Article pubs.acs.org/ac
VAMAS Interlaboratory Study for Desorption Electrospray Ionization Mass Spectrometry (DESI MS) Intensity Repeatability and Constancy Elzbieta Gurdak,* Felicia M. Green, Paulina D. Rakowska,* Martin P. Seah, Tara L. Salter, and Ian S. Gilmore National Physical Laboratory (NPL), Hampton Road, Teddington, Middlesex, TW11 0LW, United Kingdom ABSTRACT: A VAMAS (Versailles Project on Advanced Materials and Standards) interlaboratory study for desorption electrospray ionization mass spectrometry (DESI MS) measurements has been conducted with the involvement of 20 laboratories from 10 countries. Participants were provided with an analytical protocol and two reference samples: a thin layer of Rhodamine B and double-sided adhesive tape, each on separate glass slides. The studies comprised acquisition of positive ion mass spectra in predetermined m/z ranges. No sample preparation was required. Results for Rhodamine B show that very consistent craters may be generated. However, inadequacies of the spray and sample stage designs often lead to variable crater shapes. The average repeatability for Rhodamine B is 50%. Yet, repeatabilities better than 20% can be achieved. Rhodamine B proved to be an excellent reference sample to check the sample erosion crater, the sample stage movement and memory effects. Adhesive tape samples show that their average absolute intensity repeatability is 30% and the relative repeatability is 9%. The constancy of these spectra from relative intensities gives day-to-day average relative repeatabilities of 31%, three times worse than the short-term repeatability. Significant differences in the spectra from different laboratories arise from the different adventitious adducts observed or from contaminants that may cause the higher day-to-day variations. It is thought that this may be overcome by allowing some 20 ppb of sodium to be always present in the solvent, to be the dominating adduct. Repeatabilities better than 5% may be achieved with adequate control.
T
lower sensitivity24 and can provide high throughput and realtime analysis for reaction monitoring.25 One of the advantages of DESI lies in the fact that it can be coupled to any mass spectrometer type, such as time-of-flight (TOF), Orbitrap, ion trap, etc. Hence, it is also possible to combine DESI with portable MS for hand-held ambient analysis devices.26,27 To promote the wider uptake of DESI, especially by industry, it is important for measurements to be repeatable and reliable. Ideally, for laboratory conformance with accreditation, standards consistent with ISO 1702528 should be available for analysts. To achieve this, a robust measurement infrastructure needs establishment. For effective international standards, it is important that interlaboratory studies are conducted to evaluate if methods are effective and fit for purpose. It is especially important to test that methods are transferable between many different instrument designs and that analytical procedures are clear. The Versailles Project on Advanced Materials and Standards (VAMAS)29 provides an excellent mechanism for such evaluation. Within this framework, the National Physical Laboratory (NPL) has conducted an interlaboratory comparison, based on the following objectives: (i) to determine the repeatability and constancy of DESI instruments achieved in
he innovation of desorption electrospray ionization (DESI)1,2 in 2004 heralded a new era of ambient surface mass spectrometry (MS). Since then, there has been significant progress in understanding the fundamentals3−10 and a rapid expansion in the applications covering a diverse range of science and technologies.10,11 As an ambient ionization technique, DESI is uniquely suited to characterize samples in their native state. In principle, no sample pretreatment is required. Therefore, the direct sampling of surfaces in the open air is possible.2,11,12 DESI analysis can be performed either on samples predeposited on the surface such as glass or paper or performed on natural unmodified sample surfaces. It can also assist in in vivo analyses, of which an example is direct identification of bacteria.13 The ability of DESI to provide analytically useful information directly from everyday surfaces has resulted in the rapidly rising interest in its application in many areas of science.3 DESI has become very attractive in forensics where it can enable trace analysis of polymers, small molecules, drugs, and explosives directly from complex matrices, detection from fingerprints, chemical mapping, or ink analyses.14−18 Personal care products can be analyzed straight from skin, hair, or textiles.19,20 In the pharmaceutical, biomedical, and diagnostics fields, it enables detection of a wide variety of molecules, including proteins, peptides, lipids, drugs and metabolites, carbohydrates and nucleic acids, which can be done directly from plant, cell, or tissue substrates.11,21−23 DESI MS can achieve femtomolar and Published 2014 by the American Chemical Society
Received: June 5, 2014 Accepted: September 10, 2014 Published: September 10, 2014 9603
dx.doi.org/10.1021/ac502075t | Anal. Chem. 2014, 86, 9603−9611
Analytical Chemistry
Article
Table 1. Suggested DESI Source and Mass Spectrometer Operating Conditions Alongside the Minimum and Maximum Values Used by the Respondees units
suggestedd
min
max
average (n)e
mm mm mm degrees degrees μL/min V psi μm
1.7 0 4.5 35 10 2 +5000 100 50 90% acetonitrile, 10% waterb 0.1% formic acidb fused silicac
0.5 0 1.5 33 0 1.5 +20 80 50 n/a n/a n/a
2.8 0 9.0 80 25 6.7 +5000 170 75 n/a n/a n/a
1.6 (20) 0 4 (19) 48 (20) 9.8 (13) 2.2 (20) 2902 (20) 108 (18) 52.5 (10) n/a n/a n/a
°C
275 positive 30 25
19.6 n/a n/a n/a
350 n/a n/a n/a
248 (19) n/a n/a n/a
parameter a
DESI Source Parameters sprayer-to-surface distance, d1 capillary-edge-to-surface distance, d2 sprayer-to-capillary distance, d3 angle of spray to normal, α angle of capillary to surface, β electrospray solvent flow rate, S electrospray voltage, Vs gas pressure inner diameter of capillary emitter solvent volume fraction, X additives to solvent spray tip (material) Mass Spectrometer Parameters capillary temperature ion polarity data acquisition time data acquisition time taken into calculations
s s
a A diagram showing these parameters and symbols may be found in Figure 1 of Green et al.6 bSix respondees used different solvent compositions, including (i) 90% acetonitrile, 10% water; (ii) 80% acetonitrile, 20% water, 0.1% formic acid; and (iii) 90% acetonitrile, 10% water, 0.1% acetic acid. c One participant used methylated silica. dOne participant used the recommended DESI geometrical conditions, 14 used the recommended solvent composition, and 16 used the recommended electrospray flow rate. eThe parameter n represetns the number of respondees who provided the details.
instruments in the positive ion mode, according to their standard procedures and ensure that the mass scale of the spectrometer has been calibrated in the mass range from m/z 400 to m/z 1400. Furthermore, there is no single, universal set of DESI parameters that could be successfully applied to every DESI MS measurement. It is known that each analyzed substance, e.g., proteins, lipids, explosives, etc. has its optimum electrospray-to-surface distance and optimum spray angle.3 These and many other parameters can affect the signal intensity, stability, and repeatability.6,30 NPL established and provided analysts with guidance on the DESI MS operating conditions. The suggested parameters were evaluated using a DESI 2D Omni Spray Ion Source (Prosolia, Indianapolis, IN, U.S.A.), with an LTQ Orbitrap Velos mass spectrometer (Thermo Scientific, Bremen, Germany). However, the analysts were advised to operate their DESI MS instruments under conditions ensuring the most stable analytical performance, and using the recommended parameters only if suitable. The recommended settings, alongside minimum, maximum, and averaged values, applied by the respondees, can be found in Table 1. From the Rhodamine B samples, participants were asked to acquire 55 positive ion spectra in m/z 400 to m/z 500, each from a fresh area of Rhodamine B, as well as 3 blanks in the wider mass range of m/z 50 to m/z 600. The sequence for the array of analyses is shown in Figure 1. The participants were advised to take these measurements on day 1 and to maintain the same capillary-to-surface distance and continuous flow of the electrospray between the 58 acquisitions. For each mass spectrum, the peak intensities were summed over the mass range m/z 443.0 to m/z 443.5 This sum was entered into a spreadsheet provided by NPL, which automatically calculated the absolute intensity repeatability, as discussed later. The Rhodamine B sample forms a uniform layer of a distinct pink color. This allows the erosion of the sample in the DESI analysis to be clearly visible. Twelve (12) out of the 20
practice, (ii) to evaluate the equivalence of results between different laboratories and instruments, and, since this was the first such study on DESI MS, (iii) to enable a survey of measurement issues and (iv) to help establish a community of users interested in DESI MS metrology.
■
EXPERIMENTAL SECTION Sets of two reference samples were prepared and supplied to each participant. The reference materials were a thin layer of Rhodamine B (Sigma, Poole, U.K.) and double-sided adhesive tape (“Mammoth”, Everbuild), each on standard glass slides with one end frosted (76 mm × 26 mm × 1 mm; Twin Frost, Fisherbrand, U.K.). Each sample was given a unique reference number. The Rhodamine B samples were prepared according to a protocol described earlier.6 Briefly, Rhodamine B was evaporated onto a glass slides, using a vacuum coater (Edwards, Model AUTO 306) with the deposited thickness measured by spectroscopic ellipsometry (Woollam, Model M-2000DI). The average thickness of the Rhodamine B layers, measured from 28 silicon wafers used as controls in each glass slide coating, was 785 nm ± 47 nm (6% RSD). The double-sided adhesive tape was cut into 6 mm × 35 mm strips and arranged in three rows on the clean glass slide with the protective strip still on the surface. To ensure the uniform thickness of the tape, a rolling tool was used. Fresh samples of the tape, after removing the protective strip, were analyzed by DESI MS in the mass range of m/z 600−1400. Intensities in the three mass ranges(i) m/ z 670−700, (ii) m/z 774−808, and (iii) m/z 1355−1380 were extracted. The exact chemical composition of the tape is unknown, but analysis by X-ray photoelectron spectroscopy and secondary ion mass spectrometry suggest that the adhesive is rubber-based. DESI can be interfaced with a wide variety of mass spectrometers. Therefore, it was not possible to provide specific guidance for the mass spectrometer settings and operation. Participants were requested to optimize their 9604
dx.doi.org/10.1021/ac502075t | Anal. Chem. 2014, 86, 9603−9611
Analytical Chemistry
Article
Table 2. Types of DESI Models and Mass Spectrometers Used in the Study
Figure 1. Schematic showing the top view of the analysis side of the reference samples: Rhodamine B and adhesive tape. The directions of the MS inlet capillary and the sprayer are indicated. The directions of analyses are specified by numbered spots and illustrated as a directional sequence (on the right). The axes, x and y, are also defined. Spots 1a, 22a, and 44a on the Rhodamine B sample and spots 1a, 7a, and 14a on the adhesive tape sample mark the places from where blank data were recorded.
respondee
DESI model
MS model
R01 R02 R03 R04 R05 R06 R07 R08 R09 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20
commercial home-built home-built home-built commercial home-built home-built home-built commercial commercial home-built commercial home-built commercial home-built home-built home-built home-built commercial home-built
Orbitrap ion trap ion trap ion trap Orbitrap Quadrupole TOF ion trap ion trap ion trap ion trap modified TOF Quadrupole TOF ion trap ion trap ion trap Quadrupole TOF Orbitrap ion trap ion trap ion trap
participant, as a percentage standard deviation of the absolute intensity of the Rhodamine B peak. The average absolute intensity repeatability across all the 20 participants is 50%, as shown later in Table 3 and Figure S-1 in the Supporting Information. There is significant laboratory-to-laboratory variation in the repeatability standard deviation for these absolute intensities. This could be caused by many factors, such as the ability of the instrument to deliver a uniform spray, stability of the detection, and so on. In addition, contamination or variations in the ambient air may be important. Figure 2 shows how some of the measurements exhibited distinct trends that could be related to influence factors, rather than random scatters that would be more difficult to control. In Figure 2a, with regard to the data of respondee R13, there is a clear increase and decrease of the Rhodamine B intensity that is related to the analysis position along the x-direction of the sample. The vertical lines on the figures indicate the start of a new acquisition row. A similar drift was observed in intensities from participants R08 and R11. This effect may arise from a slight tilt of the sample on the stage or a similar effect that causes the sample height to be altered, relative to the DESI input capillary. These data illustrate how a reference sample may be used to verify the correct sample positioning and test the motion of the sample stage. A second group shows a gradual reduction or gradual rise in the Rhodamine B absolute intensity recorded across the 55 measurements, as illustrated in Figure 2b and Figure S-2(a) in the Supporting Information, respectively. This reduction was observed for four participants (R01, R09, R10, and R18) and an equivalent rise for two (R05 and R06). These changes could be caused by mass spectrometer performance over time. However, it is more likely that they arise from the stability of the electrospray flow. Even though participants were advised to allow 30 min for the electrospray to stabilize prior to experiments, it may require longer. The observed drifts contribute significantly to the repeatability of the measurements. As an example, the signal intensity of the Rhodamine B
participating laboratories returned their samples, after analysis, to NPL for their further examination. The samples were inspected by optical microscopy (Alicona InfiniteFocus, Sevenoaks, U.K.). The eroded area was measured using ImageJ software.31 The measurements of the adhesive tape samples were to be made over eight, not necessarily consecutive, days, from 10 separate strips of the tape. In total, participants were asked to acquire 80 positive ion spectra in the m/z 600 to m/z 1400 range (including 10 blanks in a wider mass range of m/z 50 to m/z 1400), each from a fresh area of the tape, as shown in Figure 1. The protective strip was to be removed just prior to analysis. The data from the first three strips was acquired on day 1 without any change to the DESI settings between the acquisitions. The remaining seven strips were analyzed on following days, with none on the same day as another. For all measurements, a waiting time of 30 min was recommended, before the start of the analyses, to allow the electrospray to stabilize. Further details of the acquisition setup and the sequence used to take data can be found in the study protocol.32
■
PARTICIPANTS The participants wishing to be identified are listed in the Acknowledgments section. Data were supplied from 13 ion trap instruments, 3 Orbitrap instruments, and 4 TOF instruments. Seven of the DESI interfaces were commercial, whereas 13 were home-built. Table 2 provides a summary of the type of DESI source and mass analysers used. No single participant followed all of the recommended DESI MS settings shown in Table 1.
■
RESULTS AND DISCUSSION Absolute Intensity Repeatability of Rhodamine B. The repeatability of Rhodamine B measurements was calculated, based on 55 consecutive DESI MS measurements taken by each 9605
dx.doi.org/10.1021/ac502075t | Anal. Chem. 2014, 86, 9603−9611
Analytical Chemistry
Article
The blank data sets acquired between Rhodamine B measurements for 10 of the 12 participants who returned data showed negligible memory of the Rhodamine B. The Rhodamine B was either not observed in the spectra, or its intensity was significantly lower (100-fold), in comparison to the data obtained from the Rhodamine B layer (see the Supporting Information for more details). The lack of large sample displacements during analysis and a low level of crosscontamination from the mass spectra inlet proved the effectiveness of the Rhodamine B layer as an excellent model sample for checking the instrumental setup and repeatability of DESI MS analyses. Overall, Rhodamine B provides significantly useful information. Rhodamine B Sample Erosion. Twelve (12) Rhodamine B samples were returned to NPL after analysis. Optical microscopy was then used to analyze for any correlation between the eroded crater and the absolute intensity repeatability. Eight (8) samples showed consistency in the crater shapes for the 55 points, as illustrated in Figure 3a. Figure 2. Variations in Rhodamine B signal intensity recorded from 55 fresh sample areas by participants (a) R13 and (b) R01. Vertical lines mark the end/beginning of an analysis row and, hence, a change in analysis direction.
peak, in the analysis performed by participant R01, and shown in Figure 2b, decreases over the 55 measurements by a factor of 2. If the drift was removed, the absolute intensity repeatability improves from 19% to 12%. A longer settling time may improve repeatabilities. No common pattern was observed in the data from the remaining participants. Several datasets revealed fairly consistent analyses with a few outlier points, which could have been caused by electrospray instability. For example, a clear spike in the Rhodamine B intensity was observed by participant R12. Optical examination of the returned sample indicated a larger desorption site correlated with the intensity spike (Figure S-3 in the Supporting Information) probably caused by a brief surge in the source. Removal of this result would improve the repeatability of R12, from the 71% value shown in Figure S-1 in the Supporting Information and Table 3 to 31%. Electrospray instability may account for some of the outliers seen in other datasets. However, optical examination did not provide any obvious correlations, and the outliers are all included in the repeatability calculations, since they represent current DESI MS practice. Figure S-2(b) in the Supporting Information illustrates, for participant R16, consistent Rhodamine B intensities for the first four rows on the sample, followed by a sudden decrease in the intensity and a gradual increase along row 5. This may indicate a possible issue with the sample stage movement at this edge of the glass slide. Other causes, such as intermittent, partial blockage of the MS inlet, are also possible. DESI spectra from the blank part of the glass slide, acquired in the mass range from m/z 50 to m/z 600, were returned to NPL by 12 participants. These were examined to evaluate other possible factors, such as the ambient conditions of the DESI and potential contaminants, which could affect the sensitivity and repeatability of the measurements. Various common contaminants were detected in the spectra including polysiloxane, triton, eurucamide, and phthalates. Several blank spectra also showed low-mass acetonitrile clusters.33−37
Figure 3. DESI erosion spots on Rhodamine B samples, with high consistency and regularity in shapes and sizes. Desorption across the 55 measurements for (a) R18 and single spots from (b) R08, (c) R10, and (d) R20.
Typically, the majority of the craters show similar characteristics: an inner ring of fully desorbed sample material and an outer ring of displaced material, often with a ridge of accumulated sample between the two.6 The area of the inner ring, examined on the returned samples, ranges from 0.11 mm2 to 0.75 mm2. The outer disturbed area shows more disparity, with a large variation in shapes and with total desorbed areas ranging from 0.7 mm2 to 4.5 mm2. A good example of DESI craters is shown in Figure 3b, where the spot is fairly regular in shape and elongated in the direction of the spray. Between the 8 samples with consistent crater shapes, 5 were similarly shaped, but with variations in size of the inner and outer areas, and 2 displayed outer areas with a stronger directionality, as illustrated in the example in Figure 3c. An unusual crater shape was observed on the sample from participant R20 (Figure 3d). It contained barely any outer disturbances: only the inner desorbed area and the ridging region. None of the experimental settings for R20 appeared to differ from the settings applied by other participants. This desorption shape indicates significantly less mixing of the 9606
dx.doi.org/10.1021/ac502075t | Anal. Chem. 2014, 86, 9603−9611
Analytical Chemistry
Article
repeatability of the absolute intensities is not sufficient to allow any such conclusions. Repeatability of the Adhesive Tape Measurements. Adhesive tape samples were used to measure both absolute and relative intensity repeatabilities. Representative mass spectra of the adhesive tape, and the mass ranges selected for analyses in the positive mode, are shown in Figure S-4 in the Supporting Information. Participants were asked to acquire 8 datasets from fresh areas of the adhesive tape. The first set consisted of 21 spectra on the first day, followed by 7 further sets of 7 spectra, each from a fresh area of the strip on later days. From each spectrum, the absolute intensities in the three preselected mass ranges were summed and the absolute repeatability was calculated as a relative standard deviation of the absolute intensity. The absolute and relative intensity repeatabilities (Ra and Rr, respectively) were determined for the same 8 sets of measurements, using the method previously described in Gilmore et al.38 Briefly, Ra, expressed as a percentage, is evaluated from the average of the j0 (here, 21) measures, Iij, for the summed intensity in each of the 3 mass ranges, i, where
sample material on the analyzed surface, which could bring a clear advantage in terms of spatial resolution of DESI. Four of the returned samples revealed poor uniformity in terms of the eroded spot shapes and sizes. The sample analyzed by participant R11 gives one example of an inconsistent erosion pattern across 55 consecutive measurements (see Figure 4a). A wide range of spot shapes is illustrated in Figures 4b−d.
j0
Ii̅ =
∑ j=1
Iij j0
(1)
This allows us to remove the overall intensity variation between the three different mass ranges. Ra is then given by the average of the standard deviations (si) of the j0 values of Pij, where Pij = Iij/Ii̅ :
Figure 4. Irregular DESI erosion spots on Rhodamine B samples. Desorption across the 55 measurements for (a) R11 and single spots from (b) R06, (c) R16, and (d) R14.
The optical images from the Rhodamine B samples indicate the spatial resolution of the method. The average crater is an ellipse with axis dimensions of 1.4 mm × 2.3 mm. The inner areas of the desorbed material tend to be more consistent and smaller, with average dimensions of 0.5 mm × 0.8 mm and a relative standard deviation of ∼30% for the 12 participants. The smallest dimensions observed were
VAMAS Interlaboratory Study for Desorption Electrospray Ionization Mass Spectrometry (DESI MS) Intensity Repeatability and Constancy Elzbieta Gurdak,* Felicia M. Green, Paulina D. Rakowska,* Martin P. Seah, Tara L. Salter, and Ian S. Gilmore National Physical Laboratory (NPL), Hampton Road, Teddington, Middlesex, TW11 0LW, United Kingdom ABSTRACT: A VAMAS (Versailles Project on Advanced Materials and Standards) interlaboratory study for desorption electrospray ionization mass spectrometry (DESI MS) measurements has been conducted with the involvement of 20 laboratories from 10 countries. Participants were provided with an analytical protocol and two reference samples: a thin layer of Rhodamine B and double-sided adhesive tape, each on separate glass slides. The studies comprised acquisition of positive ion mass spectra in predetermined m/z ranges. No sample preparation was required. Results for Rhodamine B show that very consistent craters may be generated. However, inadequacies of the spray and sample stage designs often lead to variable crater shapes. The average repeatability for Rhodamine B is 50%. Yet, repeatabilities better than 20% can be achieved. Rhodamine B proved to be an excellent reference sample to check the sample erosion crater, the sample stage movement and memory effects. Adhesive tape samples show that their average absolute intensity repeatability is 30% and the relative repeatability is 9%. The constancy of these spectra from relative intensities gives day-to-day average relative repeatabilities of 31%, three times worse than the short-term repeatability. Significant differences in the spectra from different laboratories arise from the different adventitious adducts observed or from contaminants that may cause the higher day-to-day variations. It is thought that this may be overcome by allowing some 20 ppb of sodium to be always present in the solvent, to be the dominating adduct. Repeatabilities better than 5% may be achieved with adequate control.
T
lower sensitivity24 and can provide high throughput and realtime analysis for reaction monitoring.25 One of the advantages of DESI lies in the fact that it can be coupled to any mass spectrometer type, such as time-of-flight (TOF), Orbitrap, ion trap, etc. Hence, it is also possible to combine DESI with portable MS for hand-held ambient analysis devices.26,27 To promote the wider uptake of DESI, especially by industry, it is important for measurements to be repeatable and reliable. Ideally, for laboratory conformance with accreditation, standards consistent with ISO 1702528 should be available for analysts. To achieve this, a robust measurement infrastructure needs establishment. For effective international standards, it is important that interlaboratory studies are conducted to evaluate if methods are effective and fit for purpose. It is especially important to test that methods are transferable between many different instrument designs and that analytical procedures are clear. The Versailles Project on Advanced Materials and Standards (VAMAS)29 provides an excellent mechanism for such evaluation. Within this framework, the National Physical Laboratory (NPL) has conducted an interlaboratory comparison, based on the following objectives: (i) to determine the repeatability and constancy of DESI instruments achieved in
he innovation of desorption electrospray ionization (DESI)1,2 in 2004 heralded a new era of ambient surface mass spectrometry (MS). Since then, there has been significant progress in understanding the fundamentals3−10 and a rapid expansion in the applications covering a diverse range of science and technologies.10,11 As an ambient ionization technique, DESI is uniquely suited to characterize samples in their native state. In principle, no sample pretreatment is required. Therefore, the direct sampling of surfaces in the open air is possible.2,11,12 DESI analysis can be performed either on samples predeposited on the surface such as glass or paper or performed on natural unmodified sample surfaces. It can also assist in in vivo analyses, of which an example is direct identification of bacteria.13 The ability of DESI to provide analytically useful information directly from everyday surfaces has resulted in the rapidly rising interest in its application in many areas of science.3 DESI has become very attractive in forensics where it can enable trace analysis of polymers, small molecules, drugs, and explosives directly from complex matrices, detection from fingerprints, chemical mapping, or ink analyses.14−18 Personal care products can be analyzed straight from skin, hair, or textiles.19,20 In the pharmaceutical, biomedical, and diagnostics fields, it enables detection of a wide variety of molecules, including proteins, peptides, lipids, drugs and metabolites, carbohydrates and nucleic acids, which can be done directly from plant, cell, or tissue substrates.11,21−23 DESI MS can achieve femtomolar and Published 2014 by the American Chemical Society
Received: June 5, 2014 Accepted: September 10, 2014 Published: September 10, 2014 9603
dx.doi.org/10.1021/ac502075t | Anal. Chem. 2014, 86, 9603−9611
Analytical Chemistry
Article
Table 1. Suggested DESI Source and Mass Spectrometer Operating Conditions Alongside the Minimum and Maximum Values Used by the Respondees units
suggestedd
min
max
average (n)e
mm mm mm degrees degrees μL/min V psi μm
1.7 0 4.5 35 10 2 +5000 100 50 90% acetonitrile, 10% waterb 0.1% formic acidb fused silicac
0.5 0 1.5 33 0 1.5 +20 80 50 n/a n/a n/a
2.8 0 9.0 80 25 6.7 +5000 170 75 n/a n/a n/a
1.6 (20) 0 4 (19) 48 (20) 9.8 (13) 2.2 (20) 2902 (20) 108 (18) 52.5 (10) n/a n/a n/a
°C
275 positive 30 25
19.6 n/a n/a n/a
350 n/a n/a n/a
248 (19) n/a n/a n/a
parameter a
DESI Source Parameters sprayer-to-surface distance, d1 capillary-edge-to-surface distance, d2 sprayer-to-capillary distance, d3 angle of spray to normal, α angle of capillary to surface, β electrospray solvent flow rate, S electrospray voltage, Vs gas pressure inner diameter of capillary emitter solvent volume fraction, X additives to solvent spray tip (material) Mass Spectrometer Parameters capillary temperature ion polarity data acquisition time data acquisition time taken into calculations
s s
a A diagram showing these parameters and symbols may be found in Figure 1 of Green et al.6 bSix respondees used different solvent compositions, including (i) 90% acetonitrile, 10% water; (ii) 80% acetonitrile, 20% water, 0.1% formic acid; and (iii) 90% acetonitrile, 10% water, 0.1% acetic acid. c One participant used methylated silica. dOne participant used the recommended DESI geometrical conditions, 14 used the recommended solvent composition, and 16 used the recommended electrospray flow rate. eThe parameter n represetns the number of respondees who provided the details.
instruments in the positive ion mode, according to their standard procedures and ensure that the mass scale of the spectrometer has been calibrated in the mass range from m/z 400 to m/z 1400. Furthermore, there is no single, universal set of DESI parameters that could be successfully applied to every DESI MS measurement. It is known that each analyzed substance, e.g., proteins, lipids, explosives, etc. has its optimum electrospray-to-surface distance and optimum spray angle.3 These and many other parameters can affect the signal intensity, stability, and repeatability.6,30 NPL established and provided analysts with guidance on the DESI MS operating conditions. The suggested parameters were evaluated using a DESI 2D Omni Spray Ion Source (Prosolia, Indianapolis, IN, U.S.A.), with an LTQ Orbitrap Velos mass spectrometer (Thermo Scientific, Bremen, Germany). However, the analysts were advised to operate their DESI MS instruments under conditions ensuring the most stable analytical performance, and using the recommended parameters only if suitable. The recommended settings, alongside minimum, maximum, and averaged values, applied by the respondees, can be found in Table 1. From the Rhodamine B samples, participants were asked to acquire 55 positive ion spectra in m/z 400 to m/z 500, each from a fresh area of Rhodamine B, as well as 3 blanks in the wider mass range of m/z 50 to m/z 600. The sequence for the array of analyses is shown in Figure 1. The participants were advised to take these measurements on day 1 and to maintain the same capillary-to-surface distance and continuous flow of the electrospray between the 58 acquisitions. For each mass spectrum, the peak intensities were summed over the mass range m/z 443.0 to m/z 443.5 This sum was entered into a spreadsheet provided by NPL, which automatically calculated the absolute intensity repeatability, as discussed later. The Rhodamine B sample forms a uniform layer of a distinct pink color. This allows the erosion of the sample in the DESI analysis to be clearly visible. Twelve (12) out of the 20
practice, (ii) to evaluate the equivalence of results between different laboratories and instruments, and, since this was the first such study on DESI MS, (iii) to enable a survey of measurement issues and (iv) to help establish a community of users interested in DESI MS metrology.
■
EXPERIMENTAL SECTION Sets of two reference samples were prepared and supplied to each participant. The reference materials were a thin layer of Rhodamine B (Sigma, Poole, U.K.) and double-sided adhesive tape (“Mammoth”, Everbuild), each on standard glass slides with one end frosted (76 mm × 26 mm × 1 mm; Twin Frost, Fisherbrand, U.K.). Each sample was given a unique reference number. The Rhodamine B samples were prepared according to a protocol described earlier.6 Briefly, Rhodamine B was evaporated onto a glass slides, using a vacuum coater (Edwards, Model AUTO 306) with the deposited thickness measured by spectroscopic ellipsometry (Woollam, Model M-2000DI). The average thickness of the Rhodamine B layers, measured from 28 silicon wafers used as controls in each glass slide coating, was 785 nm ± 47 nm (6% RSD). The double-sided adhesive tape was cut into 6 mm × 35 mm strips and arranged in three rows on the clean glass slide with the protective strip still on the surface. To ensure the uniform thickness of the tape, a rolling tool was used. Fresh samples of the tape, after removing the protective strip, were analyzed by DESI MS in the mass range of m/z 600−1400. Intensities in the three mass ranges(i) m/ z 670−700, (ii) m/z 774−808, and (iii) m/z 1355−1380 were extracted. The exact chemical composition of the tape is unknown, but analysis by X-ray photoelectron spectroscopy and secondary ion mass spectrometry suggest that the adhesive is rubber-based. DESI can be interfaced with a wide variety of mass spectrometers. Therefore, it was not possible to provide specific guidance for the mass spectrometer settings and operation. Participants were requested to optimize their 9604
dx.doi.org/10.1021/ac502075t | Anal. Chem. 2014, 86, 9603−9611
Analytical Chemistry
Article
Table 2. Types of DESI Models and Mass Spectrometers Used in the Study
Figure 1. Schematic showing the top view of the analysis side of the reference samples: Rhodamine B and adhesive tape. The directions of the MS inlet capillary and the sprayer are indicated. The directions of analyses are specified by numbered spots and illustrated as a directional sequence (on the right). The axes, x and y, are also defined. Spots 1a, 22a, and 44a on the Rhodamine B sample and spots 1a, 7a, and 14a on the adhesive tape sample mark the places from where blank data were recorded.
respondee
DESI model
MS model
R01 R02 R03 R04 R05 R06 R07 R08 R09 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20
commercial home-built home-built home-built commercial home-built home-built home-built commercial commercial home-built commercial home-built commercial home-built home-built home-built home-built commercial home-built
Orbitrap ion trap ion trap ion trap Orbitrap Quadrupole TOF ion trap ion trap ion trap ion trap modified TOF Quadrupole TOF ion trap ion trap ion trap Quadrupole TOF Orbitrap ion trap ion trap ion trap
participant, as a percentage standard deviation of the absolute intensity of the Rhodamine B peak. The average absolute intensity repeatability across all the 20 participants is 50%, as shown later in Table 3 and Figure S-1 in the Supporting Information. There is significant laboratory-to-laboratory variation in the repeatability standard deviation for these absolute intensities. This could be caused by many factors, such as the ability of the instrument to deliver a uniform spray, stability of the detection, and so on. In addition, contamination or variations in the ambient air may be important. Figure 2 shows how some of the measurements exhibited distinct trends that could be related to influence factors, rather than random scatters that would be more difficult to control. In Figure 2a, with regard to the data of respondee R13, there is a clear increase and decrease of the Rhodamine B intensity that is related to the analysis position along the x-direction of the sample. The vertical lines on the figures indicate the start of a new acquisition row. A similar drift was observed in intensities from participants R08 and R11. This effect may arise from a slight tilt of the sample on the stage or a similar effect that causes the sample height to be altered, relative to the DESI input capillary. These data illustrate how a reference sample may be used to verify the correct sample positioning and test the motion of the sample stage. A second group shows a gradual reduction or gradual rise in the Rhodamine B absolute intensity recorded across the 55 measurements, as illustrated in Figure 2b and Figure S-2(a) in the Supporting Information, respectively. This reduction was observed for four participants (R01, R09, R10, and R18) and an equivalent rise for two (R05 and R06). These changes could be caused by mass spectrometer performance over time. However, it is more likely that they arise from the stability of the electrospray flow. Even though participants were advised to allow 30 min for the electrospray to stabilize prior to experiments, it may require longer. The observed drifts contribute significantly to the repeatability of the measurements. As an example, the signal intensity of the Rhodamine B
participating laboratories returned their samples, after analysis, to NPL for their further examination. The samples were inspected by optical microscopy (Alicona InfiniteFocus, Sevenoaks, U.K.). The eroded area was measured using ImageJ software.31 The measurements of the adhesive tape samples were to be made over eight, not necessarily consecutive, days, from 10 separate strips of the tape. In total, participants were asked to acquire 80 positive ion spectra in the m/z 600 to m/z 1400 range (including 10 blanks in a wider mass range of m/z 50 to m/z 1400), each from a fresh area of the tape, as shown in Figure 1. The protective strip was to be removed just prior to analysis. The data from the first three strips was acquired on day 1 without any change to the DESI settings between the acquisitions. The remaining seven strips were analyzed on following days, with none on the same day as another. For all measurements, a waiting time of 30 min was recommended, before the start of the analyses, to allow the electrospray to stabilize. Further details of the acquisition setup and the sequence used to take data can be found in the study protocol.32
■
PARTICIPANTS The participants wishing to be identified are listed in the Acknowledgments section. Data were supplied from 13 ion trap instruments, 3 Orbitrap instruments, and 4 TOF instruments. Seven of the DESI interfaces were commercial, whereas 13 were home-built. Table 2 provides a summary of the type of DESI source and mass analysers used. No single participant followed all of the recommended DESI MS settings shown in Table 1.
■
RESULTS AND DISCUSSION Absolute Intensity Repeatability of Rhodamine B. The repeatability of Rhodamine B measurements was calculated, based on 55 consecutive DESI MS measurements taken by each 9605
dx.doi.org/10.1021/ac502075t | Anal. Chem. 2014, 86, 9603−9611
Analytical Chemistry
Article
The blank data sets acquired between Rhodamine B measurements for 10 of the 12 participants who returned data showed negligible memory of the Rhodamine B. The Rhodamine B was either not observed in the spectra, or its intensity was significantly lower (100-fold), in comparison to the data obtained from the Rhodamine B layer (see the Supporting Information for more details). The lack of large sample displacements during analysis and a low level of crosscontamination from the mass spectra inlet proved the effectiveness of the Rhodamine B layer as an excellent model sample for checking the instrumental setup and repeatability of DESI MS analyses. Overall, Rhodamine B provides significantly useful information. Rhodamine B Sample Erosion. Twelve (12) Rhodamine B samples were returned to NPL after analysis. Optical microscopy was then used to analyze for any correlation between the eroded crater and the absolute intensity repeatability. Eight (8) samples showed consistency in the crater shapes for the 55 points, as illustrated in Figure 3a. Figure 2. Variations in Rhodamine B signal intensity recorded from 55 fresh sample areas by participants (a) R13 and (b) R01. Vertical lines mark the end/beginning of an analysis row and, hence, a change in analysis direction.
peak, in the analysis performed by participant R01, and shown in Figure 2b, decreases over the 55 measurements by a factor of 2. If the drift was removed, the absolute intensity repeatability improves from 19% to 12%. A longer settling time may improve repeatabilities. No common pattern was observed in the data from the remaining participants. Several datasets revealed fairly consistent analyses with a few outlier points, which could have been caused by electrospray instability. For example, a clear spike in the Rhodamine B intensity was observed by participant R12. Optical examination of the returned sample indicated a larger desorption site correlated with the intensity spike (Figure S-3 in the Supporting Information) probably caused by a brief surge in the source. Removal of this result would improve the repeatability of R12, from the 71% value shown in Figure S-1 in the Supporting Information and Table 3 to 31%. Electrospray instability may account for some of the outliers seen in other datasets. However, optical examination did not provide any obvious correlations, and the outliers are all included in the repeatability calculations, since they represent current DESI MS practice. Figure S-2(b) in the Supporting Information illustrates, for participant R16, consistent Rhodamine B intensities for the first four rows on the sample, followed by a sudden decrease in the intensity and a gradual increase along row 5. This may indicate a possible issue with the sample stage movement at this edge of the glass slide. Other causes, such as intermittent, partial blockage of the MS inlet, are also possible. DESI spectra from the blank part of the glass slide, acquired in the mass range from m/z 50 to m/z 600, were returned to NPL by 12 participants. These were examined to evaluate other possible factors, such as the ambient conditions of the DESI and potential contaminants, which could affect the sensitivity and repeatability of the measurements. Various common contaminants were detected in the spectra including polysiloxane, triton, eurucamide, and phthalates. Several blank spectra also showed low-mass acetonitrile clusters.33−37
Figure 3. DESI erosion spots on Rhodamine B samples, with high consistency and regularity in shapes and sizes. Desorption across the 55 measurements for (a) R18 and single spots from (b) R08, (c) R10, and (d) R20.
Typically, the majority of the craters show similar characteristics: an inner ring of fully desorbed sample material and an outer ring of displaced material, often with a ridge of accumulated sample between the two.6 The area of the inner ring, examined on the returned samples, ranges from 0.11 mm2 to 0.75 mm2. The outer disturbed area shows more disparity, with a large variation in shapes and with total desorbed areas ranging from 0.7 mm2 to 4.5 mm2. A good example of DESI craters is shown in Figure 3b, where the spot is fairly regular in shape and elongated in the direction of the spray. Between the 8 samples with consistent crater shapes, 5 were similarly shaped, but with variations in size of the inner and outer areas, and 2 displayed outer areas with a stronger directionality, as illustrated in the example in Figure 3c. An unusual crater shape was observed on the sample from participant R20 (Figure 3d). It contained barely any outer disturbances: only the inner desorbed area and the ridging region. None of the experimental settings for R20 appeared to differ from the settings applied by other participants. This desorption shape indicates significantly less mixing of the 9606
dx.doi.org/10.1021/ac502075t | Anal. Chem. 2014, 86, 9603−9611
Analytical Chemistry
Article
repeatability of the absolute intensities is not sufficient to allow any such conclusions. Repeatability of the Adhesive Tape Measurements. Adhesive tape samples were used to measure both absolute and relative intensity repeatabilities. Representative mass spectra of the adhesive tape, and the mass ranges selected for analyses in the positive mode, are shown in Figure S-4 in the Supporting Information. Participants were asked to acquire 8 datasets from fresh areas of the adhesive tape. The first set consisted of 21 spectra on the first day, followed by 7 further sets of 7 spectra, each from a fresh area of the strip on later days. From each spectrum, the absolute intensities in the three preselected mass ranges were summed and the absolute repeatability was calculated as a relative standard deviation of the absolute intensity. The absolute and relative intensity repeatabilities (Ra and Rr, respectively) were determined for the same 8 sets of measurements, using the method previously described in Gilmore et al.38 Briefly, Ra, expressed as a percentage, is evaluated from the average of the j0 (here, 21) measures, Iij, for the summed intensity in each of the 3 mass ranges, i, where
sample material on the analyzed surface, which could bring a clear advantage in terms of spatial resolution of DESI. Four of the returned samples revealed poor uniformity in terms of the eroded spot shapes and sizes. The sample analyzed by participant R11 gives one example of an inconsistent erosion pattern across 55 consecutive measurements (see Figure 4a). A wide range of spot shapes is illustrated in Figures 4b−d.
j0
Ii̅ =
∑ j=1
Iij j0
(1)
This allows us to remove the overall intensity variation between the three different mass ranges. Ra is then given by the average of the standard deviations (si) of the j0 values of Pij, where Pij = Iij/Ii̅ :
Figure 4. Irregular DESI erosion spots on Rhodamine B samples. Desorption across the 55 measurements for (a) R11 and single spots from (b) R06, (c) R16, and (d) R14.
The optical images from the Rhodamine B samples indicate the spatial resolution of the method. The average crater is an ellipse with axis dimensions of 1.4 mm × 2.3 mm. The inner areas of the desorbed material tend to be more consistent and smaller, with average dimensions of 0.5 mm × 0.8 mm and a relative standard deviation of ∼30% for the 12 participants. The smallest dimensions observed were
