Analytica Chimica Acta 853 (2015) 608616

Contents lists available at ScienceDirect

Analytica Chimica Acta

journal homepage: www.elsevier.com/locate/aca

Hydrophilic interaction liquid chromatography method for measuring

the composition of aquatic humic substances
Ren-Qi Wang a,b, *, Leonardo Gutierrez b,c , Ng Siu Choon d , Jean-Philippe Crou b, **
a

College of Chemistry and Chemical Engineering, Lanzhou University, Lanzhou 730000, China
Water Desalination & Reuse Center (WDRC), 4700 King Abdullah University of Science and Technology, Saudi Arabia
c
Facultad del Mary Medio Ambiente, Universidad del Pacico, Guayaquil, Ecuador
d
School of Chemical and Biomolecular Engineering, Nanyang Technological University, 62 Nanyang Drive, 637459, Singapore
b

H I G H L I G H T S

G R A P H I C A L A B S T R A C T

Aquatic humic substances are measured by pH sensitive and neutral

HILIC columns.

We introduce statistical retention
time and disparity to dene mixture
complexity.

Aquatic humic substances are evaluated by statistical retention time and
disparity.

Compositions of aquatic humic substances are quantitatively measured.

Four humic substance samples from
very different origins are compared.

A R T I C L E I N F O

A B S T R A C T

Article history:
Received 4 June 2014
Received in revised form 12 September 2014
Accepted 17 September 2014
Available online 19 September 2014

A hydrophilic interaction liquid chromatography (HILIC) method was developed to measure the
composition of humic substances from river, reservoir, and treated wastewater based on their
physicochemical properties. The current method fractionates the humic substances into four welldened groups based on parallel analyses with a neutral and a cationic HILIC column, using mobile
phases of varied compositions and pH. The results indicate that: (i) the proportion of carboxylic acids in
the humic substances from terrestrial origins is less than half of that from treated wastewater (Jeddah,
KSA), (ii) a higher content of basic compounds was observed in the humic substances from treated
wastewater and Ribou Reservoir (Cholet, France) than in the sample from Loire River (France), (iii) a
higher percentage of hydrophobic macromolecules were found in the humic substances from Loire
River than in the other samples, and (iv) humic substances of treated wastewater contained less ionic
neutral compounds (i.e., pKa 59) than the waters from terrestrial origins. The physicochemical property
disparity amongst the compounds in each humic substances sample was also evaluated. The humic
substances from the lightly humic Loire river displayed the highest disparity, whereas the highly humic
Suwannee river (Georgia, USA) showed the most homogeneous humic substances.
2014 Elsevier B.V. All rights reserved.

Keywords:
HILIC
Humic substance
Electrostatic force
Hydrophilic interaction
Physicochemical property
Natural organic matter

* Corresponding author at: College of Chemistry and Chemical Engineering, Lanzhou University, Lanzhou 730000, China. Tel.: +86 931 891 2589; fax: +86 931 891 2582.
** Corresponding author. Tel.: +966 544 700 024; fax: +966 2 808 2984.
E-mail addresses: wangrq@lzu.edu.cn, wangrenqi@gmail.com (R.-Q. Wang), jp.croue@kaust.edu.sa (J.-P. Crou).
http://dx.doi.org/10.1016/j.aca.2014.09.026
0003-2670/ 2014 Elsevier B.V. All rights reserved.

R.-Q. Wang et al. / Analytica Chimica Acta 853 (2015) 608616

1. Introduction
Natural organic matter (NOM) is a naturally occurring assembly
of complex and heterogeneous organic compounds of varied
molecular weight, comprising of carboxylic acids, urea, and
hydrophobic compounds, collectively called humic substances
[1]. Humic substances have attracted burgeoning interests due to
their important role in geochemical and environmental processes
[2]. Measurements of humic substance are always important and
signicant, on account of its high complexity [3,4]. Conventional
analysis methods such as nuclear magnetic resonance, Fouriertransformed infrared spectroscopy, or microanalysis have provided
key information on the composition and presence of specic
functional groups. Nevertheless, chromatographic methodologies
have signicantly advanced our understanding of the distribution
of physicochemical properties in humic substances [510].
Recently, NOM has been effectively fractionated by size exclusive
chromatography (SEC). The aromaticity and hydrocarbon abundance in different mass ranges were successfully analyzed [1115].
The polydispersity of NOM samples based on molecular size
distributions (MSD) was also studied [12]. Besides the properties of
MSD and unsaturation degree of NOM matrices, their transformations (e.g., formation of disinfection by-product) are deemed
more in relation to their physicochemical properties [16]. The
distribution of hydrophobicity in humic substances has been
evaluated by reversed-phase HPLC (RPLC) [15,17]. However, further
information about the distribution of hydrophilic functionalities
was unattainable since RPLC columns retain the NOMs through
hydrophobic interactions. In addition, a quantitative study of
hydrophilic NOMs could not be achieved due to the low resolution
obtained on RPLC columns.
Hydrophilic interaction liquid chromatography (HILIC) has
been recently developed for separating polar compounds in food
and environmental analysis [18,19]. In a typical HILIC mode, polar
stationary phases are employed with the mobile phases similar to
those used in RPLC. The retention of polar analytes in HILIC usually
increases with increasing water content in the mobile phase, and
the selectivity attained is complementary to those in RPLC [20].
The HILIC stationary phases display high polarity, allowing for
strong interactions with hydrophilic analytes. The hydrophilic
interactions typically include hydrogen bonding, dipoledipole
stacking, ion-dipole interaction, and dipole-induced dipole interactions. Besides, relatively weaker intermolecular interactions
(i.e., Van der Waal forces and hydrophobic interactions) contribute
to the retention of analytes [21]. When a charged HILIC stationary
phase is used, a mixed HILIC/ion-exchange mechanism controls
the retention. Accordingly, the retention of ionic analytes is
dependent on both pH and water content of the mobile phase,
generating a specic selectivity. The mixed HILIC/ion-exchange
mechanism is also distinct from that in typical ion-exchange
chromatography (IC), because a large content of organic solvent
reduces the relative permittivity of the HILIC mobile phase.
Therein, the pKa of acidic compounds increases and the pKa of
basic compounds decreases [22]. Thus, the inuence of HILIC
interactions often predominates over the suppressed ion-exchange
effects. Consequently, raising the content of organic modier often
attenuates the retention of analytes in mixed HILIC/ion-exchange
separation mode. Due to enhanced separations attained in such
mode, the charged HILIC columns are often preferred for
separating highly complex matrices of natural products and
metabolites [23,24]. In HILIC mobile phases, insufcient organic
modier could lead to poor recovery rates of analytes from the
stationary phase. Therefore, high content of organic modier (e.g.,
6090% v/v) is usually required. In a mobile phase of low water
content, water is preferentially adsorbed onto the surface of polar
stationary phase, forming multiple water-enriched layers and

609

acting as part of the stationary phase. A partitioning process could

occur between the stationary phase and water-decient mobile
phase. Additionally, hydrogen bonding and electrostatic interaction contribute to the retention [25]. Recent progress in HILIC
fractionation has assisted in the structural elucidation work of
small natural products, pollutants, and toxin of environmental
concern [2629]. However, the reported HILIC methods focused on
the characterization of a single molecule or a group of natural
products bearing specic functionalities. On the other hand, the
HILIC methods reported for environmental studies mainly used
zwitterionic or neutral columns [20,27,28].
In this investigation, we report a HILIC method for effective
measurements of aquatic humic substances. Neutral and cationic
HILIC columns were selected for the separation of humic
substances. Since the humic substances in natural waters are
acidic, the cationic stationary phase could retain the samples with
electrostatic attraction, incorporated with hydrophilic interactions
[30]. The retention behavior of humic substances on the cationic
HILIC stationary phases was investigated by altering the composition and pH of the mobile phases.
2. Materials and methods
2.1. Reagents and solution preparation
Catechol (99%), pyrogallol (98%), hydroquinone (99%),
p-cresol (98%), resorcinol (99%), o-phenylenediamine (98%),
m-phenylenediamine (99%), p-phenylenediamine (99%), aniline (99%), N,N-diethyl-p-phenylenediamine (97%), benzylamine
(99%), benzoic acid (99.5%), phthalic acid (99%), 1,2,4-butanetricarboxylic acid (98%), and 1,2,3-benzenetricarboxylic acid (98%)
were obtained from SigmaAldrich. HPLC-grade methanol, diethylamine (>99%), and acetic acid (>99%) were purchased from Fisher
Scientic (UK). Diethylamine acetate (DEAA) buffer was prepared
by adding diethylamine into acetic acid solution (0.1% v/v DEAA),
and then ltered through a 0.2-mm membrane lter before use.
All water samples were ltered through a 0.45mm membrane
on the day of sampling, and stored at 4  C in a refrigerated tank.
Loire River was sampled upstream of the Belleville/Loire nuclear
plant (France) to avoid the presence of trace metals in the water.
Ribou Reservoir (i.e., used for potable water production) was
sampled upstream of the town of Cholet (France). Suwannee River
(Georgia, USA) is a very soft, black water river with low salt
content. The Jeddah (KSA) treated wastewater was collected from
Alrwais treatment plant (i.e., performing conventional activated
sludge treatment process). The humic substances were isolated
from the water samples by adsorption onto XAD-8 resin under
acid pH condition, and subsequently eluted using sodium
hydroxide solution [3]. The lyophilized humic substances were
characterized by elemental analysis and solid state 13C NMR
(supporting data). The humic substance from Suwannee River
shows higher C/N and C/H ratios, indicating its highly humic
character. The humic substance of treated wastewater contrasts
from those of natural waters due to its considerably higher content
of hydrocarbon (CC/C
H bond) but less content of carbohydrates (C
O bond). All the samples for HPLC analysis were
prepared by dissolving lyophilized humic substances in ultrapure
water to a nal concentration of 0.5 mg mL1. Diluted samples of
varied concentrations were prepared by diluting the sample
solution of 0.5 mg mL1 proportionally.
2.2. HPLC hardware and operation
The Breeze I-HPLC (Waters, USA) setup consisted of an isocratic
HPLC pump, an automated injector, and a multiple channel UVvis
detector. The column ZORBAX HILIC plus1 (ZORBAX, I.D. 4.6 mm

610

R.-Q. Wang et al. / Analytica Chimica Acta 853 (2015) 608616

Table 1
Calibration standards, structure of probe molecules, and chromatographic results.
Name (code)

Catechol (P1)

Structure

Retention time (min)

Peak width (min)

Tailing factor

Acidic
conditiona

Acidic
conditiona

Basic
conditiona

Acidic
conditiona

Basic
conditiona

Basic
conditiona

Column

2.22
2.89
6.62

2.17
2.98
7.18

0.56
0.16
0.70

0.66
0.12
0.66

0.81
1.00
1.22

1.42
1.00
1.44

IMPC
ZORBAX
IONPC

2.10
2.91
11.07

2.08
2.93
9.62

0.58
0.16
0.94

0.60
0.16
1.44

0.88
1.33
1.58

1.21
1.14
1.26

IMPC
ZORBAX
IONPC

Hydroquinone (P3)

2.10
2.89
6.13

2.10
2.91
6.12

0.56
0.20
0.54

0.76
0.12
0.52

0.81
1.12
1.17

1.00
1.00
1.17

IMPC
ZORBAX
IONPC

p-Cresol (P4)

2.40
2.89
6.50

2.40
2.94
6.50

0.50
0.22
1.14

0.66
0.18
0.98

0.93
1.50
2.71

1.30
1.00
2.59

IMPC
ZORBAX
IONPC

Resorcinol (P5)

2.15
2.88
6.60

2.13
2.92
7.03

0.56
0.18
0.60

0.68
0.12
0.64

0.92
1.13
1.06

1.25
1.17
1.13

IMPC
ZORBAX
IONPC

o-Phenylenediamine (B1)

2.28
3.16
11.20

2.28
3.15
5.65

0.56
0.40
35.0

0.54
0.16
0.64

0.93
0.83
2.21

1.50
1.00
1.17

IMPC
ZORBAX
IONPC

p-Phenylenediamine (B2)

1.82
3.74

2.18
3.39
5.28

0.56
0.22

0.54
0.16
0.72

0.78
1.10

1.25
1.07
0.91

IMPC
ZORBAX
IONPC

m-Phenylenediamine (B3)

2.15
3.22
9.47

2.17
3.20
5.52

0.54
0.50
7.36

0.48
0.46
1.18

0.86
1.19
2.03

1.40
0.83
0.91

IMPC
ZORBAX
IONPC

Aniline (B4)

2.43
3.09

2.38
3.10

0.68
0.38

0.92
0.38

0.86
0.95

1.71
1.00

IMPC
ZORBAX
IONPC

N,N-diethyl-pphenylenediamine (B5)

1.60
5.25

2.82
5.29

0.38
0.30

0.68
0.32

1.12
1.25

1.50
1.25

IMPC
ZORBAX
IONPC

Benzylamine (B6)

1.60
2.98

3.13
2.95

0.82
0.18

1.20
0.18

1.50
1.13

1.25
1.15

IMPC
ZORBAX
IONPC

Pyrogallol (P2)

R.-Q. Wang et al. / Analytica Chimica Acta 853 (2015) 608616

611

Table 1 (Continued)
Name (code)

Structure

Retention time (min)

Peak width (min)

Tailing factor

Acidic
conditiona

Acidic
conditiona

Basic
conditiona

Acidic
conditiona

Basic
conditiona

Basic
conditiona

Column

Benzoic acid (A1)

3.50
2.83

1.68
2.73

0.68
0.22

0.36
0.16

1.00
1.10

1.13
0.94

IMPC
ZORBAX
IONPC

Phthalic acid (A2)

4.77
2.73

1.53
2.66

0.84
0.18

0.32
0.12

0.95
1.17

1.17
1.20

IMPC
ZORBAX
IONPC

1,2,4-Butanetri-carboxylic
acid (A3)

4.38
2.69

1.45
2.62

0.22
0.18

0.28
0.14

1.05
1.00

1.33
1.00

IMPC
ZORBAX
IONPC

1,2,3-Benzenetri-carboxylic
acid (A4)

8.67
2.63

1.67
2.64

1.88
0.22

0.76
0.22

1.57
0.90

1.25
0.90

IMPC
ZORBAX
IONPC

a
The mobile phase is MeOH/buffer (0.1% DEAA) (80:20 v:v), ow rate 1.0 mL min1, column temperature 25  C, buffer pH 4.0 under acidic condition, buffer pH 9.0 under
basic condition.

 250 mm) based on neutral silica was purchased from Agilent

(USA). The polymer-based IonPac AS4A-SC Anion-Exchange1
Column (IONPC, I.D. 4.0 mm  250 mm) was purchased from
Thermo (USA). The Xbridge C18 column (I.D. 4.6 mm  150 mm)
was purchased from Waters (USA). The bonded cationic
b-cyclodextrin column (IMPC, I.D. 4.6 mm  150 mm) was prepared according to our previous study [30]. All the columns were
conditioned on the HPLC unit for 2 h after changing the mobile
phase conditions. All experiments were performed at 25  C and
with total ow rate of 1.0 mL min1. UV absorbance was measured
at 230 nm wavelength. The peak areas from 1 to 10 min were
considered for calculation because most of the investigated humic
substances samples were eluted in this range. The sample injection
volume was 10 mL. HILIC separation of humic substances was
performed on ZORBAX and IMPC columns under similar operational conditions. The chromatographic results were reproducible
for at least fty consecutive tests. The columns were eluted under
basic conditions for 30 min (buffer DEAA 0.1%, pH 9) between
injections.

3. Results and discussion

3.1. Assessment of columns with probe molecules
The chromatographic properties of HILIC columns (i.e., ZORBAX,
IMPC, and IONPC) were compared using 15 selected probe
molecules, including 5 phenols, 4 carboxylic acids, and 6 amines
(Table 1). Most of these chemicals contain a phenyl group which
allows UV detection. As summarized in Table 1, the retention times
of phenols (P1P5) on all columns were not inuenced by the pH of
the mobile phase. These compounds were strongly retained on the
polymer-based IONPC column, whilst lower retention times were
observed on silica-based IMPC and ZORBAX columns. Additionally,
phenols showed stronger interactions with IMPC and IONPC
columns, evidenced by wider chromatographic peaks and longer
peak tailing.
On IONPC column, most of the ionic analytes (A1A4 and
B4B6) failed to be eluted within the analysis time range. The
B1B3 basic compounds were also strongly adsorbed on IONPC,

Table 2
Linear correlations calculated between the humic substances concentration and HPLC peak area.
Sample origin

Buffer pHa

Columnb
ZORBAX
1

a (10
Loire River
Ribou Reservoir
Suwannee River
Jeddah Waste water

4
9
4
9
4
9
4
9

1.60
2.20
1.70
2.10
2.00
2.00
6.80
8.40

IMPC
)

3

b (10
0.6
10.0
0.1
2.3
4.0
4.0
1.5
1.3

1.00
1.00
1.00
1.00
0.98
1.00
0.99
0.98

a (101)

b (103)

R2

0.21
1.60
0.22
1.80
0.05
2.30
0.19
1.10

0.9
3.2
0.2
6.4
2.3
2.5
0.6
5.6

0.98
0.99
0.97
0.98
0.94
1.00
0.88
1.00

a
The mobile phase is MeOH/buffer (0.1% DEAA) (80:20 v:v), ow rate 1.0 mL min1 undened, column temperature 25  C, buffer pH 4.0 under acidic condition, buffer pH
9.0 under basic condition.
b
Linear regression equation: Y = aX + b; Y: peak area; X: concentration of humic substances (mg mL1); a: slope; b: intercept on Y axis; R2 is determination coefcient.

612

R.-Q. Wang et al. / Analytica Chimica Acta 853 (2015) 608616

evidenced by peak broadening. Therefore, IONPC is inappropriate

for separation of ionic organic compounds in the HILIC mode. The
retention of ionic analytes on ZORBAX was modest, but altering
buffer pH exerted very limited impact on their retention behavior.
In contrast, all the ionic compounds were eluted out of IMPC
column rapidly, and the chromatographic results indicated the
inuence of pH was obvious. When undissociated, the acid
analytes showed higher retention and larger peak widths under
acidic conditions, which favored the formation of hydrophobic
inclusions and hydrogen bonding with the stationary phase [30].
On the contrary, the basic compounds showed signicantly longer
retention times and wider peaks under basic conditions, except for
o-phenylenediamine (B1) and m-phenylenediamine (B3). It is
possible that the vicinal amino substituents form intramolecular
hydrogen bonding, consequently reducing their basicity. Thus,
IMPC column is pH-sensitive during ionic compounds separation,
which also affords selective hydrophilic retention. Conclusively,
alternative selectivities could be attained on IMPC column by
changing the composition and pH of mobile phases.
3.2. Calculation methods of recovery rates, statistical retention time,
and disparity
A ve-point calibration curve of each humic substances sample
was obtained by linear regression analysis at basic and acidic
conditions (Table 2). The chromatographic peak area was linearly
correlated to the concentration of the sample in solution (i.e.,
0.120.50 mg mL1), especially in the high concentration range.
Thus, the sample solution was prepared with a high concentration
(i.e., 0.5 mg mL1) to reduce the absolute experimental error. The
humic substances from natural waters showed similar correlation
slopes of UV peak area versus concentration (i.e., 0.160.20 at pH 4;
0.200.22 at pH 9). On the other hand, plots with much higher
slopes were observed for the sample from wastewater (i.e., 0.68 at
pH 4 and 0.84 at pH 9). It can be inferred that this wastewater
sample had a higher content of unsaturated molecules than the
natural water samples. The wastewater contains higher content of
anthropogenic and natural organic compounds, baring chromophoric groups respectively related to the unstable (oxydizable)
or condensed state of organic matter. Therefore, during the evolution
from raw wastewater to surface water, the UV absorbance of water
sample generally decreases, which is mainly related to quality
modication, especially the biodegradation process [32].
As aforementioned, the retention of most compounds on
ZORBAX column is marginally inuenced by pH of mobile phases
(Table 1). Additionally, the UV peak areas of humic substances from
ZORBAX column were nearly unchanged when they were
separated under different pH conditions (Table 2). As such, the
chromatographic peak area of a humic substances sample obtained
from ZORBAX column could be set as a reference for calculating the
recovery rate (h%) of other columns under the same chromatographic condition (Eq. (1)).

h%

PeakAreaIMPC
 100%
PeakAreaZORBAX

present at low concentrations, corresponding to small peak areas.

However, since the humic substances matrix contains a considerably large number of compounds, the selectivity calculated by the
conventional method is unattainable. Alternatively, the disparity of
physicochemical property (Dis) amongst compounds in humic
substances is dened as the deviation of the retention time (tR) of
each fraction from statistical retention time of the whole humic
substance material (Eq. (3)).
v
!
Z 10 u
u t R  t R 2
PeakAreatR
t
(3)
Dis
d
PeakAreatotal
tR
1
For example, t R and Dis of mixtures made of randomly chosen
probe molecules were calculated. As illustrated in Fig. 1(a), a
four-compound mixture of p-cresol (P4), benzoic acid (A1),
phthalic acid (A2), and 1,2,3-benzenetri-carboxylic acid (A4) has
a disparity of 0.49. The modest statistical retention time of
5.73 min represents the integral physicochemical feature of the
whole mixture. Contrastively, m-phenylenediamine (B3) has a
much shorter retention time of 2.15 min (Fig. 1(b)). Thus, the
addition of such compound into the mixture would afford a shorter
statistical retention time of 5.23 min. Meanwhile, the increase of
disparity from 0.49 to 0.55 indicates that the new component of
m-phenylenediamine (B3) has a very different physicochemical
property from the overall property of the former four-compound
mixture, resulting in a higher complexity of the ve-compound
mixture. Additionally, hydroquinone (P3) has a shorter retention
time of 2.10 min, and the addition of such compound into the
mixture reduced the overall statistical retention time towards
5.06 min (Fig. 1(c)). Similarly, the disparity of the six-compound
mixture (0.58) is higher than the former ve-compound mixture
(0.55), which indicated that the six-compound mixture is more
complex and heterogeneous. Conclusively, if the new compounds
introduced into a mixture have very different retention times

(1)

Statistical retention time t R was dened as the weighted mean

value of the retention time (tR) of each fraction in the range from
1 to 10 min, therefore showing the average retention time of the
whole sample (Eq. (2)).
Z 10
PeakAreatR
(2)
tR
tR d
PeakAreatotal
1
Humic substances consist of numerous compounds. However,
the absence of sharp peaks in the 13C NMR and FT-IR spectra of the
whole sample (results not shown) indicated that no specic
compound was predominant. In other words, each compound was

Fig. 1. Statistical retention time t R and disparity (Dis) calculated for model
systems consisting of probe molecules (mobile phase: MeOH: 0.1% DEAA (pH
4) 80:20 (v:v); stationary phase: IMPC).

R.-Q. Wang et al. / Analytica Chimica Acta 853 (2015) 608616

compared to the statistical retention time of the whole mixture,

they would result in a higher disparity of the mixture, representing
higher complexity. If the retention time of the newly added
compounds have shorter retention time than the overall statistical
retention time of the original mixture, the statistical retention time
of the newly made mixture would be decreased and vice versa.
Additionally, it is worth to mention that the statistical retention
time and disparity of mixtures were calculated simply based on
UVvis signals, assuming the compounds in mixtures have similar
UV absorbance. Although UVvis detector is most widely adapted
on HPLC in the environmental analysis, the statistical retention
time and disparity could be calculated to quantitatively dene the
overall physicochemical property and complexity of mixtures
when other detectors are used.
3.3. Investigation methods of ionic compounds in humic substances
The inuence of pH on the separation of ionic compounds in
humic substances was investigated in pH range from 4 to 10. Similar
to the retention behavior of the probe molecules (Table 1), the
retention of humic substances on IONPC column was considerably
higher than on ZORBAX and IMPC columns, which resulted in poor
recovery rates of less than 5% from IONPC column. The recovery rates
of humic substances from IMPC column were strongly inuenced by
pH of the mobile phases. The statistical retention time t R and

613

disparity of physicochemical property (Dis) calculated from IMPC

and ZORBAX chromatograms are summarized in Table 3.
The highest recovery rates from IMPC column were observed at
pH 9 for all the humic substances except for the Suwannee River
humic substance. When the pH of the solution increased beyond 9,
the dissociation of basic functional groups (pKa > 9) was suppressed, resulting in weaker repulsive electrostatic force with the
cationic stationary phase. Thus, most of the basic compounds
showed longer retention times, indicating stronger afnity with
the stationary phase. Apparently, substantial amounts of basic
compounds exist in the aquatic humic substances. This conclusion
is also in good accordance with previous literature reports [17,31].
Similarly, the dissociation of other ionic compounds (i.e., pKa < 9)
could have also been suppressed under low pH conditions, leading
to stronger attractive hydrophilic interactions with the stationary
phase. As a result, the recovery rates were gradually reduced from
pH 9 to 4. On the other hand, the shorter statistical retention time
at pH 9 always conduce to lower disparity (Dis) of the humic
substances measured on IMPC (Table 3). It has been previously
concluded that IMPC could effectively provide multi-model
interactions with analytes [30]. Therefore, it is not surprising that
IMPC showed higher disparities than the neutral ZORBAX column.
Additionally, the statistical retention times t R of all the studied
humic substances on ZORBAX reached their maximal at pH 5,
whilst the hydrophilic property disparities (Dis) are mostly at

Table 3
Evaluation of ionic compounds in humic substances from different origins.
Sample origin

Column

pH

t R (min)

Disparity (101)

Recovery (h%)

Loire River

IMPC

10
9
7
5
4
10
9
7
5
4

2.75
2.53
3.32
3.76
5.12
2.58
2.64
2.65
3.19
3.00

7.63
5.56
5.96
4.80
3.36
3.56
2.93
1.49
3.37
3.15

45.3
72.9
62.4
30.1
10.7

10
9
7
5
4
10
9
7
5
4

2.15
2.14
3.25
3.74
4.45
2.57
2.48
2.73
3.34
2.94

7.89
5.38
5.92
4.87
4.28
3.04
1.23
1.80
3.12
2.69

26.2
97.7
31.9
19.7
8.6

10
9
7
5
4
10
9
7
5
4

2.22
2.30
3.99
4.27
5.11
2.73
2.87
2.82
3.79
2.89

7.90
4.18
5.67
4.99
2.90
3.98
4.65
2.18
3.83
3.66

49.7
39.8
40.8
19.9
3.3

10
9
7
5
4
10
9
7
5
4

2.63
2.17
2.95
3.84

2.44
2.44
2.60
3.15
2.95

8.00
5.11
5.91
4.46

2.17
1.68
1.55
2.54
2.55

44.9
85.0
49.8
37.8

ZORBAX

Ribou Reservoir

IMPC

ZORBAX

Suwannee River

IMPC

ZORBAX

Jeddah waste water

IMPC

ZORBAX

The mobile phase is MeOH/buffer (0.1% DEAA) (80:20 v:v), ow rate 1.0 mL min1, column temperature 25  C.

614

R.-Q. Wang et al. / Analytica Chimica Acta 853 (2015) 608616

minimal at pH 7. Two types of silanol groups on fused-silica

surfaces had pKa values of 4.9 (isolated silanol group SiOH) and
8.5 (vicinal silanol group 
SiO
Si) [32]. A maximum
retention of humic substances occurred at pH 5, specically when
the stationary phase attained positive charges by the vicinal silanol
groups. Therefore, it is considered that positively-charged stationary phase interact through attractive electrostatic forces with the
humic substances. From pH 5 to 9, the isolated silanol group could
generate hydrogen bonding with the vicinal Si
OSi
 on
the stationary phase. As a result, signicant decreases of both t R
and Dis were observed. Thus, the attractive electrostatic forces are
vital in the separation of the humic substances compounds in this
case. The representative chromatograms of Loire River humic
substance are shown in Fig. 2. At pH 4, IMPC exerted a very strong
interaction with humic substances, which resulted in a negative
peak before 2 min [12]. Therefore, the recovery rate and t R
calculated at pH 4 could be underestimated.

the inuences of ionic interactions. The approach for the

interpretation of the separated peaks followed in this investigation
is as follows: (i) in fraction-A (1.01.9 min), the compounds were
eluted earlier with increased ratio of buffer in the mobile phases
(Fig. 3c and d). When the mobile phase was substantially organic in
composition (i.e., MeOH exceeded 75%, Fig. 3a and b), water was
preferentially adsorbed onto the surface of the stationary phase
[25]. The compounds in fraction-A would be attracted onto the
stationary phase by weak electrostatic forces, where ion-pairing
mechanism could be envisaged. Electrostatic interactions were
easily disrupted by increased water content in the mobile phase.
(ii) In fraction-B (1.92.2 min), the retention of compounds was
weakly inuenced by the large variation of mobile phase
composition. The compounds (i.e., hydrophobic macromolecules)
would be retained through weak Van der Waal interaction, size
exclusion, and steric hindrance. Apparently, the compounds in
fraction-A were well separated from those in fraction-B in the
mobile phase with 25% water (Fig. 3c). (iii) In fraction-C
(2.210.0 min), compounds attained longer retention time when
increasing the water content of the mobile phase. These
compounds could have been retained through HILIC interactions
such as hydrophobic interactions, hydrogen bonding, dipole-dipole
stacking, and ion-dipole interaction. The results indicate that it is
more convenient to use high content of organic modier (e.g.,
MeOH > 75%) to elute the humic substances for the effective
determination of their compositions.
This optimized method was used for characterization of the
aquatic humic substances from different origins. The HILIC
chromatographic peak areas of different fractions are summarized
in Table 4. Amongst the four investigated samples, Loire River
showed the lowest content of dissociated compounds at pH 9 (i.e.,
44.0% in fraction-A) whereas Ribou Reservoir sample displayed the
highest ratio (64.3%). On the other hand, the hydrophobic
macromolecules eluted out in fraction-B accounted for approximately 25% of the Loire River humic substance, and only 8.9% of the
Suwannee River sample. Generally, 2030% of the humic substance
materials were eluted out in fraction-C. More specically, humic
substances from rivers contained a larger proportion of compounds belonging to fraction-C than the humic substances from
wastewater and reservoir water.

3.4. Investigation methods of hydrophilic properties of humic

substances

3.5. Evaluation of the compositions of humic substances from different

origins

The inuence of an organic modier on the elution behavior of

Loire River humic substance was investigated at pH 9 to attenuate

Because of longer retention at pH 4, the variation in recovery

rates from IMPC (i.e., chromatograms acquired at pH 5 and 4) could
be assigned to the acids (pKa 45) retained in the column. Similarly,
the amounts of ionic compounds in three fractions were
calculated, i.e., acids (pKa 45), compounds with intermediate
pKa (59), and basic compounds (pKa 910). These three distinct
groups in humic substances are key for metal or proton binding
[32]. Both small-size neutral compounds and hydrophobic macromolecules were recognized as compounds with intermediate pKa
(59), as they do not show either acute acidity or basicity.
Specically, the amount of hydrophobic macromolecules (i.e.,
compounds in fraction-B) was found by altering the hydrophobicity of the mobile phases at pH 9 (Table 4). The measured
compositions of humic substances from different origins are
summarized in Fig. 4. The properties of ionic compounds in the
whole samples were evaluated by statistical retention time and
disparity at pH 9, specically when most of the humic substance
samples have attained the maximum recovery rates (Table 3).

Fig. 2. HILIC chromatogram of Loire River humic substances on ZORBAX and IMPC
columns at different pH conditions. The analyte has attained better separations on
IMPC, displaying lower and broader chromatographic peaks. (Mobile phase: MeOH:
0.1% DEAA 80:20 (v:v)).

Fig. 3. HILIC chromatogram of Loire River humic substances on IMPC. The

compounds are divided into three distinct fractions. From (a) to (d): fraction-A, elution
delayed; fraction-B, retention unchanged; fraction-C, elution advanced. Mobile Phase:
MeOH: 0.1% DEAA (pH 9) (v:v), (a) 25:75 (b) 50:50 (c) 75:25 (d) 80:20.

3.5.1. Loire River humic substances

The carboxylic acids (pKa 45) and basic compounds (pKa 910)
accounted for 19.4% and 27.6% of the whole matrix, respectively. On

R.-Q. Wang et al. / Analytica Chimica Acta 853 (2015) 608616

615

Table 4
Compositions of humic substances determined by hydrophilic properties.
Sample origin

Peak area percentage (%)

Fraction-A (1.01.9 min)

Fraction-B (1.92.2 min)

Fraction-C (2.210.0 min)

Loire River
Ribou Reservoir
Suwannee River
Jeddah Waste water

44.0
64.3
60.5
58.5

25.0
14.9
8.9
20.4

31.0
20.8
30.6
21.1

The mobile phase is MeOH/buffer (0.1% DEAA, pH 9) (75:25 v:v), ow rate 1.0 mL min1, column temperature 25  C.

the other hand, this sample contained approximately 25.0% of

hydrophobic macromolecules, which lacked of HILIC interactions
with the stationary phase. The rest of the humic substance
compounds (28.0%) can be considered as weakly-charged compounds with an intermediate pKa from 5 to 9 (e.g., phenol), which
was the highest amongst the four humic substances investigated.
The ionic neutral compounds could generate stronger hydrophobic
interactions with the stationary phase [30]. As a result, the Loire
River humic substance had a statistical retention time t R of 2.5 min
and disparity (Dis) of 5.6 on IMPC. Both factors were the highest
amongst the four samples. According to the 13C NMR spectrum
(supporting data), carboxylic groups represented only 11% of the
total organic carbon, which was lower than the amount of acids
detected by the HILIC method. The reason might be due to the
negative peak caused by the intensive adsorption of the humic
substances [12]. As a result, the recovery rate of acids from IMPC
column at pH 4 was underestimated, leading to higher proportion
of carboxylic acids in the HILIC measurements. Additionally, the 13C
NMR spectrum of Loire River humic substance justied its
hydrophobic character with a large amount of hydrocarbon
(49%). The result is close to the accumulative abundance of
weakly-charged compounds and hydrophobic macromolecules
determined by HILIC method (53%).
3.5.2. Ribou Reservoir humic substances
Ribou Reservoir humic substances matrix contained 11.1% of
carboxylic compounds. Additionally, it contained 71.5% of compounds bearing basic moieties, which was the highest amongst the
four samples. The elemental analysis results of Ribou humic
substance also showed the lowest C/N ratio (18.2%) amongst all
investigated samples, indicating high content of amine (supporting

Fig. 4. Compositions of the humic substances of different origins measured by

HILIC analysis method. The proportions of acids and hydrophobic macromolecules
were measured for Suwannee River humic substances. The acids, basic compounds,
hydrophobic macromolecules, and hydrophilic neutral compounds were measured
for other samples.

data). Additionally, Ribou Reservoir humic substances also contained

14.9% of hydrophobic macromolecules and 2.5% of ionic neutral
compounds with pKa from 5 to 9. Ribou Reservoir had a statistical
retention time t R of 2.1 min and a disparity (Dis) of 5.4 on IMPC.
Although the statistical retention time was the lowest amongst all
the samples, the column showed a strong separation capability.
Considering that this sample contained a substantial amount of
basic compounds, which could generate repulsive electrostatic
interactions with the cationic stationary phase, it is not surprising
to observe a high disparity with a relatively low retention time.
3.5.3. Suwannee River humic substances
Carboxylic acids accounted for 16.6% of Suwannee River sample,
which was consistent to the result obtained from 13C NMR
characterization (i.e., 16%, supporting data). The highest recovery
rate of humic substances was observed at pH 10, indicating that a
substantial amount of hydrophilic molecules were deprotonated
by basic additives in the mobile phase. On the other hand,
hydrophobic macromolecules accounted for 8.9%, which was the
lowest in all the investigated samples. Suwannee River humic
substances displayed a modest statistical retention time of 2.3 min,
and the lowest disparity (Dis 4.2). Because of the highly humic
nature of Suwannee River, its compounds could be more
homogenous in nature through various biochemical reactions,
whereas the hydrophobic macromolecules would have been
decomposed into smaller molecules.
The sample of Suwannee river humic substances (HPO) has
been widely employed as a standard sample in the environmental
science community, and reversed phase C18 columns are available
to most environmental chemists. Therefore, Suwannee river HPO
was also separated with a C18 column (Fig. SD4). As illustrated in
table SD3, the C18 column affords much less retention of the
Suwannee river HPO (i.e., t R <1.6 min) than HILIC columns (i.e.,
t R > 2.2 min). Because C18 columns mainly provide hydrophobic
interactions and the aquatic humic substances from Suwannee
river are highly hydrophilic, it is not surprising to nd the sample is
less retained on the C18 column than on HILIC columns.
Meanwhile, since the compounds in Suwannee HPO were simply
separated based on their hydrophobicities, it is found that the
disparity (Dis) of such sample measured by C18 column is much
lower (i.e., Dis < 1.5) than that evaluated by HILIC columns (i.e.,
Dis > 2.1). Conclusively, aquatic Suwannee HPO is better measured
with HILIC columns.
3.5.4. Jeddah waste water humic substances
Jeddah waste water consisted of 40.1% of basic compounds, and
37.8% of acids. Macromolecules accounted for 20.4%. The remaining
fraction (1.7%) could be assigned to ionic neutral compounds with
intermediate pKa from 5 to 9, which was signicantly lower than
the samples of terrestrial origins. The 13C NMR spectrum of the
wastewater humic substance showed that carboxylic groups
accounted for merely 12% of the total organic carbon (Supporting
data). Accordingly, a variety of acids other than carboxylic acids
(e.g., sulfonic acid, phosphoric acid, etc.) were present in the waste
water humic substances. The treated wastewater exhibited
considerably higher content of both basic compounds and

616

R.-Q. Wang et al. / Analytica Chimica Acta 853 (2015) 608616

carboxylic acids than the humic substances from rivers. The

statistical retention time of humic substances in the wastewater
was 2.2 min, which was similar to those of the samples from
Suwannee River and Ribou Reservoir. However, the disparity was
5.1, which was comparable to the water sample from Ribou
reservoir. Nonetheless, the composition of Jeddah wastewater
humic substances differed from Ribou reservoir due to its higher
content of acid compounds.
4. Conclusions
HILIC method was successfully employed in measuring the
distribution of physicochemical properties of humic substances
from rivers, reservoir, and wastewater. The analysis was convenient and consistent. Both aqueous humic substances from Ribou
Reservoir and Jeddah treated wastewater showed higher amounts
of basic compounds than the samples from rivers, while Jeddah
treated wastewater also exhibited a high content of acids. In
addition, Suwannee River humic substances displayed the lowest
disparity within the sample, indicating that it was more
homogeneous than the samples of other origins.
Acknowledgments
We thank Ms. Zhan Tong, Mr. Ali Imran for laboratory assistance,
Dr. Zhang Tao for the instrument maintenance, and all the other
colleagues working together in the same lab. The authors are
grateful to the funding from King Abdullah University of Science
and Technology (KAUST).
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.aca.2014.09.026.
References
[1] L.M. Surhone, M.T. Tennoe, S.F. Henssonow, Natural Organic Matter, Betascript
Publishing, Saarbrcken, Germany, 2010.
[2] N. Senesi, T.M. Miano, Humic substances in the global environment and
implications on human health, Proceedings of the 6th International Meeting of
the International Humic Substances Society, Amsterdam, Netherland, Elsevier,
1994, 2014.
[3] J.-P. Croue, G.V. Korshin, M.M. Benjamin, A.R. Foundation, Characterization of
Natural Organic Matter in Drinking Water, AWWA Research Foundation and
American Water Works Association, U.S.A, 1999.
[4] K. Kumada, Chemistry of Soil Organic Matter, Elsevier, Amsterdam,
Netherlands, 1988.
[5] S.A. Huber, A. Balz, M. Abert, W. Pronk, Characterization of aquatic humic and
non-humic matter with size-exclusion chromatographyorganic carbon
detectionorganic nitrogen detection (LCOCDOND), Water Res. 45 (2011)
879885.
[6] A.I. Schfer, A. Pihlajamki, A.G. Fane, T.D. Waite, M. Nystrm, Natural organic
matter removal by nanoltration: effects of solution chemistry on retention of
low molar mass acids versus bulk organic matter, J. Membr. Sci. 242 (2004)
7385.
[7] C. Jarusutthirak, G. Amy, J.-P. Crou, Fouling characteristics of wastewater
efuent organic matter (EfOM) isolates on NF and UF membranes, Desalination 145 (2002) 247255.
[8] T. Reemtsma, A. These, On-line coupling of size exclusion chromatography
with electrospray ionization-tandem mass spectrometry for the analysis of
aquatic fulvic and humic acids, Anal. Chem. 75 (2003) 15001507.

[9] A. Piccolo, M. Spiteller, Electrospray ionization mass spectrometry of

terrestrial humic substances and their size fractions, Anal. Bioanal. Chem.
377 (2003) 10471059.
[10] L. Persson, T. Alsberg, G. Kiss, G. Odham, On-line size-exclusion chromatography/electrospray ionisation mass spectrometry of aquatic humic and fulvic
acids, Rapid Commun. Mass Spectrom. 14 (2000) 286292.
[11] N. Her, G. Amy, D. McKnight, J. Sohn, Y. Yoon, Characterization of DOM as a
function of MW by uorescence EEM and HPLCSEC using UVA, DOC, and
uorescence detection, Water Res. 37 (2003) 42954303.
[12] N. Her, G. Amy, D. Foss, J. Cho, Y. Yoon, P. Kosenka, Optimization of method for
detecting and characterizing NOM by HPLC-size exclusion chromatography
with UV and on-line DOC detection, Environ. Sci. Technol. 36 (2002)
10691076.
[13] H.M. Szabo, T. Tuhkanen, The application of HPLCSEC for the simultaneous
characterization of NOM and nitrate in well waters, Chemosphere 80 (2010)
779786.
[14] F.C. Wu, R.D. Evans, P.J. Dillon, Y.R. Cai, Rapid quantication of humic and fulvic
acids by HPLC in natural waters, Appl. Geochem. 22 (2007) 15981605.
[15] K. Namjesnik-Dejanovic, S.E. Cabaniss, Reverse-phase HPLC method for
measuring polarity distributions of natural organic matter, Environ. Sci.
Technol. 38 (2004) 11081114.
[16] M. Kitis, T. Karanl, A. Wigton, J.E. Kilduff, Probing reactivity of dissolved
organic matter for disinfection by-product formation using XAD-8 resin
adsorption and ultraltration fractionation, Water Res. 36 (2002) 38343848.
[17] P.K. Egeberg, J.J. Alberts, Determination of hydrophobicity of NOM by RP-HPLC,
and the effect of pH and ionic strength, Water Res. 36 (2002) 49975004.
[18] A. Combes, S. El Abdellaoui, C. Sarazin, C.J. Vial, A. Mejean, O. Ploux, V. Pichon,
Validation of the analytical procedure for the determination of the neurotoxin
b-N-methylamino-L-alanine in complex environmental samples, Anal. Chim.
Acta 771 (2013) 4249.
[19] C. Chiaochan, U. Koesukwiwat, S. Yudthavorasit, N. Leepipatpiboon, Efcient
hydrophilic interaction liquid chromatographytandem mass spectrometry
for the multiclass analysis of veterinary drugs in chicken muscle, Anal. Chim.
Acta 682 (2010) 117129.
[20] D.V. McCalley, Is hydrophilic interaction chromatography with silica columns
a viable alternative to reversed-phase liquid chromatography for the analysis
of ionisable compounds? J. Chromatogr. A 1171 (2007) 4655.
[21] R. Kaliszan, QSRR: quantitative structure-(chromatographic) retention relationships, Chem. Rev. 107 (2007) 32123246.
[22] T.N. Bui, J.J. Verhage, K. Irgum, Tris(hydroxymethyl) aminomethane-functionalized silica particles and their application for hydrophilic interaction
chromatography, J. Sep. Sci. 33 (2010) 29652976.
[23] M.A. Strege, S. Stevenson, S.M. Lawrence, Mixed-mode anioncation exchange/
hydrophilic interaction liquid chromatographyelectrospray mass spectrometry as an alternative to reversed phase for small molecule drug discovery,
Anal. Chem. 72 (2000) 46294633.
[24] B. Buszewski, S. Noga, Hydrophilic interaction liquid chromatography (HILIC)
a powerful separation technique, Anal. Bioanal. Chem. 402 (2012) 231247.
[25] K. Kosaka, M. Asami, Y. Konno, M. Oya, S. Kunikane, Identication of antiyellowing agents as precursors of N-nitrosodimethylamine production on
ozonation from sewage treatment plant inuent, Environ. Sci. Technol. 43
(2009) 52365241.
[26] G.C. Woods, M.J. Simpson, P.J. Koerner, A. Napoli, A.J. Simpson, HILICNMR:
towards the identication of individual molecular components in dissolved
organic matter, Environ. Sci. Technol. 45 (2011) 38803886.
[27] J. Knnemeyer, L. Terborg, B. Meermann, C. Brauckmann, I. Mller, A. Scheffer,
U. Karst, Speciation analysis of gadolinium chelates in hospital efuents and
wastewater treatment plant sewage by a novel HILIC/ICPMS method,
Environ. Sci. Technol. 43 (2009) 28842890.
[28] P. Ciminiello, C. Dell'Aversano, E. Dello Iacovo, E. Fattorusso, M. Forino, L.
Grauso, L. Tartaglione, F. Guerrini, L. Pezzolesi, R. Pistocchi, S. Vanucci, Isolation
and structure elucidation of ovatoxin-a the major toxin produced by ostreopsis
ovate, J. Am. Chem. Soc. 134 (2011) 18691875.
[29] R.-Q. Wang, T.-T. Ong, W. Tang, S.-C. Ng, Cationic cyclodextrins chemicallybonded chiral stationary phases for high-performance liquid chromatography,
Anal. Chim. Acta 718 (2012) 121129.
[30] O. Thomas, F. Theraulaz, C. Agnel, S. Suryani, Advanced UV examination of
wastewater, Environ. Technol. 17 (1996) 251261.
[31] M. Takcs, J.J. Alberts, P.K. Egeberg, Characterization of natural organic matter
from eight Norwegian surface waters: proton and copper binding, Environ. Int.
25 (1999) 315323.
[32] H.-F. Fan, F. Li, R.N. Zare, K.-C. Lin, Characterization of two types of silanol
groups on fused-silica surfaces using evanescent-wave cavity ring-down
spectroscopy, Anal. Chem. 79 (2007) 36543661.