Assessment of ammonium fluoride as a mobile phase additive for sensitivity gains in electrospray ionization

Abstract

Ammonium fluoride has been shown to improve sensitivity when using electrospray ionization (ESI) coupled with mass spectrometry (MS). Recent internal investigation furthered that claim, through the observation of improved sensitivity when analyzing steroid molecules. This work focuses on extending those observations to other small molecules to understand the impact ammonium fluoride has on detection sensitivity with optimized instrument conditions. Using conventional liquid chromatography ESI‐MS we investigated sensitivity differences between ammonium fluoride, formic acid, or ammonium hydroxide as mobile phase additives. Full source optimization was performed for nine compounds at three different organic concentrations (30%, 60%, or 90%) with formic acid, ammonium fluoride, and ammonium hydroxide adjustment. Optimization results were compiled to generate individual methods by compound, polarity, mobile phase, and organic concentration. Flow injection analysis was performed with fully optimized methods to compare compounds across different solvent systems under optimal conditions. Negative ESI data showed 2–22‐fold sensitivity improvements for all compounds with ammonium fluoride. Positive ESI data showed > 1–11‐fold improvement in sensitivity for four of seven compounds and no change for three of seven compounds with ammonium fluoride. Ammonium fluoride improved ESI⁻ sensitivity for all compounds studied when using optimized source conditions. Investigation with ESI⁺ analyses showed mixed results, with four of seven compounds showing improvement and others showing equivalency or slight loss in sensitivity, suggesting potential sensitivity gains for some analogs with ESI⁺.

List of Abbreviations

ESI

electrospray ionization

HPLC

high‐performance liquid chromatography

LC/MS

liquid chromatography‐mass spectrometry

M+H⁺

protonated species

M‐H^‐

deprotonated species

SIM

single ion monitoring

MRM

multiple reaction monitoring

MS

mass spectrometry

NH₄F

ammonium fluoride

1. INTRODUCTION

High‐performance liquid chromatography (HPLC) coupled with mass spectrometry (MS) is commonly used in many quantitative applications with demonstrated analytical strengths in separation efficiency, detection, and specificity. ¹ , ² Furthermore, this is shown through the ubiquitous use of this technology in trace‐level clinical monitoring, academic, manufacturing, private and governmental research, and more, highlightlng HPLC/MS as a must‐have lab instrument in today's laboratory. ³ , ⁴

Electrospray ionization (ESI) has become the accepted standard ionization technique in the MS instrumental community, having demonstrated robust efficiency in liquid effluent introduction from the chromatography system to a variety of modern mass spectrometers. ² ESI works well with compounds across a range of polarities, charges, and sizes. There are many factors that may influence ESI ionization efficiency (mobile phase composition, co‐elution of components, source conditions, etc.). The source heat, gas, and voltage parameters are adjusted to optimize the desolvation of liquid effluent from the chromatography system. Although ESI is regarded as a standard and rugged MS introduction technique, it may still require additional optimization beyond routine instrumental tuning, especially for trace‐level analytical measurements in complex matrices.

Improving ESI efficiency is one option to improve sensitivity for LC/MS methods. ⁵ , ⁶ , ⁷ , ⁸ Ionization can be optimized by selecting appropriate mobile phases with pH adjustment to promote the formation of protonated/deprotonated species in solution. ⁹ , ¹⁰ When selecting mobile phase options, the chromatographic separation needs should be balanced with optimal ionization efficiency as both are important to the sensitivity and robustness of the method. There are only a few MS‐friendly mobile phase modifier options without lasting instrumental effects. For low pH adjustment, formic, acetic, and trifluoroacetic acids are all MS‐friendly mobile phase additives that aid in the generation of protonated species (notably for compounds with primary or secondary amines) and provide good chromatographic separatory characteristics. These acids promote the protonation of chromatographic stationary phase free silanols that can cause undesired secondary ion‐exchange interactions and lead to band broadening or peak splitting. ¹¹ For high pH adjustment, ammonium hydroxide is a volatile and common mobile phase modifier that aids in deprotonation for acidic or hydroxy functional groups. Ammonium acetate and ammonium bicarbonate are common volatile mobile phase salts used with ESI detection. ¹² , ¹³ , ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ Both allow for the formation of protonated species (M+H⁺) and NH₄ ⁺ adducts for ESI⁺ or the formation of M‐H⁻ and respective C₂H₃O₂ ⁻ or HCO₃ ⁻ species for ESI⁻. Additionally, they may help maintain a more neutral pH (compared with formic acid or ammonium hydroxide additions), allowing for additional pH stability in analyzing labile compounds. Other volatile mobile phase additives rely on ion‐pairing interactions which cause ion‐suppression or remain in the instrument after use. ¹⁸

Ammonium fluoride (NH₄F) has been reported in the literature across applications as an effective volatile mobile phase additive in improving ESI sensitivity and is hypothesized to work as a sequestration agent. ¹⁹ , ²⁰ , ²¹ , ²² , ²³ , ²⁴ , ²⁵ , ²⁶ , ²⁷ , ²⁸ Sensitivity improvements from ammonium fluoride are not fully characterized. However, ammonium fluoride has been reported to bind competing ions during the ESI process, thereby helping improve the formation of more desirable protonated or deprotonated species. ²⁷ , ²⁹ , ³⁰ Multiple application notes have been published, including analysis of fat‐soluble vitamins, veterinary drugs, pesticides, and hormones, all using ammonium fluoride‐adjusted mobile phase systems. ¹⁹ , ²⁰ , ²¹ , ²² , ²³ , ³¹ Specifically, several application notes identify improved sensitivity for hormone measurement after switching to an ammonium fluoride‐based mobile phase system. ²¹ , ²² , ²³ , ³² , ³³ , ³⁴ , ³⁵ Internal efforts in developing a trace level analytical method for steroids showed an approximate 5‐10 fold improvement in sensitivity by replacing ammonium hydroxide with ammonium fluoride and making no other changes. Additionally, recent publications describe sensitivity improvements by substituting ammonium fluoride as a mobile phase modifier (vs formic acid or ammonium acetate) in metabolism screening work. ²⁴ , ²⁵ Our experience, along with the review of multiple literary references, led us to question how ammonium fluoride might impact ionization efficiency across a variety of analogs when compared against other mobile phases.

In this study, we compare the ESI signal for nine compounds when using either ammonium fluoride, formic acid, or ammonium hydroxide as mobile phase additives across three experiments. In Experiment 1, nine compounds were injected with MS Scanning in ESI⁺ and ESI⁻, with formic acid, ammonium fluoride or ammonium hydroxide adjusted mobile phases, with three different organic compositions (30%, 60%, and 90%). In Experiment 2, each compound was analyzed to identify optimal source conditions for each respective mobile phase set. These conditions were recorded and compiled to generate optimized methods for each compound, according to mobile phase composition. In Experiment 3, a final analysis was performed (in triplicate) for all compounds with optimized source conditions from the prior experiment. Results allowed for an objective comparison of MS sensitivity attributable to mobile phase conditions.

This is the first work, to our knowledge, investigating source optimization to compare the sensitivity of ammonium fluoride as a mobile phase additive with ESI MS. All analyses were performed by flow injection analysis (no chromatographic separation) to focus attention on source optimization and mobile phase composition as key variables in ESI sensitivity enhancement or suppression.

2. MATERIALS AND METHODS

2.1. Materials

Three different sets of mobile phases were prepared for use throughout experimentation. Mobile Phase A was prepared in MS Grade water and Mobile Phase B was prepared in 1:1 Acetonitrile: Methanol (MS grade). Mobile Phase A and Mobile Phase B were adjusted accordingly for each set as described below:

• Mobile Phase Set 1: 0.1% Formic Acid (Fisher, MS Grade)

• Mobile Phase Set 2: 0.2 mM NH₄F (Sigma, 99%)

• Mobile Phase Set 3: 0.1% Ammonium Hydroxide (Fisher, MS Grade)

We chose to use 0.1% concentrations of formic acid and ammonium hydroxide respectively as these are common adjustment concentrations for LC/MS separatory conditions. Ammonium fluoride mobile phases were adjusted to 0.2 mM concentrations based on an application note we followed in our lab in steroid assay development. ²³

Reference standards for antipyrine, tetrabromobisphenol A, leucine enkephalin, thyroxine, verapamil, labetalol, estrone, and acetaminophen were purchased as powders from Sigma Aldrich, with each having purity ≥ 99%. The reference standard for aminopyralid was secured internally (Corteva Agriscience) as a certified reference standard with purity of 99%. Compound structures are shown in Figure 1.

FIGURE 1.

Structure and mass spectrometry (MS) polarity for nine compounds tested.

Stock solutions were prepared by individual weighing and dissolution to yield 1 mg/mL preparations. These stock solutions were diluted 1000‐fold to yield 1 ppm stocks (for use in Experiment 1 and 3) and as a single collective stock containing all analogs at 1 ppm (for use in Experiment 2). All stocks were prepared with dimethyl formamide (Fisher Scientific, Reagent Grade) and stored at 4°C when not in use.

2.2. Chromatographic instrumentation

Analyses were performed with an Agilent 1290 Infinity II HPLC system comprised of a binary pump (G7120A), multisampler (G7167B), and column compartment (G7116B). Isocratic flow rate of 0.5 mL/min and injection volume of 1 µL were maintained throughout all experiments. All injections were performed via loop injection (no HPLC column or guard cartridge).

2.3. MS instrumentation

Analyses were performed with an Agilent G6495A triple quadruple mass spectrometer. Instrument control was facilitated with Agilent Masshunter Workstation Acquisition software (Version 10.0.127). The mass spectrometer was equipped with an Agilent Jet Spray source and operated in positive/ negative polarities with either Q3 scanning for initial parent mass confirmation (Experiment 1), or in multiple reaction monitoring (MRM) mode for source optimization and confirmatory analyses (Experiments 2 and 3).

MRM optimization and testing were performed with Agilent Masshunter Workstation Optimizer software. The program does not allow for optimization in single ion monitoring (SIM) mode. As such, all MRM analyses were performed using the respective protonated/ deprotonated species assigned for both Q1 and Q3 masses. The collision energy was set to 5 V to reduce fragmentation and promote full precursor ion transmission through the collision cell to Q3. This approach allowed us to maximize precursor ion transmission/measurement along with the ability to use the source optimization software.

In Experiment 2, results from source optimization were reviewed for each parameter for each MRM trace using Agilent Masshunter Workstation Quantitative Analysis software (B.09.00). The chromatograms for each injection in each MRM trace were integrated, allowing for sorting by area counts in the table view to easily rank parameters from lowest to highest. The sorting capability allowed for a quick review of results and the selection of optimal conditions for each trace. This process was repeated for all optimization conditions in all mobile phase sets and organic mobile phase conditions for each respective trace. Once completed, the results were compiled and used to generate optimized methods for each compound.

3. RESULTS

3.1. Experiment 1 – ion confirmation by MS scan

Flow injections were performed with mobile phase sets 1, 2, and 3 with Organic (%B) concentration set to 30%, 60%, and 90%, respectively, resulting in nine different isocratic mobile phase conditions in all. Individual 1 ppm stock solutions were analyzed to generate spectra for Experiment 1. Initial MS Scans were performed with conditions shown in Table 1. Spectral data were reviewed against anticipated M‐H⁻ and M+H⁺ ion formation in respective negative/ positive ion mode results with Agilent Masshunter Workstation Qualitative Analysis software (Version B.08.00). If M‐H⁻ or M+H⁺ ions were observed across mobile phase conditions, the ion was included for source optimization in Experiment 2. A summary of results for confirmation of protonated/ deprotonated species by analyte is presented in Table 2.

TABLE 1.

Liquid chromatography‐mass spectrometry (LC/MS) method starting conditions for experiment 1.

LC/MS Parameter	Setting
Gas Temp	290°C
Gas Flow	14 L/min
Nebulizer	20 psi
Sheath Gas Temp	250°C
Sheath Gas Flow	11 L/min
Capillary	5000 V (+), 3000 V (‐)
Nozzle Voltage	2000 V
Scan Range	50–1000 amu
Scan Time	500 ms
Cell Acc. Voltage	5 V

TABLE 2.

Results summary from experiment 1 in confirming observed M+H⁺ and M‐H⁻ ions.

Compound	Mobile Phase Set 1			Mobile Phase Set 2			Mobile Phase Set 3			Optimization	Optimization	Optimization
	MP B Concentration			MP B Concentration			MP B Concentration
	30%	60%	90%	30%	60%	90%	30%	60%	90%	Polarity	M+H+	M‐H‐
Estrone	–	–	None	–	–	–	–	–	None	Neg Only	NA	269.1
Thyroxine	+/‐	+/‐	+/‐	+/‐	+/‐	+/‐	+/‐	+/‐	+/‐	Pos/ Neg	777.6	775.6
Acetaminophen	+/‐	+/‐	–	+/‐	+/‐	+/‐	+/‐	+/‐	–	Pos/ Neg	152.1	150.1
Aminopyralid	+/‐	+/‐	+/‐	+/‐	+/‐	+/‐	+/‐	+/‐	+/‐	Pos/ Neg	207.0	205.0
Antipyrene	+	+	+	+/‐	+/‐	+/‐	+	+	+	Pos Only	189.0	NA
Verapamil	+	+	+	+	+	+	+	+	+	Pos Only	455.3	NA
Tetrabromobisphenol A	–	–	–	–	–	–	–	–	–	Neg Only	NA	538.7
Labetalol	+/‐	+/‐	+/‐	+/‐	+/‐	+/‐	+/‐	+/‐	+/‐	Pos/ Neg	329.2	327.2
Leucine Enkephalin	+/‐	+/‐	+/‐	+/‐	+/‐	+/‐	+/‐	+/‐	+/‐	Pos/ Neg	556.2	554.2

3.2. Experiment 2 – source optimization

Experiment 2 was conducted to identify optimal source conditions for all compounds with each mobile phase with three different organic mobile phase compositions (30%, 60%, and 90%). A stock solution containing all nine compounds at 1 ppm concentration was used for all analyses in Experiment 2. Source optimization was performed with Agilent Masshunter Workstation Optimizer software (Version 10.0.127). Source parameters ranged from low to high value with steps for each parameter designated by the increment size (Table 3). Optimization was performed with a single mobile phase system (formic acid, ammonium fluoride, or ammonium hydroxide adjusted) at each organic mobile phase concentration before changing solvent systems, equilibrating, and proceeding with the next round of source optimization injections.

TABLE 3.

Source optimization parameters and ranges for experiment 2.

	Capillary (V)	Gas Flow (L/min)	Gas Temp (°C)	Ifunnel Low Pressure RF (V)	Ifunnel High Pressure RF (V)	Nebulizer (psi)	Nozzle (V)	Sheath Gas Flow (L/min)	Sheath Gas Temp (°C)
Low Value	2000	11	140	60	60	20	500	4	150
High Value	6000	20	290	200	200	60	2000	12	400
Increment	500	1	20	20	20	5	500	1	25

Example data in a selection of optimal gas flow conditions for acetaminophen in negative ion mode with mobile phase set 2 at 30% organic composition are presented in Figure 2. These optimization data were completed for the other analytes, organic compositions, mobile phase additives, and source conditions to prepare optimized methods.

FIGURE 2.

Example optimization results for acetaminophen M‐H⁻ (sorted by area response) for mobile phase set 2 – 30% organic mobile phase.

Source optimization results for all optimized analyte MRM traces with Mobile Phase Set 2 at 30% organic composition are presented in Table 4.

TABLE 4.

Compiled source optimization results—Mobile phase set 2 ‐ 30% organic mobile phase.

Compound/Polarity	Capillary	Gas Flow	Gas Temp	RF High	RF Low	Nebulizer	Nozzle Volt	Sheath Gas Flow	Sheath Gas Temp
Estrone—Neg	3000	14	280	180	140	25	2000	12	400
Thyroxine—Pos	3000	19	280	180	180	30	1500	10	400
Thyroxine—Neg	3000	20	280	200	180	25	2000	8	400
Acetaminophen—Pos	6000	19	200	100	120	60	1500	10	400
Acetaminophen—Neg	3000	12	220	120	60	60	1000	10	375
Aminopyralid—Pos	6000	16	280	200	200	60	1500	10	400
Aminopyralid—Neg	3000	19	280	60	60	60	1000	11	400
Antipyrene—Pos	3000	12	200	180	140	60	1000	10	225
Verapamil—Pos	3500	12	260	200	200	55	500	11	400
Tetrabromobisphenol A—Neg	3000	12	220	180	140	60	2000	10	400
Labetalol—Pos	4000	14	240	100	120	60	2000	12	400
Labetalol—Neg	3000	17	290	120	200	30	1500	12	400
Leucine Enkephalin—Pos	3000	19	220	120	140	30	2000	8	400
Leucine Enkephalin—Neg	3000	19	220	120	140	30	2000	8	400

An example flow injection chromatogram for acetaminophen with optimized gas flow setting with mobile phase set 2 at 30% organic composition is presented in Figure 3.

FIGURE 3.

Example flow injection chromatogram for acetaminophen M‐H⁻ with mobile phase set 2 – 30% organic composition with optimal gas flow setting (12 L/min).

3.3. Experiment 3 – analyses with optimized methods

For Experiment 3, final optimized MRM methods were created for each mobile phase set and organic condition using the results generated in Experiment 2. To generate final data with optimized methods, individual 1 ppm stock solutions were analyzed in triplicate and processed in Agilent Masshunter Workstation Quantitative Analysis software (B.09.00). Review of data included integration of the flow injection MRM peaks and export of peak areas to Microsoft Excel for final processing and comparison.

Example results for Acetaminophen M‐H⁻ analyzed with 30%, 60%, and 90% B mobile phase composition in mobile phase sets 2 and 3 are presented in Table 5. Example results for Verapamil M+H⁺ as analyzed with 30%, 60%, and 90% B mobile phase composition in mobile phase sets 1 and 3 are presented in Table 6.

TABLE 5.

Mobile phase sensitivity comparison for three replicate injections—Mobile phase set 2 versus 3 ‐ acetaminophen M‐H⁻.

Sample Name	Average Area Counts (x 1e6)	Sensitivity Improvement Amm Hydroxide versus Amm Fluoride
30% organic—Ammonium Hydroxide	0.41	22x
30% organic—Ammonium Fluoride	9.2
60% organic—Ammonium Hydroxide	0.54	20x
60% organic—Ammonium Fluoride	10.9
90% organic—Ammonium Hydroxide	1.4	7.4x
90% organic—Ammonium Fluoride	10.3

TABLE 6.

Mobile phase sensitivity comparison—Mobile phase set 1 versus 3 ‐ verapamil M+H⁺.

Sample Name	Average Area Counts (x 1e6)	Sensitivity Improvement Formic Acid versus Amm Fluoride
30% organic—Formic Acid	46.5	3.5x
30% organic—Ammonium Fluoride	165
60% organic—Formic Acid	37.3	3x
60% organic—Ammonium Fluoride	110
90% organic—Formic Acid	34.2	3.7x
90% organic—Ammonium Fluoride	125

Exported peak areas for each analog and mobile phase condition were plotted and compared. For M‐H⁻ interpretation, results from mobile phase set 3 (Ammonium Hydroxide) were compared with those from mobile phase set 2 (NH₄F). Additionally, for M+H⁺ interpretation, results from mobile phase set 1 (Formic Acid) were compared with those from mobile phase set 2 (NH₄F).

4. RESULTS

Results comparing ammonium fluoride with ammonium hydroxide optimized analyses for M‐H⁻ species are presented in Figure 4. All seven compounds showed ∼2 to 22‐fold improvement in sensitivity. Acetaminophen showed the greatest improvement in sensitivity with ammonium fluoride adjustment, with a 7–22‐fold sensitivity improvement across all organic mobile phase compositions. The greatest sensitivity gains were observed at 30% organic composition for five of the seven analogs. Two of seven analogs (thyroxine and leucine enkephalin) showed the greatest sensitivity improvements at 60% organic composition.

FIGURE 4.

Sensitivity fold improvement—ammonium fluoride versus ammonium hydroxide— M‐H⁻.

Results comparing ammonium fluoride with formic acid for M+H⁺ are presented in Figure 5. Four of seven compounds (acetaminophen, antipyrine, labetalol, and verapamil) showed improved sensitivity with ammonium fluoride. Three of seven compounds (aminopyralid, leucine enkephalin, and thyroxine) showed equivalent or slightly lower sensitivity when analyzed with ammonium fluoride. Acetaminophen showed the greatest improvement in sensitivity overall, similar to ESI⁻ results, with 4 to 11‐fold sensitivity improvement across organic compositions. Four of seven compounds (acetaminophen, antipyrine, labetalol, and thyroxine) showed outperformance in ammonium fluoride with 90% organic composition. Verapamil results showed approximately equivalent improvement across 30%, 60%, and 90% organic compositions. Data show that using ammonium fluoride versus formic acid improved sensitivity for half of the analogs, with greater sensitivity gains observed at higher organic mobile phase composition. Data were not included in the figure for aminopyralid with 90% organic mobile phase as there was a low response with formic acid modifier compared with ammonium fluoride resulting in a dramatic 29‐fold improvement in sensitivity with ammonium fluoride. These results are an outlier compared with performance from other compounds and are not included here.

FIGURE 5.

Sensitivity fold improvement—ammonium fluoride versus formic acid—M+H⁺.

M‐H⁻ results show decreased sensitivity with increased organic concentration while M+H⁺ results interestingly show an opposite trend. This difference is proposed to be due to differences in gas phase intermediates forming in positive ion mode which is not occurring in negative ion mode. NH₄F enhances signal in negative ion mode due to the strong electronegativity of free fluoride with available counter ions (sodium, other free cations, and charged species), forming NaF and resulting in promoted ionization efficiency in forming M‐H⁻. ²⁷ Similarly, in positive ion mode, NH₄F is proposed to improve sensitivity by preventing the formation of M+Na⁺ species by binding free sodium in a similar fashion. ³⁰ Organic and aqueous mobile phases in our study were adjusted with the same amount of modifier for each set in attempts to maintain pH consistency; however, the pH of the solvent system would change with adjustment of the organic content (30% vs. 60% vs. 90%). ³⁶ Additionally, small pH changes have shown differences in adduct formation rates for M+Na⁺ versus M+H⁺ in ESI⁺. ³⁰ , ³⁷ Formic acid adjusted mobile phase with 90% organic (vs. 60% or 30%) would be slightly less acidic and more readily form M+Na⁺, resulting in the reduction of M+H⁺ and the trend we observe of outperformance of ammonium fluoride with higher organic for ESI⁺. This pH drift may be occurring to a lesser extent for NH4F mobile phases (pH measured to be ∼5.2 of aqueous NH4F), but ample free fluoride in solution would help to overcome any shift in equilibrium toward M+Na⁺ adduct formation.

5. CONCLUSIONS

Results show unanimous improvement in sensitivity for M‐H⁻, with greatest gains observed at low (30%) organic mobile phase composition (four compounds) and 60% organic mobile phase composition (three compounds). Results suggest that analytical detection limits may be improved substantially with the substitution of 0.2 mM ammonium fluoride in place of 0.1% ammonium hydroxide, with elution profiles that are low to moderate organic content.

M+H⁺ results showed improvement in overall sensitivity for four of the seven analogs and an equivalent or slight reduction in sensitivity for three analogs. Mobile phase compositions with higher organic (90%) showed the greatest sensitivity improvements for three of four analogs—contrary to M‐H⁻ results. It's proposed that pH drifts to higher pH occurred with increased organic content (90% vs. 30%) and also favored M+Na⁺ formation in formic acid mobile phase at higher pH. Ammonium fluoride would show even greater outperformance, with no shift in M+Na⁺ formation with increased organic content, for this reason.

Results show improved sensitivity with ammonium fluoride under optimized conditions. Fluoride in the mobile phase scavenges for and sequesters competitive ion species, improving overall ionization efficiency for M+H⁺ and M‐H⁻.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflict of interest.

McFadden JR, Ames D. Assessment of ammonium fluoride as a mobile phase additive for sensitivity gains in electrospray ionization. Anal Sci Adv. 2023;4:347–354. 10.1002/ansa.202300031

DATA AVAILABILITY STATEMENT

Data are available on request from the authors.