The Secrets of Electrospray Ionization: Why Less is More

For this month’s Chromatography Corner, we spoke to Kevin Schug of the University of Texas at Arlington. Schug addresses a few myths that surround electrospray ionization (ESI), including why “less is more” for successful separations. Schug will be speaking on this topic in a CHROMacademy webcast on February 24, 2011.

Once you have found the right capillary (sprayer) voltage for a particular application using electrospray ionization, this voltage can then be equally applied for most separations. True or false?

Schug: This is actually true and false, depending on your point of view, and how well you want to optimize the method. While it is true that you can often find a good sprayer voltage that is suitable for the analysis of a wide range of compounds, in positive or negative ionization modes, it may not be optimal. If you are configuring a set-up for walk-up sample analysis, a set-it and forget-it approach is doable. A good rule of thumb is to err on the side of lower voltages if possible. If you are looking to get the most out of your method, then rigorous optimization of experimental variables should include the sprayer voltage.

At high spray voltages, especially in the negative ionization mode, the source can be prone to discharges. To alleviate this (which will certainly hurt reproducibility and system stability), move to lower spray voltages. Additionally, some analytes, especially in some atypical mobile phase solvents, may be prone to redox processes. Reducing the sprayer voltage can reduce contributions from unwanted side reactions that dilute signal quality.

I once received some great advice on electrospray ionization, which applies to many of its experimental parameters: If a little bit works, a little bit less probably works better. I tell this to my students all the time and it particularly relates to concentrations of mobile phase additives, flow rates, and capillary voltage.

Is it true that analytes don’t need to be ionized in the eluent solution before spraying to obtain optimum mass spec sensitivity using this technique?

Schug: Ion generation in ESI is generally regarded as a solution phase process, although gas phase processes are important to consider once the analyte ion has left the droplet in some cases. Ionization typically occurs as some combination of acid–base or charge-transfer reaction in the electrospray droplet. The direct answer to the question is that an analyte does NOT need to be ionized in the eluent solution to obtain a strong MS ion signal. Importantly, compounds that are not readily ionizable in solution, can still be observed in mass spectra generated by ESI. In such cases, adduction of ions, such as sodium, ammonium or potassium (or chloride in the negative ionization mode), could be the predominant forms of ions observed. Analytes that are not readily ionizable in bulk solution can still migrate to the surface of the ESI droplet and acquire a charge to generate a strong signal.

Another interesting thing is the notion that acidic compounds can still be ionized with good sensitivity in an acidic environment; similarly for basic compounds in a basic environment. Conventional wisdom and standard acid–base chemistry tells us that ionization of basic analytes would be suppressed in a basic environment. However, there is a growing body of literature that demonstrates higher sensitivity for some such cases (search “wrong-way-round ionization”). On an initial pass, I would still go for the use of an acidic mobile phase modifier to efficiently ionize compounds in the positive ionization mode. And even if the compound did not have a nice amine group sitting there to be protonated, I could still be pretty confident that some cationization would be possible. Of course, as with the spray voltage discussion above, for optimal sensitivity, or in cases where some atypical mobile phase additives are needed to affect separation, a thorough investigation of the effect of mobile phase additives and pH on ion response should be performed. It is important to initially monitor a wide scan range, so that different ion forms, such as adduct and dimer ions, can be tracked.

As the solvent composition changes during gradient analysis, wouldn’t it make sense to dynamically move the sprayer to ensure ions are continually released into the source sampling “sweet spot”?

Schug:While it is true that characteristics of the electrospray change with solvent composition, it is impractical to consider dynamic modification of the sprayer position during an analysis. There are a couple of reasons for this. First, the sprayer voltage and position are set so that a reproducible and stable spray is generated from the electrospray nozzle. A drastic adjustment of the position of the needle would change the magnitude of the potential gradient between the spray capillary and the counter-electrode, which could alter the spray stability. Some interesting studies of spray modes generated under different operating conditions have been published. An additional consideration is that different analyte types (for example, proteins versus small molecules) are believed to reside preferentially in different-sized droplets generated by ESI. The dynamics of an uneven droplet fissioning process relative to the surface activity of different analytes could sensibly lead to such differences. This, coupled with likely differences in how and when ions are released from droplets — for example, as a charged residue rather than by ion evaporation — means that the distribution of different analytes through the spray plume is variable. The position of the sprayer is a parameter which can — and should be — optimized, but its optimal setting is likely to be one which is pliable for providing high signal quality across a wide range of analytes.

HILIC is a popular technique. How is the ESI process affected by working in a highly organic environment?

Schug:One of the main reasons that HILIC has become popular is that it involves mobile phase conditions that are favorable for sensitive ESI-MS analysis. We published a review on this topic not too long ago (J. Sep. Sci., 31, 1465-1480, 2008) and in this review, we included a detailed section on the operation of ESI in low-aqueous and non-aqueous solvent systems. A couple of conditions are especially important.

First, the solvent environment needs to have a sufficiently high dielectric constant to allow for separation of charge. Water is, of course, very good in this regard, but it is still limited in terms of a second important parameter: vapor pressure. In HILIC, a high percentage of acetonitrile provides an optimal environment for ESI. Not only is good charge separation accommodated by the incorporation of some — even a small percentage — of water (the dielectric constant of acetonitrile is of reasonable magnitude, as well), but also, the high vapor pressure of acetonitrile ensures that ESI droplets will desolvate and shrink rapidly to allow efficient release of gas phase ions.

Acetonitrile also has a relatively low surface tension (less than half that of water). Onset voltages for the generation of stable sprays are lower in mobile phases with lower surface tension. Additionally, the point at which droplets subdivide to release smaller droplets will be reached more quickly in droplets with lower surface tension. These aspects lead to higher efficiency in gas phase ion generation during the ESI process. Finally, it doesn’t hurt that many of the analytes being analysed by HILIC are highly polar and/or ionic in nature. This can facilitate ion generation in the ESI droplet.

Ion suppression is considered important by some workers and not so by others. Can you give us a quick overview of how this occurs and some tips on how to avoid it?

Schug: Ion suppression comes in many flavors. In our lab, we deal with it most often in the form of matrix effects, where interferences in a sample co-elute with an analyte of interest, and change (most often, suppress) the response of the analyte relative to that from a pure standard solution. Imagine the relative content of lipids, proteins, and salts in different matrices such as food, plasma, and urine. Matrix effects can introduce drastic errors in quantitative analysis if not accounted for. In other cases, samples that have been subjected to some type of detergent or surfactant will likely be prone to ion suppression. Furthermore, the excessive use of some mobile phase additives (trifluoroacetic acid and triethylamine are notable ones) can lead to ion suppression.

Electrospray is a competitive ionization processes. Different chemical compounds in an electrospray droplet compete for a limited number of charged sites at the droplet surface. When interferences that are more surface active than the analytes of interest are present in the droplet (for example, detergents) these compounds outcompete the analytes for droplet surface sites. The result is ion suppression. This is the same reason that linear ranges can be limited in ESI. As higher concentrations of an analyte are reached, there may not be enough surface sites to accommodate a consistent increase ionization commensurate with concentration. This leads to decreases in ionization efficiency, and nonlinearity of calibration curves at high concentrations.

Trifluoroacetic acid (TFA) and triethylamine (TEA) can impart ion suppression in different ways. TFA should be avoided in high concentrations, because, although it is a good source of protons, the trifluoroacetate moiety has a high propensity for ion pairing with positively charged species. Ion pair formation in the ESI droplet is a solution phase process and can neutralize analyte ions of interest. TEA and other amines can be problematic because they have a high gas phase proton affinity. This means that just outside of the droplet, in the gas phase, interactions can occur where such species steal protons away from analytes of interest. This can be a good thing if you are trying to make negatively-charged ions, but it can be highly detrimental to the formation of positively charged analyte species.

In order to avoid ion suppression, proper sample preparation steps should be taken. Care should be taken to remove highly lipophilic species and detergents at all costs. Never use detergent-based soaps on glassware that will eventually hold solutions for ESI-MS analysis. Using more rigorous preparation techniques, such as solid phase extraction (SPE), can reduce the incidence of ion suppression from matrix components relative to less-selective approaches like liquid-liquid extraction (LLE) and protein precipitation (PP). For mobile phase additives, refer back to that old adage mentioned previously – if a little bit works, a little bit less probably works better. Try to keep concentrations of additives such as TFA and TEA as low as possible. I tell my students to try to stay under 0.1% v/v concentration when we need to explore the use of these or related additives.

Can you give us some tips on the most effective way to optimize source parameters for any particular application?

Schug:Most instruments are tuned regularly for either routine operations, or to accommodate specific applications. An old tune file that worked well previously for the same or a similar application is always a good place to start. However, one must acknowledge that, day-to-day, optimal operation settings on an instrument can change (due to temperature, humidity, or other things). Most instruments these days have some sort of automatic optimization function. This can be very useful, as you can enter a mass-to-charge (m/z) value of interest for your analyte, infuse a sample, and the instrument software will do a quick run though of source parameters to find settings which provide an optimal signal.

The operator should not rely on these settings alone. Interdependence of source variables and robustness should also be considered. In starting a new application, I tell my students to pick a good start point from previous operating conditions for a similar application. Next, I recommend that a systematic study of source variables be performed, starting from the outside in. This is what some software does. First, optimize the sprayer voltage, then move to transfer line variables and beyond. It is typically not necessary to adjust any parameters associated with the mass analyzer itself because this has likely been optimized during some calibration steps or routine maintenance. On a day-to-day basis, signal will be mainly controlled by source and ion transfer settings.

Find the settings which give a high and stable signal. Next, start over from these set points to see if further modification of the variables, again from outside in, can net any additional enhancements in signal. Do this manually or run the automatic optimization again. Finally, check robustness. For any variable, in an experiment, for which a response curve can be generated (i.e., all voltages, flow rates, and temperatures in your ESI source), it is not necessarily good to operate on a distinct maximal setting. Small changes in that setting will manifest a loss in performance very quickly. You would rather operate on a broad maximum (plateau) in a response curve, so that if small changes in a variable do occur, this does not have a drastic effect on measured response. The ability to resist small changes in response based on small changes in experimental variables is denoted as robustness. Robustness is key to a reproducible and reliable method.

Kevin was raised as a chromatographer under the direction of Prof. Harold McNair at Virginia Tech. However, his dissertation work centered on elucidating a better understanding of adduct ion formation in ESI-MS. In post-doctoral work, Kevin investigated the use of ESI-MS for preservation and characterization of noncovalent interactions with Wolfgang Lindner at the University of Vienna.

Currently, Kevin works with a group of 12 students at U.T. Arlington. Research in the Schug lab is highly varied and ranges across the fundamental aspects of separation and ionization processes to applied aspects in trace quantitative analysis. LC-MS and ESI-MS are common threads in all of this work. Schug is also an active member of LCGC's Editorial Advisory Board.