Use Of Noise Parameters To Optimize Data Collection Rates In Lc-ms

Michael S. Alexander, Melissa Keiser
BASi® Northwest Laboratory, McMinnville, OR

Model analyte for all experiments was paclitaxel, M.W. 854 Da, MRM transition 854>286 Da, HPLC mobile phase ACN/water/formic acid
Data collected on Micromass Quattro Ultima and Sciex 5000 platforms

Chromatographic noise is a function of dwell time
As absolute peak height decreases, noise begins to dominate the signal (Figure 1)

With all other factors constant, there is an irreducible uncertainty placed on the signal by instrument noise
The only way to reduce this noise is by increasing the dwell time

Time - intensity data were collected for background (no analyte) at different dwell times (Figure 2).

The mean signal showed no change as a function of dwell
The signal standard deviation (SD) showed a rapidly increasing value with decreasing dwell time
(Figure 3)

Signal SD vs. the reciprocal of the dwell shows characteristic 1/f noise (Figure 4)

Multiple injections (n = 6 for all points) of analyte were made at decreasing dwell times to determine the effect of noise on precision of quantification
The %CV (peak area) tracked the standard deviation of the noise (Figure 5)

Every instrument has an inherent minimum dwell time for data collection below which precision rapidly degrades

Injections of analyte were made at decreasing dwell times on two of the same model instrument, one of which was known to be badly contaminated (Figure 6)

A plot of peak area vs. dwell is a sensitive indicator of internal contamination
Replacement of the collision cell and ion optics in the high vacuum region resulted in a significant recovery of sensitivity (Figure 7)

A plot of sensitivity vs. dwell was performed on an instrument before and after a cleaning of the source only (Figure 8)

Cleaning the source resulted in an increase in absolute response with no effect on the curve shape
The shape of the curve is affected by contamination in the high-vacuum region; the vertical displacement of the curve to contamination in the source

Two instruments were tested for response to a narrow HPLC peak (3 sec) to determine suitability for use in UPLC (Figure 9)

Loss of response with decreasing dwell time indicates the MS in red would be unsuitable for use at short dwell times

Precision of quantification is limited by inherent instrument noise
Instrument noise is determined by dwell time
Instrument sensitivity (hence cleanliness) can be estimated by a series of injections with different dwell times
The shape of the dwell vs. response curve gives information about the condition of the ion optics. Pre and post cleaning curves indicate if the sensitivity problem is in the source or the high-vacuum region of the MS
For narrow peaks (such as UPLC), the need for a minimum number of points to accurately determine peak shape may require reducing the dwell time to an interval that does not allow maximum precision