MSc Chemistry
Track Analytical Sciences
Literature Thesis
Mass Spectrometry of megadalton molecules
by
Zsófia Végh
12264113
October 2019
12 ECTS
Period: 1 August - 31 October 2019
Supervisor/Examiner:
Examiner:
1
Table of contents
1. Introduction ...2
2. Ionization methods for MDa molecules ...5
2.1. MALDI ...5 2.2. ESI ...6 3. Instrumentation of conventional MS ...9 3.1. Mass analysers ...9 3.1.1. TOF ...9 3.1.2. Quadrupoles ... 10 3.1.3. Ion trap ... 11 3.1.4. Orbitrap ... 12 3.1.5. FT-ICR ... 13 3.2. Detectors ... 13
4. Applications of conventional MS methods in the MDa range... 15
5. Single-molecule MS methods ... 20
5.1. Optical detection with QIT ... 21
5.2. Single-molecule FT-ICR ... 24
5.3. Nanomechanical mass spectrometry ... 26
5.4. Cryogenic detectors ... 31
5.4.1. Microcalorimeters ... 32
5.4.2. Superconducting tunnel junctions ... 34
5.5. Charge detection mass spectrometry ... 36
5.5.1. Single-pass CDMS ... 37 5.5.2. Linear array CDMS ... 39 5.5.3. Ion trap CDMS ... 40 5.5.4. Applications of CDMS ... 43 6. Conclusions ... 49 7. References ... 52
2
1. Introduction
Mass spectrometry (MS) is a powerful analytical technique which is widely used in many scientific areas for the identification, characterisation and quantitation of macromolecules. For a long time, the analysis of molecules was possible only up to the 100s of kDa due to the limitations of the current technology. At first only volatile compounds or those that could be derivatized into volatiles could be analysed since the ionization techniques (EI and CI) required that the molecules be in the gas phase. 1 With the invention of new, softer ionization methods:
electrospray ionization (ESI) and matrix-assisted laser desorption/ionization (MALDI) the analysis of molecules with higher molecular weight became possible.
Large biomolecules with masses in the megadalton range play important roles in many scientific areas, such as biomedical science, chemistry, physics, etc. A critical parameter to characterise these molecules is the determination of their masses. 2 The ability to measure very
high masses gained much interest, since it could make the analysis of whole microorganisms possible. The biomolecules of interest in the megadalton range include large protein assemblies, intact viruses, virus capsids, bacteria and DNA ions. The measurement of individual virus particle masses and the mass variability of the populations are essential for the investigation of the structure and the properties of these genetically diverse molecules. 3 These measurements
are of great interest because of their potential ability to make fast identifications of potentially dangerous viruses and bacteria. 4 Virus capsids are also of interest because of their potential
application in the border of medicine and nanotechnology as nano-containers, -reactors or assembly scaffolds. 5 Similarly to biomacromolecules, synthetic macromolecules are also
important in many areas. Among the possible applications of pharmaceutically important polymers are polymer-drug or polymer-protein conjugates, polymer therapeutics, polymer micelles and multicomponent polyplexes. The characterization and controlling of the sequence of building blocks provides the opportunity to tailor the properties of synthetic polymers. Sequence-controlled polymers are not only important in biomedical areas but also in different fields, such as photonics, self-assemblies, solar cells, membranes, nanostructures, etc. 6
Mass spectrometry went through a lot of developments to measure higher and higher masses, but it seems that the upper achievable limit by conventional MS is reached. There are several factors which determine this high-mass limit. One of them is the low detection efficiency that commonly used detectors such as microchannel plates show for high m/z ions. MALDI produces mostly singly charged ions, in which case the m/z is higher and so the detection efficiency is lower. This problem would be solved by ESI which produces more
3
highly charged ions, however higher charges makes the resolving of charge state peaks in the spectrum more difficult. 7
The charge state resolution of very large ions is often lacking which could be explained by the intrinsic heterogeneity of analytes or by the incomplete desolvation which results in peak broadening. Another factor that limits the mass-range is that large ions can acquire a considerable amount of kinetic energy. This can pose a problem in the transmission and focusing of the ions. To improve these processes the ions needs to be cooled down, which is usually done by increasing the background pressure so the system must be able to handle large amount of gases. In exchange, the increased gas pressure also improves the desolvation. 7
Despite the many challenges, there are several experiments done in the megadalton range with conventional MS. For the successful analysis of macromolecules with such a high molecular weight the samples usually need to be homogenous or with limited heterogeneity. One way to overcome the difficulties in high mass analysis is the direct determination of the mass of each ion. It not only allows mass measurements for very large molecules but also provides the means to analyse the properties of single ions. In recent years different single-molecule methods have been developed for the direct mass determination individual single-molecules.
Figure 1 General principle of a mass spectrometer 8
The main parts of a mass spectrometer are the ion source, the mass analyser and the detector, and the gained data is analysed by a software. In this study, firstly the two main ionization methods for megadalton molecules, ESI and MALDI will be presented. Both methods are used in conventional and single-molecule mass spectrometry as well. In this study conventional MS is defined as the collection of MS methods where the m/z ratios of the ions are measured, and the ion masses are determined from the peaks in the m/z spectrum.
4
In single-molecule measurements the masses of individual molecules are directly determined, either by the simultaneous measurement of the charge and the m/z of ions (e.g. CDMS) or by the direct measurement of the mass with sensors, in which case even neutral molecules can be analysed (e.g. NEMS). After the ionization methods a short description of the different mass analysers and detectors that has been used for megadalton molecules is given. It is followed by the applications of conventional methods for megadalton molecules, including protein assemblies, viruses and synthetic polymers.
Some of these mass analysers can also be modified for single-molecule methods like FT-ICR and QIT, or they can be coupled to cryogenic detectors for the single-molecule analysis. Some instrumentations do not require analysers (NEMS), as they directly measure the electrosprayed molecules, while others function as a combination of an analyser and a detector (CDMS). The different single-molecule methods will be described in detail with their applications in many areas. Special attention will be paid to the different charge-detection MS methods. In the conclusion a comparison of the single-molecule and the conventional methods will be given, showing their advantages, disadvantages, challenges and possible applications.
5
2. Ionization methods for MDa molecules
2.1. MALDI
MALDI is a solid-phase technique used for the ionization of large, non-volatile compounds which are thermally unstable, such as proteins, polymers, large inorganic compounds, etc. It has a good tolerance for salts and other contaminations, and can make MS analysis with a high speed and sensitivity possible. 6 In a MALDI experiment the compound is dissolved in a matrix
containing small organic molecules. These molecules need to have a strong adsorption at the wavelength of the laser so typically highly conjugated molecules are used. 9 The mixture is
dried, and the liquid solvent is removed. As a result, the compound is co-crystallised with the matrix in a way that the analyte molecules are isolated from each other. The matrix needs to be chosen very carefully for the measurements. It needs to be non-volatile and must absorb the energy at used the laser wavelength. The matrixes usually dissolve in the same solvent as the sample. A good crystallization is crucial otherwise no signal would be detected. 9
The now solid solution is irradiated with a laser under vacuum conditions. The matrix absorbs the energy of the laser and then transfers it to the analyte molecules inducing desorption from the surface. The ionization mechanism is still not fully understood but the most widely accepted theories are that the proton transfer occurs in the solid phase before the desorption or that it happens in the gas phase from photoionized matrix molecules. 1
Figure 2 Diagram of the principle of MALDI 1
Differing from other ionization methods such as ESI, a big advantage of MALDI is that it produces singly charged ions even at higher mass/charge ratios which makes it easier to determine the molecular weight of the analyte even in the case of complex mixtures. 10 MALDI
has an increased sensitivity compared to other laser ionization methods. One of the reasons of this higher sensitivity is that the matrix molecules appear in a way higher number than the analyte molecules. This way the analyte molecules are isolated which prevents the formation of clusters that could inhibit the appearance of molecular ions. 1
6
Another reason is that by absorbing the incident energy, the matrix minimizes the sample damage caused by the laser pulse and makes the energy transfer from the matrix to the analyte more efficient. Since instead of the analyte it is the matrix that absorbs the laser pulse there is no need to adjust the wavelength to the different molecules which makes MALDI a fairly universal method. MALDI can be combined with different analysers, such as time of flight (TOF), ion trap or Fourier transform mass spectrometers. 1 In conventional methods MALDI
ionization in the megadalton range was applied only in a few cases, for polystyrene and polyphenylene dendrimers up until 1,5 MDa. In these cases a pulsed nitrogen laser was used.
2.2. ESI
Compared to MALDI, ESI is a softer ionization method which makes it more appropriate for labile substances such as supramolecular assemblies that are held together with noncovalent bonds, polymers with fragile end-groups, etc. 6 It is an aerosolization technique to generate
gaseous ionized molecules from liquid solutions by creating a fine spray of highly charged droplets in a strong electric field. 9 ESI creates multiply charged ions which makes the
interpretation of the MS spectra more difficult. On the other hand this method makes possible the analysis of molecules with high molar mass, that in a single charge state might not be detected. 6 Since it is a liquid ionization method it is compatible with chromatographic or
capillary electrophoresis methods.
During ESI the analyte is dissolved in a solvent which flows through a capillary tube in atmospheric pressure. A strong electric field is applied to the flowing solvent which at the end of the capillary induces a charge accumulation on the liquid surface. The electric field is the result of an applied potential difference between the capillary and a counter-electrode. At first the forming drop at the end of the capillary appears as spherical, then with the increasing voltage the surface is destabilizing, and the drop elongates under the accumulated charges. When the surface tension breaks, the drop changes to a so called “Taylor cone”, and then aerosol spray of highly charged droplets is formed. To limit the dispersion of the droplets a gas is injected coaxially to the spray at a low flowrate. To remove the remaining solvent molecules, the droplets are passed either through a heated inert gas (usually nitrogen) or through a heated capillary. 1
After leaving the capillary the charged droplets are accelerated towards the counter-electrode. The droplet shrinkage is the result of two factors: the evaporation of the solvent molecules and the disintegration of the droplets by Coulombic explosions. 9 As the droplets are
7
solvent molecules in the droplets. With the shrinkage, the surface area decreases but most of the mass remains in the droplet so the charge per surface area ratio increases and the charges at the surface area get increasingly close. This leads to the Coulombic repulsion of the charges which disintegrates the droplet into smaller droplets, because in this state bigger surface area can be achieved. 9 The solvent molecules in these small droplets continue to evaporate until the
electric field becomes large enough on the surface for the desorption of the analyte ions. 1
Figure 3 Diagram of the principle of ESI 9
Since the desorption of ions always occurs from the surface, the method is more sensitive for the compounds that has a higher concentration at the surface. In the case of mixtures, the compounds which are present on the surface at a higher concentration can mask those who are more soluble in the bulk. If the droplet contains large molecules, instead of desorbing, the molecules are freed by the solvent evaporation. While smaller molecules mostly produces singly charged ions, the bigger ones (like proteins) become multiply charged. 1
A variant of ESI is nano-ESI which we get by scaling down the flow rates to a few tens of nanoliters per minute. It is the most commonly used ionization method for native MS researches. 11 Nano-ESI has many advantages compared to normal ESI, it is more tolerant to
salts and has an enhances sensitivity. Nano-ESI is capable of evaporating the solvent in less harsh conditions so it can be directly used with aqueous solutions without the need to add volatile organic solvents for the analysis. 12 ESI can be coupled with different mass analysers,
like quadrupoles, ion traps, orbitraps, etc. These instrumentations can also be used as tandem mass spectrometers.
8
For labile substances ESI would be more applicable than MALDI, because of its softer nature, but in conventional methods it poses a challenge to resolve the spectrum of highly charged ions. In case of single-molecule methods this problem does not appear, most of these measurements are made with ESI. MALDI is more applicable in the conventional ways to produce ions with fewer charges, enabling a better charge resolution. This however reaches a high mass limit around 1,5 MDa, since the high m/z ions are difficult to detect with the usual detectors. In single-molecule methods MALDI is mostly used with cryogenic detectors. The following table represents how often were these ionization methods used in the reviewed papers.
ESI MALDI Other
Conventional MS 9 2 -
Single-molecule MS 29 6 6
9
3. Instrumentation of conventional MS
3.1. Mass analysers
3.1.1. TOF
Time of flight (TOF) is the most commonly used mass analyser for MALDI ionisations. It can analyse over an unlimited mass range and since it requires that the ions be produced in bundles, it is well suited for pulsed sources such as MALDI. It is also possible to analyse ions from continuous sources if they are introduced into the analyser in a suitable way, e.g. with orthogonal acceleration. It can be used in linear or in reflectron mode. In linear mode the ions are first accelerated in an electric field, then separated according to their velocities in a field-free flight tube before reaching the detector. In the electric field all ions acquire the same kinetic energy, so their velocity depends only on their mass. The ions are separated in time, the mass-to-charge ratio is determined by the time needed for the ions to move through the field-free region to the detector. 1
Figure 4 Principle of a linear TOF instrument 1
In reflectron mode the ions are reflected back so they pass through the flight tube twice. The reflectron, which acts as an ion mirror is positioned at the end of the field-free region and the detector is placed at the source side. The reflectron corrects the kinetic energy differences for ions with the same mass-to-charge ratio. The ions with higher velocity will penetrate the reflectron more deeply so they spend more time there and reach the detector at the same time as the slower ions with the same m/z. In this method the increased flight path increases the mass resolution, however the sensitivity is decreased and the mass range is limited. 1
10
Figure 5 Principle of a TOF instrument in reflectron mode 1
3.1.2. Quadrupoles
Quadrupole analysers separate ions according to their m/z ratios in oscillating electric fields using the stability of trajectories. The ions are electrically accelerated through four parallel metal rods on which AC and DC potentials are applied. After entering the quadrupole, the positively charged ions will move towards the negatively charged rod and vice versa. With the change of potential, the direction of the ion trajectory will change as well. Ions with stable trajectories will reach the detector, others with unstable trajectories will discharge on the rods. This way only ions with a specific m/z will reach the detector. 13
In mass selective/mass filter mode the applied voltage is increased so the analyser can scan over the entire m/z range. Quadrupoles can also serve as ion guides in RF-only mode. In this case RF voltage is adjusted so ions across a wide mass range can pass through. 14 Since the
efficiency of focusing is inversely proportional to the m/z, ions with higher masses than the monitored mass range are weakly focused and may crash into the rods. On the other hand, the ions below the mass range will be lost because of their unstable trajectories. Focusing of ions with high masses can be more efficient by using an increased RF voltage, however it will also increase the lower observable m/z limit. 1
Multipole guides (hexapole or octapole) can also be used in the place of quadrupole guides. They operate in a similar way but while they are very good at the transmission of ions, they are unable to do mass analysis. 14
11
Figure 6 Instrumentation of a quadrupole analyser 1
It is also possible to use several quadrupoles in series. The most common instrumentation is the triple quadrupole, in which case the analysis is done by three quadrupoles. By introducing a collision gas into the central one in RF mode, it can act as a collision cell, making tandem mass spectrometric measurements possible. This setup can be used in several ways, in fragment/product ion scan, parent/precursor ion scan or neutral loss scan mode. 1 The third
quadrupole can be replaced with a TOF analyser to create a hybrid quadrupole-time of flight instrument (Q-TOF). This setup is preferred in many cases thanks to its higher sensitivity, mass resolution and accuracy. 15
Figure 7 Principle of a Q- TOF instrument 14
3.1.3. Ion trap
In ion trap instruments ions are stored by using an oscillating electric field. The traps can be classified as 2D or 3D, depending on the dimensions that the ions are trapped in. 1 A 3D or
a quadrupole ion trap (QIT) uses an RF-field to trap the ions in a system of hyperbolic electrodes. This system consists of a central ring electrode and two end cap electrodes. It works as a 3D version of the quadrupole mass filter, where the ions are driven back towards the centre of the trap by the restoring forces of the electric field. 16
12
All ions are present inside the trap, and the spectrum is obtained by expelling the ions according to their masses. During trapping the ions follow a stable trajectory by rotating along the r axis and oscillating along the z axis. The trapped ions repel each other which would expand their trajectories and result in ion loss. To avoid this expansion loss, helium gas is used to maintain a given pressure and remove the excess energy by collision. 1 The spectrum is usually
obtained by operating the trap in mass selective instability mode. In this mode the amplitude of the applied RF-field is increased to increase the moving along the z axis, until the trajectory of the ion becomes unstable and the ion gets ejected through a hole in the end-cap electrode. 16
Figure 8 Schematic of a Q-IT instrument 16
3.1.4. Orbitrap
An orbitrap is an ion trap instrument without RF or magnet fields. The ions are moving around a central electrode while end-electrodes create potential barriers to axially confine the ions.17 The moving ions create an image current which is detected on receiver plates. The m/z
values of the ions are determined from the frequencies given by the Fourier transformation of the signal. The ions are injected into the orbitrap from a curved linear trap (C-trap) where they are stored and cooled beforehand.14
13
3.1.5. FT-ICR
In a Fourier transform ion cyclotron resonance (FT-ICR) analyser ions are stored in a particle accelerator known as cyclotron using combined electric and magnetic fields. 17 The
ions move in a circular motion perpendicular to the magnetic field at a frequency dependent on the m/z ratio. This characteristic frequency is called the cyclotron frequency. Irradiation by an RF pulse increases the kinetic energy of the ions forcing them to trajectories with larger radius. The ions circulating in the cell gain phase coherency which creates a detectable image current. After Fourier transformation of this current the m/z of the ions are determined from the frequencies. This method gives a very high resolution which could be further improved by increasing the magnetic field strength. 1,13
Figure 10 Two possible setups of the FT-ICR instrumentation 17
3.2. Detectors
Several detection modes have been devised to the different analytical applications and instrumentations. Currently the most widely used detector in conventional MS is the electron multiplier (EM). In this detector an electrode called conversion dynode is held at a high potential opposite to the charge on the ions. A charged particle striking into the conversion dynode causes the emission of secondary particles. When a negative ion strikes the conversion dynode, positive secondary ions are formed; when a positive ion strikes, negative secondary ions and electrons are formed. The charged secondary particles are converted into electrons at the first dynode. The formed electrons are then directed towards a second dynode by an applied potential releasing more secondary electrons. Using several dynodes, the electrons can be amplified by a cascade effect, so a single ion can generate a very large current, resulting in a high sensitivity. 1,13
14
Figure 11 Schematic diagram of the electron multiplier 1
Electron multipliers can be designed not only as discrete dynodes but as continuous dynode as well. Channeltron is a type of the continuous-dynode electron multipliers (CDEM). It is made of a lead-doped glass in the shape of a curved tube with good secondary emission properties. In this setup the cascade effect of the electrons is created by skipping across the inner surface of the detector. At the detector exit a metal anode is used to collect the stream of secondary electrons to measure the current. 1
Another type of CDEMs is the microchannel plate (MCP). In this detector parallel cylindrical channels are drilled in a plate. A semiconductor substance is responsible for the electron multiplication by covering the channels and giving off secondary ions. Since the path of the secondary electrons in the channels is very short, the response time of the detector is very fast. 1
Figure 12 Schematic diagram of a channeltron (left) and the cross-section of a MCP with the electron multiplication in a channel (right) 1
15
4. Applications of conventional MS methods in the MDa range
ESI and MALDI ionization methods gave way to the mass determination of molecules in the MDa range which makes the investigation of biomacromolecules, such as viruses or large protein complexes possible. High mass analysis however needs to overcome many difficulties, such as the heterogeneity of the sample or the charge state resolution. The most commonly used mass analyser for MDa molecules is Q-TOF because of its unlimited mass range and high sensitivity. Although it is not easy to solve the problems of high mass analysis, there are several studies about the MS analysis applied to MDa molecules.
One of the first MS studies in the MDa range was made investigating bacteriophage MS2 in intact form using ESI-TOF. 18 To maintain the intact form, dry nitrogen was used for
collisional cooling. Without it the high m/z ions dissociate to monomers and cannot be detected. Although the charge states were barely resolved, they were able to deduce an experimental mass of 2,5 MDa. This study showed that the analysis of unknown protein assemblies with masses higher than 1 MDa is possible with MS methods.
In the same year vanillyl-alcohol oxidase (VAO) protein assemblies were analysed. 19 The
gained m/z spectrum shows well resolved charge-states for all detected multimers. VAO is predominantly in its octamer form with a mass of 0,5 MDa and it showed in the spectrum with high abundance peaks. The 1 MDa 16-mer of VAO was also detected with a narrow charge state distribution and zooming into the spectrum showed that even the 1,5 MDa 24-mer form was detectable.
Figure 13 ESI spectrum of VAO (left), zoomed in part for the 16-mer and 24-mer forms of VAO (right) 19
16
ESI-QTOF was used to study the different multimers of hemocyanin (Hc) found in a deep sea crab, the Bythograea thermydron (ByTh). 20 Hc is an extracellular blue protein that acts as
an oxygen carrier and contains copper at the oxygen-binding site. It builds up of 75 kDa monomers and is usually found in its hexamer form. The m/z spectrum showed that it appears in 4 different forms in ByTh: monomer, 6-mer, 12-mer and 18-mer. The highest measured mass was 1,35 MDa of the 18-mer. In a later study Hc multimers found in 3 different species were analysed with ESI-MS and compared. 21 Mass analysis of the samples was done in both
denaturing and non-denaturing conditions. The spectrum from the denaturing measurement was used to determine the heterogeneity of the samples. The 3 investigated species were composed of 4, 5 and 9 different polypeptide chains respectively. In the non-denaturing measurements the different oligoforms were shown. The greatest variety was found in ByTh. Compared to the previous study even the 24-mer (1,8 MDa) and 30-mer (2,23 MDa) forms were detected.
Figure 14 ESI-MS spectrum of Hc in ByTh under non-denaturing condition 21
Two geometries of hepatitis B virus (HBV), composed of 90 (T=3) and 120 (T=4) dimers were analysed with high resolution MS. 22 Two variants of cp149 were observed, and both
spectrum showed very similar results. In the spectrum two well-resolved sets of ions were shown, which yielded the masses of 3 and 4 MDa respectively, clearly belonging to the T=3 and T=4 capsids. Another HBV capsid, the cp183 was analysed as well but in this case the spectrum was poorly resolved. Based on the spectrum of the cp149 capsids, they estimated the mass of this capsid between 5-6,5 MDa. The stoichiometry and stability of the capsids were further investigated by tandem MS.
17
Figure 15 Spectrum of the two cp149 capsids (left), spectrum of the cp183 capsid (right)22
The assembly and the stability of Norwalk virus-like particles (NVLP) were monitored by MS as a function of pH, ionic strength and VP1 protein concentration. 23 Native MS at
physiological pH showed only one complete capsid (T=3), composed of 180 copies of VP1, with a mass of 10 MDa. The study showed that while NVLP is stable in acidic conditions, at higher pH it disassembles into intermediates, mostly to species containing 60 and 80 copies of VP1 and weighing 3,4 and 4,5 MDa respectively. It was also shown that formation of these intermediates is reversible, and they can be reassembled into higher oligomer forms depending on the pH and the ionic strength.
The genome content of Cowpea Chlorotic Mottle Virus (CCMV) and Brome Mosaic Virus (BMV) was investigated by native MS using both Q-TOF and orbitrap mass analysers. 24 The
contained capsid proteins are very similar, they share a 70% sequence identity. The particles which cause the infectivity are the same in the two viruses: RNA1 and RNA2 separately and RNA3 and a sub-genomic RNA4 as co-package. The Q-TOF spectrum was well resolved for the charge assignment for RNA2 and RNA3+4, but not so much for the RNA1. All of the measured masses were around 4,5-4,7 MDa. In the orbitrap measurements the particles retained more charges which made the particle identification more difficult. Adding triethylammonium acetate (TEAA) to the sample reduced the charges and moved the ion signals to higher m/z. This way the masses could be assigned to the RNA2 and RNA3+4 particles, but it was not possible to resolve the RNA1. To identify the RNA1 particle, tandem MS experiments were made.
18
Although most high mass analyses are done by Q-TOF, studies were also made by orbitrap analyser as well. A modified orbitrap instrument was used to measure several virus-like particles. 25 All spectrum showed well-resolved charge state distributions. The determined
masses of these particles were between 2-4,5 MDa. The highest mass analysed with conventional MS was 18 MDa for bacteriophage HK97. 5 The T=7 capsid assembly containing
pentameric and hexameric capsomers were analysed with Q-TOF. The study concluded that the achievable mass limit with conventional MS is around 20 MDa.
Figure 16 Spectrum of virus-like particles with a modified orbitrap 25
MS measurements in the MDa range require homogenic samples. Virus capsids are good examples of this, since they are built up of identical proteins and possess only limited heterogeneity. Conventional MS methods could also be applied to polymers with high purity and extremely narrow molecular weight distributions. However, since only perfectly synthetized polymer standards comply to these requirements, there were very few studies made for the characterisation of high mass polymers.
Opposed to the preferred ESI methods for biomacromolecules, in the case of polymers usually MALDI is used for the ionization. Polystyrene standards with very small polydispersity were measured using MALDI-TOF MS. 26 The molecular weight of the standards ranged from
86 kDa to 1,5 MDa. With the increasing mass the spectrum become dominated by the multiply charged ions. In the spectrum of the 1,5 MDa standard the singly charged principal distribution does not show at all, only the doubly and triply charged ion distributions appear with very bad sensitivity.
19
A homologue series of polyphenylene dendrimers were investigated as well with MALDI-TOF up to the ninth generation. 27 In this study instead of MCP detection, a Bruker HIMAS
detector was used. This was the first experiment on analysing MDa polymers other than polystyrene. The G9 dendrimer seemed to be the limit of this measurement. The different charge distributions could barely be distinguished from the base signal, but the higher charged ions could still be used to determine a molecular weight of 1,5 MDa, which had more than 20% deviation from the theoretical mass. It can be concluded that measurements of synthetic polymers with masses higher than 1 MDa is a great challenge with conventional MS and requires the implementation of other methods.
Figure 17 Spectrum of polystyrene with a nominal mass of 1 MDa (left)26, spectrum of the
20
5. Single-molecule MS methods
Advancements in mass spectrometry has allowed the measurement of ions with higher and higher masses, however it seems that the limit of conventional MS has been approached. In conventional MS the mass-to-charge ratios (m/z) of the ions are measured and then their masses are determined from the peaks in the m/z spectrum. 7 The highest mass limit with this approach
has been estimated around 20 MDa by Heck who analysed virus capsids with nano-ESI. 5
To reach the MDa range with conventional MS the sample usually need to be homogeneous, for heterogenous samples the mass limit is significantly lower. Two types of heterogeneity can be distinguished: intrinsic and extrinsic. Samples with extrinsic heterogeneity have a single well-defined mass, however they appear heterogenous. It can be the result of incomplete desolvation, adduct formation in solution with impurities or different types of counter ions. By extensive purification of the sample the remaining solvent, the adducts and the counter ions can be removed, and well-resolved m/z spectrum can be obtained. In intrinsically heterogeneous samples the analytes do not have a single, well-defined mass. In this case the number of masses present is much higher than the number of ions in the sample, the charge states cannot be resolved. Polymers, nanoparticles, cells and aerosols can be examples of this heterogeneity. Generally, the larger the object, the more likely it is to be intrinsically heterogenous. 28
Many approaches have been considered to overcome the limitations of conventional MS. One solution is to directly determine the mass of each ion by performing single-molecule measurements. These measurements are not limited by charge-state resolution so the analysis of ions can be extended to a broad range of mass and heterogeneity. Single-molecule MS can provide valuable information that would not have been accessible by ensemble measurements. If a large number of ions are measured, then a mass distribution can be constructed which can be used to gain information on the sample heterogeneity and its origin. 28
Several methods have been developed for single-molecule measurements. Some techniques use charge stepping, in which case a very accurate m/z of a single ion is measured, then after shifting the charge the m/z is re-measured for several different values to deduce the mass. Charge-stepping measurements can be performed by FT-ICR or QIT with optical detection. 28 Among the single-molecule methods, these are the most accurate. Since the ions
can be trapped indefinitely in these instruments, their properties can be studied over a long time period and the differences from ensemble average can be investigated. However, given that
21
these measurements are time consuming, they are not well suited for the determination of mass distributions. 7
Another way to perform single-molecule measurements is with nanomechanical oscillators. When single molecules accrete on the oscillator, the resonant frequency of the device is changing. The mass of the molecules can be measured by constantly monitoring the changes in the resonant frequency. This method is independent of the charges so it can be used to measure the masses of both ions and neutral molecules, however these measurements are not as accurate as other single-molecule methods. 7
In the last class of single-molecule methods the m/z and charge are measured simultaneously and then multiplied to determine the mass. This can be done with FT-ICR where the magnitude of the induced charge on the detector plates is used to directly determine the charge of the single trapped ion. The induced charge however depends on the ion trajectory in the trap as well, so the charge determination is not accurate. 29 QIT devices can also be used for
simultaneous measurements. In axial instability mode the m/z is determined by the ejection of the ion from the trap, and when this ion strikes a detector plate the charge can be measured. This method does not give an accurate charge detection either. 7 Another technique of this class
is kinetic energy-dependent cryogenic detectors coupled to a TOF analyser. The challenge in this method is that the detector response is not linearly proportional to the charge which makes the charge determination difficult, especially for highly charged ions. 30 The final method in
this class is charge detection mass spectrometry, which makes the measurements based on the induced charge of the ion without the need to know the ion trajectories.
5.1. Optical detection with QIT
High mass species can be measured by trapping the single ions in 3D QIT systems and detecting their scattered light. In the trap large particles follow a relatively slow, so called Lissajous-like trajectories with small superimposed micromotions. Particles up to µm-size can be trapped and their scattered light can be visually observed with a microscope or detected by a CCD camera. Detecting the scattered light of trapped ions is a very precise ways to measure their m/z. The intensity of the scattered light greatly reduces for particles with smaller diameter.28 Nie et al. set a low detection limit of 50 nm while measuring monomeric virus
particles with this method. 3 This low limit excludes many biologically important species from
the analysis, such as protein complexes and many viruses. There are multiple ways to make measurements with optical detection. In early QIT measurements a small AC signal was applied to the end cap electrodes and the secular motion of the particle was visually observed. 31 When
22
the secular motion was in resonance with the signal, the applied voltage and frequency was used to determine the m/z. A single m/z however is not enough to determine the mass.
High mass particles are affected by the gravity, which forces their oscillation below the centre of the trap. The effect of gravity on the trajectory can be corrected by applying a DC bias on the top and bottom electrodes. The applied voltage provides another way to measure the m/z. Using electron-stepping, the absolute mass and charge of the trapped ion can be determined. 28
UV radiation or electron guns can be used to eject electrons and by this, changing the charge of the particle. The m/z can be remeasured by adjusting the DC bias on the electrodes. Repeating this process and measuring the m/z several times results in a more reliable mass determination.28
Figure 18 Photographs of a single ion trajectory
a) perpendicular to the axis of the end caps b) along the axis of the end caps28
A third and more accurate method to measure the m/z with optical detection is based on the trajectories of the particles. Projecting the trajectory onto the perpendicular plane shows an elliptical motion with superimposed micromotions. The elliptical Lissajous motion is caused by the secular frequency, while the micromotions are the result of the driving frequency. When the ratio of the driving frequency to the secular frequency is an integer, then the trajectory appears as a star pattern along the axis of the end caps. Measuring the applied voltage and frequency, the number of points on the star can be used to determine the m/z. By changing the drive frequency, the absolute mass can be determined using the electron stepping method. 28,32
The Cheng group has made several researches applying this detection method. They measured the absolute mass of whole E. coli K-12 bacteria cells ionized by MALDI, using the star pattern of the QIT. 4 The electron-stepping was done with an electron gun which produced
one-electron differentials in the charge states. They measured 60 E. coli cells and determined an average of 50 GDa. This was the first high-precision study of intact microorganisms in the higher than 10 GDa mass range.
23
Figure 19 a) plot of the star branch number (n) vs a function of the frequency and voltage, b) plot of the mean mass determined from different charge states vs the resulting
standard deviation 4
The group also investigated the mass distribution of two standard NIST polystyrene particles. 33 In these measurements instead of MALDI, the ionization was done by laser-induced
acoustic desorption (LIAD). 10 particles of each standard were measured, with the average mass of 2,38 ∙ 10 Da and 6,2 ∙ 10 Da respectively. The measured masses were in the expected range of the theoretical mass. Using LIAD ionization they also studied bioparticles covering the 109-1014 Da mass range. 34 This study involved the measurements of vaccinia
viruses, human red blood cells and the remeasurement of E. coli K-12 cells. The method was able to distinguish the normal red blood cells from the anemic cells which indicates that the method could be used for the detection of cancer cells.
Another approach for the m/z measurements is to apply a constant frequency and voltage during the measurements. Schlemmer et al. designed a QIT with eight cylindrical rods instead of the ring electrode to increase the collection of the scattered light. 35 A laser was used to
illuminate the trapped particles and the scattered light was detected by an avalanche photodiode (APD) and a CCD as a second detector. An applied DC voltage compensated for the gravitational force so the particles could move around the centre of the trap. The motion frequencies of the ions were analysed by FFT. This method was used to study monodisperse 500 nm SiO2 particles. Nie et al. also used an FFT based method for the measurements of whole
viruses. 3 In their setup a cylindrical ion trap (CIT) was used with transparent end-caps was
used.
Light scattering mass measurements of nano- and microparticles are possible with high precision in QIT, however these measurements usually take a long time. Analysing the motion of a trapped ion can take tens of second and sometimes the trap must be pumped out before the next measurement. The electron-stepping which is necessary for the mass determination is
24
usually done by sufficiently inefficient methods. It results in only one or a few charge stripping at a time, so it can take several seconds to finish the electron-stepping. 32 Depending on the m/z
measurement method, it might also be needed to reoptimize the settings after each charge stripping. In previous works mass distributions were studied by measuring tens of particles. In case of more heterogenous samples or for better accuracy more measurements might be necessary. The long analysis time has some advantages as well. It provides an opportunity to track the mass and charge of the ions over time and with this gain information on the properties of gas phase nano- and microparticles. 28
5.2. Single-molecule FT-ICR
Single-molecule studies with FT-ICR are based on the same method as the QIT charge-stepping method. First the m/z of a trapped single ion is measured, then after changing either the mass or the charge of the molecule, the m/z is remeasured. Repeating the cycle several times, the absolute mass of the ion can be calculated. In this case the charge-stepping is done by a gaseous reagent, which is pumped out before the measurement. Mass change can be done by adduct formation. The first measurements were done on PEG molecules around 5 MDa. The method was able to monitor the reaction steps of the ions for a long time period. 36,37
Figure 20 Signals from the measurement of a single PEG ion: stack plots of successive FT across the time-domain data (left), m/z-time plot during the charge loss (right) 37
The method was applied for electrosprayed plasmid DNA with a mass of 1,95 MDa. 38 The
charge was changed through proton transfer by introducing gaseous acetic acid. The addition and elimination of the gas-phase acetic acid was also monitored. Similarly to QIT methods, the FT-ICR measurements of single ions are time-consuming either by charge-stepping or mass change. The introduced gas must be pumped out before every analysis, and an accurate measurement requires several charge or mass stepping. Therefore this technique is more appropriate for the investigation of single ion properties over a longer time period rather than for the measurements of mass distributions. 28
25
A different approach with FT-ICR is the simultaneous determination of the charge and the m/z. The mass of the ion can be determined by their multiplication. To determine the charge, the voltage drop on the detection circuit is measured several times at increasing excitation levels. After a measurement at a certain excitation level is done, the cyclotron motion is dampened before the next excitation. If the detection circuit capacitance, response, and the ion cyclotron radius are well-defined, then the charge can be calculated from the highest measured voltage drop. The difficulty of this method is the determination of the ion cyclotron radius which is a parameter that is needed for the charge determination. The highest voltage drop belongs to the highest excitation level where the cyclotron radius reaches its maximum possible value before the ion collides with the electrode. Based on the measurements it was assumed that the ion radius at that excitation level is the 95% of the cell radius. 28,29 Electrosprayed PEG
ions around 4 MDa have been measured using this technique. 29 The method was also applied
to 110 MDa Coliphage T4 DNA ions, which is the highest mass analysed by FT-ICR. 39 In this
case the cyclotron radius was determined in a different way, using pulse sequences and measuring the peak height several times after Fourier transformation. Although in a different way, they also found that the radius is 95% of the cell radius.
Figure 21 Pulse sequences for a) ion trapping and preselection b) ion remeasurements Similarly to the charge-stepping method, the simultaneous charge and m/z determination also requires multiple measurements which makes it time-consuming as well. Because of the difficulty of the cyclotron radius determination, these measurements had an error of 10%.
26
5.3. Nanomechanical mass spectrometry
In nano(electro)mechanical mass spectrometry (NEMS) the mass of a particle is directly determined without the need to measure the charge or the m/z. In NEMS measurements the particles are introduced into the device mostly by ESI which is followed by desolvation and then they are entering the detector. The detector in these measurements is an ultrahigh-frequency nanoelectromechanical resonator. The resonant ultrahigh-frequency of a NEMS resonator is a delicate function of its full mass, even small mass changes, like the adsorption of a single molecule can measurably alter it, causing a frequency shift. In case of an adsorbing molecule it usually shows as a decrease in the frequency. The extent of these frequency shifts depends on the mass of the resonator and the adsorbing molecule. The position of the particle also influences the measurement, a larger shift can be detected if the adsorption happens to a higher-amplitude location on the resonator. It is also possible to make real time measurements, and by the monitoring of successive frequency shifts individual events can be detected. The disadvantage of this method is that only a small fraction of the electrosprayed particles can adsorb to the small surface which makes the sensitivity of the method quite low. The NEMS detector is applicable to a nearly unlimited mass range from a few Daltons to the GDa range.11,28
There were several approaches using micro- and nanomechanical oscillators for the detection of single molecules. One of the first attempts was by Ilic et al. in 2001 using low-stress silicon nitride resonating cantilevers of varying lengths.40 They investigated the binding
of single E. coli cells on the cantilevers coated by E. coli specific antibodies. The resonant frequency was measured before and after the coating and remeasured after the immersion into a solution containing E. coli cells. The position of the adsorbing cells was determined by both atomic force (AF) and scanning electron microscopy (SEM). The measurements were taken in air which caused vibrational damping and resulted in broad frequency bandwidth for the resonator. 28
27
Figure 22 a) Cantilever after antibody immobilization and cell binding b) Vibrational spectra of the resonator before (black) and after (dark grey) antibody immobilization and
single cell binding c) Frequency shifts of E. coli cells 40
Another approach for measuring single molecules is to use hollow cantilevers in vacuum.41
In this method a solution of the analysable particles flow through the cantilever whose resonant frequency is constantly monitored. The detected frequency shift depends on the mass and position of the particle and reaches its highest value at the cantilever tip. This value can be used to determine the mass excess of the particle compared to the displaced buffer. Based on the mass excess and the densities of the particle and the buffer, the absolute mass of the particle can be calculated. Hollow cantilevers can also be used for binding measurements. Bound and unbound particles both increase the mass, however those that can bind to the wall will accumulate inside the channel. The bound particles then can be detected by immobilized receptors.
Figure 23 The two measurement modes of hollow cantilevers: mass determination by peak frequency shift (left) and by immobilized receptors (right) 41
The flow-through method was used to measure gold and polystyrene particles, as well as bacteria cells. Several measurements were made for each case and histograms were constructed
28
to determine their mass. The binding method was used to measure the binding of goat anti-mouse IgG molecules to immobilized antibodies in different concentrations. These studies showed that hollow cantilevers can be used for the mass measurements of nanoparticles and bacteria as well. Since these measurements were done in vacuum, the vibrational damping of air did not affect the results. 41
Figure 24 a) Histograms of the peak frequency shifts b) Frequency shifts of goat anti-mouse IgG 41
One of the challenges in NEMS measurements is locating the adsorbing molecule, since its position effects the frequency shift. SEM can be used for this, but it also adds to the uncertainty of the method. Ensemble measurements might be a solution for this problem. Naik et al. made investigations using NEMS to make a probability histogram of the frequency shifts and then calculate an average particle mass and the width of the distribution. 42 This method was applied
to 2,5 nm gold particles and BSA molecules as well. However, since this technique includes averaging several measurements, it cannot be considered a single-molecule method. 28
The first real-time single-molecule measurement with random adsorption was based on the simultaneous measurement of the resonant frequency of multiples modes of the NEMS resonator. 43 The frequency jumps that the adsorbed particles induce are time-correlated.
Simultaneous frequency jumps of two different modes indicate a single-molecule adsorption. In order to determinate the mass and position of the particles, the frequencies of the first two vibrational modes of a doubly clamped beam were measured. 5 and 10 nm gold nanoparticles and human IgM antibody measurements were made with this method. The measurements showed that a spectrum can be built up from the individual adsorptions, allowing the real-time measurement of single particles. The spectrum of IgM isoforms clearly showed the individual pentameric complex as the highest peak around 1 MDa and its dimer around 2 MDa. The spectrum was decomposed to show the different polymerization levels of IgM.
29
Figure 25 Frequency shifts of the two modes (left), acquisition of mass spectrum for human IgM (middle), decomposition of the IgM spectrum (right) 43
The possible analysis of neutral particles by NEMS was investigated by Sage et al. 44
A hybrid setup was used to analyse neutral and ionized tantalum nanoclusters in the MDa-range. In this setup NEMS and TOF-MS measurements were made simultaneously and then the mass spectra of NEMS and the m/z spectra of TOF-MS were compared. The spectra made by the different techniques showed great similarities which demonstrated that NEMS measurements are insensitive to charge and the mass measurements of neutral molecules is possible with NEMS. 44
Figure 26 Comparison of TOF-MS and NEMS spectra of tantalum nanoclusters 44
The results also showed that while in current MS systems the resolving power decreases with the increasing of the mass, the mass resolution in NEMS remains constant over the whole mass range and so the relative uncertainty for heavier particles is lower. It is also apparent that in the TOF-MS spectra there two peaks present in every mass distribution. It can be attributed to the multiple charge states of the ion which is a common occurrence in current MS methods.
30
However, the NEMS spectra provides a clear, single peak which shows that nanoparticle populations can be represented by a well-defined peak corresponding to a single distribution. NEMS measurements can be useful in avoiding information loss caused by peak overlapping for multiply charged species. It also provides the means to analyse compounds in their native state, regardless of their charge. 44
If the adsorbing particle is not spherical then its stiffness and orientation also affect the resonant frequency, a stiff particle can even increase it. Based on this feature, not only the mass, but the mechanical proprieties of the particles could also be measured. 28 To accomplish this,
the simultaneous frequency measurements of multiple vibration modes are needed. Malvar et al. was able to determine the mass and the stiffness of gold nanoparticles and E. coli bacteria cells by measuring the frequency shifts of four vibration modes of the resonator. 45
Figure 27 Frequency shifts of four vibrational modes (left), mass spectra of E. coli bacteria with and without of the stiffness effect (right)
Considering the small size of the resonators, only a small fraction of the particles can be adsorbed, most of the particles will not be detected. Since the neutral particles are more difficult to focus, an even smaller fraction of them can adsorb. Only a part of the resonator can be effectively used, because the frequency shift produced by particles adsorbing at low-amplitude regions are usually not large enough to be detected. The effective area could be increased by using an array of oscillators. 28
This theory was proven very recently by arrays of 20 nanomechanical resonators measuring metallic aggregates in the MDa range. 46 The resonators in the array are interconnected via two
metal levels and each one of them is individually addressed by a distinct resonance frequency. Each detector is operated simultaneously on its first two resonance modes. The frequencies of the resonators are monitored sequentially over time by operating the array in closed loop. Initial frequencies and phase references are recorded for each resonator. A phase lock loop (PLL) locks onto a certain resonator, measures its current frequency and after a given time switches
31
to the next. Three tantalum nanocluster population was analysed with this method and compared with a simultaneous TOF measurement. Their results are very similar to their previous study, but the capture efficiency is highly improved due to the array. Imaging at the single-particle level was also attempted with this configuration. The array was moved to scan the 4 cm diameter particle beam to obtain spatial mapping for two locations in the beam. Each resonator obtained a spectrum with 4 min acquisition time which was used to make an interpolated surface map.46
In a different study, arrays consisting of 20 resonators and operating in the same principle were used to measure virus capsids above 100 MDa. 47 The frequency shifts of two vibration
modes were detected and a mass distribution was constructed for empty and filled T5 bacteriophage capsids.
Figure 28 Imaging of a particle beam when the array is a) in the centre, b) at the edge 46
5.4. Cryogenic detectors
Cryogenic detectors are one of the solutions developed for the poor detection efficiency problem of MCPs in the analysis of high m/z ions, usually coupled to TOF-MS. Cryogenic detection methods depend on the kinetic energy of the particles, instead of their velocity. The ions gain their kinetic energy by their acceleration through a fixed potential difference, so ions with the same charge will have the same kinetic energy. As a result the detection efficiency will not depend on the m/z. The method is also able to detect neutral molecules. One of the disadvantages of cryogenic detectors is that is requires the system to be cooled to <4 K. These
32
detectors are also much slower than MCPs and have a small active area. There are three types of cryogenic detectors: microcalorimeters, superconducting tunnel junctions (STJ) and superconducting stripline detectors (SSLD). Among these methods only microcalorimeters and STJs are capable of charge detection, which makes them applicable to single-molecule measurements. 28
5.4.1. Microcalorimeters
In microcalorimeter detectors an insulating substrate (e. g. Si) is coated with a metal absorber (e. g. Ag) and held at extremely low temperature (<4 K). An ion colliding into the metal surface deposits its thermal energy and heats up the absorber. The sharp rise and decay of the temperature can be monitored by a very precise thermometer. The temperature rise is then used to determinate the deposited energy. 28
Figure 29 Cross-section of a microcalorimeter 48
One of the first measurements with microcalorimeters were done in by Hilton to analyse 14 kDa lysozyme and 66 kDa bovine serum albumin (BSA) ions. 49 The detector was coupled
to a MALDI-TOF spectrometer and a liquid-helium cryostat with an adiabatic demagnetization refrigerator was used to keep the temperature at 4 K. The temperature rise was measured by normal insulator superconductor (NIS) tunnel junctions. In the spectrum the singly charged monomer and doubly charged dimer BSA ions could be successfully separated, however they realised that not all of the kinetic energy was deposited during the ion striking. They plotted the ratio of the deposited kinetic energy to the kinetic energy of each analysed ion and found that BSA and lysozyme ions deposit around 54% of their kinetic energy, while the lighter sinapinic acid which was used as a matrix deposited 72%.
33
Figure 30 a) Detector signal measuring BSA ions, b) Scattered plot of the deposited energy of BSA ions in sinapinic acid matrix, c) Plot of the ratio of the deposited energy to the
kinetic energy of ions 49
After the successful separation of the singly and doubly charged BSA ions, the group investigated if this separation is possible for higher charge states as well. 48 ESI was used to
create BSA ions from +15 to +21 charges. In the m/z spectrum the +16 to +21 charge states are clearly visible, and small peaks are showing for the +15 and +21 charge states. Plotting the pulse heights of the ion strikings, they found that they follow is non-linear dependence. Their calculations found that the pulse height is proportional to z0,4. In a different study of the group,
they found that this non-linearity is an effect of the interactions between the absorber and the ion. 50 In this study a protein mixture of myoglobin fragments and glucagon was analysed. The
results showed that the constant mass lines of the contour plot can be used to increase the separation in complex mixtures.
34
Figure 32 ESI-MS analysis of a protein mixture a) m/z spectrum, b) scatter plot of the signal heights, c) contour plot of signal heights 50
In a recent study the strong influence of the absorber material on the detector response was investigated.51 In this study they also made single-molecule experiments to investigate the
fragmentation pathways of acetone and CH3+ ions. They found that the method is capable of
directly detecting neutral molecules, although it is only applicable for smaller molecules.
5.4.2. Superconducting tunnel junctions
In STJ detectors an insulating layer is used to separate two superconducting layers. The temperature of the detector is held around 1 K, although in this case the temperature controlling does not need to be as precise as in microcalorimeters. An energy bias is formed between the superconducting layers when high-energy ions impact on one of the layers and deposit their energy, breaking millions of Cooper pairs. This bias will induce a tunnelling current between the layers by the liberated electrons, which then can be measured. The detected height of the signal depends on the deposited kinetic energy, so it can be used to determinate the charge. Measuring the flight time of the ions, their mass can be deduced. 28
35
Human serum albumin (HSA) ions were analysed both with STJ and MCP to compare the two techniques. 52 In a single record of the detector signal, it could be seen that the signal height
for the doubly charged ion was about twice as large as the singly charged ion and the dimer. 500 measurements were made, and the results integrated into one spectrum. Although the spectrum showed that the singly and doubly charged ions were not resolved, comparison with MCP demonstrated that the STJ method showed significant improvements in performance and sensitivity compared to the MCP.
Figure 34 Single record of the signal (left), integrated spectrum of 500 measurements (right) 52
The possibility to analyse neutral molecules with STJs was demonstrated in the fragment-analysis of polystyrene and BSA molecules. 53 Neutral and ionic fragments produced by
post-source decay (PSD) were detected. The highest detected fragment was the 1 MDa pentamer of the polystyrene particle.
The active area of cryogenic detectors is very small, but it can be increased by using an array. A 4x4 array of STJ was developed in order to measure proteins in the megadalton range.30
Each detector in the array had their own preamplifier and data acquisition system. IgM and von Willenbrand factor (vWF) proteins were analysed with this array. A high signal was detected for singly charged 1 MDa IgM, smaller signals up until IgM4+ charge were detected. For vWF
a strong signal was shown for the single ion and the dimer, multimers were detected up until 2 MDa. The method was able to clearly distinguish the charge states for both analytes.
36
Although the detection efficiency of cryogenic detectors is greatly improved for higher m/z molecules compared to MCPs, there are multiple factors which limit the applicability of STJs in the higher mass range. Even though the signal height should be proportional to the kinetic energy of the particles, it seems that the relation is non-linear. It can be explained by the possible recombination of the electrons into Cooper pairs. Another limitation is the poor energy resolution of this method. Many approaches have been made to improve it, and newer setups has been successfully used for the analysis of ions in the MDa range. The most remarkable result by STJs is the 13 MDa and 18 MDa capsid measurements of bacteriophage HK97. Despite the few successful studies for ions with higher mass, highly charged ions are still difficult to measure, which makes the cryogenic detectors more applicable to MALDI-TOF instrumentations. 28
5.5. Charge detection mass spectrometry
Charge detection MS (CDMS) is a single-molecule method for the simultaneous measurement of the m/z and the charge. In CDMS measurements an ion passes through a conducting tube and induces a charge on it which is picked up by a charge-sensitive amplifier. The induced charge is maintained until the ion leaves the tube and then it dissipates. By increasing the length of the tube, the magnitude of the induced charge approaches the charge of the ion. Using a length to diameter ratio of 4 or higher, the magnitude of the two charges can be considered equal. 7,54
In contrast with FT-ICR where the induced charge depends on the ion trajectory, in CDMS the knowledge of the ion trajectory is not required to measure the charge. This method is capable of accurate charge measurements because the induced charge is independent from the location of the ion in the tube. The m/z ratio can be determined from the flight time of the ion, and the mass then can be calculated by multiplying the charge and the m/z. CDMS can be operated in 3 modes: single-pass, linear array and ion trap. 7,28
37
5.5.1. Single-pass CDMS
Single-pass is the simplest and first developed mode of CDMS. The first CDMS measurements were made in 1960 by Shelton to study the impact of micrometer-sized charged iron particles on various surfaces. 56 The particles were charged to 10000 e and then accelerated
through a 100 kV potential difference to 3 km/s. The accelerated particles went through two charge detector tubes in series. An amplifier picked up the signals from the detector tubes and displayed it on an oscilloscope. Since the induced charge only lasts until the particle leaves the detector tubes, the velocity, which provides the m/z was determined by the width of the signal from the time the particle entered the first tube until it left the second one. The amplitude of the signal was used to determine the charge. 7
Figure 37 Detector for measuring particle velocity, charge and position 56
In 1995 Fuerstenau and Benner combined this method with ESI, which extended the measurements into the MDa range and gave way to the study of nanoparticles and biological macromolecules.57 They were the first to call this technique charge detection mass
spectrometry. In this setup the detector tube is connected to a JFET transistor at the input of a charge sensitive preamplifier. The charge induced by the entering ion is detected by the preamplifier and a Gaussian shaping amplifier processes the signal. A pulse is produced by the amplifier when an ion enters the detector tube, and a pulse of the opposite polarity when it exits.28
The charge is usually determined from the amplitude of the pulses. The induced charge does not instantly appear with the entering ion. It builds up while the ion approaches the detector tube and reaches its full value when it is already a short distance in the tube, so the ion velocity and the tube diameter determine the rise and fall times. The charge can only be determined from the pulse amplitudes if the rise and fall times are much shorter than the shaping time constants. If this condition is not fulfilled, then the pulse amplitudes become dependent of the ion velocity and the pulse area must be used to determine the charge. 7
38
Figure 38 Charge detector and amplifier set up 57
In single-pass measurements the m/z is determined from the flight time of the ions in the detector tube. The flight time is obtained from the ion pulses at the entrance and the exit. Since the induced charge starts to build up even before entering the tube, the apparent length of the tube can be different from the physical length. 58 The flight time depends on the ion energy
which comes from multiple sources. The ions gain energy when they are accelerated through a potential difference. Since this potential difference is known, the energy gained by it can be determined. Another source of energy is the expansion in the electrospray interface. An average excess energy can be determined by grounding the electrodes so the ions would pass the tube without accelerating and then measuring the average ion velocity. However, the gas expansion also creates a distribution of ion energies which cannot be determined by this method and so it causes tailing of the peaks. 7,57
Two methods have been developed for the correction of the excess energy. In the first case an ion is passed through the detector tube and the flight time is used to measure its velocity. Then the ion enters a grounded acceleration tube where it is pulsed to a high potential, so it is accelerated when it exits. After the acceleration the ion passes through a second detector tube. The m/z can be determined by the acceleration potential and the difference in the velocities. In the second method the ion velocity in the first tube is measured from the flight time as usual, then the ion is accelerated and decelerated through a potential ramp before arriving to the second tube. To determine the m/z the arrival time of the ion is compared to the theoretical arrival time that would have been without the potential ramp. Although the excess energy makes the m/z measurements more difficult, in single-pass measurements it is still about an order of magnitude more precise than the charge measurements. 7
In single-pass measurements the charge uncertainty is limited by the detector noise which makes these measurements quite imprecise. The lowest value the uncertainty could be reduced to was about 50 e, with a limit of detection of 250 e, since the signal needs to be at least 5 times
