113591 Handbk FLD 3000 Method Develop 22May2012 LPN2988
-
Upload
yeisson-mora -
Category
Documents
-
view
219 -
download
0
Transcript of 113591 Handbk FLD 3000 Method Develop 22May2012 LPN2988
-
7/27/2019 113591 Handbk FLD 3000 Method Develop 22May2012 LPN2988
1/16
Fluorescence Method Development
HandbookHolger Franz and Verena Jendreizik; Germering, Germany
Handbook
ey Words
UltiMate 3000
System
3D Synchro,
Excitation, and
Emission Scans
Variable
Emission Filter
Sensitivity
Xenon Flash
Lamp
Eluent Quality
Test
Introduction
Fluorescence spectroscopy is the most sensitive optical
detection technique used with high-perormance liquid
chromatography (HPLC). Fluorescence o a molecule can
occur naturally or can be achieved by labeling the molecule
with a uorescent tag in a derivatization reaction. A classic
application o such a derivatization is or amino acid analysis.
The chemical modifcation can be perormed in an automated
way prior to the chromatographic separation (precolumn)1 or
ater the separation (postcolumn).2
In the ow cell o a uorescence detector (FLD), the
active molecule is exposed to light o a defned wavelength
originating rom a high energy light source, typically a xenon
lamp. Ater a lietime o up to several nanoseconds (ns), the
excited analyte relaxes and emits the energy at a less energetic,
longer wavelength.
A wavelength-selective FLD usually utilizes a photo-
multiplier which is positioned at an angle o 90 to the
light source and detects the light that is emitted rom the
uorescing compounds.
In contrast to UV-vis detectors, a uorescence detector
measures a very weak light signal rather than the dierence
between light intensities (absorbance). A related challenge
is to separate high-intensity excitation radiation rom low-intensity emission light; a related beneft o uorescence
detection compared to UV detection is its ability to
discriminate an analyte rom intererences or background
peaks.3 Generally speaking, uorescence detection has
dierent requirements compared to the almost ubiquitous
UV detection.
This handbook demonstrates how to optimize
uorescence detection using the Thermo Scientifc Dionex
UltiMate 3000 FLD-3000 Fluorescence Detector series.
This document guides the analyst through the dierent
method development and optimization steps, recommends
the most suitable procedures, and explains optical eects
that might be observed. The handbook does not discussspecial requirements when combining uorescence detection
with UHPLC. For additional inormation on this topic, see
Reerences 4 and 5. A number o application notes using the
FLD-3000RS series uorescence detector are also available.
See Reerences 6 to 10 or details.
Equipment
UltiMate 3000 Binary Rapid Separation System including
SRD-3400 Solvent Rack with our degasser channels
HPG-3400RS Binary Pump
WPS-3000RS Wellplate Sampler
TCC-3000RS Thermostatted Column Compartment
DAD-3000RS Diode Array UV-vis Detector with
semi-micro ow cell
FLD-3400RS Fluorescence Detector with
Dual-PMT and micro ow cellThermo Scientifc Dionex Chromeleon Chromatography
Data System (CDS) sotware version 7.1
All modules were connected using 0.005 in. (0.13 mm) i.d
Thermo Scientifc Dionex Viper fttings.
Chromatographic Conditions
Column: Thermo Scientifc Acclaim RSLC 120, C18
2.2 m, 3.0 100 mm (P/N 071604)
Standard: PAH-Mix 221, Dr. Ehrenstorer GmbH,
P/N XS20950221AL, Lot 91124AL
12/2011 including naphthalene,
anthracene, uoranthene, pyrene anddibenzo(a,h)anthracene; 100 ng/L in
CH3CN, Dilution 1:10 (standard 1) or
1:100 (standard 2) injection volumes 1 L
or 10 L
Eluent A: Water, Fisher Scientifc Optima LC/MS
(P/N W6-212)
Eluent B: Acetonitrile, Fisher Scientifc Optima
LC/MS (P/N A955-212)
Gradient: Time (min) % Acetonitrile
0.0 75
2.0 75
4.0 955.0 95
5.5 75
7.5 75
Column Temp.: 45 C
Fluorescence Detector Flow
Cell Temp.: 45 C
Flow Rate: 1 mL/min
Inj. Volume: 1 L or 10 L
-
7/27/2019 113591 Handbk FLD 3000 Method Develop 22May2012 LPN2988
2/162
Experimental
Initial Steps and Overview
As or all scientifc work, method development or UHPLC
with uorescence detection starts with a literature review.
This typically provides valuable inormation on the column,
eluents, and gradients used by other researchers. One must
consider i the analyte uoresces, and i so, what are the
published excitation and emission wavelengths, and what isthe UV absorption spectrum? An absorption spectrum is a
good starting point or selecting the excitation wavelength
o an analyte because the spectrum indicates which energy
is absorbed to excite an electron to a higher quantum state.
Ultraviolet-active substances transorm electromagnetic
radiation into heat, uorescent compounds emit light o
longer wavelength during the relaxation process.
In preparation or the initial experiments, make sure
that the mobile phase is compatible with highly sensitive
uorescence measurements. Incompatible solvents may
create background uorescence or stray light. This reduces
the dynamic range o the detector, directly contributes to
baseline noise, and thereore adversely aects the limit o
detection (LOD). Solvent suppliers typically speciy i
solvents are suitable or uorescence spectroscopy.
Figure 1 provides an overview o the method
development steps discussed in this document.
Initial Experiments
A suitable UHPLC system may use an FLD as the only
detector, or an FLD in combination with a diode array
detector (DAD). This combination is not required or routi
use, but is useul or method development. The ollowing
sections explain how to work efciently with both system
confgurations.
System with DAD and FLD
The frst step is to optimize the chromatographic separatio
The easiest way is to use a DAD, which can be considered
universal detector or uorophores, because the absorption
spectrum or the molecule is typically similar to the
excitation spectrum. Fluorophores can be detected by UV-
detectors and at the same time the obtained spectra indicat
how to excite them. Note that the applied standard must b
typically 50100 times more concentrated than or the late
uorescence detection.
This handbook does not ocus on optimization o the
separation but uses constant separation conditions with
respect to column selection and gradient profle. All workis perormed using a mixture o fve polycyclic aromatic
hydrocarbons (PAHs). The PAHs are easy to separate and
detect at higher concentrations with a DAD, as shown
in Figure 2. Note that the our inital peaks are eluted
isocratically; the gradient is applied to speed the elution o
the fth peak.
Absorption and fluorescence spectra
Excitation and emission wavelengths
Chromatographic separationof compounds
Literature Review
29094
Initial Experiments
3D Emission Scan
3D Excitation Scan
Variable EmissionFilter
Sensitivity Settingof the PMT
Data Collection Rateand Response Time
Flash LampFrequency
Eluent Quality Test
Obtain UV spectra with DAD
Without DAD: Use zero order mode or
3D synchro scan of FLD
Use suitable scan speed and PMT
Specific test using optimized parameters
Smart use improves noise andlamp lifetime
Enter peak width in Chromeleoninstrument method wizard, optimizingnoise, precision, and data storage size
Adjust the gain of the photomultiplier
for best S/N
Use the auto setting or optimize manually
Scan for optimum excitation wavelengths
Scan for optimum emission wavelengths
Figure 2. Separation o ve PAHs with UV detection at 240 nm, standard 1,10 L injection.
0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 7
Minutes
0
600
ACN:
75.0 %
95.0
75.01
2
3
4
5
mAU
Peaks:
1. Naphthalene
2. Anthracene
3. Fluoranthene
4. Pyrene5. Dibenzo-
(a,h)anthracene
290
Figure 1. Overview o fuorescence method development steps discussed in
this document.
-
7/27/2019 113591 Handbk FLD 3000 Method Develop 22May2012 LPN2988
3/16 3
The DAD now provides easy access to UV spectral
inormation or each compound. Figure 3 shows all
fve spectra (obtained at peak apex) and determines the
wavelength o the absorption maximum. As the absorption
maximum is almost identical with the excitation maximum o
the analyte, the excitation wavelength is typically chosen close
to the absorption maximum.
System with FLD Only
The FLD-3000 series uorescence detectors support two
operation modes that are recommended or initial detection
when little is known about the nature o the analytes and their
retention times. Note that all the ollowing FLD experiments
were perormed with the standard photomultiplier tube
(PMT) that can be used in the emission wavelength
range o 220650 nm because all compounds o the test
application all within this range. The FLD-3000 detectors
are also available in a unique Dual-PMT confguration.
This confguration adds a second PMT which expands the
emission wavelength range up to 900 nm. The beneft o theDual-PMT is that the wavelength range can be expanded
without sacrifcing sensitivity in the UV range, as commonly
occurs with other FLD detectors (Figure 4).
The Dual-PMT can be ordered with the detector or as an
upgrade. I analytes are present that require detection in the
visible or near-inrared range, use the extended wavelength
range o the Dual-PMT.
200 250 300 350 400 450 500
200
1600 218.6
nm
mAU
NaphthaleneMax Abs: 219 nm
200 250 300 350 400 450 500
200
1400 251.7
nm
mAU
AnthraceneMax Abs: 252 nm
200 250 300 350 400 450 500
50
350 235.4
nm
mAU
FluorantheneMax Abs: 235 nm
200 250 300 350 400 450 500
50
450 239.9
nm
mAU
PyreneMax Abs: 240 nm
200 250 300 350 400 450 500
100
800 297.3
nm
mAU
Dibenzo(a,h)-anthraceneMax Abs: 297 nm
29096
Figure 3. The UV absorption spectra o the ve analytes, which also indicate the
excitation maximum o each.
Figure 4. The Dual-PMT option o the FLD-3000 detector is a unique solution
or providing optimum sensitivity across all wavelengths.
Zero Order Mode
In zero order mode, the grating o the emission mono-
chromator reects the entire emission spectrum o the sample
onto the PMT, rather than only a defned wavelength. Theexcitation monochromator is set to a single wavelength, as
usual. To detect peaks in zero order mode, the excitation
light must have a wavelength somewhere near the excitation
maximum o the target analytes based on literature review
or UV-absorption inormation. For instance, PAHs can be
excited at 220250 nm, although the optimum excitation
wavelength may be quite dierent.
Figure 5 compares two chromatograms obtained in zero
order mode with (A) excitation at 220 nm and (B) excitation
at 250 nm. The responses or the fve peaks in chromatogram
(A) are similar, whereas chromatogram (B) is dominated by
the size o the second peak, anthracene. This compound has
its excitation maximum at 250 nm; however, all other peaksstill can be detected. At both 220 and 250 nm, all analytes
can be excited and their retention times can be determined
and optimized. Zero order mode allows simultaneous
detection o all emitted light, independent o the emission
spectrum.
0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
Minutes
0
20 105
0
55 105
Response
(counts)
29097
B
A
Response
(counts)
Figure 5. Dierent excitation wavelengths can be used or detection in zero
order mode, as long as the analytes can be excited at all. Depending on the
characteristics o the analytes, the response varies.
200 300 400 500 600 700 800 900
Standard Wavelength RangeExtended Wavelength RangeDual PMT
LogarithmS
ensitivity
26787
-
7/27/2019 113591 Handbk FLD 3000 Method Develop 22May2012 LPN2988
4/164
3D Synchro Scans
The 3D synchro scan is a tool that provides uorescence
response as long as any combination o excitation and oset
results in a measurable light emission. This mode is even more
generic than the zero order mode because knowledge about
the excitation range o the analyte is not required. Use this
mode in the early chromatographic development stage: or
instance, to optimize the chromatographic resolution. It is not
a tool or discovering the most suitable excitation or emission
wavelengths.
A user-defned excitation wavelength range is scanned,
whereas the emission wavelength is synchronously scanned
with a fxed distance to the excitation wavelength. This scan
is repeated over the entire data acquisition time, resulting
in a DAD-like 3D feld. The oset must take a reasonable
wavelength shit, such as 30100 nm, into account. This is
another approach to determine the retention times o the
target analytes. In addition, rough excitation and emission
wavelengths pairs can be obtained; however, optimum
excitation and emission wavelengths can only be determined
by perorming individual excitation and emission scans.
Figure 6. Contour plot o a synchro scan with 30 nm oset between excitation
and emission wavelengths.
Minutes
ExcitationWavelength(nm)
0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5
250
450
29098
Figure 6 shows how a uorescence 3D scan data is
displayed in Chromeleon sotware. In a synchro scan, the
y-axis represents the excitation wavelength. The emission
wavelength is Em=Ex+oset. In the example shown in
Figure 6, the oset is 30 nm. Using this plot, wavelength
specifc chromatograms can be extracted.
Figure 7. Synchro scan o pyrene at 30 nm oset (black), overlaid with the
respective excitation (blue) and emission spectra (green). The synchro scan
maximum represents the excitation wavelength at which the best response is
obtained with the given oset.
Figure 7 shows an example o a single synchro scan.
The black synchro spectrum is a result o the combination
o the excitation and the emission spectra also provided
in Figure 7. The synchro spectrum maximum at 337 nm
represents the most suitable excitation and emission wavelength
combination at the given oset o 30 nm. The excitation
wavelength o 337 nm is close to pyrenes excitation
maximum at 333 nm. The emission wavelength o 367 nm
(337 nm + 30 nm) provides a air light intensity.
A shit to a shorter wavelengthor instance, to
333 nmimproves the excitation but decreases the emissio
response. A slight shit toward 345 nm does the opposite:excitation reduces as emission improves. Given the 30 nm
oset, an excitation at the excitation maximum o 238 nm
would result in almost zero response.
For all 3D scans, the selection o a suitable scan speed is
important. In Chromeleon sotware, the scan can be set to
our dierent speeds: slow, medium, ast, and superast. Slow
provides the lowest noise because it efciently combines sma
wavelength steps with a signifcant number o data point
repetitions at each wavelength selection. This speed is ideal
monitoring the quality o a mobile phase, but it is typically to
slow to be used or scanning peaks.
386.7
238.1
270.5
332.9
336.9
280 320 360 400 440 480 52
Wavelength (nm)
240200
390.0
375.5
Emission Spectrum Pyrene, Ex 240 nm Excitation Spectrum Pyrene, Em 390 nm Synchro Spectrum Pyrene, Offset 30 nm
290
-
7/27/2019 113591 Handbk FLD 3000 Method Develop 22May2012 LPN2988
5/16 5
Figure 8 identifes the recommended scan speed based
on the peak width and the scan range. I the peak width/scan
range coordinates are within a certain scan speed range, a ew
repeated scans per peak are guaranteed, which is enough to
detect a peak and to assign related spectra.
Figure 9. Overlay o dibenzo(a,h)anthracene excitation spectra obtained with
medium (blue), ast (black), and superast scan speed (red). Medium and ast
spectra are similar; the superast spectrum is not as precise but supports
scanning o very narrow peaks.
Thereore, the majority o UHPLC and HPLC peaks,
even at sub-2 second peak widths, can be detected by the ast
scan speed.
Varying scan speeds will impact data quality.
Figure 9 compares dibenzo(a,h)anthracene excitation spectra
obtained at dierent speeds. The very similar medium
and ast excitation spectra vary to some extent rom the
superast spectrum. Medium and ast scan acquisitions
dier by a wavelength step dierence o only 0.3 nm. Theirmain dierentiation is the number o ashes obtained per
wavelength step, inuencing noise.
For the sake o speed, the superast scan provides only
1/10 o the wavelength steps o the ast scan. The lack o
wavelength steps explains why the shape o the superast
excitation spectrum in Figure 5 is similar but not very close
to the spectra obtained at slower scan speeds. The excitation
maxima, however, only vary by 5 nm. This dierence is
close enough to the correct maximum, considering the
20 nm spectral resolution o the FLD-3000 series. The
superast scan is thereore suited or an initial screening o
very narrow peaks obtained in ultraast separations. I very
narrow peaks require superast scans, it is possible to repeat
the experiment at ast scan speed and a reduced wavelength
range o 4060 nm around the maximum.
In general, the ast scan speed is a good deault setting. It
is suitable to scan typical peak widths o both conventional
HPLC and UHPLC separations. By selecting a smaller
wavelength range, the scan rate and thereore the support o
narrower analyte bands can be urther improved.
Optimizing the Emission Wavelength
Once the chromatographic method development is per-
ormed with the help o a UV detector, the zero order mode,
or the 3D synchro scans o the FLD-3000 detector series,
the next step is to optimize the detection parameters o the
uorescence detector to achieve the best signal-to-noise (S/N)
ratio. The frst tool or this optimization process is the 3D
emission scan eature.
The FLD-3000 Series supports the acquisition o
permanent 3D emission scans. During the scan, the excitation
remains constant at a defned wavelength. Across the entire
acquisition time, the emission monochromator repeatedly
scans a user-defned emission wavelength range. Similar to
the synchro 3D scan, this results in a DAD-like 3D feld.
However, the synchro scan is a more generic approach
because an excitation wavelength does not have to be defned.
Instead, only a resonable Stokes shit is entered (also see: 3DSynchro Scans).
Figure 8. The 3D scan speed can be selected based on expected peak width and
scan range. Even UHPLC peaks are easily supported. Note that the y-axis uses a
logarithmic scale.
40 100 160 220 280 340 400 460 520 580
Scan Range (nm)
640
29100
7000
Stop-Flow Scan Recommended
Fluorescence 3D Scan Speeds
Slow ScanRecommended for Solvent Monitoring
Medium Scan
Fast Scan
Super-Fast Scan
1000
100
10
1
2
5
0.5
PeakWidth(s)
50
20
500
200
200 300 400
Wavelength (nm)
0
55Medium
Fast
Super-Fast
Rel.Intensity(%)
29101
-
7/27/2019 113591 Handbk FLD 3000 Method Develop 22May2012 LPN2988
6/166
Figure 10 shows emission scans extracted rom a single
separation. The data were obtained by setting the emission
scan speed to ast and the constant excitation wavelength to
219 nm, because the naphthalene UV spectrum in
Figure 3 suggested that 219 nm is a good excitation
wavelength or this compound. However, it is likely that
exciting at 219 nm is not ideal or the other analytes. When
compounds show quite dierent excitation spectra, it is
questionable whether the excitation wavelength chosen has
an impact on the resulting emission spectra.
Figure 11. Dibenzo(a,h)anthracene emission spectra extracted rom our
dierent 3D emission scans. Each spectrum was obtained with a dierent
excitation wavelength. The relative emission spectra are not infuenced by th
excitation wavelength selection, except or the highlighted areas which are a
consequence o second order diraction light.
360 400 450 500 550 60
Wavelength (nm)
0
60391.6
412.5
Dibenzo(a,h)anthraceneEx: 324 nm
0
55 391.5
412.5
Dibenzo(a,h)anthracene
Ex: 235 nm
0
55 391.9
412.7
573.5
Ex: 287 nm
0
55
Rel.Intens.
(%)
391.8412.7
553.4
Dibenzo(a,h)anthraceneEx: 277 nm
Rel.Intens.
(%)
Rel.Intens.
(%)
Rel.Intens.
(%)
Dibenzo(a,h)anthracene
29
Figure 11 shows our emission spectra o dibenzo(a,h)-
anthracene obtained in our dierent 3D emission scans. The
excitation wavelength was alternated between 235 nm,
277 nm, 287 nm, and 324 nm. All spectra look qualitatively
the same, except or the peaks indicated by the blue rectangle
These peaks are a result o second order diraction light o
the applied excitation wavelengths. Light diraction at a
grating occurs as a frst, second, or higher order process. Thes
diraction orders requently overlap. Second order diraction
light, or example, is always visible at twice the excitation
wavelength; so when exciting at 287 nm, second order light isvisible at 574 nm.
In an optical detector, higher order diraction light
is typically suppressed by optical long-pass flters. The
FLD-3400RS uorescence detector eatures a unique variab
emission flter that automatically selects a suitable emission
cut-o flter element.
Figure 10. Emission spectra obtained rom a single chromatographic run with
3D emission acquisition. The optimum emission wavelength or each compoundis at the absolute maximum o the emission spectrum.
250 300 400 500
Wavelength (nm)
0
55 327.9
441.3
Naphthalene
500250 300 400
0
55 399.8380.0
423.2
Anthracene
250 300 400 500
0
55 443.2Fluoranthene
250 300 400 500
0
55
Rel.Intensity(%)
390.0376.0
Pyrene
250 300 400 500
0
55 396.0
417.1
438.5
Dibenzo(a,h)anthracene
Ex Wvl: 219 nm
Em Wvl Range: 250500 nm
Scan Speed: Fast
Scan Sensitivity:1
Emission Filter: Open
Wavelength (nm)
Wavelength (nm)Wavelength (nm)
Wavelength (nm)
Rel.Intensity(%)
Rel.Intensity(%)
Rel.Intensity(%)
Rel.Intensity(%)
29102
-
7/27/2019 113591 Handbk FLD 3000 Method Develop 22May2012 LPN2988
7/16 7
The ideal flter cut-o wavelength is between the
excitation and the emission wavelengths. In the case o
Figure 11, the flter was set to 280 nm or all runs to provoke
the occurence o second order diraction light. For excitation
at 277 nm, the second order eect is visible because the flter
does not cut o at 280 nm sharp but has a certain transition
range. This allows light o 277 nm to partially pass. The ideal
emission flter depends on the excitation and the emission
wavelength. The nominal flter wavelength should be betweenthe excitation and the emission wavelength. Thereore, or
excitation wavelengths o 277 nm or 287 nm, the selection o
the next higher emission flter (370 nm) would suppress higher
order eects. The emission range must then be adapted not to
overlap with the excitation. During 3D scans, light o higher
diraction orders typically can be recognized by the band
that is present during the complete separation (Figure 12).
The FLD-3100 uorescence detector uses a fxed
280 nm emission flter, similar to other FLD detectors. With
this confguration, second order light rom excitation wave-
lengths around 280 nm and higher cannot be suppressed.
Returning to the 3D emission scan example o
dibenzo(a,h)anthracene, Figure 13 shows an excitation
spectrum o this molecule. The excitation maximum is
at 293 nm. Below this wavelength, excitation decreases
but remains present. It is thereore possible to excite
dibenzo(a,h)anthracene across the range o 220300 nm.
Typically, the obtained emission intensity will vary with the
selected excitation wavelength, but the qualitative appearanceo the emission spectrum remains unchanged.
In summary, optimum emission wavelengths can easily be
extracted rom 3D emission scans. The FLD-3400RS detector
is most efcient in suppressing higher order diraction light
because it automatically selects the best suitable emission
flter. The emission spectrum is typically not aected by
the excitation wavelength (also known as Kashas rule).11
Customers using the FLD-3000 detector series can thereore
typically extract optimum emission wavelengths or all analytes
rom a single 3D emission scan.
Optimizing the Excitation Wavelength
The second part o the wavelength optimization is the
identifcation o suitable excitation wavelengths. The easiest
way to do this is to acquire permanent 3D excitation
scans. During the scan, the emission wavelength remainsconstant. Across the entire acquisition time, the excitation
monochromator repeatedly scans a user-defned excitation
wavelength range. Similar to synchro and emission 3D
scans, this results in a DAD-like 3D feld.
The detector logic requires constant settings o the
scan range and the emission wavelength across the entire
chromatogram. For a separation o quite dierent analytes,
as in the example, there are options to run a 3D scan
sample with the optimized emission wavelength or each
compound, or to use a generic emission wavelength across
all analytes (i.e., a single run or all peaks). This leads to a
question similar to that raised or the emission scan: does
the selection o the constant emission wavelength inuencethe shape o the observed excitation spectrum?
Figure 12. A 3D emission scan with an excitation wavelength o 287 nm.
Second order stray light is visible as a permanent band at 554 nm.
Figure 13. Excitation spectrum o dibenzo(a,h)anthracene. The molecule
can be excited best at 293 nm but also in the range down to approximately
220 nm.
Minutes
nm
1.5 2.0 2.5 3.0 3.5 4.0 4.5
360
600
29104-01
200 220 240 260 280 300 320 340 360 380 400 423
Wavelength (nm)
0
55
R
el.Intensity(%)
293.3
228.7
246.0
Excitation low
but present
29105
-
7/27/2019 113591 Handbk FLD 3000 Method Develop 22May2012 LPN2988
8/16
-
7/27/2019 113591 Handbk FLD 3000 Method Develop 22May2012 LPN2988
9/16 9
applying the auto setting o the VEF, the built-in logic o
the detector selects the best flter or the vast majority o
analytes and wavelength combinations. Figure 17 provides
an example. The graph shows the relationship between flter
setting and the S/N ratio or two analytes. The flter which
achieves the highest S/N is labeled with the related ratio. In
both cases, the 370 nm flter provides the best result. This
flter also would have been selected by the auto setting.
Minutes
Log.MaxPMTSaturation
Respo
nse(
counts)
ClearPMTSaturation
Log.MaxPMTSaturation
ClearPMT-
Saturation
29110
Figure 19. MaxPMTSaturationis a useul parameter to optimize the sensitivity
o the PMT.
Note that the 435 nm cut-o flter still allows a certain
part o the light to pass, although the emission wavelengths
o uoranthene and pyrene are below the cut-o flter
wavelength (443 nm and 390 nm). The reason or this is the
20 nm bandwidth o the emission monochromator and the
transmission characteristics o the flter. At the specifed cut-
o wavelength, 50% o the light can pass. As indicated in
Figure 18, the usable emission wavelength range starts at
15 nm below the nominal flter wavelength (i.e., or a 280
nm flter, at 265 nm).
80007000600050004000300020001000
0280
9000
no filter
Signal-to-
NoiseRatio
29108
435370
Fluranthene (Ex 232, Em 433) Pyrene (Ex 238, Em 390)
2423
8504
Figure 17. Signal-to-noise ratio or two analytes with dierent cut-o ltersettings. The best result is obtained with the 370 nm lter, which would also be
selected by the auto setting.
50%
15 nm
Useable Em wavelength range
220 240 260 280 300 340 360
100%
Wavelength (nm) 29109
320
Filtertransmission
Figure 18. At the specied cut-o wavelength o the lter, 50% o the light
is transmitted. The usable emission wavelength range starts at 15 nm below
the nominal lter wavelength.
In summary, the FLD-3400RS uorescence detector
automatically picks the best suitable flter or a selected
wavelength pair, reduces stray light, and thereore
improves S/N due to its unique variable emission flter.
Optimizing the Sensitivity Setting
All commercial FLDs or use with HPLC use a PMT or
measuring the intensity o the emitted light. It is thereore
important to understand the capabilities and limitations o
this device. A PMT can be regarded as a current source that
generates a current proportional to the light intensity to
which it is exposed.
A PMT consists o a photocathode, several dynodes, and
an anode. When photons hit the photocathode, electrons
are ejected rom its surace. Between the photocathode and
the frst dynode is an electrostatic potential dierence that
accelerates these electrons towards the dynode. Based on
the phenomenon o secondary emission, this dynode emits a
larger number o electrons. A ollowing cascade o dynodes
unctions as an electron multiplier. Depending on the voltage
dierence between the dynodes, more or less electrons
are emitted at each incident (i.e., the gain o the chain o
dynodes can be adjusted by the applied voltage). Finally, a
current pulse arrives at the anode.12
Depending on the wavelength o the detected light and its
intensity, it is useul to tune the voltage applied to the PMT.Too large a voltage can cause an electron overow or even
damage the light-sensitive photocathode. Too low a voltage
can result in increased noise and thereore in a reduced S/N
ratio. The FLD-3000 series property or adjusting the gain o
the PMT is called Sensitivity. At any time o data acquisition,
this setting can be altered between 1 and 8; 1 represents the
lowest and 8 the highest settable voltage (i.e., the sensitivity
o the PMT to light). Typically, the best limit o detection can
be obtained with a setting o 5 or 6.
The intensity o the emission light (measured by the
PMT) is normalized with the intensity o excitation light
passing through the sample measured by the reerence diode
(Figure 16). Thereore, the values (counts) that the emissionchannel displays cannot be used or optimizing the sensitiv-
ity. Instead, a dierent procedure is required to tune the
sensitivity setting or the best S/N ratio. The schematic o
this procedure is depicted in Figure 19.
-
7/27/2019 113591 Handbk FLD 3000 Method Develop 22May2012 LPN2988
10/1610
At this point o the method development, elution
windows o peaks o interest are known. Between the elution
o the peaks, a pair o commands can be placed to determine
the PMT saturation, i.e. the ratio between light actually
received and the maximum allowed light level. Use a sample
or standard with highest expected concentration to optimize
this setting.
The parameterMaxPMTSaturation indicates the highest
saturation o the PMT signal measured since the last reset.Logging this value ater the elution o a peak with the
command log.MaxPMTSaturationwill create an entry in
the sample audit trail. Beore the next eluting peak, reset
the PMT saturation with the ClearMaxPMTSaturation
command. Repeat the process ater each peak.
The saturation o the PMT is ideal when it is between
3080% (Table 1). I the saturation is below 30%, increase
the sensitivity so that it reaches the optimum range. A
one-step increase o the sensitivity increases the peak height
by a actor o approximately 2. When the PMT is 100%
saturated, the detector automatically reduces the sensitivity
setting. This ensures that the PMT is not damaged; however,
the peak which causes the PMT overow cannot be used or
quantitative analysis.
Note that a PMT overow at sensitivity 1 does not lead
to a reduced sensitivity, but reports an error in the audit
trail, and potentially leads to a non-optimal peak shape and
probably incorrect quantifcation o the peak. Thus, or a
successul quantifcation, reinject the sample and reduce the
sensitivity setting as suggested by the Chromeleon sotware
sample audit trail, or dilute the sample i sensitivity 1 is still
too sensitive.
Note that it is not mandatory to optimize the sensitivity;
it is possible to operate the PMT at a lower-than-optimal
setting.
Figure 20 provides an example o how much the signal
height, noise, resulting S/N, and PMT saturation change,
depending on the sensitivity setting. When changing rom
1 to 5, the noise increases rom 2501 to 10,759 counts,
equating to a actor o 4 (calculated with equation 1). At th
same time, the peak height increases by a actor o almost
As a consequence, the S/N ratio improves by a actor o 4.
Sensitivity 5 provides the best S/N ratio; however, it is only
slightly better than sensitivity 4. So sensitivity 4 may be the
better choice because it provides a larger dynamic range; it
can handle unknowns with concentrations 2 higher than
those o sensitivity 5.
Equation 1:
Table 1. Recommended actions to optimize the sensitivity setting.
S/N = 2.Noise
Peak Height
10 106
70 106
0
Height Noise PMTSensitivity dibenzo(a,h)anthracene (counts) S/N Saturatio
(counts) (%)
1 3.2 106 2501 2537 4
2 7.0 106
1754 7929 9
3 15.1 106 3518 8576 19
4 28.0 106 5749 9755 36
5 53.8 106 10759 9998 69
6 PMT overflow
Minutes
Response(
counts)
291
Figure 20. Eect o the sensitivity setting on the peak height, noise, and
S/N ratio o dibenzo(a,h)anthracene. The S/N improves by a actor o 4 rom
sensitivity 1 to sensitivity 5. At sensitivity 6, the PMT overfows
(standard 2, 1 L injection).
Value Action to Optimize
-
7/27/2019 113591 Handbk FLD 3000 Method Develop 22May2012 LPN2988
11/16 11
In summary, the sensitivity setting allows tuning o the
PMT or the best S/N ratio. I dierent sensitivities provide the
same S/N ratio, it is benefcial to use the lower setting because
it supports higher concentrated samples. The optimization can
easily be done with the command pair oMaxPMTSaturation
and ClearMaxPMTSaturation. I a PMT overows in routine
operation, the detector automatically reduces the sensitivity
setting. Rerun the same sample with the adjusted sensitivity
in order to quantiy the peak that previously caused thePMT overload.
Data Collection Rate and Response Time
Data collection rate (DCR) and response time are settings
which inuence data storage size and S/N ratio.
The DCR is the number o data points per second gener-
ated by the detector electronics. For good precision during
quantitative analysis, a peak should be defned by 2030
data points. I the data collection rate is too low, the start
and end points o peaks will not be determined accurately.
I the data collection rate is too high, data fles may occupy
excessive disk space and post-run analyses may require more
processing time.
The DCR is typically paired with a mathematical data
flter called response time (or simply response or rise time).
The response time is a measure o how quickly the detector
responds to a change in signal. It is defned as the time it
takes the detectors output signal to rise rom 10% to 90%
o its fnal value. The input value is postulated to directly
jump rom 0% to 100% (Figure 21).
Initial experiments should be perormed at relatively
high data collection rates and short response times. Once
separations are optimized, the widths o relevant peaks can
be extracted rom the chromatogram. As in Chromeleon
sotware version 6, Chromeleon sotware version 7 provides
guidance on the selection o data collection rates and response
times. Calculated based on the peak width, Chromeleon
sotware makes sure that a sufcient number o data points
across a peak are obtained and optimizes the related response
time. To access this guidance, open the instrument method
(Chromeleon sotware version 7), click on the FLD symbol
on the let, and select the rider Channel Settings. In the Data
Collection section, the peak width can be entered (Figure 22).
Select DCR and response times as suggested by the sotware.
Selecting appropriate data collection settings is thereore aneasy task.
Signal
10%
Time
90%
Response Time (or Rise Time)
Filtered Signal
Input Signal
29112
Figure 21. The response time measures the time until the detector responds to
a change in signal. It is dened as the time it takes the detector's output signalto rise rom 10% to 90% o its nal value ollowing an instantaneous change o
the input signal.
29113
Figure 22. The Chromeleon instrument method wizard calculates suitable DCR
and response times based on the width o the narrowest peak o interest.
-
7/27/2019 113591 Handbk FLD 3000 Method Develop 22May2012 LPN2988
12/1612
Effect of Changing the Xenon Flash Lamp Frequency
The FLD-3000 detector series uses a xenon ash lamp as a
light source. The expected lietime o this lamp is related to
the lamp operation time and the used ash requency. Xenon
ash lamp operation time is similar to data acquisition
time because the lamp only requires a ew seconds to warm
up, unlike continuous xenon and mercury lamps that need
time to reach a thermal equilibrium and typically must be
switched on 1 h prior to the frst injection; these lamps
usually remain on between samples. Leaving the lamp on
obviously reduces the number o samples analyzed during the
lietime o a continuous lamp. A ash lamp with a lietime
o 4000 h can thereore process ar more than 4 times the
number o samples o a 1000 h continuous lamp.
The FLD-3000 detector series xenon ash lamp oers
three dierent ash requencies: HighPower (300 Hz),
Standard (100 Hz), and LongLie (20 Hz). A higher ash
requency improves the S/N ratio; a lower ash requency
improves the lietime o the lamp. Operating the lamp in
LongLie mode results in >15,000 h o chromatography.
Note that the lietime o the lamp is defned as such that thespecifcations o the Raman S/N (an internal measurement to
judge lamp health) will still be achieved. Thus, at the end o
the lamp lietime, the sensitivity o the detector is expected
to reach a limit, but the lamp may be used longer without
any problem.
Next, consider the inuence o the lamp mode on
both the analyte peak and the baseline. Figure 23 shows
an overlay o three dibenzo(a,h)anthracene peaks detected
using HighPower, Standard, and LongLie modes. All other
settings remained constant. The lamp mode change does
not aect the peak height or area, but the baseline noise
increases with a smaller ash requency.
3.65 4.15
0
HighPowerStandard
LongLife
10 106
Minutes
Response
(counts)
29114
17 104
15 104
Figure 23. Changing the fash lamp requency has an impact on the baseline
noise, but the peak height remains constant. The lamp mode thereore only
infuences the noise and respective S/N (standard 2, 1 L).
Figure 24. Signal-to-noise ratio or dibenzo(a,h)anthracene obtained with
HighPower, Standard, and LongLie lamp modes.
Noise (F2) = Noise (F1)F2
F1
The related S/N ratios are shown in Figure 24. A rough
estimate o the noise change can be calculated using
Equation 2.
Equation 2:
With
F1= Previous ash lamp requencyF2= Planned ash lamp requency
Changing the lamp mode rom HighPower (300 Hz) to
Standard (100 Hz) thereore increases the baseline noise by
approximately 3 which is a actor o ~1.7. In the example
the noise changes by a actor o 1.53. From Standard
(100 Hz) to LongLie (20 Hz), the predicted actor is 2.24.
The actor calculated rom real results is 1.7. In summary,
equation is not precise, but allows to estimate the eect o
the lamp requency change.
9482
29
36326178S/N
8000
7000
6000
5000
4000
3000
2000
1000
0
9000
10000
HighPower LongLifeStandard
-
7/27/2019 113591 Handbk FLD 3000 Method Develop 22May2012 LPN2988
13/16 13
Time (min) Lamp Mode Wavelengths Optimized For Excitation (nm) Emission (nm) Sensitivity Filter
0.00 LongLie Naphthtalene 219 330 6 280
0.75 HighPower
1.50 Anthracene 247 400 3 370
1.85 Fluoranthene 232 443 6 370
2.13 Pyrene 238 39 4 370
2.60 Dibenzo(a,h)anthracene 294 396 4 370
4.25 LongLie
Table 2. Timed settings across the chromatogram.
Smart Use of the Xenon Lamp Mode
The ash lamp o the FLD-3000 detector series is so exible
that the lamp mode can even be changed during a chromato-
graphic run. This eature gives the analyst a smart way to
optimize their method.
A chromatogram typically shows parts in which no peaks
o interest elute. In gradient separations, this includes the time
until the elution o the frst peak and the wash/re-equilibration
time at the end o the run. A ash lamp could even be
switched o during this time, but the variable lamp modes o
the FLD-3000 detector series allow an even smarter way to
achieve both the best S/N ratio and an increased lietime o
the light source. During chromatographic idle times, the lamp
mode can be switched to LongLie; during peak elution, the
mode can be changed to Standard or HighPower.
Figure 25 applies this approach to the separation
used throughout this document. Total run time, including
re-equilibration, is 7.5 min. The authors applied LongLie
mode rom 00.75 min and rom 4.257.5 min. In total,
the LongLie mode was active or 4 min. During the
remaining 3.5 min o the chromatogram, the lamp modewas set to HighPower. The number o ashes decreased by
50%, compared to i HighPower had been used across the
entire chromatogram. Thus, this approach leads to 100%
longer lietime o the ash lamp without any sacrifce to the
obtained data quality. Data acquisition is always on, and
during idle times, the baseline is monitored at reasonable
sensitivity.
Summary of Optimizations
Several optimized settings were obtained or the separation
o the PAH standard. The data collection rate and the
response time are settings which are applied across the entire
chromatogram:
Data collection rate: 20 Hz
Response time: 0.4 s
Other settings are optimized based on the characteristics
o a uorophore such as excitation and emission wave-length,
sensitivity setting, and the emission flter. The lamp mode is
also varied to apply HighPower mode during peak elution
and LongLie mode to sections o the chromatogram with no
peaks o interest. Table 2 summarizes these timed settings.Sensitivities were evaluated or 1 L injections o standard 2.
Figure 25. Smart use o the lamp modes urther extends the lietime o the
xenon fash lamp (standard 2, 1 L) injection.
0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 7.5
0
HighPower,
300 Hz
LongL
ife,
20Hz
4.5 106
Minutes
Respo
nse(
counts)
29116
LongLife, 20 Hz
-
7/27/2019 113591 Handbk FLD 3000 Method Develop 22May2012 LPN2988
14/1614
Test of Eluent Quality
This chapter demonstrates a way to test the eluent quality,
particularly when switching to a new lot or even a new
supplier.
As mentioned at the beginning o the Experimental
section, it is important to use solvents o compatible quality
or uorescence detection. Noncompatible solvents may
create background uorescence and thereore stray light.
This stray light reduces the dynamic range o the detector,
directly contributes to baseline noise, and thereore adversely
aects the limit o detection (LOD).
It is not only important to buy a solvent product that is
nominally uorescence compatible, but it is also useul to
make a relative comparison between dierent suppliers. In
addition, a brie test when switching to a new batch or
a new supplier gives confdence that the solvent quality
is suitable.
Results may dier with stock time o the water, so reshly
purifed lab water may be the best choice.
In a typical run, the signal is set to zero by an autozero
command. The disadvantage o this procedure is that thelevel o background uorescence created by the solvent
cannot be evaluated; or any solvent quality, the baseline
starts at zero counts. With the FLD-3000 detector series,
use the ClearAutoZero command to release the baseline. A
certain level o stray light is always present, so the baseline
will rise to a higher level. The worse the eluent quality,
the higher the baseline level observed. Figure 26 shows an
overlay o three experiments with water rom three dierent
sources, obtained with a backpressure capillary instead o a
column. A backpressure o a ew bar ensures that the pump
logic operates seamlessly. As the stray light level is inuenced
by the sensitivity setting, it must be kept constant during
comparison experiments. Here, the sensitivity was set to 5throughout the run.
At 0 min, the run starts with the excitation and emission
wavelengths or the frst analyte, naphthalene. At 0.5 min,
the baseline is released with the ClearAutoZero command.
Deionized water reaches the highest background uorescence
with ~400,000 counts under these conditions. Lab water
and the LC/MS grade water are o similar quality at ~100,000
counts.
At 1 min, wavelengths are switched to the ones used or
detecting the second analyte, anthracene, ollowed by an
AutoZero. As a consequence, the baseline returns to zero
counts response. At 1.5 min, the baseline is released again,
resulting in increased baseline levels. This process is repeatedor all the wavelength/flter combinations.
Figure 26. Three dierent water qualities are tested or their background
fuorescence at analyte-relevant detection parameters. This quick test makes
eluent quality monitoring an eective yet easy task.
0.17 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.10100000
0
500000
Naphthalene
Anthracene
Fluoranthene
Pyrene
Dibenzo(a,hanthracen
Deionized tap water
Purified lab waterLC/MS grade water
Minutes
Response(
counts)
29
This test ensures that the quality o the solvent is not just
checked at a single wavelength combination that might not
even be relevant or a given analysis. All relevant parameters
are tested quickly and can easily be compared when overlayinthem with earlier results or other eluent products.
Note that the peak at 2 min is a consequence o the
wavelength switch to uoranthene detection conditions. Fo
a short time, the background uorescence already reaches
the levels that are permanently shown between 2.5 and
3 min. The AutoZero just ater the wavelength switch mak
the baseline return to zero counts. The consequence is a
peak-like signal.
-
7/27/2019 113591 Handbk FLD 3000 Method Develop 22May2012 LPN2988
15/16 15
Conclusion
UltiMate3000FLD-3100andFLD-3400RSuorescencedetectors are designed or highest perormance and methoddevelopment exibility.
Zeroordermodeand3Dsynchroscansareeffectivetools or initial experiments without the need o a diodearray detector.
3Demissionand3Dexcitationscansreducetheeffortto determine the most suitable wavelength pair toa minimum.
Thevariableemissionlterandthesensitivitysettingo the photomultiplier can signifcantly improve theS/N ratio (FLD-3400RS only).
Chromeleonsoftwarecalculatesoptimaldatacollectionrates and response times based on the narrowest peakwidth or the best S/N ratio and data storage size.
Thisstudydemonstratestheeffectofdifferentxenonlampash requencies on noise and shows how tocombine the modes within a chromatogram or best limito detection and urther-increased lamp lietime.
Asimpleyetapplication-tailoredtestforevaluatingtheeluent quality is explained.
References1. Technical Note 107: Automated In-Needle Derivatizat ion Applying a User-
Defned Program or the Thermo Scientifc Dionex WPS-3000 Split-LoopAutosampler, LPN2849, 2011. Thermo Scientifc: Documents, TechnicalNotes. www.dionex.com/ (accessed Dec 5, 2011).
2. Pickering, M.V. Ion-Exchange Chromatography o Free Amino Acids,LC-GC19897(6), 484490.
3. Snyder, L.R.; Kirkland, J.J.; Glajch, J.L. Practical HPLC MethodDevelopment, 2nd ed.; John Wiley & Sons, Inc.: New York, 1997; p 82.
4. Technical Note 92: Combining Fluorescence Detection with UHPLC:An Overview o the Technical Requirements, LPN2716-01, 2011.Thermo Scientifc: Documents, Technical Notes. www.dionex.com/(accessed Dec 5, 2011).
6. Application Brie 115: Determination oN-Methylcarbamates, LPN2587,2010. Thermo Scientifc: Documents, Applications. www.dionex.com/(accessed Dec 5, 2011).
7. Application Update 177: Faster and More Sensitive Determination oN-Methylcarbamates in Drinking Water by HPLC, LPN2675, 2010.Thermo Scientifc: Documents, Applications.www.dionex.com/. (accessed Dec 5, 2011).
8. Application Note 266: Determination o Sialic Acids Using UHPLC withFluorescence Detection, LPN2662, 2011. Thermo Scientifc: Documents,Applications. www.dionex.com/ (accessed Dec 5, 2011).
9. Application Note 272: Faster Yet Sensitive Determination oN-Methylcarbamates in Rice, Potato, and Corn by HPLC, LPN2728, 2011.
Thermo Scientifc: Documents, Applications. www.dionex.com/(accessed Dec 5, 2011).
10. Application Note 278: UHPLC Determinat ion o Sialic Acids withFluorescence Detection, LPN2817, 2011. Thermo Scientifc: Documents,Applications. www.dionex.com/(accessed Dec 6, 2011).
11. Kasha, M. Characterization o Electronic Transitions in Complex MoleculesDisc Faraday Soc.1950,9, 1419.
12. Lakowicz, J.R. Principles of Fluorescence Spectroscopy, 3rd ed.; Springer
Science+Business Media, LLC: New York, 2006; p. 44.
-
7/27/2019 113591 Handbk FLD 3000 Method Develop 22May2012 LPN2988
16/16
Thermo Scientifc Dionex produ
designed, developed, and manu
under an ISO 9001 Quality Syste
www.thermoscientific.com
All trademarks are the property o Thermo Fisher Scientifc Inc. and its subsidiaries. This inormation is presented as an example o the capabiliti es o ThermoFisher Scientifc Inc. products. It is not intended to encourage use o these products in any manners that might inringe the intellectual property rights o others.Specifcations, terms and pricing are subject to change. Not all products are available in all countries. Please consult your local sales representative or details.
FLDHandbook_E 02/12 S
In addition to these
ofces, Thermo Fish
Scientifc maintains
a network o represe
tative organizations
throughout the worl
Dionex Products(408) 737-0700
North America:U.S./Canada(847) 295-7500
South America:Brazil(55) 11 3731 5140
Europe:Austria(43) 616 51 25
Benelux
(31) 20 683 9768(32) 3 353 4294
Denmark(45) 36 36 90 90
France(33) 1 39 30 01 10
Germany(49) 61125 991 0
Ireland(353) 644 0064
Italy(39) 02 51 62 1267
Sweden(46) 8 473 3380
Switzerland(41) 62 205 9966
United Kingdom(44) 1276 691722
Asia Pacific:Australia(61) 2 9420 5233
China(852) 2428 3282
India(91) 22 2764 2735
Japan(81) 6885 1213
Korea(82) 2 2653 2580
Singapore(65) 6289 1190
Taiwan(886) 2 875 6655