26101 Miles Road Cleveland, OH 44128 USA
Automated Quantif ication of Compounds in a Mixture: An Example Using the Universal Detection System (UDS) March 2006
Summary The Universal Detection System (UDS) is an automated platform for flexible quantification of multiple compounds in solution. The UDS combines HPLC with total nitrogen equimolar detection to provide unparalleled capabilities for truly hands-off accurate quantification of complex mixtures. This brief Technical Note illustrates the operational capabilities of the system, including instrument calibration and automated quantification of an unknown mixture containing two organic compounds in different amounts.
UDS Background The UDS combines two instruments into an automated platform: Agilent’s 1100/1200 HPLC system running ChemStation control and analysis software, and Antek Instruments’ CLND 8060 chemiluminescence detector. The system thus provides customizable chromatographic separation of peaks corresponding to different compounds in the mixture with their UV chromatograms, together with equimolar detection which results in quantification of all separated peaks on a mass nitrogen concentration basis. The latter could readily be converted to mass/volume units if each individual compound’s molecular weight and number of nitrogen per mole are known. Key technical innovations underlying the UDS operation include total automation and user-selected compound quantification optimization algorithms. The first innovation provides for total control over 10 system gain settings of the CLND and a variable HPLC injection volume. Such control enables “gain transparency” resulting in simultaneous resolution of parent peaks with very small impurity peaks. The second innovation allows user-selected flexibility to select not only optimization algorithm specific for the sample, but also results in the smallest possible number of injections and thus total analysis time. From a user perspective, the entire UDS operation is controlled via several additional menus embedded within the ChemStation software. This easy-to-use control and analysis user interface provides access to all of the UDS functionality, including equimolar calibration and sample analysis, within the familiar ChemStation software providing user-friendly operation suitable for custom and open-access operations alike.
Example The following example illustrates two of the basic operations of the UDS: 1. Equimolar calibration: The process by which a universal calibration curve in terms of peak area vs. mass nitrogen concentration is constructed and saved for further use by the system. 2. Sample quantification: The process combining chromatographic separation and iterative peak quantification for achieving optimal quantification of each identified peak.
© All Rights Reserved
2
Equimolar Calibration: Achieving quantification without standards requires the use of a universal calibration curve for the detector. In the UDS such a curve is constructed by preparing standards, injecting them into the HPLC, obtaining the response signal from the nitrogen detector, and constructing of a calibration curve. Once the calibration curve is constructed and electronically stored, it is available within the ChemStation operating software for quantification, in terms of mass nitrogen concentration, of any mixture which is chromatographically resolved. Thereafter, it can be used to quantify peaks with retention times other than that belonging to the original calibrator. Caffeine (C8H10N4O2, MW = 194.19) solutions were prepared in HPLC grade water and injected into the UDS system. The following data was obtained using repeated injections: Area Under The Curve #
mg/ml
ppmN
Inj. #1
Inj. #2
Inj. #3
Avg.
1
0.0005
0.144
113
114
96
108
2
0.001
0.289
198
163
171
177
3
0.0025
0.721
389
399
393
394
4
0.005
1.443
802
803
786
797
5
0.01
2.886
1617
1629
1608
1618
6
0.025
7.214
4284
4283
4255
4274
7
0.05
14.428
8871
8818
8866
8852
8
0.1
28.856
18549
18502
18340
18463
9
0.25
72.141
47908
47484
47484
47625
10
0.5
144.281
97254
96869
95795
96639
11
1
288.563
193570
193361
193109
193347
12
2.5
721.407
460138
462725
466960
463274
13
5
1442.814
849204
841916
852793
847971
14
10
2885.627
1494800
1486750
1476160
1485903
The above table illustrates the high repeatability of the system, and the typical very large dynamic range – over 4 orders of magnitude – which can be covered by the UDS. The calibration curve is stored within the ChemStation software by one or more user-selected formats, for example, using a piece-wise linear approximation. More typically, calibration and subsequent run data are exported and analyzed in
© All Rights Reserved
3
Microsoft Excel using automated procedures, providing for user-specific export capabilities consistent with different compound database formats. The equimolar limitation of the detection are primarily encountered for compounds containing adjacent nitrogens, azides (1/3 response), and nitrosamine. These cases are handled, if so desired, using separate calibration curves. A typical ChemStation screen in the calibration mode is depicted below.
Sample Quantification: Quantification of unknown separated peaks in the chromatogram makes use of several user-selected mathematical optimization algorithms. Essentially, the goal of such algorithms is to achieve an acceptable signal to noise ratio for peaks of interest in the assay. The optimization variables are the 10 instrument gain settings, and the variable total injection volume. The user has a choice of 4 principal optimization algorithms. The first two algorithms are aimed at automated detection and quantification of all peaks in the chromatogram, from the parent peak to the limit of detection. These algorithms run in either aggressive or conservative modes when calculating the next gain setting based primarily on evaluation of the noise (signal outside recognized peaks) properties. The other two algorithms are especially suitable for quantification of a parent peak alone, e.g., when constructing a calibration curve, aiming at the best signal to noise ratio.
© All Rights Reserved
4
In the following example, mixtures of caffeine (C) and acetaminophen (A) were prepared: Sample Name
Acetaminophen, mg/ml
Caffeine, mg/ml
Acetaminophen, ug/ml N
Caffeine, ug/ml N
AC 0.025:0.05 AC 0.01:1 AC 0.1:10
0.0250 0.0083 0.0980
0.050 0.833 9.804
2.316 0.772 9.084
14.428 240.47 2829.0
In ChemStation, UDS-specific settings are context-accessible in various modes. For example, once a standard run sequence table is set up, UDS sequence parameters are accessible as shown below:
This table allows the user to set up various parameters for each sample, describing, for example, the type of optimization algorithm to use, whether HPLC injection volume could be changed during optimization, and whether to use “smart repeats” – duplicates of the same sample using the gains identified as suitable during the first injection. Following the run, the same sequence table contains additional rows for each sample. These rows were automatically inserted during the optimization algorithm and are denoted as “dynamic” rather than “original”, which are rows originally set up by the user. Additionally, “dynamic-R” are subsequent injections related to the number of repeats requested by the user. It is noted that the number of the “dynamic-R” lines varies, depending on how many individual gains where required to resolve all of the peaks in the original optimization sequence.
© All Rights Reserved
5
Experience demonstrates that between 2-4 injections per samples are necessary to completely find and optimally resolve all peaks in the chromatogram, with a choice of 11 optimization parameters available for the UDS algorithms.
The operational sequence during optimization cycle of the UDS system is illustrated below. Following optimization of the parent peak, the algorithm dynamically increased the UDS gain in subsequent injection to resolve minor peaks, resulting in a saturated (clipped) signal in the region corresponding to the parent peak. The
© All Rights Reserved
6
saturated peak has no effect on subsequent quantification of other peaks, since once a peak is resolved, the chromatogram is effectively divided into regions already quantified and are thus ignored, and regions where additional peaks may be found and quantified.
Conclusion The Universal Detection System is a flexible, user-friendly platform for automated quantification of complex real-world samples. Using a combination of two orthogonal detection methods (spectroscopy and equimolar nitrogen detection), unknown samples could be chromatographically separated and quantified, the latter without requiring standards and in a totally automated manner where multiple optimization algorithms ensure proper and efficient peak quantification. The UDS provides the first practical solution for analyzing samples containing multiple compounds, such as impurities – a confounding bottleneck issue today in the modern analytical laboratory.
Additional UDS Functionality Planned in UDS V.2 (Q3 of 2006) 1 UDS V.2 will include: 1. Full support for latest ChemStation 32-bit release. 2. Full support for the new Agilent 1200 HPLC system. 3. Ability to automatically calculate UV response factor for each peak. This capability is highly useful for repeated quantification of the same compounds, recognizing the ease of using standard UV-only quantification once the individual peak response factors have been calculated in the UDS.
1
UDS V.1 software owners will receive V.2 upon release.
© All Rights Reserved
7
