Matrix-enhanced secondary ion mass spectrometry (ME SIMS) using room temperature ionic liquid matrices

Article abstract of DOI:10.1021/ac100133c

Room temperature ionic liquids (ILs) have many applications including as matrices in MALDI. We wished to investigate the efficacy of ILs as matrices in time-of-flight secondary ion mass spectrometry and in mass spectrometric imaging (MS imaging). Two ILs

Full text of DOI:10.1021/ac100133c

Anal. Chem. 2010, 82, 4413–4419
Matrix-Enhanced Secondary Ion Mass
Spectrometry (ME SIMS) Using Room Temperature
Ionic Liquid Matrices
Jennifer J. D. Fitzgerald, Paul Kunnath, and Amy V. Walker*
Department of Materials Science and Engineering, University of Texas at Dallas, 800 West Campbell Road,
RL 10, Richardson, Texas 75080
Room temperature ionic liquids (ILs) have many applica-
tions including as matrices in MALDI. We wished to
investigate the efficacy of ILs as matrices in time-of-flight
secondary ion mass spectrometry and in mass spectro-
metric imaging (MS imaging). Two ILs derived from
r-cyano-4-hydroxycinnamic acid (CHCA) were synthe-
sized and tested using phospholipids, cholesterol, and
peptides. The molecular ion intensities of 1,2-dipalmitoyl-
sn-glycero-3-phosphocholine (DPPC), 1,2-dipalmitoyl-sn-
glycero-3-phosphoethanolamine (DPPE), cholesterol, and
bradykinin were greatly increased using IL matrices.
Further, detection limits were also improved: for DPPC
and DPPE detection, limits were at least 2 orders of
magnitude better using IL matrices. However, these IL
matrices were not effective for the enhancement of angio-
tensin I ions. The data also indicate that IL matrices are
suitable for imaging MS. The IL matrices did not cause
changes to the sample surface via matrix crystallization
or other processes; no “hot spots” were observed in the
mass spectra. As a demonstration, an onionskin mem-
brane was imaged. In the matrix-enhanced MS images,
ions characteristic of proteins and other biomolecules
were observed which could not otherwise be observed.
Clearly ionic liquids deserve further investigation in SIMS
and MS imaging.
molecules within a cell or nanoscale object. One disadvantage of
MALDI imaging is that the sample must be carefully prepared to
maintain the spatial arrangement of the compounds present.9
SIMS has some advantages over MALDI for imaging. The
lateral resolution is better, and can be as high as 200 nm.6 It also
does not require the careful application of a matrix. In SIMS, a
high energy primary ion beam impinges on the sample surface
leading to the ejection of secondary speciessneutrals, electrons,
and ions.10 Imaging is usually performed in a microprobe config-
uration in which a focused primary ion beam is rastered across
the sample surface.5,10 In this configuration, the lateral resolution
of the image is determined by the analyte ionization probability
and concentration, and the ion beam diameter as well as other
instrument characteristics.10 One of the major issues in SIMS is
that the ionization probability of intact molecules with m/z > 500
is extremely low leading to lateral image resolutions that are much
lower than the ion beam diameter.5,11 For example, for 25 keV
Ga+ primary ions, the beam diameter is 100 nm but the lateral
11,12
resolution .1 µm.11 Cluster ion beams, including Aun
,
+
11,13-15
11,16,17
x+
x+
Bin
,
and C60
,
greatly improve the efficiency of
secondary ion formation from molecules particularly at high mass.
(1) McDonnell, L. A.; Heeren, R. M. A. Mass Spectrom. Rev. 2007, 26, 606–
643
(2) Caprioli, R. M.; Farmer, T. B.; Gile, J. Anal. Chem. 1997, 69, 4751–4760
(3) Stoeckli, M.; Chaurand, P.; Hallahan, D. E.; Caprioli, R. M. Nat. Med. 2001,
7, 493–496
(4) Todd, P. J.; Schaaff, T. G.; Chaurand, P.; Caprioli, R. M. J. Mass Spectrom.
2001, 36, 355–369
(5) Pacholski, M. L.; Winograd, N. Chem. Rev. 1999, 99, 2977–3005
(6) Walker, A. V. Anal. Chem. 2008, 80, 8865–8870
(7) Rubakhin, S. S.; Jurchen, J. C.; Monroe, E. B.; Sweedler, J. V. Drug Discovery
Today 2005, 10, 823–837
(8) Chaurand, P.; Schwartz, S. A.; Caprioli, R. M. Curr. Opin. Chem. Biol. 2002,
6, 676–681
(9) Schwartz, S. A.; Reyzer, M. L.; Caprioli, R. M. J. Mass Spectrom. 2003, 38,
699–708
.
.
.
Imaging mass spectrometry (MS) provides a unique op-
portunity to obtain insights into a wide range of samples from
biological tissues to molecular electronics. In this method, the
spatial distribution of atoms and molecules, including polymers,
pharmaceuticals, lipids, proteins, and semiconductors, is acquired
without the use of chemical labels such as fluorescent tags.1-8
Currently, two ionization techniques are used to record MS
images: matrix-assisted laser desorption ionization (MALDI)1-4,7,8
and secondary ion mass spectrometry (SIMS).1,5-7
.
.
.
.
.
.
(10) Walker, A. V. In The Encyclopedia of Mass Spectrometry: Molecular Ionization
Methods; Gross, M. L., Caprioli, R. M., Eds.; Elsevier Science & Technology
Books: Burlington, VT, 2007; Vol. 6, p 535-551.
MALDI is often employed to image biological samples because
it can easily ionize large nonvolatile biomolecules, such as proteins
and nucleotides.4 In imaging MALDI, the image is obtained by
scanning the sample through the laser beam.1,7 The spatial
resolution is determined by the laser spot size, typically ∼25
µm,3,7,8 which is insufficient to obtain the spatial distribution of
(11) Kollmer, F. Appl. Surf. Sci. 2004, 231-232, 153–158
(12) Nagy, G.; Gelb, L. D.; Walker, A. V. J. Am. Soc. Mass Spectrom. 2005, 16,
733–742
(13) Touboul, D.; Kollmer, F.; Niehuis, E.; Brunelle, A.; Lapre´vote, O. J. Am.
Soc. Mass Spectrom. 2005, 16, 1608–1618
(14) Nagy, G.; Walker, A. V. Int. J. Mass Spectrom. 2007, 262, 144–153
(15) Winograd, N. Anal. Chem. 2005, 77, 143A–149A
(16) Wong, S. C. C.; Hill, R.; Blenkinsopp, P.; Lockyer, N. P.; Weibel, D. E.;
Vickerman, J. C. Appl. Surf. Sci. 2003, 203-204, 219–222
(17) Carado, A.; Kozole, J.; Passarelli, M.; Winograd, N.; Loboda, A.; Bunch, J.;
Wingate, J.; Hankin, J.; Murphy, R. Appl. Surf. Sci. 2008, 255, 1572–1575
.
.
.
.
.
.
* Corresponding author: (e-mail) amy.walker@utdallas.edu; (phone) 972 883
5780; (fax) 972 883 5725.
.
10.1021/ac100133c  2010 American Chemical Society
Published on Web 05/12/2010
Analytical Chemistry, Vol. 82, No. 11, June 1, 2010 4413

This has led to images with improved contrast and increases the
experimental lateral resolution from molecular ions to ∼400 nm.13
However, this lateral resolution is still much larger that the
diameter of the primary ion beam, ∼100 nm.11
A second method by which to enhance secondary ion intensi-
ties (by increasing the ionization probability of the analyte) is to
employ sample pretreatment methods. The combination of sample
pretreatment methods with a cluster ion beam source has the
potential to greatly improve the capabilities of imaging SIMS. In
matrix-enhanced SIMS (ME SIMS), a matrix, such as 2,5 dihy-
droxybenzoic acid (2,5 DHB), glycerol, or hydrochloric acid, is
mixed with, or sprayed onto, the analyte.18-23 Using this method,
quasimolecular ions from proteins up to m/z 10 000 have been
detected.18 However, ME SIMS is unsuitable for use in imaging
applications because the application of a matrix can lead to
changes in the sample surface due to crystallization of the applied
matrix.18,22,23 This causes the recorded distribution of the atoms
and molecules present on the surface to be different from the
original (“true”) distribution. This phenomenon is often manifested
via the formation of “hot spots”, which are areas where the
recorded ion intensities are much larger than elsewhere on the
sample surface. The addition of metals, such as gold and silver,
to a sample can also increase the intensity of secondary ions
detected (meta-SIMS).24-26 Although meta-SIMS is suitable for
imaging, it can be difficult to assign peaks and to interpret the
images obtained.1,27 Further, Adriaensen et al.27 have reported
that meta-SIMS ion yields are time dependent and do not
correspond to the concentration of the analyte.
Figure 1. The structure of (a) MI CHCA, (b) trip CHCA, (c) DPPC,
(d) DPPE, (e) cholesterol, (f) Lys-(Des-Arg9-Leu8) bradykinin, and
(g) angiotensin I.
matrices for use in imaging MS because they have a very low
,33
vapor pressure and so can be used in very high vacuum.28
Recently, room temperature ionic liquids (ILs) have been
shown to be effective MALDI matrices. They can be employed
with a wide variety of analytes including polymers,28,29 oligonucleo-
tides,30,31 peptides,28,31 proteins,28,31 lipids,32 oligosaccarides,29 and
glycoconjugates.29 In MALDI, the performance of IL matrices has
been demonstrated to be at least as good as, if not better than,
the analogous solid matrix. For example, using a solid R-cyano-
4-hydroxycinnamic acid (CHCA) matrix, the detection limit of
bradykinin was 10 pmol/mL (10-5 mg/mL).28 Using an IL
matrix, the detection limit was more than 3 orders of magnitude
better, <1 fmol/mL (10-9 mg/mL).28 Ionic liquids are attractive
Second, they are liquids and so do not alter the sample surface
chemistry by crystallizing. Thus, no “hot spots” are observed
across the sample surface.28 Finally, there is little, or no, mass
interference observed at m/z < 200 due to fragmentation of the
matrix.28 This is because ILs are composed of preformed ions,
and so few fragment ions are observed in the mass spectrum.
In this paper, we investigate the use of ILs as a novel class of
matrices in SIMS. Two different room temperature ILs derived
from the MALDI matrix CHCA were tested because they have
been demonstrated to be effective MALDI matrices for the analysis
of polyethylene glycol (PEG MW 2000), bradykinin and human
insulin.28 Figure 1 displays the structures of the matrices chosen,
1-methylimidazolium R-hydroxycinnamate (MI CHCA) and tripro-
pylammonium R-hydroxycinnamate (trip CHCA). The analytes
chosen were 1,2 dipalmitoyl-sn-glycero-3-phosphocholine (DPPC)
(m/z 734.04), 1,2 dipalmitoyl-sn-glycero-phosphoethanolamine
(DPPE) (m/z 691.95), cholesterol (m/z 386.65), Lys-(Des-Arg9-
Leu8) bradykinin (m/z 997.6), and angiotensin I (m/z 1295.3)
(Figure 1). These analytes represent three classes of biomol-
eculessphospholipids, sterols, and peptides. DPPC, DPPE, and
cholesterol are found in cell walls.34-36 DPPC and DPPE are the
most abundant lipids in the inner leaflet of the plasma mem-
brane,35,36 while cholesterol plays a critical role in the formation
of lipid domains and helps establish proper membrane fluidity
and permeability.34 Bradykinin and angiotensin I are critical to
(18) Wu, K. J.; Odom, R. W. Anal. Chem. 1996, 68, 873–882
(19) Nicola, A. J.; Muddiman, D. C.; Hercules, D. M. J. Am. Soc. Mass Spectrom.
1996, 7, 467–472
(20) Seki, S.; Kambara, H. Shitsuryo Bunseki 1987, 35, 84–89
(21) Tantsyrev, G. D.; Pronchev, G. B. J. Anal. Chem. 1996, 51, 961–963
(22) Hanton, S. D.; Cornelio Clark, P. A.; Owens, K. G. J. Am. Soc. Mass Spectrom.
1999, 10
(23) Luxembourg, S. L.; McDonnell, L. A.; Duursma, M. C.; Guo, X.; Heeren,
R. M. A. Anal. Chem. 2003, 75, 2333–2341
.
.
.
.
.
.
(24) Heile, A.; Lipinsky, D.; Wehbe, N.; Delcorte, A.; Bertrand, P.; Felten, A.;
Houssiau, L.; Piraux, J.-J.; De Mondt, R.; Van Vaeck, L.; Arlinghaus, H. F.
Appl. Surf. Sci. 2008, 255, 941–943
(25) Delcorte, A.; Bour, J.; Aubriet, F.; Muller, J.-F.; Bertrand, P. Anal. Chem.
2003, 75, 6875–6885
(26) Delcorte, A.; Me´dard, N.; Bertrand, P. Anal. Chem. 2002, 74, 4955–4968
(27) Adriaensen, L.; Vangaever, F.; Gijbels, R. Anal. Chem. 2004, 76, 6777–
6785
(28) Armstrong, D. W.; Zhang, L.-K.; He, L.; Gross, M. L. Anal. Chem. 2001,
73, 3679–3686
(29) Mank, M.; Stahl, B.; Boehm, G. Anal. Chem. 2004, 76, 2938–2950
(30) Carda-Broch, S.; Berthod, A.; Armstrong, D. W. Rapid Commun. Mass
Spectrom. 2003, 17, 553–560
(31) Li, Y. L.; Gross, M. L. J. Am. Soc. Mass Spectrom. 2004, 15, 1833–1837
(32) Li, Y. L.; Gross, M. L.; Hsu, F.-F. J. Am. Soc. Mass Spectrom. 2005, 16,
679–682
.
.
.
.
.
.
(33) Smith, E. F.; Rutten, F. J. M.; Villar-Garcia, I. J.; Briggs, D.; Licence, P.
Langmuir 2006, 22, 9386–9392
(34) Leah, E. Lipidomics Gateway 2009, 1, DOI: 10.1038/lipidmaps.2009.3.
.
.
.
(35) Roddy, T. P.; Cannon, D. M., Jr.; Ostrowski, S. G.; Ewing, A. G.; Winograd,
N. Anal. Chem. 2003, 75, 4087–4094
(36) Ko¨hler, U.; Mantsch, H. H.; Casal, H. L. Can. J. Chem. 1988, 66, 983–988
.
.
.
4414 Analytical Chemistry, Vol. 82, No. 11, June 1, 2010

vascular signaling and the regulation of blood pressure.37,38 The
data clearly show that the use of ILs as matrices in SIMS greatly
enhances quasimolecular ion intensities and substantially improve
analyte detection limits. Ion intensities are at least an order of
magnitude better, while detection limits are ∼1000× improved.
Further, ILs do not cause changes to the sample surface via matrix
crystallization or other processes; no “hot spots” are observed.
The results indicated that IL matrices are suitable for use in
imaging. To demonstrate this application, an onionskin membrane
was imaged. It is observed that a thin spun-coat layer of the IL
matrix on the membrane greatly improves the MS images
obtained.
200 µL) was spun coat onto an ∼0.5 cm2 Si substrate for 6 s at
500 rpm and 20 s at 2000 rpm using a KW-4A spin coater
(Chemat Technology, Inc., Northridge, CA).
Angiotensin and bradykinin samples were prepared by dis-
solving 1 mg/mL of peptides in 1:1 methanol/chloroform (v/v)
solution. A total of 100 µL of the peptide solution was then mixed
with 100 µL of the matrix (concentration 0.5 M). Control samples
were prepared by mixing 100 µL of the peptide solution with 100
µL of the methanol/chloroform solution. Spin coating on silicon
substrates was performed in a similar manner as for the phos-
pholipid and cholesterol samples.
For imaging, three red onionskin samples were prepared by
placing a piece of the membrane onto a ∼1 cm2 Si substrate. A
total of 500 µL of matrix solution (concentration 0.2 M, either
trip CHCA or MI CHCA) was spun coat (30 s at 2000 rpm) on
top of two onionskin samples. The control sample was prepared
by spin coating 500 µL of the methanol/chloroform solvent onto
the membrane.
Time-of-Flight Secondary Ion Mass Spectrometry. TOF
SIMS spectra were obtained using a TOF SIMS IV (ION TOF,
Inc., Chestnut Ridge, NY) equipped with a Bi liquid metal ion
gun. Briefly, the instrument consists of a loadlock, preparation
chamber, and analysis chambers, each separated by a gate valve.
During experiments, the analysis chamber was maintained at
e8 × 10-9 mbar to prevent sample contamination. The primary
Bi+ ion had a kinetic energy of 25 keV. The primary ion dose
was less than 1 × 1010 ions·cm-2, which is within the static
regime.10 The secondary ions were extracted from the surface
using 2000 V and reaccelerated to 10 keV before reaching the
detector.
EXPERIMENTAL SECTION
Materials. 1-Methylimidazole (99%), tripropylamine (99%),
R-cyano-hydroxycinnamic acid (97%) (CHCA), and cholesterol
(99%) were purchased from Sigma Aldrich (St. Louis, MO). 1,2-
Dipalmitoyl-sn-glycero-3-phosphocholine (DPPC) and 1,2-dipalmi-
toyl-sn-glycero-3-phosphoethanolamine (DPPE) were obtained
from Avanti Polar Lipids, Inc. (Alabaster, AL). Lys- (Des-Arg9-
Leu8) bradykinin and angiotensin I were acquired from American
Peptide (Sunnyvale, CA). HPLC grade methanol and chloroform
were obtained from Omnisolv (Charlotte, NC). All chemicals were
used without further purification except the silicon wafer sub-
strates (〈111〉 orientation, single-side polished, Addison Engineer-
ing, San Jose, CA). These were etched using RA1 etch (NH4OH/
H2O2/H2O, 1:1:2), rinsed in copious amounts of water and
ethanol, and dried using nitrogen gas prior to use. The
onionskin membrane was obtained from a red onion (Schnuck
Markets, Inc., St. Louis, MO).
High-resolution mass spectra and images were obtained using
analysis areas of (100 × 100) µm2 and (500 × 500) µm2,
respectively. The mass spectra have a resolution (or resolving
power; m/∆m) of 4850 at m/z 29, while the mass resolution
(∆m) of the images is unity. The secondary ion intensities
reported are the averages of the data obtained from one sample
with at least three randomly selected areas per sample
analyzed. The detection limit of the analyte is given by the
lowest concentration for which the quasimolecular ion has a
signal-to-noise ratio of 2.
To investigate the spatial variation of the analyte and matrix
signal, the DPPC and DPPE (1 mg/mL) mass spectra were
investigated in more detail. For each (100 × 100) µm2 spot, five
(20 × 20) µm2, nonoverlapping areas were then selected and
the intensities of the quasimolecular, characteristic fragment
and matrix ions were then obtained (see Supporting Informa-
tion). For each (20 × 20) µm2 area, the data presented are the
average of the ion intensities over the three spots examined.
Preparation of Ionic Liquids. The ionic liquids were syn-
thesized using the method described by Armstrong et al.28 Briefly,
1-methylimidazolium R-hydroxycinnamate (MI CHCA) and tripro-
pylammonium R-hydroxycinnamate (trip CHCA) were prepared
by first dissolving 0.5 g (0.00265 mol) of CHCA in 15 mL of
methanol. To synthesize MI CHCA, an equimolar amount of
1-methylimidazole (212 µL) was dissolved in the CHCA solution.
Similarly, to prepare trip CHCA, an equimolar amount of tripro-
pylamine (505 µL) was added to the CHCA solution. After
sonication of the resulting solutions for 2 min, the methanol was
evaporated leaving the ionic liquid. Prior to use, the ILs were
characterized using 1H NMR, infrared spectroscopy, and time-
of-flight secondary ion mass spectrometry (see Supporting
Information).
Sample Preparation. To investigate the ion intensity en-
hancements and the detection limits of DPPC, DPPE, and
cholesterol, three sets of samples were prepared one for each ionic
liquid matrix (MI CHCA and trip CHCA) and a control (no matrix
added). For the “matrix enhanced” samples, 100 µL of IL solution
(concentration 0.5 M) was mixed with 100 µL of the analyte
solution. The concentration of the analyte was varied from 1 to
10-5 mg/mL. The solvent employed was 1:1 methanol/
chloroform (v/v). For the control samples, a 100 µL aliquot of
the methanol/chloroform solution was mixed with 100 µL of
the analyte solution. Once mixed, the sample (total volume:
RESULTS AND DISCUSSION
The data clearly show that both MI CHCA and trip CHCA
are effective matrices. In Figure 2, it can be clearly seen that
the intensities of the quasimolecular ions of DPPE ((M - H)-
,
m/z 690.5), DPPC ((M + H)+, m/z 734.6), and cholesterol
((M + H)+, m/z 385.4) have significantly increased when an
IL matrix is employed. Further, in general, the ion intensity
enhancement is larger when MI CHCA is employed as a matrix.
The enhancement observed is dependent on the concentration
of the analyte; that is, it is dependent on the matrix-to-analyte
(37) Dendorfer, A.; Wolfrum, S.; Wagemann, M.; Qadri, F.; Dominiak, P. Am. J.
Physiol.: Heart. Circ. Pysiol. 2001, 280, H2182–H2188
(38) Paul, M.; Poyan Mehr, A.; Kreutz, R. Physiol. Rev. 2006, 86, 747–803
.
.
Analytical Chemistry, Vol. 82, No. 11, June 1, 2010 4415

Figure 3. The enhancement of the protonated molecular ion intensity
of DPPC vs DPPC concentration. The subscripts denote which matrix
was employed.
respectively. Similar behavior is observed for DPPE (see Sup-
porting Information). For cholesterol, the enhancement in the
deprotonated molecular ion intensity is much smaller, typically
3-10×, over the concentration range investigated (see Supporting
Information).
The ion intensity enhancements also appear to be affected by
other experimental variables. We observe variations in the ion
intensity enhancements for different sample series particularly at
high analyte concentrations. For example, at a DPPC concentra-
tion of 1 mg/mL, the enhancement of the quasimolecular ion
varies from 100× to 300× and from 50× to 70× using MI CHCA
and trip CHCA, respectively. This is most likely due to uneven
mixing of the analyte and matrix in solution or small changes in
the thickness of the spun coat film. No “hot spots” are observed
across the samples.
The detection limit of DPPC, DPPE, and cholesterol was also
improved using these IL matrices. It was determined by finding
the analyte concentration for which the quasimolecular ion had a
signal-to-noise ratio of at least 2. For DPPC, without matrix the
detection limit was 4 × 10-2 mg/mL corresponding to 5.3 nmol.
Note: this is the total amount of DPPC in the sample solution
and not the amount present on the Si substrate which is less
(due to the spin coating process). Using the matrices, the
detection limit is 1000× better: for MI CHCA, the limit of
detection is 2.7 pmol (2 × 10-5 mg/mL), while for trip CHCA,
it is 5.3 pmol (4 × 10-5 mg/mL). Assuming that all of the DPPC
is deposited evenly across the substrate (∼0.5 cm2), for a (100
× 100) µm2 analysis area, this corresponds 0.54 and 1.1 fmol
of DPPC for MI CHCA and trip CHCA, respectively. This is
∼10× lower than the reported smallest amount of analyte
detected using ME SIMS and a traditional MALDI matrix (50
fmol bovine insulin using 2,5 DHB as a matrix). Similarly, for
DPPE, the detection limit was improved. Without a matrix, the
detection limit was 5.8 nmol (4 × 10-2 mg/mL). Using MI
CHCA, the detection limit was 1000× better, 5.8 pmol (4 × 10-5
mg/mL), while it was slightly higher for trip CHCA, 29 pmol
(2 × 10-4 mg/mL), a 200× improvement. Finally, for cholesterol,
using matrices the detection limit was <10-5 mg/mL, which
was the smallest analyte concentration employed. Without a
matrix, the detection limit was 6 × 10-5 mg/mL (16 pmol).
Figure 2. TOF SIMS spectra of the quasimolecular ions of (a) DPPE,
(b) DPPE, and (c) cholesterol in MI CHCA and trip CHCA matrices,
and with no matrix applied. Concentration of DPPE and DPPC: 0.5
mg/mL. Concentration of cholesterol: 1.0 mg/mL.
ratio. To investigate this effect, the molar matrix ratio was
varied from ∼350:1 to ∼3.5 × 107 (matrix/analyte). The samples
were prepared in the following way. To 100 µL of 0.5 M matrix
solution, 100 µL of analyte of varying concentration (1 to 10-5
mg/mL was added and thoroughly mixed. The resulting
solution was then spun coat onto a clean Si substrate. Figure
3
displays an example of data obtained: the variation of the
enhancement in the quasimolecular molecular ion intensity of
DPPC ((M + H)+, m/z 734.6) with DPPC concentration. The
ion intensity enhancement is calculated using
I((M + H)+matrix
I((M + H)+no matrix
)
enhancement )
)
where I((M + H)+) is the intensity (peak area) of the protonated
molecular ion of DPPC. In Figure 3, it can be seen that the
largest ion intensities are observed for concentrations g0.1 mg/
mL. At concentrations e0.01 mg/mL, the ion intensity enhance-
ment is 10-20× and 3-15× for MI CHCA and trip CHCA,
4416 Analytical Chemistry, Vol. 82, No. 11, June 1, 2010

Figure 5. High resolution SIMS spectra of Lys-(Des-Arg9-Leu8)
bradykinin using (a) trip CHCA as a matrix and (b) no matrix.
Concentration of bradykinin solution: 1 mg/mL.
DPPE: the intensity enhancement of the characteristic fragment
ions of the head and tail groups ((C2H9PO4N)+ m/z 142 and
(C35H67O4)+ m/z 552, respectively) is much smaller than for
the quasimolecular ion ((M - H)- m/z 690) (Figure 4b).
However, in contrast to DPPC, the enhancement of the charac-
teristic fragment ions decreases with decreasing DPPE concentra-
tion. In contrast to both DPPC and DPPE, which are both
phospholipids, for cholesterol (Figure 4c), the enhancement of
the molecular ion at m/z 369 ((M - H2O + H)+) is ap-
proximately, or about the same as the quasimolecular ion
((M - H)+ m/z 385) intensity enhancement. Similar behavior
is observed using trip CHCA (see Supporting Information).
TOF SIMS analyses were also performed for bradykinin (m/z
997.6) and angiotensin I (m/z 1295.3) using trip CHCA and MI
CHCA matrices. A positive ion mass spectrum of bradykinin using
trip CHCA as a matrix is shown in Figure 5a. The concentration
of the analyte solution was 1 mg/mL. Assuming that all of the
analyte is deposited on the Si substrate (∼0.5 cm2) using spin
coating, the amount of analyte is 10 pmol in the (100 × 100)
µm2 analysis area. In Figure 5a, it can be clearly seen that an
intense quasimolecular ion ((M + H)+ m/z 998.6) as well as a
sodiated ion ((M + Na)+ m/z 1020.6) is observed. For the neat
analyte, a very low intensity quasimolecular ion is also observed
but no sodiated ion. The enhancement of the bradykinin
quasimolecular ion is 30× and 40× using trip CHCA and MI
CHCA, respectively. Little fragmentation is observed in either
the ME SIMS or SIMS spectra.
Figure 4. The intensity enhancements for ions characteristic of
DPPC, DPPE, and cholesterol at four different concentrations: 1, 0.5,
0.1, and 0.05 mg/mL. The subscripts denote whether a matrix was
employed.
The intensity of the characteristic fragment ions was also
increased using MI CHCA and trip CHCA matrices. Figure 4
displays the enhancement of secondary ions characteristic of
DPPC, DPPE and cholesterol at several different analyte concen-
trations using MI CHCA as a matrix. As in Figure 3, the ion
intensity enhancement is calculated using
I(X(matrix
I(X(no matrix
)
enhancement )
)
where I(X ) is the intensity (peak area) of X . For DPPC, the
enhancement of fragment ions characteristic of the headgroup
((C5H15NPO4)+ m/z 184 and (C8H19NPO4)+ m/z 224) as well
as the tail group ((C35H67O4)+ m/z 552) are substantially smaller
than the enhancement observed for the quasimolecular ion
((M + H)+ m/z 734), suggesting that the extent of fragmenta-
tion decreases using an IL matrix (Figure 4a). Further, in
contrast to the quasimolecular ion intensity, the intensity enhance-
ments of these fragment ions remain approximately constant with
decreasing DPPC concentration. Similar behavior is observed for
However, IL matrices are not always effective in SIMS. For
angiotensin I, no quasimolecular ions were observed in the neat
analyte spectrum or using MI CHCA as a matrix. Using trip CHCA
as a matrix, a quasimolecular ion was observed in the positive
Analytical Chemistry, Vol. 82, No. 11, June 1, 2010 4417

Figure 6. Negative ion TOF SIMS images of DPPE ((M - H)- m/z
690) and the matrix ((CHCA - H)-, m/z 188). The images are shown
on a heat scale, and the scale bar shows the minimum and maximum
counts per pixel. Concentration of DPPE solution: 1 mg/mL. Area of
analysis: (100 × 100) µm2.
ion mass spectrum. However, it has a low intensity (signal-to-noise
ratio ) 5).
Imaging. The data also indicate that the analyte ion intensities
are uniform across the sample surface. In Figure 6, it can be
clearly seen that the ion intensities of both the analyte, DPPE
((M - H)- m/z 690) and the matrix ((CHCA - H)-, m/z 188;
CHCA ) C10H7NO3) are approximately constant indicating that
there are no “hot spots” (regions of greatly increased secondary
ion intensity). To further investigate the spatial variation of the
analyte and matrix ions, the DPPC and DPPE spectra were
examined in more detail. For each analysis area (100 × 100
µm2), five nonoverlapping (20 × 20) µm2 areas were selected.
For each (20 × 20) µm2 area, the intensities of the molecular,
characteristic fragment and matrix ions were averaged over
the three analysis areas used per sample. Figure 7 displays the
results obtained using trip CHCA as a matrix. Similar results are
obtained for MI CHCA (see Supporting Information). The data
clearly show that the ion intensities are constant across the
analysis area indicating that ionic liquids are suitable for imaging
applications.
Figure 7. Normalized ion intensities of the molecular and charac-
teristic fragment ions of DPPC, DPPE, and trip CHCA ions across
the spun coat samples. The concentration of the phospolipid solutions
was 1 mg/mL. For the DPPE sample (a), the ion intensities were
studied (C5H8O3)- (m/z 116) and (M - H)- (m/z 690), characteristic
ions of DPPE, and the matrix ion (CHCA - H)- (m/z 188). For DPPC
(b), the ion intensities examined were the matrix ion (C(CH3)2H)3-
NH+ ((trip + H)+, m/z 144), and m/z 184 ((C5H15NPO4)+), 224
(C8H19NPO4)+ and 734 ((M + H)+), characteristic ions of DPPC. The
ion intensities were normalized to the total number of ion counts to
compensate for any small changes in primary ion dose.
To demonstrate this application, three onionskin membranes
were imaged using SIMS. Two samples were prepared by spin
coating either MI CHCA or trip CHCA on top of the onionskin
while the third sample, the control, was prepared with application
of only the solvent. In the positive ion mass spectra for the MI
CHCA and trip CHCA coated samples, the ion intensities of ion
intensities of (MI + H)+ (m/z 83) and (trip + H)+ (m/z 144)
are approximately constant over the sample suggesting that
the matrix is evenly coated across it (see Supporting Informa-
tion). Figure 8 displays optical and negative ion TOF SIMS images
of the onion skins. For the control sample, very low ion intensities
are observed and there is no localization of ions such as CN-
(m/z 26) which is characteristic of proteins present in the cell
nuclei. In contrast for the MI CHCA and trip CHCA spun-coat
samples, ions characteristic of proteins including CN- (m/z
26), and the b ions of (Gly-Ser) (m/z 144) and (Ser-Cys) (m/z
189) are observed. These ions are colocalized in well-defined
areas and most likely arise from the cell nuclei. In some cases,
these areas are within the cell while other areas are located at
the cell wall, suggesting that some of the cells “burst” upon
introduction into the vacuum. It is interesting to note that these
features are not clearly defined in the optical images but can
be easily observed in ME SIMS.
Figure 8. Optical and negative ion SIMS images of onion mem-
branes. SIMS images centered at m/z 26 (CN-), m/z 144 (b ion of
Gly-Ser), and m/z 190 (b ion of Ser-Cys), characteristic ions of
proteins. The samples were prepared with no matrix or by spin coating
500 µL of either MI CHCA or trip CHCA on the onionskin. As in Figure
7, the SIMS images are shown using a heat scale. SIMS: ion source
Bi+; area of analysis 500 × 500 µm2, 128 × 128 pixels.2
Mechanism. It is clear that the mechanism of ion yield
enhancement using ionic liquids is different in MALDI and SIMS.
This is because the secondary ions are generated in a different
way. In MALDI, a pulsed laser is employed. Secondary ions are
generated via collisions between neutral analytes and matrix ions
4418 Analytical Chemistry, Vol. 82, No. 11, June 1, 2010

in the expansion plume as the sample (matrix/analyte mixture)
is ablated from the surface.39 In contrast, in TOF SIMS, a pulsed
primary ion beam is employed. Primary ions impinging on the
surface cause a collision cascade within the sample which leads
to the ejection of secondary speciessneutrals, electrons, and
secondary ions.10 However, there are some similarities in the
behaviors observed using ionic liquid matrices in SIMS and
MALDI.28,29,32 First, the detection limit of phospholipids is several
orders of magnitude better using IL matrices (Figure 3 and
Supporting Information). Second, the extent of fragmentation is
significantly reduced for most analytes (Figure 4). In MALDI, it
is believed that ionization is promoted by available matrix protons;
the donation of protons, but this is a vanishingly small effect.
Thus, the pKa of the matrix cannot explain the observed matrix
effects. One possible reason for the observed behavior is that
there is aggregation of the analyte and IL in the solution which
affects the proton transfer process. Further work is required
to understand the mechanism of ME SIMS using ionic liquid
matrices.
CONCLUSIONS
We have investigated the use of two different room tempera-
ture ionic liquids (ILs), derived from the MALDI matrix CHCA,
in TOF SIMS. Our data indicates that molecular ion intensities of
DPPC, DPPE, cholesterol, bradykinin, and angiotensin I are
greatly increased using IL matrices, indicating that ILs are a
broadly applicable effective matrices for ME SIMS. Further, using
IL matrices, the detection limits of DPPC, DPPE, and cholesterol
are improved by several orders of magnitude than those achieved
using no matrix. We also observe that the formation of fragment
ions is suppressed for the phospholipids and peptide samples. In
contrast, for cholesterol, the ion intensity enhancement of the
characteristic fragment ion, (M - H2O + H)+, is similar to that
observed for the quasimolecular ion, (M - H)+.
In contrast to solid matrices, such as CHCA and 2,5 DHB, the
data also show that the ion intensities are uniform across the
sample surface indicating that ILs are suitable matrices for SIMS
imaging. To demonstrate this application, onionskin membranes
were imaged. Using IL matrices, fragment ions characteristic of
proteins, most likely from the cell nuclei, were observed which
were not observed in SIMS images of control samples (no matrix
applied).
ionic liquids with an acidic proton promote the production of the
28
desired gas phase protonated quasimolecular ion, (M + H)+
.
In SIMS, it is also likely that ionization is promoted via the
transfer of protons from (or to) the matrix. We observe that
the intensities of protonated, or deprotonated, molecular ions
of DPPC, DPPE, cholesterol, and bradykinin are significantly
increased using IL matrices (Figures 1 and 5). Further, for
cholesterol, similar ion intensity enhancements are observed for
the characteristic fragment ion (M - H2O + H)+ (m/z 369) as
for the quasimolecular ion (M - H)+ (m/z 385) (Figure 4).
The formation of this ion also requires the transfer of a proton
from the surrounding matrix to the analyte. For DPPC, we note
that the formation of the fragment ion characteristic of the
headgroup (m/z 184) also involves the transfer of a proton from
the molecular environment. However, its ion intensity is only
increased ∼5-10× which is much smaller than the enhancement
observed for the quasimolecular ion, (M + H)+ (m/z 734) (Figure
4
). The most likely reason for this is that the formation of this
The data indicate that ionization is promoted by proton transfer
from (or to) the matrix. However, there appears to be no
relationship between the pKa of the matrix and the secondary
ion intensity enhancement observed. Further work is underway
to understand the mechanism of ionization in ME using ionic
liquid matrices. We are studying the use of other ionic liquids
as matrices for SIMS including those not derived from MALDI
solid matrices. Finally, we are also investigating the use of IL
matrices to improve the secondary ion intensities of proteins,
peptides, polymers, and other high molecular weight species,
as well as optimizing the preparation method for imaging
tissues and other samples.
fragment ion involves an intramolecular rearrangement, and not
just a proton transfer.40
The chemical nature of the cation portion (protonated amine)
of the matrix also appears to have an influence on secondary ion
intensities. However, there appears to be no relationship between
the pKa of the matrix and secondary ion intensity. The pKa of
methylimidazole is 6.95, while the pKa of tripropylamine is
,42
10.89.28 The pKa of CHCA (matrix base) is ∼2.41 Upon
formation of the ionic liquid, the matrix base (CHCA) becomes
almost completely deprotonated, and the matrix acid (trip or MI)
is almost completely protonated. Upon addition of DPPC (pKa
)
2 - 3),36 there will be a competition between CHCA and DPPC
to donate protons to the MI or trip amine groups (cation). Thus,
one would expect to see an increase in deprotonated ions, such
as (M - H)-. This is observed for DPPE, but not DPPC where
a large increase in the intensity of the protonated molecular
ion ((M + H)+) is observed (Figure 2). Further, the pKa of MI
and trip cannot account for the observed differences in the ion
intensity enhancements observed. Upon formation of the ionic
liquid, the matrix base with the lower pKa (MI) will be slightly
less protonated than the base with the higher pKa (trip). This
suggests that there will be more uncharged MI available for
ACKNOWLEDGMENT
The authors acknowledge the National Science Foundation and
a DuPont Young Faculty Award for financial support.
SUPPORTING INFORMATION AVAILABLE
NMR, transmission IR and TOF SIMS analyses of trip CHCA
and MI CHCA; the enhancement of the quasimolecular ions of
DPPE and cholesterol using trip CHCA and MI CHCA vs analyte
concentration; normalized ion intensities of DPPC and DPPE and
MI CHCA ions across the sample surface. This material is
available free of charge via the Internet at http://pubs.acs.org.
(39) Zenobi, R.; Knochenmuss, R. Mass Spectrom. Rev. 1998, 17, 337–366
(40) Sostarecz, A. G.; Cannon, D. M., Jr.; McQuaw, C. M.; Sun, S.; Ewing, A. G.;
Winograd, N. Langmuir 2004, 20, 4926–4932
(41) Guo, Z.; He, L. Anal. Bioanal. Chem. 2007, 387, 1939–1944
(42) Lou, X.; van den Hout, K. P.; van Houtem, M. H. C. J.; van Dongen, J. L. J.;
Vekemans, J. A. J. M.; Meijer, E. W. J. Mass Spectrom. 2006, 41, 659–669
.
.
Received for review January 15, 2010. Accepted April 29,
2010.
.
.
AC100133C
Analytical Chemistry, Vol. 82, No. 11, June 1, 2010 4419