Two fluoroalcohols as components of basic buffers for liquid chromatography electrospray ionization mass spectrometric determination of antibiotic residues

Dynamic Article LinksC<Analyst

Cite this: Analyst, 2011, 136, 4587

www.rsc.org/analyst PAPER

Two fluoroalcohols as components of basic buffers for liquid chromatographyelectrospray ionization mass spectrometric determination of antibioticresidues†

Karin Kipper,*a Koit Herodes,a Ivo Leitoa and Lembit Neib

Received 14th February 2011, Accepted 9th August 2011

DOI: 10.1039/c1an15123a

Two fluoroalcohols—1,1,1,3,3,3-hexafluoro-2-propanol (HFIP) and 1,1,1,3,3,3-hexafluoro-2-methyl-

2-propanol (HFTB)—were evaluated for the first time as volatile buffer acids in the basic mobile phase

for reversed-phase chromatography with electrospray ionization-mass spectrometric (LC-ESI-MS)

detection of five antibiotics. Chromatographic separation as well as positive and negative ion ESI-MS

intensities using these novel buffer components were compared to traditional buffer systems. Overall,

the highest signal intensities and best chromatographic separation for the five antibiotics (ciprofloxacin,

norfloxacin, ofloxacin, sulfadimethoxine and sulfamethoxazole) were achieved using 5mMHFIP as the

buffer acid to methanol : water mobile phase (pH of the aqueous component adjusted to 9.0 with

ammonium hydroxide). Comparable results were achieved using 5 mMHFTB (pH adjusted to 9.0 with

ammonium hydroxide). The suitability of HFIP for analysis of antibiotic residues in lettuce is

demonstrated.

Introduction

The analysis of basic compounds with reversed-phase (RP)

chromatographic separation in the low pH range often presents

difficulties due to strong interactions with the residual silanol

groups in the silica-based column packing.1 Silanol groups cause

poor peak shapes and low efficiency as well as retention and

column-to-column reproducibility problems.2 The impact of

these problems is wide-ranging, because most of the pharma-

ceuticals (estimated over 70%) have basic properties. At the same

time about 20% are acids.1 Basic compounds are present

predominantly in their protonated form if the pH of the solution

is lower than the pKa value of the base. Protonated, i.e. cationic,

form is polar and has poor retention in the RP column. At a pH

value higher than the pKa of the base, the basic center is

deprotonated. As a result better retention behavior is expected.

Therefore, for the separation of basic compounds by RP liquid

chromatography, a basic buffer solution would be preferable.

Buffer solution components provide the separation of analytes

using the pH or ion pairing effect.3 Buffer solutions more

frequently used in the liquid chromatography-mass spectrometry

(LC-MS) analysis should not suppress the ionization of the

analyte and must be volatile. Use of non-volatile buffer

components causes contamination of the electrospray ionization

aUniversity of Tartu, Institute of Chemistry, 14a Ravila Street, 50411Tartu, Estonia. E-mail: [email protected]; Tel: +372 5666 7504bDepartment of Environmental Protection, Tartu College of TallinnUniversity of Technology, Puiestee 78, 51008 Tartu, Estonia

† Electronic supplementary information (ESI) available. See DOI:10.1039/c1an15123a

This journal is ª The Royal Society of Chemistry 2011

(ESI) source.4 There are only a limited number of basic buffer

systems available for LC-MS analysis with suitable properties5–7

and additional suitable buffer systems would be highly welcome.

As an example, selection of basic buffer components for LC-MS

use (on the example of the Waters XBridge column) is presented

in Table S1 in the ESI†.

Working in a high pH range also sets requirements for the

column. Selection of the column for working in a high pH range

should be made according to its resistance to high pH.

Fluorinated alcohols are a potentially promising class of

compounds to be used as weak acids for preparing buffers of

pH value above 7. 1,1,1,3,3,3-Hexafluoro-2-propanol (HFIP,

pKa ¼ 9.3)8 has been used in several studies8–15 as an additive to

the LC mobile phase at neutral or at slightly basic pH. In these

studies the pH of the buffer was adjusted to 7.0 (ref. 8–10), 7.5

(ref. 13), 7.9 (ref. 11 and 12), 8.2–8.4 (ref. 15), 8.5 (ref. 14) with

triethylamine where oligonucleotides and oligosaccharides were

analyzed. However, these pH values are in most cases signifi-

cantly different from the pKa value of HFIP and the concen-

trations of HFIP were in the range of 100 mM to 800 mM (ca.

2% to 15% by mass), by far exceeding the buffer concentration

levels commonly used for LC-MS applications. Thus the role of

HFIP was rather that of an additional solvent component than

a buffer acid. Using the HFIP as the weak acid and triethyl-

amine (TEA) as the weak base in buffer systems resulted in high

ESI intensities, high efficiency of dissociation of the oligonu-

cleotide–TEA ion pairs and good chromatographic separa-

tion.8,9 Interestingly, this promising approach has not been

extended to analysis of other compounds or to the use of other

polyfluorinated alcohols.

Analyst, 2011, 136, 4587–4594 | 4587

Fig. 2 Acid–base equilibrium of fluoroquinolones at basic pH.

In this study LC-ESI-MS analysis of five antibiotics in lettuce

plant was investigated using acidic and basic buffer solutions.

The basic buffer solutions used volatile compounds HFIP and

1,1,1,3,3,3-hexafluoro-2-methyl-2-propanol (HFTB) as addi-

tives. The pH was adjusted to 9.0 or 10.0 with triethylamine or

NH4OH. The results were compared with those obtained using

mobile phases with more common volatile buffer solutions.

Separation of antibiotics in environmental samples (sewage

sludge and wastewater)16–20 and in plants21 has been recently

under study by several groups. Sewage sludge may be regarded as

hazardous waste but it can also be used as a fertilizer. Regarding

the latter, its safety with respect to pharmaceutical residues (in

addition to other potential factors, e.g. pathogens, heavy metals,

etc.) must be assessed before use.

The selection of analytes for this study from the possible range

of antibiotics was made considering the stability and potential of

accumulation of the residues. The following five antibiotics were

chosen for the study: three fluoroquinolones (FQ-s): cipro-

floxacin (CIP), norfloxacin (NOR), ofloxacin (OFL) and two

sulfonamides (SA-s): sulfadimethoxine (SDM) and sulfame-

thoxazole (SMX). Chemical structures of these antibiotics are

shown in Fig. 1.

For SMX pKa1 and pKa2 values are 1.49 and 5.41, and for

SDM 2.11 and 6.17, respectively.22 Acid–base behavior of FQ-s

has been studied by several researchers but there is still no

agreement in published data. The number of pKa values deter-

mined for FQ-s is two,23 three24 or four.25 Also the assignment of

pKa values to acidic/basic sites is a topic of controversy. These

difficulties with studies of acid–base properties of FQ-s are also

mentioned in the review article.26 As the pH range from 9 to 10 is

investigated in this work, only the last (most basic) pKa is of

importance. Therefore, in order to avoid confusion, we denote

this pKa as pKax. Respective acid–base equilibrium is presented

in Fig. 2. The pKax values are adapted from the work of Barbosa

et al.:27 CIP 8.62, NOR 8.38 and OFL 8.11.

Chromatographic separation of these compounds using acidic

buffer solution has been problematic,23 but since mass-spectro-

metric detection was used in selected reaction monitoring mode,

poor chromatographic separation was not considered an issue.23

Fig. 1 Chemical structures of used antibiotics.

4588 | Analyst, 2011, 136, 4587–4594

Poor separation, however, can result in the serious matrix effects.

A change of chromatographic conditions did not give satisfac-

tory separation.23 Therefore, the buffer solution pH change as

a possible remedy was taken into consideration in this study.

Experimental

Instrumentation

Chromatographic separation of the analytes was carried out on

the Agilent Series 1100 LC-MSD Trap XCT (Agilent Technol-

ogies, Santa-Clara, CA, USA) equipped with a binary pump,

a degasser, an auto-sampler and a column thermostat. For

detection diode array detector and ESI-MS were used in series.

ESI-MS detection was carried out in alternating positive

and negative ion detection mode. In positive mode protonated

[M + H]+ forms were detected: SDM atm/z 311; SMX 254; NOR

320; CIP 332; OFL 362. In negative mode deprotonated

[M � H]� ions were detected: SDM at m/z 309; SMX 252; NOR

318; CIP 330; OFL 360. Default parameters for ESI andMSwere

used for all the experiments (nebulizer gas pressure was 40 psi,

dry gas flow was 10 L min�1, dry gas temperature was 350 �C,capillary voltage was 5000 V, detected mass range was from m/z

100 to 1000 and target mass for compounds was m/z 350).

LC-UV and MS instruments were controlled by Agilent Chem-

station for LC 3D rev. A.10.02 (Agilent Technologies) and LC/

MSD TrapControl ver. 5.2 (Bruker Daltonik GmbH, Germany).

Data analysis (including determination of retention and peak

shape parameters) was carried out using Chemstation software

(Agilent Technologies) and Data Analysis for LC/MSD Trap

Version 3.2 (Bruker Daltonik GmbH).

Chemicals

Pharmaceuticals were purchased from Riedel-de-Ha€en (Seelze,

Germany)—three FQ-s: CIP (purity 99.8%), NOR (purity

99.9%) and OFL (purity 99.3%); two SA-s: SDM (purity 99.4%)

and SMX (purity 99.9%). Acetonitrile and methanol were

obtained from J. T. Baker (Deventer, The Netherlands), formic

acid and ammonia fromRiedel-de-Ha€en, ammonium acetate and

1-methylpiperidine from Fluka (Buchs, Germany). TEA, HFIP

and HFTB were purchased from Sigma (St Louis, MO, USA).

All solvents were of reagent grade or higher quality. Water was

purified in-house using a Milli-Q plus system from Millipore

(Bedford, USA).

Chromatographic conditions

Five antibiotics were chromatographed using a Waters XBridge

C18 column (150 mm � 3 mm, 3.5 mm) equipped with a Waters

Guard Cartridge (20 mm � 4.6 mm) (Waters, Milford, USA).

This stationary phase was used as it is usable in the pH range

This journal is ª The Royal Society of Chemistry 2011

from 1 to 12.Mobile phases composed of different buffers (see the

next section) and methanol (solvent B) were used. Gradient

elution at a flow rate of 0.3mLmin�1 started at 10%methanol and

was raised to 50% within 45 min, after that the methanol

concentration was raised to 100% within 5 min. The methanol

concentration was kept at 100% for 5min and lowered to 10% in 5

min and was equilibrated at 10% for 5 min. Column temperature

was set to 30 �C and the injection volume was 10 mL (Table 1).

Standard and buffer solutions

Stock solutions of the analytes at 1 mg mL�1 in the appropriate

solvent were prepared. The stock solution for SDM was 0.5 mg

mL�1 due to its poor solubility. The working standard solution

contained 5 antibiotics at 0.1 mg mL�1 and dilutions (10 mg mL�1

and 1 mg mL�1) with water from the working standard were

made. The stock solution was stored at �20 �C. Fresh working

standard solutions were made daily.

Composition of used mobile phases is presented in Table 2.

The method

Standard mixtures, containing almost equal (weighed) amounts

of five antibiotics, were prepared at two concentration levels, 1

mg mL�1 and 10 mg mL�1 in water. The influence of different

buffer solutions on the chromatographic separation and subse-

quent ESI ionization was investigated. The lower standard

concentration was used to compare differences in ESI signal

intensities in the positive ionization mode. For comparison in the

negative mode higher concentrations were required.

HFIP is miscible with water, methanol, 2-propanol and hexane

but is claimed to be immiscible with acetonitrile.8 Our studies

showed that in the concentration range 1 mM to 10 mM HFIP

buffer solutions with pH 9 and 10 are miscible with acetonitrile

and can be used as buffer components for LC-MS analysis.

When the concentration of HFIP in the buffer solution exceeded

20 mM, then the solution appeared to be immiscible with

acetonitrile. Consequently HFIP buffer solution can be used in

a mobile phase using acetonitrile as organic modifier; however, in

our study methanol was used because better separation was

achieved with methanol. With methanol, gradient elution started

at 10% of organic component. To achieve comparable retention

with acetonitrile even lower organic content proved to be

necessary. However, low content of organic modifier in eluent is

not recommended for C18 columns due to the possibility of

stationary phase collapse. The low content of organic modifier

also hinders the ionization process in the ESI source.

Table 1 ESI-MS compatible (volatile) buffer components recommended for

Additive/buffer pKa

4-Methylmorpholine �8.4Ammonia (NH4OH) 9.2Ammonium bicarbonate 10.3 (HCO3

�) 9.2 (NH4+)

Ammonium (acetate) 9.2Ammonium (formate) 9.21-Methylpiperidine 10.2Triethylamine (as acetate salt) 10.7Pyrrolidine 11.3

This journal is ª The Royal Society of Chemistry 2011

Results and discussion

Buffer solution influence on the chromatographic separation of

compounds

The initial separation of antibiotics was carried out using elution

under acidic conditions with AAF 2.8 (see Table 2 for designa-

tion of buffer solutions) and methanol as our in-house standard

method. Chromatographic separation of the antibiotics was

problematic, and the peaks of CIP, SMX and NOR overlapped.

As the change of the organic solvent to acetonitrile and modifi-

cation of gradient conditions did not provide better separation,

the possibility of shifting the mobile phase pH into the basic

range was taken into consideration. Alternatives of basic buffer

components are presented in Table S1 in the ESI†. Buffer solu-

tion pH range from 9 to 10 was carefully studied and buffer

components 1-MePip 9.85, TEAA 10.0, CH3COONH4 9.0 and

10.0 were selected for further study as reference buffers for the

HFIP/NH4OH and HFTB/NH4OH systems. Separation and ESI

signal intensities in positive and negative ion mode (expressed as

peak heights) of analytes are presented in Table S3 in the ESI†.

Overlapping of some analyte peaks occurred when using

1-MePip 9.85, TEAA 10.0 and CH3COONH4 10.0. Satisfactory

separation was achieved using CH3COONH4 9.0, HFIP/NH4OH

9.0 and HFTB/NH4OH 9.0. Chromatographic separation of

antibiotics using four different buffers is presented in Fig. 3.

HFIP and HFTB as weak acids and TEA and ammonia as

weak bases have acidic dissociation constants (pKa) 9.3, 9.6 and

10.7, 9.2, respectively.

In further discussion all changes in retention of analytes are

presented as per increase of mobile phase pH from 9 to 10.

For the CH3COONH4 buffer the retention times of SA-s did

not change. This observation is easy to rationalize—pKa values

of SA-s are much lower than 9 and the pH increase does not

cause a change in protonation equilibrium of the SA-s. Using

HFIP/NH4OH and HFTB/NH4OH buffers SA-s retention times

increased. This change of retention times must be caused by the

nature of HFIP and HFTB. HFIP and HFTB are predominantly

protonated at pH 9 and are predominantly deprotonated at pH

10, e.g. at pH 9, they are less polar than at pH 10. Therefore, at

pH 9 the fluoroalcohols effectively compete with the analytes for

the stationary phase surface, which is indicated by shorter

retention times of SA-s at pH 9. With both pH values retention

times of SA-s are longer in the case of CH3COONH4 as

compared to HFIP and HFTB. This also indicates that the flu-

oroalcohols compete with analyte molecules for the stationary

phase surface.

Waters XBridge columns at high pH by waters

Buffer range Recommended concentration

7.4–9.4 10 mM or less8.2–10.2 Below 10 mM6.8–11.3 5–10 mM range8.2–10.2 1–10 mM range8.2–10.2 1–10 mM range9.3–11.3 1–10 mM range9.7–11.7 0.1–1.0% range10.3–12.3 —

Analyst, 2011, 136, 4587–4594 | 4589

Table 2 Composition of the buffer solutions used in this study

Designation Composition pH

AAF 2.8 1 mM ammonium acetate in 0.1% formic acid 2.8TEAA 10.0 5 mM ammonium acetate, pH adjusted to 10.0 with triethylamine 10.0CH3COONH4 9.0 5 mM ammonium acetate, pH adjusted to 9.0 with ammonia 9.0CH3COONH4 10.0 5 mM ammonium acetate, pH adjusted to 10.0 with ammonia 10.01-MePip 9.85 5 mM 1-methylpiperidine, pH adjusted to 9.85 with ammonia 9.85HFIP/NH4OH 9.0 5 mM HFIP, pH adjusted to 9.0 with ammonia 9.0HFIP/NH4OH 10.0 5 mM HFIP, pH adjusted to 10.0 with ammonia 10.0HFTB/NH4OH 9.0 5 mM HFTB, pH adjusted to 9.0 with ammonia 9.0HFTB/NH4OH 10.0 5 mM HFTB, pH adjusted to 10.0 with ammonia 10.0HFIP/TEA 9.0 5 mM HFIP, pH adjusted to 9.0 with triethylamine 9.0

When using CH3COONH4 buffer, the retention times of FQ-s

decreased. At pH 10 the FQ-s exist mostly in anionic form

(Fig. 2) while at pH 9 some zwitterionic forms are still present.

Similar trends in retention behavior of FQ-s have been noted in

the pH range from 6 to 7.5 (ref. 28) and 7.5 to 10 (ref. 23) using

non-ion-interaction buffer components.

In the case ofHFTB/NH4OHbuffer the retention times of FQ-s

increased. This is in contrast to the effect observed in the case of

CH3COONH4 buffer. In the case of HFIP/NH4OH buffer,

retention times of FQ-s are nearly the same. The pH of a solvent

has a similar effect on the solute regardless of the compounds used

to create the pH, e.g. CIP is protonated to the same extent in

CH3COONH4 9 as in HFTB/NH4OH 9 buffer. However, these

pH values refer to the buffer solution before mixing with the

Fig. 3 Chromatographic separation of five antibiotics. Used eluent

buffer solutions: (A) AAF: 1 mM ammonium acetate and 0.1% formic

acid, pH 2.8. (B) TEAA: 5 mM triethylammonium acetate buffer, pH

10.0. (C) CH3COONH4: 5 mM ammonium acetate, pH 9.0. (D) HFIP: 5

mM hexafluoroisopropanol, pH adjusted to 9.0 with ammonium

hydroxide.

4590 | Analyst, 2011, 136, 4587–4594

organic solvent. The addition of the organic solvent may have

different effect on the buffer solutions created using different

acid–base systems. In theCH3COONH4buffer solution ammonia

acts as the weak base, while acetic acid is virtually fully deproto-

nated. In theHFTB/NH4OH andHFIP/NH4OHbuffer solutions

both compounds are present as mixtures of protonated and

deprotonated forms but at different ratios. Thus the effective pH

in the eventual mobile phase can be different in all three cases.

Further effects to consider are the competition of the alcohols for

the active sites of the stationary phase and the complex acid–base

behavior of the FQ-s (present partly as zwitterions at the used pH

level), which in turn also depends on the organic solvent. The

observed retention time changes are probably due to a complex

interplay of all these effects.

Buffer systems of weak acid (HFIP) and another weak base

(TEA) were compared with fluoroalcohol/NH4OH buffers at pH

9 and 10. Separation was achieved using HFIP/TEA at pH 9, but

very low ESI-MS signal intensities were observed and analytes

remained undetected at 1 mg mL�1 level.

During the separations with TEAA and HFIP/TEA buffers

TEA ions can form ion-pairs with FQ-s deprotonated carboxyl

groups and SA-s sulfonamide groups and retention of antibiotics

increases. Retention increase is more significant for TEAA

buffer. Retention increased using HFIP/TEA buffer compared to

the HFIP/NH4OH buffer due to the analyte ion-pairing effect

with TEA. Using ammonium hydroxide as an additive for the

buffer solution the ion-pair formation between negatively

charged antibiotics and ammonium ion does not alter retention

times to a significant extent.

Buffer solution influence on the ESI signal intensities

As limit of detection (LOD) and limit of quantitation (LOQ) are

directly related to the height of chromatographic peak rather

than its area, the influence of eluent composition and pH on the

heights of extracted ion chromatogram peaks was assessed.

In the positive mode ESI (+ESI), the highest signals for all the

analytes were observed using the AAF 2.8 buffer (Fig. 4). This

observation can be easily rationalized as at pH 2.8 FQ-s are

present as cations already in the solution phase. SA-s seem to be

easily ionized although their pKa1 values are lower than 2.8.

When considering the basic pH range, HFIP/NH4OH 9.0 and

HFTB/NH4OH 9.0 provide equally high signal intensities for all

the analytes. Very low analyte signal intensities are observed for

1-MePip and buffers containing TEA: SMX was not detected

using HFIP/TEA 9.0 buffer; none of the five analytes were

This journal is ª The Royal Society of Chemistry 2011

Fig. 4 Comparison of normalized signal intensities in positive ESI mode

between different buffer compounds.Fig. 5 Comparison of normalized signal intensities in negative ESI

mode between different buffer compounds.

Fig. 6 Background spectra using HFIP and HFTB buffer components

in negative ESI mode. Flow rate 0.3 mL min�1. Eluent composition 50%

HFIP or HFTB buffer with 50% methanol.

detected at 1 mg mL�1 concentration level with 1-MePip 9.85. In

contrast to CH3COONH4 9.0 and 10.0 buffers, for which

comparable signal intensities for all the analytes were obtained,

HFIP/NH4OH and HFTB/NH4OH cause very low signal

intensities at pH 10. In the positive ESI mode no interfering ions

were observed.

In the negative mode ESI (�ESI) at pH 2.8 the signal intensity

of FQ-s is among the lowest of all (Fig. 5). FQ-s are in cationic

form at this pH and conversion of solution-phase cations into

gas-phase ions is not an efficient process. For SA-s the most

intense signals are observed at pH 2.8, which may be attributed

to the ease of penetration of neutral SA molecules to the ESI

droplet surface.

In the basic buffers NOR, OFL and SDM exhibit most intense

signals in HFIP/NH4OH 9.0 buffer and SMX in HFTB/NH4OH

9.0. Unexpectedly, CIP yields most intense signal in 1-MePip.

The weakest signals were observed in the case of HFIP/NH4OH

10.0 and HFTB/NH4OH 10.0 as was the case in +ESI.

Higher signals (higher chromatographic peaks) result in

improved LODs and LOQs only if accompanied by an increase

in signal-to-noise ratios (s/n-s). The observed s/n-s were higher

for higher signals in +ESI and �ESI mode. For example in the

case of +ESI HFIP/TEA 9.0 buffer peaks of all the analytes were

low and s/n-s ranged from only 5 to 166, but in the case of HFTB/

NH4OH 9.0 buffer, which resulted in high analyte peaks, s/n-s

were from 405 to 2435.

1-MePip andTEAbuffers generated background signal in+ESI

mode; [M + 1] is present atm/z 100 and 102 for 1-MePip and TEA

respectively. Both 1-MePip and TEA gave adduct ions with FQ-s.

Adduct ions were not formed with SA-s and 1-MePip and TEA.

For �ESI HFIP and HFTB buffer systems have a disadvan-

tage—the background signal generated by the reagents. Fig. 6

shows the background spectra of the eluents with the HFIP and

HFTB buffers. In the negative ESI mode [M � 1]� and the dimer

[2M� 1]� are present atm/z 181, 363 and 167, 335 for HFTB and

HFIP respectively. HFIP and HFTB did not give adduct ions

with analytes.

Buffer solution influence on peak shapes

Peak widths and symmetry factors were compared for all used

buffer solutions and are shown in Table 3. Peak symmetry

factors were obtained using Chemstation software, values lower

This journal is ª The Royal Society of Chemistry 2011

than unity indicate a tailing peak, and fronting peaks are char-

acterized by peak symmetry factors exceeding unity.

Using HFIP/NH4OH 9.0 and HFTB/NH4OH 9.0 peak

symmetry factors were similar. Peak symmetries using HFIP/

TEA 9.0 buffer were comparable with HFIP/NH4OH 9.0 and

HFTB/NH4OH 9.0, except for SDM with lower symmetry value

Analyst, 2011, 136, 4587–4594 | 4591

Table 3 Comparison of antibiotic separation and ESI signal intensity using different buffers

Buffer solution

Concentration of an analyte

1 mg mL�1 10 mg mL�1

c/mM pH tR/min AnalytePeak height(+ESI)

Peak height(�ESI)

Peak height(+ESI)

Peak height(�ESI)

Peakwidth/min Symm.

AAF 2.8 1 2.8 18.9 OFL 6.8 � 107 6.9 � 103 6.6 � 108 1.8 � 105

19.8 NORa 3.2 � 107 4.0 � 108 6.6 � 104

20.3 CIPa 4.0 � 107 3.7 � 108 5.7 � 104

20.3 SMXa 2.9 � 107 1.8 � 105 1.5 � 108 2.5 � 106

26.1 SDM 6.0 � 107 7.9 � 105 3.9 � 108 5.7 � 106 0.5333 1.3TEAA 10.0 5 10 12.1 SMX 7.7 � 104 7.6 � 105 0.1867 0.47

19.6 SDM 2.1 � 105 2.1 � 105 2.1 � 106 2.2 � 106 0.1956 0.5720.9 NORa 4.6 � 105 1.2 � 104 4.5 � 106 1.1 � 105

21.2 CIPa 3.6 � 105 1.5 � 104 3.6 � 106 1.4 � 105

34.5 OFL 5.9 � 105 7.8 � 104 6.0 � 106 8.0 � 105 0.2356 0.59HFTB/NH4OH 9.0 5 9.0 4.1 SMX 4.1 � 107 1.2 � 105 2.1 � 107 2.6 � 105 0.2 1.25

7.5 SDM 1.5 � 107 5.2 � 105 1.4 � 108 2.0 � 106 0.2 1.88.9 NOR 7.6 � 106 1.6 � 104 9.4 � 107 2.1 � 105 0.2111 1.1810.3 CIP 1.4 � 107 1.9 � 104 2.0 � 108 3.2 � 105 0.1644 1.1616.5 OFL 4.7 � 107 5.8 � 104 5.1 � 108 9.9 � 105 0.1633 1.61

HFIP/TEA 9.0 5 9.0 4.4 SMX 3.8 � 104 3.2 � 105 0.1533 1.348.6 SDM 1.5 � 106 4.5 � 105 5.1 � 106 1.5 � 106 0.1833 1.0410.8 NOR 2.8 � 106 1.1 � 107 1.7 � 105 0.2467 1.2312 CIP 4.8 � 106 1.4 � 107 2.6 � 105 0.2378 1.1218.9 OFL 1.2 � 107 1.9 � 106 0.2244 1.61

CH3COONH4 10.0 5 10.0 7.8 SMX 8.8 � 105 9.5 � 104 1.9 � 107 5.0 � 105 0.3933 1.4712.2 SDMa 5.7 � 106 1.5 � 106 1.4 � 108 3.6 � 106 0.1667 1.5612.4 NORa 5.6 � 106 3.1 � 104 5.1 � 107 1.5 � 105 0.1833 1.0113.9 CIP 7.6 � 106 3.2 � 104 7.3 � 107 2.0 � 105 0.1622 1.1321.2 OFL 2.2 � 107 6.3 � 104 2.3 � 108 5.9 � 105 0.1583 1.52

CH3COONH4 9.0 5 9.0 7.8 SMX 2.5 � 106 1.8 � 105 1.1 � 107 5.5 � 105 0.2756 1.2912.4 SDM 1.8 � 107 1.1 � 106 6.5 � 107 3.6 � 106 0.1467 0.7714.8 NOR 3.7 � 106 8.0 � 103 7.7 � 107 2.2 � 105 0.1889 0.3615.3 CIP 6.3 � 106 9.0 � 103 6.0 � 107 2.6 � 105 0.1778 0.4522.8 OFL 2.5 � 107 2.8 � 104 2.8 � 108 1.0 � 106 0.1822 0.55

1-MePip 9.85 5 9.85 11.3 SMX 4.7 � 104 3.8 � 105 0.1844 0.6616.2 SDM 3.2 � 105 5.4 � 106 2.3 � 106 0.1667 0.5217.0 NORa 9.2 � 104 6.5 � 104 0.2333 0.2817.2 CIPa 4.3 � 106 2.7 � 106

23.5 OFL 4.8 � 106 5.9 � 105 0.1467 0.53HFTB/NH4OH 10.0 5 10.0 5.6 SMX 6.7 � 104 1.4 � 103 2.2 � 105 3.2 � 104 0.305 —

9.8 SDMa 1.9 � 105 1.9 � 104 1.0 � 106 1.2 � 105 0.1781 1.6310.3 NORa 1.7 � 105 4.1 � 103 1.3 � 106 3.7 � 104 0.3267 1.0311.6 CIP 2.5 � 105 6.3 � 103 1.1 � 106 6.4 � 104 0.34 1.1618 OFL 7.7 � 105 2.9 � 104 8.5 � 106 1.9 � 105 0.3133 1.53

HFIP/NH4OH 10.0 5 10.0 5.2 SMX 6.9 � 104 2.0 � 103 2.8 � 105 3.1 � 104 0.2633 —9 SDM 3.7 � 105 1.6 � 104 1.4 � 106 9.4 � 104 0.2717 1.6310.1 NOR 3.3 � 105 3.9 � 103 1.7 � 106 2.4 � 104 0.4233 1.0411.3 CIP 5.4 � 105 5.5 � 103 3.9 � 106 3.8 � 104 0.3 1.1718.4 OFL 1.6 � 106 1.4 � 104 1.1 � 107 1.4 � 105 0.4167 1.62

HFIP/NH4OH 9.0 5 9.0 4.5 SMX 4.4 � 106 1.4 � 105 3.4 � 107 5.6 � 105 0.2044 1.018.2 SDM 1.8 � 107 1.2 � 106 1.6 � 108 3.8 � 106 0.2067 0.89.9 NOR 1.0 � 107 3.3 � 104 1.1 � 108 5.5 � 105 0.2156 0.411.1 CIP 1.9 � 107 5.7 � 104 2.0 � 108 6.1 � 105 0.1822 0.4618.6 OFL 3.9 � 107 1.1 � 105 4.3 � 108 4.1 � 106 0.1867 0.57

HFIP/NH4OH 9.0 1 9.0 3.7 SMX 2.3 � 105 1.0 � 104 6.5 � 105 9.1 � 104 0.14 —6 SDM 2.5 � 105 3.7 � 104 1.7 � 106 1.6 � 105 0.3444 1.510 NORa 1.3 � 105 2.1 � 103 8.2 � 105 3.5 � 104 0.73 1.4410.5 CIPa 1.4 � 105 3.5 � 103 1.4 � 106 5.5 � 104 0.5133 1.1117.4 OFL 4.0 � 105 1.8 � 104 8.4 � 106 1.8 � 105 0.68 1.66

HFIP/NH4OH 9.0 10 9.0 4.6 SMX 9.8 � 105 2.7 � 103 3.4 � 105 4.9 � 104 0.2567 —8.2 SDM 2.7 � 105 1.7 � 104 1.6 � 106 2.0 � 105 0.33 1.8310.1 NOR 2.5 � 105 3.2 � 103 1.5 � 106 2.8 � 104 0.64 1.1911.6 CIP 3.1 � 105 5.5 � 103 3.1 � 106 4.8 � 104 0.5067 1.1619.3 OFL 1.9 � 106 3.2 � 104 1.8 � 107 1.2 � 105 0.4867 1.67

a Overlapping peaks.

4592 | Analyst, 2011, 136, 4587–4594 This journal is ª The Royal Society of Chemistry 2011

Fig. 7 Lettuce sample spiked with antibiotics. (A) 5 mM HFTB buffer,

pH 9.0 adjusted with ammonium hydroxide. (B) 5 mM HFIP buffer, pH

9.0 adjusted with ammonium hydroxide. (C) 5 mMCH3COONH4 buffer,

pH 9.0 adjusted with ammonium hydroxide.

and SMX and NOR having higher value. For CH3COONH4 9.0

symmetries were lower for all compounds compared with other

buffer solutions at same pH. Peak widths using all buffer solu-

tions at pH 9 were under 0.28 min and the narrowest peaks for

CIP and OFL were achieved using HFIP/NH4OH 9.0. The

narrowest peaks for SDM and NOR were achieved using

CH3COONH4 9.0. For SMX the narrowest peaks were achieved

using HFIP/TEA 9.0.

Buffer concentration influence on chromatographic separation

and ESI signal intensities

Since the buffer containing HFIP and TEA was recommended

for oligonucleotide separation8–14 and better ESI signal

This journal is ª The Royal Society of Chemistry 2011

intensities were achieved using this buffer, differences in our

study could originate from the used buffer concentration. For

oligonucleotide separation the concentration of HFIP used was

400 mM; in our study the concentrations were 1 mM to 10 mM as

regular concentrations for LC-MS.

HFIP/NH4OH buffer solutions were prepared at three

concentration levels 10 mM, 5 mM and 1mM, pH adjusted to 9.0

using NH4OH. Comparison of the three concentrations is

demonstrated in Table 3.

At higher concentration ESI signal intensities decrease both in

the negative and positive ionization mode. If the HFIP concen-

tration in buffer solution decreases to 1 mM, the ESI signal

decreases as well as the overlapping occurs for CIP and NOR.

Application

The possibility of accumulation of pharmaceutical residues from

soil into plants must be assessed before sewage sludge can be used

as fertilizer.29 Determination of antibiotic residues in plant

material was demonstrated in the example of lettuce sample

spiked with the five antibiotics. Lettuce samples were obtained

from a local supermarket. Dried and ground samples were

extracted with 1% acetic acid. The extract was purified and

concentrated by solid phase extraction (SPE) using hydrophilic–

lipophilic balance solid phase cartridges. After concentrating the

SPE samples in N2 stream the samples were dissolved with

buffer/methanol (90/10), spiked and analyzed by LC-MS.

Chromatographic separation using three buffer solutions (HFIP,

HFTB and ammonium acetate, for all pH adjusted to 9.0 with

NH4OH) was compared. Out of the three buffer compositions

tested for lettuce, matrix HFIP and HFTB provided the best

resolution and signal intensity. Antibiotic separation in lettuce

sample using positive ESI mode is demonstrated in Fig. 7. A

blank lettuce sample was also analyzed (not shown in the figure).

Conclusions

HFIP and HFTB were evaluated as volatile buffer acids in basic

mobile phase for RP chromatography with LC-ESI-MS. Chro-

matographic parameters and positive and negative ion ESI-MS

intensities were evaluated using novel buffer additives compared

with traditional buffer additives in basic conditions detecting five

antibiotics.

Both HFIP and HFTB showed the highest signal intensities

and best chromatographic separation using 5 mM fluoroalcohol

as the buffer acid in methanol : water mobile phase (pH of the

aqueous component adjusted to 9.0 with ammonium hydroxide).

As an application the suitability of 5 mM HFIP buffer (pH 9.0)

was demonstrated for analysis of antibiotic residues in lettuce.

Acknowledgements

This work was supported by the grant no. 7127 from the Esto-

nian Science foundation, by the target financing project no.

SF0180061s08 from the Ministry of Education and Science of

Estonia and by Estonian Environmental Investment Centre.

Notes and references

1 D. V. McCalley, J. Chromatogr., A, 2010, 1217, 858–880.

Analyst, 2011, 136, 4587–4594 | 4593

2 A. Espada and A. Rivera-Sagredo, J. Chromatogr., A, 2003, 987, 211–220.

3 J. Dai and P. W. Carr, J. Chromatogr., A, 2005, 1072, 169–184.4 L. R. Snyder, J. J. Kirkland and J. W. Dolan, Introduction to ModernLiquid Chromatography, John Wiley & Sons Inc, New York, 3rd edn,2010, p. 810.

5 J. J. Zhao, A. Y. Yang and J. D. Rogers, J. Mass Spectrom., 2002, 37,421–433.

6 K. Petritis, H. Dessans, C. Elfakir and M. Dreux, LC GC Eur., 2002,2, 98–102.

7 L. R. Snyder, J. J. Kirkland and J. L. Glajch, Practical HPLCMethodDevelopment, John Wiley & Sons Inc, New York, 2nd edn, 1997, pp.296–310.

8 A. Apffel, J. A. Chakel, S. Fischer, K. Lichtenwalter andW. S. Hancock, Anal. Chem., 1997, 69, 1320–1325.

9 A. Apffel, J. Chakel, S. Fischer, K. Lichtenwalter andW. Hancock, J.Chromatogr., A, 1997, 777, 3–21.

10 C. G. Huber and H. Oberacher, Mass Spectrom. Rev., 2001, 20, 310–343.

11 M. Gilar, K. J. Fountain, Y. Budman, U. D. Neue, K. R. Yardley,P. D. Rainville, R. J. Russell, II and J. C. Gebler, J. Chromatogr.,A, 2002, 958, 167–182.

12 M. Beverly, K. Hartsough, L. Machemer, P. Pavco and J. Lockridge,J. Chromatogr., B: Anal. Technol. Biomed. Life Sci., 2006, 835, 62–70.

13 G. Zhang, J. Lin, K. Srinivasan, O. Kavetskaia and J. N. Duncan,Anal. Chem., 2007, 79, 3416–3424.

14 Y. Zou, P. Tiller, I.-W. Chen, M. Beverly and J. Hochman, RapidCommun. Mass Spectrom., 2008, 22, 1871–1881.

15 C. E. Doneanu, W. Chen and J. C. Gebler, Anal. Chem., 2009, 81,3485–3499.

4594 | Analyst, 2011, 136, 4587–4594

16 J. R. Marengo, R. A. Kok, G. K. O’Brien, R. R. Velagaletti andJ. L. Stamm, Environ. Toxicol. Chem., 1997, 16, 462–471.

17 T. Christian, R. J. Schneider, H. A. F€arber, D. Skutlarek,M. T. Meyer and H. E. Goldbach, Acta Hydrochim. Hydrobiol.,2003, 31, 36–44.

18 S. Thiele-Bruhn and M. O. Aust, Arch. Environ. Contam. Toxicol.,2004, 47, 31–39.

19 R. H. Lindberg, P. Wennberg, M. Johansson, M. Tysklind andB. A. V. Andersson, Environ. Sci. Technol., 2005, 39, 3421–3429.

20 M. Lillenberg, S. Yurchenko, K. Kipper, K. Herodes, V. Pihl,R. Lohmus, M. Ivask, A. Kuu, S. Kutti, S. V. Litvin and L. Nei,Int. J. Environ. Sci. Technol., 2010, 7, 307–312.

21 L. Migliore, G. Brambilla, S. Cozzolino and L. Gaudio, Agric.Ecosyst. Environ., 1995, 52, 103–110.

22 M. Lillenberg, S. Yurchenko, K. Kipper, K. Herodes, V. Pihl,K. Sepp, R. Lohmus and L. Nei, J. Chromatogr., A, 2009, 1216,5949–5954.

23 V. Sanz-Nebot, I. Toro and J. Barbosa, J. Chromatogr., A, 2001, 933,45–56.

24 C.-E. Lin, Y.-J. Deng, W.-S. Liao, S.-W. Sun, W.-Y. Lin andC.-C. Chen, J. Chromatogr., A, 2004, 1051, 283–290.

25 Z. Qiang and C. Adams, Water Res., 2004, 38, 2874–2890.26 S. Babi�c, A. J. M. Horvat, D. M. Pavlovi�c and M. Ka�stelan-Macan,

TrAC, Trends Anal. Chem., 2007, 26, 1043–1061.27 J. Barbosa, R. Berg�es, I. Toro and V. Sanz-Nebot, Talanta, 1997, 44,

1271–1283.28 C. Pistos, A. Tsantili-Kakoulidou and M. Koupparis, J. Pharm.

Biomed. Anal., 2005, 39, 438–443.29 A. J. Moffat, A. T. Armstrong and J. Ockleston, Biomass Bioenergy,

2001, 20, 161–169.

This journal is ª The Royal Society of Chemistry 2011