WO2000045155A1

WO2000045155A1 - Near infrared spectrum measurement of gases

Info

Publication number: WO2000045155A1
Application number: PCT/US1999/029399
Authority: WO
Inventors: Joanne Elizabeth Anderson; John Joseph Partridge; Kimberly Ann Mcclintock; Dwight Sherod Walker
Original assignee: Glaxo Group Ltd
Current assignee: Glaxo Group Ltd
Priority date: 1999-01-29
Filing date: 1999-12-10
Publication date: 2000-08-03
Anticipated expiration: 2001-07-29
Also published as: AU2174800A

Abstract

The present invention relates to a novel method of determining HBr levels in an effluent stream comprising measuring the near infrared (NIR) absorbance of HBr in said effluent stream. The present invention also relates to a novel method of determining the onset of a chemical reaction which is indicated by the production of HBr comprising measuring near infrared (NIR) absorbance of HBr in an effluent stream and further relates to a method of determining removal of HBr in a chemical reaction comprising measuring near infrared (NIR) absorbance of HBr in an effluent stream wherein disappearance of HBr bands indicates removal of HBr in said reaction.

Description

NEAR INFRARED SPECTRUM MEASUREMENT OF GASES

The present invention relates to a novel method of determining HBr levels in an effluent stream comprising measuring the near infrared (NIR) absorbance of HBr in said effluent stream. The present invention also relates to a novel method of determining the onset of a chemical reaction which is indicated by the production of HBr comprising measuring near infrared (NIR) absorbance of HBr in an effluent stream and further relates to a method of determining removal of HBr in a chemical reaction comprising measuring near infrared (NIR) absorbance of HBr in an effluent stream wherein disappearance of HBr bands indicates removal of HBr in said reaction. Many industrially significant chemical syntheses require bromination in closed systems, i.e., systems in which there is no access to withdraw an aliquot of the sample. For example, bromination of an organic compound may be run neat by adding liquid bromine to molten starting material (an activated organic compound where R is any organic moiety) resulting in formation of the brominated organic compound and substantial amounts of hydrogen bromide (HBr) as a byproduct of the reaction:

Reaction 1

Chemical syntheses involving brominations present a number of safety challenges such as handling liquid bromine, containment and neutralization of HBr gas, and equipment inertness to bromine compounds. See Manufacturers Safety Data Sheet (MSDS) for bromine. Thus monitoring HBr would ideally be carried out in an on-line / closed system operation in order to minimize operator exposure to liquid bromine or HBr. This neat bromination reaction in Figure 1 requires strict temperature control due to the exothermic nature of the chemistry and the formation of a strong lachrymator. Moreover, neat brominations pose an explosion hazard if liquid bromine accumulates in a closed system without reacting with the selected reactants. Activation information is thus critical from a safety perspective. Thus by monitoring the level of HBr in such a closed system, one can determine the onset of the bromination or detect whether Br₂ is simply accumulating to levels posing an explosion hazard. Normally, the analyst must employ an open system to watch for a color change in the solution to determine when the activation energy has been reached, thus increasing exposure to the analyst. Since non-transparent closed systems are more suitable for large- scale manufacturing, a method of monitoring HBr production other than watching for color change must be employed. Finally, in many reaction schemes, the HBr byproduct will react with the components in the next stage of the reaction and therefore must be completely removed. Being able to monitor the level of HBr in a closed system would allow the scientist to safely and effectively determine when HBr removal is complete so that the next stage of the reaction can begin.

Determining HBr concentration may be accomplished by removing a sample and precipitating the bromine ion in the form of silver bromide. Capillary electrophoresis may also be used to determine HBr levels. However, neither of these techniques is specific for HBr but only for bromide ions. Moreover, neither technique is amenable to on-line measurement in a closed system. Thus the manufacturing process/synthesis must be suspended until HBr levels have been determined. This in turn would require opening the system in order to recover the sample, which would itself present an increased safety hazard to the analyst. Currently available, ion-selective electrodes that are capable of measuring HBr are unlikely to survive the caustic environment of most reactions where HBr is produced and therefore are not suitable. Moreover, ion-selective electrodes are not capable of distinguishing HBr in vapor phase. Mass spectrometry is not suitable for measuring HBr in neat brominations because it is not specific for HBr, not amenable to on-line monitoring in closed systems, not continuous, corrosive to the ion source and cost prohibitive. Fourier Transform-IR (FIR) may be used to measure atmospheric HBr but requires a pathlength of 10 meters making it unsuitable for measurement in neat brominations. Moreover, FTIR gas cells require use of salt plates that are not compatible with HBr. Thus a safe, practical and quantitatively accurate method of measuring HBr concentration has to date been wanting.

A first aspect of the present invention comprises a method of determining HBr levels in an effluent stream comprising measuring the near infrared (NIR) absorbance of HBr in said effluent stream. An effluent stream may be in gas or liquid phase.

A second aspect of the present invention comprises a method of determining the onset of a chemical reaction that is indicated by the production of HBr comprising measuring near infrared (NIR) absorbance of HBr in an effluent stream. A third aspect of the present invention comprises a method of determining removal of HBr in a chemical reaction comprising measuring near infrared (NIR) absorbance of HBr in an effluent stream wherein disappearance of HBr bands indicates removal of HBr in said reaction. A particular embodiment of each aspect of the present invention comprises a method of determining HBr levels in an effluent stream comprising measuring the near infrared (NIR) absorbance of HBr in said effluent stream with a NIR spectrometer having a cell with a path length of between about 1 cm and 100 cm.

Another particular embodiment of each aspect of the present invention comprises a method of determining HBr levels in an effluent stream comprising measuring the near infrared (NIR) absorbance of HBr in said effluent stream with a NIR spectrometer having gas cell with a path length of about 25 cm.

Another particular embodiment of each aspect of the present invention comprises a method of determining HBr levels in an effluent stream comprising measuring the near infrared (NIR) absorbance of HBr in said effluent stream wherein the lower limit of detectable HBr levels is 2000 ppmV.

Another particular embodiment of each aspect of the present invention comprises a method of determining HBr levels in an effluent stream comprising measuring the near infrared (NIR) absorbance of HBr in said effluent stream wherein the lower limit of detectable HBr levels is 4000 ppmV.

Another particular embodiment of each aspect of the present invention comprises a method of determining HBr levels in an effluent stream comprising measuring the near infrared (NIR) absorbance of HBr in said effluent stream wherein the lower limit of detectable HBr levels is 8,000 ppmV. Another particular embodiment of each aspect of the present invention comprises a method of determining HBr levels in an effluent stream comprising measuring the near infrared (NIR) absorbance of HBr in said effluent stream wherein the absorbance spectrum measured is between about 1 100 nm to 2500 nm.

Another particular embodiment of each aspect of the present invention comprises a method of determining HBr levels in an effluent stream comprising measuring the near infrared (NIR) absorbance of HBr in said effluent stream wherein the absorbance spectrum measured is between about 2100 nm to 2300 nm.

Another particular embodiment of each aspect of the present invention comprises a method of determining HBr levels in an effluent stream comprising measuring the near infrared (NIR) absorbance of HBr in said effluent stream wherein HBr is produced in the following reaction (Reaction 1):

and wherein R is an organic moiety.

Another particular embodiment of each aspect of the present invention comprises a method of determining HBr levels in an effluent stream comprising measuring the near infrared (NIR) absorbance of HBr in said effluent stream wherein HBr is produced in the following reaction (Reaction 2a):

wherein X, Y and Z are independently hydrogen, halogen or organic moieties.

Another particular embodiment of each aspect of the present invention comprises a method of determining HBr levels in an effluent stream comprising measuring the near infrared (NIR) absorbance of HBr in said effluent stream wherein HBr is produced in the following reaction (Reaction 2b):

wherein X, Y and Z are independently halogen or hydrogen.

Another particular embodiment of each aspect of the present invention comprises a method of determining HBr levels in an effluent stream comprising measuring the near infrared (NIR) absorbance of HBr in said effluent stream wherein HBr is produced in the following reaction (Reaction 3):

1 -(3-chlorophenyl)- 1 -propanone 2-bromo-l-(3-chlorophenyl)-l-propanone

Another particular embodiment of each aspect of the present invention comprises a method of determining HBr levels in an effluent stream comprising measuring the near infrared (NIR) absorbance of HBr in said effluent stream wherein HBr is produced in the following reaction (Reaction 4):

1 -(3 ,5 -difluorophenyl)- 1 -propanone 2-bromo-l -(3,5-difluorophenyl)-l -propanone

The embodiments of each aspect may be independently or cooperatively employed. Detailed Description of the Invention

Instrumentation. An on-line near infrared (NIR) spectroscopic method was developed to monitor and quantitate the flow rate of HBr gas generated during this reaction. See Reaction 1 illustrated earlier. Samples were removed at various points during the reaction for capillary electrophoresis (CE) analysis. The CE data correlated residual bromide levels in the reaction mixture with NIR responses to HBr in the vapor phase. The NIR method enabled monitoring of the reaction in a closed system, allowing the scientist to safely determine when bromination began and when the HBr removal was complete.

Near Infrared Spectroscopy. A Rosemount AOTF-NIR Process Analyzer (Rosemount Analytical, Orville, OH) was used for the measurements and has been discussed in detail by Coffey, et al. See Coffey, C.K., Predoehl, A., Walker, D.S., Appl. Spec, 52, 5, 7689, 1998. Instrument resolution is specified at 25 cm^"1 between 1 100 and 2400 nm. The AOTF-NIR consists of two units, a source and a detector unit which houses the AOTF. Data from the AOTF-NIR can be collected via a computer capable of logging both spectra and PLS results, or via four 4-20 mA outputs at the instrument. Light from the instrument's tungsten-halogen source was transmitted to a high pressure 25-cm pathlength gas cell (Galileo Electro-Optic, Sturbridge, MA) by a single, 10 meter fiber optic cable (low -OH fused silica, 550/650/710, numerical aperture of 0J5 from Polymicro Technology, Phoenix, AZ).

PLS Calibration. To perform the Partial Least Squares (PLS) calibration, NIR spectra and the corresponding HBr flow rates (mL/min) were entered into GRAMS/32 running PLS/386

(both from Galactic Industries Corp., Salem, NH). The details of the PLS method are described by Massart et al. See Massart, D.L, Vandeginste, B.G.M., Deming, S.S., Michotte, Y., Kaufman,

L, Chemometrics: a textbook, Elsevier, New York, 1988. The calibration parameters included mean centering and cross validation. Mean centering is used to normalize the spectra so that small shifts in the instrument's performance will not unduly affect calibration. Cross validation is a calibration technique that removes one sample from the calibration set and predicts its value based on a model of all the other samples. This technique establishes a

Predicted Residual Error Sum of Squares (PRESS) value for each calibration point. A plot of the PRESS values allows for visual determination of the number of factors necessary for the model. For this reaction (Reaction 1) a single factor model was generated. Although only one channel was used for this work, the instrument allows up to four output channels corresponding to four predicted variables. The calibration model was then uploaded to the instrument for real-time use. Once the calibration model is initiated, the instrument continuously fits the current collected spectrum to the PLS model, yielding a predicted HBr flow rate for each scan. The initial predicted values were noted. All residual HBr was presumed gone once the predicted value returned to initial levels and the NIR spectra no longer exhibited the definitive HBr doublet.

Capillary Electrophoresis. A Hewlett Packard 3D-CE Model G1600 (Waldbronn, Germany) was used to determine residual bromide levels. The separation conditions and information about the method for the analysis can be found in Table 1. Three main features of this method differ from normal CE-anion analyses. A DB-WAX PEG modified capillary was used rather than a standard capillary, the separation potential was reverse polarity, and a low wavelength range was used for UV detection. It is also important to note the pH of the solutions, as the organic product containing the HBr was pH sensitive.

Example Gas mixing for calibration. The calibration set for HBr was developed by manually mixing nitrogen and HBr gases. Gas streams of a constant flow rate of N2 and a variable flow rate of HBr were merged and channeled through the NIR gas cell. Five calibration points were generated by assigning the intensity of a NIR peak to the flow rate of HBr that had been used while the scan was acquired. The five calibration spectra and the actual vs. predicted plot for the calibration are shown in Figures 2 and 3, respectively. The peak labels indicate the flow rate of HBr in mL/minute.

Chemistry. The bromination of an organic compound was run neat by adding liquid bromine to the molten starting material according to the generic Reaction 1. HBr gas is produced as a byproduct of the reaction. Due to the hazardous nature of this vapor and its reactivity with components of the next stage, it was imperative that all of the HBr was eliminated from the reaction system. Vacuum or a nitrogen purge was applied after completion of the bromine addition to accomplish this. However, it was unknown how long either method had to be applied to drive off all detectable HBr. Therefore, a quantitative method had to be developed to determine when all of the HBr had been removed. Two absorbance spectroscopic techniques were used to monitor the course of the reaction: NIR and UV-vis. The UV-vis measurements were taken using a fiber optic UV/vis spectrometer (Carl Zeiss Jena GmbH, Germany) with a fiber optic coupled Hastelloy^® attenuated total reflectance (ATR) probe (Equitech International Corp., Aiken, SC) inserted into the reaction mixture. Spectra were collected at set intervals during the course of the reaction. The NIR spectra were generated by monitoring the vapor phase of the reaction mixture. The NIR spectrum of neat HBr collected prior to the reaction exhibited a doublet with a maximum at 2195 nm and a shoulder (of about half the intensity) at 2246 nm. (See Figure 1) These peaks correspond to the first overtone of the mid-IR bands of HBr. The intensity of this doublet directly relates to the amount of HBr evolving from the reaction. The NIR data collected during the reaction showed the same pattern as that observed in the neat HBr spectrum, and the intensity of the doublet varied over the course of the reaction. Only minutes after pulling a vacuum on the reaction system, the doublet was no longer observed, suggesting that the HBr was gone. In order to prove that the absence of a visible HBr peak from the vapor phase indicated the absence of bromine in the reaction mixture, capillary electrophoresis was utilized.

Capillary electrophoresis (CE) was utilized as a means of validating the present invention's accuracy in measuring HBr, ;^'.e., for correlating the vapor phase and solution data and for quantifying the NIR data. Samples were pulled at several points during the reaction, after bromination, and after pulling vacuum on the system. The CE data correlated well with NIR vapor phase data; the HBr vapor peaks in the NIR and bromide ion by CE were there during the reaction and gone within minutes of pulling a vacuum. Gases were mixed manually to build a partial least squares (PLS) calibration for HBr while capillary electrophoresis (CE) proved effective for determining residual bromide in the reaction mixture. Agreement between CE and NIR data confirmed that the measurements of HBr in the effluent stream did correlate with the bromide levels in solution.

For larger scale manufacturing representing Reaction 1 , the production procedure was revised. Removal of HBr would be accomplished using a nitrogen purge rather than vacuum.

The lab scale reaction was repeated using a nitrogen purge, and again, the CE and NIR data correlated. The NIR monitoring technique proved successful on a scale from 100 mL to 12 Liter round bottom flasks, and for reactions purged by vacuum or nitrogen. Table 1. Table of Capillary Electrophoresis Conditions.

Description of the Figures

Figure 1. Neat HBr absorbance spectrum collected using the 25cm gas cell connected to a lecture bottle of HBr.

Figure 2. HBr calibration spectra labeled with the corresponding HBr flow rate (ml/min).

Figure 3: Predicted versus actual plot for the calibration data.

Claims

1. A method of determining HBr levels in an effluent stream comprising measuring the near infrared (NIR) absorbance of HBr in said effluent stream.

2. A method of determining the onset of a chemical reaction that is indicated by the production of HBr comprising measuring near infrared (NIR) absorbance of HBr in an effluent stream.

3. A method of determining removal of HBr in a chemical reaction comprising measuring near infrared (NIR) absorbance of HBr in an effluent stream wherein disappearance of HBr bands indicates removal of HBr in said reaction.

4. A method of determining HBr levels in an effluent stream comprising measuring the near infrared (NIR) absorbance of HBr in said effluent stream wherein HBr is produced in the following reaction

and wherein R is an organic moiety.

5. A method of determining HBr levels in an effluent stream comprising measuring the near infrared (NIR) absorbance of HBr in said effluent stream wherein HBr is produced in the following reaction

wherein X, Y and Z are independently hydrogen, halogen or organic moieties.

6. A method as claimed in claim 5 wherein X, Y and Z are independently halogen or hydrogen.

7. A method of determining HBr levels in an effluent stream comprising measuring the near infrared (NIR) absorbance of HBr in said effluent stream wherein HBr is produced in the following reaction

1 -(3 -chlorophenyl)- 1 -propanone 2-bromo-l-(3-chlorophenyl)-l -propanone

8. A method of determining HBr levels in an effluent stream comprising measuring the near infrared (NIR) absorbance of HBr in said effluent stream wherein HBr is produced in the following reaction

l-(3,5-difluorophenyl)-l-propanone 2-bromo-l -(3,5-difluorophenyl)-l -propanone

9. The method of any one of claims 1 to 8 wherein said HBr levels are measured with a NIR spectrometer having a cell with a path length of between about 1 cm and 100 cm.

10. The method of claim 9 wherein said HBr levels are measured with a NIR spectrometer having a cell with a path length of about 25 cm.

1 1. The method of any one of claims 1 to 10 wherein HBr levels may be detected at a lower limit of 2000 ppmV.

12. The method of any one of claims 1 to 10 wherein HBr levels may be detected at a lower limit of 4000 ppmV.

13. The method of any one of claims 1 to 10 wherein HBr levels may be detected at a lower limit of 8000 ppmV.

14. The method of any one of claims 1 to 13 wherein said absorbance is measured between about 1 100 nm to 2500 nm.

15. The method of claim 14 wherein said absorbance is measured between about 2100 nm to 2300 nm.