Environmental Engineering Reference
In-Depth Information
with former published results and is vital for theoretical and technical fluorescence
measurement, which has been expressed by means of three methods in earlier stud-
ies. (i) The typically used standard quinine sulfate method is the fluorescence inten-
sity normalized to that of aqueous solutions of quinine sulfate monohydrate (at
ppm or ppb levels) in either 0.05-0.1 M or 0.1 N solution of sulfuric acid (H
2
SO
4
).
However, aquatic scientists use different scaling units to express the fluorescence
intensity, such as millifluorescence (mFI) (Dorsch and Bidleman
1982
; Hayase
et al.
1987
), the fluorescence unit (flu) (Chen and Bada
1992
), Raman-normalized
quinine sulfate equivalents (QSE) (Coble et al.
1998
; Kowalczuk et al.
2009
), and
the quinine sulfate unit (QSU) (Coble
1996
; Mopper and Schultz
1993
; Yamashita
and Tanoue
2003a
; Mostofa et al.
2005a
; Nagao et al.
2003
; Burdige et al.
2004
;
Zhang et al.
2009a
,
2009b
; Coble et al.
1993
; Obernosterer and Herndl
2000
).
(ii) The arbitrary unit method is the direct fluorescence intensity that is primar-
ily detected by the fluorescence spectrophotometer (Mayer et al.
1999
; Cory and
McKnight
2005
; Fu et al.
2007
,
2006
; Baker and Curry
2004
; Chen et al.
2003
;
Klapper et al.
2002
; Yue et al.
2006
). The Raman peak intensity of Milli-Q water
at Ex/Em
=
348 or 350 or 275/303 nm over the analysis period (or the QS solu-
tion as mentioned before) is used to monitor the stability of light emitted by the
xenon lamp in the fluorometer. (iii) The Raman Unit (RU) method is the corrected
fluorescence intensity, where the Raman signals are corrected by the baseline and
integrated over the entire Raman peak for each excitation wavelength. Then, the
fluorescence intensities are divided by the Raman area for the corresponding exci-
tation wavelength to obtain a RU (nm
−
1
) (Determann et al.
1994
,
1996
; Stedmon
et al.
2003
; Fulton et al.
2004
; Mostofa et al.
2005b
; Matthews et al.
1996
; Nieke
et al.
1997
; Hayakawa et al.
2003
; Yoshioka et al.
2007
; Huguet et al.
2009
). The
RU calibration processes are difficult at the shorter wavelength regions due to
noise, whilst the fluorescence intensities at longer excitation wavelengths tend to be
enhanced (Mostofa et al.
2005b
). Because of RU calibration on fluorescence inten-
sities over all the excitation wavelengths, it is possible to have artifacts that produce
unusual fluorescent components that are not shown in the original EEM spectra.
For example, the results of PARAFAC modeling on QSU and RU EEM data of
upstream (Nishi-Mataya) and lake waters (Lake Biwa) show the presence of fulvic
acid-like substance (components 1 and 2) with two fluorescence peaks for compo-
nent 1, at Ex/Em
=
320/442 (peak C) and 255/442 nm (peak A) (Fig.
2
a). There is
no specific peak for component 2, which only has strong fluorescence intensity at
Ex/Em
=
225-230/428 nm (Fig.
2
b) in QSU of upstream waters. Photobleached
fulvic acid-like compounds have two fluorescence peaks at Ex/Em
=
310/450 nm
(peak C) and 250/450 nm (peak A) (component 1, Fig.
2
e), while component 2 is
associated to photobleached autochthonous fulvic acid-like material with a weak
peak at Ex/Em
=
280/442 nm (component 2, Fig.
2
f) in QSU of lake water. On the
other hand, in RU units the respective fluorescent components are composed of the
two fluorescence peaks at Ex/Em
=
370/474 nm (peak C) and 270/474 nm (peak
A) for component 1 (Fig.
2
c). Component 2 has a peak at Ex/Em
=
330/427 nm
(peak C) in upstream waters. In lake waters, the fluorescence peaks for component
1 are 350/461 nm and 260/461 nm (Fig.
2
g), and the peaks for component 2 are
