Design Characteristics to Eliminate the Need for Parameter Optimization in Nanoflow ESI-MS

Abstract

The sampling efficiency in electrospray ionization-mass spectrometry (ESI-MS) can be improved by decreasing the liquid flow rate to the nanoflow regime, where it is possible to draw a large fraction of the ESI plume into the mass spectrometer. This mode of operation is typically more difficult than ESI-MS at higher flow rates because it requires careful optimization of a number of parameters to achieve optimal sampling efficiency. In this work, we screened the relative impact on signal intensity and spray stability of factors that included sprayer position, spray electrode protrusion, sprayer tip shape, spray angle relative to the MS inlet, nebulizer gas flow rate, ESI potential, and means for generating the electric field to initiate electrospray. Based on the screening results, we explore the possibility of providing fixed optimal values for many of the key source parameters to eliminate much of the tuning that is required for conventional nanoflow sources. This approach has potential to greatly simplify nanoflow ESI-MS, while providing optimized sensitivity, stability, and robustness, with decreased variability between analyses.

Introduction

Electrospray ionization (ESI) is the most widely utilized ionization technique in mass spectrometry (MS) due to high sensitivity and broad compound coverage [1]. In a typical ESI source, a flowing stream of liquid is transported through a capillary that is maintained at high potential relative to a counter electrode (typically a mass spectrometer inlet). Charged droplets emanate from the “Taylor cone” [2,3,4] at the tip of the capillary, and gas phase ions are released as the solvent evaporates. Ions are directed to the sampling inlet of the MS and pass through a series of differentially pumped vacuum stages prior to mass analysis. Efforts to improve mass spectrometer sensitivity are limited by the efficiency of ionizing analytes in the sample flow stream and the efficiency of transferring ions into the vacuum system; the combination of ionization and transfer efficiency is referred to as sampling efficiency. Sampling efficiency can be determined by the ratio of ions transferred into the vacuum system of the mass spectrometer to molecules entering the spray electrode, and this improves significantly as the solvent flow rate in ESI drops from approximately 1 mL/min to below 1 μL/min [1, 5,6,7,8]. The nanoflow regime can generally be defined as using solvent flow rates less than 1 μL/min and extending to flows below 1 nL/min [3, 9]. In nanoflow ESI, the initial droplet diameters are in the hundreds of nm range, which is 100–1000 times smaller than droplets formed by high-flow ESI. This significantly improves the ionization efficiency and allows the capillary tip to be positioned much closer to the entrance of the mass spectrometer [4] to improve transfer efficiency. Under ideal nanoflow ESI conditions, the entire vaporized spray can be drawn into the mass spectrometer with sufficient desolvation to liberate declustered ions. At flows in the range of 50–500 nL/min, sampling efficiencies greater than 70% have been achieved [7] and reproduced [10].

High-efficiency nanoflow ESI requires careful optimization of all source and inlet parameters, and therefore, wide variations in sampling efficiency have been reported in the literature [6]. Sampling efficiency is also affected by changes of solvent viscosity and conductivity [6, 7]. Typical source parameters that need to be optimized include the physical positioning of the sprayer [11] relative to the sampling inlet in terms of angle and 3-dimensional location, emitter tip protrusion from the nebulizer probe, ESI potential, and nebulizer gas flow rate. Given the parametric nature of optimization for these parameters, retuning is typically required whenever an emitter tip is replaced. Researchers have worked hard in this area to derive preferred combinations of the various parameters to achieve results specific for their applications [12]; however, a nanoflow ESI source that does not require extensive optimization is desirable.

It has been shown previously that the optimal inner diameter of the spray emitter decreases as the liquid flow rate decreases [5, 8, 13]. For example, one study showed that the optimal emitter inner diameters were 20 μm, 5 μm, and 1 μm for flow rates of 300 nL/min, 50 nL/min, and 10 nL/min [5], respectively. Fused silica capillary is often used as the nanospray emitter because it is durable, inexpensive, inert to most solvents, and available in a range of inner and outer diameters. Two general approaches have been used to apply the ESI potential to a nanoflow liquid stream [14]: The outer terminus of the fused silica sprayer can be coated with conductive material (direct contact approach), or the flow stream can be electrically contacted with a conductive member upstream of the sprayer tip (liquid junction approach). While these approaches have been described in the literature [14, 15], it remains unclear whether either approach provides a benefit in terms of analytical performance.

The goal of our work is to achieve optimal signal intensity, stability, and robustness for nanoflow ESI without the need for extensive optimization. To simplify the process, we included the introduction of a nebulizer gas [16] which helps to stabilize the spray and a standard nanoflow interface to provide improved sampling efficiency with effective desolvation [7, 17, 18]. A series of experiments were conducted to screen the relative significance of the key source parameters that typically need to be tuned to achieve optimal performance in nanoESI operation. These experiments provide information on the criticality of these parameters for nanoESI performance, as well as the broadness of the optimization range for tuning. Based on these results, we have developed a new nanoESI configuration having fixed optimal values for sprayer position, tip protrusion, tip geometry, and sprayer angle. The ESI potential and nebulizer gas flow were adjustable; however, the optimization range for each of them was sufficiently broad that careful optimization was not required. The new approach eliminates the need for substantial tuning while providing performance equivalent to a fully articulated conventional nanoflow ESI source. In addition, we studied two approaches for applying the ESI potential to the liquid stream and systematically compared the analytical performance using coated sprayers with direct electrical contact and liquid junction approaches in both positive and negative modes. Finally, we conducted highly accelerated contamination testing [19] for a mass spectrometer using sprayers oriented on the sampling axis and at a 25° angle relative to the sampling inlet, to compare the system robustness.

Experimental

Mass Spectrometers

Experiments were conducted on two hybrid triple quadrupole/linear ion trap mass spectrometers (QTRAP® 6500 system, SCIEX) and a hybrid quadrupole/time of flight mass spectrometer (TripleTOF® 6600 system, SCIEX). The systems incorporated a nanoESI-compatible interface with a curtain chamber consisting of a curtain plate with 10-mm aperture, a heated laminar flow chamber with i.d. of 2 mm, and an inlet orifice with i.d. of 0.6 mm (TOF platform) or 0.72 mm (triple quadrupole platform). On the triple quadrupole system, data were acquired in multiple reaction monitoring (MRM) mode. On the TOF platform, data were acquired in TOF MS, IDA, and SWATH® acquisition modes following a standard procedure [20]. The IDA raw data were searched through ProteinPilot™ v5.0 software to generate the peptide spectral library for identifications using a false discovery rate (FDR) threshold of 1%. The SWATH® raw data were processed, and the spectra of confidently identified species were extracted (1% FDR) using SWATH® 2.0 application in a custom version of PeakView® software. Automated peak detection, transitions/peptide selection, and area extraction were performed for quantitation [21].

NanoLC Platforms

The nanoLC-MS experiments were conducted on two nanoLC systems: (1) an Eksigent nanoLC-Ultra™ system combined with the cHiPLC® system (SCIEX) operated in a trap-elute configuration with a trap column (200 μm × 6 mm ChromXP™ column C18-CL, 3 μm, 120 Å) and a nano cHiPLC® column (75 μm × 15 cm ChromXP™ column C18-CL, 3 μm, 120 Å); (2) the ekspert™ nanoLC 400 System (SCIEX) operated in direct injection configuration with the use of a nano column (75 μm × 15 cm ChromXP™ column C18-CL, 3 μm, 120 Å).

Nanospray Ion Source

Nanoflow ESI experiments utilized a fully articulated ion source (NanoSpray® III Ion Source, SCIEX), hereafter referred to as “baseline source,” and a source configuration constructed in-house. Ions were generated from solutions eluted from LC at a flow rate of 300 nL/min or infused from a syringe pump at a flow rate of 500 nL/min. The fused silica capillary connected to the LC or syringe was coupled to the spray emitter using either a conductive low-dead volume union (C360US62, Valco Instruments Co. Inc.) or a non-conductive low-dead volume union (P-779, IDEX Health & Science LLC.). For the liquid junction approach, the union also served as the electrical contact for ESI. When using the coated tip approach, the union was non-conductive and the ESI potential was applied to a platinum coating on the outer surface of the emitter. When using the baseline source, the sprayer was positioned with an angle of 25° towards the sampling axis, and the x-y-z stage was used to position the emitter tip relative to the curtain plate aperture. In the modified source, the sprayer was positioned with an angle of 0° towards the sampling axis, and the tip positioning was not adjustable; therefore, this source will be referred to as “fixed source” throughout this paper. Two different nanoflow ESI emitters were used for this work. The commercial nanoflow ESI source used emitters with 20-μm inner diameter pulled down to 10 μm at the tip (FS360-20-10-N-20-C12, New Objective Inc.). The fixed source used custom sprayers (SCIEX) with 20-μm inner diameter and 360-μm outer diameter ground at the tip to an outer dimension of approximately 50 μm, with grind length of 1 mm. These tips did not plug as frequently as the pulled emitters because they maintained a constant internal diameter from the inlet to the outlet. A pneumatic nebulizer was used to stabilize the spray, emanating from a nozzle with 500-μm diameter.

Design of Experiments (DOE)

Signal intensity was measured for a reserpine standard using a 3-level, 3-factor DOE (SigmaZone, Quantum XL 2013) with three input factors: sprayer radial position, ion spray voltage, and nebulizer gas flow. Reserpine ion signals were recorded with different combinations of the input factors (see details in “Results and Discussion”, Table 1), with a total of 27 runs in one dataset. This set of experiments was repeated in triplicate. The DOE model was generated by regression analysis of the experimental results to predict the ion transmission across a range of settings for sprayer radial position, ion spray voltage, and nebulizer gas flow. The model was verified with additional experiments that were conducted using input values other than those used to generate the model, and predicted values were compared with experimental results.

Table 1 Three-Level and Three-Factor DOE Design Parameters

Highly Accelerated Contamination Testing

Experiments were conducted on the triple quadrupole system using the 2 ion source configurations, with continuous infusion of either diluted olive oil solution comprising ~ 3 mM lipids or crashed horse plasma, at a flow rate of 1 μL/min for 17 h and 24 h, respectively. The system was baselined prior to infusion of samples, using a reserpine standard. The analytical performance of the instrument was remeasured after contamination experiments, and the reserpine MRM signal before and after each of the contamination tests with the 2 sources was compared. Various ion optics elements were examined for the presence of visible debris, and digital photographs were taken after each contamination experiment.

Chemicals

Reserpine standards (Sigma-Aldrich Co.) were diluted to 10 pg/μL in a solvent mixture containing ethanol, methanol, water, and isopropanol (volume ratio of 25:51:23:1) with 0.1% formic acid. Taurocholic acid (Sigma-Aldrich Co.) was diluted to 50 pg/μL in solution comprising 90/10 water/methanol with 0.1% formic acid. TOF tuning solution (SCIEX) containing cesium iodide and a peptide was prepared in solutions comprising 50:50 methanol/water (v/v) with 0.1% acetic acid. Concentrated lipid solutions were made from a sample of extra virgin olive oil diluted by 1000× in solvent comprising 2:1 methanol/chloroform with 5 mM ammonium acetate. A solution of crashed horse plasma was prepared by adding 50:50 acetonitrile/water (v/v) into horse plasma samples (Sigma-Aldrich Co.), followed by centrifugation to precipitate serum proteins. The supernatant was used with no further dilution. Bovine serum albumin (BSA, Sigma-Aldrich Co.) was trypsin-digested following a standard procedure [18], to produce stock solutions containing 30 pmol/μL of digested BSA, which were frozen at − 20 °C in aliquots. Prior to analysis, each BSA stock was diluted to 5 fmol/μL in solvent comprising water with 0.1% formic acid. A mixture of 20 peptides (PepCalMix, SCIEX) was diluted to 10 fmol/μL in water containing 5% acetic acid and 2% acetonitrile. E. coli Digest (Waters) and K562 Protein Extract Digest (SCIEX) were diluted to 0.5 μg/μL in water with 0.1% formic acid.

Results and Discussion

DOE Study: Sprayer Position Relative to the Sampling Inlet

Routine analysis on a conventional ESI platform requires tuning of the sprayer position, ion spray potential, and nebulizer gas flow to achieve optimized sensitivity and spray stability. To investigate the relative importance of these factors for ion transmission, we conducted a 3-level, 3-factor DOE. A total of 27 runs were conducted with different combinations of values for the sprayer radial position, ion spray potential (ISV), and nebulizer gas flow (Neb), with the settings listed in Table 1, and the DOE experiments were replicated 3 times (81 runs). Transmission was measured by monitoring the signal for reserpine ions using the baseline source to permit positional adjustment. The optimal value for each of the factors was first determined (settings defined as Level-2 in Table 1), and then a suitable offset was selected to define Level-1 and Level-3. In the case of the radial position, a setting of 14 mm on the stage micrometer defined the optimum and the radial position was then shifted by ± 1.0 mm to set the other positions tested in the DOE. The results were independent of whether the radial position was adjusted horizontally or vertically. For all experiments, the sprayer tip was positioned 1 mm out from the curtain plate aperture to ensure maximum sensitivity while maintaining sufficient curtain gas flow to effectively decluster the ions. The ion spray potential and nebulizer gas flow settings were selected to span a typical range of settings for this type of source. The validity of the model was verified with 8 verification runs using radial positions of 13.5 mm and 14.5 mm, ion spray potential of 2550 V and 3050 V, and nebulizer gas settings of 13 psi and 23 psi, respectively. The experimental values matched the model predictions to within 10% in all cases.

Figure 1 shows “surface plots” generated from the DOE to demonstrate the change in signal intensity with different sprayer positions, ion spray potentials, and nebulizer gas pressures. Sprayer position had the greatest effect on signal intensity in these experiments. Shifting the radial position of the sprayer by ± 1.0 mm from the optimal location decreased the reserpine ion signal by approximately 2-fold. Given that the heated inlet diameter was 2 mm for these experiments, a shift of 1 mm from the center axis of the inlet would position the sprayer such that 50% of the plume would overlap the inlet and 50% would not, explaining the 2-fold decrease in signal intensity. Offsetting the sprayer from the central axis by ± 0.5 mm gave smaller decreases in intensity, on the order of ~ 20 to 30%. These results are encouraging as they provide a relatively easy engineering challenge for a fixed position configuration. Tolerance stack-up analysis for the proposed configuration suggests that a worst-case scenario of 0.332 mm offset from the center axis is possible. A series of 12 fixed position probes were built to verify the experimental offset from the central sampling axis, and the average offset from the center in the horizontal and vertical dimensions was 0.0183 ± 0.1162 mm and − 0.0374 ± 0.1038 mm, respectively. The DOE results predict negligible signal loss due to fixing the position of the sprayer within this range of offsets. Nebulizer gas flow pressure and ion spray potential had broad tuning optima relative to radial position. As shown in Figure 1a, when the nebulizer gas flow was adjusted over a range from 8 to 28 psi, the reserpine ion signal varied by less than 10%. When the ion spray potential was adjusted over a 1-kV range, the reserpine ion signal varied by less than 20% (Figure 1b). There was a gradual signal reduction at the highest nebulizer gas pressure or lowest ion spray potential. The tuning ranges for nebulizer pressure and ion spray potential measured in these tests cover the typical operational ranges for this kind of device. The optimal tuning range for ion spray potential and nebulizer gas flow rate was sufficiently broad that there was no need to restrict the operational range. In practice, it was possible to use a single ion spray potential and nebulizer gas flow rate across multiple source housings and nebulizer probes with no significant loss of performance, enabling “plug-and-play” performance.

Figure 1
figure1

(a) DOE plot of reserpine ion intensity as a function of radial position and nebulizer gas flow pressure, with the ion spray voltage fixed at 2800 V. (b) DOE plot of reserpine ion intensity as a function of radial position and ion spray voltage, with the nebulizer gas flow fixed at 18 psi. Sensitivity bands were marked in different colors

The distance from the nebulizer nozzle to the tip of the emitter is referred to as the electrode protrusion, and its optimal value has previously been determined to be approximately 1 mm [18]. Using the proposed ESI configuration with a fixed sprayer position and a triple quadrupole instrument, we observed ~ 30% signal difference when varying the protrusion over a total range from 0.681 to 1.502 mm, with an optimum value of approximately 0.9 mm that was selected for all further experiments. Confining the protrusion range from 0.7 to 1.1 mm reduced the signal spread to ~ 11%. A protrusion tolerance of ± 0.2 mm is easily achieved with the proposed ESI configuration with fixed sprayer position and tip protrusion by controlling the tolerances for the probe and the sprayer electrode. A series of 11 probe and electrode combinations were tested to determine the typical spread in protrusion around the optimum value, and the results showed ± 0.079-mm spread. Such a small protrusion variance would be expected to give less than 5% variation in signal response.

Sprayer Angle to Sampling Inlet and Robustness

The conventional nanoESI configuration used for these studies oriented the sprayer with a 25° vertical angle relative to the axis of the sampling inlet. A small spray angle can be beneficial for interface configurations that capture only a small portion of the nanoESI plume. However, a spray angle adds additional tuning complexity because any variation in the axial position requires a change in the angled radial dimension. In situations where the majority of the nanoflow ESI plume is consumed at the nanoflow interface to provide maximum transmission, a spray angle presents an unnecessary complication, particularly for a sprayer configuration designed to have a fixed position. When the sampling inlet size approaches the ESI plume size, a similar fraction of the plume can be captured with the sprayer placed on axis or with a small angle (less than 30°). Experiments were conducted using a reserpine sample to compare the sampling efficiency [7] for sprayer configurations with a 25° angle and a 0° angle using the same pulled emitter. Under conditions to provide optimal transmission, the reserpine ion count rate for repeated measurements varied over a range of 3.0 × 106 to 3.2 × 106 cps in both configurations. In addition to signal intensity for targeted analytes, indistinguishable full scan Q1 spectra were achieved using these 2 angles, further indicating equivalent sampling efficiency. These results support the premise that a slight difference in sprayer orientation was not sufficient to affect the sampling efficiency of charged species using the present interface, and this is consistent with our previous results (unpublished data).

To investigate if there was any difference in system robustness, back-to-back highly accelerated contamination tests were conducted in duplicate, using nanoESI configurations with a 25° sprayer angle and a 0° sprayer angle, following the procedures described in the “Experimental” section. One test involved continuous infusion of diluted olive oil, and the other test involved continuous infusion of crashed plasma, representing contamination from hydrophobic and hydrophilic matrices, respectively. The signal decrease for reserpine ions from the initial baseline, after the contamination tests, is listed in Table 2. The extent of signal decrease after contamination experiments was similar regardless of the angle of the emitter relative to the inlet. The accumulation of debris on critical ion path elements resulted in charging and over-resolving of peaks to a similar extent regardless of the sprayer orientation.

Table 2 Signal Decay After Contamination

In addition to the experimental measurements of ion signal, digital photographs were taken from all lens elements from the inlet to the Q1 mass analyzer after the contamination experiments, as illustrated in Figure 2(a–d) for lipid analysis with the 25° angle, (e–h) for lipid analysis with the 0° angle, (i–l) for crashed plasma analysis with the 25° angle, and (m–p) for crashed plasma analysis with the 0° angle. Dried residues from the oil sample had a similar appearance on the heated inlet (Figure 2a, e), inlet orifice (Figure 2b, f), IQ0 lens (Figure 2c, g), and IQ1 lens (Figure 2d, h), providing no indication that sprayer orientation contributed to the extent or appearance of deposits.

Figure 2
figure2

Schematic of the mass spectrometer system used for these studies. Ions were generated by nanoESI in a source region. They passed through a countercurrent flow of nitrogen (curtain gas) directed by a curtain plate prior to entering the mass spectrometer inlet with the heated laminar flow chamber inlet tube and the orifice (OR). The system was differentially pumped, and a quadrupole ion guide was installed in the first vacuum stage (QJET® ion guide). The ions passed through an aperture (IQ0) prior to the second differentially pumped vacuum stage. The second vacuum stage included an additional quadrupole ion guide (Q0), and an additional aperture (IQ1) was located prior to the first quadrupole mass analyzer. The digital photographs across the top of the figure show the comparison of four of the ion path elements (heated inlet tube, OR, IQ0, and IQ1) after a highly accelerated robustness test that involved infusion of concentrated lipid solution with the 25° (ad) and 0° orientation (eh). The digital photographs across the bottom of the figure show the comparison of the four ion path elements after crashed plasma contamination tests, with the 25° (il) and 0° (mp) configuration

The crashed plasma matrix gave a more substantial buildup of solid debris in the inlet region, as compared with the diluted oil sample, particularly on the heated inlet (Figure 2i, m) and the inlet orifice (Figure 2j, n). The total amount of material and shape of the deposits was similar for the 2 orientations with only slight differences in symmetry for the deposits on the heated inlet. Given that ion transport is gas flow dominated in this region, debris accumulation on the heated inlet did not significantly contribute to signal reduction; cleaning only this ion path element did not restore substantial signal. The difference in dryness for the deposit on the orifice (Figure 2j, n) reflects a different length of time between the end of the experiment and capturing the digital photographs (i.e., the orifice plate that looks drier was exposed to a longer period of time at high temperature). Deposits on the IQ0 lens (Figure 2k, o) and the IQ1 lens (Figure 2l, p) appeared similar. Equivalent sensitivity and robustness can be achieved regardless of sprayer angle orientation from 0 to 25°, and this finding is not surprising since under conditions for optimized transmission, the same fraction of the plume was consumed into the MS inlet in either case. Orienting a nanoflow sprayer with an angle other than 0° relative to an MS inlet presents an unnecessary complication for an ion source with fixed position.

Means of Applying the Ion Spray Potential

An investigation was undertaken to evaluate both “direct contact” and “liquid junction” approaches for applying the ion spray potential in nanoESI. The ground emitters were used in both cases to eliminate any variation due to emitter shape. A batch of the emitters were coated on the external surface with platinum, and the “direct contact” connection was made to the coating using a modified probe with internal dimple to connect the ion spray potential. For liquid junction experiments, probes with no internal dimple were used with non-coated emitters and the electrical connection was made to a low-dead volume stainless steel union upstream of the emitter.

The analytical signal intensity for positive ion mode analytes was evaluated by measuring a variety of samples in a range of different solvent compositions, including infusion of reserpine standards and TOF tuning solution, LC-MS injections of BSA digest, and PepCalMix on both triple quadrupole and TOF systems. Distribution plots were generated from multiple experimental data points using Capability Analysis tool (SigmaZone, Quantum XL 2013), to compare nanoESI performance between different approaches, including the baseline source. The ratio of signal intensity for the fixed source/signal intensity for the baseline source is plotted in Figure 3, where equivalent performance gave a value of 1.0. The distribution plots also provide a means to evaluate reproducibility of the relative signal for the 2 approaches. As shown in Figure 3a, a total of 10 sets of comparisons were analyzed to generate a statistical perspective of the intensity distributions, and the results indicated that the liquid junction approach gave slightly better sensitivity (≈ 10–20%) than the coated tip approach in addition to slightly better reproducibility. The ratios to baseline were of 1.060 ± 0.048 and 0.884 ± 0.125 for positive ion mode analysis using the liquid junction and direct contact approaches, respectively. Figure 3b compared the quality in SWATH® and IDA analysis measured on a TOF system. The IDA results used the numbers of digested peptides of E. coli or K562 protein identified at 1% FDR, and the SWATH® results used the numbers of confidently extracted peptides (1% FDR) with good quantitation reproducibility (CV% < 20%) [20, 21]. A total of 9 sets of data were analyzed, and consistent with the sensitivity comparison, the liquid junction approach gave slightly better results (≈ 10%) than the coated tip approach for proteomic analysis. The results relative to the baseline were 0.972 ± 0.031 and 0.900 ± 0.041 for the liquid junction and coated tip approaches, respectively.

Figure 3
figure3

NanoESI performance comparisons between “liquid junction” and “direct contact with Pt coating”, for (a) sensitivity measured in positive mode; (b) proteomics analysis in IDA and SWATH® acquisition; and (c) sensitivity measured in negative mode. The x-axis represents the ratio to the baseline which was measured using a conventional fully articulated ion source

The biggest differentiator for the direct contact and liquid junction approaches was operation in the negative ion mode, where performance is normally limited by the onset of corona discharge [18]. Figure 4 shows digital photographs of a non-coated sprayer tip (Figure 4a) and a coated sprayer tip under 3 different operational conditions (Figure 4b–d). Corona discharge occurs at the emitter tip when the ESI potential increases above a certain threshold in the negative ion mode, and under corona discharge conditions, the tip of the sprayer will appear to glow as shown in Figure 4b, c. In the negative ESI tests, the sprayer was located 1 mm outside of the curtain plate aperture, and tests were conducted under conditions with varying orientations, ion spray potentials, and nebulizer gas flow rates. The onset for corona discharge at the sprayer tip was independent of the spray angle. When using the liquid junction approach (Figure 4a) with a nebulizer setting of 8 psi, the ion spray potential could be increased to − 4000 V with no visible indication of corona discharge. Conversely, when using the coated emitter, visible corona discharge was apparent when the spray potential increased to − 3400 V with the same nebulizer pressure (Figure 4b). Under these corona discharge conditions, a significant (> 10×) signal drop was observed. Increasing the nebulizer gas flow to very high values can help minimize the occurrence of corona discharge (Figure 4d) but at the expense of substantially reduced transmission relative to the liquid junction. Figure 3c demonstrates the results (N = 4) for measured intensity of deprotonated taurocholic acid from samples prepared in solvents comprising 30% acidified water to 95% acidified water. The liquid junction approach provided signal intensity improvements of ~ 3× (ratio to the baseline of 0.988 ± 0.035) compared with the direct contact approach which required high nebulizer gas pressures (ratio to the baseline of 0.371 ± 0.065).

Figure 4
figure4

Digital photograph of tip emitters operated in negative mode, using electrical contact approaches of (a) liquid junction at ISV = − 4000 V and Neb = 8 psi for infusing solutions consisting of 50/50 methanol/water with 0.1% formic acid and (bd) direct contact at ISV = − 3400 V for the same solution compositions. The nebulizer gas settings were 8, 15, and 45 psi for (b), (c), and (d), respectively

Another important factor to consider is the broadness of the tuning range to achieve optimized performance. When the sprayer position and electrode protrusion are fixed, two variables are accessible to tune the nanoflow ESI system: spray potential and nebulizer gas flow rate. In the positive ion mode, the direct contact and liquid junction approaches had similar tuning ranges for the ion spray potential for a given nebulizer pressure; however, the nebulizer gas flow was optimized over a wider tuning range for the liquid junction approach. In the negative ion mode, the coated tip approach gave a very narrow tuning range for the ion spray potential and required very high nebulizer gas flows to mitigate corona discharge problems. Changing the ion spray voltage by 150 V resulted in a 2-fold signal drop for the direct contact approach, while using the liquid junction, the potential could be varied from − 2300 to − 3400 V with less than 30% change in signal.

A final consideration for the optimal method to apply the ESI potential is system durability. Platinum coated tips generally exhibited shorter lifetimes than the liquid junction approach due to loss of the coating at the tip after extensive or improper use. In addition, the dimple that is required to make contact between the probe and the coated tip can scrape the coating if the force and mechanical shape are not ideal. In summary, the liquid junction approach outperforms the direct contact approach since it provides better signal intensity in both positive and negative modes, better spray stability in the negative ion mode, and a wider tuning range.

Optimized nanoESI with No Need for Tuning

The experimental results discussed above illustrate the relative impact of many key nanoESI parameters on sampling efficiency and disclose the operational values or ranges of these features for maximizing nanoESI performance. The results suggest it should be possible to develop a nanoESI configuration with fixed values for tip position, electrode protrusion, and means for providing the electrospray potential, with the capability to achieve optimized performance. This fixed nanoESI source configuration includes a liquid junction connection, on-axis spray orientation, custom ground emitters with 20-μm inner diameter, protrusion fixed at 0.9 mm, and the x-y-z position fixed in the optimal location to eliminate the need for any positional adjustment. Figure 5 shows one example of the sprayer design in this proposed nanoESI configuration. A custom probe assembly was designed to hold the emitter and contain the conductive low-dead volume union upstream of the sprayer tip for application of the ion spray potential. The union is embedded in a plastic holder such that the high voltage is not exposed to the operator. This proposed configuration only has 2 adjustable parameters, ion spray potential and nebulizer gas flow rate, and these have sufficiently broad optimal ranges that it is possible to use default values rather than conducting extensive optimization.

Figure 5
figure5

Schematic of a probe assembly used in the fixed nanoESI configuration

A series of back-to-back experiments were conducted to compare the performance of the fixed source with the baseline source, and the results are summarized in Figure 3. In all cases, values in the optimized range of the ESI potential and nebulizer gas flow were applied, and in the case of the baseline source, the sprayer position and protrusion were carefully optimized. In general, no substantial difference in sensitivity, stability, or robustness was observed between the two configurations. The fixed source provided comparable ion signal intensity and stability to the baseline when operated in either positive (Figure 3a) or negative ion mode (Figure 3c), with the ratio to baseline of 1.060 ± 0.048 and 0.988 ± 0.035 for positive and negative modes, respectively. In addition, comparable peptide identification and quantitation have been achieved in both configurations, with a ratio of 0.972 ± 0.031 to the baseline (Figure 3b). An example of comparative LC-MS data using these two configurations for a peptide mixture eluted from the same LC column is presented in Figure 6. Averaged results for peak areas, peak widths, and peak tailing factors for 20 peptides measured in the new configuration were 1.10×, 0.99×, and 0.97× relative to the baseline, respectively. These results provide proof in principle for a fixed nanoESI source configuration.

Figure 6
figure6

Comparison of LC-MS chromatograms of PepCalMix using (a) a fixed nanoESI configuration and (b) a conventional nanoESI

Conclusion

We have conducted a series of experiments such as DOE and highly accelerated contamination testing to investigate the relative importance of various design characteristics that affect nanoflow ESI performance. These results have enabled us to design a novel nanoflow ESI configuration with all key operational factors fixed, except for the ion spray potential and the nebulizer gas flow rate. The fixed source configuration provides sensitivity, stability, and robustness equivalent to a fully optimized conventional nanoflow ion source, while the “plug-and-play” nature simplifies nanoflow ESI to make it amenable to less-experience users. The results suggest that it should be possible for future nanoflow ESI sources to provide high performance in addition to user-friendly operation, enabling nanoESI as a useful tool for routine analysis.

References

  1. 1.

    Covey, T.R., Schneider, B.B., Javaheri, H., Yves LeBlanc, J.C., Ivosev, G., Corr, J.J., Kovarik, P., ESI, APCI, and MALDI: A comparison of the central analytical figures of merit: sensitivity, reproducibility, and speed. In: Cole, R.B. (ed.) Electrospray and MALDI Mass Spectrometry: Fundamentals, Instrumentation, Practicalities, and Biological Applications, pp. 443–490. John Wiley & Sons, Inc, Hoboken (2010)

    Google Scholar 

  2. 2.

    Gaskell, S.J.: Electrospray: principles and practice. J. Mass Spectrom. 32, 677–688 (1997)

    CAS  Article  Google Scholar 

  3. 3.

    Wilm, M.S., Mann, M.: Electrospray and Taylor-Cone theory, Dole’s beam of macromolecules at last? Int. J. Mass Spectrom. Ion Process. 136, 167–180 (1994)

    CAS  Article  Google Scholar 

  4. 4.

    Kebarle, P., Verkerk, U.H.: Electrospray: from ions in solution to ions in the gas phase, what we know now. Mass Spectrom. Rev. 28, 898–917 (2009)

    CAS  Article  Google Scholar 

  5. 5.

    Covey, T.R., Thomson, B.A., Schneider, B.B.: Atmospheric pressure ion sources. Mass Spectrom. Rev. 28, 870–897 (2009)

    CAS  Article  Google Scholar 

  6. 6.

    El-Faramawy, A., Siu, K.M., Thomson, B.A.: Efficiency of nano-electrospray ionization. J. Am. Soc. Mass Spectrom. 16, 1702–1707 (2005)

    CAS  Article  Google Scholar 

  7. 7.

    Schneider, B.B., Javaheri, H., Covey, T.R.: Ion sampling effects under conditions of total solvent consumption. Rapid Commun. Mass Spectrom. 20, 1538–1544 (2006)

    CAS  Article  Google Scholar 

  8. 8.

    Geromanos, S., Freckleton, G., Tempst, P.: Tuning of an electrospray ionization source for maximum peptide-ion transmission into a mass spectrometer. Anal. Chem. 72, 777–790 (2000)

    CAS  Article  Google Scholar 

  9. 9.

    Emmett, M.R., Caprioli, R.M.: Micro-electrospray mass spectrometry: ultra-high-sensitivity analysis of peptides and proteins. J. Am. Soc. Mass Spectrom. 5, 605–613 (1994)

    CAS  Article  Google Scholar 

  10. 10.

    Krutchinsky, A.N., Padovan, J.C., Cohen, H., Chait, B.T.: Maximizing ion transmission from atmospheric pressure into the vacuum of mass spectrometers with a novel electrospray interface. J. Am. Soc. Mass Spectrom. 26, 649–658 (2015)

    CAS  Article  Google Scholar 

  11. 11.

    Marginean, I., Kelly, R.T., Prior, D.C., LaMarche, B.L., Tang, K., Smith, R.D.: Analytical characterization of the electrospray ion source in the nanoflow regime. Anal. Chem. 80, 6573–6579 (2008)

    CAS  Article  Google Scholar 

  12. 12.

    Manisali, I., Chen, D.D., Schneider, B.B.: Electrospray ionization source geometry for mass spectrometry: past, present, and future. Trends Anal. Chem. 25, 243–256 (2006)

    CAS  Article  Google Scholar 

  13. 13.

    Schmidt, A., Karas, M., Dülcks, T.: Effect of different solution flow rates on analyte ion signals in nano-ESI MS, or: when does ESI turn into nano-ESI? J. Am. Soc. Mass Spectrom. 14, 492–500 (2003)

    CAS  Article  Google Scholar 

  14. 14.

    Gelpí, E.: Interfaces for coupled liquid-phase separation/mass spectrometry techniques. An update on recent developments. J. Mass Spectrom. 37, 241–253 (2002)

    Article  Google Scholar 

  15. 15.

    Amirkhani, A., Wetterhall, M., Nilsson, S., Danielsson, R., Bergquist, J.: Comparison between different sheathless electrospray emitter configurations regarding the performance of nanoscale liquid chromatography–time-of-flight mass spectrometry analysis. J. Chromatogr. A. 1033, 257–266 (2004)

    CAS  Article  Google Scholar 

  16. 16.

    Bruins, A.P., Covey, T.R., Henion, J.D.: Ion spray interface for combined liquid chromatography/atmospheric pressure ionization mass spectrometry. Anal. Chem. 59, 2642–2646 (1987)

    CAS  Article  Google Scholar 

  17. 17.

    Schneider, B.B., Baranov, V.I., Javaheri, H., Covey, T.R.: Particle discriminator Interface for nanoflow ESI-MS. J. Am. Soc. Mass Spectrom. 14, 1236–1246 (2003)

    CAS  Article  Google Scholar 

  18. 18.

    Schneider, B.B., Guo, X., Fell, L.M., Covey, T.R.: Stable gradient nanoflow LC-MS. J. Am. Soc. Mass Spectrom. 16, 1545–1551 (2005)

    CAS  Article  Google Scholar 

  19. 19.

    Kang, Y., Schneider, B.B., Covey, T.R.: On the nature of mass spectrometer analyzer contamination. J. Am. Soc. Mass Spectrom. 28, 2384–2392 (2017)

    CAS  Article  Google Scholar 

  20. 20.

    Collins, B.C., Hunter, C.L., Liu, Y., Schilling, B., Rosenberger, G., Bader, S.L., Chan, D.W., Gibson, B.W., Gingras, A.-C., Held, J.M.: Multi-laboratory assessment of reproducibility, qualitative and quantitative performance of SWATH-mass spectrometry. Nat. Commun. 8, 291 (2017)

    Article  Google Scholar 

  21. 21.

    Kang, Y., Burton, L., Lau, A., Tate, S.: SWATH-ID: an instrument method which combines identification and quantification in a single analysis. Proteomics. 17, 1500522 (2017)

    Article  Google Scholar 

Download references

Acknowledgements

We appreciate the help of Deolinda Fernandes for the preparation of the samples. We also appreciate Stan Potyrala for the prototype source design.

Author information

Affiliations

Authors

Corresponding author

Correspondence to Yang Kang.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Kang, Y., Schneider, B.B., Bedford, L. et al. Design Characteristics to Eliminate the Need for Parameter Optimization in Nanoflow ESI-MS. J. Am. Soc. Mass Spectrom. 30, 2347–2357 (2019). https://doi.org/10.1007/s13361-019-02301-8

Download citation

Keywords

  • Nanospray ESI
  • Optimization
  • Tuning
  • Sensitivity
  • Robustness