ToF-AMS DAQ. Joel Kimmel Univ. of Colorado at Boulder & Aerodyne Research, Inc AMS Users Meeting 2007

Transcription

1 ToF-AMS DAQ Joel Kimmel Univ. of Colorado at Boulder & Aerodyne Research, Inc AMS Users Meeting

2 Web Resources Downloads Release Notes Supplemental Files Directions for New Features Questions Requests Problems / Bugs Reporting Problems jkimmel@colorado.edu Copies of Menu Files PowerPoint File with Screenshots Text from Error Message (If Bug) 2

3 Outline Function and Structure of Code Configuring Acquisition Parameters Structure of Data Files V2.1 Tuning Window Fundamental of TOFMS calibration and tuning BitWise Fundamentals of ion detection Looking Ahead to V2.2 Diagnostics ToF-AMS DAQ Data Save Acquisition Timing & Triggering Sequence of ToF or Ion Types (Menu Switching) Data Averaging (AP240 + RAM) Sequence of Acquisition Modes (GenAlt) 3

4 Menu parameter files drive the DAQ. (You define the menu file) Menu Parameter Files Diagnostics ToF-AMS DAQ Data Save Acquisition Timing & Triggering Sequence of ToF or Ion Types (Menu Switching) Data Averaging (AP240 + RAM) Sequence of Acquisition Modes (GenAlt) 4

5 Menu Parameter Files MENU PARAMETER FILES save all adjustable parameters relevant to the operation of the instrument, e.g., timing and saving controls Menu Window Text Editor VIEW/UPDATE Menu Values through DAQ or Text Editor Menu Parameter Files SAVE VALUES in archive, to operating files, and with data Archive At Load of Software Auto-Save Changes During Use of Software Meta-Data Values for Run included in all AMS data files (ParVal) To design an experiment is to configure runs. 5

6 The Run is Our Base Unit Run Unique ID Number Active Menu File A RUN is a user defined duration of data averaging, having operating parameters determined by the ACTIVE MENU FILE Hardware Configuration Acquisition Timing and Triggers Time to Average (Run Duration) / Synchronization with Clock Sequence of Acquisition Modes (GenAlt) Structure for Saved Data GenAlt allows flexibility within a single run. 6

7 Sequence of Acquisition Modes (GenAlt) When configuring a run, the user defines ACTIVE ACQUISITION MODES Acquire Start New Run, t = 0 During a single run acquisition CYCLES THROUGH ACTIVE MODES PTOF While Acquisition in a mode lasts for a user-defined DWELL TIME BFSP MS t < T save Between modes, software checks whether total acquisition time has reached the user-defined SAVE TIME (T Save ) Save AMS Data Stop Start New Run, t = 0 PTOF 10 s BFSP 5 s MS 10 s While t < 50s (2 cycles) Save Run EXAMPLE GenAlt Acquisition may run with ANY COMBINATION of PToF, MS, and BFSP During the run, acquisition CYCLES through active modes, spend the user-defined DWELL TIME in each The cycle continues until the acquisition time exceeds SAVING TIME 7

8 MS 10 s Start Avg Data on AP240 for Time Spent on Board (2 sec) Move Chopper If t > T DWELL (10 sec) Stop Transfer Data AP240 to RAM; Average with existing Data (Open or Closed) CLOSER LOOK AT MS DWELL User-defined Time Spent on Board Transfer and Chopper Move = DEAD TIME For MAX DUTY CYCLE, keep ON BOARD high (but less than dwell) Number of choppers determines the on board averaging time for PToF mode. This value is an estimate based on input chopper frequency. 8

9 Magnitudes for proper settings: (On Board Time) =< (DWELL) AND S(dwell times) <= (Saving Time) Menu switching allows flexibility in a sequence of runs. 9

10 Sequence of Unique Runs TYPICAL OPERATION one collects a continuous sequence of runs with identical operating parameters (i.e., a fixed menu file). Run 1 Run 2 Run N M1 MENU SWITCHING allows the user to define a cyclic sequence of menu files to use. Allows continuous experiment, with alternating configuration Most commonly applied to V/W SWITCHING with HR-ToFMS Run 4n - M4 Run 1n - M1 Run 3n - M3 Run 2n - M2 GenAlt + Menu Switching = flexible experiment design. 10

11 (i) V (PM) (ii) V (MP) 5-min W (M) V (PM) W (M) V (M) 30 sec V (M) 30 sec W (M) V (M) 30 sec EXAMPLE DESIGNS (i) 2 menus; Vary ToF type; Different active modes for each ToF type (ii) 3 menus; Vary ToF Type; 2 different V-mode configurations SEQUENCE is DEFINED BY Active Menu Numbers Start Menu Successive Runs at each menu number Stop Acquire Run Number +1 Select Menu Save Active Menu File Update Hardware and Timers Save AMS Data Acquire & Average AMS Data PTOF BFSP MS 11

12 V (PM) W (M) V (PM) Menu Summary W (M) ToF-AMS data are written as HDF files. One file contains all data from multiple runs. 12

13 HDF File Format A single HDF FILE can contain multiple DATASETS, each of unique size and dimension (equivalent of an Igor wave) Dataset 1 Dataset 2 Dataset 3 Dataset 4 All datasets have an EXTENDIBLE DIMENSION An existing dataset is written to by adding a LAYER along the extendible dimension Layer 4 Layer 3 Layer 2 Layer _001000_m.hdf Run 1003 Run 1002 Run 1001 Run 1000 ToF-AMS DAQ writes to TWO TYPES of files The MAIN (_m) file contains MS data and PTOF-stick data. The PTOF (_p) file contains only raw PTOF data. Data from each RUN ARE WRITTEN AS LAYERS of the datasets. Both file types include parameter and info values, and a RUNINFO dataset that SQUIRREL uses for indexing of the data. 13

14 Run as Layer of HDF Datasets Run The datasets written for a run depend on active modes and save preferences. Data are sorted as V or W; Menu Number is NOT part of the save structure RunInfo guides SQUIRREL in indexing of the layers. RunInfo ParVal ComPar Val InfoVal MSSClosed_v MSSDiff_v If MS Active If PTOF Active If PTOF Active AND Save RAW If V If W If V If W MSSClosed_w MSSDiff_w ptof_stick_v ptof_stick_w If V ptof_v ptof_w If W RunInfo ParVal ComPar Val MSClosed_v MSClosed_w InfoVal MSOpen_v MSOpen_w Estimating Data Size Data values are singles (4 Bytes) MS: Array of length = number of samples MSS: Array of length = MaxMass MinMass PTOF: Matrix of dimension = (ToFperChop/Co-Adds) x (number of sample) PTOF Stick: Matrix of dimension = (ToFperChop/Co-Adds) x (MaxMass - MinMass) Example. V-mode Run. 30,000 samples, m/z = 10 to 410 Th 100 Tof/chop, 2 co-adds MS = 30,000 * 4 = 120 KB MSS = 400 * 4 = 1.6 KB PTOF Stick = 50 * 400 * 4 = 80 KB PTOF = 50 * 30,000 * 4 = 6 MB _m = 2*120KB + 2*1.6KB + 80KB = 323 KB/Run _p = 6 MB/Run 14

15 V2.1 Implementation of Common Menu Structure Expanded m/z Calibration and Tuning Window BitWise replaced Threshold Analysis and m/z Ratio Windows Go to DAQ Calibration and Tuning 15

16 Time-of-Flight Mass Spectrometry To determine m/z values A packet of ions is accelerated by a known potential and the flight times of the ions are measured over a known distance. Data are collected in time (ns). User assigns m/z values to known peaks in time spectrum. Mass Calibration based on LSF allows determination of m/z from time of flight. 16

17 TOFMS Source, S Drift Region, D a Detector V E = V/S E = 0 t D U = qv = qes qesa = mv 2 D a a 1 2 D 2qEsa vd = m D D = = v 2qEs a m Ideally U is known exactly, T is exactly predictable. IN REALITY: Initial U And position dependent V a and Causes broadening of ToF peak. Reflectron Series of electrodes, forming a linear field in direction opposite of initial acceleration. Penetration depth depends on U s, which is function of U 0 and acceleration field, E. From: Reflectron voltages are tuned to create a SPACE FOCUS at the plane of the detector. 17

18 For ToF-AMS Ion source voltages are tuned to maximize transmission into TOFMS extraction region Reflectron is tuned to compensated for?e of accelerated ion beam. Go to window BitWise 18

19 Ion Detection Signals in AVERAGED TOF mass spectra can be divided into two classes: (i) LOW ABUNDANCE, which have shape and intensity originating from the accumulation of stochastic single ion detection events (ii) HIGH ABUNDANCE, which have shape and intensity originating from the accumulation of signals generated by the simultaneous detection of multiple ions. Time-to-Digital Converter (TDC) Converts a signal of sporadic pulses into a digital representation of their time indices. Neither the height nor the area of the pulse is recorded. D A TDC usually follows a discriminator, which sets the minimum accepted pulse amplitude. ToF-MS experiments measuring ONLY LOW ABUNDANCE IONS can use TDC for ION COUNTING Peaks heights in MS develop through averaging Time NOT LINEAR IN HIGH ABUNDANCE REGIME (Saturation) 19

20 Analog-to-Digital Converter (ADC) Converts continuous signals to discrete digital numbers. Records time and amplitude of waveform events. Acqiris AP240 = 8-bit, up to 2 GHz Demo from Learning by Simulation. By Hans Lohninger ( Particle Detection bursts necessitate ADC. Quantification relies on (1) The detector having a response that is proportional to the number of incident ions (2) The ADC recording the signal waveform with fidelity (3) Knowledge of the relative detector response in order to calculate the ion count rate. Measure response of ToF-MS detector for single ion and use this value as the base unit for quantification Total integrated signal at m/z 30 and 46 for single ammonium nitrate particle detection events. Total recorded signal grows as cube of diameter, indicating that recorded intensity is proportional to the number of ions. 20

21 To determine the SI area in the ToF-AMS we record the average shape of LOW ABUNDANCE m/z PEAK DETECTION EVENTS 0.10 implies that 10% of recorded events will be double ion signals. Peak height and Areas have broad distributions Calculate Average Area as sum of average element values Magnitude of response depends on MCP gain 21

22 Use of the calculated SI value for quantification of AMS data requires that it is not m/zdependent. Kimmel et al., 2007, In preparation for JASMS Noise Noise in the ToF-AMS has ELECTRONIC and CHEMICAL components. Chemical originates from scattered ions, and thus has an intensity equivalent to real, low-intensity signals (i.e., single ion arrival events). 22

23 Threshold m/z 32 AP240 includes noise rejection thresholding Any portions of the signal waveform with amplitude greater than a user-defined threshold are recorded relative to a real and exact baseline. All portions with amplitude below user defined amplitude discarded (i.e., recorded as 0 amplitude). Choosing the ADC threshold setting is a balance of SNR and dynamic range. The threshold is ideally set at a value where electronic noise is rejected and all ion signal intensity is recorded. Signal (Bits) Before determining best threshold, must determine BASELINE In voltage space, ToF-AMS signals are negative going. Setting baseline that is MORE POSITIVE than actual INFLATES signals Setting baseline that is MORE NEGATIVE than actual DEFLATES signals 23

24 BASELINE is average value of recorded signal with no detector gain With baseline set, we can then determine the distribution of intensities of noise signals FIRST GUESS (ideal) THRESHOLD will be value discards all electronic noise Affect of threshold on single ions is determined by applying a software threshold to all events used to calculate average shape AGAIN, setting threshold is a balance of eliminating noise and maintaining AREA OF LOW ABUNDANCE signals. 24

25 To observe effect of threshold on average data, we monitor the relative intensities of HIGH and LOW ABUNDANCE ions The total area of 40 decreases, while m/z 28 is unaffected until high voltages Baseline setting is AS CRITICAL. Inflation and deflation have clear effect on LOW ABUNDANCE signals. Effect of baseline setting on 40/28. Black = Area measured with threshold applied. Grey = Raw Area. From: Kimmel et al, 2007, In preparation for JASMS 25

26 Use 40/28 to determine degree of degradation at Ideal Threshold. Balance low abundance signals with electronic noise If not possible, consider raising threshold. (Avoid saturation more in a couple slides) 1950V 2050V 2150V 2250V 2350V 26

27 ADC saturation can be detected by comparing the signal of ambient O 2+ (m/z 32) to the signal of ambient N 2+ (m/z 28) in unthresholded ToF- aerosol mass spectra. The N 2 + signal will saturate at far lower MCP gain than O 2+. With BitWise, saturation can be identified with the peak analysis routines by watching m/z 28 with the chopper open. 1. Set Baseline 2. Measure electronic noise to determine ideal threshold 3. Measure peak shapes to determine saturation probability and SI Area 4. Use m/z ratio to determine if ideal threshold is detrimental to small peaks 27

28 Peak-to-Noise Ratio compares the probability signal events and noise events within the m/z integration area at the user-defined peak threshold. A low value suggests that SI area calculation will include noise events low bias. Want to run at lowest Measured Areas settle after 300 PEAKS 28

29 V2.2 Within the first half of 2008 Emphasis: More Accurate Run Timing LS triggered SP mode More tuning functionality Burst Modes (Up to 2 Hz PTOF and MS) Suggestions welcome!! 29