Online Determination of Boron Isotope Ratio in Boron Triﬂuoride by Infrared Spectroscopy

: Enriched boron-10 and its related compounds have great application prospects, especially in the nuclear industry. The chemical exchange rectiﬁcation method is one of the most important ways to separate the 10 B and 11 B isotope. However, a real-time monitoring method is needed because this separation process is difﬁcult to characterize. Infrared spectroscopy was applied in the separation device to realize the online determination of the boron isotope ratio in boron triﬂuoride (BF 3 ). The possibility of determining the isotope ratio via the 2 ν 3 band was explored. A correction factor was introduced to eliminate the difference between the ratio of peak areas and the true value of the boron isotope ratio. It was experimentally found that the inﬂuences of pressure and temperature could be ignored. The results showed that the infrared method has enough precision and stability for real-time, in situ determination of the boron isotope ratio. The instability point of the isotope ratio can be detected with the assistance of the online determination method and provides a reference for the production of boron isotope.


Introduction
Boron in naturally occurring compounds is composed of two isotopes, one of mass 10 and the other of mass 11. 10 B at a relative abundance of 18.8%, has an unusually large cross section for the capture of low energy neutrons [1]. Therefore, in high concentrations, it is particularly valuable for neutron shielding [2,3].
The design and realization of the isotopes separation processes of boron began as early as the Second World War. A large industrial plant with a productivity of up to 300 kg per year of elemental boron with 95% 10 B content was put into operation in 1944 [4]. At present, the most common methods for the separation of boron isotopes include the distillation method [5], chemical exchange rectification [6,7], ion exchange [8,9], and laser isotope separation [10,11], among others. Chemical exchange rectification is the simplest and most productive method for the separation of 10 B and 11 B isotopes at concentrations of 95% and over [12]. Figure 1 shows the principal scheme of the apparatus for the separation of boron isotopes using the chemical exchange rectification method. 10 B and 11 B can be extracted from the separation and the complex parts of the apparatus, respectively [13]. The internal variations in this separation process tend to be very complex, making it necessary to frequently track any changes in the isotope ratio obtained. The mass spectrometry (MS) method is one of the most important ways of analyzing the boron isotope ratio. Analysis comes from the peaks with mass-to-charge ratios (m/e + ) equal to 10 and 11 [14]. Porteous [15] et al. determined the isotopic ratio of boron in samples of groundwater by inductively coupled plasma mass spectrometry (ICP-MS) to evaluate possible levels of boron pollution from anthropogenic inputs into natural aqueous systems. The matrix effect was reduced by the preconcentration and ion-exchange of the sample. Compared with the standard material, the precision of the method achieved 0.13% (natural abundance). However, the sampling and preparation for this method are very complex, so it is difficult to achieve real-time tracking for this process during actual production. Infrared spectroscopy (IR) is a traditional technology that has been used for on-line detection. For instance, Hepburn [16] et al. established a method using online FTIR spectroscopy to determine the siloxane content in bio-gas. Both the precision and the detection limit for siloxane were satisfactory when using this online FTIR technology.
The vibrational spectra of BF3 has been studied by Herrebout [17,18] et al.and Sluyts [19] et al.. The results were obtained by dissolving BF3 in liquid Argon. The infrared absorption bands changed depending on the change in the boron isotope ratio. Thus, the bands could be used to measure the boron isotopic ratio in BF3. Despite having a good signal-noise ratio in the ν3 region, such a high dilution factor was difficult to realize via changing optical paths in the industrialized process. Although the 2ν3 is a weak IR absorption band for BF3, the difference in absorbance peaks was about 112 cm −1 between 10 BF3 and 11 BF3 at the 2ν3 bands, which means that there is good resolution for this measurement. In this paper, the IR method was applied to determine the boron isotope ratio based on calculating the BF3 infrared spectrum at 2ν3. Further, an infrared spectrophotometer was installed in a boron separation device to monitor the separation process online.

Instruments and Reagents
The BF3 was purchased from LiuFang Gas Corporation (Dalian, LN , China) with a stated purity of 99.95%. Small amounts of SiF4 were present as an impurity in the BF3 but these were ignored in this study. NIST SRM 951a isotopic reference material of boric acid ( 10 B/ 11 B is 0.2473 ± 0.0002) was purchased from the National Institute of Standards and Technology (Gaithersburg, MD, USA). A gas cell (TianGuang Optical Instrument Co. Ltd., Tianjin, TJ, China) was equipped with the CaF2 windows (φ40 × 5 mm). The optical path length of this gas cell was 100 mm. An infrared spectrophotometer (Xintian Optical Analytical Instrument Co. Ltd., Tianjin, TJ, China) with an accuracy of transmittance lower than or equal to ±0.2%T. (TJ270-30A, Tianjin, China). The 10 B/ 11 B ratio was measured as a reference by an ICP-MS X7 (Thermo Electron Corporation, Waltham, MA, USA) series mass spectrometer manufactured by the Thermo Electron Corporation.

Procedure for Offline Measurement
The procedure for the offline experiment is shown in Figure 2. BF3 gas samples with different amounts of boron were collected from the chemical exchange rectification device or feed gas cylinder. The outlet of the gas cell was linked to a buffer bottle and an absorber bottle filled with sodium hydroxide solution to prevent BF3 gas from escaping. BF3 was filling into the gas cell, which was replaced by N2 to remove the air (especially the water in air) inside of the gas cell beforehand. This The mass spectrometry (MS) method is one of the most important ways of analyzing the boron isotope ratio. Analysis comes from the peaks with mass-to-charge ratios (m/e + ) equal to 10 and 11 [14]. Porteous [15] et al. determined the isotopic ratio of boron in samples of groundwater by inductively coupled plasma mass spectrometry (ICP-MS) to evaluate possible levels of boron pollution from anthropogenic inputs into natural aqueous systems. The matrix effect was reduced by the preconcentration and ion-exchange of the sample. Compared with the standard material, the precision of the method achieved 0.13% (natural abundance). However, the sampling and preparation for this method are very complex, so it is difficult to achieve real-time tracking for this process during actual production. Infrared spectroscopy (IR) is a traditional technology that has been used for on-line detection. For instance, Hepburn [16] et al. established a method using online FTIR spectroscopy to determine the siloxane content in bio-gas. Both the precision and the detection limit for siloxane were satisfactory when using this online FTIR technology.
The vibrational spectra of BF 3 has been studied by Herrebout [17,18] 3 in liquid Argon. The infrared absorption bands changed depending on the change in the boron isotope ratio. Thus, the bands could be used to measure the boron isotopic ratio in BF 3 . Despite having a good signal-noise ratio in the ν 3 region, such a high dilution factor was difficult to realize via changing optical paths in the industrialized process. Although the 2ν 3 is a weak IR absorption band for BF 3 , the difference in absorbance peaks was about 112 cm −1 between 10 BF 3 and 11 BF 3 at the 2ν 3 bands, which means that there is good resolution for this measurement. In this paper, the IR method was applied to determine the boron isotope ratio based on calculating the BF 3 infrared spectrum at 2ν 3 . Further, an infrared spectrophotometer was installed in a boron separation device to monitor the separation process online.

Instruments and Reagents
The BF 3 was purchased from LiuFang Gas Corporation (Dalian, LN , China) with a stated purity of 99.95%. Small amounts of SiF 4 were present as an impurity in the BF 3 but these were ignored in this study. NIST SRM 951a isotopic reference material of boric acid ( 10 B/ 11 B is 0.2473 ± 0.0002) was purchased from the National Institute of Standards and Technology (Gaithersburg, MD, USA). A gas cell (TianGuang Optical Instrument Co. Ltd., Tianjin, TJ, China) was equipped with the CaF 2 windows (ϕ40 × 5 mm). The optical path length of this gas cell was 100 mm. An infrared spectrophotometer (Xintian Optical Analytical Instrument Co. Ltd., Tianjin, TJ, China) with an accuracy of transmittance lower than or equal to ±0.2%T. (TJ270-30A, Tianjin, China). The 10 B/ 11 B ratio was measured as a reference by an ICP-MS X7 (Thermo Electron Corporation, Waltham, MA, USA) series mass spectrometer manufactured by the Thermo Electron Corporation.

Procedure for Offline Measurement
The procedure for the offline experiment is shown in Figure 2. BF 3 gas samples with different amounts of boron were collected from the chemical exchange rectification device or feed gas cylinder. The outlet of the gas cell was linked to a buffer bottle and an absorber bottle filled with sodium hydroxide solution to prevent BF 3 gas from escaping. BF 3 was filling into the gas cell, which was replaced by N 2 to remove the air (especially the water in air) inside of the gas cell beforehand. This was allowed to continue for at least 1-2 min, and the operation was stopped when the inside of the gas cell was fully replaced by BF 3 . The valves of the gas cylinder and gas cell were closed in turn for about 10 min before use. The infrared spectra were acquired from scanning the contents of the gas cell at a scanning speed of 1 cm −1 /s and a scanning range of 4000 to 400 cm −1 .
Appl. Sci. 2018, 8, x FOR PEER REVIEW 3 of 10 was allowed to continue for at least 1-2 min, and the operation was stopped when the inside of the gas cell was fully replaced by BF3. The valves of the gas cylinder and gas cell were closed in turn for about 10 min before use. The infrared spectra were acquired from scanning the contents of the gas cell at a scanning speed of 1 cm −1 /s and a scanning range of 4000 to 400 cm −1 .

Figure 2.
Diagram of offline experiment.

Procedure for Online Measurement
The online isotopic ratio determination system was set on the device used for boron isotope separation, as shown in Figure 3. The BF3 gas was extracted from a pipe where the gas flow and the pipeline can be controlled by valve. The BF3 flowed through a gas cell that was equipped with two CaF2 windows because the weatherability and corrosion resistance of CaF2 are superior to KBr under atmospheric and BF3 conditions. The whole gas cell was fitted to the IR spectrophotometer by which the infrared spectrogram could be obtained directly.

Testing method of ICP-MS
The ICP-MS was used to determine the isotope ratio as a reference. The BF3 gas was absorbed by a small amount of ethanol and diluted to samples containing 0.1 μg/mL of boron with ultrapure water. The operation method and experimental conditions of the ICP-MS experiments were similar to the method found in [15]. After each test, the system pipeline was alternately washed three times with 2% HNO3 and 0.1 mol/L ammonia water solution to reduce any errors due to memory effect. The specific parameters used for the ICP-MS experiments to determine the boron isotope ratios ( 10 B/ 11 B) are shown in Table 1.

Procedure for Online Measurement
The online isotopic ratio determination system was set on the device used for boron isotope separation, as shown in Figure 3. The BF 3 gas was extracted from a pipe where the gas flow and the pipeline can be controlled by valve. The BF3 flowed through a gas cell that was equipped with two CaF 2 windows because the weatherability and corrosion resistance of CaF 2 are superior to KBr under atmospheric and BF 3 conditions. The whole gas cell was fitted to the IR spectrophotometer by which the infrared spectrogram could be obtained directly. was allowed to continue for at least 1-2 min, and the operation was stopped when the inside of the gas cell was fully replaced by BF3. The valves of the gas cylinder and gas cell were closed in turn for about 10 min before use. The infrared spectra were acquired from scanning the contents of the gas cell at a scanning speed of 1 cm −1 /s and a scanning range of 4000 to 400 cm −1 .

Figure 2.
Diagram of offline experiment.

Procedure for Online Measurement
The online isotopic ratio determination system was set on the device used for boron isotope separation, as shown in Figure 3. The BF3 gas was extracted from a pipe where the gas flow and the pipeline can be controlled by valve. The BF3 flowed through a gas cell that was equipped with two CaF2 windows because the weatherability and corrosion resistance of CaF2 are superior to KBr under atmospheric and BF3 conditions. The whole gas cell was fitted to the IR spectrophotometer by which the infrared spectrogram could be obtained directly.

Testing method of ICP-MS
The ICP-MS was used to determine the isotope ratio as a reference. The BF3 gas was absorbed by a small amount of ethanol and diluted to samples containing 0.1 μg/mL of boron with ultrapure water. The operation method and experimental conditions of the ICP-MS experiments were similar to the method found in [15]. After each test, the system pipeline was alternately washed three times with 2% HNO3 and 0.1 mol/L ammonia water solution to reduce any errors due to memory effect. The specific parameters used for the ICP-MS experiments to determine the boron isotope ratios ( 10 B/ 11 B) are shown in Table 1.

Testing method of ICP-MS
The ICP-MS was used to determine the isotope ratio as a reference. The BF 3 gas was absorbed by a small amount of ethanol and diluted to samples containing 0.1 µg/mL of boron with ultrapure water. The operation method and experimental conditions of the ICP-MS experiments were similar to the method found in [15]. After each test, the system pipeline was alternately washed three times with 2% HNO 3 and 0.1 mol/L ammonia water solution to reduce any errors due to memory effect. The specific parameters used for the ICP-MS experiments to determine the boron isotope ratios ( 10 B/ 11 B) are shown in Table 1.  Figure 4 shows the comparison of the IR spectrograms in the 2ν 3 regions between natural abundance BF 3 and enriched 11 BF 3 . The peaks at these regions are more complex than peaks in the ν 3 regions. Fortunately, the peak changes had no significant overlap of the two bands. By contrast, the absorption band of 2ν 3 for 10 BF 3 (the range being about 3270-3170 cm −1 ) was significantly reduced, which proved that this absorption band is influenced by the amount of 10 BF 3 that is present. In contrast, the absorption peaks in the 2ν 3 band for 11 BF 3 (the range being about 3170-3070 cm −1 ) varied regularly with changing concentrations of 11 BF 3 . Thus, it was feasible to calculate the isotope ratio based on comparing the peak areas of these two absorption bands.   Figure 4 shows the comparison of the IR spectrograms in the 2ν3 regions between natural abundance BF3 and enriched 11 BF3. The peaks at these regions are more complex than peaks in the ν3 regions. Fortunately, the peak changes had no significant overlap of the two bands. By contrast, the absorption band of 2ν3 for 10 BF3 (the range being about 3270-3170 cm −1 ) was significantly reduced, which proved that this absorption band is influenced by the amount of 10 BF3 that is present. In contrast, the absorption peaks in the 2ν3 band for 11 BF3 (the range being about 3170-3070 cm −1 ) varied regularly with changing concentrations of 11 BF3. Thus, it was feasible to calculate the isotope ratio based on comparing the peak areas of these two absorption bands. The NIST SRM 951a boric acid standard sample was used to investigate the accuracy of the ICP-MS. Before testing by ICP-MS, the standard boric acid sample was dissolved and diluted to a solution containing 0.1 μg/mL of boron with ultrapure water. The relative error between the result from the ICP-MS and the standard value is shown in Table 2. The NIST SRM 951a boric acid standard sample was used to investigate the accuracy of the ICP-MS. Before testing by ICP-MS, the standard boric acid sample was dissolved and diluted to a solution containing 0.1 µg/mL of boron with ultrapure water. The relative error between the result from the ICP-MS and the standard value is shown in Table 2. The relative error between the experimental result from the ICP-MS and the isotopic ratio of the standard material was less than 0.16%. This result indicated that the ICP-MS, which was used to determine the boron isotope ratio in solution has a high accuracy, and could be used to calibrate the isotope ratio of boron in BF 3 . In Figure 4, the isotopic ratio values of the samples were calibrated by ICP-MS. The 10 B/ 11 B ratios of the natural abundance BF 3 and enriched 11 BF 3 samples were 0.2473 ± 0.0009 and 0.0526 ± 0.0011, respectively. 10 BF 3 and 11 BF 3 could be regarded as two distinct forms of matter. As is shown in Figure 5, the graph shape of A 10 represents peaks of 2ν 3 of 10 BF 3 . Similarly, A 11 represents 2ν 3 peaks for 10 BF 3 . The baselines of the spectrograms were adjusted to zero before calculation. The experiments were processed at room temperature (25 ± 3 • C). The relative error between the experimental result from the ICP-MS and the isotopic ratio of the standard material was less than 0.16%. This result indicated that the ICP-MS, which was used to determine the boron isotope ratio in solution has a high accuracy, and could be used to calibrate the isotope ratio of boron in BF3. In Figure 4, the isotopic ratio values of the samples were calibrated by ICP-MS. The 10 B/ 11 B ratios of the natural abundance BF3 and enriched 11 BF3 samples were 0.2473 ± 0.0009 and 0.0526 ± 0.0011, respectively. 10 BF3 and 11 BF3 could be regarded as two distinct forms of matter. As is shown in Figure 5, the graph shape of A10 represents peaks of 2ν3 of 10 BF3. Similarly, A11 represents 2ν3 peaks for 10 BF3. The baselines of the spectrograms were adjusted to zero before calculation. The experiments were processed at room temperature (25 ± 3 °C). It is obvious that the difference in the peak integration value is significant between the A10/A11 and the true value of 10 B/ 11 B, and this is most likely due to 10 BF3 and 11 BF3 having different molar absorption coefficients at the 2ν3 band. According to Lambert-Beer theory:

Calculation and Correction of the Isotope Ratio
where, ε is the molar absorption coefficient, c is the molarity of the substance (mol/L), b is the optical length (cm), and lg(I0/I) is the absorbance at the specific wavenumber. A10 or A11 could be regarded as the addition of several fixed number of absorptions roughly: For b10 = b11, c10i/c11i is a constant for the same gas (fixed concentration). Then: It is obvious that the difference in the peak integration value is significant between the A 10 /A 11 and the true value of 10 B/ 11 B, and this is most likely due to 10 BF 3 and 11 BF 3 having different molar absorption coefficients at the 2ν 3 band. According to Lambert-Beer theory: where, ε is the molar absorption coefficient, c is the molarity of the substance (mol/L), b is the optical length (cm), and lg(I 0 /I) is the absorbance at the specific wavenumber. A 10 or A 11 could be regarded as the addition of several fixed number of absorptions roughly: For b 10 = b 11 , c 10i /c 11i is a constant for the same gas (fixed concentration). Then: which is the correction factor, then: K = A 11 /A 10 × c 10 /c 11 (4) and c 10 /c 11 = A 11 /A 10 × K The correction factor, K, is the ratio of the addition of a series of molar absorption coefficients in essence according to Equations (1)- (5). So, boron isotope ratio can be calculated by the peak areas of A 10 and A 11 . The correction factor can be obtained by testing the BF 3 gas of a known isotope ratio. Table 3 shows the continuous determination results of the A 10 /A 11 ratio for the normal, natural abundance BF 3 sample. The ICP-MS result of 10 B/ 11 B for this sample is 0.2473. While, the isotope ratio of 10 B/ 11 B is equal to c 10 /c 11 , based on the ICP-MS result and Equation (5), the value of correction factor K is 0.5431.

Precision Experiment for the IR Method
The precision experiment was achieved by continuous measurements of a group of samples of a known isotope ratio, and the results are shown in Table 4. The relative standard deviation (RSD) was calculated by Equation (6), where, S is the standard deviation, x is the mean value of the results of n tests, and n is the number of tests. The RSD of the samples with the higher levels of isotope ratio were less than 2.00%. Although the RSD of the sample where the 10 B/ 11 B ratio was 0.0506 (by ICP-MS) was more than 5.00%, the mean value of the IR method was close to the results from the ICP-MS experiments. These results indicated that the IR method is more precise when determining the boron isotope ratio.

Influence of the Pressure and Temperature
Due to the effect of the concentration of BF 3 , absorbance values grew with an increase in pressure. In addition, an increase in the temperature made the peak pattern wilder. Figure 6A,B show the influence of pressure and temperature on the isotope ratio of 10 B/ 11 B, with the ranges of pressure and temperature being the possible interval values for the chemical exchange rectification operation. It was found that both the pressure and temperature have little impact on the isotope ratio determination results. Nevertheless, the optimum operation temperature was in the range of 278 to 298 K, because the errors might increase significantly when the temperature is higher than 303 K.
Appl. Sci. 2018, 8, x FOR PEER REVIEW 7 of 10 pressure and temperature being the possible interval values for the chemical exchange rectification operation. It was found that both the pressure and temperature have little impact on the isotope ratio determination results. Nevertheless, the optimum operation temperature was in the range of 278 to 298 K, because the errors might increase significantly when the temperature is higher than 303 K.  Figure 7 shows the IR spectrograms for an enriched 10 BF3 sample and an enriched 11 BF3 sample. It is clearly observed in Figure 7A that when the concentration of 10 B decreased, the absorption peaks of A10 were reduced and the A11 peaks increased. Similarly, the absorption peaks of A11 were reduced and the A10 peaks were increased in Figure 7B. The calculation results for the samples in Figure 7A,B were 0.0120 and 0.7536, respectively. All of the results were in good agreement with the ICP-MS as a reference standard (the isotope ratios for Figure 7A,B were 0.0102 and 0.7546, respectively) for each sample, which proved the validity of the IR method. A series of enriched BF3 samples, with their isotopic ratios determined by IR, are listed in Table 5, with the mean and standard deviation obtained from 10 determinations. By comparing the results of the IR method and the ICP-MS method, the errors of measurement increased slightly when determining the ratio of high abundance samples. This could be attributed to the decline in the signal-noise ratio (SNR), which was caused by the decrease in the absorbance value.   Figure 7 shows the IR spectrograms for an enriched 10 BF 3 sample and an enriched 11 BF 3 sample. It is clearly observed in Figure 7A that when the concentration of 10 B decreased, the absorption peaks of A 10 were reduced and the A 11 peaks increased. Similarly, the absorption peaks of A 11 were reduced and the A 10 peaks were increased in Figure 7B. The calculation results for the samples in Figure 7A,B were 0.0120 and 0.7536, respectively. All of the results were in good agreement with the ICP-MS as a reference standard (the isotope ratios for Figure 7A,B were 0.0102 and 0.7546, respectively) for each sample, which proved the validity of the IR method. A series of enriched BF 3 samples, with their isotopic ratios determined by IR, are listed in Table 5, with the mean and standard deviation obtained from 10 determinations. By comparing the results of the IR method and the ICP-MS method, the errors of measurement increased slightly when determining the ratio of high abundance samples. This could be attributed to the decline in the signal-noise ratio (SNR), which was caused by the decrease in the absorbance value. pressure and temperature being the possible interval values for the chemical exchange rectification operation. It was found that both the pressure and temperature have little impact on the isotope ratio determination results. Nevertheless, the optimum operation temperature was in the range of 278 to 298 K, because the errors might increase significantly when the temperature is higher than 303 K.  Figure 7 shows the IR spectrograms for an enriched 10 BF3 sample and an enriched 11 BF3 sample. It is clearly observed in Figure 7A that when the concentration of 10 B decreased, the absorption peaks of A10 were reduced and the A11 peaks increased. Similarly, the absorption peaks of A11 were reduced and the A10 peaks were increased in Figure 7B. The calculation results for the samples in Figure 7A,B were 0.0120 and 0.7536, respectively. All of the results were in good agreement with the ICP-MS as a reference standard (the isotope ratios for Figure 7A,B were 0.0102 and 0.7546, respectively) for each sample, which proved the validity of the IR method. A series of enriched BF3 samples, with their isotopic ratios determined by IR, are listed in Table 5, with the mean and standard deviation obtained from 10 determinations. By comparing the results of the IR method and the ICP-MS method, the errors of measurement increased slightly when determining the ratio of high abundance samples. This could be attributed to the decline in the signal-noise ratio (SNR), which was caused by the decrease in the absorbance value.    Figure 8 illustrates the online determination results from the separation part of a chemical exchange rectification device, showing the enriched stage, the stability stage, and an instability point (where the isotope ratio suddenly drops). With the assistance of IR spectroscopy, the instability point can be found and thus encourage correction of the operation. Figure 9 shows the isotope ratio curve of the complex part where the 11 BF 3 was enriched on the same isotope separation device. The errors of determination results were a little larger when the isotope ratio become smaller with the decrease in SNR. The sampling stage experiments lasted for 12 h and the sampling interval was 2 h. The curve shows the change of the isotope ratio in the process of product collection. The overtopping (exceeding limit value) point indicated that the sampling operation should be stopped at that moment, because the limit of the isotope ratio for 11 B product ( 10 B/ 11 B ≤ 0.0204) has been exceeded.  0.0120 ± 0.0079 0.0102 ± 0.0007 Figure 8 illustrates the online determination results from the separation part of a chemical exchange rectification device, showing the enriched stage, the stability stage, and an instability point (where the isotope ratio suddenly drops). With the assistance of IR spectroscopy, the instability point can be found and thus encourage correction of the operation. Figure 9 shows the isotope ratio curve of the complex part where the 11 BF3 was enriched on the same isotope separation device. The errors of determination results were a little larger when the isotope ratio become smaller with the decrease in SNR. The sampling stage experiments lasted for 12 h and the sampling interval was 2 h. The curve shows the change of the isotope ratio in the process of product collection. The overtopping (exceeding limit value) point indicated that the sampling operation should be stopped at that moment, because the limit of the isotope ratio for 11 B product ( 10 B/ 11 B ≤ 0.0204) has been exceeded.    0.0120 ± 0.0079 0.0102 ± 0.0007 Figure 8 illustrates the online determination results from the separation part of a chemical exchange rectification device, showing the enriched stage, the stability stage, and an instability point (where the isotope ratio suddenly drops). With the assistance of IR spectroscopy, the instability point can be found and thus encourage correction of the operation. Figure 9 shows the isotope ratio curve of the complex part where the 11 BF3 was enriched on the same isotope separation device. The errors of determination results were a little larger when the isotope ratio become smaller with the decrease in SNR. The sampling stage experiments lasted for 12 h and the sampling interval was 2 h. The curve shows the change of the isotope ratio in the process of product collection. The overtopping (exceeding limit value) point indicated that the sampling operation should be stopped at that moment, because the limit of the isotope ratio for 11 B product ( 10 B/ 11 B ≤ 0.0204) has been exceeded.