high-resolution capabilities (better than 2 cm-1 ) NIR spectroscopy.

NIR spectra are usually characterized by high spectral absorption, which does not require high resolution. At that time, there are often situations where the desired calibration model from low-resolution spectra cannot be created. In addition, high resolution directly affects the wavelength accuracy of the instrument and, consequently, the stability of results and the “transportability” of calibration models.

Experimentally, to demonstrate the value of increasing resolution in spectral analysis, the NIR spectra of 5 tablets with various low concentrations of the active ingredient were measured. The spectra were measured at a resolution of 8 cm-1 and 2 cm -1 , after which an identification model for the tablets was created using OPUS. With a resolution of 2 cm-1 , the model could only distinguish between placebo and tablets with active ingredients, while at a higher resolution of 8 cm-1 , all concentrations are clearly distinguishable.

Figure 10a shows the spectra and plot obtained for the first two principal components of the measurements at 8 cm-1 . Figure 10b shows the spectra and plot obtained for the first two principal components of measurements at 2 cm-1 . The 5 areas in the last graph indicate that the higher resolution model can clearly distinguish the 5 concentration levels of the active ingredient.

DETERMINING THE THICKNESS OF THE COVERING LAYER

FTIR spectroscopy has also been successfully used to determine layer thickness on pharmaceutical tablets. Several tests were performed in this study, including experiments with nonlinear relationships between the light absorption measure and layer thickness, similarity of core and coating material composition, and the lack of sufficient calibration samples for standard LSM calibration. Peak at 7184 cm-1 , which differentiates the core material from the coating material, was identified when high resolution NIR spectra were collected (2 cm-1, 0.4 nm at 7184 cm-1 ) on a Fourier-NIR spectrometer IFS-28/N from Bruker (see Figure 11).

Research shows that layer thickness can be modeled as a polynomial fit to the peak region of that sample peak (see Fig. 12), while least squares calibration of the same data is not possible due to the lack of sufficient calibration samples. Also, this calibration has been successfully used for a number of pellets, but is unacceptable for fiber optic diffuse reflectance measurements due to insufficient penetration of the fiber into the core.

TRANSFER CALIBRATION

Developing a stable and reliable calibration model is a very labor-intensive, resource-intensive task that involves preparing and analyzing a large number of samples using a standard method, and then analyzing them using the Fourier-NIR method. Thus, it is important that a calibration model be developed that can be used over time, and for which it does not matter what kind of instrument, type of sources, detectors, sensors, etc. are used.

In addition, several factors influence the transfer of calibration from one instrument to another. This includes, for example, the wavelength and photometric accuracy of various instruments. Therefore, for all calibration models transferred from one instrument to another, it is necessary to re-measure at least the original set of calibrations (or the complete set of calibrations) on the new instrument to determine the correction factors that will allow the model to work on the new instrument.

Sometimes this leads to difficulties in transferring the calibration model, and sometimes, in the case of rare or changing calibration samples, such transfer is not possible at all.

Typically, the difficulty in transferring the calibration model is the wavelength accuracy of these two instruments. The lack of a stable wavelength axis is a factor that greatly limits the ability to transfer the calibration model among dispersive instruments. Therefore, Brooker's high-resolution instrumentation FT-NIR spectrometer production line has the great advantage of using the wavelength axis as a calibration method.

To do this, a narrow region in the spectrum of atmospheric water vapor with a known constant wavelength is considered, which is used as a wavelength standard. This allows FT-NIR spectrometers (manufactured by Bruker Optik GmbH, Germany) to provide much higher wavelength accuracy than any dispersive instrument. As a result, direct transfer of calibration from one Fourier-NIR instrument to another is possible. The benefit of this feature, which avoids costly recalibration while saving time, money and effort, cannot be underestimated.

One such example of transfer of a calibration model for quantifying the alcohol content of spirits is shown in Table 1. The calibration was performed on an IFS-28/N Brooker spectrometer with immersion probe A, and was subsequently transferred to a Vector 22/N Brooker spectrometer with immersion probe B. After transmission, comparison R 2 and standard deviation errors showed the success of direct calibration transfer. Additional tests have shown the success of direct transfer of other calibration models from instrument to instrument, as well as direct transfer of models on one instrument, after replacing all major system components, including the NIR source, HeNe laser, detector, sensors and electronics.

CONFORMITY TEST

It is often necessary to determine whether the final product meets a certain standard. This is easy to do on Bruker spectrometers, using Compliance Test . A series of spectra are measured for a few selected samples of each substance and will be checked against the spectra determined independently by a standard method. For each substance, along with the standard deviation spectrum, an average spectrum is generated. New samples of the substance are then analyzed, their spectra compared with the stored average spectrum, and an assessment is made of whether the new spectrum is within the acceptable limits defined by the standard deviation spectrum and the customer-adjustable factor. A typical compliance test report is shown in Figure 13.

MIXTURE ANALYSIS

In many pharmaceutical processes, analysis of the mixing process of two or more components is often necessary. Mixture analysis plays an important role when mixing powders, where samples are characterized by heterogeneity. The optimal ratio in the mixture determines the final product. The mixing process must be verified in real time using FTIR spectroscopy. Spectra are taken from the correct reference mixtures, and then the average spectrum and the standard deviation spectrum are calculated. After this, the spectra are taken during mixing, processed and compared with the average spectrum. The mixing process is stopped if the resulting spectrum falls below a user-defined threshold for the average spectrum of the desired mixture.

CONCLUSION

FT-NIR spectroscopy is a fast, easy-to-use and reliable tool for quality assurance and quality control in the pharmaceutical industry. The advanced performance of Fourier transform technology enables more complex studies and allows calibration to be transmitted directly. In addition, methods such as raw material identification and quality testing, determination of active ingredient concentration, conformity testing of final products, and mixture analysis of products are common among consumers in the pharmaceutical industry.

As a manuscript

DOLBNEV DMITRY VLADIMIROVICH

IDENTIFICATION OF MEDICINES BY NEAR INFRARED SPECTROSCOPY

04/14/02 – pharmaceutical chemistry, pharmacognosy

dissertations for an academic degree

candidate of pharmaceutical sciences

Moscow – 2010

The work was carried out at the State Educational Institution of Higher Professional Education First Moscow State Medical University named after

Scientific supervisors:

Doctor of Pharmaceutical Sciences, Academician of the Russian Academy of Medical Sciences, Professor

Doctor of Pharmaceutical Sciences, Professor

Official opponents:

Lead organization:

All-Russian Scientific Center for the Safety of Biologically Active Substances (VSC BAV)

The defense will take place “___”____________________2010 at ____ o’clock at a meeting of the Dissertation Council (D 208.040.09) at the First Moscow State Medical University named after Moscow, Nikitsky Boulevard, 13.

The dissertation can be found in the library of Moscow State Medical University named after. Moscow, Nakhimovsky prospect, 49.

Scientific secretary of the dissertation

council D 208.040.09

Doctor of Pharmaceutical Sciences,

Professor

Relevance of the research topic. Over the past 15 years, near-infrared (NIR) spectroscopy has been rapidly developing and has found application in a wide variety of industries. NIR spectroscopy is known as an effective method for qualitative and quantitative analysis. This method is widely used in agriculture (to determine the quality of soils, the content of protein, fat, etc. in food products), in industry (to determine the composition of petroleum products, the quality of textile products, etc.), in medicine (to determine fat, oxygen in the blood, studies of tumor development). Currently, NIR spectroscopy is becoming one of the in-process control methods in the pharmaceutical industry in Europe and the USA.

It is used to check input raw materials, mixing uniformity, determination of granulation end point, drying moisture content, tableting uniformity, coating thickness measurement.

The NIR spectroscopy method is described in the European Pharmacopoeia and the US Pharmacopoeia, but it is still used relatively rarely in pharmacopoeial analysis: mainly when determining the water content in preparations obtained from blood.

In this regard, the development of unified methods for the analysis of pharmaceutical substances and drugs for their further use in pharmacopoeial analysis is of great importance.

This issue is of particular importance in connection with the publication of the 12th edition of the State Pharmacopoeia of the Russian Federation.

It is also necessary to note the ongoing problem of counterfeit medicines, one of the ways to solve which is the development of rapid analysis methods.

Considering the above, an urgent problem is the development of unified methods for analyzing substances and preparations and identifying counterfeit medicines using the NIR spectroscopy method.

Purpose and objectives of the study. The purpose of the study was to develop unified methods for analyzing substances and preparations and identifying counterfeit medicines using the NIR spectroscopy method.

To achieve this goal, the following tasks were solved:

– to study the possibility of obtaining NIR spectra of substances, tablets and capsules using a fiber optic sensor and an integrating sphere;

– compare the NIR spectra of substances and drugs;

– compare the NIR spectra of drugs with different contents of the active substance;

– study the possibility of using NIR spectroscopy to establish the authenticity of substances and preparations from specific manufacturers, as well as to identify counterfeit medicines;

– develop an electronic library of NIR spectra of substances and drugs.

Scientific novelty of the research results. For the first time, it has been shown that the NIR spectroscopy method can be used both to determine the authenticity of pharmaceutical substances and finished medicinal products (tablets and capsules). It has been shown that, in general, the NIR spectra of substances and drugs differ. Spectra can be obtained using a fiber optic sensor and an integrating sphere. It has been shown that if the capsule shell or tablet packaging (blister) is transparent, a spectrum can be obtained without removing the capsules or removing the tablets from the packaging. It has been shown that the NIR spectroscopy method can be used to identify counterfeit drugs, provided that the spectra of the original and test drugs are compared. Spectra of substances and drugs can be stored as an electronic library. It has been established that for a more reliable comparison of the spectrum of the test drug and the standard spectrum, the use of mathematical data processing is required.

Practical significance of the work. Developed methods for analyzing drugs using NIR spectroscopy are proposed to establish the authenticity of pharmaceutical substances, drugs in the form of tablets and capsules. The techniques allow the use of an integrating sphere and a fiber optic sensor (“gun”).

The developed methods can also be used for express identification of counterfeit medicines and for incoming and outgoing control of pharmaceutical substances and intermediates at pharmaceutical enterprises. The methods allow in some cases to carry out non-destructive quality control without opening the primary packaging.

The developed library of NIR spectra can be used to identify substances, tablets and capsules using a fiber optic sensor (“gun”) and an integrating sphere.

The results of the work have been tested and used in the quality control department.

Approbation of work. The main provisions of the dissertation work were reported and discussed at the XII Russian National Congress “Man and Medicine” (Moscow, 2005), the International Congress on Analytical Chemistry ICAS (Moscow, 2006) and the XIV Russian National Congress “Man and Medicine” (Moscow , 2007). The work was tested at a scientific and practical meeting of the Department of Pharmaceutical Chemistry with the course of toxicological chemistry of the Faculty of Pharmaceutical Sciences of Moscow State Medical University. March 22, 2010

Publications. 5 printed works have been published on the topic of the dissertation.

Linking research to the problem design of pharmaceutical sciences. The dissertation work was carried out within the framework of a complex topic of the Department of Pharmaceutical Chemistry of Moscow State Medical University named after. “Improving quality control of medicines (pharmaceutical and environmental aspects)” (state registration No. 01.200.110.54.5).

Structure and scope of the dissertation. The dissertation is presented on 110 pages of typewritten text, consists of an introduction, a literature review, 5 chapters of experimental studies, general conclusions, a list of references, and also separately includes 1 appendix. The dissertation work is illustrated with 3 tables and 54 figures. The list of references includes 153 sources, of which 42 are foreign.

Provisions for defense:

– results of studying the possibility of obtaining NIR spectra of substances, tablets and capsules using a fiber optic sensor and an integrating sphere;

– results of a comparative study of NIR spectra of substances and drugs, as well as NIR spectra of drugs with different contents of the active substance;

– results of studying the possibility of using NIR spectroscopy to establish the authenticity of substances and preparations from specific manufacturers, as well as to identify counterfeit medicines.

1. Objects of study

Substances and preparations of a number of drugs have been studied. A total of 35 substances were used in the study: aluminum hydroxide, amikacin sulfate, ascorbic acid, sodium ascorbate, warfarin sodium, vitamin B12, gemfibrozil, magnesium hydroxide, glurenorm, D-biotin, iron gluconate, zopiclone, calcium D panthenoate, clindamycin phosphate, lidocaine hydrochloride , metoprolol tartrate, nicotinamide, paracetamol, pyridoxine hydrochloride, piperacillin, ranitidine hydrochloride, riboflavin, thiamine mononitrate, tyrothricin, famotidine, folic acid, cefadroxil, cefazolin sodium salt, ceftizoxime sodium salt, ciprofloxacin hydrochloride, cyanocoblamine, various manufacturers and 59 drugs from various manufacturers containing: isoniazid, meloxicam, omeprazole, ranitidine hydrochloride, rifampicin, famotidine, ciprofloxacin, esomeprazole, ethambutol, as well as 2 falsified samples (OMEZ 20 mg, Dr. Reddy`s Lab. and Rifampicin 150 mg,).

2. Equipment and test conditions

An MPA device was used in the work - a near-infrared Fourier spectrometer (Bruker Optics GmbH, Germany). Recording parameters: spectral range from 800 nm to 2500 nm (cm-1 to 4000 cm-1), number of scans 16, spectral resolution 4 cm-1. The instrument was controlled and the obtained spectra were processed using the OPUS 6.0 software package (Bruker Optics GmbH, Germany). NIR spectra were obtained in two ways:

1) using a fiber optic sensor (“gun”),

2)

Both methods were used to obtain NIR spectra of substances, tablets and capsules.

The fiber optic sensor (“gun”) allows for reflection measurements only, while the integrating sphere allows both reflection and transmission measurements. In this work, NIR reflectance spectra were obtained.

2.1. Methods for obtaining NIR spectra:

using a fiber optic sensor (“gun”).

2.1.1. Substances . The powder substance was poured into a transparent cuvette with a layer thickness of 1 to 3 cm. Then the fiber optic sensor was pressed perpendicular to the surface of the powder. The spectrum registration procedure was started by pressing a button on the fiber optic sensor. The measurement of the spectra was repeated 3–5 times from different areas to obtain statistically reliable analysis results.

2.1.2. Tablets removed from the blister . The fiber optic sensor was pressed perpendicular to the tablet. The spectrum registration procedure was started by pressing a button on the fiber optic sensor. The measurement of the spectra was repeated 3–5 times from different areas of the tablet to obtain statistically reliable analysis results.

2.1.3. Tablets in blister . If the blister is transparent, the measurement was carried out as follows, the fiber optic sensor was pressed perpendicular to the surface of the tablet in the blister. The spectrum registration procedure was started by pressing a button on the fiber optic sensor. The measurement of the spectra was repeated 3–5 times from different areas of the tablet in the blister to obtain statistically reliable analysis results. If the blister was opaque or aluminum, the tablet was first removed from the blister and then the NIR spectrum was obtained.

2.1.4. Capsules . If the capsule shell is transparent, then the measurement was carried out as follows: the fiber-optic sensor was pressed perpendicular to the surface of the capsule in the blister. The spectrum registration procedure was started by pressing a button on the fiber optic sensor. The measurement of the spectra was repeated 3 - 5 times from different parts of the capsule in the blister to obtain statistically reliable analysis results. If the capsule shell was not transparent, then the capsule was first opened and then the spectrum of the contents was measured in a glass cuvette.

2.2. Methods for obtaining NIR spectra:

using an integrating sphere.

Obtaining NIR spectra in reflection mode

2.2.1. Substances . The powder substance was poured into a transparent cuvette with a layer thickness of 1 to 3 cm. Then the cuvette was placed on top of the optical window of the integrating sphere. The measurement process was started on a computer using the OPUS program or directly on the device itself (the “Start” button). The measurement of the spectra was repeated 3–5 times to obtain statistically reliable analysis results.

2.2.2. Tablets removed from the blister . The tablet was placed in a special holder. A holder with a tablet was installed on top of the optical window of the integrating sphere. The measurement process was started on a computer using the OPUS program or directly on the device itself (the “Start” button). The measurement of the spectra was repeated 3–5 times from different areas of the tablet to obtain statistically reliable analysis results.

2.2.3. Capsules . If the capsule shell is transparent, then the measurement was carried out as follows: the capsule was placed in a special holder. A holder with a capsule was installed on top of the optical window of the integrating sphere. The measurement process was started on a computer using the OPUS program or directly on the device itself (the “Start” button). The measurement of the spectra was repeated 3–5 times from different parts of the capsule to obtain statistically reliable analysis results. If the capsule shell was not transparent, then the capsule was first opened, and then the spectrum of the contents in a glass cell was measured by placing the cell on top of the optical window of the integrating sphere.

3. Mathematical processing of NIR spectra.

Mathematical processing of the obtained spectra was carried out using the OPUS IDENT program, included in the OPUS 6.0 software package (Bruker Optics GmbH, Germany). The unknown spectrum was compared with the reference library spectrum by calculating the spectral distance. IDENT identifies those comparison spectra that are closest to the analyzed spectrum and determines the deviations between these spectra and the analyzed spectrum. This allows IDENT to identify unknown substances and assess the degree to which the substance meets the reference standard.

We used two methods of mathematical processing of NIR spectra: 1) Ident analysis, which correlates the spectrum and a specific substance, and 2) cluster analysis, which correlates the spectrum and a group of substances.

Once the spectra are measured, an average spectrum of each material is generated and a library of all such average spectra is created, along with statistically determined acceptance criteria (or thresholds) for all substances in the library. The test spectrum was compared with all reference spectra located in the electronic library. The result of the comparison between spectra A and B ends with the output of the spectral distance D, which is called the “match quality factor” in the IDENT program. Spectral distance indicates the degree of spectral similarity. Two spectra with a spectral distance equal to zero are completely identical. The greater the distance between two spectra, the greater the spectral distance. If the spectral distance is less than the threshold for one substance and greater than the threshold for all other substances, the unknown substance is identified.

Cluster analysis allows you to examine NIR spectra for similarity and divide similar spectra into groups. These groups are called classes or clusters. This type of analysis was carried out for a more convenient presentation of data in graphical form.

Hierarchical cluster algorithms are performed according to the following scheme:

First, calculate the spectral distances between all spectra,

· then the two spectra with the highest similarity are merged into a cluster,

· calculate the distances between this cluster and all other spectra,

· two spectra with the shortest distance merge again into a new cluster,

· calculate the distances between this new cluster and all other spectra,

· two spectra merge into a new cluster

This procedure is repeated until only one large cluster remains.

4 . Research results

The possibility of using the NIR spectroscopy method to identify substances and drugs from a number of domestic and foreign manufacturers has been studied.

As a result of the research, six different electronic libraries of NIR spectra were created:

1) NIR spectra of capsule contents obtained using a fiber optic sensor (“gun”),

2) NIR spectra of capsule contents obtained using an integrating sphere,

3) NIR spectra of tablets obtained using a fiber optic sensor (“gun”),

4) NIR spectra of tablets obtained using an integrating sphere,

5) NIR spectra of substances obtained using a fiber optic sensor (“gun”),

6) NIR spectra of substances obtained using an integrating sphere.

4.1. Dependence of the NIR spectra of substances and preparations on the method of preparation (using a “gun” and an integrating sphere).

In Fig. Figure 1 shows the NIR spectra of ranitidine hydrochloride substance from Vera Laboratories (India), obtained using a “gun” and an integrating sphere. The figure shows that the spectra differ in the intensity of the absorption bands, but the absorption bands themselves coincide in wavenumber values.

The main difference between NIR spectroscopy and mid-range IR spectroscopy is that the spectra cannot be compared visually. The fact is that, in general, there is an insufficient number of bands in the NIR spectrum, and the intensity of many bands is low (especially the second and third overtones), so mathematical processing of the spectra is required.

https://pandia.ru/text/78/375/images/image003_173.jpg" width="624" height="388">

Rice. 2. Result of IDENT analysis of the NIR spectrum of Ulfamid 40 mg tablets, KRKA (Slovenia), obtained using a “gun” using an electronic library of NIR spectra obtained using an integrating sphere.

Rice. 3. Result of IDENT analysis of the NIR spectrum of Ulfamid 40 mg tablets, KRKA (Slovenia), obtained using an integrating sphere using an electronic library of NIR spectra obtained using a “gun”.

4.2. Identification of the active substance by the NIR spectrum of preparations containing this substance.

https://pandia.ru/text/78/375/images/image008_152.gif" width="648" height="234"> .gif" width="648" height="244">.jpg" width="649" height="235 src=">

Rice. 7. Result of IDENT analysis of the NIR spectrum of Ciprofloxacin 250 mg tablets, Cypress Pharmaceutical Inc. (USA), using a library consisting of NIR spectra of various substances.

Thus, we have established that with a high content of the active substance (at least 40%) in the drug, it is possible to establish the authenticity of the drug by the NIR spectrum of the substance.

4.3. Identification of drugs with different dosages using NIR spectra.

In the third part of the study, we found that the NIR spectroscopy method can be used to determine various dosages of a particular drug, if they are available in the electronic library of NIR spectra. For this purpose, an electronic library of NIR spectra was created from drugs containing famotidine as an active ingredient, which included 27 samples from 7 different manufacturers in dosages of 10 mg, 20 mg and 40 mg (Fig. 8).

https://pandia.ru/text/78/375/images/image016_63.jpg" width="648" height="216 src=">

https://pandia.ru/text/78/375/images/image018_70.jpg" width="648" height="223 src=">

Rice. 9. Results of IDENT analysis, quamamg tablets, 20 mg and 40 mg, Gedeon Richter Plc. (Hungary) using a library consisting of NIR spectra of various drugs in various dosages.

4.4. Identification of drugs through the blister.

To establish the possibility of identifying drugs using NIR spectroscopy through a blister, two additional libraries of NIR spectra No. 7 and No. 8 were created:

7) NIR spectra of capsules obtained using a fiber optic sensor (“gun”) directly through the blister,

8) NIR spectra of tablets obtained using a fiber optic sensor (“gun”) directly through the blister.

During the analysis, the NIR spectra of the drugs obtained through the blister were compared with the NIR spectra obtained from the surface of tablets or capsules without the blister. In Fig. Figure 10 shows such a comparison of the spectra for rifampicin capsules.

https://pandia.ru/text/78/375/images/image020_58.jpg" width="624" height="268 src=">

Rice. 11. The result of IDENT analysis of the NIR spectrum of rifampicin 150 mg capsules, (Russia), obtained using a “gun” directly through the blister using an electronic library obtained through the blister.

https://pandia.ru/text/78/375/images/image013_124.gif" width="14" height="136">

Rice. 13 NIR spectra of the contents of omeprazole 20 mg capsules from 14 different manufacturers in comparison with a falsified sample, obtained using an integrating sphere.

From the data obtained it is clear that without mathematical processing, only the spectrum of counterfeit can be reliably distinguished.

Using the “OPUS IDENT” software for a three-dimensional model of statistical processing of spectra (“cluster analysis”), we obtained the distribution of NIR spectra of generic omeprazole 20 mg capsules, which can be presented in the form of a dendrogram (Fig. 14).

Rice. 14. Cluster analysis of the studied samples taken in triplicate from 14 different manufacturers.

As a result of the cluster analysis, all drugs were well divided into their classes and according to their manufacturer (Fig. 14).

Mathematical processing of the results obtained using IDENT analysis showed the presence of a counterfeit drug. The OPUS program determined that this sample X is indeed falsified and its “match quality coefficient” (spectral distance) is much higher than the threshold for all drugs in this group (omeprazole, 20 mg capsules) from 14 different manufacturers, from which an electronic library was created (Fig. 15).

Rice. 15. Result of IDENT analysis for a falsified sample of OMEZ 20 mg, Dr. Reddy's Lab. (India).

As a result of the IDENT analysis, a series of all original samples of omeprazole 20 mg capsules were uniquely identified, and we compiled a summary table of results for all samples, including the falsified sample (Table 1).

Table 1. Summary table of IDENT analysis results in the omeprazole group, 20 mg capsules.

Sample name	Spectral distance
Falsified sample
Sample from KRKA
Sample from Akrikhin company
Sample from Ranbaxy Laboratories
Sample from Dr. Reddy's Lab.
Sample from M. J. Boipharm
Sample company
Sample company
Sample company
Sample of the company "Pharma"
Sample of the Obolenskoye company"
Sample company. vit. factory"

Thus, as a result of the research carried out to identify medicinal products of omeprazole from various manufacturers using NIR spectroscopy, we were able to obtain results on identifying counterfeit products for the counterfeit drug OMEZ 20 mg, Dr. Reddy's Lab. (India), and also uniquely identify each generic according to its manufacturer. We also obtained positive IDENT analysis results for all tablets containing ranitidine hydrochloride (12 samples) and famotidine (9 samples), allowing us to uniquely identify the manufacturer of each sample.

GENERAL CONCLUSIONS

1. It was shown that NIR spectra of substances, tablets and capsules can be obtained using a fiber optic sensor and an integrating sphere. In this case, to establish authenticity, you should use an electronic library obtained in the same way as used to take the NIR spectrum of the test sample.

2. It has been shown that with a high content (at least 40%) of the active substance in the drug, it is possible to establish the authenticity of the drug based on the spectrum of the substance. However, in general, to identify drugs, one should use an electronic library compiled on the basis of the NIR spectra of the corresponding drugs.

3. It has been established that the NIR spectroscopy method can be used to differentiate drugs from a specific manufacturer that contain the same active substance in different dosages. At the same time, it is difficult in some cases to quantitatively determine the active substance in drugs from different manufacturers using the NIR spectroscopy method.

4. It has been shown that the NIR spectroscopy method can be used to identify the manufacturer of a substance or drug. In this case, a parallel analysis of the tested product of a specific series and a known product of the same series should be carried out.

5. An electronic library of NIR spectra of substances and preparations containing various active ingredients and manufactured by different manufacturers has been developed.

1. , Comparative assessment of the quality of drugs using near-infrared spectroscopy // Abstracts. report XII Russian National congr. “Man and Medicine.” – M., April 18-22. 2005.– P. 780.

2. , Detection of counterfeit medicines using NIR spectroscopy // Proc. report XIV Russian National congr. “Man and Medicine.” – M., April 16-20. 2007.– P. 17.

3. , The method of near-infrared spectroscopy as a promising direction in assessing the quality of medicines // Questions of biological, medical and pharmaceutical chemistry. – 2008. – No. 4. – P. 7-9.

4. , Application of the method of near-infrared spectroscopy for the identification of drugs // Questions of biological, medical and pharmaceutical chemistry. – 2008. – No. 6. – P. 27-30.

5. Arzamastsev A. P., Dorofeyev V. L., Dolbnev D. V., Houmoller L., Rodionova O. Ye. Analytical methods for rapid counterfeit drug detection. International Congress on Analytical Sciences (ICAS-2006), Moscow, 2006. Book of abstracts. V. 1. P. 108.

Modern methods for assessing the quality of medicinal raw materials and finished products include near-infrared spectrometry. The method has a number of significant advantages, including:

Simplicity of sample preparation or complete absence of its need. Eliminating this step allows you to save up to 80% of the time spent on sample examination. High speed of analysis. When using the latest generation analyzers, such as, for example, the PT IM100 NIR spectrometer, the entire process takes only 15 minutes. Possibility of studying the drug without opening the package. This feature of NIR spectrometry is especially valuable when analyzing expensive drugs, toxic substances (for example, chemotherapy drugs), etc. Drugs in transparent plastic or glass packaging can be examined without opening.

• Simultaneous analysis of various components of complex mixtures, including information about their concentrations. For example, using this method it is possible to analyze the percentage of water, organic solvents and other components in microheterogeneous systems, such as oil-in-water or water-in-oil emulsions.
• Possibility of organizing remote control of samples in real time directly in the process flow (remote control). For these purposes, stationary or portable spectrometers are used. Stationary devices are installed in production facilities of pharmaceutical enterprises, where they are integrated directly into production lines, mounting sensors above conveyor belts, in chemical reactors, and mixing chambers. This allows you to receive information online and use the received data in the automated control system. Mobile drug quality control laboratories are most often equipped with portable battery-powered NIR spectrometers.

Methods for obtaining spectra in the NIR region

In the near-infrared region, spectra are obtained using transmission or diffuse reflection.

The transmission method can be used to analyze both liquid and solid substances. In this case, liquids are placed in cuvettes or other specialized containers that are supplied with the device. Such measuring vessels can be made of ordinary or quartz glass. For transmission testing of solid samples, a probe or sphere can be used.

However, probe-based diffuse reflectance analysis has a number of significant advantages, as it provides a more detailed spectrum and more accurate results. This is achieved due to the fact that the inclined plane of the tip of the fiber optic probe minimizes the specular effect, allowing more light to be scattered. In addition, a module can be integrated into the fiber optics to read barcodes from sample packaging. It should also be noted that only with the help of a probe is it possible to identify samples remote from the device itself.

To test samples with low scattering and reflectivity, a combined transmission-reflection method is used. This requires cuvettes and sensors of a special design, thanks to which the beam flow passes through the analyzed sample twice.

In addition, “interaction” spectra can be obtained in the near-infrared region.

Problems of NIR spectrometry and ways to solve them

The main problems of this analytical method in the pharmaceutical industry for a long time have been the difficulty of analyzing the spectrum, characterized by less intense and relatively wider absorption bands compared to the fundamental bands in the mid-infrared region.

Combining mathematical methods of data processing (chemometrics) with the results of instrumental analysis made it possible to eliminate this drawback. For these purposes, modern analyzers are equipped with special software packages based on a cluster or discriminant method of processing results.

In order to be able to take into account various possible sources of changes in the spectrum in chemometric analysis, special libraries of spectra are created at pharmaceutical enterprises, taking into account the manufacturer of raw materials, the technological process of its production, the homogeneity of the material from different batches, temperature, mode of obtaining the spectrum and other factors.

According to European regulatory requirements, to compile libraries, it is necessary to study at least 3 samples of the drug substance to obtain 3 or more spectra.

Another possible problem - the possibility of a change in the spectrum due to the design features of the NIR spectrometer - is solved by qualifying the device in accordance with pharmacopoeial requirements.

Things to remember when conducting research

In NIR spectroscopy of liquid and other thermally labile samples, the nature of the spectrum depends on the degree of its heating. A difference of just a few degrees can significantly change the spectrum. This point must be taken into account when developing the recipe and testing the technology. For example, when creating a new drug or cosmetic product using a pilot laboratory homogenizer, it is often necessary to heat the homogenized mixture. A sample of the emulsion obtained in this way must be cooled before examination in an NIR spectrometer.

When studying powder raw materials, the presence of residual amounts of solvents (water, etc.) can affect the analysis results. Therefore, pharmacopoeial monographs indicate the need and technology for drying such samples. The results of near-infrared spectroscopy are influenced by the thickness of the powder layer, which directly affects the degree of transmittance. The thicker the layer, the higher the absorption. Therefore, if the testing task is to compare different samples using the transmission method, then it is necessary to prepare samples with the same layer thickness or take this indicator into account when comparing the results obtained. If the degree of reflection is analyzed, then the thickness of the layer can be any (but not less than the depth of penetration of the beam). To analyze a sample of powder using the diffuse reflection method, the layer thickness of which is less than the depth of penetration of the beam, the sample must be shielded. In addition, the characteristics of the spectrum depend on the optical properties, density, and polymorphism of the materials under study.

Benefits of NIR spectroscopy

• Easy to measure
• High accuracy and reproducibility of the analysis (the accuracy of the analysis is determined by the quality of spectrum processing, backlash and accuracy of calibration of mechanical parts, calibration of the radiation source)
• No pollution
• Possibility of taking measurements through glass and plastic packaging
• Automation of measurements. The OPUS program is used. Working with this program requires a highly qualified user
• Transferring a method from one device to another
• Analysis of physical and chemical properties

Benefits of Raman Spectroscopy

• No sample preparation required
• Due to the absence of mechanical parts and more defined spectral characteristics, measurements of Raman spectra are significantly simpler than NIR
• Raman spectroscopy measurements are considered to be chemical fingerprints (i.e., the most accurate available today). The absence of moving parts and the independence of the Raman spectrum from fluctuations in the frequency and intensity of the emitter provide ultra-high repeatability of measurements.
• No pollution
• It is possible to carry out measurements through glass (including colored glass) and plastic packaging, and the identification of individual elements (packaging and medicines) is much more reliable than in the NIR method
• Automation of measurements. A user software interface has been created that allows an untrained user to operate the device. The program is easily adapted to the end user. This point is very important for the work of pharmacists and doctors
• Raman spectra recorded on two different instruments with the same spectral resolution always coincide. Therefore, there is no problem of method transfer
• A more accurate analysis of the physical and chemical properties of the substances under study is possible, since the NIR technique measures overtones of fundamental vibrations, the direct obtaining of physical information from the energy and scattering cross sections of which is very difficult, if not impossible. Raman spectroscopy analyzes the very fundamental vibrations of chemical molecules, complete information about which is either already available or can be obtained by simple experimental and theoretical methods

Device characteristics

BIC

• Speed ​​(usually 5 – 10s)
• Compact dimensions
• Resolution determined by the width of the studied lines (about 100 cm-1)
• The minimum amount of substance for analysis is approximately 0.1 mg
• There is no database. The method has appeared recently and there are extremely few calibrated NIR spectra. This means that a huge amount of work must be done (performed by qualified personnel) to create an appropriate drug database

InSpektr

• Fast (usually less than 1 s)
• The InSpectr portable Raman complex has significantly smaller dimensions and weight than the NIR spectrometer
• Resolution determined by the width of the studied lines (about 6 cm-1). This means that a significantly larger number of substances can be identified
• The minimum amount of substance for analysis is approximately 0.001 mg (i.e. 100 times less). This is due to the better sensitivity of the receiving system in the visible range
• The method is well developed. A database of calibrated spectra of a large number of drugs and chemicals has been accumulated