Analytical and Forensic Chemistry
Courses
Forensic Chemistry Laboratory I (CHEM371W)
Forensic Chemistry Laboratory II (CHEM372W)
General Chemistry II and Analytical Chemistry
Research Interests
Prior to my appointment at Duquesne University , I spent several years in the Polymers Division at NIST. I worked to develop synthetic polymer characterization by mass spectrometry. Currently, my research interests are to continue in this research, but also to apply these polymer characterization methods to forensic polymer analysis, using LC-MS, GC-MS and MALDI-MS.
Influence of Matrix and Laser Energy on the Molecular Mass Distribution of Polyethylene glycol Obtained by MALDI-TOF-MS . Polyethylene glycol was analyzed in all trans-retinoic acid (RA), 2,5-dihydroxybenzoic acid (DHB), and dithranol. The mass spectra of PEG in these matrixes are shown in the figure below. The laser energy used to obtain the mass spectra of PEG ranged from 2.0 m J to 7.5 m J in DHB, between 1.5 m J and 5.0 m J in dithranol and from 1.0 m J to 1.8 m J in RA.
In DHB and in dithranol, the moments obtained as a function of the laser energies were found to vary significantly. In DHB the number-average molecular mass (Mn) decreases as laser energy is increased, while in dithranol the Mn increases as laser energy is increased. The moments of PEG analyzed in RA were not found to vary significantly as implied by an ANOVA test. These results are reflected in the moments shown in Table 1. Both DHB and dithranol have lower mean moments than RA.
The PEG samples analyzed in dithranol and RA appear to show few signs of fragmentation in the analysis of the moments and bins. These results indicate that effects of laser energy are seen for both the PEG analyzed in DHB and dithranol. If the effect of laser energy were independent of the matrix, the expected effect of the laser energy would be the same for each matrix, when the laser energy ranges overlap. However, these results show vastly different effects when PEG was analyzed in DHB and dithranol at the same laser energies. An effect of laser energy exists, but it is matrix dependent.
Matrix |
Mn |
SD |
Corrected Mn |
SD |
|
(u) |
(u) |
(u) |
(u) |
RA |
4350 |
60 |
4390 |
60 |
Dithranol |
4080 |
140 |
4210 |
130 |
DHB |
4210 |
140 |
4420 |
60 |
PEG produces a secondary series of peaks arising from the fragmentation of the polymer. This is a result of the fragmentation occurring at the oxygen-carbon bond in the repeat unit of the PEG molecule. Each polymer molecule fragmentation produces two discernible fragment molecules. One PEG fragment molecule will have the same end groups as the original polymer molecule, and the second fragment ion will have a different end group mass. This secondary peak series is shifted sixteen mass units from the main series in the mass distribution. If the assumption is made that the PEG molecule is equally likely to fragment anywhere along the polymer chain, then the distribution of the fragmentation peaks with no end group change should be the same as the distribution of the fragment ions with an end group change.
The PEG mass spectra were integrated into peak values using Polymerix (Sierra Analytics, Modesto , CA ) analysis software. This software has the ability to identify individual peak series and calculate separate distribution information such as moments and polydispersity for the secondary series. Using this software the
PEG data were analyzed and then the fragmentation peaks were subtracted from the mass spectrum. The resulting mass spectra were then analyzed and compared with the original PEG data.
The molecular mass moments of the original data and those calculated after the fragmentation peaks were subtracted from the molecular mass distribution shown in Table 1 are averaged over all laser energies for each matrix. Regardless of matrix, all the moments increase after the fragmentation is subtracted. Changes to the effect of laser energy on the molecular mass distribution due to the subtraction of the fragment peaks are considered in Figure 3. Each matrix seems to have a different effect on the molecular mass distribution of PEG. When PEG was analyzed in RA, no effect of laser energy on the molecular mass distribution is seen. The fragmentation and change in the Mn for the PEG in RA is the same for all laser energies.
When PEG is analyzed in DHB, the effect of laser energy is different from that in RA after the fragmentation peaks are subtracted. The Mn of PEG in DHB decreased as laser energy increases. However, when the fragmentation peaks are subtracted from the mass distribution, the Mn remains constant, independent of laser energy. Therefore the effect of changing laser energy on the PEG MMD in DHB can be attributed to fragmentation.
The PEG in dithranol has a different result from the other two matrixes. The original data revealed an increase in Mn as laser energy increases. This trend in the data remains after the fragmentation peaks are subtracted from the PEG molecular mass distribution. There appears to be a uniform amount of fragmentation at all laser energies. This indicates that although there is an effect of matrix on the PEG mass distribution, it is not due to fragmentation. The increase in the Mn that is observed as laser energy increases is due to an increase in the number of high mass molecules getting into the gas phase. The analysis of the bins indicates that a higher number of high mass molecules are present at higher laser energies.
The Optimization of MALDI-TOF-MS for Synthetic Polymer Characterization by Factorial Design. One of the most significant issues in any analytical practice is optimization. Optimization and calibration are key factors in quantitation. In matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF-MS), one of the limitations restricting quantitation is instrument optimization. Understanding which parameters are most influential and the effects of these parameters on the mass spectrum is required for optimization. This understanding is especially important when characterizing synthetic polymers by MALDI-TOF-MS due to the complex nature of synthetic polymers.
The two matrixes studied were all-trans retinoic acid (RA) and dithranol. The polymer concentrations studied initially were 1:75:1, 1:150:1 and 1:225:1 by mass of polystyrene (PS):RA:silver trifluroacetate (AgTFA) and 1:40:1, 1:80:1 and 1:120:1 by mass of PS:dithranol:AgTFA. The spectra were taken from random spots on the target. Each mass spectrum was the sum of 100 laser shots. The signal-to-noise ratio was calculated for each polymer distribution using the central peak in the polystyrene molecular mass distribution (MMD) at a mass of approx. 8900 u.
A 2 5-1 fractional factorial design was employed to study the optimization of MALDI-TOF-MS. This is an orthogonal design, which is a highly efficient design. Orthogonal designs yield precise effect estimates with minimal bias in addition to the parameter effects and interaction effects of any factorial design. The 2 5-1 design is beneficial due to its ease in interpreting the results. Because there are only two options at which each parameter is run, it is easy to interpret parameter effects.
The optimization of a 9 ku polystyrene is different than for a low mass polymer. Isotope peaks are not resolvable at 9 ku in our MALDI-TOF-MS. As a result we chose to consider only signal-to-noise as a response variable. At lower mass both signal-to-noise and resolution can be considered in optimization. Our five factors in our design are all instrument parameters. Laser energy and delay time are factors that influence how much sample is desorbed and cationized in the MALDI process. The lens and extraction voltages influence the ion envelope. The detector voltage increases the sensitivity of the detector, so that all the ions reaching the detector are detected, but if the detector voltage is too high, a loss in signal-to-noise may result. Two sample preparation parameters were also considered in this analysis; the matrix used in the analysis, as well as the polymer concentration. These have both been shown to influence the measured molecular mass distribution in synthetic polymers.
The mean plots for each test reveal the effects of the instrument parameters within each matrix and polymer concentration. From these plots, it can be seen that in both dithranol and retinoic acid, the highest polymer concentration (1:5:1, PS:Matrix:AgTFA) yields the highest mean signal-to-noise. Overall, the mass spectra obtained in retinoic acid provide higher mean signal-to-noise values than those run in dithranol. A more in-depth analysis was performed on Test 4, where the highest signal-to-noise was obtained in each matrix. Test 4 is polystyrene run in RA at the highest polymer concentration. The further analysis reveals the effects of the factors on the response, as well as interaction effects.
The main effects plot reaffirms that the detector voltage and delay time are the most influential parameters on signal-to-noise. The maximum delay time and detector voltages yielded higher signal-to-noise values. Laser energy also influences the signal to noise, although lower laser energies resulted in better signal-to-noise.
Interactions occur when the effect of one factor on a response depends on the level of another factor. The interaction effects matrix shows the effect of each individual factor as well as the interaction of each factor with every other factor. The greatest effect is seen for the detector voltage. Delay length is the next highest effect and the interaction between the detector voltage and the delay length is also high. The laser energy and the interaction between laser energy and detector voltage influence the signal-to-noise to a lesser extent.
We found that the detector voltage and the time delay were the most influential of the instrument parameters for polystyrene; longer delay times and higher detector voltages cause the signal-to-noise to increase. Both the matrix and the polymer concentration influenced the optimization as well. All-trans retinoic acid yielded higher signal-to-noise values than dithranol. And higher polymer concentrations increased the signal-to-noise as well. Factorial design is a promising technique for understanding and optimizing MALDI-TOF-MS.
Selected Publications
