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


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.


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


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


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


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.

(%)


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


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


7/16 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


8/16


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.


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


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.


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


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


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.


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.


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

113591 Handbk FLD 3000 Method Develop 22May2012 LPN2988

Documents

Transcript of 113591 Handbk FLD 3000 Method Develop 22May2012 LPN2988