MASS SPECTROMETRY CHARACTERI ZATION OF ACYCLIC DIENE METATHESIS POLYMERS AND ETHYLENE OXIDE/PROPYLENE OXIDE COPOLYMERS By VIOLETA ILIEVA PETKOVSKA A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2005
Copyright 2005 by Violeta Ilieva Petkovska
To my parents, my brother Tony, and my late grandparents Violeta and Atanas.
iv ACKNOWLEDGMENTS Having a graduate degree is an achievem ent not only of the person who carries the honor but just as much an achievement of the people who helped and guided this person throughout the journey of educa tion. Thus, I would like to take a moment to honor and thank these people for they changed my life forever. My deepest thanks and gratitude are a ddressed to my advisor, Dr. Kenneth Wagener. Throughout my gradua te studies, he was my biggest supporter and encourager. Without him by my side, my education would ha ve never had the successful finish it has now. He is a true blessing that I will always be grateful for since he helped me obtain one of the things I value most in life: Education, a nd therefore, Independence. An advisor and a teacher who also greatly contributed towards my Ph.D studies is Dr. David Powell. He not only taught me what I love most about chemistry, mass spectrometry, but he also taught all the st udents working with hi m how to be real professionals by respecting the work we do and the people we work with. I am grateful for the training I received from him, for the excellent working atmosphere he created, for his graciousness, and care for all of us. I would also like to thank Dr. Reynolds for introducing me to Polymer chemistry in his class. Throughout the five years at the Univ ersity of Florida, I enjoyed the interaction with him (especially his sense of hum or!) and with the students in his group.
v Very many special thanks to Dr. Winefordner who, as a Chair of the Analytical Division, let me join the divisi on with a year delay and made all necessary adjustments to ease this transition and make me feel welcome. I also thank Dr. Laurie Gower, the only lady in my committee, for her patience and true interest in my research. I would like to acknowledge everyone in the mass spec lab: Dr. Jodie Jonson, Dr. Lidya Matveeva, Dr. Danielle Dickinson, Dr. Danielle Anderson, Dr. Li Qian, Dr. Jeremiah Tipton, Dr. Cris Dancel, and Ms. Lani Cardasis for their help in every aspect of my work, for their friendship, and support. Thanks to everyone in the Wagenerâ€™s group, especially Dr. Tim Hopkins who taught me polymer chemistry, guided my work in the â€œbio-olefinâ€ project, and who continues to help with my studies to this da y even though it has been a year since he left the group. Also, thanks to Dr. John Swor en who improved my writing style and Mr. Travis Baughman who, even t hough a â€œstrangerâ€ in the begi nning, eventually became my polymer synthesis â€œconsultant.â€ My graduate experience would have never been as productive and as fun without the people I shared it with. So, I thank Maria, Valerie, Samaret, Camille, and Ani for the good times we had together a nd for being true friends. In the end, I would like to thank my pare nts and my brother for letting me follow my dreams; I know that not just me but they too have paid the pric e for my education. My late grandparents Violeta and Atanas ar e the people who raised me and who have always had the most special place in my hear t. I thank them and I miss them immensely.
vi I keep the very end for the Lord Jesus Christ . The blessings He gave me and the miracles He did in my life have brought me where I am today. So, thank you Jesus for letting me finish this journey and be with me and my family in the next one.
vii TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iv TABLE OF CONTENTS..................................................................................................vii LIST OF TABLES.............................................................................................................ix LIST OF FIGURES.............................................................................................................x ABSTRACT.....................................................................................................................xi ii CHAPTER 1 INTRODUCTION TO POLY MER MASS SPECTROMETRY.................................1 Polymer Characterization.............................................................................................1 MALDI Ionization Mechanism....................................................................................4 Instrumentation.............................................................................................................6 MALDI Source......................................................................................................6 Linear Time of Flight (TOF) Analyzer.................................................................7 Delayed Extraction................................................................................................9 Reflectron Mode..................................................................................................11 Detection..............................................................................................................12 Fourier Transform Ion Cyclotron Resonance (FTICR) mass analyzer...............12 MALDI Sample Preparation.......................................................................................15 Methods for MALDI Sample Preparation...........................................................15 Choice of Matrix, Solvent, and Ionization Agent................................................16 MALDI Sample Preparation for Polyethylene Analyses....................................20 Limitations of MALDI-TOF Mass Sp ectrometry for Polymer Analyses...................21 Determination of the Average Molecula r Weight of Polydisperse Polymers.....21 Quantitative Polymer Analyses by MALDI-TOF...............................................24 Research Outlook........................................................................................................26 2 DESORPTION/IONIZATION ON SILICON MASS SPECTROMETRY: OPTIMIZATION OF THE SAMPLE PREPARATION AND THE ETCHING CONDITIONS FOR POLYMER ANALYSES.........................................................28 Introduction.................................................................................................................28 Experimental Section..................................................................................................31
viii Materials..............................................................................................................31 Silicon Surface Etching.......................................................................................32 Sample Preparation..............................................................................................33 Mass Spectrometry Analyses..............................................................................33 Results and Discussion...............................................................................................33 Conclusions.................................................................................................................37 3 MALDI-TOF DETECTION OF OLEFIN STRUCTURAL ISOMERIZATION IN METATHESIS CHEMISTRY...................................................................................38 Introduction.................................................................................................................38 Experimental Section..................................................................................................43 Materials..............................................................................................................43 MALDI Sample Preparation................................................................................43 MALDI-TOF Analyses.......................................................................................44 Results and Discussion...............................................................................................44 Conclusions.................................................................................................................54 4 FUNCTIONALITY DEPENDENT OLEFIN ACTIVITY IN ACYCLIC DIENE METATHESIS POLYMERIZATION: A MASS SPECTROMETRY CHARACTERIZATION OF AMINO AC ID FUNCTIONALIZED OLEFINS.......56 Introduction.................................................................................................................56 Experimental Section..................................................................................................60 Materials..............................................................................................................60 MALDI Sample Preparation................................................................................61 DIOS Etching Conditions and Sample Preparation.............................................62 Mass Spectrometry Analyses..............................................................................63 Results and Discussion...............................................................................................63 Conclusions.................................................................................................................77 5 MASS SPECTROMETRY ANALYSES OF ETHYLENE OXIDE/PROPYLENE OXIDE COPOLYMERS............................................................................................79 Introduction.................................................................................................................79 Experimental Section..................................................................................................91 Materials..............................................................................................................91 MALDI Sample Preparation................................................................................91 Mass Spectrometry Analyses..............................................................................91 Results and Discussion...............................................................................................92 Conclusions...............................................................................................................104 LIST OF REFERENCES.................................................................................................105 BIOGRAPHICAL SKETCH...........................................................................................115
ix LIST OF TABLES Table page 3-1 Theoretical and experimental average masse s of the dimer (n = 1), trimer (n = 2) and the tetramer (n = 3) of polymer P1.....................................................................55 3-2 Theoretical and experimental average masses of isomerized at one site (1I) dimer (n = 1), trimer (n = 2) and tetramer (n = 3) of polymer P1............................55 4-1 Theoretical average masses of the polymersâ€™ dimers...............................................66
x LIST OF FIGURES Figure page 1-1 A MALDI-TOF spectrum of poly(ethylene glycol), PEG.........................................3 1-2 A schematic of a linear MALDI-TOF instrument......................................................6 1-3 Principle of dela yed ion extraction. .........................................................................10 1-4 Operating principle of the refl ectron mode of the TOF analyzer.............................11 1-5 An FTICR analyzer cell...........................................................................................14 2-1 Anodic etching apparatus.........................................................................................32 2-2 DIOS-TOF full spectra and zooms of PEG standards with Mn varying from 900 to 3400......................................................................................................................35 3-1 An example of the ADM ET equilibrium reaction...................................................38 3-2 Structural isomerizat ion as a side reaction...............................................................40 3-3 Metathesis catalysts: Me s = 2,4,6-Trimethylphenyl, R = C(CF3)2(CH3).................41 3-4 ADMET conversion of the cy steine diene to polymer P1........................................45 3-5 MALDI-TOF spectrum of polymer P1.....................................................................45 3-6 Some possible mechanistic pathways for the formation of the metathesis, isomerization and non-productive metathesis products...........................................49 3-7 Formation of the polymerâ€™s dimers..........................................................................51 3-8 A high mass region MALDI-TOF spectrum of polymer P1.....................................53 4-1 Olefin metathesis......................................................................................................56 4-2 Olefin metathesis catalysts: Mes = 2,4,6-Trimethylphenyl, R = C(CF3)2(CH3)......57 4-3 The ADMET reaction...............................................................................................58 4-4 Formation of ADMET dimers...................................................
xi 4-5 Amino acid functiona lized ADMET polymers........................................................64 4-6 MALDI-TOF full spectra of the an alyzed amino acid functionalized ADMET polymers...................................................................................................................65 4-7 Comparison of TOF and DIOS spectra of Polymer P1............................................66 4-8 A zoom into the trimer (n = 2) and tetramer (n = 3) MALDI-TOF spectra of polymer P1 obtained in two sa mple preparation techniques....................................67 4-9 Zooms into the MALDI-TOF oligomeric peaks of Polymer P2 and Polymer P3....68 4-10 DIOS-TOF spectra of Polymer P2...........................................................................69 4-11 Zooms into the MALDI-TOF oligomeric peaks of Polymer P4 and Polymer P5....71 4-12 Bar diagrams representing the two tr ends of oligomeric product distribution calculated as a ratio of the sum of peak intensities of â€œlight â€ or â€œheavyâ€ products to NI products corresponding to each oligomer.......................................................72 4-13 A schematic representation of a po ssible sequence of metathesis (M) and isomerization (I) events leading to symme trical product distribution in the mass spectra.......................................................................................................................7 6 5-1 A MALDI-TOF spectrum of â€œUCON Lubricant 50-HB-600â€................................92 5-2 A zoom into the low mass region of the MALDI-TOF spectrum of â€œUCON Lubricant 50-HB-600â€.............................................................................................93 5-3 A zoom into the high mass region of the MALDI-TOF spectrum of â€œUCON Lubricant 50-HB-600â€.............................................................................................94 5-4 Theoretical isotopic dist ribution calculated for (EO)9(PO)34 and (EO)38(PO)12 at Rs of 5000................................................................................................................95 5-5 Theoretical isotopic dist ribution calculated for (EO)9(PO)34 and (EO)38(PO)12 at Rs of 100 000...........................................................................................................96 5-6 A zoom into the MALDI-TOF spectrum of â€œUCON Lubricant 50-HB-2000â€.......97 5-7 A MALDI-FTICR spectrum of PEG 600.................................................................98 5-8 A MALDI-FTICR spectrum of PEG 1500...............................................................99 5-9 A zoom into the MALDI-FTICR spectrum of PEG 1500......................................100 5-10 A MALDI-FTICR spectrum of â€œUCON Lubricant 50-HB-600â€...........................101
xii 5-11 A zoom into the low mass region of the MALDI-FTICR spectrum of â€œUCON Lubricant 50-HB-600â€...........................................................................................102 5-12 A zoom into the 2225-2265 Da region of the MALDI-TOF spectrum of â€œUCON Lubricant 50-HB-600â€...........................................................................................103 5-13 A zoom into the 2225-2263 Da regi on of the MALDI-FTICR spectrum of â€œUCON Lubricant 50-HB-600â€.............................................................................103
xiii Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy MASS SPECTROMETRY CHARACTERI ZATION OF ACYCLIC DIENE METATHESIS POLYMERS AND ETHYLENE OXIDE/PROPYLENE OXIDE COPOLYMERS By Violeta Ilieva Petkovska December, 2005 Chair: Kenneth B. Wagener Major Department: Chemistry Incorporation of amino acid pendant gr oups along the polyolefin backbone of acyclic diene metathesis (ADMET) polymers l eads to the formation of chiral polymer structures with potential applications as bi omaterials. On the other hand, small changes in the chemical structure of ethylene oxi de/propylene oxide (EO/ PO) copolymers induce variations in their pr operties, and in turn, allow for th eir use as surface-active agents, detergents, foams, lubricants, etc. As a re sult, studying structure-pr operty relationships in these polymers is important and mass spectro metry characterization becomes an essential step in their analyses. The goal of the research is structural characterization of a series of amino acid functionalized ADMET polymers and a set of tw o molecular weight versions of EO/PO copolymers. In addition to st ructural elucidation, the par ticular objectives of the mass spectrometry analyses target determina tion of the polymersâ€™ molecular weight distribution and polidispersity index, confirmation of previ ous analytical data about
xiv possible isomerization taking place duri ng ADMET polymerization, correlation of the structure of the ADMET polymers to the polym erization mechanism, sequencing of the EO/PO copolymers. To achieve thes e objectives, MALDI (matrix-assisted laser/desorption ionization) and DIOS (desor ption/ionization on porous silicon) were interfaced either to the time-of-flight (TOF) or the Fourier transform ion cyclotron resonance (FTICR) mass analyzer and used for polymer analyses. The mass spectrometry data of the amino acid ADMET polymers indicate that isomerization occurs as a si de reaction of metathesis pol ymerization with the second generation Grubbsâ€™ catalyst at elevated temper atures; consequently, precise structural control of the polymers under these conditions is impossible. In addition, the data show that due to their different structural characteri stics, the olefins display different selectivity towards main metathesis or isomerized products. For EO/PO mass analyses, only the resolv ing power of the FTICR mass analyzer provides resolution and accuracy that are high enough to achieve fine structural characterization of the copolymers; overlappi ng oliogmeric peaks in the low resolution MALDI-TOF spectra prevent accurate dete rmination of the unknown oligomer end groups and of the exact number of the EO and PO units in each oligomeric chain.
1 CHAPTER 1 INTRODUCTION TO POLYMER MASS SPECTROMETRY Polymer Characterization Selective change in chemical and physical properties of chemical compounds can be achieved through controlled and deliberate ch ange in chemical stru cture; as a result, detailed structural charac terization is essential in studying structure â€“ property relationships. This is especially true for polymeric materials, where the chemical structure, the extent of crystallinity, the mo lecular weight, and the distribution of polymer chain lengths determine a lot of the polymers physical properties.1;2 The distribution of chains in polymers is a ssociated with the statistical nature of polymerization, which leads to production of ch ains with different size in contrast to proteins, for example, that consist of a number of chains with the same size. As a result, polymer molecular weight terminology diffe rs from the terminology used for other compounds. In polymer science, both aver age molecular weight and polydispersity rather than a single molecular mass describe the molecular weight characteristics of the polymer. A variety of analytical tec hniquesâ€”direct (absolute) th at do not require calibration and indirect (relative) requiri ng calibration standardsâ€”are us ed in polymer science to measure polymersâ€™ molecular weight distribut ions. Light scattering and osmometry, for instance, are direct methods, whereas intrinsic viscosity and gel permeation chromatography (GPC) are indirect. Regard less of the method used, the techniques measure the number (number average molecular weight, Mn) or the weight (weight
2 average molecular weight, Mw) of the polymer chains according to the following equations: n w i i i i w i i i nM M PDI N M N M M N N M M 2 where Ni is the number of chains with mass Mi; the polydispersity index ( PDI ) is calculated as a ratio of the weight average to the number average molecular weight. From all techniques used for polymer charact erization to date, mass spectrometry is one of the newest and most powerful ones.3-9 This is an absolute method for determination of the polymer molecular wei ght and polydispersity that provides also structural information about polymerâ€™s repeat units, end gr oups, chain lengths, including chain sequence in copolymers. To obtain a ll the information that a single polymer mass spectrometry analysis gives, a combination of several other analyti cal techniques has to be used. For example, a molecular weight characterization by GPC has to be followed by NMR (nuclear magnetic resonance) and IR (infrared) spectroscopy for structural characterization. Mass spectrometry charact erization of high molecu lar weight polymers and proteins, however, was not possible until 1988, when independently of each other, the Hillenkamp and Karas group in Germany10;11 and Koichi Tanaka in Japan12 introduced a new ionization technique, matrix assisted laser desorption/ionizat ion (MALDI), that allows for desorption and ionization of high mass compounds. MALDI is a â€œsoftâ€ ionization method that provides a mechanism for transferring large molecular weight
3 molecules into gas phase with little or no fragmentation. In addition to MALDI, electrospray ionization (ESI),13;14 introduced by John Fenn at about the same time as MALDI, provided another means for high mass mo lecule mass spectrometry analyses. In ESI, however, multiply charging of the analytes occurs, which complicates the mass spectra, especially when large number of different size chains is present, like in polymers. As a result, ESI is mostly used for analysis of monodisperse proteins. Polymer analysis by ESI is also performed by in terfacing mass spectrometry to separation techniques, like GPC for example; the prior separation redu ces the polydispersity of the polymers by producing monodisperese fractions of chains, which later are individually submitted for ESI analyses.15-17 For the contributions that MALDI and ES I made towards analyses of high mass molecules, and the following breakthrough in proteomics, John Fenn and Koichi Tanaka received the 2002 Nobel Prize in Chemistry. A typical MALDI spectrum of a pol ymer is shown in Figure 1-1. Figure 1-1. A MALDI-TOF spectrum of poly(ethyl ene glycol), PEG (a) Structure of the PEG oligomeric chains (b ) A MALDI spectrum of PEG
4 The repeat unit of the polymer, poly(ethylen e glycol) (PEG) in this case, is shown in brackets, whereas the end groups dete rmined by the obtained MALDI masses are H and OH (Figure 1-1a). Depending on the de gree of polymerization, oligomeric chains with different length, and in turn mass, are produced. According to the representation we have chosen, an oligomeric chain with n = 1 corresponds to a polymer dimer, with n = 2 to a trimer, with n = 3 to a tetramer, etc. Chains of the same type are detected in the MALDI spectrum as a single mass with difference between two adjacent peaks corresponding to the mass of the polymer repeat unit, 44 Da in the case of PEG (Figure 11b). The single mass obtained for each oligom eric type chain, on the other hand, allows for determination of the degree of polymeriza tion, the repeat unit, and the end groups of the polymers. Based on the masses, the peak intensities obtained fr om the spectrum, and the equations presented earlier, computer soft ware calculates experimental values for Mn, Mw, and PDI . This is the basic information that a MA LDI spectrum could give about the polymer characteristics. In addition to that, numerous examples will be given in the following discussion for MALDI polymer data that also allow for evaluation of the polymerization mechanism, product selectiv ity, preferred chain termin ation, and product purity. MALDI Ionization Mechanism In MALDI, analyte molecules (proteins/ polymers) are mixed in a solution with matrix molecules (usually UV absorbing organi c acids) in a molar ratio of matrix:analyte ranging from 100:1 to 10 000:1 and even higher. Often ionization agents like sodium iodide (NaI) or trifluor oacetic acid (TFA) are also added in very low concentrations to aid the ionization process. About one micro liter of the matrix/analyte/ionization agent mixture is spotted on a stainless steel MALDI plate to achieve picomole range of analyte
5 deposit. Upon evaporation of the solvent, matrix and analyte co-crystallize and are subjected to irradiation by a UV laser (usual ly a nitrogen laser emitting at 337 nm), which induces desorption and ionization. A lot of research has been dedica ted to studying the MALDI ionization mechanism.18-28 The scientific outlook on the subject is divide d into two major views presented by two different theories for MALDI ion formation. In the earlier years, it was believed that ionization is due to a gas-phase proton exchange between the matrix and the analyte molecules with prior formation of a â€œgas jetâ€ by immediate transition from solid to gaseous state. For the neut ral analyte molecules to ionize, their proton affinity has to be higher than the proton affinity of the matrix,18 where proton affinity is defined as the negative enthalpy of the reaction: A + MH+ AH+ + M A â€“ analyte, M matrix The gas-phase proton exchange theory is based on the following three-step model: Incorporation and isolation of the analyte molecules in the matrix by formation of a matrix â€“ analyte â€œsolid solutionâ€ Release of the matrix and the analyte mo lecules into vacuum upon laser irradiation Ionization in the gas phase by ion-molecule reactions The critics of the gas-phase model argue that the theory is weakly supported by experimental proof of the occurring ev ents and suggest a model for MALDI ion formation that includes formation of matrix-analyte clusters. According to the â€œclusterâ€ theory, the analyte ions are preformed; in ot her words, the analyte receives charge during sample preparation while in solution with the acidic matrix. Proof of cluster formation was obtained by raising the pr essure in the MALDI source during analysis of myoglobin
6 and detection of myoglobin-matrix clusters. In addition, most of the background peaks in the MALDI spectra matched masses corresponding to matrix clusters.19 The â€œclusterâ€model explains the different velocity displayed by analyte ions depending on the matrix used in sample prepar ation. The velocity of a single analyte can range between 300 and 800 m/s depending on the t ype of matrix cluster. In addition, cluster formation is necessary for transforma tion of ionic analytes into gas phaseâ€”unlike neutrals, ionic species would decompose upon ev aporation without bein g incorporated in clusters. According to th is theory, ionization is a ccomplished mainly by charge separation during the laser induced cluster breakage; possible photoi onization reactions that follow lead to neutralization and ev aporation of some of the ionic adducts. Instrumentation A typical mass spectrometer has three majo r parts: ion source, mass analyzer, and detector. Figure 1-2 shows a basic schematic of a liner MALDI-TOF instrument. Figure 1-2. A schematic of a linear MALDI-TOF instrument; white circles represent ions of lower mass whereas black circle s represent ions of higher mass MALDI Source The source compartment accommodates the MALDI plate where the sample is deposited, optics that focus the laser beam into a desired spot of th e MALDI plate, and an
7 extraction plate situated immediately after the MALDI plate. The MALDI plate is usually held at + 20 kV potenti al so that the ions formed upon irradiation with the laser could be expelled from the plate. To accelerat e the desorbed ions into the TOF analyzer, the following extraction plate is grounded, t hus providing a total of +20 kV acceleration voltage. Linear Time-of-Flight (TOF) Analyzer The time-of-flight analyzer measures29;30 the time required for an ion to travel from the ion source to the detector according to the equation: zeV m L tof 2 where m is the mass of the ion, z is the number of charges on the ion, V is the acceleration voltage, e is the magnitude of the electronic charge, L is the length of the drift tube, and tof is the time-of-flight.31 Since the acceleration voltage and the lengt h of the flight tube are constant, by measuring the time of flight, the mass to char ge ratio can be determined from the above equation. The ions of different mass will travel different times into the drift region (field free region) of the TOF analyzer. Smaller ions will travel faster and reach the detector sooner than the larger ions that have higher mass. In Figure 1-2, the lighter ions (white) reach the detector first since th eir velocity is higher than th e velocity of the heavier ions (black). Based on this principle, the two type s of ions are separated and detected in order of increasing mass to charge ra tio. However, due to uncerta inties associated with the determination of the time-of-flight, more exac t values for the mass to charge ratio are obtained through internal or external calib ration of the instrument with standard compounds according to the equation:
8 b tof a z m 2) ( where m is the mass of the ion, z is the number of charges, tof is the time-of-flight, a, b are constants.7 The constants a and b are determined by measuring th e time of flight of two or more ions with known mass (standards). The start of the ion flight is considered to be the beginning of the laser pulse. However, ions form a short time after the applie d laser irradiation, and as a result, the real start of the ions flight is unknown. The method of calibration described above compensates for this uncertainty. During acceleration, all ions receive the same kinetic energy since the same acceleration voltage is ap plied to all of them: 22mv zeV KE where KE is the kinetic energy, z is the number of charges on ion, V is the acceleration voltage, v is the ionâ€™s velocity, m is the mass of the ion. Even though the kinetic energy is the same, the ion separation is possible due to the different ion masses, and thus, di fferent ion velocity in the time of flight mass analyzer. The linear TOF analyzer, however, is ch aracterized with ve ry poor resolving power. The resolving power for this analyzer is given by the equation: t t m m RS 2 1 where Rs is the resolving power, m is the mass of the ion, m is the width of the mass peak at half maximum, t is the time of flight of the ion, and t is the full width of the signal at half maximum (FWHM).
9 Three factors determine the low resolv ing power of the lin ear TOF analyzer: Different kinetic energies of the ions leading to kinetic energy distributions Different positions of ion formation associated with the size of the volume where the ions are formed lead ing to space distribution Different time of ion formation due to the pulse length of the laser irradiation leading to time spread and time distribution As a result of the above mentioned factor s, ions with the same mass might receive different time-of-flight in the field free drif t region, which results in broad, not well resolved mass peaks. The resolution of a lin ear MALDI-TOF instrument is usually about 300 (FWHM). Two modifications of the MALDI-TOF inst rument that address all the issues associated with kinetic, time, and space di stribution of ions improve the resolution and allow for better resolved peaks. These ar e the delayed extraction and the reflectron TOF analyzer. Delayed Extraction The principle of delayed extraction32-34 is presented in Figure 1-3. In contrast to the single stage extraction and accel eration described previous ly, where a single grounded extraction plate situated at a certain distance from the MALDI plate extracts the ions into the drift free region, the de layed extraction principle is based on a two stage extraction process. In addition to th e grounded extraction plate, anot her extraction plate is added right after the MALDI plate and before the grounded extraction plate. During delayed extraction, the acceleration vo ltage is applied after a ti me delay of a few hundred nanoseconds after the laser pulse rather than imme diately upon irradiation.7
10 Figure 1-3. Principle of dela yed ion extraction. White and black circles represent ions with the same mass having kine tic/special/time distribution According to Figure 1-3, the delayed extraction process begins with laser irradiation upon the sample applied on the MA LDI plate. Out of the three electrodes necessary for delayed extracti on, the MALDI plate plays the role of the first electrode with applied potential of + 20 kV. During th e time delay, the second electrode IS2 is held at the same potential as the first electrode IS1. At that time, ions with the same mass but different kinetic energies (represented by white and black circles) move in directions and velocities determined by their kinetic energies, thus convert their kinetic energy differences into special differences, or special distribution. Next, the voltage difference between IS1 and IS2 is increased by about 5 kV (IS2 = 15 kV), which is a fraction of the full acceleration voltage of +20 kV set by the grounded extraction plate. During this first stage of slow acceleration between IS1 a nd IS2, the applied acceleration potential transfers more energy to the ions that remain in the source for a l onger time, that is, the ions that are placed at a greater distance from IS2 (white ions in Figure 1-3). As a result, these lower energy ions receive more kinetic energy than the higher energy ions (black
11 circles in Figure 1-3) and upon application of the full acceleration potential by the ground extraction plate, the two ions will level a nd reach the detector at the same time. Reflectron Mode Apart from delayed extraction, ion reflect ron is another means to compensate for the kinetic energy distributions observed in MALDI ionization that leads to increase of the resolving power of the instrument. Introducing an ion mirror at the end of th e drift region that reflects the ions by introducing a retarding voltage through a seri es of grids and ring electrodes was first introduced by Mamyrin (Figure 1-4).35 Figure 1-4. Operating principle of the reflectron mode of the TOF analyzer Ions with the same mass but different velo cities (the two ions shown in Figure 1-4) have different flight times in linear mode (ions having higher kinetic energy (black ions) travel faster, whereas ions with lower kinetic energy (white ions) travel slower). Due to this, peak broadening and low resolution occu r in linear mode. If ions are focused through the reflectron, however, the faster i ons travel deeper into the reflectron region compared to the slower ions that travel shor ter a flight time until they are reflected. The reflection of ions is achieved through a gr adually increasing poten tial applied to the
12 reflectron grids with final pot ential having the same polarity and higher value than the applied acceleration voltage. As a result, slower ions (white ) catch up with faster ions (black) in the reflectron region of the TOF an alyzer and both reach th e detector (detector 2) at the same time. The typical resoluti on provided by a reflectron TOF analyzer is about 1000 (FWHM). The reflectron, howev er, increases the mass resolution but decreases the sensitivity a nd the detectable mass range.7;31 Detection The widely used detector in mass spectrometry is the electron multiplier.31 It consists of a series of conversion dynodes th at convert ion current to electron current where the intensity of the emitted secondary el ectrons is proportional to the intensity of the ion beam focused on the dynode. By focusing the secondary electrons emitted from the first dynode into a sec ond dynode, another electron current can be produced that, in turn, can induce a third electron curren t through a third dynode, etc., creating a multiplying effect of gradually increasing elect ron current, which is at the end amplified and detected. Fourier Transform Ion Cyclotron Resonance (FTICR) mass analyzer The TOF analyzer, in both linear and re flectron mode, provides limited resolution and mass accuracy. High mass accuracy analys es are achieved with the FTICR mass analyzer.36-41 A charged particle precess in a magnetic field with a frequency called cyclotron frequency, which is related to the mass to charge ratio of the ion according to the following equation: m qB f 2
13 where f is the cyclotron frequency, q is the charge present on an ion, B is the strength of the magnetic field. While in magnetic field, the ion experiences a force that is perpendicular to the direction of the magnetic field and to the veloci ty of the ion. Under this force, the ion moves in a circular, cyclotron motion that ha s frequency described by the above equation. The FTICR mass analyzer measures the fre quencies of the ions when subjected to magnetic field. In contrast to the time-of-flight, the cyclotron frequency does not depend on the kinetic energy of the ions or their velocities . This phenomenon is the basis for the ultra high resolution and mass accuracy displayed by the FTICR analyzer. The analyzer consists of an analyzer cubic cell as shown in Figure 1-5. The magnetic field B is oriented from the back panel of the cell to the front. When an electric field with a certain frequency is applied to the excitation plates of the cell (an opposite pair of plates parallel to th e magnetic field), the ions with cyclotron frequencies that are with resonance with the applied frequency of the electric field will increase the radius of their cyclotron oscillation. This results fr om the increased kinetic energy of the ions, which is directly related to the radius of thei r cyclotron motion. The rest of the ions in the cell, whose frequencies are not in resonance with the app lied frequency of the electric field, will remain in the center of the cell. When the radius of oscillation of the excited by the applied electric field ions exceeds the dimensions of the cell, they will collide with the detection plates of the cel l (opposing plates that are pe rpendicular to the excitation plates). By moving in the vici nity of the detection plates, th e ion packet carrying ions of
14 specific cyclotron frequency will produce an electric current that is related to the abundance of the ions. The curren t is detected and amplified. Figure 1-5. An FTICR analyzer cell The detected signal is converted by Fourie r transformation from time domain to frequency domain. The recorded frequencies are then converted to mass to charge ratios by the equation for cyclotron frequency. Th e image current produced by several different ions is called a transient. The resolution of the FTICR mass analyz er is proportional to the cyclotron frequency of the ions and to th e duration of the transient according to the equation: 2 fT R where f is the cyclotron frequency and T is the duration of the transient. As a result, higher resolution can be achie ved at higher magnetic fields and longer duration of the transient signa l. Increased collision between ions and neutrals in the analyzer cell decreases the dur ation and the amplitude of the transient. As a result,
15 FTICR instruments require ultra high vacuum for operation (10-10 Torr) in contrast to the other mass spectrometers where the high vacuum requirement is lower. MALDI Sample Preparation Methods for MALDI Sample Preparation Three methods for MALDI sample preparation exist:3 Dried droplet method. With this method, separate so lutions of the matrix and the analyte are prepared in the same solvent. The more concentrated matix solution is then mixed with the analyte solution such that a matrix to analyte molar ratio of 100:1 to 10 000:1 is achieved. A characteristic of the dried droplet method is the formation â€œsweet spots.â€ After the mixture of matrix and an alyte is deposited onto the MALDI plate and the solvent allowed to evaporate, the matrix and the analyte co-crystallize on the surface of the stainless steel plate. Some of the areas of crystallization, however, are better than others, which results in better defined crystals and higher concentration of analytes in these regions, also called â€œsweet spot s.â€ Upon irradiation with a la ser, the â€œsweet spotsâ€ will produce a better MALDI signal, wh ereas the rest of surface mi ght not give any signal at all. The presence of â€œsweet spotsâ€ aff ects the stability of the MALDI signal and complicates data interpretation and quantitation analyses. Layered method. In the layered method, the matr ix solution is applied to the MALDI plate first and the solvent allowed to evaporate. Next, the analyte solution is deposited on the top of the matrix layer. Electrospray sample deposition. This is the best method for MALDI sample preparation known to date.42 Instead of manually applying solutions of the matrix and the analyte on the MALDI plate, like in the dried droplet and the layered methods,
16 analyte/matrix solution is sprayed onto the pl ate using elsectrospray apparatus that has a high voltage power supply, a syringe pump, and a glass syringe. The syringe is filled in with the polymer/matrix solution, which is sprayed onto the plate by applying about 5 kV to the syringe needle while th e sample plate is held at gr ound. The distance between the end of the needle and the plate is between 1 and 2 cm. This method of sample deposition produces homogeneous layers of matrix/analy te on the MALDI plate where the presence of â€œsweet spotsâ€ is minimized. The layer homogeneity allows for continuous and stable MALDI signal that does not fluctuate with change of the area of lase r irradiation. By improving the strength and the stability of signal intensity, electrospray deposition allows for more reliable MALDI quantitation analyses. Choice of Matrix, Solvent, and Ionization Agent The role of the matrix in MALDI sample preparation includes isolation of the previously separated by the so lvent oligomeric chains, absorbtion of the energy of the laser, desorbion, and ionization of the analyte. As a result, the choice of an appropriate matrix is an important step in MALDI sample preparation and determines, to a big extent, the successful outcome of the MALDI experiment. Small organic acids that absorb the energy of the laser (337 nm for nitrogen laser) serve as matrix molecules. During sample prep aration, the matrix and the analyte have to form a â€œsolid solutionâ€ in which the analyt e molecules are isolated and do not form aggregates. To achieve this goal of the samp le preparation, the c hoice of matrix should be made based on matching polarity of matrix and analyte molecules. The closer the polarity of the two compounds, the better their compatibil ity to form a homogeneous â€œsolid solutionâ€. The choice of a matrix is, fo r the most part, a tria l and error process. Recently, however, Hoteling et al. used reversed-phase high-performance liquid
17 chromatography (HPLC) as a tool to determin e the polarity of the matrix and the analyte by measuring their retention times.43 In the study, HPLC was used to determine the relative polarity of 15 MALDI matrices, a set of peptides, and a number of polymer materials with different polar ity and molecular weight. The polymers of choice were poly(ethylene glycol) (PEG), polystyrene (PS) , and poly(methyl met acrylate) (PMMA). For HPLC analyses, only the low molecular we ight versions of th e polymers were used, even though later, MALDI an alyses were performed with both high and low molecular weight analytes. Hydrophobic/ non-polar matrices and analyt es have longer retention times than hydrophilic/polar compounds and el ute later in the chromatograms. The polarity order of the three polymers was dete rmined to be as expected PEG > PMMA > PS, where the polarity within a polymer specie s decreases with increase of the molecular weight. The reason for the last observati on was associated with decreased end group influence on the polymer polarity in longer chains since in this case the end groups account for a smaller percentage of the oligom er. As a result, the hydrophobicity of the polymer and its retention time increases with an increase of the molecular weight. After determination of the retention times, s ubsequent MALDI analyses showed that a combination of matrix and polymer that have close values of their HPLC retention times (matching polarity) produced the best MA LDI signals. On the other hand, big differences in the retention times of a matrix and an analyte led to poor signal intensity. Even though prior HPLC screening of the polarity of the com pounds used in MALDI sample preparation is a useful source of in formation, time and sample limitations affect its practical applications. The study, howev er, confirms the importance of the matching polarities of the matrix and the analyte for the quality of the MALDI signal.
18 Another important factor that needs to be considered during sample preparation is the pH of the matrix solution.44 MALDI analyses of a mixt ure of PMMA and a protein, with dixydroxybenzoic acid (DHB) serving as a matrix, showed that as the pH of the matrix solution was gradually increased by titration with KOH, the appearance of the corresponding MALDI spectra changed. The sa mple preparation produced best signal for both PMMA and the protein when DHB was used at its intrinsic pH of 1.8. The polymer is ionized by Na+ ions possibly present as impurities, whereas the prot ein was protonated by a proton exchange with the matrix. The good signal produced for both the polymer and the protein at pH 1.8 shows that the matrix functions as a desorption and an ionization agent under these conditions. With in crease of the pH of the matrix solution, however, the protein signal gradually decr eases relative to the polymer signal, an indication that the ionization ability of the ma trix decreases. No signal was observed at all in the region where the pH of the matrix changes fast and the solution displays no buffer capacity. In this region, the matrix form s salts, which are unable to either ionize or desorb the analytes. With further increase of the pH close to the ba sic region, only signal for the polymer is observed. In this case, the polymer probably precipitates out of the matrix/polymer solid solution, or matrix salt, and produces signal induced by laser desorption processes. In conclusion, when choosing a matrix for polymer MALDI sample preparation, two major points have to be considered: first, the polarity of the matrix and the analyte has to match, and second, the matrix solution has to be kept at its intrinsic pH during sample preparation.
19 Studies dedicated to optimization of th e MALDI sample preparation conditions also address the issue of solven t selection. In general, the matrix and the analyte have to dissolve in the same solvent. Using different solvents for the analyte and the matrix, and subsequent mixing of the solutions leads to problems during the crys tallization process. In 2005, Hoteling at al.45 studied the effect of solubility of poly (ethylen e terephthalate), PET, by using dithranol (DTH) as a matrix, and an 70:30 azeotrope mixture of methylene chloride/1,1,1,3,3,3-hexafluoro-2-propanol (b.p . 36 C) as a PET solvent. An alternative solvent, tetrahydrofuran (THF), was used to solubilize the matrix (b.p. 66 C). The study showed that the best MALDI signal was obt ained when azeotrope solvent was used to dissolve both the matrix and the polymer. Th e addition of the second solvent affects the MALDI signal as follows: if th e amount of THF in the final matrix/polymer solution is increased, the detectable MALDI mass distribu tion range of PET chai ns decreases due to decreased detection of the higher mass chains . This phenomenon is attributed to the faster evaporation of the azeotrope solvent compared to THF due to its lower boiling point. As a result, upon deposition of th e matrix/PET solvent mixture on the MALDI plate, the PET molecules will precipitate before the matrix crystallizes since PET is more soluble in the lower boiling point azeotrope wh ereas the matrix is more soluble in the higher boiling point THF. This effect of segregation is especi ally pronounced with higher mass PET chains, which precipitate firs t, while the lower mass PET chains remain in solution longer. In conclusion, finding a solvent that has good solubility characteristics for both the polymer and the matrix is a key in prepari ng a MALDI sample that produces high quality signal.
20 Many synthetic polymers ionize through alka li metal ion attachment rather than protonation. The gas-phase ioni zation theory explains this fa ct by the low proton affinity of these polymers, which prevents proton tr ansfer from the matrix to the analyte.18 As a result, in polymer MALDI sample preparati on, after selection of proper matrix and solvent, choice of ionization sa lt is the next step determin ing the outcome of the MALDI experiment. The basis of choosing proper ioni zation agent is based on the hard and soft acids and bases (HSAB) theory.46;47 Polymer chains act as bases due to the available electrons that form lone pairs or occupy -orbitals. The polymer bases share their electrons with electron deficient acids that in our case are the cations. The HSAB theory defines hard bases as molecules that are sm all and electronegative, and thus, do not tend to give their available electr ons easily. For example, PEG and PMMA are hard bases due to the electronegative oxygen pl aying the role of a hard ba se site. These types of polymers are easily ionized by hard acids like Na+ that have no unshared electrons in the outer layers. Soft base s like polystyrene PS, however, are be tter ionized by soft acids like Ag+ that posses unshared electr ons in the outer orbitals. In conclusion, when choosing a proper ionization salt for po lymer MALDI sample preparation, the principle â€œlike dissolve likeâ€ is applied to the acids and bases: hard bases react with hard acids whereas soft acids couple with soft bases. MALDI Sample Preparation for Polyethylene Analyses Ionization of synthetic polymer involves att achment of protons or alkali metal ions to functional groups present in the polymer chains. Po lymers that do not posses functionalities like polar or unsaturated functional groups ha ve been difficult to analyze by MALDI mass spectrometry due to the difficulties associated with th e attachment of an ionization cation to the non-f unctionalized polymer chains. As a result, polymers like
21 polyethylene, polypropylene, and polyisobut ylene present a ch allenge for MALDI analyses. In 2001, however, the group of Bill Wall ace from the National Institute of Standards and Technologies, reported a tw o step synthetic procedure for chemical modification of polyethylene that results in covalent attachment of a cation to the polyethylene chains.48;49 The modification method involve s bromination of the residual vinyl groups of polyethylene, which are fo rmed during chain transfer termination, followed by conversion of the brominated compound to a phosponium salt that can be ionized during MALDI. Later, in 2002, the group reported a successful substrate assisted laser/desorption ioni zation (SALDI) of polyethylene that utilizes metal powder rather than a matrix to aid the desorption and the ionization of the an alyte. In the study, cobalt powder was used as substrate assisting agent and a silver salt as an ionization agent. First, a solution of the metal powde r is spotted on a sample holder plate where it forms a thin layer after evaporation of the so lvent. Next, a solution of the ionization salt is deposited on the top of the powder laye r followed by a third layer formed by the solution of the analyte. The method detected polyethylene chains with masses up to 5000 Da. Limitations of MALDI-TOF Mass Spectrometry for Polymer Analyses Determination of the Average Molecula r Weight of Polydisperse Polymers It is well known that for polymers with PD I > 1.25 (some studies report PDI > 1.1), MALDI-TOF mass spectrometry does not give ac curate representation of the molecular weight distribution. MALDI analyses us ually underestimate the high mass components of the polymer distribution resu lting in weight average and number average values of the
22 molecular weight that are lower than the sa me values obtained by other techniques, like GPC, for example. Schriemer at al . from University of Alberta, Canada, suggest that multimer formation that occurs during analyses of polydisperse polymers affects the MALDI mass distribution, where a multimer is defined as an aggregation of two or more polymer distributions.50 MALDI analyses of a blend of PS with Mn of 5050 and 11 600 showed that the addition of a higher mass PS component ( Mn 35 000) to the blend leads to a decreased Mn value for the first two components. Comparison of the MALDI spectra of the two and the three component blends showed that the latter spectrum contains unexpected peaks at m/z around 60 000 and 40 000; the author s believe that the observed characteristic peaks in the three bl end MALDI spectrum correspond to multimer formation between the chains of the highest mass component and between the chains of the highest and the mid mass components, resp ectively. Based on the obtained evidence, it was concluded that due to multimer formation, MALDI analyses of polydisperse polymers report lower values of the average mol ecular weight compared to the real ones. As a partial solution of the observed effect, th e authors suggest that the concentration of the polymer in the matrix be decreased by increasing the matrix:analyte ratio during MALDI sample preparation. The lower concentration of polymer chains will decrease the probability for multimer formation. It is also believed that mass discrimina tion during MALDI anal yses occurs due to the different ionization efficiency of high a nd low mass chains. Ionization of very high mass chains is challenging due to the difficulties associated with transferring these chains into gas phase. Very low molecular wei ght polymers also present a challenge in
23 obtaining their accurate averag e molecular weight since lo w mass polymer chains do not cationize well. Even though different ionization efficiencies might be a factor in accurate molecular weight MALDI measurements, McEwen et al . performed studies showing that instrumental limitations of the MALDI-TOF in strument rather than ionization/desorption efficiencies of MALDI are of primary con cern in accurate molecular weight analyses.51 Several observations were made during analys es of polydisperse PMMA sample (17 K, PDI = 1.8). First, reflectron mode of the MA LDI-TOF instrument give s closer values for the polymer average molecular weight to th e reported GPC values compared to linear mode. Second, MALDI-TOF is able to dete ct polymer oligomers with masses greater than 25 000 Da which are present in the PMMA 17 K sample but not observed during routine MALDI analyses in ei ther linear or reflectron m ode. Detection of high mass chains was achieved by two methods. The fi rst one includes prior separation of the polydisperse PMMA sample into lower disp ersity fractions by GPC analyses. MALDI analyses of the early eluting GPC fractions allowed for mass spect rometry detection of the 25 000 Da oligomeric chains. The sec ond method applies deflection of the low mass ions during reflectron mode TO F analyses of the sample. Upon deflection of the low mass ions, the instrument detects only a se lected, narrow range of high mass ions with masses up to 25 000 Da. Both techniques for hi gh mass oligomeric chain detection show that the MALDI-TOF instrument is able to produce ions over the entire polymer mass distribution but not able to detect high mass ions when the low mass ions interfere with the detection of the high mass chains. Thes e observations were attr ibuted to detector response limitations due to satu ration caused by the low mass ions. In linear mode, both neutral molecules and ions reach the detect or. In reflectron mode, however, only the
24 ionized species reach the detector while the ne utrals are not reflected , a result associated with lower saturation of the reflectron detector compared to the linear detector. Consequently, the reflectron mode of the MALDI-TOF instrument demonstrates better abilities for high mass detection. The deflection of the low ma ss ions also confirmed that the instrumental response of the MALD I-TOF instrument rather than the ionization/desorption ability of the MALDI technique is responsible for inaccurate determination of the molecular we ight of polydisperse polymers. Apart from multimer formation, ionizati on efficiencies, and detector response, other factors like sample preparation and laser power might also affect the final distribution of chains in th e MALDI spectra of polydisper se polymers. Higher laser power, for example, is known to promote de tection of higher mass polymer chains. Extremely high values of the laser power, how ever, lead to fragmentation and detection of low average molecular weight values.3;7 The best approach known to date for dete rmination of accurate average molecular weight of polydisperse polymers by mass spect rometry is prior separation by GPC. The hyphenation of GPC with MALDI corrects for a lot of the limitati on associated with MALDI-TOF determination of the molecu lar weight of polydisperse polymers.52-54 Quantitative Polymer Analyses by MALDI-TOF For successful quantitation by MALDI-TOF, th e instrument has to produce a stable signal and representative of th e actual concentration of the polymer. However, it shown that the quality of the MALDI signal depe nds on a number of factors like sample preparation, laser intensity, detection saturati on, etc. As a result, accurate quantitation analyses by MALDI-TOF are difficult, and if to be performed, they have to be
25 accompanied by, first, optimization of the para meters that influence the signal intensity, and second, statistical analyses of the da ta obtained under the optimized conditions. Chen at al. studied the effects of the MALD I experimental conditions on the quantitative analysis of ethylene oxide/propylene oxide (EO/PO) copolymers.55 The study showed that the choice of solvent, matr ix, laser intensity, and matrix to analyte ratio affects the relative inte nsity of the MALDI signal. In addition, the signal was found to be mass dependent with smaller fluctuat ions of the signal intensities for lower oligomeric masses and bigger signal fluctu ations observed for higher masses. The experimental parameters described above, however, affect the absolute intensity of the MALDI signal. By introducing an internal standard and reporting the relative intensity of the analyte signal, calculated as a ratio of the absolute intensities of the analyte and the standard, it could be a ssumed that the effects of the experimental parameters on signal intensities are eliminated, at least to a certain extent. The reason for this is based on the similar acquisition parame ters and sample preparation conditions used for MALDI-TOF analyses of both the internal standard and the analyte. Nevertheless, problem arises again if the internal standa rd and the analyte diffe r in structural and chemical properties. In this case, they will ha ve different crystallization efficiency in the matrix, and eventually, produce different MALDI signal. If another oligomeric distribution of the same kind but different molecular weight is used as an internal standard, the probl ems associated with quantitation could be reduced even more. This method was used by Yan at al . to perform quantitative measurements of polydimethylsiloxane (PDMS) with Mn 2200 and 6140.56 For quantitation, the calculations were done by us ing the ratio of the intensities of the two
26 oligomeric distributions where the total si gnal intensity of each polymer was calculated according to the equation:56;57 R C EZ EI p p tM M M M M I I where It is the total signal intensity, Ip is the intensity of an oligomeric peak, Mp is the mass of the oligomeric peak, MEI is the mass of the two end groups, MR is the mass of the repeat unit, and Mc is the mass of the ionization cation. The obtained experimental data were in linear relation with the ratios based on stoichiometry calculations, which confirmed the accuracy of the analyses. A further improvement in the MALDI-TOF quantitation settings was achieved by applying electrospray sample deposition of ma trix/analyte solution on the target stainless steel plate.58 As discussed previously, in contrast to the dried droplet and the layered methods, electrospray deposition allows for uniform and homogeneous layering of the analyte on the plate, which in turn, improve s the reproducibility of the MALDI signal. Quantitation analyses of polymers by MALDI-TOF is a rapidly developing field in polymer mass spectrometry and an ongoing re search continuously develops improved experimental conditions for obt aining a stable MALDI signal.59;60 Research Outlook The goal of our research is mass spect rometry characterization of amino acid functionalized Acyclic Diene Metathes is (ADMET) polymers and ethylene oxide/propylene oxide (EO/PO) copolymers by the mass spectrometry. Both MALDITOF and MALDI-FTICR were used for structur al characterization of the polymers. To complement the MALDI data, an alternativ e ionization techniqueâ€”D esorption/Ionization on Silicon (DIOS)â€”was applied for analyses. In contrast to MALDI, DIOS is a matrix-
27 free ionization method that does not introdu ce background signal in the low mass region, and yet, provides means for â€œsoftâ€ ioni zation of polymers and proteins by producing singly charged ions with little fragmentation. The research started with experiments ev aluating the applicability of DIOS for polymer analyses and optimization of the DI OS sample preparation conditions, which are followed by the actual mass spectrometry char acterization of the polymers. The obtained structural data give insight into the m echanism of the polymer ization reactions and provide valuable information necessary for determining the properties of the polymers.
28 CHAPTER 2 DESORPTION/IONIZATION ON SILICON MASS SPECTROMETRY: OPTIMIZATION OF THE SAMPLE PREPARATION AND THE ETCHING CONDITIONS FOR POLYMER ANALYSES Introduction Desorption/Ionization on Silicon (DIOS) is a relatively new mass spectrometry ionization method introduced by Gary Siuzdak et al. from the Scripps Research Institute, California, in 1999.61;62 In DIOS, porous silicon is used as a substrate to absorb the UV energy of the laser, trap the analyte molecules, desorb, and ionize them. In general, this ionization technique is coupled to the TOF analyzer and produ ces singly charged, easy to interpret spectra. A number of advantages th at DIOS has over other deso rption/ionization techniques, like MALDI, for example, make it a preferred choice of ionization for analyses of certain analytes. First, DIOS requi res little sample preparation since choosing a proper matrix, one of the challenges of MALDI, is eliminat ed. The analytes are simply dissolved in a solvent and spotted on an etched spot. Second, DIOS does not produce background peaks in the low mass region like MALDI si nce it is a matrix-free technique. Consequently, DIOS provides means for de tection of low mass molecules with the benefits of soft ioniza tion given by MALDI. Porous silicon is usually produced by a nodic etching where the silicon wafer is deposited in a hydrofluoric acid/ethanol solution while constant current is passing through the solution. The morphology of the pr oduced etched silicon surface depends, to a big extent, on the etching conditions.63-66 Longer etching times produce larger, deeper
29 pores, which were found to initiate desorpti on and ionization of larger molecules. Smaller pores, on the other hand, promot e desorption and ionization of smaller molecules. Varying the concentration of the hydrofluoric acid et ching solution and the type of solvent used to prepare it have littl e or no effect on the performance of the DIOS chip. It is known, however, that higher concen trations of hydrofluoric acid in the etching solutions produce deeper pores. The upper mass detectable limit for routine DIOS analyses is about 3000 Da even though polypept ides with mass to charge of 18 000 were also observed. Despite the be nefits that DIOS displays, the limitation in detecting high mass analytes restricts its use only to low mass compounds. In addition, the following additional disa dvantages further limit, at the present degree of development, the application of DIOS for routine analyses of a wider variety of analytes, including polymers. First, backgr ound signal in the low mass range (less than 300 Da) is still observed by DIOS. Analyses of the DIOS background signal showed that some of the peaks might be due to alipha tic hydrocarbons absorbed as impurities by the porous silicon either from the air or from the pump oil under vacuum. The background DIOS signal can be reduced by post-etching washing of the etched silicon chip in ethanol. Not only is the etched silicon chip washed in ethanol, but it is also stored in it. Special storage of the etched silicon is necessary due to oxidation of the initially hydrophobic etched Si surface, dominated by silicon hydride (Si-H) groups, to a hydrophilic one, dominated by oxide (Si-O-Si) and hydroxide (Si-OH) moieties. Surface oxidation degrades the DIOS performance a nd presents another major disadvantage of the DIOS technique. The special handling a nd storage requirements not only affect the desorption/ionization process, but also introdu ce time constraints on silicon chip usage.
30 For best results, freshly etched chips have to be used for DIOS analyses. After a couple of days, even stored in ethanol, the etched chipsâ€™ efficiency to ionize the sample decreases. Surface modification is used to stabil ize the silicon chip and improve its performance. It has been s hown that replacement of the labile Si-H bonds with Si-C bonds protects the silicon surface from oxida tion. This transformation is usually achieved by Lewis acid surface functiona lization with alkynes and alkenes.67;68 The hydrophobic silicon surface presents a chal lenge in depositing an aliquot of the analyte solution on the small area defined by an etched spot (~800 m in diameter). Hydrophilic solvents, like water, are easily concentrated on a small area of the chip, whereas hydrophobic solvents, like tetrahydrofur an (THF), have small contact angle with the silicon surface and spread ea sily upon deposition. As a re sult, analyses of synthetic polymer samples by DIOS present solvent limitati on, in addition to th e other restrictions described previously, since most of the synthetic polymers ar e soluble in hydrophobic organic solvents rather than water. In th e case of solvent spread, the analyte cannot be applied on the surface at a concentration that meets the detection limit requirements of DIOS-TOF, which results in poor signal. The sample droplet evaporation pattern de pends on the shape of the etched spot.69 Rectangular and star-shaped sample wells lead to analyte concentr ation in the corners rather than in the center of th e etched silicon spots. Round spots prove to be the best for sample deposition since the analyte is concentr ated in the spot center upon evaporation of the solvent.
31 The ionization mechanism of DI OS is believed to be similar to the mechanism of substrate assisted laser desorption (SALDI) wh ere the surface promotes charge separation of analyte ions from counter ions.70 Upon laser irradiation, the preformed ions are desorbed into gas phase by heating deso rption. While in SALDI desorption and ionization are considered to depend on th e roughness of the surface and occur on sharp edges of it, in DIOS, the porosity of the surf ace is a determining factor for desorption and ionization. The number and the variety of compounds analyzed by DIOS range from small organic molecules to natural product s and peptides, including polymers.71-74 Despite the diversity of studied applicat ions, little research has been done on DIOS analyses of polymers. The goal of our DIOS work includes optim ization of the etching and the sample preparation conditions and determination of the detectable upper mass limit for polymer analyses. This was achieved by DIOS analyses of poly(ethylene glycol) standards (PEG) with different average molecular weights pe rformed under different etching and sample preparation conditions. Experimental Section Materials An n-type silicon wafer with resistivity 1-10 â€¢cm and 525 microns thickness was purchased from Silicon Sense (Nashua, NH). Ethanol (200 proof) was purchased from Aaper Alcohol and Chemical Co. (Shelbyville , KY). Hydrofluoric ac id (48 to 50%) and HPLC grade chloroform was obtained from Fisher Scientific (Fair Lawn, NJ).
32 Poly(ethylene glycol) PEG standards w ith number-average molecular weights ( Mn) of 600, 1500, 2000, 3400, 4600, and 8000 were purchased from Sigma-Aldrich (Milwaukee, WI). Silicon Surface Etching Porous silicon was produced by anodic et ching. The silicon wafer was cut into small pieces (2 cm 2 cm) and sealed in a Teflon cell with a gold electrode lying on the bottom of the cell and a platinum electrode (cathode) placed into a 1:1 (v/v) HF/Ethanol etching solution filling in the cell cavity (Figure 2-1). Figure 2-1. Anodic et ching apparatus An O-ring positioned between the wafer and the top part of the cell prevents the etching solution from leaking out side the cell. The wafer is irradiated with a white light that is focused through a series of lenses to pass through a patterned mask so that the image of the mask is transf erred onto the silicon wafer. Upon completion of the etching
33 process by inducing a 30 s c onstant current of 4 mA/cm2 between the electrodes, a sequence of numbered spots is produced as an image of the mask (a few hundred micrometers in diameter) with nanometer sized pores. Sample Preparation The PEG standards were dissolved in dei onized water at differe nt concentrations and spotted as 1 L aliquots on an etched spot . No matrix and an ionization agent were added to the solution. All de tected ions were sodiated. Mass Spectrometry Analyses Bruker Reflex II TOF (Ballerica, MA) instrument was used for DIOS-TOF analyses. The DIOS etched chip was att ached to the sample holder with double sided conductive type. The instrument is equipped with a 337 nm nitrogen laser and delayed extraction. Positive ions were detected in linear mode with accelerat ion voltage of 20 kV and 100 shots acquired for every spot. Br uker Xmass software was used for data processing and theore tical calculations. Results and Discussion Analyses of PEG standards with number average molecular weights of 900, 1500, 2000, 3400, 4600, 8000, and 8000 were performed at varying concen trations of the analytes (110-4 M, 5-3 M, 1-2 M , 4-2 M ) on silicon wafers etched for 10 s, 30 s, 90 s, and 120 s. Based on previously publis hed results, the expected outcome of the experiments was that the l onger etched silicon chips w ould produce better signal for higher molecular weight analytes, whereas the shorter etched silicon surface will be more suitable for lower mass poly(ethylene glycol). For PEG standards with Mn of up to 3400, the best signal was obtained from silicon chips etched for 30s at 1-2 M concentration of the analytes. Figure 2-2 shows the
34 DIOS-TOF spectra of PEG 900, 1500, 2000, and 3400 obtained under the described optimized etching and sample preparation c onditions. For compar ison, the acquisition parameters of the instrument were also kept constant for all the analytes throughout the analyses. Zoom into the spectrum of each standard shows the presence of a characteristic sequence of peaks with ma ss difference of 44 Da betw een two adjacent peaks, corresponding to a PEG repeat unit. The DI OS spectrum of PEG 900, however, displays two sets of peaks corresponding to two di fferent distributions, each having difference between adjacent peaks of 44 Da with the second distribution shifted by 16 Da from the first distribution. The displayed arrangeme nt of peaks found in the spectrum of PEG 900 is most probably due to ionization by Na and K ions, which gives the characteristic double distribution with 16 Da mass difference. The obtained DIOS data for the lower mol ecular weight PEG st andards show that: first, background signal in the low mass region is observed in DIOS-T OF spectra of the polymers, second, the intensity of the bac kground signal increases with increase of the average molecular weight of the polymer, and third, previously re ported trends of pore size correlation of the DIOS pe rformance to the molecular we ight of the analytes were not confirmed. In other words, larger por e-sized silicon chips do not produce better DIOS signal for higher molecula r weight analytes. Instead, an optimal etching time of 30 seconds, corresponding to a medium sized pores , seems to generate best signal for all analytes, independently of their molecular weight.
35 Figure 2-2. DIOS-TOF full spectra and zooms of PEG standards with Mn varying from 900 to 3400 The increase of the backgr ound signal with increase of the molecular weight is probably due to increased fragmentation of th e higher mass polymer chains. Since all the acquisition parameters, laser power including, are kept the same for DIOS analyses of all analytes, the increased fragmentation produced by the higher molecular weight analytes is probably due to desorp tion/ionization processes.
36 The DIOS-TOF data obtained for the highe r molecular weight PEG standards, PEG 4600 and PEG 8000, are shown in Figure 2-3. Figure 2-3. DIOS-TOF spect ra of PEG 4600 and PEG 8000 Out of all different etching times and anal yte concentrations used to optimize the DIOS signal of these higher molecular weight polymers (same varia tions in etching and sample preparation parameters as for the lo wer molecular weight analytes); no optimal conditions for DIOS analyses were found. Figure 2-3 represent data obtained at 1-2 M concentration of the analytes from a silic on wafer etched for 30 s. Analyses performed at other concentrations of th e analytes and other etching conditions did not lead to a significant improvement of the DIOS spectra. The data were acquired under the same acquisition parameters used for analyses of the lower molecular weight polymers with the exception of an increased laser power (1.6 ). In all cases, almost no signal was observed at the expected mass range around 4600 or 8000, respectively. Instead, the analytes fragment in the low mass region producing a sequence of fragme nt peaks (Figure 2-3). In
37 this case, the high level of the obtained fragment ation is due to the increased laser power. The increase of the laser pow er for analyses of the higher mass analytes is inevitable since the laser threshold genera lly increases with increase of the molecular weight of the analytes in both MALDI and DIOS. Conclusions DIOS analyses of PEG standards with diffe rent average molecular weight led to the following conclusions. First, the detect able upper mass region for DIOS polymer analyses determined with PEG standards is about 3500 Da, not too different from the detectable upper mass region obtained for othe r types of analytes like small molecules and peptides. Second, the a pplicability of DIOS for polym er analyses decreases with increase of the average molecular weight of the analyte due to increased fragmentation and background in the low mass region. Thir d, the best DIOS signal for any of the different molecular weights was observed fr om silicon wafers etched for 30 seconds.
38 CHAPTER 3 MALDI-TOF DETECTION OF OLEFIN STRUCTURAL ISOMERIZATION IN METATHESIS CHEMISTRY Reproduced with permission from V.I. Pe tkovska, T.E. Hopkins, D.H. Powell and K. B. Wagener Macromolecules 2005, 38, 5878-5885. Copyright 2005 American Chemical Society. Introduction Synthesis of polymers with desired prope rties and morphologies can be achieved through variations in their stru ctural design; precis e structural control of macromolecular structure is preferred, of course, but is not easily ach ieved. For example, chain propagation synthesis of polyeth ylene leads to chain transfer and formation of randomly branched polymer chains, a result which has been exploited for more than 50 years to generate a large family of polyethylene materi als. While random br anching can be put to good use, it is of value to systematica lly study precise branching such that structure/property relationships in polymers can be more carefully identified. Recently, we have embarked on such a study using step polymerization acyclic diene metathesis (ADMET) chemistry rather than chain chemis try, which results in the preparation of polyethylene having branches placed at specific positions along the polymer backbone (Figure 3-1).75;76 R R n catalyst + CH2CH29 9 9 9 Figure 3-1. An example of the ADMET equilibrium reaction Being a step-growth condensation polyme rization, ADMET permits the synthesis of linear and/or branched pol yolefins with double bonds and br anches placed repeatedly
39 at exact positions along the polymer chain. Precisely branched pol yethylene structures with branches placed on every 9th, 11th, 15th, 19th, and 21st carbon have been made, where such structures display different thermal prope rties from those exhibited for linear and the randomly branched polyolefin analogues. In particular, ADMET polyethylene exhibits sharp, better defined melting and recrystalliz ation peaks in the DSC (differential scanning calorimetry) curves as compared to broa d melts in randomly branched copolymers.77;78 However, ADMET does not always produce pr ecise structural design when olefin isomerization competes with condensation meta thesis chemistry, an issue that has been related to catalyst chemistry. As a result, metathesis isomerization has been subject of extensive research. Double bonds can migrate (isomerize) along the polymer backbone and/or in the monomer during olefin metathesis chemistry (F igure 3-2), and as a result the number of the methylene spacer units se parating the branch point on the backbone can become irregular. This process also can affe ct the precise positioning of any additional functionalities the po lyolefin might have, since the distance between them is also determined by the number of the methylene sp acers. As a further effect on structural order in ADMET polymerization, isomerizat ion changes not only the regularity of methylene spacing within chains of the same length, but it also changes the length of these chains. For example, dimers with different lengths are produced when isomerized versions of the monomer enter the ADMET polymerization. As a result, the ADMET mechanism leads to release of higher mass molecules like propylene and butylene, instead of just ethylene. The identity of the molecules released relates to the structure of the dimer; that is, isomerized dimers vary in length by multiples of fourteen Daltons (a
40 methylene group), corresponding to the differenc es in mass between the olefin molecules released. R R R R R R (a) (b)isomerization isomerization 9 9 9 8 9 9 9 9 9 10 8 9 Figure 3-2. Structural isomerization as a si de reaction (a) in the monomer (b) in the polymer backbone Research suggests that isomerization is pr omoted by the catalyst used to initiate metathesis.79-82 Out of the three most common metath esis catalysts (Figure 3-3), complex 2 is reported to promote significant isomerization, whereas 1 and 3 give little to no isomerization side products.80;82 In general, Catalyst 3 demonstrates the highest activity among the three, but at the same it exhibits high sensitivity to oxyge n/moisture and polar functional groups.83;84 On the other hand, catalyst 1 and 2 show less air sensitivity.85;86 Among these three catalysts, complex 2 is the most active towards metathesis of functionalized monomers.87-92 As a result, this catalyst plays an important role in polymerization and ring closing metathesis reactions, and research efforts are dedicated to elimination of its isomerization activity.
41 Ru Ph PCy3Cl Cl PCy3Ru Ph PCy3Cl N N Mes Mes Cl N Mo Ph RO RO 1 3 2 Figure 3-3. Metathesis catalysts: Mes = 2,4,6-Trimethylphenyl, R = C(CF3)2(CH3) As part of these efforts, Bourgeois et al . reported that proper selection of solvents and additives can eliminate isomerization wi th Ru-based metathesis catalysts in ringclosing metathesis (RCM).79 The addition of tricycl ohexylphosphine oxide or oxygen inhibits isomerization, whereas using more coordinating solv ents favors it. Additional research in this area reports that other types of additives can also enhance or reduce the catalyst activity.93-95 Even though additive introduc tion and careful selection of reaction conditions can provide a degree of control over the catalyst activity, metathesis isomerization with complex 2 remains a challenge. Consequently, a vari ety of analytical methods have been used to study isomerization, the catalysts that promote it, and the reaction products that result from it. Among the most common t echniques are GC, GC-MS, NMR, and X-ray crystallography;80-82;85 Matrix-Assisted Laser Deso rption/Ionization (MALDI) mass spectrometry has also been applied to study me tathesis but not with respect to olefin isomerization.96-98 As a â€œsoftâ€ ionization method, MALDI provides a means to transform large, synthetic macromolecules from solution or so lid state to a gas pha se with little or no
42 fragmentation. Singly charged ions, as oppos ed to species with multiple charges, are created in the gas phase, which pr oduce easy to interpret mass spectra.10-12 In contrast to other analytical techniques, such as ge l permeation chromatography (GPC), MALDI permits mass measurements and struct ural characterization of the polymers.3;5;6;99 The spectra give information not only about pol ymersâ€™ molecular wei ght distribution and polydispersity index (PDI) but also about the nature of their repeat units and end groups.100;101 Polymer analysis by MALDI, however, usually is limited to structures having heteroatoms or unsaturated f unctionalities that can serve as ionization sites, such as polyethylene glycol or polystyrene (even though MALDI analysis of non-functionalized polyethylene has also been reported49;102). In addition, MALDI does not provide accurate determination of the molecular weight dist ribution of polydisperse samples since it usually underestimates the molecular weight, as heavier species are harder to vaporize and detect. Coupling chromatography (GPC ) and mass spectrometry overcomes this limitation and is widely used for polymer characterization.12;15;17;50;52-54;60;103 This chapter describes MALDI-TOF dete ction of isomerization found in ADMET polymerization of amino acid bear ing diene when using catalyst 2. The amino acid functionalities make the polymer chains easy to ionize by MALDI and provide mass spectra that not only complement other an alytical data used for characterization91 but also give insight into certain aspects of metathesis isomerization. In the following discussion, the MALDI data are used to confir m the structure of the polymer, relate this structure to the mechanism of ADMET polym erization and isomerization, and compare
43 the conclusions and hypothesis drawn from these data with previously published analysis of metathesis isomerization. Experimental Section Materials Dithranol (1,8-Dihydr oxy-9[10H]-antracenone), HPLC grade, was purchased from Sigma-Aldrich (St. Louis, MO) and used w ithout further purifica tion in MALDI sample preparation. Certified sodium iodide a nd HPLC grade tetrahydrofuran (THF) were obtained from Fisher Scientific (Fair La wn, NJ) and used as received. For mass calibration, the standard poly( ethylene glycol) with number average molecular weights ( Mn) of 600, 1500, 2000, 3400, 4000, and 8000 was purchased from Sigma-Aldrich (Milwaukee, WI). Literature reports a vari ety of procedures for synthesis of polymers having amino-acid moieties;104-107 the procedure for the ADMET synthesis of the analyzed amino acid polymer is previously published as well.89-91;108 MALDI Sample Preparation Solutions of the matrix (dithranol), th e polymer, and the ionization salt were prepared in THF as follows: 4 mg/mL matrix solution, 1 mg/mL polymer solution, and 1 mg/mL sodium iodide. Appropr iate volumes of the three so lutions were mixed to obtain matrix:polymer:ionization salt rati os of 1:1:1, 5:1:1, and 7:1: 1. One microliter of each mixture was spotted on a stainle ss steel plate and the solvent was allowed to evaporate. The data presented in this paper were obtained from spots having ratio of matrix:polymer:ionization salt 1:1:1 since this ratio produ ced the best MALDI signal. Solutions of the PEG standards in deioni zed water were prepared as follows: 1 L/mL PEG 600, 1 mg/mL PEG 1500, 1 mg/mL PEG 1500, 1 mg/mL PEG 3400, 2 mg/mL PEG 4600, and 3 mg/mL of PEG 8000. A mixture of each standard PEG solution
44 with 4 mg/mL dithranol water solution was pr epared in matrix:PEG ratio of 10:1; one microliter of each mixture was spotted on a MALDI plate. MALDI-TOF Analyses All spectra were acquired with Bruker Reflex II MALDI-TOF (Billerica, MA) instrument retrofitted with delayed extrac tion and a nitrogen laser emitting at 337 nm. All ions producedâ€”from the analyte and the PEG standardsâ€”were sodiated. We speculate that the sodiated ions of the PEGs are due to impurities possibly present in the matrix, the solvents, or the glassware. The positive ions were detected in linear mode with acceleration voltage of 20 kV after 100 laser shots were summed for every spot. Analyte detection in reflectron mode was also attempted but with no success. External calibration with PEG standards having mol ecular weights covering the mass range of interest provided a mass accuracy of 0.1% in linear mode. All data processing and theoretical mass calculations were done with Bruker Xmass software. Results and Discussion The ADMET polymer P1 (Figure 3-4) is designed to have precisely-placed amino acid branches along the polymer backbone, wher e the protected amino acid (cysteine in this case) induces polymer chirality.91 A zoom into the first three oligomeric p eaks [n = 1 (dimer), n = 2 (trimer), n = 3 (tetramer)] of the MALDI spectrum of the polymer (Figure 3-5a) shows that every oligomeric peak consists of a cluster of peaks sepa rated by fourteen daltons, corresponding to a methylene group.
45 9 9 n 9 9 HN O S O O catalyst + H2CCH2 NH O O HN O S O O NH O O P1 Figure 3-4. ADMET convers ion of the cysteine diene to polymer P1 Were isomerization not to occur, the MA LDI spectrum would cons ist only of single peaks representing the oligomeric chains of the dimer, the trimer, or the tetramer, etc., with masses 1409, 2088, and 2767 (see Table 3-1 at the end of this ch apter). Obviously, more chemistry is occurring than just ADMET polymerization. Figure 3-5. MALDI-TOF spectrum of polymer P1 (a) Full spectrum (b) A zoom into the first three oligomeric peaks of the spectrum; NI â€“ non-isomerized product; 1I, 2I, 3I, 4I, 5I â€“ products with isomerization at one, two, three, etc., sites â€“ â€œlight productsâ€, (red); NP â€“ products resulting from isomerization with nonproductive metathesis-like reacti ons â€“ â€œheavy productsâ€, (blue)
46 The balance of the peaks in Figure 3-5 resu lt from olefin structural isomerization (olefin migration) occurring along with ADMET polymerization, which leads to formation of oligomers having the same numb er of monomer units attached, but yet, a different chain length. This translates into having a cluster of peaks for each oligomer in the MALDI mass spectrum. For example, the dimer of the polymer is represented not only by the peak at m/z 1409 at the MALDI spectrum, but also by the peaks at m/z 1339, 1352, 1366, 1380, 1395, and 1423. When olefin isomerization occurs only at one site 1I, that is, the double bond moves one position away from its original place in the non-isomerized product NI, the isomerized product formed ( m/z 1395, 2074, 2753, etc.) could ha ve a structure such as the one shown in Table 3-2 (table at the end of this chapter). To confirm, the experimental masses from the MALDI spectrum are shown to match the theoretical masses calculated based on the proposed structure. The isomerized product 1I has one methylene group le ss in the polymer backbone compared to the non-isomerized product NI, which in the MALDI mass spectrum shows as a difference of 14 Daltons. Likewise, olefin isomerization that occurs at more than one site forms products (2I, 3I, 4I), each having masses lower than the mass of the nonisomerized product by multiples of fourteen. In addition, every oligomeric cluster of p eaks in the spectrum also has peaks with masses higher than the mass of the non-isomer ized product by multiples of fourteen. The mechanism of ADMET polymerization93;109 can be used to explain the formation of all products observed in the MALDI spectrum (Figure 3-6) when c onsidering the presence of olefin bond migration. Apparently, thr ee major types of products form during
47 polymerizationâ€”main metathesis products (NI), â€œlight productsâ€ ha ving mass lower than the mass of the corresponding main produc ts, and â€œheavy productsâ€ with mass higher than the mass of the corresponding main meta thesis products. Three pathways, all consistent with the ADMET mechanism, play a role in the observed metathesis polymerization: productive metathesis (M), olefin isomerization followed by productive (condensation) metathesis (I+M), and olefin isomerizati on followed by non-productive (trans metathesis) pathways (I+NP). The MALDI-TOF data cl early can discern each one of them. During productive metathesis (M), Figure 3-6â€”middle rectangle, monomer A first reacts with the catalyst to fo rm the cyclobutane intermediate B, which later rearranges to C. Upon the addition of another monomer molecule to C, the cyclobutane intermediate D is formed. In the next step, the productive ADMET mechanism releases the polymerâ€™s dimer NI ( m/z 1409) with a structure co rresponding to a non-isomerized, main metathesis product (also see Table 1). Next, anothe r monomer molecule forms the cyclobutane E with the modified version of the catalyst followed by release of ethylene gas. The catalytic cycle repeats in the same manner to produce the trimer ( m/z 2088), tetramer ( m/z 2767), etc. When isomerization occurs (Figure 3-6â€”last rectangle, red), isomerized versions of the monomer A (such as A1 and A2) take part in the metathesis cycle; these events could be described by the second major process occurring during polymerizationâ€” isomerization followed by productive metathesis (I+M). In this case, the metathesis cycle maintains the same course but the produced intermediate D1 differs from the intermediate D created in the cycle that forms the non-isom erized products. As a result, instead of
48 release of ethylene during the metal carbene reaction with an olefin, the catalytic cycle releases propylene and produces a dimer 1I ( m/z 1395) with one methylene group less in the chain. As the cycle repeats, the corresponding trimer ( m/z 2074), tetramer ( m/z 2753), etc., are produced (â€œlight productsâ€). Si milarly, with the isomerized monomer A2, the catalytic cycle releases butylene and produces a dimer 2I ( m/z 1380) with two methylene groups less in the backbone. The third pathway (Figure 3-6â€”first recta ngle, blue), isomer ization followed by non-productive (trans metathesis) chemistry (I+NP), is initiated by the same isomerized versions of the monomer as the second pathway (I+M); in essence, these two pathways represent different consequences caused by the same olefin isomerization event. The difference arises from the alternate way the starting materials arrange to form a cyclobutane intermediate dur ing productive and non-productive metathesis. For example, when the isomerized monomer A1 and C react, they can form a cyclobutane intermediate in several ways. If they arrange to intermediate D1, productive metathesis follows isomerization and an isomerized dimer forms ( m/z 1395). However, if they arrange to form D2, instead of forming a dimer, the intermediate breaks in two to form structures A3 and C1, a process referred to as non-productive metathesis that results in no net reaction when non-isomerized molecules are reacting. This non-productive reaction is similar to interchange reactions which acco mpany all polycondensation chemistry, thus the term â€œtrans-metathesisâ€. Compared to the original monomer A, A3 has a methyl group in place of one of the hydrogens of the te rminal olefin, which corresponds to a mass increase of fourteen Daltons (methylene unit). If the new monomer-like structure A3 enters the productive catalytic cy cle with another molecule of C, then a dimer having
49 R MLn Ph R LnM Ph 9 9 9 9 R 9 9 MLn+ R 9 9LnM R 9 9 R 9 9 R 9 9 R 9 9LnM R 9 9LnM R 9 9 H2C CH2 R 9 8 MLn R R 9 9 8 9NI m/z 1409 n = 1 m/z 2088 n = 2 m/z 2767 n = 3 1I m/z 1395 n = 1 m/z 2074 n = 2 m/z 2753 n = 3Ph n( I + M ) R 9 7 R 9 9 MLn+ R 8 99 9 or or R R 9 9 7 9 ..... .... 2I m/z 1380 n = 1 m/z 2059 n = 2 m/z 2738 n = 3 The formation of the lower mass "light" products R n ( M ) ( I + NP ) R 9 8 R 9 9 MLn+ M Ln R R 9 9 8 9 R 9 9 R 9 8 MLn R 9 9 MLnLnM R 9 9 R 9 9 R R 9 9 9 9NP m/z 1423 n = 1 m/z 2101 n = 2 m/z 2780 n = 3H2CCH2 n The formation of the higher mass "heavy" products A BC D E D1 D2 A1 A2 A1 C C A3 C1n ( NP ) ( M ) C The repetition of the cycle leads to the formation of all main metathesis, nonisomerized products . . . . . . . . . . . . . . . . . . . ..... ....( M ). . . . . Figure 3-6. Some possible mech anistic pathways for the formation of the metathesis (NI), isomerization (I) and non-productive metathesis (NP) products
50 an additional methylene group on the chain forms ( m/z 1423). The repetition of the cycle forms the corresponding trimer ( m/z 2101), tetramer ( m/z 2780), etc., the so-called â€œheavy productsâ€. All products formed by the (I+NP) pathway are referred for short as nonproductive metathesis products NP, as they are in Figure 3-5. While Figure 3-6 shows rationalizations made about the possible mechanistic pathways leading to the formation of the products observed in the MALDI spectrum, it does not represent all the possibilities. For instance, apart from the examples given in Figure 3-6, numerous comb inations of different monomers, formed either by isomerization (I) or by isomerization followed by non-productive metathesis (I+NP), may be paired to produce dimers through metathesis (Figure 3-7a). Of course, some structures in Figure 3-7 are more probable to occur than others. The point, however, is that every monomer shown in the schematic (as well as countless others) may undergo metathesis with another monomer of either the same or a different kind. As a result, dimer chains with different numbers of methylene groups ar e generated which lead to the first cluster of peaks in the MALDI spectru m with masses ranging from m/z 1339 to m/z 1423 (Figure 3-7b). At the same time, each peak within every cluster in the spectrum represents a certain number of isomerization events that have occurred. Consequently, several possible molecular configurations could be assigned to a single mass. Note that even the peak that represents a non-isomerized structure NI (m/z 1409 in the case of dimer) also corresponds to several isomerized structures that ha ve the same mass (Figure 3-7c). Even though the probability of formation for every is omer is different, the information and the data available so far do not allow for elimin ation of any of the possible structures. Consequently, every possibility should be considered (Figure 3-7 b,c).
51 R 9 9 R 9 8 R 9 7 R 8 8 R 9 6 R 8 7 9,9 a R 9 5 R 9 9 R 9 9 R 9 9 9,8+C1 9,7+C2 9,6+C3 9,5+C4 C1+8,8+C1 C1+8,7+C2 9,9+C1 9,9+C2 9,9+C3 R 9 9 C1+9,9+C2 I+NP I R 9 9 9,9+C4 R 9 9 C1+9,9+C3 m / z 1 3 3 9 9,8-9,5 9,7-9,6 m / z 1 3 5 2 9,9-9,5 9,8-9,6 m / z 1 3 6 6 9,9-9,6 9,8-9,7 m / z 1 3 8 0 9,9-9,7 9,8-9,8 9,9+C1-9,6 m / z 1 3 9 5 9,9-9,8 9,7-9,9+C1 9,8-9,8+C1 m / z 1 4 0 9 9,9-9,9 9,8-9,9+C1 C1+9,8-9,8+C1 m / z 1 4 2 3 9,9-9,9+C1 9,8-9,9+C2 9,8+C1-9,9+C1.... .... .... .... .... .... .... .... ....a) b) R 9 8 R 9 9 9,9-9,9 b 9 9 R 9 9 R 9 8 R 9 8 R 9 8 R 9 8 9,8-9,9+C1 9,8-9,8+C2 C1+9,8-9,8+C1 m / z 1 4 0 9 : R ....c)continuescontinues Figure 3-7. Formation of the polymerâ€™s dimers a) Monomers available for metathesis b) Dimers formed during metathesis of these monomers, arranged by their mass c) Dimer structures corresponding to m/z 1409 In addition, Figure 3-6 and Figure 3-7 represent the possible pathways of oligomeric chains formation based on the a ssumption that isomerization occurs prior to
52 metathesis of the monomer. However, previous studies80;82 and the MALDI data presented here indicate that isomerization occu rs concurrently with metathesis. This idea is supported by the peak distribut ion pattern that emerges in every oligomeric cluster in the MALDI spectrum (Figure 3-5 and Figure 3-8). Just like in the gas chromatograms,82 the MALDI peaks in every cluster are not distributed evenly around the peak corre sponding to the non-i somerized product (NI); in other words, the peaks represen ting the non-isomerized products (NI) are not centered within the distributions. Had any such order been observed, the data would imply that metathesis occurs prior to isomerizati on. Instead, the non-isomerized peaks (NI) are shifted to the right of the center in each distributi on. Even though no attempts for quantitative analyses were made to determ ine the relative abundance of the different chains, in general, the MALDI data allow for following conclusions. First, regardless of the degree of polymerizati on, the amount of isomer ized chains produced (1I, 2I, 3I, 4I, etc.) dominates over the amount of non-isomerized (NI) products for each type of oligomer. Second, the amount of â€œheavy product sâ€ produced is relatively low. The information that MALDI gives about isomerization is similar to and yet different from the data received by gas chromatography.80;82 In both cases, a certain pattern of GC or MALDI peak s reveals the presence of isom erization; however, while every GC peak indicates the number of car bon atoms associated with the structure corresponding to the peak, every MALDI peak indicates the average mass of the structure corresponding to that peak. As a result, MA LDI provides valuable information about the possible structure of the oligom eric chains, and in particular , about the possible structure of the polymerâ€™s end groups, informa tion not accessible by gas chromatography.
53 Figure 3-8. A MALDI-TOF spectrum of polymer P1 (a) A zoom into the second three oligomeric peaks of the spectrum (b) A zoom into the third three oligomeric peaks of the spectrum. Two possible mechanisms explain isomeri zation occurring during transition metal catalyzed reactions.110 Both mechanisms suggest that isomerization proceeds through the formation of metal-hydride complexes of the metathesis catalyst. In the case of ruthenium catalyzed metathes is reactions, research indi cates that ruthenium-hydride species formed during metathesis serve as a ca talyst that initiates isomerization. The source and the formation of the hydride specie s are unclear. Some reports assume that isomerization might result dur ing purification of the fi nal metathesis product by distillation,111 others suggest that impurities in the metathesis catalyst lead to isomerization.112 A lot of data also implies that ruthenium-hydride species might form during heating of the ruthen ium metathesis catalyst.85;113-117 One such species was obtained during heating at 55 C of an Nheterocyclic-based ruthenium metathesis catalyst in benzene; the crystal structure conf irmed the formation of a hydride complex of
54 the catalyst that was proven to initiate metathesis isomerization.81 Since the ADMET polymer analyzed in our study was formed at about the same temperature with a similar catalyst, it is possible that analogous decompos ition species of the catalyst is responsible for the isomerization observed during ADMET conditions. Conclusions MALDI analysis of an amino-acid ADM ET polymer complements and confirms previous analytical data about olefin isomerization in metath esis chemistry. It confirms the presence of isomerization in metathesis polymerization when using complex 2 at elevated temperature and it complements previous GC data by giving the mass of the polymer chains rather than just the number of carbon atoms corresponding to each peak. Consequently, these mass spectrometry data increase the level of certainty when assessing the possible chemical structure of the polymer chains corresponding to each peak. Current and future work includes coupling GPC and MALDI to obtain more reliable data about the polymerâ€™s molecular we ight distribution and pol ydispersity index. Fourier Transform Ion Cyclotron Resonance (MALDI-FTICR) analysis is also being investigated due to its higher resolution and mass accuracy to obtain the exact mass of the oligomeric chains.
55Table 3-1. Theoretical and experimental aver age masses of the dimer (n = 1), trimer (n = 2) and the tetramer (n = 3) of polymer P1 Table 3-2. Theoretical and experimental aver age masses of isomerized at one site (1I) dimer (n = 1), trimer (n = 2) and tetrame r (n = 3) of polymer P1 n Average Mass Theoretical (Da) Average Mass Experimental (Da) Corresponding Ions Isomerized Product ( 1I ) 1 1394.98 1395 [C81H118S2O10N4+Na]+ 2 2073.96 2074 [C121H176S3O15N6+Na]+ 3 2752.94 2753 [C161H234S4O20N8+Na]+ 99 R n 9 8 R n Average Mass Theoretical (Da) Average Mass Experimental (Da) Corresponding Ions Non-isomerized Product ( NI ) 1 1409.01 1409 [C82H120S2O10N4+Na]+ 2 2087.99 2088 [C122H178S3O15N6+Na]+ 3 2766.97 2767 [C162H236S4O20N8+Na]+ 99 R n 9 9 R
56 CHAPTER 4 FUNCTIONALITY DEPENDENT OLEFIN ACTIVITY IN ACYCLIC DIENE METATHESIS POLYMERIZATION: A MASS SPECTROMETRY CHARACTERIZATION OF AMINO ACID FUNCTIONALIZED OLEFINS Reproduced with permission from V.I. Pe tkovska, T.E. Hopkins, D.H. Powell and K. B. Wagener. Copyright American Chemical Society. Introduction Olefin metathesis is a powerful, metal-cata lyzed reaction, used in organic synthesis for formation and transformation of C-C double bonds (Figure 4-1).118-120 R1 R1 R2 R2 R1 R2 R2 R1 LnM Figure 4-1. Olefin metathesis The olefin functionality (R1, R2) steric and electronic properties and catalyst selection, determine the olefin activity during metathesis. Certain sterically hindered terminal olefins, for example, do not homodi merize since the large functionality present does not allow binding.121;122 Apparently, olefin propertie s affect product selectivity in metathesis, a tendency that is essential in ol efin cross metathesis where olefin activity determines whether heterodimers or undesired homodimers are produced.123;124 The steric and electronic effects of functionality on olefin activ ity, and as a result on product selectivity, have been studied extensively;125-129 based on these results, Grubbs et al. created a general model for selectivity in olefin cross metathesis, where the olefins are divided into four groups base d on their relative ability to homodimerize and the ability of the homodimers to participate in additional metathesis reactions.130 As Type I are classified sterically unhindered, electron rich olefins that are most reactive in metathesis.
57 The gradient of reactivity d ecreases as the group number in creases with Type IV being least reactive, sterically hindered, electron deficient olefins. St yreneâ€™s reactivity, for example, could be modified by variati on in styreneâ€™s structure achieved through alternations of the a ttached functional groups. Even though styrene classifies as a Type I olefin, an insertion of bromine in the ort ho position decreases styreneâ€™s activity due to both the steric and electron-withdrawing effects of the bromine, which makes 2bromostyrene a Type II olefin. In addition, the catalyst chosen has also been used successfully to alter selectivity in cross me tathesis. Styreneâ€™s reactivity towards cross metathesis with -conjugated olefins increa ses when using catalyst 1, whereas cross metathesis with catalyst 2 is not existent (Figure 4-2).83;92;130;131 Ru Ph PCy3Cl N N Mes Mes Cl N Mo Ph RO RO 2 1 Figure 4-2. Olefin metathesis catalyst s: Mes = 2,4,6-Trim ethylphenyl, R = C(CF3)2(CH3) Olefin properties determine olefin activity by influencing the r eaction rate, and as a result, relate to metathesis kinetics. Term inal olefins, for example, display different metathesis rates depending on the steric and el ectronic properties of their functionalities; no linearity in the electronic e ffects of the functional groups ex ists but olefin steric bulk increase leads to a decrease in the metathesis rate.132 As a continuation of our previous work133 and previous studies on olefin metathesis activity, this investigation employs Matr ix Assisted Laser/Desorption Ionization (MALDI) mass spectrometry,10-12 complemented with Desorption/Ionization on Silicon (DIOS)65;134;135 data, to characterize amino acid functionalized Acyclic Diene Metathesis
58 (ADMET)75;76;136 polymers, formed during self-metat hesis of terminal dienes with Catalyst 1 (Figure 4-3).89-91;108 R R + n-1 catalyst n n-1 CH2CH2 Figure 4-3. The ADMET reaction Metathesis initiated at elevated temperat ures (~50 C) by cert ain ruthenium-based catalysts, including Catalyst 1, is characterized by the presence of isomerization side products in the reaction mixture.79-82;95;137 As a result of such structural isomerization (double bond migration) occurring during ADM ET polymerization, th ree major products are formed and detected in the MALDI spectrum of an amino acid functionalized ADMET polyolefinâ€”non-isomerized, main meta thesis products, â€œlight productsâ€ with masses lower than the mass of the main products, and â€œheavy productsâ€ with masses higher than the mass of the main products (Figure 4-4).133 The formation of each product follows the metathesis mechanism;109 the differences arise from the different starting materials entering the reaction. When th e reaction starts w ith a non-isomerized monomer, the formed dimer is the exp ected non-isomerized, main ADMET product (NI) (Figure 4-4a). However, if an isomerizat ion event takes place on the monomer, and the double bond moves one, two, three, etc. (x = 1, 2, 3) positions away from its original place, the metathesis reaction releases higher mass molecules than ethylene, like propylene, butylene, etc., depending on the numbe r of isomerization ev ents (x) that have taken place (Figure 4-4b). As the mass of the released small molecule increases by multiples of 14 Da, a methylene group, the mass of the produced dimer decreases accordingly, â€œlightâ€ products form (1I, 2I, 3I, etc., where the numbers indicate the number of isomerization events).
59 R n nLnM R n R n n R n nMLn Ph (a) H2CCH2 R n n-x R n n R n nNI LnM R n R n n R n nMLn Ph n n-x R n n R n-x n"light products" 1I, 2I, 3I, etc. (b) R n nLnM R n R n n R n nMLn Ph (c) H2CCH2 R n n R n n n H x H x H x"heavy products" NP1, NP2, NP3, etc. H x H H x xx = 1, 2, 3, etc. x = 1, 2, 3, etc. Figure 4-4. Formation of ADMET dimers Apart from low mass products, the ADM ET reaction also forms higher mass dimers, with masses higher than the mass of the corres ponding non-isomerized, main product by multiples of 14 Da (Figure 4-4c). If the isomerization event that takes place on the monomer (x = 1, 2, 3, etc.) is followed by non-productive metathesis-like processes (NP),133 a modified version of the monomer forms that has mass higher than the mass of the original starting material; again, the mass difference correlates to a different number of methylene groups. As a result, the produced dimer chains have masses higher than the mass of the main-metathesis, non-isomerized dimer by multiples of 14 Da, â€œheavyâ€ products form (NP1, NP2, NP3, etc.). Due to isomerization and non-productive metathesis accompanying ADMET, the MALD I spectrum of each oligomer (dimer, trimer, tetramer, etc.) is represented by a sequence of peaks co rresponding to main products (NI) and isomerized, â€œlightâ€( I) and â€œheavyâ€(NP), products distributed around NI.
60 Even though ADMET is homopolymerization, isomerization leads to formation of undesired products and the subject of product se lectivity emerges again. A set of five amino acid functionalized ADMET polymers w ith amino acid functional groups differing in bulk size and electronic properties, attached either as branches or incorporated in the polymer backbone, were chosen to detect possible functional group effect on olefin activity in ADMET polymerization by mass sp ectrometry. Previous studies on amino acid activity in olefin metathesis,138;139 including studies on the effect of amino acids inserted as a remote functionality to the reactive metathesis center,140 indicated the role of the amino acid groups in selective metathes is. Our interest in mass spectrometry characterization of the selected amino acid ADMET polyolefins, however, targets not only a mass spectrometry study of the functiona lity effect on product se lectivity but also structural characterization of the polymers due to their potential applications as biomaterials.105;106;141 This chapter presents MALDI-TOF analyses of the polyolefins, complemented with DIOS-TOF and MALDI-FTICR37-39 data. The peak distribution in every spectrum was used to evaluate the olefin activity towards metathesis and isomerization, and in turn, se lectivity towards â€œlightâ€, â€œh eavyâ€, and non-isomerized, main metathesis products. In a following discus sion, speculations were made about the possible reasons causing pref erential formation of e ither of the products. Experimental Section Meterials MALDI matrices Dithra nol (1,8-dihydroxy-9[10 H ]-antracenone) and -CHCA ( cyano-4-hydroxycinnamic acid) were purchased from Sigma-Aldrich (St. Louis, MO). Trifloroacetic acid (97% TFA) , certified sodium iodide, and HPLC grade tetrahydrofuran (THF) were obtained from Fisher Scientific (Fair Lawn, NJ). Poly(ethylene glycol) PEG
61 standards with number-average molecular weights ( Mn) of 600, 1500, 2000, 3400, 4600, and 8000 were purchased from Sigma-Aldr ich (Milwaukee, WI). The synthetic procedure for the synthesis of the ADM ET polyolefins is previously published.89-91;108 For DIOS analysis, an n-type silicon wafer with resistivity 1-10 cm and 525 microns thickness was purchased from Silic on Sense (Nashua, NH). Ethanol (200 proof) was purchased from Aaper Alcohol and Chem ical Co. (Shelbyville, KY). Hydrofluoric acid (48 to 50%) and HPLC grade chloroform was obtained from Fisher Scientific (Fair Lawn, NJ). MALDI Sample Preparation The main MALDI sample preparation t echnique applied for analysis of the polymers uses DTH as a matrix, NaI as an ioni zation salt, and THF as a solvent. Matrix, polymer, and ionization salt were dissolved in THF to achieve concentrations of 4 mg/mL DTH solution, 1 mg/mL polymer solution, and 1 mg/mL solution of NaI. Aliquots of the three solutions were mixed in appropriat e volumes in matrix:polymer:ionization salt ratios of 7:1:1, 5:1:1, and 1:1:1. On e microliter of each mixture was spotted on a MALDI plate. For data comparison, MALDI spectra of one of the polymers were obtained in CHCA matrix. A saturated solution of the matrix and a 1 mg/mL solution of the polymer were prepared in 0.1% TFA solution in THF. Appropriate volumes of the two solutions were mixed in ratios of matrix:polymer 10:1, 7:1, and 1:1; one microliter of each mixture was spotted on a MALDI plate. For external mass calibtation, PEG standa rds were prepared for MALDI analyses either in DTH or in -CHCA matrix, depending on the matrix used for analyte detection.
62 In ether case, PEGs were dissolved in TH F at the following concentrations: 1 L/mL PEG 600, 1 mg/mL PEG 1500, 1 mg/mL PEG 2000, 1 mg/mL PEG 3400, 2 mg/mL PEG 4600, and 3 mg/mL PEG 8000. When DTH was us ed as a matrix, aliquots of the PEGs solutions in THF were mixed with appropriate volumes of 4 mg/mL DTH/THF solution and 1 mg/mL NaI/THF solution to achieve a ma trix: PEG: NaI ratio of 10:1:1 in the mixture. The sample preparation with -CHCA used saturated solu tion of the matrix and a matrix:PEG:NaI ratio of 1:1:1 in the final mi xture. In either case, 1L of the mixture was applied on the stainless steel MALDI plat e and the solvent allowed to evaporate. DIOS Etching Conditions and Sample Preparation Porous silicon was produced by anodic etching.64-66;142 The silicon wafer was cut into small pieces (2 cm 2 cm) and sealed in a Teflon cell with a gold electrode lying on the bottom of the cell and a platinum electrode (cathode) placed into a 1:1 (v/v) HF/Ethanol etching solution filling in the ce ll cavity. An O-ring positioned between the wafer and the top part of the cell prevents the etching solu tion from leaking outside the cell. The wafer is irradiated with a white light that is focused through a series of lenses to pass through a patterned mask so that the imag e of the mask is transferred onto the silicon wafer. Upon completion of the etching proce ss by inducing a 30 s constant current of 4 mA/cm2 between the electrodes, a sequence of nu mbered spots is produced as an image of the mask (a few hundred micrometers in diameter) with nanometer sized pores. For DIOS analyses, the analyzed ami no acid ADMET polymer was dissolved in 0.1% TFA acid chloroform solution at a concentration 10-2 M and 1 L of the solution was spotted on an etched silicon spot.
63 The calibration was done with PEG standard s dissolved in deionized water at 10-2 M concentration and spotted as 1 L aliquots on an etched spot. Mass Spectrometry Analyses Bruker Reflex II TOF (Ballerica, MA) in strument was used for both MALDI and DIOS time-of-flight analyses. The instrument is equipped with a 337 nm nitrogen laser and delayed extraction. Positive ions were detected in linear mode with an acceleration voltage of 20 kV and 100 shots acquired for every spot. No signal in reflectron mode was obtained. Bruker Xmass software was used for data processing and theoretical calculations. Bruker Apex II FTICR instrument providing 4.7 T magnetic field was used for high mass accuracy MALDI analyses. Ions were inje cted into the infinity cell by applying broadband excitation and detection, 0.01 s hexapole ion accumulation, and 1.14 V trapping potential using the Dinamic trapping me thod with a time domain data sets of 512 K. All ions detected, regardle ss of sample preparation t echnique or analyses method applied, were sodiated. Results and Discussion Figure 4-6 represents the MALDI-TOF spect ra of the analyzed polymers shown in Figure 4-5. As expected, the spectra confirm that every oligomer (dimer (n = 1), trimer (n = 2), tetramer (n = 3), etc.) consist of a sequence of peaks rather than a single mass representing the oligomeric chains, a result associated with the fo rmation of three major products during ADMET polymerizationâ€”main metathesis, non-isomerized oligomeric chains (NI), and â€œlightâ€ and â€œheavyâ€ products resulting from isomerization.
64 Polymer 1 (P1) Polymer 2 (P2) Polymer 3 (P3) Polymer 4 (P4) Polymer 5 (P5) HN O O NH O O HN O O 9 9 n NH O HN O O 3 3 n O N H O 8 8 n O HN O O 9 n 9 O HN O NH O O 9 9 n Figure 4-5. Amino acid f unctionalized ADMET polymers To assess product selectivity in ADMET by MALDI-TOF MS, the relative ratio of â€œlightâ€ (I) and â€œheavyâ€ (NP) versus non-isomerized (NI) products has to be determined for every oligomeric cluster. The peak assignments of the NI products for each oligomer reveal the product distribution pattern of â€œlightâ€ and â€œheavyâ€ chains around NI. Table 4-1 shows the theoretical average masses for the dimers corresponding to main metathesis, non-isomerized NI products, calculated based on the e xpected repeat unit and end groups produced by ADMET.
65 Figure 4-6. MALDI-TOF full spectra of th e analyzed amino acid functionalized ADMET polymers
66 Table 4-1. Theoretical average masses of the polymersâ€™ dimers Polymer Corresponding Dimer Ion (n = 1) Theoretical Average Mass (Da)a P1 [C82H134N6O12Na]+ 1419.00 P2 [C36H64N4O6Na]+ 671.92 P3 [C54H98N2O6Na]+ 894.38 P4 [C66H106N2O6Na]+ 1046.58 P5 [C60H108N4O8Na]+ 1036.54 A zoom into the dimer MALDI-TOF spectru m of P1 (Figure 4-7a), for example, shows that experimental match to the theoretical NI average mass of 1419.00 Da is present in the spectrum. In addition, higher reso lution (~ 50,000) MALDI-FTICR analysis confirmed the presence of monoisotopic NI dimer mass of 1417.99 Da (Figure 47b) where the theoretical P1 monoisotipic dimer mass is 1417.9957 Da. Higher and Figure 4-7. Comparison of TOF and DIOS spect ra of Polymer P1. (a) A zoom into the MALDI-TOF spectrum of P1 dimer (sample prepared in DTH/NaI) (b) Dimer chains of P1 detected by MALDI-F TICR (sample prepared in DTH/NaI)
67 lower mass products, differing from the NI peak by multiples of 14 Da, are also detected in agreement with the expected isomerizati on events occurring al ong with metathesis. The peak distribution in Figur e 4-7a shows that for P1 di mer, the relative amount of formed â€œlightâ€ products dominates ov er the amount of â€œheavyâ€ ones. To further explore the validity of the peak distribution pattern obtained for the P1 dimer, the polymer was analyzed by MA LDI-TOF using an alternative sample preparation to the DTH matrix and NaI salt; instead, -CHCA served as a matrix and 0.1 % TFA acid as an ionization agent. Comparison of the trimer (n = 2) and tetramer (n = 3) distribution pattern displayed by â€œl ightâ€ (red), â€œheavyâ€ (blue), and NI products in both spectra shows very similar distribution trends (Figure 4-8a and 4-8b) detected by MALDI regardless of the applied sample preparation. Figure 4-8. A zoom into the trimer (n = 2) and tetramer (n = 3) MALDI-TOF spectra of polymer P1 obtained in two sample pr eparation techniques (a) matrix: DTH; ionization salt: NaI (b) matrix: -CHCA; ionization agent: 0.1% TFA acid. Determination of the NI dimer mass by both TOF and FTICR (higher mass NI peaks are obtained by adding the polymerâ€™s repeat unit mass to the NI dimer mass), and detection of similar product distribution pa ttern by two different sample preparation
68 techniques allowed for confidence in drawing the conclusion that ADMET polymerization of polymer P1 with catalyst 1 at 50 C leads to dominant formation of â€œlightâ€ products over â€œheavyâ€ and NI. The detailed analyses of P1 ensured th at either of the c hosen MALDI sample preparation techniques produces spectra whic h give accurate MALDI representation of the actual distribution of polymer chains, and as a result, the rest of the polymers were analyzed only in DTH/NaI. Zooms into the oligomeric MALDI peaks of polymers P2 and P3 reveal similar product distribution to the one observed for polymer P1 (Figure 49). The match between the experimental and the theoretically calculated average masses for P2 and P3 additionally supports the accuracy of the obtained data (Table 4-1 gives a mass of 894.38 Da for P3 dimer, whic h matches the experimental mass of 895 Da observed in the detected in linear mode (0.1% accuracy) polymer MALDI spectrum). Figure 4-9. Zooms into the MALDI-TOF oli gomeric peaks of (a) Polymer P2â€”trimer and tetramer (b) Polymer P3â€”dimer, trimer, and tetramer. The matrix peak interference in the lo w mass region of the MALDI mass spectrum of polymer P2 (Figure 4-6c) does not allow for detection of the NI dimer mass, and
69 consequently, for evaluation of the peak dist ribution of the P2 dimer (n = 1). As an alternative to MALDI, DIOS-TOF was used to analyze the low mass region of this polymer, for this is a matrix-free ionization technique known for its ab ility to give little background signal in contrast to MALDI. Polymer analysis by DIOS allowed for isolation of the dimer chains of P2 (Figure 4-10) even though a certain extent of DIOS low mass background signa l still interferes with analyte detection (background signal possibl y due to impurities in the silicon wafer, fragmentation of the analyte, etc.). The obt ained DIOS data allowed for detection of the NI dimer peak for P2 of 671 Da, informa tion not accessible by MALDI. Consequently, the assignment of â€œlightâ€ and â€œheavyâ€ peak s around m/z 671 in DIOS revealed a dimer product distribution that matches the dist ribution trend observed for the higher mass oligomers of P2 by MALDI (Figure 4-9a), with â€œlightâ€ chains dominating the spectrum. Figure 4-10. DIOS-TOF spectra of Polymer P2. (a) DIOS-TOF of Polymer P2 (b) A zoom into the dimer cluster of the DIOS-TOF spectrum By complementing each other, the MALD I and DIOS data of P2 allowed for evaluation of the oligomeric P2 product dist ribution throughout the detectable mass scale for this polymer. In DIOS, higher mass oligom eric peaks (n > 1) were not detected; only part of the â€œlightâ€ trimer chains appeared in the spectrum. The upper mass detectable limit for polymer DIOS characte rization, however, is higher than what was detected for
70 P2. Analyses of poly(ethylene glycol) st andards (PEG) with varying number average molecular weight determined that under opt imized conditions (silicon wafer etched for 30s, 4 mA/cm2 applied current between the Pt and Au electrodes) a number average molecular weight of up to 3, 400 Da is readily detected. It was shown in the previous chapters that with an increase of the analyte molecular weight the background signal in the DIOS lo w mass region increases probably due to increased analyte fragmentation, which could also explain the observed background DIOS signal for polymer P2. In contrast to the PEG DIOS spectra, however, the DIOS detection of P2 is limited to the lower mass chains possibly due to solubility issues. While the PEG standards were dissolved in deionized water for DIOS analyses, the hydrophobic polymer P2 dissolves only in orga nic solvents like chloroform, which was used for DIOS sample preparation. Deposition of hydrophobic solvents on the hydrophobic porous silicon surface is challenging due to the resulting small contact angle, which in turn, does not allow for application of the analyte on a small DIOS s pot at a concentration meeting the detection limits of DIOS-TOF. The above presented mass spectrometry data of the three amino acid functionalized ADMET polymers display similar product distri bution pattern for each polymer oligomer with â€œlightâ€ products dominati ng over the â€œheavyâ€ ones. Howe ver, the last two of the five polymers exhibit different oligomeric peak arrangement. For polymers P4 and P5, the peaks corresponding to ma in metathesis products NI appear in the middle of the oligomeric clusters with â€œlightâ€ and â€œhea vyâ€ products equally distributed around them (Figure 4-11).
71 Figure 4-11. Zooms into the MALDI-TOF spect ra of oligomeric peaks of (a) Polymer P4â€”trimer and tetramer (b) Polymer P5â€”trimer and tetramer The trimer and tetramer experimental NI masses of polymers P4 and P5 correspond to the theoretically calculated masses of these oligomers (dimer NI mass (Table 4-1) + repeat unit mass = trimer NI mass, etc.). For polymer P4, the MALDI spectrum detects fifteen peaks corresponding to â€œlightâ€ products and fifteen p eaks related to the formed â€œheavyâ€ products, whereas for polymer P5 , the number of analogous products is seventeen. This information indicates th at during ADMET of polymers P4 and P5, an equal number of isomerization events leads to formation of both â€œlightâ€ and â€œheavyâ€ products, a result that could be associated with the olefinsâ€™ me tathesis activity and selectivity. To summarize the results, the analyzed polymers were divided into two groups depending on the oligomeric product distribution their mass spectra display (Figure 412). In the first Group A (Figure 4-12a), polym ers P1, P2, P3, and P6 (polymer P6 spectra reported in the previous chapter) mass spectra display dominant formation of â€œlightâ€ chains over â€œheavyâ€ and NI products for each polymer oligomer. Polymers P4 and
72 P5 form the second group, Group B, where the amount of formed â€œlightâ€ and â€œheavyâ€ products is similar (Figure 4-12b). Figure 4-12. Bar diagrams representing the two trends of oligomeric product distribution calculated as a ratio of the sum of p eak intensities of â€œlightâ€ or â€œheavyâ€ products to NI products corresponding to each oligomer (a) Bar diagram representing the oligomeric product di stribution observed for Polymers P1, P2, P3, and P6 based on peak intensities of three spectra of polymer P1 acquired in -CHCA (b) Bar diagram repres enting the oligomeric product distribution observed for Polymers P4 and P5 based on peak intensities of three spectra of Polymer P4 acquired in DTH The goal of our study was not a detailed research on the possible reasons causing the different olefin activity the polymers disp lay but rather structural characterization and mass spectrometry detection of possible produc t selectivity. Howe ver, a look at the literature published to date on olefin product sele ctivity in metathesis allows for speculations to be made about the causes of the diverse olefin activity displayed by the amino acid functionalized ADMET polymers.
73 According to the Grubbsâ€™ ge neral model for selectivity,130 increasing of the olefinâ€™s steric bulk and electron deficien cy decreases its reactivity in metathesis. The different olefin metathesis activity shown by the substr ates in Figure 4-12 coul d be attributed to the different bulk size propert ies their functional groups posse ss. Out of the six polymers presented in Figure 4-12 (and their similar m onomer units), polymers P1 and P6 exhibit large steric bulk functional groups present, whereas polymers P2, P4, and P5 have relatively smaller sized functionalities. The di fference in the latter set of polymers relates to the P2 polymer higher functional group popul ation density compared to P4 and P5, which in turn, should lead to a stronger functional group influence on the P2 olefin activity. For polymer P3, however, the displayed activity relates more to the electronic properties of the olefin rath er than to steric bulk. The olefinâ€™s activity, determ ined by steric or electroni c functional group properties, affects metathesis kinetics, which in our case would determine the ratio of the rate of metathesis versus the rate of isomerizati on and establish the fina l distribution of ADMET products. Previously performed kinetics st udies on olefin metathesis isomerization examined geometric (cis/trans) isomeriza tion and suggested a correlation between the metathesis and the isomerization relative ra tes and the manner of olefin coordination to the catalyst.143-145 Kinetics studies of olefin structural (double bond migration) isomerization in ADMET polymerization, on th e other hand, led to the conclusion that metathesis and isomerization occur concurrently while competing in the initial stages of the polymerization, but after metathesis finishes, only isomerization takes place.80;82;146 Metathesis of the model co mpound 1-octene with catalyst 1 confirmed that the described kinetic pathway is preferred in ADMET polymer ization since it led to the formation of
74 7-tetradecene as a non-is omerized main product (NI) and revealed an asymmetric distribution of isomerized chains around NI in the gas chromatograms (GC).82 This kinetic outcome, where the distribution is shif ted towards the â€œlightâ€ products, with few â€œheavyâ€ chains formed, assures that isomeriza tion occurs at the same time as metathesis on both the monomer and the polymerization pro ducts. If an alternative kinetic pathway took place, where metathesis occurred before isomerization, the product distribution of isomerized chains around NI would be symmetricâ€”in this case, only the non-isomerized symmetric olefin 7-tetradecene would form in the initial stages of the reaction without presence of any isomerized monomers or other products. If post metathesis, isomerization occured on 7-tetradecene, the displacement of the double bond from its original internal position would create an asymmetric olefin with different number of carbon atoms on the two sides of the double bond. As a result, in a subsequent metathesis reaction, the 1-octane monomer would have an equal opportunity to dimerize with either the lower or th e higher carbon number side chai n of the asymmetric dimer creating an equal number of â€œlightâ€ or â€œhea vyâ€ isomerized chains and a symmetric GC product distribution. Mass spectrometry analyses of the am ino acid functionalized ADMET polymers, however, detect product distribut ions corresponding to both ki netic situations described above, i.e. for polymers of Group A isomer ization and metathesis seem to occur concurrently complementing the previ ous reports on metathesis structural isomerization,82 but for polymers of Group B, the ma ss spectrometry data indicate that metathesis probably occurs prior to isomeriza tion. As a result, the questions arise of
75 whether the latter kinetic situation is possible and what happens during ADMET polymerization to cause the forma tion of both product distributions. We speculate that even when symmetr ic product distributi on forms during ADMET (Group B), metathesis and isomerization st ill occur concurrently but the rate of metathesis is higher than the rate of isomerization, and as a result, no significant and noticeable isomerization occurs on the monomer. In this case, formation of dimer chains, for example, would probably follow the sequen ce of metathesis (M) and isomerization (I) events shown in Figure 4-13. Due to the fast metathesis rate, in the initia l stages of polymer ization, only non isomerized, main metathesis dimer oligomers (NI) form. A following isomerization event on the NI product would create an asymmetric isomerized dimer with different number of carbon atoms on both sides of the internal double bond. When secondary metathesis (also called non productive (NP) or â€œtransâ€ metathesis sin ce it does not form higher mass oligomeric chains) occurs on the displaced internal double bond, the reaction will produce an equal amount of lighter and h eavier mass products th an the mass of the original NI dimer, and thus, will create a symmetric distribution of isomerized products around the NI oligomers in the mass spectra. Even though only the â€œheavyâ€ oligomeric chains in Figure 4-11 are labeled as non productive (NP) metathesis products, the suggested kinetic situation for polymers P4 and P5 would produce both â€œlightâ€ and â€œheavyâ€ products by non productive metathesis events.
76 R R R R R 10 9 9 9 9 9 9 9 8 9 R LnM 9 9 R R 9 8 9 9 R R 9 9 10 9 "light" "heavy" M I M (NP) "NI" Figure 4-13. A schematic representation of a possible sequence of metathesis (M) and isomerization (I) events leading to sy mmetrical product distribution in the mass spectra For the polymers in Group A, however, more probable kinetic sequence would be the one described in Figure 4-4, where isom erization affects the monomer due to slower metathesis rate by producing a variety of modified monomer versions that enter metathesis and non productive metathesis to create a distribution shifted towards the isomerized â€œlightâ€ products. The difference in the metathesis rate displayed by the different olefin substrates is associated with the mechanism of metathesis and the different coordi nation of the olefin double bond to the metathesis catalyst. Metathes is of terminal olefins, differing in steric bulk and electronic properties , with a ruthenium metathesis catalyst (first generation Grubbs) showed that increase in the steric bulk leads to a decrease of the metathesis rate.132 The authors associate th is trend with the preferre d formation of methylidene rather than alkylidene complex from the cataly st during the reaction wi th bulkier olefins,
77 even though the study determines other fact ors to also influence the kinetics; the methylidene complex is known to promote slower metathesis than the more electron rich alkyledene. In our case, the olefins classified as Group A ar e believed to display slower metathesis rate than the ones in Group B due to either steric bulk (Polymers P1, P6) or higher population density of the functional group (Polymer P2) that eventually creates the effect of steric bulk. Group B, on the othe r hand, includes sterical ly less hindered olefin substrates, which follow faster metathesis. The difference in the metathesis rates leads to the different product selectivity displa yed by the two groups of olefins. Metathesis activity of olefins that have remote amino acid functionality incorporated in the backbone, like po lymer P3, has been studied before.140 The data indicate that homoallyl amides with no aromatic side chains do not undergo self metathesis due to possible coordination of oxygen functionality to the active catalyst center and formation of non productive catalys t species. For other olefins, similar coordination to the catalyst was found to extr emely decrease the rate of the metathesis reaction.147 Based on these findings, we speculate th at the slow metathesis activity of the P3 monomers and oligomeric ch ains is probably associated w ith the formation of similar modified catalyst species that affects the metathesis rate and leads to dominant formation of isomerized â€œlightâ€ products. Conclusions Mass spectrometry charact erization of the amino acid functionalized ADMET polymers provided details a bout the structural design and the uniformity of the oligomeric composition, valuable inform ation needed for assessing the polymer properties and potential bioappli cations. In addition, this study detected dependence of the product selectivity on the olefin functional group pr operties. Two possible product
78 distributions caused by different olefin ac tivities were related, based on previous literature reports, to different kinetics of the reaction due to different coordination of the olefins to the catalyst center.
79 CHAPTER 5 MASS SPECTROMETRY ANALYSES OF ETHYLENE OXIDE/PROPYLENE OXIDE COPOLYMERS Introduction Ethylene oxide/propylene oxide (EO/PO) copolymers are surface active agents used as detergents, foams, lubricants, and drug-delivery systems. O O x y The ethylene oxide part of the copolymer has hydrophilic properties whereas the propylene oxide part is hydrophobic. The ratio of the two monomers determines whether hydrophilic or hydrophobic properties will dominat e in the copolymer. Moreover, small changes in the chemical structure of the copolym ers lead to variations in their properties, which results in a variety of applications. Consequently, complete characterization of the copolymers that includes determination of the relative abundance of each monomer unit, identification of the end groups, establishmen t of the sequence of the EO and the PO units along the polymer backbone, and determ ination of the molecular weight and polydispersity is important. The traditional analytical techniques used to study EO/PO-type copolymers like gel permeation chromatography (GPC), nuclear ma gnetic resonance (NMR), and infrared spectroscopy (IR), for example, provide info rmation about the average molecular weight and the average moles of EO and PO units present in the sample; functional groups and other structural information could also be determined. However, fine structural characterization of complex polyol mixtures is limited by thes e techniques. In contrast,
80 mass spectrometry is an excellent analyti cal tool for exact mass measurements and sequencing of the polymersâ€™ repe ating units since it allows mass determination, isolation, and subsequent characterization by tande m mass spectrometry of the individual oligomeric chains. In general, MALDI-TOF mass spectrometry is used for analyses of simple homopolymers, where high mass resolv ing power is not requ ired for complete characterization. For analyses of copolyme rs, Fourier transform i on cyclotron resonance (FTICR) mass analyzer should provide the nece ssary resolution and mass accuracy. This type of analyzer gives a mass resolving power of approximately 50,000 at molecular mass 5,000 and a mass accuracy of 5 ppm. In comparison, a TOF analyzer gives a resolving power of less than 1,000 in linear mode at the sa me molecular weight and a mass accuracy of about 0.1%. As discussed previously in Chapter 1, the very high resolution and mass accuracy provided by the FTICR mass analyzer are relate d to the measured cyclotron frequency, which is independent of the kinetic ener gy of the ions. Kinetic energy and space distributions are main causes for the low re solving power displaye d by the TOF analyzers even though improvements such as dela yed extraction and reflectron detection compensate for these defects to a certain ex tent. To illustrate the powerful resolving power of the FTICR analyzer, Marshall et al . give the following example.148 At room temperature, a singly charged i on with mass of 100 Da that is put in a magnetic field of 3 Tesla will have an ICR radius of 0.08 mm. If the same ion is excited to a radius of 1 cm in the ICR infinity cell (see Chapter 1), it will have a velocity of 2.974 m/s and a kinetic energy of 434 eV. For a time period of 1 s this ion will travel a distance of 30 km.
81 Compared to the distance that the ions travel in a linear TOF analyzer (about 1.5 m), a distance of 30 km will allow the ion current to pass by the detection plates and be detected numerous times, which is the ba sis of the high resolving power and high accuracy provided by the FTICR analyzer. Also, in contrast to the TOF instrume nts, where the detection by the electron multiplier is mass dependent, the FTICR anal yzers detect image current, which is not mass dependent.149 Another important advantage of the FTICR analyzers is their capability to perform tandem mass spectrometry (MSn). In MS/MS, a precursor ion with a certain mass is selected, isolated by ejecting the other ions from the cell, and fragmented into pieces that display a secondary characteris tic mass spectrum of the precurs or ion. One of the most common methods used for ion selection in FTICR is stored-wavef orm inverse Fourier transform (SWIFT) excitation.41;150-152 In SWIFT, an excitation waveform is constructed based on an acquired spectrum. Some of the ions in the spectrum are excited to a given radius whereas others are not. The ions that ne ed to be excited are selected in the mass to charge domain of the spectrum and their ranges are subsequently transformed back to the frequency domain. The obtained frequency spec trum is inverse Fourier-transformed into the time domain, which in turn, is represented by a waveform that is applied to the excitation plates of the ICR ce ll via a digital-to-analog conver ter. Once the precursor ion is isolated, a couple of options exist for disso ciation. The most used are collision-induced dissociation (CID) and electro n-capture dissociation (ECD).150 In CID, the isolated ion is excited to a larger cyclotron radius in the pr esence of a natural gas. The increased ion radius and kinetic energy lead to a reduced path length and increased ion velocity, which
82 allows for collisions with the natural gas molecules to occur. Upon collisions, energy transfer occurs to both collision partners. Ev entually, the transferred energy is converted to internal energy of the ion. If the in ternal energy exceeds the required dissociation energy, the ion breaks into a character istic pattern producing a mass spectrum. Several techniques are used to increase the ionâ€™s cy clotron radius and kinetic energy. Two of the widely used methods are resonant excitation and sustained offresonance ion (SORI) excitation. In resonant excitation, the ionâ€™s kinetic energy and cyclotron radius are increased by applying the exact cyclotr on frequency of the ion on the excitation plates, which results in excess of internal energy imposed on the ion and significant fragmentation. In SORI, however, the collisions with the neutral gas do not lead to a big increase of the internal en ergy of the precursor ion since the excitation frequency applied is slightly offset from th e ion cyclotron frequency. As a result, the ionâ€™s radius and kineti c energy increase and decrease in an alternate way which prevents accumulation of excess energy and excess fragme ntation. Apart from SORI, low-energy excitation of the precursor ion could be achieved by infrared multiphoton dissociation (IRMPD). In IRMPD, the ion is excited to a hi gher radius by irradiation with an IR laser passing through the center of th e ICR cell. An advantage of IRMPD is that higher pressure introduced by neutral gas pulses is not necessary. This allows for fast detection since the absence of neutral ga s eliminates the pumpdown time. As an alternative to CID, electron capture dissociation is also used for dissociation of the isolated precursor ions. In ECD, however, the isolated ions should be multiply charged cations that are subjected to a lo w energy beam of electrons. The main difference between ECD and the other technique s, such as CID and IRMPD, is that ECD
83 does not increase the internal energy of the ions in small steps. Instead, during an electron-capture, ener gy of about 5-7 eV is releas ed, which leads to extensive fragmentation and formation of an odd-electron molecular ion [M+nH](n-1)+. Along with the advantages that the FTI CR mass analyzer offers in terms of resolution, mass accuracy, detector response, and MS/MS capabilities, the ICR detection suffers from gradually decreasing resolution wi th increasing mass to charge ratio, a trend observed with the TOF analyzer as well. In FTICR, the resolution is related to the mass of the ions by the following two equations:149 m qB m m 2 m zB vc2 where B is the magnetic field strength, m/z is the mass to charge ratio, is the time constant for signal decay, m/ m is the resolution described as full width ad half-height, c is the cyclotron frequency of the ion. It is obvious from the firs t equation, that resolution is inversely proportional to mass, whereas the second equation indicat es that the frequency of the signal c decreases as the mass increases. The frequency equation shows that higher mass ions will have lower frequency due to their higher radius of oscillation (see the next equation) which results in lower frequency of detection and decreased resolution and mass accuracy.148 qB mv rxy where r is the radius of i on oscillation in meters, and xy is the ionâ€™s velocity in m/s.
84 High mass compounds, however, could be de tected with high resolution and accuracy by electrospray ionization (ESI). In ESI, multiple charging occurs, which brings the high masses down to the low mass to charge region where the ICR instruments exhibit their best resolving power. The multip le charging received in ESI also benefits the potential for tandem mass spectrometry since th e higher the ion charge, the easier it is to fragment it in MS/MS; singly charged ions with mass higher than 3 kDa are not easy to dissociate.153 ESI-FTICR has been mainly used for characterization of biomolecules such as proteins and peptides. In 1995, Oâ€™Conner and McLafferty used ESI-FTICR for analyses of poly(ethylene glycol)s, PEGs, with molecular weights of 4.3, 13, and 23 kDa.154 For PEG 23 000, the authors isolated 47 oli gomers having about 5000 isotopic peaks altogether in 10 charge st ates. Apart from hydrophilic analytes like PEG, hydrophobic polymers like poly(dimethylsiloxane)s, PD MS, have also been successfully characterized by ESI-FRICR.155 Despite its advantages, however, ESI of polymers produces very complex mass spectra due to the multiple charging states produced for a single oligomer. Consequently, the data pr ocessing and structure elucidation of ESI polymer spectra are time and labor consuming. As an alternative to ESI, MALDI is a nother â€œsoftâ€ ionization technique that produces singly charged ions with very little fragmentation. As a result, in contrast to ESI, coupling of MALDI to the FTICR anal yzer should produce high resolution and high accuracy polymer mass spectra that are easy to interpret. The first MALDI-FTICR experiments were reported by Hettich and Buchman in 1991.156;157 They analyzed small peptides and oligonucleotides in negative mode. The
85 first polymer MALDI-FTICR characterizati on, however, was reported by Castro et al. for PEG 10 000.158 MALDI-FTICR experiments are performed e ither in the MALDI source inserted directly into the ICR cell (the so-called â€œin-ce llâ€ or internal MALDI) or in a source that is placed at a certain distance from the ICR cell (external source). The Bruker instrument used in the presented research has an extern al source whose instrumental characteristics are described in detail.159 External MALDI sources ar e preferred in high throughput analyses since they operate at higher pressures and allow for fast and easy sample change. The Bruker FTICR instrument has th e following elements aligned in a sequence: a MALDI target plate where the samples are deposited, a hexapole ion guide situated 1 mm away from the plate where the ions can be stored, an extraction plate placed at the rear end of the hexapole, ion opt ics that transfer the ions to the ICR cell, and finally, an ICR cell situated within the magnet. In a ddition, a device for pulsi ng collision gas and a laser are situated on both sides, in close proximity, to the MALDI target. The ions form upon irradiation with the laser of the sample target and ar e trapped in the hexapole ion guide. Ion trapping is achieved through applying a +20 V poten tial to both the target the extraction plate. While in the hexapole ion gu ide, the ions are cooled by an inert collision gas, usually argon. The cooling is necessa ry due to the high ion kinetic energy (about several electron volts) generated during laser irradiation, whic h is associated with space and velocity distributions that make the pr oper transfer, capture, a nd detection of ions difficult. Once the ions are cooled while trappe d in the ion guide, they are transferred to the ICR cell by applying about -2 to -5 V to the extraction plate. According to the experience obtained in Bruker,159 MALDI-FTICR spectra of compounds with mass
86 below 3000 Da could be obtaine d without storing and cooli ng in the hexapole guide. However, for higher mass molecules, pulsed gas should always be used. Along with storage purposes, the ion guide also serv es as an accumulation region where ions produced during several laser s hots could be accumulated for an increased signal to noise ratio. As an alternative to the ion guide, the ICR cell is also used for ion cooling by introduction of a pulsed gas during trapping, which is known as a gas assisted dynamic trapping. In-cell ionization with an in ternal MALDI source for analyses of bovine insulin and gramicidin D was employed by Knobeler and Wanczek.160 The instrumental arrangement employs a MALDI target plate that is directly placed at one of the trapping electrodes of a cylindrical ICR cell. On the other hand, the laser beam enters the ICR cell through a fused silica window positioned within the opposite trapping electrode. The in-cell trapping method displays high sensitivity due to the large amount of trapped ions (less than 1 fmol sample was consumed per experi ment during analyses of gramicidin D after 10 laser shots were acquired). Another in-cell MALDI-FTICR an alysis was described by Mize and Amster.161 In their instrumental setup, the MALDI targ et was placed ~1 cm away from the front trapping plate. The internal ionization was used to develop a method of ion accumulation that is not based on signal av eraging but rather on a repeat accumulation of ions from different laser shots in the ICR cell and a subs equent excitation and de tection of all ions together. A comparison between signal aver aging of 50 laser shots during analyses of bovine serum albumin (BSA) and a single measurement of the BSA signal after 50
87 accumulations showed a dramatic difference in resolution and signal to noise in favor of the accumulation method. MALDI-FTICR with an external source wa s used for determination of unknown polymer end groups.162;163 By using a regression method, the mass of the end groups is determined by correlation of the measured oligomeric mass with the degree of polymerization according to the following equation: cation groups end unit rep measM M M n M . . where Mmeas is the measured oligomeric mass, n is the degree of polymerization, Mrep.unit is the mass of the repeat unit, Mend.groups is the mass of the end groups, and Mcation is the mass of the cation used for ionization. Linear regression of the meas ured mass of isotopically resolved oligomeric peaks as a function of the degree of polymerization would yield the mass of the repeat unit (the slope) and the sum of the masses of the two e nd groups and the cation (the intercept). A method for accurate MALDI-FTICR dete rmination of the polymer molecular weight and polydispersity was develope d and described by the Wilkinsâ€™ Group.164;165 In contrast to biomolecules, which have a modisperse mass corresponding to a single molecular ion, polymers are polydisperse and have a range of different masses corresponding to a variety of oligomeric chai ns. As a result, accurate molecular weight determination of polymers by FTICR is mass dependent since no single constant deceleration time (time delay between the lase r trigger and the application of a trapping potential on the rear trapping pl ate) could be chosen such th at all masses representing the polymer distribution could be trapped and dete cted. Since all MALDI ions have a mean velocity of about 750 m/s, shorter gated deceleration times will trap lighter ions that
88 travel faster, while longer gated deceleration times will suit better for heavier ions that travel slower from the ion source to the I CR cell. To account for this mass dependent signal response, Wilkins introduced a me thod where a number of mass spectra are collected at different gated deceleration times and integrated by adding the corresponding time domain transients. Increasing the deceler ation time by 10 s increments leads to an accurate detection of the entire mass range. The method was used for determination of the molecular weight of a si mulated polydisperse sample co mposed of a mixture of PEG 1000, PEG 3000, PEG 6000, and PEG 8000. For example, a gated de celeration time of 60 s allowed only for detection of PEG 1000. PEG 3000 was detected at a deceleration time of 100 s, PEG 6000 at 170 s, and PEG 8000 at 260 s. However, summation of the time domains of the data obtained between 50 and 280 s at 10 s increments allows for detection and reconstruction of the en tire FTICR spectrum representing all four components of the mixture. Before the introduction of MALDI and ES I, a number of polymer analyses of ethylene oxide/propylene oxi de (EO/PO) type copolymers were done by fast atom bombardment. In particular, Lattimer et al. studied the fragmentation patterns of polyols under collision activated disso ciation performed with a ma gnetic sector Finnigan MAT H-SQ 30 mass spectrometer.166-168 The group of Liang Li used EO/PO copol ymers to study the effects of MALDI experimental conditions on th e mass spectral appearance.55 The presented experimental MALDI-TOF data proved that va riations in analyte concentration (matrix to analyte ratio), the type of matrices and solvents us ed, and the laser power, affect the appearance of the mass spectra of the copolymers. For exam ple, the use of three different matrices in
89 MALDI sample preparation, all-trans-Retin oic acid (RA), dithranol (DTH), and 2-(4hydroxyphenylazo) benzoic acid (HABA), leads to different peak intensities displayed by a single oligomer in spectra acquired in the di fferent matrices; for some of the peaks, a relative standard deviat ion (RSD) of 31.5% was calculated. Similar results were obtained for solvent and laser power variations. In a ddition, the authors noticed that the matrix, solvent, and laser power effects are mass de pendent: with the increase of the molecular weight the variations in peak intensity and spectral appearance decrease. In other words, these effects are more pronounced in the lower mass range of the copolymer distribution. In conclusion, quantitative analyses of copolymers, and of any other compound, by MALDI should be done based on re lative rather than absolute peak intensities due to variations in the spectral appearances associ ated with different sa mple preparation and acquisition parameters. Further, due to th e same reasons, care should be taken in assessing the average molecular weight and the molecular weight distributions of the polymers by MALDI. As discussed in Chapte r 1, for accurate quantitative analyses, mass spectra have to represent the actual compos ition of the analytes a nd MALDI data should be viewed with care. Later, Liang Liâ€™s group continued studi es on another set of EO/PO copolymers with ESI MSn analyses on Ion Trap LC/MSn instrument.169 Four samples altogether displayed an average molecular weight determined by GPC between 2200 and 2400 Da with a polydispersity index (PD I) of 1.1. The ESI spectra showed that neither of the samples is actually a copolymer but rather a mixture of EO and PO. In addition, the MS/MS spectra of oligomeric ions with th e same nominal mass but isolated from two
90 different samples, displayed th e same fragmentation pattern, a sign that the ions belong to the same species. Full structural characteri zation of amine-terminated EO/PO copolymers was done by MALDI-FTICR on a Bruker Apex inst rument with a 7 Tesla magnet.170 In the study, the obtained data allowed for determination of the average molecular weight of the copolymers, their repeat units , and end groups; regression an alyses were used for end group determination. Also, the mass spectra de tected additional oli gomeric distributions that were associated with contaminations and different rearrangements and terminations. Determinations of the block length of triblock poly(oxypropylene) and poly(oxyethylene) copolymers was done by both MALDI-FTICR171 and MALDI-TOF172. Due to the large number of produced peaks, th e authors that performed the TOF analyses used a homemade software that automatically assigned the peak composition after a deisotoping process. The software determin es the relative contribution of each monomer in the observed oligomer and generates 2D plots of the relativ e ion abundance as a function of the number of EO and PO units. The goal of our research is mass spectrome try characterization of two samples of EO/PO copolymers, â€œUCON Lubricang 50HB-660â€ and â€œUCON Lubricant 50-HB2000â€ with average molecular weights of about 2000 and 4000 Da, respectively. The specific aims of the analyses are as follows: Identification of unknown oligomer end groups Determination of the exact number of EO and PO units in every oligomer Sequencing of the individual oligomers by MS/MS
91 To achieve these objectives, we used the combined capabilities of MALDI-TOF, DIOS-TOF, and MALDI-FTICR to obtain comp lementary data that will allow for thorough and accurate mass spectrometry analyses. Experimental Section Materials Poly(ethylene glycol) PEG standards w ith number-average molecular weights (Mn) of 600, 1500, and 2000, the Brilliant Green dye, and the 2,5-Dihydroxybenzoic acid (DHB) were purchased from Sigma-Aldrich (M ilwaukee, WI). Certified sodium iodide was obtained from Fisher Scientific (Fair Lawn, NJ) while the two polymer samples were synthesized and provided by th e Dow Chemical Company. MALDI Sample Preparation The EO/PO copolymers were prepared as 4.6-4 M solutions in deionized water whereas a saturated solution of the matrix was prepared in the same solvent. Sample and matrix solutions were mixed in a 1:1 (v/v) rati on, and applied to a st ainless steel plate as 1 L aliquots. PEGs were dissolved in water at the following concentrations: 1 L/mL PEG 600, 1 mg/mL PEG 1500, 1 mg/mL PE G 2000 and mixed with the saturated DHB solution and a 1 mg/mL NaI/H2O solution in a matrix:analyte:NaI ratio of 1:1:1. Mass Spectrometry Analyses Bruker Reflex II TOF (Ballerica, MA) instrument was used for MALDI-TOF analyses. The instrument is equipped w ith a 337 nm nitrogen laser and delayed extraction. Positive ions were detected in reflector mode with an acceleration voltage of 20 kV and 100 shots acquired for every spot. Bruker Xmass software was used for data processing and theore tical calculations.
92 Bruker Apex II FTICR instrument providing 4.7 T magnetic field was used for high mass accuracy MALDI analyses. Ions were inje cted into the infinity cell by applying broadband excitation and detection, 0.01 s hexapole ion accumulation, and 1.14 V trapping potential using the Dinamic trapping me thod with a time domain data sets of 512 K. All detected ions were sodiated. Results and Discussion Mass spectrometry character ization of the low (2000 Da) and the high (4000 Da) molecular weight versions of the EO/PO c opolymers started with MALDI-TOF analyses. Figure 5-1 shows a MALDI-TOF spectrum of the â€œUCON Lubricant 50-HB-600â€, which was acquired in the reflectron mode of the Br uker Reflex II instrument. The TOF data of this copolymer provide limited structural info rmation especially in the higher mass region of the spectrum where the resolution (Rs) is lower. A zoom into the low mass region, however, allows for determination of the ex act number of EO a nd PO units of each oligomeric chain (Figure 5-2). Figure 5-1. A MALDI-TOF spectrum of â€œUCON Lubricant 50-HB-600â€.
93 For mass to charge of about 1700, for example, the MALDI-TOF instrument provides a resolution of approximately 5000, wh ich is enough to assign the exact number of EO and PO units. As shown in Figure 52, the spectrum consists of clustered peaks and each cluster represents at leas t three different oligomers. Figure 5-2. A zoom into the low mass re gion of the MALDI-TOF spectrum of â€œUCON Lubricant 50-HB-600â€. The first clustered peak in the spectr um, for example, corresponds to the overlapping isotopic distributions of the resolved oligomers (EO)14(PO)18, (EO)18(PO)15, and (EO)22(PO)12. The detected masses correspond to a butanol initiated polymerization with sodium promoted ionization in the MA LDI source. With the same success, peak assignment could be done for the rest of the oligomeric clusters in the low mass region. However, a look into the higher mass region of the spectrum (Figure 5-3) shows that as the mass increases, it becomes more difficult to assign the exact structure of the oligomer corresponding to a certain mass.
94 For example, the achieved accuracy for m/z 2466.6 does not allow for unambiguous determination of the oligomer ic structure corresponding to this mass. Two different oligomeric chains, with ma sses differing by 0.2 Da [(EO)9(PO)34 and (EO)38(PO)12]could both contribute to a peak at 2466.6 Da. Figure 5-3. A zoom into the high mass re gion of the MALDI-TO F spectrum of â€œUCON Lubricant 50-HB-600â€. In this case, the question becomes what re solution should be achieved to obtain a mass accuracy that is high enough to resolve this problem. The theoretical isotopic distributions calculated for (EO)9(PO)34 and (EO)38(PO)12 at a resolution of 5000, provided in the low mass region of the MALD I-TOF spectrum, is shown in Figure 5-4. The theoretical spectrum of the two oligom ers shows that if analyzed separately, the oligomeric ions could be detected with resolved isot opic peaks at a resolution of 5000. However, detection of the two oligom ers in a mixture at a resolution of 5000 will not provide separation of their ove rlapping isotopic distributions.
95 Separation of the overlapping isiotopic se ries corresponding to the two oligomers could be achieved at a resolution of 100 000 (F igure 5-5). In this case, the achieved high mass accuracy leads to a distinct separation of the two distributions. As a result, if both oligomer structures are present in the mixture, they will be separated and detected. If only one of them is present, the mass analysis will determine which one it is with very high accuracy. Figure 5-4. Theoretical isotopic distribution calculated for (EO)9(PO)34 and (EO)38(PO)12 at Rs of 5000. As the molecular weight of the polymer in creases, it becomes hard to even guess the possible oligomeric structures of the MALDI-TOF peaks due to the poor resolution and mass accuracy provided by the TOF analyzer. A zoom into the MALDI-TOF spectrum of â€œUCON Lubricant 50-HB-2000â€ demonstrates the difficulties associated with TOF data processing of the high molecular weight version of the EO/PO copolymers (Figure 5-6). The presented MALDI-TOF data and theoretic al calculations indicate that time of flight analyses will not provide the necessa ry resolution and mass accuracy to assign the
96 exact number of EO and PO units in every oligomeric chain. In copolymers, and polymer mixtures in general, the large amount of ions make mass spectrometry analyses at low resolution difficult in terms of ion se paration, data proce ssing, and unambiguous peak assignment. The only mass analyzer available today that provides ultra high resolution and mass accuracy is the FTICR mass analyzer. Figure 5-5. Theoretical isotopic distribution calculated for (EO)9(PO)34 and (EO)38(PO)12 at Rs of 100 000. As a result, our further efforts to ch aracterize the EO/PO copolymers were concentrated on achieving high resolution and mass accuracy by FTICR analyzes. Due to the large amount of ions produced by the copolymers, the ionization method of choice was MALDI; as stated in previous chapte rs, electrospray ionization induces multiple charging of the analytes, which co mplicates the spectra even more.
97 Figure 5-6. A zoom into the MALDI-TOF spectrum of â€œUCON Lubricant 50-HB-2000â€ The MALDI-FTICR analyses started with tuning the instrument by detection of Brilliant Green, a dye purchased as a marker that was easily applied on a stainless steel MALDI plate, and erased afterwards. N CH3 H3C N CH3 CH3 + SOO O HO Brilliant Green Following tuning, the MALDI-FTICR analyzes targeted detection of different molecular weight poly(ethylene glycol)s that would be used as bases for calibration. Detection of the low molecular weight vers ions of the glycols, like PEG 600 and PEG 1500, provided the spectra show n in Figures 5-7 and 5-8. Compared to the TOF spectra, the FTICR spectra show very well defined peaks with high signal to noise ratio in additi on to the high resolution and mass accuracy. For example, a zoom into the FTICR spectrum of PEG 1500, Figure 5-9, shows that for m/z 1494.64177 the achieved resolution is 95 872. 6 with signal to noise ratio S/N
98 of 469.8. Detection of higher molecular weig ht poly(ethylene glycol ) analytes was also achieved but with lower resolution and lower signal to noise ratio. In addition to that, the signal fluctuated, which led to difficulties in spectra averaging (usually ten to fifty laser shots were acquired for every spectrum). Figure 5-7. A MALDI-FTI CR spectrum of Peg 600
99 Figure 5-8. A MALDI-FTICR spectrum of PEG 1500.
100 Figure 5-9. A zoom into the MALDI-FTICR spectrum of PEG 1500. The full MALDI-FTICR spectrum of â€œUCON Lubricant 50-HB-600â€ is shown in Figure 5-10. Initial evaluation of the spect rum indicates its low signal to noise ratio; however, a zoom into the spectrum shows that despite the poor signa l to noise ratio, the achieved resolution is about 30 000.
101 Figure 5-10. A MALDI-FTICR spectrum of â€œUCON Lubricant 50-HB-600â€. For example, a zoom into the m/z 1600-1900 region of the spectrum provides an accuracy of about 0.003 Da with a resolution for m/z 1774 of 26 795 (Figure 5-11). The higher resolution and mass accuracy provided by the FTICR analyzer allows for increased confidence in unambiguous peak assignment of the oligomeric structures. However, compared to the structural in formation obtained MALDI-TOF, FTICR data do not add any new details to the mass spectrometr y characterization. In addition, a closer look at the data shows that the peak distri bution and spectra appearance obtained by TOF and FTICR differ to a certain extent. For example, according to the MALDI-TOF data, the oligomeric cluster around 1774 Da corresp onds to three oligomeric structuresâ€” (EO)14(PO)18, (EO)18(PO)15, and (EO)22(PO)12 (Figure 5-2). The MALDI-FTICR spectrum of the same region, however, consists of fewer peaks that seem to correspond to two oligomeric structures at most (Figure 5-11).
102 Figure 5-11. A zoom into the low mass region of the MALDI-FTICR spectrum of â€œUCON Lubricant 50-HB-600â€. The same discrepancy in TOF and FTICR spectral appearances are noticed in the higher mass region of â€œUCON Lubricant 50-HB-600â€, which could be seen from Figure 5-12 and 5-13. The accuracy provided by the MALDI-TOF spectrum (Figure 5-12) does not distinguish between the two di fferent oligomeric structures that could be assigned for m/z 2230.7 [(EO)30(PO)14 and (EO)1(PO)36] and 2244.8 [(EO)29(PO)15 and (EO)0(PO)37]. In contrast, the FTICR data provi de accuracy that unambiguously detects one of the possible oligomeric structures as present in the mi xture rather than the other (Figure 5-13). However, the 2230 and 2244 Da masses detect ed in the FTICR spectrum appear in cluster positions that
103 Figure 5-12. A zoom into the 2225-2265 Da region of the MALD I-TOF spectrum of â€œUCON Lubricant 50-HB-600â€. do not match the positions in the TOF clusters . Moreover, no isotopic peaks are detected by FTICR for these masses. These discrepancie s introduce uncertainty in the validity of the obtained MALDI-FTICR data. Figure 5-13. A zoom into the 2225-2263 Da region of the MALDI-FTICR spectrum of â€œUCON Lubricant 50-HB-600â€. Optimization of the FTICR acquisition para meters should allow not only for data that have better resolution and mass accuracy but also for peak distributions that are not limited by defects associated with the number of peaks or peak positions. Most of the
104 challenges faced during the work with the FTICR instrument, however, were instrumental. Due to laser misalignment, easily deposited matrix and analyte in the source, and in particular, on th e rods of the hexapole ion guide, and a number of other issues, no stable signal was obtained out of the MALDI source of the FTICR instrument for more than a couple of days at a time. Consequently, future work on this project should be focused on optimizing the opera tional parameters of the MALDI source interfaced to the FTICR instrument. Conclusions MALDI-FTICR analyses of the analyzed EO/PO copolymers prove to be a promising alternative to ESI-FTICR due to the lack of multiply charging of the analytes and the tolerance to a variety of solvents for sample preparation; electrospray ionization is limited to more polar solvents, which used in analyses of biomolecules rather than polymers. Despite its advantages, MA LDI-FTICR analyses of complex polymer mixtures like copolymers are challenging due to the large number of produced ions and the difficulties associated with ion accumulati on, space charge effect, and detection. For polymer FTICR analyses, MALDI should be co upled to a separation system like GPC to reduce the number of accumulated ions. As an alternative to sepa ration interfacing, the acquisition might be achieved by collecting sp ectra at varying deceleration times and signal averaging the spectra.
105 LIST OF REFERENCES 1. Carraher, C. E., Jr. Polymer Chemistry, sixth edition; Marcel Dekker, Inc.; New York, Basel: 2003. 2. Odian, G. Principles of Polymerization, third edition; John Wiley and Sons, Inc.; New York, Chichester, Brisbane , Toronto, Singapore: 1991. 3. Hanton, S. D. Chemical Reviews 2001, 101, 527-69. 4. Montaudo, G.; Lattimer, R. P. Mass Spectrometry of Polymers, first edition; CRC Press; Boca Raton, London, Ne w York, Washington D.C.: 2002. 5. Murgasova, R.; Hercules, D. M. International Journal of Mass Spectrometry 2003, 226, 151-62. 6. Nielen, M. W. F. Mass Spectrometry Reviews 1999, 18, 309-44. 7. Rader, H. J.; Schrepp, W. Acta Polymerica 1998, 49, 272-93. 8. McEwen, C. N.; Peacock, P. M. Analytical Chemistry 2002, 74, 2743-48. 9. Montaudo, M. S. Mass Spectrometry Reviews 2002, 21, 108-44. 10. Bahr, U.; Deppe, A.; Karas, M.; Hillenkamp, F.; Giessmann, U. Analytical Chemistry 1992, 64, 2866-69. 11. Karas, M.; Hillenkamp, F. Analytical Chemistry 1988, 60, 2299-301. 12. Tanaka, K.; Waki, H.; Ido, Y.; Akita, S.; Yoshida, Y.; Yoshida, T. Rapid Commun.Mass Spectrom. 1988, 2, 2299. 13. Fenn, J. B.; Mann, M.; Meng, C. K.; Wong, S. F.; Whitehouse, C. M. Science 1989, 246, 64-71. 14. Yamashita, M.; Fenn, J. B. Journal of Physical Chemistry 1984, 88, 4451-59. 15. Aaserud, D. J.; Prokai, L.; Simonsick, W. J. Analytical Chemistry 1999, 71, 479399. 16. McEwen, C. N.; Simonsick, W. J.; Larsen, B. S.; Ute, K.; Hatada, K. Journal of the American Society for Mass Spectrometry 1995, 6, 906-11.
106 17. Prokai, L.; Aaserud, D. J.; Simonsick, W. J. Journal of Chromatography A 1999, 835, 121-26. 18. Burton, R. D.; Watson, C. H.; Eyler, J. R.; Lang, G. L.; Powell, D. H.; Avery, M. Y. Rapid Communications in Mass Spectrometry 1997, 11, 443-46. 19. Karas, M.; Kruger, R. Chemical Reviews 2003, 103, 427-39. 20. Karas, M.; Gluckmann, M.; Schafer, J. Journal of Mass Spectrometry 2000, 35, 112. 21. Knochenmuss, R.; Zenobi, R. Chemical Reviews 2003, 103, 441-52. 22. Karbach, V.; Knochenmuss, R. Rapid Communications in Mass Spectrometry 1998, 12, 968-74. 23. Yingling, Y. G.; Garrison, B. J. Chemical Physics Letters 2002, 364, 237-43. 24. Handschuh, M.; Nettesheim, S.; Zenobi, R. Applied Surface Science 1999, 137, 125-35. 25. Fournier, I.; Brunot, A.; Tabet, J. C.; Bolbach, G. International Journal of Mass Spectrometry 2002, 213, 203-15. 26. Knochenmuss, R. Journal of Mass Spectrometry 2002, 37, 867-77. 27. Zhigilei, L. V.; Garrison, B. J. Journal of Applied Physics 2000, 88, 1281-98. 28. Zeifman, M. I.; Garrison, B. J.; Zhigilei, L. V. Journal of Applied Physics 2002, 92, 2181-93. 29. Guilhaus, M. Journal of Mass Spectrometry 1995, 30, 1519-32. 30. Wollnik, H. Mass Spectrometry Reviews 1993, 12, 89-114. 31. Watson, T. J. Introduction to Mass Spectrometry, third edition; Lippincott-Raven; Philadelphia, New York: 1997. 32. Wiley, W. C.; Mclaren, I. H. Review of Scientific Instruments 1955, 26, 1150-57. 33. Brown, R. S.; Lennon, J. J. Analytical Chemistry 1995, 67, 1998-2003. 34. Spengler, B.; Cotter, R. J. Analytical Chemistry 1990, 62, 793-96. 35. Mamyrin, B. A. International Journal of Mass Spectrometry and Ion Processes 1994, 131, 1-19. 36. Amster, I. J. Journal of Mass Spectrometry 1996, 31, 1325-37.
107 37. Comisaro.M.B; Marshall, A. G. Chemical Physics Letters 1974, 26, 489-90. 38. Comisaro.M.B; Marshall, A. G. Chemical Physics Letters 1974, 25, 282-83. 39. Comisaro.M.B; Marshall, A. G. Canadian Journal of Chemistry-Revue Canadienne de Chimie 1974, 52, 1997-99. 40. Dienes, T.; Pastor, S. J.; Schurch, S.; Scott, J. R.; Shulin, C.; Wilkins, C. L Mass Spectrometry Reviews 1996, 15, 163-211. 41. Wilkins, C. L.; Gross, M. L. Analytical Chemistry 1981, 53, A661-&. 42. Hanton, S. D.; Hyder, I. Z. ; Stets, J. R.; Owens, K. G.; Blair, W. R.; Guttman, C. M.; Giuseppetti, A. A. Journal of the American Soc iety for Mass Spectrometry 2004, 15, 168-79. 43. Hoteling, A. J.; Erb, W. J.; Tyson, R. J.; Owens, K. G. Analytical Chemistry 2004, 76, 5157-64. 44. Dogruel, D.; Nelson, R. W.; Williams, P. Rapid Communications in Mass Spectrometry 1996, 10, 801-04. 45. Hoteling, A. J.; Mourey, T. H.; Owens, K. G. Analytical Chemistry 2005, 77, 75056. 46. Llenes, C. F.; Omalley, R. M. Rapid Communications in Mass Spectrometry 1992, 6, 564-70. 47. Thomson, B.; Suddaby, K.; Rudin, A.; Lajoie, G. European Polymer Journal 1996, 32, 239-56. 48. Bauer, B. J.; Wallace, W. E.; Fanconi, B. M.; Guttman, C. M. Polymer 2001, 42, 9949-53. 49. Lin-Gibson, S.; Brunner, L.; Vanderhart, D. L.; Bauer, B. J.; Fanconi, B. M.; Guttman, C. M.; Wallace, W. E. Macromolecules 2002, 35, 7149-56. 50. Schriemer, D. C.; Li, L. A. Analytical Chemistry 1997, 69, 4169-75. 51. McEwen, C. N; Jackson, C.; Larsen, B. S. International Journal of Mass Spectrometry and Ion Processes 1997, 160, 387-94. 52. Byrd, H. C. M.; Mcewen, C. N. Analytical Chemistry 2000, 72, 4568-76. 53. Hanton, S. D.; Liu, X. M. Analytical Chemistry 2000, 72, 4550-54. 54. Montaudo, M. S.; Puglisi, C.; Samperi, F.; Montaudo, G. Rapid Communications in Mass Spectrometry 1998, 12, 519-28.
108 55. Chen, R.; Zhang, N.; Tseng, A. M.; Li, L. Rapid Communications in Mass Spectrometry 2000, 14, 2175-81. 56. Yan, W. Y.; Gardella, J. A.; Wood, T. D. Journal of the American Society for Mass Spectrometry 2002, 13, 914-20. 57. Chen, H.; He, M. Y.; Pei, J.; He, H. F. Analytical Chemistry 2003, 75, 6531-35. 58. Hensel, R. R.; King, R. C.; Owens, K. G. Rapid Communications in Mass Spectrometry 1997, 11, 1785-93. 59. Murgasova, R.; Hercules, D. M. Analytical Chemistry 2003, 75, 3744-50. 60. Williams, J. B.; Chapman, T. M.; Hercules, D. M. Analytical Chemistry 2003, 75, 3092-100. 61. Stewart, M. P.; Buriak, J. M. Advanced Materials 2000, 12, 859-69. 62. Wei, J.; Buriak, J. M.; Siuzdak, G. Nature 1999, 399, 243-46. 63. Kruse, R. A.; Li, X. L.; Bohn, P. W.; Sweedler, J. V. Analytical Chemistry 2001, 73, 3639-45. 64. Lewis, W. G.; Shen, Z. X.; Finn, M. G.; Siuzdak, G. International Journal of Mass Spectrometry 2003, 226, 107-16. 65. Shen, Z. X.; Thomas, J. J.; Averbuj, C.; Broo, K. M.; Engelhard, M.; Crowell, J. E.; Finn, M. G.; Siuzdak, G. Analytical Chemistry 2001, 73, 612-19. 66. Thomas, J. J.; Shen, Z. X.; Blackledge, R.; Siuzdak, G. Analytica Chimica Acta 2001, 442, 183-90. 67. Asao, N.; Sudo, T.; Yamamoto, Y. Journal of Organic Chemistry 1996, 61, 765455. 68. Buriak, J. M.; Allen, M. J. Journal of the American Chemical Society 1998, 120, 1339-40. 69. Tuomikoski, S.; Huikko, K.; Grigoras, K.; Ostman, P.; Kostiainen, R.; Baumann, M.; Abian, J.; Kotiaho, T.; Fransilla, S. Lab Chip 2002, 2, 247-53. 70. Alimpiev, S.; Nikiforov, S.; Karavanskii, V.; Minton, T.; Sunner, J. Journal of Chemical Physics 2001, 115, 1891-901. 71. Cuiffi, J. D.; Hayes, D. J.; Fonash, S. J.; Brown, K. N.; Jones, A. D. Analytical Chemistry 2001, 73, 1292-95. 72. Kruse, R. A.; Rubakhin, S. S; Romanova, E. V.; Bohn, P. W.; Sweedler, J. V. Journal of Mass Spectrometry 2001, 36, 1317-22.
109 73. Li, Q.; Ricardo, A.; Benner, S. A.; Winefordner, J. D.; Powell, D. H. Analytical Chemistry 2005, 77, 4503-08. 74. Zou, H. F.; Zhang, Q. C.; Guo, Z.; Guo, B. C.; Zhang, Q.; Chen, X. M. Angewandte Chemie-International Edition 2002, 41, 646-+. 75. Lindmark-Hamberg, M.; Wagener, K. B. Macromolecules 1987, 20, 2949-51. 76. Wagener, K. B.; Nel, J. G.; Konzelman, J.; Boncella, J. M. Macromolecules 1990, 23, 5155-57. 77. Smith, J. A.; Brzezinska, K. R.; Valenti, D. J.; Wagener, K. B. Macromolecules 2000, 33, 3781-94. 78. Sworen, J. C.; Smith, J. A.; Berg, J. M.; Wagener, K. B. Journal of the American Chemical Society 2004, 126, 11238-46. 79. Bourgeois, D.; Pancrazi, A.; Nolan, S. P.; Prunet, J. Journal of Organometallic Chemistry 2002, 643, 247-52. 80. Courchay, F. C.; Sworen, J. C.; Wagener, K. B. Macromolecules 2003, 36, 823139. 81. Hong, S. H.; Day, M. W.; Grubbs, R. H. Journal of the American Chemical Society 2004, 126, 7414-15. 82. Lehman, S. E.; Schwendeman, J. E.; O'Donnell, P. M.; Wagener, K. B. Inorganica Chimica Acta 2003, 345, 190-98. 83. Schrock, R. R.; Murdzek, J. S.; Bazan, G. C.; Robbins, J.; Dimare, M.; Oregan, M. Journal of the American Chemical Society 1990, 112, 3875-86. 84. Oskam, J. H.; Fox, H. H.; Yap, K. B.; Mcconville, D. H.; Odell, R.; Lichtenstein, B. J.; Schrock, R. R. Journal of Organometallic Chemistry 1993, 459, 185-98. 85. Louie, J.; Grubbs, R. H. Organometallics 2002, 21, 2153-64. 86. Schwab, P.; Grubbs, R. H.; Ziller, J. W. Journal of the American Chemical Society 1996, 118, 100-10. 87. Bielawski, C. W.; Grubbs, R. H. Angewandte Chemie-Inter national Edition in English 2000, 39, 2903. 88. Chatterjee, A. K; Grubbs, R. H. Organic Letters 1999, 1, 1751. 89. Hopkins, T. E.; Pawlow, J. H.; Koren, D. L.; Deters, K. S.; Solivan, S. M.; Davis, J. A.; Gomez, F. J.; Wagener, K. B. Macromolecules 2001, 34, 7920-22. 90. Hopkins, T. E.; Wagener, K. B. Advanced Materials 2002, 14, 1703-15.
110 91. Hopkins, T. E.; Wagener, K. B. Macromolecules 2003, 36, 2206-14. 92. Scholl, M.; Ding, S.; Lee, C. W.; Grubbs, R. H. Organic Letters 1999, 1, 953-56. 93. Dias, E. L.; Nguyen, S. T.; Grubbs, R. H. Journal of the American Chemical Society 1997, 119, 3887-97. 94. Huang, J. K.; Schanz, H. J.; Stevens, E. D.; Nolan, S. P. Organometallics 1999, 18, 5375-80. 95. Salvini, A.; Piacenti, F.; Frediani , P.; Devescovi, A.; Caporali, A. Journal of Organometallic Chemistry 2001, 625, 255-67. 96. Martinez, V.; Blais, J. C.; Bravic, G.; Astruc, D. Organometallics 2004, 23, 86174. 97. Scherman, O. A.; Rutenberg, I. M.; Grubbs, R. H. Journal of the American Chemical Society 2003, 125, 8515-22. 98. Tezuka, Y.; Komiya, R.; Washizuka, M. Macromolecules 2003, 36, 12-17. 99. Peacock, P. M.; Mcewen, C. N. Analytical Chemistry 2004, 76, 3417-27. 100. Liu, J. S.; Loewe, R. S.; McCullough, R. D. Macromolecules 1999, 32, 5777-85. 101. Puglisi, C.; Samperi, F.; Alicata, R.; Montaudo, G. Macromolecules 2002, 35, 3000-07. 102. Yalcin, T.; Wallace, W. E.; Guttman, C. M.; Li, L. Analytical Chemistry 2002, 74, 4750-56. 103. Montaudo, M. S.; Puglisi, C.; Samperi, F.; Montaudo, G. Macromolecules 1998, 31, 3839-45. 104. Biagini, S. C. G.; Davies, R. G.; Gibson, V. C.; Giles, M. R.; Marshall, E. L.; North, M. Polymer 2001, 42, 6669-71. 105. Koyama, E.; Sanda, F.; Endo, T. Journal of Polymer Science Part A-Polymer Chemistry 1997, 35, 345-52. 106. Koyama, E.; Sanda, F.; Endo, T. Macromolecules 1998, 31, 1495-500. 107. Maynard, H. D.; Okada, S. Y.; Grubbs, R. H. Macromolecules 2000, 33, 6239-48. 108. Hopkins, T. E.; Wagener, K. B. Macromolecules 2004, 37, 1180-89. 109. Herrison, J. L.; Chauvin, Y. N. Makromol.Chem. 1971, 141, 161. 110. McGrath, D. V.; Grubbs, R. H. Organometallics 1994, 13, 224-35.
111 111. Maynard, H. D.; Grubbs, R. H. Tetrahedron Letters 1999, 40, 4137-40. 112. Sutton, A. E.; Seigal, B. A.; Finnegan, D. F.; Snapper, M. L. Journal of the American Chemical Society 2002, 124, 13390-91. 113. Bielawski, C. W.; Scherman, O. A.; Grubbs, R. H. Polymer 2001, 42, 4939-45. 114. Hillmyer, M. A.; Nguyen, S. T.; Grubbs, R. H. Macromolecules 1997, 30, 718-21. 115. Jazzar, R. F. R.; Macgregor, S. A.; Mahon, M. F.; Richards, S. P.; Whittlesey, M. K. Journal of the American Chemical Society 2002, 124, 4944-45. 116. Trnka, T. M.; Morgan, J. P.; Sanford, M. S.; Wilhelm, T. E.; Scholl, M.; Choi, T. L.; Ding, S.; Day, M. W.; Grubbs, R. H. Journal of the American Chemical Society 2003, 125, 2546-58. 117. Ulman, M.; Grubbs, R. H. Journal of Organic Chemistry 1999, 64, 7202-07. 118. Grubbs, R. H.; Tumas, W. Science 1989, 243, 907-15. 119. Grubbs, R. H.; Miller, S. J.; Fu, G. C. Accounts of Chemical Research 1995, 28, 446-52. 120. Grubbs, R. H.; Chang, S. Tetrahedron 1998, 54, 4413-50. 121. Feher, F. J.; Soulivong, D.; Ek lund, A. G.; Wyndham, K. D. Chemical Communications 1997, 1185-86. 122. Brummer, O.; Ruckert, A.; Blechert, S. Chemistry-A European Journal 1997, 3, 441-46. 123. Schuster, M.; Blechert, S. Angewandte Chemie-International Edition in English 1997, 36, 2037-56. 124. Connon, S. J.; Blechert, S. Angewandte Chemie-International Edition 2003, 42, 1900-23. 125. Schwab, P.; Grubbs, R. H.; Ziller, J. W. Journal of the American Chemical Society 1996, 118, 100-10. 126. Dias, E. L.; Nguyen, S. T.; Grubbs, R. H. Journal of the American Chemical Society 1997, 119, 3887-97. 127. Blackwell, H. E.; O'Leary, D. J.; Chatterjee, A. K.; Washenfelder, R. A.; Bussmann, D. A.; Grubbs, R. H. Journal of the American Chemical Society 2000, 122, 58-71. 128. Chatterjee, A. K.; Morgan, J. P.; Scholl, M.; Grubbs, R. H. Journal of the American Chemical Society 2000, 122, 3783-84.
112 129. Grela, K.; Bieniek, M. Tetrahedron Letters 2001, 42, 6425-28. 130. Chatterjee, A. K.; Choi, T. L.; Sanders, D. P.; Grubbs, R. H. Journal of the American Chemical Society 2003, 125, 11360-70. 131. Crowe, W. E.; Goldberg, D. R. Journal of the American Chemical Society 1995, 117, 5162-63. 132. Ulman, M.; Grubbs, R. H. Organometallics 1998, 17, 2484-89. 133. Petkovska, V. I.; Hopkins, T. E.; Powell, D. H.; Wagener, K. B. Macromolecules 2005, 38, 5878-85. 134. Wei, J.; Buriak, J. M.; Siuzdak, G. Nature 1999, 399, 243-46. 135. Lewis, W. G.; Shen, Z. X.; Finn, M. G.; Siuzdak, G. International Journal of Mass Spectrometry 2003, 226, 107-16. 136. Wagener, K. B.; Boncella, J. M.; Nel, J. G. Macromolecules 1991, 24, 2649-57. 137. Sworen, J. C.; Pawlow, J. H.; Case, W.; Lever, J.; Wagener, K. B. Journal of Molecular Catalysis A-Chemical 2003, 194, 69-78. 138. Pernerstorfer, J.; Schuster, M.; Blechert, S. Chemical Communications 1997, 194950. 139. Gibson, S. E.; Gibson, V. C.; Keen, S. P. Chemical Communications 1997, 110708. 140. McNaughton, B. R.; Bucholtz, K. M.; Camaano-Moure, A.; Miller, B. L. Organic Letters 2005, 7, 733-36. 141. Methenitis, C.; Morcelletsauvage, J.; Morcellet, M. Polymer Bulletin 1984, 12, 141-47. 142. Canham, L. T. Applied Physics Letters 1990, 57, 1046-48. 143. Bilhou, J. L.; Basset, J. M. ; Mutin, R.; Graydon, W. F. Journal of the American Chemical Society 1977, 99, 4083-90. 144. Courtier, J. L.; Paillet, C.; Lecont e, M.; Basset, J. M.; Weiss, K. Angewandte Chemie-International Edition 1992, 31, 628-31. 145. Grubbs, R. H.; Hoppin, C. R. Journal of the American Chemical Society 1979, 101, 1499-508. 146. Lehman, S. E.; Wagener, K. B. Macromolecules 2002, 35, 48-53.
113 147. Harrity, J. P. A.; La, D. S.; Cefalo, D. R.; Visser, M. S.; Hoveyda, A. H. Journal of the American Chemical Society 1998, 120, 2343-51. 148. Marshall, A. G.; Hendrickson, C. L.; Jackson, G. S. Mass Spectrometry Reviews 1998, 17, 1-35. 149. Brown, R. S.; Weil, D. A.; Wilkins, C. L. Macromolecules 1986, 19, 1255-60. 150. Heeren, R. M. A.; Kleinnijenhuis, A. J.; McDonnell, L. A.; Mize, T. H. Analytical and Bioanalytical Chemistry 2004, 378, 1048-58. 151. Marshall, A. G. Accounts of Chemical Research 1985, 18, 316-22. 152. Nibbering, N. M. M. Analyst 1992, 117, 289-93. 153. McLafferty, F. W. Accounts of Chemical Research 1994, 27, 379-86. 154. O'Connor, P. B.; McLafferty, F. W. Journal of the Americ an Chemical Society 1995, 117, 12826-31. 155. Maziarz, E. P.; Baker, G. A.; Wood, T. D. Macromolecules 1999, 32, 4411-18. 156. Hettich, R.; Buchanan, M. Journal of the American Society for Mass Spectrometry 1991, 2, 402-12. 157. Hettich, R. L.; Buchanan, M. V. Journal of the American Society for Mass Spectrometry 1991, 2, 22-28. 158. Castro, J. A.; Koster, C.; Wilkins, C. Rapid Communications in Mass Spectrometry 1992, 6, 239-41. 159. Baykut, G.; Jertz, R.; Witt, M. Rapid Communications in Mass Spectrometry 2000, 14, 1238-47. 160. Knobeler, M.; Wanczek, K. P. Journal of the American Society for Mass Spectrometry 1996, 7, 1026-33. 161. Mize, T. H.; Amster, I. J. Analytical Chemistry 2000, 72, 5886-91. 162. de Koster, C. G.; Duursma, M. C.; vanRooij, G. J.; Heeren, R. M. A.; Boon, J. J. Rapid Communications in Mass Spectrometry 1995, 9, 957-62. 163. van Rooij, G. J.; Duursma, M. C.; Heeren, R. M. A.; Boon, J. J.; deKoster, C. G. Journal of the American Soc iety for Mass Spectrometry 1996, 7, 449-57. 164. Dey, M.; Castoro, J. A.; Wilkins, C. L. Analytical Chemistry 1995, 67, 1575-79. 165. Pastor, S. J.; Wilkins, C. L. Journal of the American Society for Mass Spectrometry 1997, 8, 225-33.
114 166. Lattimer, R. P.; Munster, H.; Budzikiewicz, H. International Journal of Mass Spectrometry and Ion Processes 1989, 90, 119-29. 167. Lattimer, R. P. Journal of the American Society for Mass Spectrometry 1992, 3, 225-34. 168. Selby, T. L.; Wesdemiotis, C.; Lattimer, R. P. Journal of the American Society for Mass Spectrometry 1994, 5, 1081-92. 169. Chen, R.; Tseng, A. M.; Uhing, M.; Li, L. Journal of the American Society for Mass Spectrometry 2001, 12, 55-60. 170. van der Hage, E. R. E.; Duursma, M. C.; Heeren, R. M. A.; Boon, J. J.; Nielen, M. W. F.; Weber, A. J. M.; deKo ster, C. G.; deVries, N. K. Macromolecules 1997, 30, 4302-09. 171. van Rooij, G. J.; Duursma, M. C.; de Kost er, C. G.; Heeren, R. M. A.; Boon, J. J.; Schuyl, P. J. W.; van der Hage, E. R. E. Analytical Chemistry 1998, 70, 843-50. 172. Terrier, P.; Buchmann, W.; Cheguillaume, G.; Desmazieres, B.; Tortajada, J. Analytical Chemistry 2005, 77, 3292-300.
115 BIOGRAPHICAL SKETCH Violeta started her chemistry educati on while attending high school in her home town of Sandanski, Bulgaria, where she was enrolled in a Chemistry major class. The four year chemistry high school experience was followed by a five year college at the Chemistry Department, University of Sofia, Bulgaria. In 2000, sh e joined the group of Dr. Kenneth Wagener at the University of Florida to complete her Ph.D. degree in analyical chemistry.