Title: PROCESS ANALYTICAL TECHNOLOGIES:
1PROCESS ANALYTICAL TECHNOLOGIES Chemical
Separations and Related Techniques Robert E.
Synovec Center for Process Analytical Chemistry
(CPAC) Department of Chemistry, Box
351700 University of Washington, Seattle, WA
98195, USA synovec_at_chem.washington.edu http//syn
oveclab.chem.washington.edu 206-685-2328 (Present
ed by M. Koch to FDA trainees Dec. 2002)
2Abstract This short course will overview current
separation methods applicable to PAT, and lead
into a discussion of emerging hyphenated
technologies coupled with chemometric data
analysis algorithms. The technologies will
include liquid chromatography, gas
chromatography, capillary electrophoresis, flow
injection analysis, and comprehensive
two-dimensional gas chromatography (GC x GC).
Sampling issues and recent advances in NeSSI will
be put into context along with recent advances in
micro-scale separation-based chemical analyzers.
Traditional data analysis methods will be put
into context with emerging chemometric data
analysis approaches. Chemometric approaches for
the multidimensional data analysis will include
the generalized rank annihilation method (GRAM),
three-way partial least squares (Tri-PLS) and
pattern recognition methods such as principal
component analysis (PCA).
3 OUTLINE PART ONE This
didactic session Basic Principles, Definitions
and Examples Chromatography Capillary
Electrophoresis Flow Injection Analysis Data
Analysis, a Historical Perspective Microfluidics
Sampling (NeSSI) MicroScale Chemical
Analyzers
4 OUTLINE PART TWO
Practicum session Data Analysis in
Chromatography Traditional Approach Chemometr
ic Approach Toward high-throughput analysis of
complex mixtures Hyphenated instrumentation
coupled with chemometrics 2D Separations
Comprehensive Two-Dimensional Gas
Chromatography (GC x GC) Instrumentation Chemo
metric Application of GRAM to GC x GC Data
5CHROMATOGRAPHY DEFINITIONS AND
PRINCIPLES Chromatography is a chemical analysis
method based upon the physical separation of the
chemical species in a mixture (i.e., analytes
and interferences). Separation occurs in a
column filled with a stationary phase working
with a mobile phase. Differential migration, and
thus separation, occurs due to the relative time
a given species spends in the stationary phase
relative to the mobile phase, governed by a
distribution constant, KD. Thus, separated
species have different distribution constants.
The retention time, tR, is the total time a
given species spends in the mobile and
stationary phases. The separated species are
detected by a suitable device and the resulting
data is called a chromatogram. Information is
obtained from chromatograms.
6(No Transcript)
7Chromatography Method Summary HPLC High
performance liquid chromatography. Utilizes a
liquid mobile phase and various stationary
phases depending upon the separation mechanism
Partitioning, Adsorption/Desorption, Chiral,
Hydrophobic Interaction, Size Exclusion, Ion
Exchange, etc. GC Gas Chromatography. A
carrier gas is the mobile phase. The
stationary phase is generally a thin polymer.
Separation is based upon analyte volatility
(boiling point) and analyte interaction
with the stationary phase. SFC
Supercritical Fluid Chromatography. The mobile
phase is generally supercritical carbon
dioxide possibly with a polar additive
such as methanol. The stationary phase
is similar to those used in GC.
8- Chromatography Detectors
- HPLC and SFC
- Absorbance
- Fluorescence
- Electrochemical and Conductivity
- Mass Spectrometry (MS)
- Refractive Index (RI)
- Polarimetry (optical activity)
- GC
- Flame Ionization (FID)
- Electron Capture (ECD)
- Thermal Conductivity (TCD)
- Mass Spectrometry (MS)
9Chromatography Equations - - - - - Relating
Thermodynamics to Separation (1) Distribution
Constant analyteStationary Phase KD
--------------------------------- analyteMobi
le Phase where species molar
concentration (2) Retention Time tR
tStationary Phase tMobile Phase tSP
t0 ..where t0 dead
time (3) Retention Factor (aka, capacity
factor) k (tR - t0) / t0 KD
(VolumeSP / VolumeMP)
10(No Transcript)
11(No Transcript)
12(No Transcript)
13Information in Chromatograms Analyte
identification is provided by the retention time
tR, i.e., the thermodynamic basis for selectivity
in separation. A unique KD provides a unique
tR. Match the tR of anunknown peak in a
sample chromatogram with known peak in a
standard chromatogram. Also, it may be coupled
with the Standard addition method.
Identification is strengthened by using a
detector that provides additional analyte
selectivity (eg., unique absorbance spectrum,
mass spectrum, or optical activity). Analyte
quantification is provided by the peak height or
area, which are proportional to the injected
analyte concentration (the quantity of interest
in many applications). Alternatively, complex
chromatograms may provide information using
pattern recognition approaches.
14(No Transcript)
15(No Transcript)
16Anion exchange separation (DIONEX), using NaOH
gradient with post-column micro-membrane
suppression prior to conductivity detection
17Cation exchange separation (DIONEX) using
counter- ligand side-equilibria in mobile phase
for separation selectivity followed by
post-column derivatization with PAR for sensitive
detection
Time, minutes
18Capillary Electrophoresis (CE)Separation method
based upon differential migration in an electric
field
- Separation is performed in
- narrow-bore (25 75 ?m id),
- fused silica capillary
- Electroosmotic flow (EOF) is
- superimposed on electrophoretic
- flow and carries all analytes to the
- cathode
- Separation is based on q/r,
- the analyte charge-to-size ratio
- Well suited for biological samples
19 CE Instrumentation
Separation Capillary
Cathode
Anode
Detector
Power Supply
Electrolyte
Electrolyte
HV
HV 30,000 V / meter capillary
20Capillary Electrophoresis for Standard Proteins
(3 runs offset, to show reproducibility)
1. ?-Lactalbumin 2. ?-Lactoglobulin 3. Trypsin
inhibitor 4. Carbonic anhydrase 5. Ovalbumin 6.
Bovine serum albumin 7. Conalbumin 8.
?-Galactosidase MW 14400 -116000 Separation
buffer 0.1MTris0.1MCHES 0.1SDS8pullulan
7
6
1
2
5
4
3
8
40
35
30
25
20
15
10
Time (min)
21FLOW INJECTION ANALYSIS
22FLOW INJECTION ANALYSIS (FIA) Automated
Sample Handling Coupled with Highly Selective
Reagent Chemistry
Evolving Technology
23(No Transcript)
24Sequential Injection System
25Lab-on-Valve (LOV) small volume sequential
injection
26µSI-LOV for Process Control
- Monitoring Ammonium for Fermentation
Micro sequential injection fermentation
monitoring of ammonia, glycerol, glucose, and
free iron using the novel lab-on-valve system,
Chao-Hsiang Wu, Louis Scampavia, Jaromir
Ruzicka, Bruce Zamost, Analyst, 2001, 126(3),
291-297.
27Analytes Determined by LOV
- Environmental Monitoringb
- Nitrate 3.91 ppb(N)
- Nitrite 4.53 ppb(N)
- Phosphate 0.10 ppb(P)
- Real-time, on-line fermentation monitoring
- Nitrate
- Nitrite
- Phosphate
- Ammonium
- Glucose
- Lactate
- Glycerol
- Free Iron
bMicro Sequential Injection Environmental
Monitoring of Nitrogen and Phosphate in Water
Using Lab-on-Valve System Furnished with a
Microcolumn, Chao-Hsiang Wu, and Jaromir
Ruzicka, Analyst, 2001, 126, 1947-1952.
Page 13
CPAC Spring 2002 Meeting
28The µSI-LOV-Capillary Electrophoresis System
CE as an Advanced Detector for µSI-LOV System
29The µSI-LOV-Capillary Electrophoresis System
Micro sequential injection anion separations
using Lab-on-Valve coupled with capillary
electrophoresis Chao-Hsiang Wu, Louis
Scampavia, Jaromir Ruzicka, Analyst, 2002, 127,
898-905.
30(No Transcript)
31(No Transcript)
32(No Transcript)
33Peak Height Analyte 1
34(No Transcript)
35Is the common practice of chromatography always
in our best interest? Here is another view The
common practice of chromatography is to achieve
complete separation between analytes and
interferences, at the cost of analysis time and,
in some cases, applicability. This practice
evolved from the belief that more or complete
chromatographic resolution, Rs, always translates
into a better chemical analysis. While this
belief has served us well, is it always true,
and how may have it limited us? How much
resolution is really necessary to solve the
problem at hand?
36What should be our Goal? Answer Optimize the
Information per Time gleaned from the chemical
system of interest that is, solve chemical
analysis problems efficiently. Achieving this
GOAL is likely to require a change in how we
think about applying chromatography, specifically
working with analytical methods that do not
always rely upon complete resolution of
components of interest.
37 We shall consider achieving this GOAL in
terms of two analytical problem scenarios,
specifically, the need to... rapidly analyze
more samples in high throughput applications
...and the need to address emerging
applications of more complex samples.
38These two types of analytical problems can be
addressed by instrumentation development and data
analysis strategies in the following ways (1)
Strategies to rapidly analyze more samples in
high throughput applications can be optimized by
considering the consequence of keeping sample
complexity constant while substantially reducing
analysis time. (2) Strategies for emerging
applications of more complex samples can be
optimized by considering the consequence of
having sample complexity increase while keeping
analysis time constant.
39(No Transcript)
40Example Calculation The relative difference
in run time on going from a Rs of 1.5 to a Rs of
0.25 is time 1.5 1.5
2 --------- --------
62 36 time 0.25
0.25 2 Therefore, a 36-fold potential
savings in analysis time can be achieved if the
information can be extracted at Rs 0.25 versus
at Rs 1.5. The savings in run time for this
example is going from 20 min to 1/2 min. Can
this potential benefit be achieved and what are
the issues involved in making the
instrumentation provide reliable data? How best
can the less-resolved data be analyzed to extract
useful information?
41Lets take a lesson from spectroscopy Chemometri
cs (CLS and PCR) is readily and successfully
applied in the analysis of individual xylenes in
mixtures, even though the data is poorly
resolved. Here, resolution in chromatography
terms for the three xylene peaks is very low.
However, quantitative precision (RSD) was 2.
Chemometrics should also be more heavily used for
chromatography!!
Wavelength (nm)
NIR specta. C. L. Erickson, M. J. Lysaght and J.
B. Callis, Analytical Chemistry, 1992, 64, 1155A
- 1163A.
42Key Point The human eye sees only peak
resolution while a computer-based data analysis
algorithm (chemometrics) sees both peak
resolution and precision along the measurement
axis (either wavelength for spectroscopy or
retention time for chromatography).
43(No Transcript)
44Zero Order Instruments The earliest analytical
instruments were zero order. They produce a
single data per sample measurement. Examples pH
electrode, temperature sensor, absorbance at one
wavelength, etc. Key point There is no way to
detect errors due to the presence of interferents
with a zero order instrument. It is not even
possible to tell if there is an interferent
present.
45First Order Instruments First order instruments
produce a vector of data for each sample run.
Often referred to as a 1-D method. Examples Mult
i-wavelength spectroscopy (IR, MS,UV-vis
absorbance, etc), chromatography (LC, GC, SFC),
capillary electrophoresis, sensor arrays,
etc. Key point The presence of interferents can
detected with first order instruments but can
only be corrected if the interferents are known,
and therefore part of the calibration procedure.
46Second Order Instruments Second order
instruments (2-D) produce a matrix of data for
each sample run. These are hyphenated
instruments obtained by combining two first order
instruments. Examples GC/MS,
HPLC/multi-wavelength UV-vis absorbance, GC x GC,
LC x CE, etc. Key point The presence of UNKOWN
interferents can detected and mathematically
eliminated using second order CHEMOMETRIC
methods. Thus, interferents do not have to be
part of the calibration standard. Also, some
forms of instrumental run-to-run drift can be
corrected.
47(No Transcript)
48(No Transcript)
49- Chemometric Methods for Hyphenated Separations
- Generalized Rank Annihilation
Method(GRAM) with objective retention time
alignment - Analysis of Variance Feature Selection followed
by Principal Component Analysis (ANOVA - PCA) - Trilinear Partial Least Squares
(tri-PLS) - These three chemometric methodologies have
been adapted and applied with GC x GC, etc.
50(No Transcript)
51MicroFluidics the ability to move, mix, pump and
otherwise control liquids on a microscopic scale
Potential advantages Smaller more compact
systems More amenable to integration Potentially
lower overall cost New approaches to analysis
Challenges Sample introduction Pumping Material
s of construction New approaches to analysis
52Fully Integrated Positive-PressureDriven
Microfabricated Liquid Analyzer
The basic idea is to incorporate on-chip near
real-time sampling, capillary separation, and
multivariate detection in a single integrated
microfabricated device. Soft lithography is
used for rapid prototyping for proof of
principle design and testing. References (1)
Initial analyzer development - P. G. Vahey, S.H.
Park, B. J. Marquardt, Y. Xia, L. W. Burgess and
R.E. Synovec, Talanta, 2000, 51, 1205 - 1212.
(2) Incorporation of GLRS detection - P.G.Vahey,
S. A. Smith, C. D. Costin, Y. Xia, A. Brodsky,
L.W. Burgess and R. E. Synovec, Analytical
Chemistry, 2002, 74, 177 - 184.
53(No Transcript)
54(No Transcript)
55Microfluidic Chip and Instrument Setup
56(No Transcript)
57(No Transcript)
58(No Transcript)
59Microfluidic Analysis Chip Network
60?-RIG Detection Mechanism
61Protein detection without derivatization. Well
suited to micro-reactor monitoring.
- 100 ?L off chip sample injection
- Concentration injected
- BSA 15 ?M
- RNAse 73 ?M
62Using the RIG System as a Micro-scale Molecular
Weight Sensor (?-MWS)
63?-MWS Microfluidic Network
64?-MWS data collected for Linear PEGs at the
Upstream and Downstream Positions
Upstream Signal Area
Downstream Signal Area
65Molecular Weight Calibration
- Ratio of downstream to upstream area signals
- Ratio is independent of Concentration
- MWS tuned to MW range 1,000 20,000
- Flow rates and probe positions can be tuned to
obtain desired MW range
66(No Transcript)
67ADVANCES IN SENSORS AND CONTROLSHIGHLIGHT THE
NEED FOR IMPROVED SAMPLING AND DATA HANDLING
TECHNOLOGIES
68SAMPLING
? Systems are unique But they could be
Modularized ? Smart Systems are Needed
69Todays Typical Technology - Sasquatch Size
Courtesy of a major petrochemical facility
70NeSSI(New Sampling/Sensor Initiative)
ExxonMobil, Dow, CPAC
71- Facilitate the acceptance/implementation of...
- modular, miniature smart sample system
technology - based on ISA SP76 standardization work
- Promote the concept of...
- field-mounted smart sampling/analytical systems
- integration of sample systems with
physical/chemical sensors - Lay the groundwork for...
- open connectivity architecture for sensor
communications (e.g. Ethernet, DeviceNet, etc.) - industry standard communication protocols (e.g.
OPC, DeviceNet, CAN, etc.) - web enabled technologies (e.g. Browser based HMI)
- Provide a technology bridge to the process for...
- sensor/lab-on-a-chip micro-analytical devices
72The drivers for change...
- Reduce Cost to Build
- Design and Engineering
- Manufacturing and Installation
- Reduce Cost of Ownership
- Design Standardization
- Simpler, less expensive maintenance
- Enhance Reliability
- Validate sample, data, communication
73Whats going on in the world?
Market Acceptance Lower Cost Ease of
Design System Reliability
- Technology Explosion!
- Fast networks robust electronics
- Lab-on-a-chip lasers wireless
- Open software platforms
- What does the Market Need?
- Reliable, rugged, low cost fast
- What are the Critical Issues?
- Adoption of standard footprint
- Component Availability
NeSSI
Analyzer House
On the Post
1938
2000
Time
74What is NeSSI?
Picture courtesy Swagelok
75(No Transcript)
76Petrochemical installations
Coutesy of a gulf coast petrochemical facility
77Rosemount-Swagelok
- Field mounted gas chromatograph with a
miniature/modular sample system - Example of synergy that is driving NeSSI
Courtesy of Rosemount Swagelok
78NOW, THE ANALYZER IS LARGE COMPARED TO THE
SAMPLING DEVISE
- WHY CANT THE ANALYZER BE LOCATED ON THE SAMPLING
PLATFORM??
79NeSSI with Analyzer on board and Wireless
SENSOR MANAGER
ANALYZER
SensorManager(I/OController)
80NeSSI has become an Enabling technology for...
- Miniature physical sensors
- Miniature chemical composition sensors
Rosemount Analytical
Courtesy of Panametrics Swagelok
Porter Instruments and the Swagelok Co.
Courtesy of Applied Analytics
81Miniature Gas Chromatography (GC)
Preconcentrator/ Thermal Desorber
4 SAW Array
One Meter GC Column
82A fully autonomous hand-held microanalyzer
Containing Multiple coated/packed GC columns
w/multiple coated SAW detectors
and Multiple LC columns with
fluorescence electrochemical detection
83Vision The End of the Analyzer House System
84NeSSI with an Array of Analytical Techniques will
Impact
- Process Control
- Process Optimization
- Analytical Method Development
- Product Development
85(No Transcript)
86Parameter Detection with Fringing Electric Field
Dielectrometry Sensors
- Prof. Alexander Mamishev
- Electrical Engineering
- UW
87Multiple Resolution Levels
- Electric field lines extent into space beyond the
distributed sensor - Through varying excitation patterns, multiple
levels of proximity are possible - In-plane resolution is proportional to proximity
sensing depth
88Sensing Possibilities
- Fringing electric fields can detect various
characteristics of a sample.
89Chemometric Challenge
- Most samples vary in multiple ways all of which
affect the output signal.
90Examples Cookie Samples Paper Samples Salted
Fish Samples
- Changes in Texture and Density
- Non-homogeneous Material Distribution
- (eg. Chocolate Bits, Oils, Salt, Fish Sections
- Temperature Distribution
- Moisture Distribution
- Contact Specifics
91Current Research Approach
- Our approach is to collect data from a complex,
unknown system and expect chemometrics to make
sense of it.
92Anticipated Future Approach
- Observe the response to individual unknowns by
analyzing well understood materials with
reproducible variations - Try to determine the spectral content based on
known individual responses
93(No Transcript)
94Development and Application of On-Line Particle
Analyzers
- Richard Gustafson, Jeff Mathews
- Paper Science and Engineering
- Brian Mayers, James Callis, Younan Xia
- Department of Chemistry
- University of Washington
95Objective
- To develop methods for on-line particle analysis
- Geometry
- Chemistry
- Mechanical Properties
- Examples
- Surface charge of pulp fibers
- Nanocrystal growth
96Approach to Method Development
- Apply flow cytometry technology to develop
general particle analysis method - Identify fluorescence stains that are specific
for property of interest - Develop imaging flow cytometer that can be used
on a broad range of samples (e.g. pulp fibers)
97Single Fiber Properties
Charge Kappa Length Kink Curl
Flexibility Cell
Wall Thickness
Fiber Performance and Pulp Behavior
98Fiber Surface Charge
- Surface charge originates from the disassociation
of carboxyl groups on cellulose - Charge is affected by wood species, recycle
content, bleaching mechanisms, water source - Charge matters
- Retention of fines, chemical additives, fillers,
dyes, starch - Flocculation, formation
- Current measures of charge
- Bulk streaming potential-based methods
- Electrophoresis of colloidal-sized particles (no
fibers)
99Charge Measurement by Fluorescence Staining
- Benefits
- Measure charge distributions of fibers and fines
simultaneously - View and control papermaking wet-end
electrochemistry at the scale at which the
critical interactions take place - Incorporate into current kappa-morphology fiber
analyzer
100Charge Measurement Results
- Identified blue charge-specific stain
- Identified red reference stain
- Established calibration between fluorescence
intensity and surface charge
101Charge-Sensitive Blue Stain - MQAE
NOT Charge-Sensitive Red Stain - AO
LOW CHARGE FIBERS 0.03meq/g
HIGH CHARGE FIBERS 0.29meq/g
102Emission Intensity of Stains on High and Low
Charged Fibers
103Correlation Between Fluorescence Emission
Intensity Ratio and Fiber Charge