The ACQUITY UPLC PDA combined with the improved resolution and speed of the ACQUITY UPLC System make it the ideal solution for quantitative impurity analysis.
Some of the most challenging methods for modern LC detectors are in performing quantitative impurity analysis. These methods entail the simultaneous analysis of a large amount of parent compound and low-level impurities, thus requiring a detector with low noise levels and a wide linear dynamic range. Combining these needs with the desire to perform the separations rapidly, the detector must also be able to perform at fast data acquisition rates to reproducibly quantitate narrow peaks while not significantly increasing noise levels. It is also desirable to collect spectral data with a narrow bandwidth to maintain high spectral quality for purity analysis and library matching.
The low noise characteristics and large linear range of the Waters ACQUITY UltraPerformance LC System equipped with an ACQUITY UPLC Photodiode Array (PDA) Detector allows for easy quantitation of low-level impurities and high levels of a parent compound, with fast run times and without sacrificing spectral integrity.
LC system: |
ACQUITY UPLC System and ACQUITY UPLC PDA Detector with Empower Software |
Column: |
ACQUITY UPLC BEH C8, 2.1 mm x 50 mm, 1.7 μm |
Detection: |
Analytical flow cell 225 nm Normal time constant 10 pts/sec 1.2 nm bandwidth |
Mobile phase: |
65/35 10 mM Phosphate buffer (pH 7.0)/Acetonitrile |
Flow rate: |
0.800 mL/min |
Temp.: |
45 °C |
Samples: |
Oxybutynin and Oxybutynin Related Compounds B and C in 50/50 Water/Acetonitrile, 0.001% to 0.1% of 4.0 mg/mL standard |
Injection volume: |
4.8 μL characterized full loop |
Strong wash: |
200 μL 25/75 Water/Acetonitrile |
Weak wash: |
800 μL 75/25 Water/Acetonitrile |
Sensitivity limits of low-level impurities are characterized by the limits of quantitation (LOQ) and detection (LOD). There are several methods for determining these parameters which are discussed in the USP Chapter 1225.1 If the detection method has a noise signature, the noise level can be used to estimate LOQ and LOD. A signal-to-noise ratio of 10:1 is generally considered acceptable for estimating to LOQ and a signal-to-noise ratio of 3:1 is generally considered acceptable for estimating the LOD. To determine these values, signals of low-level samples are compared to blank injections. ASTM noise (average peak-to-peak noise) from the blank injection is used as the estimation of the noise levels for the theseparation. Other methods use the slope of thecalibration curve and the standard deviation of the response.
To establish the sensitivity limits of the ACQUITY UPLC with PDA detection, the LOQ and LOD of Oxybutynin and two Oxybutynin Related Compounds B (methyl ester of phenylcyclohexylglycolicacid) and C (4-(ethylmethylamino) but-2-ynyl (±)-2-cyclohexyl-2-hydroxy-2-phenylacetate) were determined. The chromatographic method (based upon the USP method) was scaled to the UPLC column dimensions and particle size. This method transfer step resulted in a run time reduction from 35 minutes to just 10 minutes, with equivalent resolution (2.4 between oxybutynin and the major impurity). Signal-to-noise ratios were used to determine the LOQ and LOD by generating a calibration curve near the limits of detection. The concentrations of the standards were based upon the response of a 4.0 mg/mL oxybutynin standard at ~2.0 AU full scale (Figure 1), which was within the linear range of the detector.2 The appropriate concentration range for the standards was established between 0.001% and 0.1% by mass of the main oxybutynin peak.
The chromatograms for the 0.005% and 0.01% standards, which are close to the LOQ and LOD, are shown in Figure 2. Standards were injected in triplicate. The calibration curves generated from these standards (Figure 3) were linear with R2 values in excess of 0.999. Curves were based upon height in order to easily determine the LOQ and LOD (ratio to noise). The equation for these curves used to calculate the LOQ and LOD can be found in Table 1. Blank injections were performed and the average peak-to-peak noise for these chromatograms was 27.5 μAU. From this, the LOQ and LOD were calculated for oxybutynin and its two related compounds, Table 1.
For impurity analysis, the USP recommends identification of all peaks at 0.1% of the parent compound for a drug product and 0.05% for a drug substance.3 The LOQ for oxybutynin was determined to be 0.0105%, well below the minimum requirement, while the related compounds had quantitation limits of 0.0049% and 0.0085% for B and C, respectively. To verify these results, replicate injections were performed. Typical acceptance criteria for % RSD at the LOQ is 10%. For oxybutynin, replicate injections at LOQ had a % RSD of 3.1%. The corresponding % RSD values for related compounds B and C were 3.7% and 5.8%. Limits of detection were calculated using the same method and are listed in Table 1.
Typically, a photodiode array detector is employed to gain spectral information for identification of unknowns, library matching and peak purity analysis. Therefore, it is important to maintain high spectral resolution to perform these functions. To minimize noise levels, some PDA/DAD detectors must increase bandwidth (the number of averaged diodes) or slit width (optical resolution) to improve limits of detection for low level impurities. However, increasing these parameters reduces spectral quality and can also negatively impact detector linearity. The highest quality spectra are obtained when using a narrow spectral bandwidth, typically 1–2 nm. By using a narrow bandwidth (1.2 nm), the highest resolution spectra for oxybutynin and its related compounds was achieved (Figure 4). Higher spectral resolution yield better library matching and peak purity results, especially for highly related compounds (Figure 5).
The low noise characteristics and wide linear dynamic range of the ACQUITY UPLC PDA combined with the improved resolution and speed of the ACQUITY UPLC System make it the ideal solution for quantitative impurity analysis. Limits of quantitation below 0.005% and limits of detection as low as 0.0015% are achievable without sacrificing spectral information.
720001595, March 2006