Electrospray Mass Spectroscopy - Advances in Chemistry (ACS

---

7 Electrospray Mass Spectroscopy MALCOLM DOLE, H. L. COX, Jr., and J. GIENIEC

Downloaded by NORTH CAROLINA STATE UNIV on October 24, 2012 | http://pubs.acs.org Publication Date: June 1, 1973 | doi: 10.1021/ba-1973-0125.ch007

Department of Chemistry, Baylor University, Waco, Tex. 76703

Electrospray mass spectroscopy, mass spectroscopy of ions created by an electrospray process, should be significant for the study of nonvolatile substances such as high polymers or thermally unstable lower molecular weight species. Intact gas-phase macroions of polystyrene, polyvinylpyrrolidone, and lysozyme were prepared by this technique, and low­ -resolution M/z distributions were inferred from data obtained using a nozzle beam system with repeller-grid, Faraday-cage system as an analyzer-detector. Use of a time of flight mass spectrometer to improve resolution is not feasible, since the magnetic electron multipliers used to detect ions in TOF spectrometers have been shown to have negligible response to macroions. Applicability of the Flasma Chromatograph to determination of the charge states of gaseous macroions is also being investigated. /Currently available techniques of determining molecular weights and ^ molecular weight distributions permit only low-resolution analyses. Miiller ( 1 ) has recently stated "If we could determine molecular weights [of macromolecules] to a tenth of a percent or better, a load could be taken from the microanalysis' shoulders/' The primary ( although not the only) problem which has heretofore prevented the development of a mass spectrometer for use with macromolecules is that of producing intact gas-phase macroions. Owing to the low vapor pressures of macromole­ cules, macroions cannot be produced in the gas phase by conventional techniques without extensive degradation and/or fragmentation. Electro­ spray mass spectroscopy provides a solution to this basic problem. Electrospray mass spectroscopy ( E M S ) is the mass spectroscopy of gaseous ions produced by electrospraying into a suitable gas at atmos­ pheric pressure a dilute solution containing as solute the macromolecules in question. (Although the use of other gases is possible, we have ob­ tained our best results with nitrogen gas.) Although the technique is 73 In Polymer Molecular Weight Methods; Ezrin, M.; Advances in Chemistry; American Chemical Society: Washington, DC, 1973.

74

POLYMER MOLECULAR WEIGHT METHODS

still i n its infancy, it was first proposed (2) i n 1966, and demonstrated in the case of polystyrene (3,4) i n 1968. Further studies ( 5 - 7 ) appeared in 1970 and 1971.

Downloaded by NORTH CAROLINA STATE UNIV on October 24, 2012 | http://pubs.acs.org Publication Date: June 1, 1973 | doi: 10.1021/ba-1973-0125.ch007

The Electrospray

Technique

The prime requisite for any type of solution-spraying process suitable for use as an ion source in a mass spectrometer is that it produce droplets small enough so that on final evaporation of the solvent single macroion species w i l l result. Various methods of aerosol production have been considered for this application (8). Mechanical and ultrasonic methods of aerosol production are generally limited to droplet diameters greater than 1 μ. In order to produce a reasonable proportion of single macro­ molecules, it is necessary that the solute concentration be such that an average of no more than one macromolecule be present i n each of the final droplets. F o r a homogeneous droplet diameter of 1 μ, this requires a maximum solute concentration of the order of 3 pg m l " amu" . This represents a very low concentration. W e have demonstrated experimen­ tally (5) that such techniques do not produce individual intact macroions at reasonable solute concentrations. If, however, we can produce droplets of 0.1 μ diameter, solute concentrations up to 3 ng m l " amu" are per­ mitted. This represents a reasonable concentration for macromolecules. It was demonstrated in 1952 ( 9 ) that the electrospray process can produce aerosols with rather homogeneous droplet diameters i n the neighborhood of 0.1 μ or less. The electrospray process consists of feeding a liquid through a metal capillary which is maintained at a high electrical potential with respect to some nearby surface. As the liquid reaches the capillary tip, the liquid is dispersed into fine electrified droplets by the action of the electric field at the capillary tip. If the liquid is volatile, as the liquid evaporates the droplets shrink in size, become electrically unstable, and break down into smaller size droplets. This process has been experimentally demon­ strated by Doyle, Moffett, and Vonnegut (10) and by Abbas and Latham (11). If the liquid contains macromolecules, after the solvent has evapo­ rated completely the macromolecules are left as electrically charged particles i n the gas phase, that is, as gaseous macroions. Although numerous attempts to provide theoretical explanations of the electrospray process have been made (see, for example, ref. 12-17), 2L good quantitative theory of the phenomenon would require simulta­ neous solutions of the hydrodynamic and electrostatic differential equa­ tions, and, to our knowledge, no such theory has yet been proffered. However, experimental observations have provided some insight into the process. 1

1

1

1

In Polymer Molecular Weight Methods; Ezrin, M.; Advances in Chemistry; American Chemical Society: Washington, DC, 1973.

7.

Electrospray Mass Spectroscopy

DOLE ET AL.

75

Experimental observations (19) have indicated that liquids with molecular dipole moments less than 10" d y n c m or with specific con­ ductivities outside the range 10" -10~ ohm" cm" cannot be dispersed electrically. Schultze (18) investigated the electrospraying quality of various liquids, and observed that the liquids producing the finest and most stable sprays had specific conductivities i n the range 2 χ 10" to 5.6 Χ 10" ohm" cm" . H e also observed that application of hydrostatic pressures to force the liquid out of the capillary resulted i n a coarser spray with the coarseness of the spray increasing as the pressure increases. Gieniec (7) has obtained data indicating that the dielectric constant of the liquid is also an important factor. Liquids with low dielectric con­ stants such as dioxane cannot be electrosprayed while liquids with high dielectric constants such as distilled water cannot be electrosprayed easily. However, a liquid having a high dielectric constant can be mixed with a liquid having a low dielectric constant to produce a liquid with an inter­ mediate dielectric constant which can be electrosprayed successfully. A n example is 7 0 % dioxane (dielectric constant 2.209 at 25°C) and 3 0 % distilled water (dielectric constant 80.37 at 25°C) by volume. 18

5

13

1 / 2

1

2

1

8

Downloaded by NORTH CAROLINA STATE UNIV on October 24, 2012 | http://pubs.acs.org Publication Date: June 1, 1973 | doi: 10.1021/ba-1973-0125.ch007

6

1

1

N

2

MIXTURE OUT

GROUNDED METAL PLATE

Figure 1. Schematic of a typical electrospray apparatus as used in electrospray mass spectroscopy (see text for explana­ tion of dimensions) It has also been observed that differences i n the physical structure of the spraying system can influence the spray. A typical spraying sys­ tem as used in E M S is illustrated schematically i n Figure 1. The N gas flow is required to dry the droplets and to help sweep the ions away from the needle. In the system employed by Dole et al. ( 3 - 6 ) , the dimensions used were usually a = 13 inches, b = 100 mm, c = 24 inches, and d = 0.1 mm. The orifice of dimension d is used to sample the gaseous mix­ ture and is not functional in the production of the spray. Using voltages 2

In Polymer Molecular Weight Methods; Ezrin, M.; Advances in Chemistry; American Chemical Society: Washington, DC, 1973.

Downloaded by NORTH CAROLINA STATE UNIV on October 24, 2012 | http://pubs.acs.org Publication Date: June 1, 1973 | doi: 10.1021/ba-1973-0125.ch007

76

POLYMER MOLECULAR WEIGHT METHODS

on the needle of 4.5 k V and up they were able to obtain good sprays of both positive and negative polarity. However, best results were obtained with negative sprays. In a similar system used by Gieniec ( 7 ) , the dimen­ sions were a = 44 mm, b = 5.5 mm, c = 25 cm, and d = 9 mm. I n this system the orifice of diameter d is a collimating orifice, and sampling takes place 9 cm downstream from this orifice. In addition, a l l of the gaseous mixture exits through this orifice rather than through a separate orifice. Again, using high voltages of 4.5 k V , good sprays are obtained with positive voltages, but negative sprays could not be formed. Using variations of these basic systems, Drozin (19) and Vonnegut and N e u bauer (9) obtained both positive and negative sprays but observed that a spray with smaller droplets is obtained when the capillary is charged positively. A l l of the work with macroions reported to date has been done using mixtures of either ethanol and water or acetone and benzene. Further investigations into the electrospray process should permit optimization of the system and production of a wide variety of solvent mixtures suitable for use in the electrospray process. Formation of Macroion Beams After the macroions are produced by the electrospray process, they are injected into a vacuum by use of a nozzle-beam system of the type first suggested by Kantrowitz and Grey (20) and later modified by Becker and Bier (21). Articles describing i n detail the principles of operation of such a system have been published, among them being articles by Anderson, Andres, and Fenn (22, 23). A typical nozzle-beam system with associated electrospray apparatus is shown schematically i n Figure 2. Although nozzle-beam systems differ widely i n the details of construction, the basic principles of operation are the same. The gaseous mixture at the stagnation pressure ( p ) (atmospheric pressure in our systems ) expands nearly isentropically through the nozzle orifice (100 μ diameter i n our systems) into the region between the nozzle and skimmer, kept at pressure p i (normally 25 μ i n our systems). As the expansion takes place, much of the energy of random thermal motion is converted into energy of forward-directed mass motion. Super­ sonic flow and aerodynamic shocks result. The shock region is barrel shaped and closed off at the downstream end. Inside the barrel shock, near molecular flow results, and, for points more than a few nozzle d i ­ ameters downstream of the nozzle orifice and away from the shock boundaries, the flow approximates that of a radial source with its point of origin located a few nozzle diameters downstream of the nozzle orifice 0

In Polymer Molecular Weight Methods; Ezrin, M.; Advances in Chemistry; American Chemical Society: Washington, DC, 1973.

Downloaded by NORTH CAROLINA STATE UNIV on October 24, 2012 | http://pubs.acs.org Publication Date: June 1, 1973 | doi: 10.1021/ba-1973-0125.ch007

7.

77

Electrospray Mass Spectroscopy

DOLE ET AL.

(24). That is, the gas density varies as the inverse square of the distance downstream of this point. Since the shock region is essentially a closed surface, the shock surfaces disrupt this flow. It has been shown [see, for example, Anderson et al. (22, 23)] that this disruption can be prevented, and molecular beams of rather uniform velocities and high intensities formed by i m ­ mersing a conical "skimmer" into the shock region. The shocks become attached to the skimmer surface, and molecular flow through the skimmer orifice results. This also permits differential pumping of the system so that lower background pressures (approximately 5 Χ 10" mm i n our systems ) can be achieved in the analyzing and detecting region. Anderson et al. (23) among others have discussed the optimum design and place­ ment of the skimmer. 5

Faraday Cage.

Sk

'«T'

N o z

zle

N gas out 9

t

Spray chamber

1

N gas fj-j—1 « - in 2

Liquid in

Figure 2.

Electrospray chamber-nozzle beam-mass analyzer assembly {after Dole et al. (4)]

Since, during the expansion, thermal energy is converted into directed motion, the gas i n the jet becomes very cold with local temperatures of tens of degrees Kelvin being possible. Because of this low temperature, quite narrow velocity distributions result (24). F o r a homogeneous gas with ratio of specific heats 1.4 (such as nitrogen) the final beam velocity is approximately 2.17ϋ where ϋ is the speed of sound i n the gas at pressure p . F o r nitrogen gas, the final beam velocity is approximately 750 meters/sec. If the gas is a mixture of several pure gases of different molecular weights, with one of the gases representing only a minor constituent of the mixture, then the minor constituent takes on the velocity of the major constituent. This is the "seeded beam" technique as first verified experi­ mentally by Becker et al. (26, 27). Since the macroions constitute only about 10" mole % of the final gaseous mixture, intermediate energy beams of macroions can be produced with the macroions having narrow velocity distributions (2). 0

0

0

8

In Polymer Molecular Weight Methods; Ezrin, M.; Advances in Chemistry; American Chemical Society: Washington, DC, 1973.

78

POLYMER MOLECULAR WEIGHT METHODS

The seeded-beam technique has a further advantage i n that the final beam is enriched i n the heavier constituent (25, 26). This is the so-called " M a c h focussing factor." Methods of Macromass

Analysis

In principle, a conventional magnetic analyzer type of mass spec­ trometer could be used to measure the M/z (mass to charge) ratios of the macroions resulting from the electrospray process. However, for singly charged ions of mass 10 amu (atomic mass units), a magnetic field strength of about 450,000 G would be required to bend the ion path into a 10-cm radius circle if the initial ion velocity were that produced by an accelerating voltage of 1000 V . Thus, although use of magnetic analyzers with low mass macroions (15-50 kamu) is feasible, it is not practical for heavier macroions. In the case of a time of flight mass spectrometer, the estimates are more reasonable. Inasmuch as the time of flight is proportional to the square root of the mass, one can calculate that for an accelerating voltage of 3000 V and a flight path of 100 cm the flight time for an ion of M/z equal to 10 would be 1.4 msec. Measurement of these flight times is entirely feasible. However, two difficulties remain. The first difficulty concerns itself with introducing the macroions into the mass spectrometer in such a way that the initial velocity of the macroions is as near zero as possible. In our work we have attempted to do this by slowing the molecular beam (4) of macroions down to thermal velocities and then by pulsing the ions at right angles into the flight tube of the mass spec­ trometer. T o slow the ions down and to deflect them at right angles without loss of intensity represent two difficult problems. A t the moment we are working on these two problems but without positive results as yet. The second major difficulty is that of detecting the macroions since the magnetic electron multiplier ( M E M ) on which commercial time of flight mass spectrometers are based do not respond sufficiently to macroions. O u r observations show no response to negative ions and a small response, amplification factor of about 20, to positive ions. A factor of about 10 is needed for operation of the time of flight mass spectrometer. The response of the M E M to positive ions is due to an Auger effect (28) which can exist when the ionization potential of the ion is more than twice the work function of the sensitive metal surface of the cathode of the M E M . In general the M E M detects ions because of either a kinetic or potential effect. Schram et al. (29) have obtained data on the rare gas ions which demonstrate that for operation of the M E M a critical velocity of about 10 cm sec" , independent of the mass, is required. If the extremely long extrapolation of these results is made to ions of macromolecular size, the conclusion can be reached that the kinetic effect would

Downloaded by NORTH CAROLINA STATE UNIV on October 24, 2012 | http://pubs.acs.org Publication Date: June 1, 1973 | doi: 10.1021/ba-1973-0125.ch007

6

6

6

6

1

In Polymer Molecular Weight Methods; Ezrin, M.; Advances in Chemistry; American Chemical Society: Washington, DC, 1973.

Downloaded by NORTH CAROLINA STATE UNIV on October 24, 2012 | http://pubs.acs.org Publication Date: June 1, 1973 | doi: 10.1021/ba-1973-0125.ch007

7.

DOLE ET AL.

79

Electrospray Mass Spectroscopy

not function when using the M E M to detect macroions. In the case of negative ions an Auger or potential effect is not possible and, since i n our experiments neither the positive nor the negative ions attained the thresh­ old velocity, the kinetic effect could not contribute to the response of the M E M . A n ion detector of the sensitivity of the M E M but applicable to macroions is critically needed. The quadrupole mass filter ( Q M F ) is a mass analyzer on whose operation use of an M E M is not necessarily dependent. The ion currents produced are of sufficient magnitude to be measured by means of a Faraday cage and a suitable amplifier such as a vibrating-reed electro­ meter. The Q M F is a true M/z filter which requires no magnetic fields. Since first being proposed by Paul and Steinwedel (30), the Q M F has been investigated extensively, and the principles and methods of operation are well known ( see, for example, ref. 31 ). The Q M F is comprised of four parallel rods excited by a combination of static and oscillating fields. Ions of a selected M/z range traverse the entire axial length of the mass filter while other ions are forced into un­ stable trajectories and are ejected from the filter. Usually the ratio of the oscillating and static fields is kept constant, and mass scanning is accomplished by varying this ratio. Frequency scanning could also be used. For stable oscillations of an ion passing through the quadrupole mass filter the values of the ratio V/mr