1. Limit of Linearity (LOL)
1.1. 5% Deviation from linearity
1.2. Linear Dynamic Range
1.2.1. range over which the calibration remains linear
1.2.2. From LOQ to LOL
1.2.3. Should be at least 2 orders of magnitude
1.2.4. Good = 5 or 6 Orders
2. Signals
2.1. Analog
2.1.1. Continuous
2.2. Digital
2.2.1. Count
2.2.2. Binary
2.2.2.1. Serial
2.2.2.2. Parallel
2.3. Time
2.3.1. Frequency
2.3.2. Pulse Duration
2.3.3. Time or Average rate of Pulses
3. Calibration Methods
3.1. External Calibration
3.1.1. Relate Concentration of Known Standards to Response
3.1.2. Plot response vs Concentration
3.1.3. Linear Regression
3.1.3.1. Use to determine concentration of unknown
3.1.3.2. Uncertainty
3.1.3.2.1. Uncertainty about the Regression (Sr)
3.1.3.2.2. Uncertainty in Concentration (Sc)
3.1.4. When to Use?
3.1.4.1. Calibrate Instruments
3.1.4.2. No Matrix Effects
3.1.5. Use Blank
3.1.5.1. Real Blank
3.1.5.1.1. Solvent Blank
3.1.5.1.2. Reagent Blank
3.1.5.2. Ideal Blank
3.2. Standard Addition
3.2.1. When to Use?
3.2.1.1. Matrix cannot be duplicated
3.2.2. Add small aliquot of a high concentration standard to the unknown
3.2.3. Plot Response vs Concentration Added
3.2.4. Uncertainty
3.2.4.1. Uncertainty in Concentration (Sx)
3.3. Internal Standards
3.3.1. Add substance in constant amount to all samples, blanks, and calibration standards
3.3.2. When to use?
3.3.2.1. Sample volume is not constant
3.3.2.2. Instrument Response varies
3.3.3. Plot ratio of analyte signal to internal-standard signal vs Concentration
3.3.4. Compensates for Random and Systematic errors
4. Figures of Merit
4.1. Precision
4.1.1. Absolute Standard Deviation
4.1.2. Relative Standard Deviation
4.1.3. Coefficient of Variation
4.1.4. Variance
4.1.5. Standard error of the mean
4.2. Accuracy
4.3. Signal to Noise Ratio
4.3.1. Analytical Signal
4.3.1.1. Instrument Response to analyte
4.3.1.2. Baseline to top of peak
4.3.2. Noise
4.3.2.1. Random Fluctuation in baseline
4.3.2.2. peak to peak
4.3.2.3. RMS
4.3.2.3.1. RMS = 5 p-p
4.3.2.4. 68-95-99.7 Rule
4.3.3. S/N = 1/RSD
4.3.4. Proportional to the square root of the # of scans
4.3.5. Want to be large
4.4. Limit of Detection (LOD)
4.4.1. Smallest concentration of analyte that can distinguished from the blank
4.4.2. 10-20 measurements of blank
4.4.3. yL=ybl+3Sbl
4.4.3.1. LOD=3Sbl/m
4.5. Sensitivity
4.5.1. slope
4.5.2. analytical sensitivity
4.5.2.1. y=m/Ss
4.5.2.1.1. Ss= Standard Deviation of the measurement
4.6. Limit of Quantification (LOQ)
4.6.1. the level above which quantitative results can be obtained
4.6.2. yQ=ybl+10Sbl
4.6.2.1. LOQ=10Sbl/m
4.7. Selectivity
4.7.1. the ability of a technique to respond specifically to an analyte
4.7.2. influenced by matrix
4.7.3. K=mb/ma
5. Selecting an Analytical Method
5.1. Speed
5.2. Ease & Convenience
5.3. Cost
5.4. detection limit
5.5. size of sample
6. Noise
6.1. Chemical
6.1.1. Effect of temp or pressure on chemical equilibrium
6.1.2. Change in moisture content
6.1.3. Lab fumes condensing or reacting
6.1.4. Photosensitive materials
6.2. Thermal
6.2.1. "Johnson Noise"
6.2.2. Agitation of electrons (Charge carriers)
6.2.3. Independent of Frequency
6.2.4. Dependent on temperature, resistance, and bandwidth
6.3. Shot
6.3.1. "Quantum", "Poisson", "Schottky","Dark Current"
6.3.2. Type of White Noise
6.3.3. Because charge and energy are quantized
6.4. Flicker
6.4.1. "Pink Noise"
6.4.2. Significant at frequencies lower than 100Hz
6.4.3. Proportional to signal and inversely dependent on absolute frequency
6.4.4. 1/f
6.5. Environmental
6.5.1. Reduction
6.5.1.1. shield circuits and wires
6.5.1.2. good grounding
6.5.1.3. Reduce vibrations
6.5.1.4. Insulate
6.5.1.5. Reduce radiant energy
6.5.2. Interferes with analytical measurements
6.6. Noise Reduction
6.6.1. Grounding/Shielding
6.6.1.1. Surround circuits with a conducting material
6.6.1.2. Absorb electromagnetic radiation by the shield
6.6.1.3. Trial and error for optimal configuration
6.6.2. Filtering
6.6.2.1. Noise at high frequencies, signals at low frequencies
6.6.2.2. Take the output voltage across the capacitor
6.6.2.3. Lets low frequencies pass
6.6.3. Ensemble Average
6.6.3.1. Noise is randomly distributed but signal is not
6.6.3.2. Average successive sets of data
6.6.3.3. Noise smooths out
6.6.4. Boxcar Averaging