A. Selecting an Analytical Method

1

• A. Selecting an Analytical Method

2

• Choice of method
• define problem
• compromise between detection, accuracy,
  precision, costs.
• How many samples? One-off job?
• Speed? Equipment available?
• Skill required?

3

• Choice of method
• Must satisfy users requirements
• 1. specify determinands
• then for each determinand
• 2. concentration range of interest
• 3. max. tolerable bias s.d. of results
• 4. required frequency/times of sample collection
• 5. max. tolerable time between collection to
  receipt of result
• 6. format of report.

4

• 2. Sampling
• - representative
• 3. Preliminary sample treatment
• - sample mass/vol. measurement
  crushing/mixing/ ignition/drying/dissolution.
• 4. Separations
• - remove interferences, or mask them.

5

• 5. Measurement
• - calibration validation (CRM/blank..)
• - Sample measurement standard addition,
  direct calibration, internal standard
• 6. Assessment of results report
• - statistical tests

6
Internal standard method

• Internal standard method
• Na to be
  determined
• Add fixed amount of Li to all
  solutions
• - Calibration Graph
• Intensity Na
• Intensity Li
• 
  Conc. Na

Measure intensity ratio
?
?
?
?
Unknown sample Can find Na
7

• Internal standard method
• Test element and i. s. element affected in same
  way by fluctuations during measurement
• e.g. flame temp., nebulizer rate.. in AAS
• Method of external calibration is usual for large
  number of samples.

8

• Method of standard addition
• e.g. Sr in urine using ICP-AES
• Add different amounts
• of Sr to sample
• containing Sr.
• Value of Sr in sample is x-intercept.
• Unknown standard solns. measured under same
  conditions.
• viscosity
• surface tension
• flame effects
• matrix effects

Emission conc. Sr added ?g/L
9
Method of standard addition

• Useful when not possible to suppress physical or
  chemical interferences in sample matrix.
• -Emission must be linear with Concn. over range
  used.
• -must be zero intercept, Emission 0, when
  Concn. 0.

10
B. Performance criteria of instruments and
figures of merit

• Precision s.d. variance r.s.d c.v.
• Bias systematic error
• Sensitivity calibration and analytical
  sensitivity
• Detection LOD
• Concn. Range LOQ to LOL
• Selectivity Coeff. of selectivity

11
Definitions

• Precision
• standard deviation, s
• rsd s / mean
• CV rsd 100
• variance s2
• Concn. Bias
• population concn. mean true concn. mean

12
Definitions

• Sensitivity
• S measured signal c analyte concn.
• Sbl signal measured for blank
• m slope of straight line calibration
  sensitivity
• Most calibration curves
• S mc Sbl
• ? analytical sensitivity m standard
  deviation of signals

13
Definitions

• Limit of detection
• cm ksbl/m where k 3
• Concn. range
• LOQ limit of quantitative measurement
• 10 sbl
• LOL limit of linearity (linear response)

14
Definitions

• Selectivity
• selectivity coefficient for analyte B with
  respect to A,
• kB,A mB / mA
• gives relative response of the method to species
  B compared with A.

15
C. Environmental Analysis Intercomparison of
analytical data
16
1. What is EAGLE? 2. What is the purpose of
EAGLE? 3. What is the major reason for the
difference of analytical results for different
laboratories? 4. What is a z-score? 5. What
specific factors cause differences in results for
a) Zn b) As c) PAH d) CN? 6. What criteria
must an analytical method have? 7. What factors
need to be included in a method specification?
17
D. Data presentation Environmental indices
What is the use of providing accurate and
sophisticated analytical data to the public if
they do not understand it? How can we get around
this problem?
18
Pollutant variable (PV) physical, chemical or
biological quantity intended as a measure of
environmental pollution. We try simplify data
presentation by 1. Translating PV to
dimensionless, scaled quantities Environmental
indicator (EIc) is a single quantity derived from
one PV,used to reflect a particular environmental
attribute.
19
2. Elc can be aggregated mathematically to form
an environmental index (Elx). 3. Relating EIc,
PV or Elx to damage (dose-effect)functions plot
effect or damage vs Elx or dosage or pollutant
concentration. Frequently the shape has a
threshold, linear portion and saturation level.
EPI increase with increasing pollution EQI
decrease...Generally, WPI I 0 pure water, I100
severe pollution AQI reverse convention. To
compute the Eix,we change PV into subindices
EIc, then combine to give EIx.
20
Uses of EIx -Resource allocation
(funds/priorities of environmental
decisions) -Ranking of locations (compare
environmental conditions) Enforcing standards
(are environmental criteria being met?) -Trend
analysis (time series of changes) -Public
information (inform about environmental
conditions) -Research (reduce large datasets to
simple form).
21
HK API represents a simple measure of the air
pollution in HK, and to predict it on a daily
basis. Based on measurements from 6 pollutants
for which HKAQO have been established. RSP
PM10, TSP, CO, O3, NO2, SO2. For each one, a
subindex is calculated by linear interpolation
using a segmented scale of 0 to 500, with
breakpoints for descriptive categories, over 9
pollutant variables.
22
Breakpoints for the HK API calculation (All
concentrations in µg m-3 except for CO in mg
m-3)
23

• For example
• 24h NO2 120 µg m-3, then subindex for NO2 would
  be NO2 subindex
• 50 (120-80)/70 x 50 79
• 1hO3 590 µg m-3, then the subindex is 200
  (190/400)x100 248.

24
API is a maximum index, current index is just the
maximum value from all the 9 subindices. If the
API exceeds 100, then the critical pollutant is
the one with the highest subindex. The
descriptor categories for the API are 0-50 Good
51 100 Moderate 101 200 Unhealthy 201
Very unhealthy. Advice is given to general public
for API gt 100.
25

• The API for each monitoring station (about 13 in
  all ) is calculated from the subindices. The
  values are aggregated for similar stations, into
  categories urban, new development and industrial
  zones and roadside.

26
Comments API only constructed from a few
pollutant variables some such as SO2 and PM10
are synergistic, TSP is harmless breakpoints
are arbitrary unclear whether acute or chronic
effects are predicted.
27
Question

• Calculate the HK API for the following scenario,
  and comment upon the critical pollutant (if any)
  and the health descriptor
• (units µg m-3, except CO in mg m-3)
• TSP 24h 190 RSP 24h 110 SO2 24h 120 SO2 1h
  300 NO2 24h 170 NO2 1h 430 CO 8h 9 CO 1h 22
  O3 1h 100 benzene 24h 1 dioxin 1h 0.01 H2O2 1h
  1.

28
Compare the HK AQO with USA NAAQS

• http//www.epa.gov/air/criteria.html
• about PSI
• http//www.epa.gov/ttn/amtic/psi.html
• http//www.epa.gov/ttn/oarpg/t1/fr_notices/psiprpf
  r.pd

29
About units!!!

• 1 µg m-3 (10-6/M.Wt.) mol m-3
• (10-6/M.Wt.) (2410-3 m3) m-3
• (24/M.Wt.) ppb
• so 1 ppb (M.Wt./24) µg m-3
• Example 1 ppb O3 ? µg m-3