Instrument-Independent MS/MS Database for XQQ Instruments: A Kinetics-Based Measurement Protocol

A detailed kinetics-based measurement protocol is proposed for the development of a standardized MS/MS database for XQQ tandem mass spectrometers. The technical basis for the protocol is summarized. A CAD database format is proposed.


Introduction
Tandem mass spectrometry (MS/MS) is widely used for structure elucidation and for the analysis of multicomponent mixtures [1]. The analysis makes use of the collisionally-activated dissociation (CAD) of "parent" ions. A parent ion may be a molecular radical cation, a protonated molecule, or a "progeny" fragment ion (daughter, granddaughter, etc. produced by the fragmentation of a larger precursor parent ion). The MS/MS technique consists of three steps: (i) a parent ion is mass selected by the first mass analyzer (designated MS-I); (ii) the selected parent ion collides with a target gas which is located within a reaction region between MS-I and the second mass analyzer (designated MS-II); and (iii) the undissociated parent ions and the progeny fragment ions which were produced within the reaction region are channeled into MS-II for mass analysis. MS/MS instruments thus produce a CAD spectrum of each initially-selected parent ion.
Reference [1] provides an excellent up-to-date review of the capabilities and advantages of the MS/MS technique. Among the advantages enumerated in [1] for MS/MS are: (a) the functional group specificity of neutral loss experiments (for which there is no analogue in GC/MS) provides the unique ability to rapidly screen unknown mixtures for compound classes (e.g., nitroaromatic cations can be inferred by their loss of NO). (b) the ability to perform very rapid analyses of targeted compoimds in complex mixtures based on unique progeny ions (especially when "soft" ionization techniques are used) or on neutral loss experiments [e.g., the nitroaromatics screened above in (a)], thus avoiding the need to identify every component. (c) very low limits of detection (<ppt; better than a single stage of mass spectrometry) can be attained with enhanced signal/noise ratios be-cause of the elimination of "chemical noise" [1}. (d) unstable or extremely reactive species, which cannot be studied directly by optical spectroscopies, NMR, or GC/MS, have been studied directly by MS/MS (e-g., radical trapping of polyatomic free radicals; neutralization-reionization mass spectrometry [1]; etc.). In principle, standard CAD spectra of a variety of ions (fragment ions, molecular ions, protonated molecules, etc.) could be generated and collected as reference libraries, to be used for comparison against unknown spectra in a manner analogous to the use of reference libraries in the data handling systems of ordinary electron impact mass spectrometers. Further, it should be possible to infer the identity of an unknown molecule by identifying the ionic substructures of fragment ions generated in its CAD spectrum. However, to date reference libraries of CAD spectra have not been collected because of a lack of standardization of operating conditions of MS/MS instruments [IJ.
This paper addresses the standardization of one of the major classes of commercially-available analytical MS/MS instruments, the so-called "XQQ" tandem mass spectrometers (there are currently more than 400 XQQ MS/MS instruments worldwide, representing a capital investment in excess of $170 million). XQQ instruments have three components: 2^ 0101-2Li designates the first mass analyzer, which can be a quadrupole mass filter (represented by a Q), a reversed-geometry magnetic/electrostatic sector instrument (represented by BE), etc. Q2 designates a quadrupole mass filter operated with only radiofrequency (rf) potentials [no direct current (dc) potentials are used]. Q2 contains the target gas, and provides efficient ion containment of parent ions and of progeny fragment ions. Q3 stands for the second mass analyzer, a quadrupole mass filter.
Because XQQ instruments (QQQ, BEQQ hybrid, etc.) are complex ion-optical devices [2]- [9], the observed spectra can be extremely dependent on the experimental parameters used during the analysis. That is, one observes instrument-dependent CAD spectra. This was clearly demonstrated in a 1983 international round robin [10] wherein very different CAD spectra were observed for the same molecule. That is, the relative intensities measured in different QQQ instruments for any given pair of progeny ioris differed by factors ranging into the hundreds, even though the same nominal operating conditions were supposedly used in each of the QQQ instruments. Therefore, until now it has not been possible to use a CAD spectrum of a given species in one XQQ instrument to identify and quantitate that same species in a different XQQ instrument.
To develop an instrument-independent MS/MS database (or library) for XQQ instruments, instrument users must be able to tune their instruments to obtain undistorted, dynamically-correct (i.e., instrument-independent) representations of any reactions studied within such instruments [9]. That is, the measurements must be substantially free from kinetic interferences (viz., every effort must be made to ensure a well-defined gas target and to prevent back reactions, impurity reactions, scattering losses, fringing fields, mass discrimination, etc.). The branching-ratio measurements must be precise and accurate (± 10%).
Recent work from this laboratory has explored the prerequisite conditions for obtaining instrument-independent dynamically-correct branching ratios for the CAD of ions within XQQ instruments [9]. In this paper a kinetics-based measurement protocol is described for the determination of standard CAD spectra within XQQ instruments. The technical basis for the protocol is summarized. Table 1 includes a typical CAD spectrum measured using this protocol.
This protocol can also be used for the development of a standardized MS/MS database for XQQ instruments. One of the goals of this paper is to promote discussion about the database format best suited to the needs of the analytical mass spectrometry community. The spectrum of table 1 illustrates the database format proposed here. The precepts of the protocol are also applicable to other types of tandem mass spectrometers which have strong focusing properties (e.g., quadrupolehexapole-quadrupole MS/MS), so long as the energy range used for the pertinent ion chemistry is the same as for XQQ instruments.
The kinetics notation and nomenclature developed in reference [9] are used throughout this paper.

General Background
As has been demonstrated [11]- [16], dynamically-correct branching ratios can be measured in XQQ instruments when the key MS/MS instrument parameters [9] are properly selected to account for reaction-induced mass discrimination [9] within Q2 and the intrinsic mass discrimination within the Q3 mass analyzer. for the CAD of (C2H3O+) from the source compound biacetyl [16]. ^ The numbers in the square brackets represent the maximum uncertainty in the cross section or and in the branching ratios (cc -ij), expressed as a percentage of each cr and of each branching ratio {e.g., for biacetyl at £CM = 2.4 eV, the maximum uncertainty in y is ±([2%]/100)y; i.e., y=0.999±0.02}. " A reference citation would be provided for each CAD spectrum to identify the source of the data. ' For the CAD of any given parent ion (e.g., C2H3O+), appropriate corrections must be made for contributions from the concurrent CAD of isobaric ions (e.g., C3H7+), regardless of their source. The isobaric ions may be co-produced in the ion source (i) from the source compound (e.g., ionization of CH3C(0)C3H7 will produce both CH3CO+ and C3H7+) and/or (ii) from a neutral impurity in a source compound (e.g., for a butanol impurity in a 2-butanone source compound, the butanol generates C3H1+, while the 2-butanone generates CH3CO+). Fortunately, the CAD spectra of CH3CO+ and C^Ht are easily distinguishable [16]. That is, for the CAD of CiW, C2H3+ {m/z 27) is the major CAD fragment for £'CM-2-80 eV. By contrast, for the CAD of C2H3O+, C2H3+ is not produced at EcM. = 2.4 eV, and is only a very minor fragment for ^CM > 2.4 eV. Unfortunately, even minor impurities can contribute disproportionately to the CAD spectrum of a source compound because of differences in the CAD dynamics of isomeric and/or isobaric ions. Because of this problem, it is advisable that both the source compound and the target gas be of high purity (>99.95%). Otherwise, the impurities must be characterized so that appropriate corrections can be made for their contribution to the observed CAD spectrum. ' The Notes field would be used to refer a user of the database to any information of special significance about the parent ion (e.g., structure of the ion, etc.). The notes enumerated in all the Notes fields of the CBRIS database would be collected together in a separate "Notes Appendix". For the example given here. Notes 9, 99, and 999 would be found in the Notes Appendix, and might contain the following types of information. Note 9: For a given Ecu, the branching ratio for each fragment ion is substantially the same for all CH3CO-X source compounds (e.g., biacetyl, acetone, acetophenone, etc.) [16]. The reactant ion entering Q2 appears to be pure CH3CO+ in every case [16]. Note 99: The energy dependence (magnitude and direction) of the branching ratios is distinctly different for the isobars C2H3O+ and C3H7+ [16]. Hence, one can readily distinguish C2H3O+ from C3H7+. Note 999: One can readily distinguish ethanol, oxirane, and cis-2,3-epoxybutane from each other, and from CH3CO-X source compounds on the basis of the energy dependence of the branching ratios for the CAD of C2H3O+ [16].
The systems studied in this laboratory include charge transfer and dissociative charge transfer reactions, and CAD.
Charge Transfer: Ne++Ne-vNe+Ne+ [11], Ar+-j-Ar->Ar-f Ar+ [12], and Ar+ + N2->Ar-fN2+ [13] Dissociative Charge Transfer: (where x = 1-5) [14] CAD: Ar CAD (CH3)2CO+ > CH3CO+ -t-CHj [15] Ar CAD CH3CO+^ ^CHj+HhCO [16] Note that for collision energies £LAB=1 -200 eV in the Laboratory (LAB) frame of reference, charge transfer reactions are experimentally equivalent to a "worst-case" CAD reaction system because they take place at large impact parameters with nearzero momentum transfer. That is, the product ions of a charge transfer reaction are formed essentially at rest (thermal energies) within Q2 [17]- [20]. The studies in our laboratory led to the development of a kinetics-based measurement protocol for the generation of standard XQQ spectra. Using the protocol, one can obtain, for the first time, accuracy and precision for CAD measurements within XQQ tandem mass spectrometers. A round robin exercise, the NIST-EPA International Round Robin [21], was organized to ascertain which commercially-available XQQ instruments are capable of generating dynamically-correct (i.e., instrumentindependent) spectra. Appendix 1 contains the actual roimd-robin protocol which was disseminated to 22 laboratories worldwide (including the six XQQ instrument manufacturers). Analysis of the round robin data from six participants indicates that at least 50% of the QQQ instruments which have been sold and are currently in the field can provide such standard spectra [21].
Instruments which can generate dynamicallycorrect results with the round robin protocol of Appendix 1 can be used to develop a generic, standardized CAD spectral database (or library) based on Characteristic Branching Ratios of Ionic Substructures (CBRIS) (analogous to the use of group frequencies in infrared spectroscopy; see section 5.1). Hence, members of the analytical mass spectrometry community who have such instruments could use the kinetics-based measurement protocol detailed in this paper to generate and contribute instrument-independent CAD spectra of species studied during the course of their work. Contributed CAD spectra (to be sent to this author) would be included in the NIST-EPA Standardized CAD Database currently being developed in our laboratory. The latter would be disseminated by NIST to the analytical mass spectrometry community. To facilitate the critical evaluation of contributed spectra, a dynamically-correct CAD spectrum of a well-studied "model" compoimd (e.g., n-butylbenzene) should also be submitted.

Technical Basis for Protocol
Here we summarize the technical basis for the kinetics-based measurement protocol. For a more detailed treatment, the reader is referred to references [2]- [9].

Kinetics
With reference to the following general reaction sequence: etc.
In the case of CAD, A"*" corresponds to the parent ion; B corresponds to the target gas; C"*", D"*", etc. are the progeny fragment ions; and S, T, etc. are the neutral fragments complementary to C"*", D"*^, etc. Under dynamically-correct conditions, the rate of reactant ion decay equals the rate of product ion formation. That is In y=ln W^=\3X W^, etc. Then the product ion intensities C"*", D+

Dynamically-Correct Representation
If the key MS/MS parameters [9] of an XQQ instrument can be tuned so that data generated by the instrument conform to equations (l)-(5) [viz., so that cc+fi+y+... = 1.00], then the instrument can provide a dynamically-correct representation of any reaction studied within it. That is, by using eqs (4'). (5')» etc., the instrument would provide a measure of the branching ratios equivalent to those that, in principle, would be observed at the scattering center of a molecular beam machine.

Kinetics Constraints
The selection of key MS/MS parameter settings is constrained by the kinetics prerequisite for <x -|-y8-|-y+-=1 (see reference [9]). This prerequisite necessitates that: (a) each product ion be formed only by the primary reaction (no secondary sources; pseudofirst order, single-collision conditions) (b) all ions be detected with equal sensitivity (requires Conversion Gain corrections for each detector; see sec. 4) (c) there be no scattering losses because of unreactive collisions (must have high ion containment within Q2; approximately 100% collection efficiency for product ions and unreacted projectiles) (d) corrections be made for differences in ion containment (transmission) within the Q2Q3 structure.

Instrumental Parameters
To accomplish <x-f/3-t-'y + ...= 1.00, the key MS/MS parameters must be properly selected to obviate discrimination against product ions because of: (5) We shall use the tern^ "RIMD-free" and "IMDfree" to refer to branching ratios oc, yS, etc. which have been measured in concordance with the precepts detailed next for (a)-(c). That is, to be RIMD-free and IMD-free, the instrument parameters discussed below in sections 3.2.1-3.2.3 have to be tuned and returned iteratively until K+)S+y-f...=^1.0±0.1.

Reaction-Induced Mass Discrimination (RIMD) Within Q2
The rf amplitudes of Ql, Q2, and Q3 are characterized here by the Mathieu parameters qi, qj, and qz, respectively (for further information, see references [2]- [5]). In this discussion q^i and q^^ are used to represent, respectively, the values of q2 which correspond to the (4') maximum ion transmission through Q2Q3 for the reactant (projectile) ion of mass mreact and for each (5') product ion of mass mprod-The subscripts "react" and "prod" designate, respectively, the reactant ion A+ and the product ion C^ (or D+, etc.). Within Q2, m^,^m^^=^q,^.Jq^,^ [9]. Therefore, low-mass daughters are not detected when '"prod/'Wreact<?react/0.908 since lou trajectorfes are unstable when g'2>0.908 [5]. This means that the signal for each product ion must be measured at its respective q^^. For each product ion, its corresponding {[A+jo-[A+]} must also be measured at that same q^^. That is, with reference to eq (4'), oc must be determined by measuring In principle, if XQQ instruments were well behaved, all measurements could be made at the q^^ for the progeny ion of lowest mass (m/z 15 for the example above [5]). However, because of ion-optical imaging (focusing) problems within the Q2Q3 structiire, one usually observes oscillations in the ion intensity as ^2 is varied [5]

The Kinetic Energy of Product Ions Entering Q3
It is now recognized that in the LAB frame of reference the translational energy of product ions ^prod is generally related to the collision energy of the projectile (reactant) ion jBreact by: where n=:l-2 for CAD processes studied to date [8]. Hence, if mprod/'Wreact < < 1. low-mass daughters exiting Q2 will have low kinetic energies. Consequently, the Q3 rod offset (pole bias) must be set sufficiently low with respect to the Q2 rod offset so that no low-mass, low-energy product ions are denied entry to Q3 by the Q3 potential barrier. On the other hand, the Q3 potential barrier must be sufficiently high to provide adequate resolution within Q3 [8]. The reader may wish to consult references [7] and [8] for a more detailed exposition with reference to the selection of the appropriate Q3 rod offset.

Kinetics-Based Protocol for MS/MS Measurements
In this section we detail the generic protocol used to measure instrument-independent CAD spectra in XQQ instruments. The reader is referred to the caveats detailed in items 1-9 of the Precautions in Appendix 1. (a) The protocol of Appendix 1 (for the NIST-EPA International Round Robin) must first be used to validate that any particular XQQ instrument can provide a dynamically-correct (i.e., instrument-independent) representation of ion-molecule reactions. If the instrument cannot pass the protocol (i.e., it cannot provide a dynamically-correct representation), one should not proceed any further until one has eliminated the ion-optical defect(s) so that the instrument can pass the protocol. If the instrument can provide a dynamically-correct representation, then the measurements for Part 1 of the round-robin protocol of Appendix 1 will provide an in situ calibration of the target thickness of the XQQ instrument.
(b) For all MS/MS measurements, one must use a gas target thickness within the single-collision regime [<2 cm-mtorr (<0.267 cm-Pa) for Ar]. For CAD, use an Ar gas target since it has been demonstrated [22] that the equivalent excitation energy (i.e., the equivalent internal energy imparted to a parent ion) is significantly less for a polyatomic target than it is for a monatomic target. However, one may also use other inert gas targets (He, Ne, Kr, or Xe) as necessitated by the coUisional energy requirements since it has been demonstrated that the equivalent excitation energy is the same for the inert gases at any given center-of-mass energy where m\ and mi are, respectively, the masses of the reactant ion and of the gas target}. (c) One must use a. fixed operating voltage for the ion detector (multiplier, Daly detector, etc.). The fixed voltage should provide the best compromise between, the signal-to-noise ratio (5/ N) of A+ and the S/N of C+, D+, E+, etc, (d) One must make Conversion Gain measurements if an instrument uses analog current measurements. Conversion Gain is the ratio of the electron current output from an electron multiplier relative to the ion current input. The Conversion Gain measurements are used to correct for differences in mass-dependent detector response. If the instrument uses true ion pulse counting, Conversion Gain measurements are not needed [i.e., ignore instruction (d) and continue with instruction (e)]. Warning: Some instruments use analog current measurements, but report the ion intensities as equivalent ion count rates within their data systems. Such instruments still require Conversion Gain measurements. Note: If possible, one should adjust the Conversion Gain of the ion detector so that one attains a mass-independent detector response. That is, so that the Conversion Gain is the same for a projectile ion and for its lowest-mass product ion (e.g., for CAD) or highest-mass product ion (e.g., for neutral gain reactions). This might be accomplished, for example, by adjusting the operating voltage of a high-voltage (e.g., 20-30 kV) conversion dynode which is used in conjunction with an electron multiplier. If it is not possible to attain a mass-independent detector response, then one must make corrections for difference in Conversion Gain for each parent ion and for each product ion.

Jovirnal of Research of the National Institute of Standards and Technology
To measure Conversion Gains, one may use a procedure analogous to that used for instructions 30-39 in Part 2 of the NIST-EPA round robin protocol in Appendix 1. However, to ensure that the most reliable correction factors are obtained, the mass-dependent Conversion Gain measurements would have to be made under actual operating conditions (viz., with the conversion dynode and electron multiplier at high voltages). This might be done by measuring the ion current striking the conversion dynode and relating it to the output current from the electron multiplier. One would have to demonstrate, however, that there is a one-to-one correspondence between the ion current input and the electron current output, independent of ion structure. Note: For the rest of the instructions which follow, it is presumed that the ion signals correspond to a mass-independent detector response. That is, if necessary, the ion signals have been corrected for differences in Conversion Gain efficiency. surements determine £50 (corresponds to the Q2 rod offset that stops 50% of the reactant ions). Then set the Q2 rod offset equal to ^50-£^LAB. where £LAB is the requisite collision energy (in LAB coordinates). (g) One must use only a "daughter-scan" mode.
That is, Ql is tuned to the peak position which corresponds to the maximum ion intensity at the m/z of the projectile ion while Q3 scans over the peak which corresponds to the m/z of a product ion. The q^ must be referenced to q^ (i.e., to 9react) rather than to q^ (i.e., to q^^oi)-Use a scan window of ca. 4-10 amu so one can see the entire mass peak (baseline-to-baseline). Note: To ensure consistent measurements for the duration of each experiment, make sure Ql is staying timed to the peak position for the projectile ion intensity; this can be done by varying the Ql mass command to maximize the ion signal being viewed within the Q3 scan window. Use the same fixed operating voltage for the ion detector as was used for instruction 4 (c). (h) One then adjusts iteratively the resolution parameters for Q3, and the Q3 rod offset as necessary to approximate a-f-;8+y-f .. etc. which may be observed in the absence of added target gas. Such background contributions can arise from decompositions of A"*" (to produce C""", D+, etc. via CAD or RIF [9]) occurring between the rear of Ql and the front of Q3. Such background CAD or RIF may be induced by the interaction of A^ with plumes of gas emanating from the ion source, (i) In practice, instructions (b)-(h) provide IMDfree and RIMD-free estimates for the branching ratios of fragment ions, except for very minor fragment ions adjacent to major ions. The resolution is not adequate for obtaining dynamically-correct estimates for the branching ratios of very minor fragment ions adjacent to major ions. Consequently, the resolution must be increased for each group of ions which contains a minor ion adjaent to a major ion, and complementary measurements must be made for the branching ratios of minor fragment ions within each group. Each group contains only fragment ions which are in close proximity (e.g., m/z 13-15), so that the measurements for each ion in the group are still IMD-free and RIMD-free within that group.
To relate these individual group measurements to the overall CAD spectrum [viz., to the entire range of masses from the m/z of the reactant (parent) ion down to the m/z of the lowest-mass product ion], the measurements for each group must be referenced to the IMDfree and RIMD-free estimate for the branching ratio of a major fragment ion within that group. For example, for ^CM^^Z-40 eV, the CAD of C2H3O+ {m/z 43) from biacetyl pro-

G)
duces fragment ions at the following m/z: 13 Table 1 shows a typical CAD spectrum (measured with the protocol of sec. 4 [16]) in the format proposed for presentation of CBRIS.

Advantages
The advantages of this CBRIS database format are: (a) The partial cross sections acr, ^cr, etc. can characterize both known and unknown species (so long as the unknown species contain ionic substructures for which the CAD cross sections and product identities are known). Therefore, it may be possible to assign the structure of an unknown species on the basis of the absolute cross sections o-jj, o-jt, etc. for the CAD of known ionic substructures i, j, k, etc. That is, this proposed use of CBRIS in MS/MS is analogous to the use of group frequencies in infrared spectroscopy. (b) Characterization of an unknown compound by using CBRIS does not require that the compound be in the CBRIS database. By contrast, to characterize an unknown by spectral matching within a "library", the compound must usually be in the library. In this regard, development of a database of CBRIS for all substructures (or a subset thereof) would be very much more tractable than the development of a library of CAD spectra for all the source compounds that contain all the substructures (or a subset thereof). For example, consider the simpUstic analogy where compounds correspond to words, and substructures correspond to letters: more than 500,000 words can be composed with only 26 letters of the alphabet. (c) The format is compatible with its use in expert systems. This should facilitate rapid real-time analysis of unknowns within computer-controlled field instruments. (d) End users are involved directly in its evolution by using critically evaluated cross sections already in the database and by submitting new cross sections for inclusion in the database.

Conclusions
A kinetics-based measurement protocol can provide accuracy and precision for CAD measurements within XQQ tandem mass spectrometers. This protocol can be used to develop a dynamically-correct (i.e., instrument-independent) MS/ MS database (or library) for XQQ instruments.

Appendix 1. NIST-EPA Round Robin for XQQ Instruments (QQQ, BEQQ, etc.): A Test Protocol to Assess Which Instrument Designs Provide Instrument-Independent (Dynamically-Correct) Performance
Note: This appendix is the full text of the actual document which was sent to, and was used by, the participants in the NIST-EPA International Round Robin. The round-robin results have been published elsewhere [21]. This document may be photocopied. Prospective participants may submit test data to the author anytime for this ongoing evaluation project.

Objective
To assess which XQQ instrument designs can be used to generate a generic, instrument-independent MS/MS database. To do so, an XQQ instrument (QQQ, BEQQ, etc.) must provide a dynamically-correct (instrument-independent) representation of any ion-neutral interaction (e.g., CAD, etc.) studied within its Q2Q3 structure. The prerequisites for obtaining dynamically-correct branching ratios within XQQ instruments have been detailed in reference [9].

Data Analysis
To ensure consistency, all data will be analyzed by the round-robin coordinator. The participants will provide only "raw data" on the forms provided with the test protocol. To perform the analysis, the raw data will have to include information about such technical specifications as field radius, rf frequency, etc. Please submit one completed copy of the Questionnaire for each XQQ instrument included in the round robin.

Confidentiality
All information provided by participants will be confidential. No proprietary information will be divulged. No participant or instrument manufacturer will be identified (the data files will be encrypted with code letters, numbers, and symbols known only to the round-robin coordinator). All test results will be discussed only in generic terms.

Summary of Test Protocol
The test protocol is based on (i) the concepts detailed in reference [9], and (ii) the use of the ion-neutral reaction kinetics of references [12] and [14].
The following is a synopsis of the two major components of the test protocol. The nomenclature used herein is defined in references [9], [12], and [14].

Parti
Purpose (i) To determine if the reaction kinetics are well controlled (i.e., no back reactions, no scattering losses, minimal fringing fields, no mass discrimination, well-defined gas target, etc.) within each XQQ instrument. The symmetric charge transfer reaction ^*Ar"'"-f''°Ar->'*Ar-f-''°Ar'^ is used (see reference [12]). In this case there is no reaction-induced mass discrimination (defined in reference [9]).
(ii) To calibrate the response of the pressure sensor closest to the participant's collision region in terms of the effective target thickness [Ar]Z/efr within the collision region {L^f is the effective pathlength of the complex oscillatory trajectory traversed by a projectile ion through the CAD target gas within Q2). This is done to ensure all participants use the same effective target thickness measurements.

Glossary
Here we describe some terms used below. NISTCALPLOT A calibration curve of In Y (=hi W) vs effective target thickness [Ar]Leff.
NISTCALPLOT ( fig. 1)   Using several different pressures (P/XQQ) of Ar target gas, each participant measures In Y and In W for the reaction "Ar+-t-*Ar->^*Ar-f "^Ar^. NISTCALPLOT is then used to convert the participant's In Fand In W measurements to the corresponding target thicknesses ARLEFF/LNY and ARLEFF/LNW, respectively, These data are then vised to prepare the ealibration graph XQQCALPLOT, If the kinetics are well controlled within the participant's XQQ instrument, then the two plots of ARLEFF/LNY vs P/XQQ and ARLEFF/LNW vs P/XQQ should be superimposable.

Purpose
To assess how well one can correct for reaction-induced mass discrimination and for strong ion-optical imaging (focusing) problems within the Q2Q3 structure (i.e., how well one can control the key MS/MS parameters in different XQQ instrument designs) (see reference [9] for further information).

Method
Each participant measures In Y and In W for the reaction N2^+SF6->-N2+SF5"'' as a function of the pressure P/XQQ of the SFs gas target (cf. reference [14]).

Data Format
Different types of data are requested within the test protocol. Most of the test data to be provided by the participants will consist of "filling-in-the-blanks" in the spaces provided with the test protocol.
For laboratories with more than one XQQ instrument participating in the round-robin test, please submit a separate set of data for each XQQ instrument (use photocopies of this test protocol document to record each separate set of data).
To provide the best possible assessment of instrument Hardware design (without obfuscation by software artifacts), and to ensure consistency among the data submitted, participants must not use any of the automated capabilities.
Instrument control and data acquisition should be as close to "manual" control as is possible with your instrument. You may use a "hybrid" control system wherein you are "manually" controlling the parameters [e.g., programmable "soft-knobs" (EXTREL') or "tuning tables" (Finnigan TSQ70) are used to vary the dc voltage which is being applied to a lens via a digital-to-analog converter (DAC)]. However, use only a single "tuning table" with "flat" (constant) parameter settings; don't use "programmed ion optics". Please niark the top of the first page of the submitted data with the following label: Manual Instrument Control and Data Acquisition.
However, for any one XQQ instrument being tested, participants are welcome to submit a second set of data to assess possible artifacts in the instrument control and data acquisition software. Please mark the top of the first page of the submitted data with the following label: Automated Instrument Control and Data Acquisition.
If your system has only automated capabilities (i.e., there is absolutely no way that it can be made to function in a manual or hybrid mode of operation for instrument control and data acquisition), please mark the top of the first page of the submitted data with the following label: Non-Manual Instrument Control and Data Acquisition.
' Certain commercial equipment, instruments, and materials are identified in this paper in order to adequately specify the experimental procedure. In no case does such identification imply recommendation or endorsement by the National Institute of Standards and Technology, nor does it imply that the material, instruments, or equipment identified is necessarily the best available for the purpose.

Actual Voltage Measurements
The actual voltages of all essential ion-optical elements [rod offset (pole bias) of Ql, Q2, Q3; lens potentials; etc.] must be measured with a voltmeter at the appropriate measurement test points closest to each ion-optical element. Consult the manufacturer for the location of each test point and the method of measurement that you should use to ensure correct values (e.g., to measure actual rod offsets, you may have to disconnect the inputs for mass command and/or resolution control and/or LM control). Please make actual voltage measurements to validate any voltage information provided by a computer system via an analog-todigital converter (ADC).

Ql Rod Offset and Projectile Energy Distribution
To be kinetically meaningful, participants must ensure: (a) that [B]o> > 100[A"'"]o (in molecules cm~^ within Q2; ion intensity data are converted to the approximate ion concentrations with reference to the defming aperture which provides entry into Q2), (b) that the measured reactivity can be attributed to a well-characterized state of the projectile ion A^, and (c) that the translational energy distribution of the projectile ion entering Q2 is reasonably narrow (cf Q2 stopping curve data in references [12] and [14]). [Unfortunately, narrowing the energy distribution of the projectile ion can produce ion-optical imaging (focusing) problems within Q2 if the XQQ instrument has restrictive interquadrupole apertures (see reference [9]).] For QQQ instruments, tasks (a)-(c) can be accomplished by over-resolving Ql (i.e., increasing the resolution of Ql to produce a sharp, narrow peak) and by raising the Ql rod offset close to the ion source potential. The effect of these actions is to: (i) discard a large fraction (>90%) of A"*", (ii) delay the transit of A"*" through Ql (this provides sufficient time for excited electronic states of the N2^ used for Part 2 to be quenched by radiative transitions prior to their entry into Q2), and (iii) reduce the energy spread of the ions exiting Ql by cutting out low-energy ions. Unfortunately, rf-pickup within each quadrupole mass filter tends to broaden the energy distribution of the projectile ion as it traverses Ql, Q2, and Q3. Therefore, a high-energy tail is usually observed.

Q3 Rod Offset
To achieve In W=hx Y, one must have very high ion-coUection-efficiency (=:100%). Consequently, the Q3 rod offset must be biased very negatively with respect to the Q2 rod offset to draw ions out of Q2. This will result in very broad, poorly resolved mass peaks (reference [8] shows examples of how the Q3 rod offset affects peak shapes). Throughout each Part of the test protocol, you must keep the Q3 rod offset biased as negatively as necessary so that [C+]=:[A''"]o-[A"^].

Tuning q2
The rf amplitude of Ql, Q2, and Q3 is characterized here by the Mathieu parameters qi, qj, and q^, respectively (for further information, see references [2]- [5]). Some instruments have a separate rf/dc generator (quadrupole controller/power supply) for each of the three quadrupole rod assembHes (Ql, Q2, and Q3); some do not. Some can reference q^ to qi and/or to q^ (i.e., one can vary ^2/?! or q^i/q^, from 0-1; in some instruments the ratio is fixed at q^qi=0.5 or q2/q3=0.5). Some instruments reference ^2 to the ion transmission cut-off which occurs at ^2=0.908 [this latter mode of tuning is somewhat risky; the rf-pickup within Q2 can make it difficult to distinguish the true cut-off at 5'2==0.908 from the small, but significant contribution of ions in the high energy tail which are transmitted through Q2Q3 even though the nominal (apparent) value is 92>0.908].
In general, tune 92 by following the recommendations of the manufacturer of your XQQ instrument (please complete the questionnaire regarding modes of operation for your XQQ instrument). However, to ensure experimental consistency among participants, please validate your q2 values relative to qi and/or q^ (i.e., remember that whenever Ql or Q3 are operated in a well-resolved mode and are tuned to any m/z, then the mass peak position corresponds to g'i=::0.706 or g'3=::0.706 for that m/z).
One can accomplish the validation of q2 values with a simple potentiometer divider circuit if the instrument can be operated in a manual or hybrid mode. For example, in some instruments a Mass Command input dc voltage (typically 0-10 V) is used to control the rf amphtude of a quadrupole mass filter. This 0-10 V is usually supplied by a DAC which may be controlled by a computer via a programmable soft-knob. Here we use MCI, MC2, and MC3 to designate, respectively, the 0-10 V Mass Command input voltages for Ql, Q2, and Q3. Then the divider potentiometer can be used to vary 92/91 (==MC2/MC1) or g'l/g's (^MC2/MC3). Consult your manufacturer for the location of the inputs analogous to MCI, MC2, and MC3.
If the manufacturer asserts there is absolutely no possible way to vary qi within your instrument (i.e., one cannot vary q2/gi or 92/^3), please indicate how qi is controlled by the manufacturer (i.e., does the manufacturer use a fixed value of qj/qi or qi/g^^'O.S, or ?). Also, please mark the top of the first page of the submitted data with the following label: 92 Not Varied.

Acquisition of Ion Intensity Data (i) Mode of Operation
Use only a "daughter-scan" mode [i.e., 92 referenced to qi; Ql tuned to the peak position which corresponds to the maximum ion intensity at the m/z of the projectile ion; Q3 scanning over the peak which corresponds to the m/z of the product ion (use a scan window of ca. 4-10 amu) so you can see the entire mass peak (baseline-to-baseline)].
Note: To ensure consistent measurements for the duration of each experiment, make sure Ql is staying tuned to the peak position for the projectile ion intensity; this can be done by varying the Ql mass command to maximize the ion signal being viewed within the Q3 scan window.

(ii) Signal and Background Measurements
For the reasons discussed in 1 above, all ion signal intensities must be acquired manually. To do so, you must use a display which shows the complete mass spectral peak shape for a single mass peak (i.e., baselineto-baseline display of ion intensity vs m/z for a single mass peak). For this, you can use either an external oscilloscope connected to the output of your detector chain, or whatever mass spectral display your system provides on its CRT.
For each mass peak, use arbitrary units (e.g., CRT divisions) to measure the ion intensity at the peak. However, report the ion intensities in relative units by making appropriate corrections for differences in amplifier gain, scope attenuation, etc. [e.g., an ion intensity of 1.00 relative units could correspond to (a) 1 CRT division with an amplifier gain of X1 and a scope attenuation of X1, etc. or (b) 10 CRT divisions with an amplifier gain of X100 and a scope attenuation of X10].
For the measurement of background alone, the ion intensity is measured at the mass peak position in the absence of the ion of interest. For the measurement of signal+background, the ion intensity is measured at the mass peak position in the presence of the ion of interest.
Note: Please follow the following guidelines to ensure the ion intensity measurements are kinetically meaningful, and to minimize long-term variations in gas flows, detector response, etc. Here source gas refers to the gas introduced into the ion source, and target gas refers to the gas introduced into Q2.
(a) Before making any ion intensity measurements in Part 1 or Part 2, allow sufficient "warm-up" time to ensure all instrumental components, ion source, gas flows, etc. have stabilized. (b) Leave everything (source gas, ionizer filament, etc.) turned on at all times for the duration of all measurements in any one Part (the target gas is the only thing that should be turned on and off during the measurements). (c) If possible, use a high-quality toggle valve or solenoid shut-off valve to turn the target gas on/off.
(The shut-off valve should be located between the Q2 collision region and a high-precision needle valve). Thus, the target gas flow can be reset reproducibly to any given P/XQQ by presetting the needle valve. This will help to ensure high-accuracy "back-to-back" units for In F/ln H^ measurements.
(d) To make meastirements of the background intensity at any m/z, block the Projectile Ion Beam (PIB). To do so, raise to a high positive voltage the potential of any ion-optical element (e.g., the rod offset of Ql or Q2). The following abbreviations are used throughout the protocol: PIBOFF=ProjectUe Ion Beam OFF (projectile ion beam blocked; no ions reach the detector) PIBON = Projectile Ion Beam ON (projectile ion beam unblocked; ions enter Q2) TGOFF = Target Gas OFF (not flowing into Q2) TGON = Target Gas ON (flowing into Q2)

(Hi) In Y and In WMeasurements
All In Y and In W data should be collected in back-to-back units to minimize long-term variations in gas flows, detector response, etc. For example, with reference to equations (1) and (2), one back-to-back unit would consist of items (a)-(g) below. Here q^i and q^ represent, respectively, the values of qi which correspond to the maximum ion transmission through Q2Q3 for the reactant (projectile) ion A"*^ and for the C"*" product ion. Note that q^^x and q^^ must be known prior to making the In Fand In ^measurements.

(iv) Detector Conversion Gain
Use a fixed operating voltage for your ion detector (multiplier, Daly detector, etc.). The fixed voltage used for measurements within Part 1 can differ from the fixed voltage used for Part 2. But you must use the same fixed voltage for all back-to-back units within any one Part. For each Part, the fixed voltage should provide the best compromise between the S/N of A"*" and the S/N of C^.

Target Gas Purity
The purity of the Ar and SF^ target gases should be >99.99%.

P/XQQ Measurements
There are only two requirements for P/XQQ measurements: (1) that the sensor being used must provide a reproducible, single-valued response for any given target gas pressure of Ar, and (2) that you know the relative sensitivity of the sensor (i.e., its response factor) for SFe relative to Ar. For example, some instruments use a mass flowmeter to measure the flow rate of the target gas into the collision region. You could use the flowmeter response if you know its response factor for SF^ and for Ar.

High-Vacuum Materials
No plastic tubing (teflon, etc.) should be used to introduce the high-purity target gas into the collision region within Q2. The target gas should flow only through materials normally used in high-vacuum applications (bellows valves, etc.). This will ensure that the In Y and In W measurements are not conapiicated by reactions of the projectile ion with impurities inadvertently admixed with the target gas. If you do have plastic parts in contact with the target gas, and can't substitute high-vacuum materials for the duration of this test, please mark the top of the first page of the submitted data with the following label: Plastic Tubing Used.

Test Protocol
Note: Before proceeding with the test protocol, please read the Precautions above, familiarize yourself with all the steps of the test protocol, and refer to references [9], [12], and [14] as necessary. To ensure consistency among participants, all tuning procedures and measurements must be carried out in accord with the Precautions. Please be sure to observe Precaution 4. If you have any questions, please call the round-robin coordinator. Note: Record all ion intensities in relative units (Precaution 6.) in the spaces provided with each instruction. For the sensor you are using to make P/XQQ measurements, record here the type of sensor (Bayard-Alpert Ion Gauge, mass flowmeter, etc.) and its relative sensitivity (i.e., the response factor) for SFs relative to Ar. Type of sensor Response Factor (SFe/Ar) Note: For instructions 1-7 below, maintain a constant difference of ca. 40 eV between the nominal Ion Source Potential (ISP) and the Q2 Rod Offset (Q2RO); i.e., ISP-02RO=40 eV.

Parti. ='*Ar+-|-*'Ar-v"'Ar-t-«Ar+
Warning: If your instrument can only achieve a nominal collision energy which is less than 40 eV (LAB), then use your maximum achievable collision energy in lieu of 40 eV, as you did for Part 1.
30. In general, follow the manufacturer's recommendations for conversion gain measurements. However, if your detector uses a conversion dynode, please observe the following additional precautions to ensure reproducible results. Note: When making absolute current measurements with the detector turned on, use all the settings used for instructions [18][19][20][21][22][23][24][25][26][27][28][29]. When you are using the Faraday cup (for absolute ion current measurements), turn off the detector's high voltage but use all the other settings used for instructions 18-29. 31. Make sure both the N2 and the n-butylbenzene are flowing together into the ion source at all times. Use a fixed flow rate. Do 32-39 after the flows have stabilized (constant pressure in the ion source region). 32. PIBON/TGOFF: Set Ql and Q3 at the mass peak position for N2+ with q2=q?£t for m/2 28. Leave everything turned on and monitor [Na^lo untU it reaches the final (highest or lowest) stabilized ion current for [N2"*^]o (ca. 10-15 minutes). 33. PIBON/TGOFF: Set Ql and Q3 at the mass peak position for 134+ with qi^q^ for m/z 134. Immediately record the initial absolute current measured for [134"*"] with the detector turned on. 38. The values from 36 and 37 provide an estimate for the ratio CG\ys/CG^= . 39. Repeat 32-38 at least one more time to ensure your estimates for CG127, CG28, and CG127/CG28 are reproducible.

Conclusion:
40. For each XQQ instrument, please submit: (1) this completed test protocol (i.e., with the blank spaces filled in) (2) a copy of the completed Questionnaire (3) any other data requested herein. Thank you.
How many lenses are there betwen the Ql quad rod assembly and the Q2 quad rod assembly (include the Ql exit aperture and the Q2 entrance aperture in your count)? What is the nominal diameter (in cm) of the aperture of each of the LI2 interquadrupole lenses? L12a L12b L12c L12d L12e If you are using a hybrid XQQ instrument (e.g., BEQQ, etc.), indicate the diameter (in cm) of the Q2 entrance aperture . How many lenses are there between the Q2 quad rod assembly and the Q3 quad rod assembly (include the Q2 exit aperture and the Q3 entrance aperture in your count)? What is the diameter (in cm) of the aperture of each of the L23 interquadrupole lenses? L23a L23b L23c L23d L23e

Modes of Operation:
Parent-scan mode: Q3 is set to a fixed mass while Ql scans over a range of masses; one can thus assess which parent ion masses are the progenitors of a given daughter ion mass. Daughter-scan mode: Ql is set to a fixed mass while Q3 scans over a range of masses; one can thus assess which daughter ion masses are the progeny of a given parent ion mass.
For the parent-scan mode, how is the value of 92 set in your instrument? Check all that apply: It can be referenced to ^i? It can be referenced to g'3?
Other? (Explain) For the daughter-scan mode, how is the value of 92 set in your instrument? Check all that apply: It can be referenced to qi"! It can be referenced to qi"! Other? (Explain) Comments: