WATER POLLUTION CONTROL RESEARCH SERIES
16020 FEN 03/71
Characterization and Separation
of Secondary Effluent Components
by Molecular Weight
o
VS. ENVIRONMENTAL PROTECTION AGENCY
-------
WATER POLLUTION CONTROL RESEARCH SERIES
The Water Pollution Control Research Series describes the
results and progress in the control and abatement of pollu-
tion of our Nation's waters. They provide a central source
of information on the research, development, and demon-
stration activities of the Environmental Protection Agency
through inhouse research and grants and contracts with
Federal, State, and local agencies, research institutions,
and industrial organizations.
Inquiries pertaining to the Water Pollution Control Research
Reports should be directed to the Head, Publications Branch,
Research Information Division, R&M, Environmental Protection
Agency, Washington, D.C. 20460.
-------
CHARACTERIZATION AND SEPARATION OF SECONDARY EFFLUENT
COMPONENTS BY MOLECULAR WEIGHT
by
Arthur D. Little, Inc.
Cambridge, Massachusetts 02140
for the
ENVIRONMENTAL PROTECTION AGENCY
Project #16020 FEN
Contract #14-12-886
March, 1971
For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington, D.C. 20402 - Price 55 cents
-------
EPA Review Notice
This report has been reviewed by the Water
Quality Office, EPA, and approved for publi-
cation. Approval does not signify that the
contents necessarily reflect the views and
policies of the Environmental Protection
Agency, nor does mention of trade names or
commercial products constitute endorsement
or recommendation for use.
ii
-------
ABSTRACT
Membrane ultrafiltration with a 1000 MW cut-off membrane is an
effective means of separating the high and low molecular weight
fractions of effluent and provides useful analytical samples
for further study. Solvent extraction techniques and thin layer
chromatography show promise as appropriate methods for isolating
and resolving the low molecular weight effluent subfraction.
While infrared spectrometry yields good spectra of various effluent
fractions, most are still much too complex to be able to infer much
specific information from the spectra of the complex mixture. The
same is true of ultraviolet spectrometry. High resolution mass
spectrometry (HUMS) shows promise for being able to obtain a great
deal of specific chemical information even from still very complex
mixtures.
Using HRMS the following classes of compounds have been identified
in a solvent extract of the low molecular weight effluent fraction:
fatty acids, aromatic acids, alkylphenyl-
ethylene oxides, chlorinated phenols,
aliphatic and aromatic amines, and amides.
This report was submitted in fulfillment of Project Number 16020FEN,
Contract 14-12-886, under the sponsorship of the Water Quality Office,
Environmental Protection Agency.
iii
-------
CONTENTS
Section
I
II
III
IV
V
VI
VII
VIII
IX
X
XI
Conclusions
Recommendations
Introduction
General Sample Preparation Approach
Comparison of the Solids in Brockton
and Rockland Effluent
Separation Based on Molecular Size
Solvent Extraction
Thin Layer Chromatography
Page
1
3
5
7
9
17
29
33
High Resolution Mass Spectrometry
(HRMS) of Low Molecular Weight Fractions 35
Acknowledgment a
References
111
-------
FIGURES
Page
1 Sample Preparation Procedure 8
2 Infrared Spectrum of Total Brockton
Solids (Sample 4AS) n
3 Infrared Spectra of Solvent Extracts
of Brockton Solids ]_o
A Schematic Representation of GPC System 22
5 GPC Curves on Sephadex G-10 23
6 GPC Curves on Sephadex G-25 of Brockton
Effluent after Ultrafiltration
(Millipore) 2U
7 Brockton Sample 11A — GPC Curves on
Sephadex G-25 25
8 GPC Curves on BioGel P-30 of Brockton
Sample 4A 26
9 GPC Curves on Brockton Sample on
Coming-Glass CPG-10 27
10 GPC Curves for Brockton Effluent from
Eastman Membrane Ultrafiltration 28
11 Solvent Extraction Procedure 30
vi
-------
TABLES
No. Page
1 Infrared and Ultraviolet Spectra of
Solvent Extractable Organics 31
2 TLC Behavior of CHLORH Fraction 3)4
3 Summary of HRMS Results from Solvent Extracts
of As-Is Sample 33
4 Summary of HRMS Results from Solvent Extracts
of Acidified Sample 39
5 HRMS Results on TLC Separated CHLORH Sample I^Q
vii
-------
SECTION I
CONCLUSIONS
1. Municipal secondary waste effluent consists of a variety of
chemical species ranging from simple molecules to high poly-
mers. Very little is known at the present about the specific
chemical composition of these species.
2. Membrane ultrafiltration is an effective means of separating
the high and low molecular weight fractions of the effluent
for further analytical studies.
3. Solvent fractionation and thin layer chromatographic sepa-
ration of the low molecular weight fraction provides suitable
samples for chemical identification.
4. High resolution mass spectrometry is a very effective means
of obtaining a specific identification of the many chemical
species present in the isolated analytical fractions.
-------
SECTION II
RECOMMENDATIONS
The purpose of this program was to explore the appropriateness of var-
ious methods for separating and characterizing municipal secondary
waste effluent. After examining several methods a combination of mem-
brane ultrafiltration, solvent extraction, thin layer chromatography,
and high resolution mass spectrometry was found to be most appropriate.
The methodology developed during this study should be applied to a
comprehensive study of the low molecular weight effluent species.
Each of the solvent extracts should be separated by thin layer chroma-
tography for chemical identification of the component mixtures by
high resolution mass spectrometry. The aqueous phase should also be
examined for any remaining organic species.
Derivative formation techniques, such as silylation or acetylation,
should be studied for their appropriateness in examining the non-
volatile effluent species.
-------
SECTION III
INTRODUCTION
The overall goal of the program was to develop techniques for sepa-
rating and characterizing the soluble organic components of secon-
dary effluent from municipal waste treatment plants. The primary
emphasis at the outset of the program was on physical means of
achieving separation of the sample based on the molecular size and/
or weight of the dissolved components. During the course of the
study it became apparent in discussions with the project monitor
that it would be more appropriate to shift the emphasis to chemical
identification of the soluble species, especially the low molecular
weight fraction.
To pursue these goals we explored ultrafiltration, gel permeation
chromatography, and thin layer chromatography as separation tools
and used X-ray and infrared spectrometry for general identification
studies and high resolution mass spectrometry for identification of
specific chemical species.
-------
SECTION IV
GENERAL SAMPLE PREPARATION APPROACH
The general procedure developed for sample preparation is shown in
Figure 1. The secondary effluent sample was collected from points
just before the chlorinator and stored at 4°C prior to filtration.
The primary filtration was achieved with a W and R Balston Ltd. 2-
micron cellulose filter cartridge held in a Commercial Filters
Corporation "Fulflo" water filter. The sample was pressurized in a
5 gallon, stainless steel pressure vessel (Millipore) prior to fil-
tration. This procedure removes most of the large particulate.
Further filtering, plus sterilization by bacteria filtration, was
accomplished with a Millipore 0.45 micron filter held in a Milli-
pore Model No. YY30-142-05 holder. Ultrafiltration was then carried
out In a Millipore Model No. XX42-142-50 ultrafiltration cell with
a series of membranes of different characteristics.
Samples were freeze dried in normal laboratory equipment at various
stages for measurement of total dissolved solids and concentration
of the sample.
The fractions from the membrane ultrafiltration stage were then studied
further by various separation and spectrometric techniques in exploring
ways of characterizing the effluent.
-------
Coding Final
Symbols Code
Collect Secondary Effluent Sample Notebook 4 (11, etc.)
page
i
Primary Filtration (2y)
I
Sterilization Filtration (0.45y) A 4A
I
1
Freeze Dry for Solids S 4AS
Membrane Ultrafiltration M* 4A-M
I
Filtrate
Retentate R 4A-M-R
Fb 4A-M-F
i
Gel Permeation Chromatography
Thin Layer Chromatography
Spectrometry
a. Or other appropriate membrane filter code
(M - Millipore, A - Amicon, E - Eastman).
b. Sequential fractions numbered thus, Fl,
F2, etc.
Figure 1. Sample Preparation Procedure
8
-------
SECTION V
COMPARISON OF THE SOLIDS IN BROCKTON AND ROCKLAND EFFLUENT
Samples of water after secondary waste treatment and before chlori-
nation were obtained from the municipal treatment plants in Brockton,
Massachusetts and Rockland, Massachusetts. Both plants operate an
activated sludge process. The Brockton facility has a normal pro-
cess capacity of about 12 mgd with a design limit of 29 mgd, while
the Rockland plant presently has a normal load of about 0.8 mgd with
a design limit of 2 mgd. The two plants represent a different resi-
dential/industrial influx balance and have experienced quite differ-
ent efflux levels. The two locations, therefore, provided different
types of samples for initial evaluation.
Total Solids
In order to characterize the solids content of the samples, the water
was removed via freeze drying as follows:
One liter of the sterile liquid was coated on the wall of a
5-liter, single necked, round oottom flask by continuous
rotation of the flask until all the liquid was solidified.
The flask was then connected by glass tubing to a second
5-liter flask immersed in a dry ice acetone bath.
Then, a vacuum pump was applied and a vacuum of < 1 mm was
maintained during the freeze-drying. After the majority
of water was removed from the flask (requiring about 24
hours continuous operation), the remaining liquid was
transferred into a one-liter flask, and the freeze drying
procedure was repeated.
The percent freeze dried solids for each sample was found to
be:
Sample
Source solids (mg/fc)
Brockton 4 AS 280
Rockland 5 AS 360
-------
Ash Content
The solids obtained via the freeze dried process were subjected to an
ash analysis at 800°C. The results were:
Sample
Source Code % Ash
Brockton 4 AS 71
Rockland 5 AS 61
* percent of solids after water removed via freeze-drying
Infrared Analysis of Freeze-Dried Solids
An infrared spectrum of the total solids (4 AS and S AS) showed pre-
dominantly water and inorganic salts — probably sulfates and nitrates.
There was only a faint trace of organics (CH2 bands) indicating that
most of the sample was inorganic in nature. The spectrum for sample
4 AS is shown in Figure 2. Therefore, it was decided that to better
characterize the organics in these samples, it would be necessary to
remove the organic portion by solvent extraction.
A 15 mg portion of solids was extracted sequentially with hexane,
chloroform, and methanol using the following procedure:
a) The 15 mg portion was extracted twice (vortex mixer)
with 8 ml of hexane.
b) The extract solutions were centrlfuged and filtered
through Whatman fl paper.
c) The filtrates were dried under N£ and the solids
mixed with KBr, pressed into a pellet and an IR
spectrum obtained.
d) Undissolved solids were rinsed off the filter with
chloroform and extracted twice with 5-ml portions
of chloroform; steps (b) and (c) were repeated.
e) Remaining solids were then finally extracted with
two 8 ml portions of methanol in a similar manner.
10
-------
WAVELENGTH (MICRONS)
2
0.0
0.2
0.4
0.6
0.8
1.0
1 ft
40
5
S
\
\
00
1
\
s
35
\
*—
3
S
f
'
i
i
,
/*
.1
t
i
„
^
x
00
^>
^
\J
i
^
'
f*"
r*
3000
- t
\
I
i
g
a
^
i,..
1
u
2500
FREQUENCY
1
>
I
1
N
\
6
*
^
2000
(CM-1)
^
^
—
1500
r,
n
g
g
-8.
r
<4_
i
4
J
*\
\
\
\
\
\
\
\
\
\
\
V
>
I
1
1
j-
i
—
7
/
/
0
/
12
•^-v
i —
, .T5.,,?
•N,
^
1000
_1
„
-\
|
ft
r
I
o 39
j-
J
500
M
25
FIGURE 2 INFRARED SPECTRUM OF TOTAL BROCKTON SOLIDS (SAMPLE 4AS)
-------
The results of using this approach are summarized below:
Solids from Brockton Effluent - Sample 4 AS
Hexane extract - approximately 0.05 mg of the original 15 mg
of solids (0.3%) were soluble in hexane, and
this extract, after removal of solvent, ap-
peared greasy. IR absorptions (Figure 3a)
were noted at:
Absorption (cm" ) Intensity Comment
3350 medium hydroxyl (acids and alcohols)
2930, 2855, 1465 strong methylene
2960, 2870, 1380 weak methyl
1730 medium-weak ester carbonyl (lipids)
1650 medium-weak amide carbonyl, or aromatic
1250 weak , , /„ «\
} carbon-oxygen (C-0)
1140 strong
610 weak carbon-halogen
chloroform extract - approximately 0.5 mg (3%) of white, powdery,
non-odorous solids were removed. The spec-
trum (Figure 3b) was very similar to that
for the hexane extract except that ester
carbonyl was slightly more intense in re-
lation to other bands.
methanol extract - almost 0.5 mg (32) of material was extracted
by the methanol. The resultant solids were
pale yellow and waxy in appearance with a
sweetish odor. The IR spectrum of this
sample (Figure 3c) was different from the
other solvent extracts — notably more
hydroxyl, almost no- ester carbonyl, and
indications of salts (carbonate, sulfate,
nitrate). The large hydroxyl band (probably
water or methanol) makes meaningful inter-
pretation difficult. It appears that the
methanol extracts a large portion of the
inorganic fraction suggesting that further
extraction should be done only with hexane
and chloroform.
12
-------
VAVf LENGTH (MICRONS)
5
15 20 3044
2900 2000
FREQUENCY (CM1)
500 250
A. Hexane
WAVELENGTH (MICRONS)
4 s • JL
7 • 9 10 12 15 20 30 a
2500 2000
FREQUENCY (OK1)
500 250
B. Chloroform
WAVELENGTH (MICRONS)
4 5
7 8 V 10 12 15 20 30 40
3 0.2
06
08
1.0
JJ t:
4000
2500 2000
FREQUENCY (CM')
500 250
C. Methanol
FIGURE 3 INFRARED SPECTRA OF SOLVENT EXTRACTS OF BROCKTON SOLIDS
13
-------
Solids from Rockland Effluent - Sample 5 AS
The percentage of the original sample which dissolved in each of the
three solvents was the same as observed for the Brokton sample. The
IR spectra were also the same with the following exceptions:
• The infrared spectrum of the hexane solubles (0.3%)
showed little amide (1650_cm~1), much less ester
absorption at 100-1150 cm'1, and no C-halogen at
610 cm'1.
• The chloroform solubles (3%) also showed much less
ester absorption.
X-Ray Analysis of Freeze Dried Solids
X-ray fluorescence - Both samples (4 AS and 5 AS) showed only
the presence of chlorine, sulfur, calcium
and potassium. The X-ray technique is
capable of detecting all elements of atomic
number greater than 13 (Al). Sodium is not
easily detected by this technique.
X-ray diffraction - In examining the patterns for the presence
of specific crystalline species, both samples
showed the presence of sodium chloride. In
addition, the freeze dried solids from the
Brockton effluent contained potassium chloride,
sodium sulfate, and possibly calcium carbonate.
Chemical-Oxygen Demand
Estimates of the chemical-oxygen demand values for these samples and
membrane ultrafiltration fractions (discussed in the following section)
were also obtained for the purpose of comparing the two samples:
—COD Value (mg Q£ consumed/liter)—
* .**„,, j**
Sample Type Brockton Brockton Rockland
0.45 micron filtrate (A) 40 53 72
Ultrafiltration retentate (M-R) 109 133 103
Ultrafiltration filtrate (M-F) 9 -
* 1965 APHA Procedure including correction for chloride
interference.
** 1955 Procedure without chloride correction.
-------
It is not known at this time how much of the COD value is due to
organics and how much is from inorganics. The data show the two
sources of sample to be similar in terms of oxygen demand.
Conclusions
The conclusions reached from this portion of the study were that
the soluble portion of the water samples from both waste treat-
ment plants are predominantly inorganic. This is in accord with
other published results O) where organics averaged 55 mg/X, with
a total dissolved solids being 730 mg/fc.
By the procedures described above we have found the bulk of the
inorganics to be sodium, calcium, carbonate, chloride, and sulfate
with lesser amounts of potassium, ammonium, magnesium, nitrate,
silica, and phosphate. Our data did not suggest any particular
differences in the two sources of the samples, and.we chose to
pursue our studies with samples of the Brockton effluent primarily
because of access convenience. The solubility/IR spectral data
from the solids also suggested that the greatest amount of organic
chemical detail could be obtained from studies of the hexane/
chloroform extracts.
15
-------
SECTION VI
SEPARATION BASED ON MOLECULAR SIZE
Several studies have shown that secondary effluent consists of or-
ganic materials ranging from low molecular weight to natural poly-
mers. Separation techniques were examined to determine whether
classification by effective molecular size could be a useful char-
acterization technique and could also provide samples useful in the
successive analysis steps.
ULTRAFILTRATION
Several membranes were evaluated for their effectiveness in modifying
the filtered effluent samples in a manner which would be most useful
for further characterization studies. Since most specific chemical
analysis techniques tend to have a cutoff in the 500-1000 MW range,
it seemed most appropriate to study those membranes which could
separate the sample based on effective molecular size at about
MW 1000. The Millipore PSAC 142-05 (M), Amicon UM-2 (A), and
Eastman HT-00 (E) membranes were examined. All of these membranes
were reported to have nominal 1000 MW cut-offs.
The Millipore filter is made from mixed cellulose esters, while the
Amicon UM-2 is a polyelectrolyte complex which can have ion exchange
properties. The Eastman membrane is an open Loeb cellulose acetate
type (wet gel) with low salt rejection. These three filters were
selected because they have different properties, and the samples
could be separated in different ways.
Before use, each membrane was washed with sterilized distilled water
to remove any residual trace constituents left during the manufac-
turing process and to obtain some measure of the membrane's per-
formance. The latter was accomplished by observing the water flux
(flow per unit time) at 50 psi, an indirect measure of pore size and
distribution.
The primary means of judging the effect of the ultrafiltration on the
effluent samples was by comparison of the gel permeation chromato-
gram (GPC) of the filtrate and the retentate obtained after filtering
with each of these membranes. The details of the GPC results are
given in the following section.
IT
-------
Millipore PSAC 142-05
Samples 4A and 11A of the Brockton effluent were subjected to ultra-
filtration on the Millipore membrane. A 700 ml aliquot of effluent
was placed in the cell and 480 ml were forced through the membrane
(requiring 50 psi nitrogen). The resultant samples were character-
ized by chemical-oxygen demand, gel permeation chromatography and
ultraviolet spectrophotometry. Unfortunately, the UV analyses were
interfered with seriously, because it was found that the membrane
bleeds a UV absorber unless washed with large amounts of water.
The GPC tracings (Figures 6-9) obtained on these samples gave some
indication of a molecular size separation, but the contamination
problems associated with the membrane suggested that it was generally
as useful for the purposes of the program as the Eastman membrane.
The COD values reported earlier indicate that most of the oxidizable
sample was retained by the membrane.
Amicon UM-2
Our efforts to employ the Amicon membrane were unsuccessful due to
blistering of the membrane in the Millipore separator system. We
have not resolved whether this is a typical problem with this com-
bination or only a bad lot of membranes. Due to the change in em-
phasis during the program, we did not pursue this matter.
Eastman HT-00
The Eastman membrane was placed in the Millipore cell and washed
with distilled water until the filtrate showed no absorbance at
250 my in a 1 cm cell. Then the cell was charged with 1000 ml of
Brockton sample. The cell was pressurized to 100 psi with nitrogen
and water forced through the membrane. The first 50 ml of filtrate
were collected and coded Fl. Successive fractions of 250 ml, 200 ml
and 200 ml were collected and coded F2, F3, and F4. Then the cell
was depressurized, an additional 700 ml of water added, and the
latter forced through the membrane. This filtrate was coded FW1.
The GPC tracings for the initial solution, retentate and filtrate
are shown in Figure 10. It is evident from these results that the
Eastman membrane is effectively separating the sample into two
fractions, presumably on the basis of molecular size, and this mem-
brane was chosen for sample preparation for the remainder of the
project. Several samples of filtrate and retentate were obtained
by filtering approximately 800 ml samples of Brockton sample 11A
with the Eastman membrane in the Millipore ultrafiltration apparatus.
Resulting fractions are coded 11A-E-R or 11A-E-F for retentate and
filtrate.
18
-------
Solids were isolated from the membrane fractions using the freeze
drying technique. Solids content and ash contents of the solids
were as follows:
Ash Content of
Sample Solids Content solids (%)
11AS (S for solids) 240 mg/l
11A-E-RS (retentate) 370 mg/i 50.0
11A-E-FS (filtrate) 180 mg/£ 44.7
With the samples corrected to original volume — based on the approxi-
mately 75% of sample volume which passed through the membrane —
(90 mg/£ for RS and 140 mg/£ for FS), the data suggest that approxi-
mately 61Z of the solids pass through the Eastman filter. The re-
tentate and filtrate solids do not differ appreciably in ash content.
A sequential solvent extraction of the solids revealed the following
solubilities:
—Z Extracted—
Solvent RS FS
hexane 3 1
chloroform 1 trace
methanol 26 49
An examination of the infrared spectra of the total solids and the
extractable portions, while suggesting small amounts of organic
species, was not especially useful in identifying classes of organic
compounds, X-ray fluorescence examination of the RS and FS samples
revealed the normally observed presence of calcium, chlorine, sulfur
and potassium with possible traces of phosphorus, iron and copper.
19
-------
GEL PERMEATION CHROMATOGRAPHY (GPC)
Several studies were carried out to examine the usefulness of GPC
both as a means of sample characterization itself and as a means
of further sample separation to aid the chemical identification
efforts.
A number of columns have been evaluated, including those prepared
from Sephadex (a polycarbohydrate), BioGel (a polyamide) and Corning
CPG (porous glass). Nominal molecular weight cutoffs range from 700
to 30,000 (dextrose). We recognized in initiating these studies that
GPC behavior in aqueous systems is not well understood at the present
time. High concentrations of salts will affect the elution volume of
organics, and the variety of charged species in an effluent sample
will interact in different ways with the polar gel sites. The GPC
studies were also used partly to evaluate the effect of the membrane
filtration.
The GPC studies reported were performed using the appratus shown
schematically in Figure 4. The solvent was water, with flow rates
ranging from 0.2 to 0.5 ml/min. Sample size injected into the
column has ranged from 20 yl to 200 y£. A variety of columns and
samples have been studied, the results of which are given below:
Sephadex G-10 (Exclusion limit of 700)
Some preliminary experiments were run using a Sephadex column with a
low exclusion limit (nominal molecular weight of 700). For this we
compared an inorganic salt (NaCl) with the soluble portion of the
total solids recovered by freeze drying the 0.45 micron filtered
waste water (Sample 11 AS). Figure 5 shows the GPC curves for these
samples using a Laboratory Data Control refractive index detector.
It can be seen that a separation of the soluble filtrate solids has
occurred. However, it also indicates that more adequate differen-
tiation might be obtained on a column with a higher exclusion limit.
Sephadex G-25 (Exclusion limit of 5.000)
A comparison of the GPC behavior of Brockton retentate and filtrate
on Sephadex G-25 is shown in Figure 6 (ultraviolet absorption detec-
tor). From this, it can be seen that the Millipore pellicon ultra
filter (nominal MW cutoff of 1,000) is separating the waste water
sample into two distinctive fractions — the filtrate carrying mostly
low molecular weight material, while the retentate has a wider spec-
trum of material.
20
-------
The patterns for a fresher sample of Brockton water (lla), before
and after ultrafiltration (Millipore), are given in Figure 7. This
series of curves was obtained the same day and is useful for com-
parison of relative amounts in each fraction. The large difference
in appearance between the results reporesented in Figures 6 and 7
is unknown but could be due to differences in initial sample compo-
sition.
BioGel P-30 (Exclusion limit of 30,000)
Figure 8 presents GPC curves for sodium chloride and two samples on
BioGel P-30. Once again, a number of peaks are evident in the ef-
fluent samples. The freeze-dried filtrate (Curve c) shows better
resolution here than in Figure 5 on Sephadex G-10 (the detector
difference could influence this comparison) — something to be ex-
pected if larger molecules are present. The curve for the retentate
(Curve b) is very similar to the one for that sample on Sephadex
G-25 (Figure 7).
Corning Glass CPG-10 (75A° limit of 28,000)
The column prepared from the particular porous glass studied did not
demonstrate much distinctive separation for the retentate sample
(Figure 9). This is a result similar to that observed in some pre-
liminary work done at Waters Associates.
GPC Studies of Eastman Membrane Samples
In order to compare the efficiency of the Eastman (E) and Millipore
(M) membranes in the ultrafiltration process, the original sample
and the ultrafiltration retentate and filtrate were examined on a
Sephadex column. The column actually consisted of three, three-
foot sections in series containing Sephadex G-10, G-15, and G-25
successively; these have nominal high molecular weight cutoffs at
750, 2500 and 5000 respectively. A comparison of these curves in
Figure 10 with the Sephadex curves in Figures 6 and 7 using the
samples from the Millipore membrane suggest that the Eastman mem-
brane gives a more distinct separation between the retentate and
filtrate and would, therefore, be more useful for preparing ana-
lytical samples.
The net conclusion of these GPC studies is that it does appear
possible to achieve some resolution and characterization of the
samples using this approach. However, as a characterization aid
there does not appear to be enough detailed information to be truly
useful, and there does not appear to be any really significant sepa-
ration advantages to the approach as an analytical aid. It appears
that the membrane ultrafiltration step alone gives the greatest as-
sistance in terms of sample fractionation.
21
-------
1
t
K
1
a. Solvent Reservoir - Glass bottle
b. Pump - Milton Roy, Milroyal, capacity 123 ml/hr
c. Sample Injector - Septum and syringe
d. Column - 1/4" x 3' stainless steel
e. Refractive Index Detector - Laboratory Data Control UV
monitor, peak wave length @ 254 my, 8 1 flow-through
cell
f. Refractive Index Detector - Laboratory Data Control
refractomonitor (temperature stabilized 25°C by water
from a Hoake circulating bath) flow-through cell with
a volume of 3 yl
g. Recorder - Hewlett-Packard, Mbseley 7100B - dual channel
recorder
h. Fraction collector or discard
Figure 4: Schematic Representation of GPC System
22
-------
ro
oo
a
v-/
01
o
«
0)
(a) 25 yg/5 yl NaCl
(b) 200 g/20 1 soluble
portion of freeze dried
filtrate (AAS-Brockton)
10
12.5
Volume (ml)
(Flow - 0.5 ml/mln)
Figure 5: GPC Curves on Sephadex G-10
-------
100 yl Retentate
(AA-M-R)
Filtrate
7.5
10
12.5
Volume (ml)
(Flow -0.5 ml/mln)
Figure 6. GPC Curves on Sephadex G-25 of Brockton Effluent After Ultrafiltration (Millipore)
-------
200 yl allquots of
(a) Effluent (11A)
(b) Retentate (1LA-M-R)
(c) Filtrate (11A-M-F2)
ro
P.
CO
M
O
4J
U
2
2
(b)
(c)
5
7l5 ib
Volume (ml)
(Flow -0.2 ml/oin)
12 .'5
15
Figure 7. Brockton Sample 11A — GPC Curves on Sephadex G-25
-------
(a) 200 yg/40 pi NaCl
(b) 40 ill Retentate
(4A-M-R)
(c) 200 ug/20 yl Freeze Dried
(4AS)
(a)
ro
ON
0»
o.
a
O
8
I
2.5
I
5
I
10
12.5
I
15
17.5
Volume (ml)
(Flow -0.5 ml/mln)
Figure 8. 6PC Curves on BloGel P-30 of Brockton Sample 4A
-------
M
O
4J
O
fl
at
O
(a) 100 Ml (4A-M-R)
(b) 100 Ml NaCl
~l 1
7.5 10
Volume (ml)
(Flow - 0.5 ml/min)
2.5
T
5
12.5
15
17.5
20
Figure 9. GPC Curves on Brockton Sample on Coming-Glass CFG-10
-------
ro
CD
10
40
0.45 filtered sample
(11A)
First Membrane Filtrate
(11A-E-F1)
Retentate
(11A-E-R)
50
20 30
Volume (ml)
Figure 10. GPC Curves for Brockton Effluent from Eastman Membrane Ultrafiltration
-------
SECTION VII
SOLVENT EXTRACTION
In our experience with previous problems of this sort^' dealing with
dilute mixtures of a large variety of structure types, we have found
that extraction of the aqueous phase with a non-polar solvent such as
hexane and a polar solvent such as chloroform achieves a nearly com-
plete extraction of all non-polymeric organic materials, except possibly
the very small molecular weight acids and alcohols, and provides an
efficient fractionation of the sample on the basis of structure polarity.
We have also found that species such as acids and phenols may be present
as salts and require acidification before they can be extracted into t
the organic sblvents.
Accordingly, the Brockton effluent samples (11A) were extracted first
with hexane and then with chloroform, acidified to a pH of 1-2 and
extracted a second time with each of the two solvents. The procedure
is shown schematically in Figure 11 for the filtrate fraction. The
same procedure was used for the retentate. The infrared and ultra-
violet absorption spectra of each of the extracts were examined for
evidence of structural features. A brief summary of the observations
from these samples is given in Table 1.
No clear differences were observed in the spectra of the extracts of
the as-is and acidified samples, although we expected some change
since the original pH was about 6-7.
It is difficult to satisfactorily interpret the significance of a com-
posite spectrum of such a complex sample. However, the infrared
spectra do indicate several major functional groups. Some of the ob-
served hydrocarbon absorption bands observed in the hexane extracts
may be due to Impurities later found to be present in the hexane sol-
vents. The infrared spectra on the CHLOR extract indicate the pre-
sence of an alkylated phenyl-ethylene oxide adduct of the type used
as a base for detergents, cleaning agents or emulsifiers.
Successful comparison has been made of spectra of known compounds such
as Tergitol, Igepal, and Hostapol. It could have its origin in some
of the industrial makeup of the treatment plant intake.
The ultraviolet absorption maxima indicate generally the presence of
aromatic or conjugated species in each of the fractions, including some
with fairly extensive chromophores extending the absorption out to
340 nm.
The lower molecular weight filtrate fractions were each studied by
resolution mass spectrometry as reported later.
-------
SOLVENT EXTRACTION PROCEDURE
Sample 11A-XE-F
(2 800 ml samples of Eastman HTOO filtrate)
Iflexane (100 ml)
Hexane .
Solubles Aque<
(HEX)
Chloroform
Solubles
(CHLOR)
Hexane
Solubles
(HEXH)
>us phase
Chloroform (100 ml)
Aqueous phase
1) Acidify (HC1) to pH 1
2) Hexane (100 ml)
Aqueous phase
Chloroform (100 ml)
Chloroform Aqueous phase
Solubles Discard
(CHLORH)
Figure 11. Solvent Extraction Procedure
30
-------
TABLE 1
Infrared and Ultraviolet Spectra of Solvent Extractable Organics
(JO
H
Sample
Filtrate
Filtrate
Filtrate
Filtrate
Retentate
Retentate
Retentate
Retentate
Treatment
as is
as is
acidified
acidified
as is
as is
acidified
acidified
Extracting Solvent
hexane
chloroform
hexane
chloroform
hexane
chloroform
hexane
chloroform
Spectral Features
Infrared Ultraviolet (maxima run)
hydrocarbon
carbonyl, unsaturation
hydrocarbon
carbonyl, unsaturation
hydro carb on, carbony1
hydrocarbon, carbonyl
hydrocarbon
acid
340, 325, 280, 270, 247
270
270
275
288, 260, 245
298, 260
265, 260, 230
(insufficient sample)
-------
SECTION VIII
THIN LAYER CHROMATOGRAPHY (TLC)
Since most of the more polar and reactive species were likely to be
present in the chloroform extract, we explored the suitability of
thin layer chromatography (TLC) for resolution of this extract prior
to identification by mass spectrometry.
Behavior of the acidified filtrate chloroform extract (CHLORH) was
examined on Silica Gel HF using ultraviolet absorption and iodine
detection to evaluate the separation. Mixtures of chloroform and
methanol were evaluated as eluting solvents including 100% chloro-
form as well as 5, 7.5 and 10% methanol in chloroform.
The 10% methanol system had the best overall behavior and gave the
TLC observations reported in Table 2. The spots were extracted with
a 20% methanol in chloroform solution, while the origin was extracted
with a 50% methanol solution. The presence of amines in this extract
was indicated by the positive response to ninhydrin reagent in some
of the TLC tests. Each of the coded sample areas was isolated for
study by high resolution mass spectrometry.
33
-------
TABLE 2
Eluting
TLC Behavior of CHLORH Fraction
Areas*
D
Code for HUMS f Ultraviolet Maxima (nm)
TLC 6
TLC 5
TLC 4
TLC 3
TLC 2
TLC 1
_
0.8 (major) 258
0.7 (major) 255
0.5
0.4 260
0.3 255
0.25
origin (major) 270
*Adsorbent - Silica Gel HF
Solvent - 10% Methanol/90Z Chloroform
-------
SECTION IX
HIGH RESOLUTION MASS SPECTROMETRY (HUMS) OF LOW
MOLECULAR WEIGHT FRACTIONS
High resolution mass spectrometry (HRMS) studies have been completed
on the solvent extracts of the membrane filtrate and also on the
regions separated from the acidified chloroform extract by thin layer
chromatography (TLC), obtained as described in the last two sections.
The samples examined at this stage have proven to still be extremely
complex mixtures and detailed chemical structure assignments can only
be made on a tentative basis. The inherent resolving power in the
HRMS technique was able, however, to demonstrate the power of using
this approach in characterizing such samples.
Experimental
Attempts were first made to obtain HRMS on the four solvent extracts
by drying (in a stream of nitrogen) an aliquot in a small glass con-
tainer and placing the dried sample in the glass reservoir inlet of
the spectrometer, where the vaporization temperature is 200°C. No
volatile species were observed in any of the samples in this manner,
indicating the lack of low molecular weight volatile species in the
samples. This observation could be the result of our sample prepa-
ration procedures, but further studies would be needed to check this
point.
Spectra were finally obtained by drying portions of the solvent ex-
tracts and extracted TLC regions in small glass melting point capil-
laries and placing these in the direct insertion probe. The probe
was then positioned in the cooled (about 80 C) spectrometer source
where high resolution spectra were recorded on photoplates as the
temperature of the source was gradually increased to a limit of
about 250 C. After examining the photoplates, we concluded that all
of the information from each sample could be obtained by processing
the spectra obtained about 100°C and 200°C. No additional infor-
mation was observed at the other temperatures. Spectra were also
obtained on hexane and chloroform solvent blanks.
The photoplate spectra were processed under digital computer (Hewlett-
Packard (2116B) control on a Grant comparator-microdensitometer. The
basic digitized peak profile data were stored on magnetic tape and
later processed for spectral output on an IBM 360/65.
35
-------
RESULTS
The output from each spectrum represented a still highly complex
mixture with typical results showing 2-6 different specific chemical
compositions at each nominal mass value. We have extracted from
these data the observations that represent the major species present
in the samples and the species for which some tentative assignments
could be made.
It is important to remember in interpreting complex spectra such as
from these samples, that one can only make initial estimates as to
the structure types represented by the computed elemental composi-
tions. More detailed sample resolution will be required before the
structures for most of the species can be assigned.
The results from the solvent extracts are given in Tables 3 and 4,
and the results from the TLC separated regions are in Table 5. The
data represent a summary of the results from the spectra of the
samples at the two temperatures. Large amounts of both aliphatic
and aromatic hydrocarbon were observed in all of the samples.
Unfortunately, the background interference from the hexane solvent
blank was sufficiently high that we were unable to determine whether
or not any hydrocarbon species could be attributed to the sewage
effluent, and these data have been eliminated in the tables. The
chloroform blank had only a small amount of hydrocarbon background.
The data in Tables 3 to 5 represent only a small amount of the data
potentially available from the effluent samples. It serves best to
demonstrate the type of information which may be gained. While most
structure assignments are very tentative, the evidence is quite good
for a few specific choices such as the aromatic ethyleneoxide, nico-
tine, and fatty acids. Carbohydrates and amino acids would not have
been observed with the approach used because of their lack of vola-
tility. Lipids (triglycerides) would have been observed but appear
not to have been present in the samples.
Many aromatic amines were observed and a few aliphatic ones. Some
of the species containing nitrogen and oxygen may be amides. An
alkylated phenyl ethylene oxide, previously identified from the infra-
red spectra, is frequently present in many of the fractions. Its
high concentration may be due to its use as an industrial wetting
agent. The interesting series of chlorinated phenols may also have
their origin in Industrial use.
-------
In summarizing these findings, the chemical classes tentatively iden-
tified as being present in the membrane filtered effluent include:
fatty acids, aromatic acids, alkyl-
phenylethylene oxides, chlorinated phenols,
aliphatic and aromatic amines, amides, and
several still unidentified oxygen and nitro-
geneous species.
The samples do not appear to contain low molecular weight volatile
species, and no lipids were observed. A large hydrocarbon background
interference problem resulted in not being able to assess the pre-
sence of this class in the samples. Of course, the species still
present in the aqueous phase have not been examined in this analysis,
but our experience has been that those chemicals amenable to HUMS
studies would have been observed in the work reported. Those still
present in the aqueous phase usually are not volatile enough for
study. Through suitable derivative formation, however, it may be
possible to study additional parts of the sample.
The preliminary results, represented in these samples, clearly demon-
strate that the HRMS technique is a powerful means of obtaining
specific chemical information on complex samples. The potential of
the method is still comprised, however, by the greatest number of
species present in the samples. We recommend that in future programs
an approach be developed for Identifying specific compounds based on
the application of column, thin layer, and gas chromatography to ob-
tain sample resolution, followed by identification using high reso-
lution mass spectrometry.
37
-------
TABLE 3
Summary of HRMS Results from Solvent Extracts of As-Is Sample
Extract - HEX
228
240
203
193
179
183
147
121
Elemental
Composition
C 1 1*^2 8^2
C16H20S2
Ci5H9N
cli»HllN
C13H13N
C8H5N02
C8HnH
Possible Structure (or type)
fatty acid
aromatic amine
aromatic amine
methyl quinoline
quinollne
aromatic amine
phthaliaide (?)
alkyl aniline
Extract - CHLOR
438
256
176
121
98
C8HHH
C5H8NO
? (379-C26B19°3 fragmentation
fatty acid
? (aromatic)
alkyl aniline
? possible amide fragmentation
* both samples have series of oxygenated peaks consistent
with an alkylated phenylethylene oxide.
-------
TABLE 4
Summary of HRMS Results from Solvent Extracts of
Acidified Sample
*
Extract - HEXH
MW
264
230
196
146
Extract-CHLORH
284
256
236
170
Elemental
Composition
C6C15HO
CeCl^O
C6C13H30
C6H,C12
C18H36°2
C16H3202
C16H280
Ci2Hi00
Possible Structure (or type)
pentachlorophenol
tetrachlorophenol
trichlorophenol
dichlorobenzene
fatty acid
fatty acid
»
diphenyl ether (?)
* both samples have series of oxygenated peaks
consistent with an alkylated phenylethylene
oxide.
39
-------
TABLE 5
HUMS Results on TLC Separated CHLORH Sample
Elemental
MW Composition Possible Structure (or type)
TLC-1
very weak spectrum - traces of aliphatic acids
TLC-2
TLC-6
176 C10H12M20 ? (aromatic)
164 C9H12H20 ? (*romatic)
157 CUHUN dimethyl quinoline
143 CigHgN methyl quinoline
101 CHN aliphatic amine
TLC-3
very weak spectrum - traces of oxygenated species
TLC-4
aliphatic acid fragment ions from C6-C10
TLC-5
162 CioHmN2 nicotine
93 C6H7M aniline
284 C18H3602 fatty acid
256 C16H3602 fatty acid
148 C10H120 ?
bromine fragment ions at 79/81(Br), 80/82(HBr)
indicating presence of brominated species
267 C15H23Oi, }
205 C13H1702 I ml°r component. These
177 C8H1704 ( are a11 fragment ions
161 CnH^O f ' from an alkylated phenyl-
135 C9HnO ) ethyleneoxide.
133 C6H1303 J
311 C2oH25N°2 ? (aromatic)
176 C10H12N20 ? (aromatic)
175 CnH13NO
142 C9H20N aliphatic amine fragmentation
170 C12H100 diphenylether
-------
SECTION X
ACKNOWLEDGMENTS
The cooperation and assistance of Mr. Timothy Murphy, Supervisor of
the City of Brockton Water Pollution Control Facilities is sincerely
acknowledged.
We wish to thank Dr. Aaron A. Rosen and Dr. Barry M. Austern, the
Project Officer of the Water Quality Office, Environmental Protection
Agency for their guidance in providing direction of the program, and
we also wish to thank the Office for its support of this project.
The laboratory studies were conducted by Mr. Clifford H. Summers,
Miss Ann L. Hynes and Mr. Arthur A. Massucco. Drs. John T. Funkhouser
and James E. Oberholtzer contributed to design of the laboratory
experiments and interpretation of the data.
-------
SECTION XI
REFERENCES
1. Cleaning Our Environment, The Chemical Basis for Action,
The American Chemical Society, 1969 (page 109).
2. "Chemical Identification of the Odor Components in Diesel
Engine Exhaust", final report by Arthur D. Little, Inc.
on AFRAC Project No. CAPE-7-68 to Coordinating Research
Council and Environmental Protection Agency, June 1970.
NTIS Accession No. PB 194-144.
•National Technical Information Service
U. S. Department of Commerce
Springfield, Virginia 22151
-------
i 1 .Access/on Number
W
n I Sabject Field & Group
05A
SELECTED WATER RESOURCES ABSTRACTS
INPUT TRANSACTION FORM
Organization
Arthur D. Little, Inc.
Title
Characterization and Separation of Secondary Effluent Components by
Molecular Weight
Levins, Philip L.
EPA, WQO Contract No. 14-12-886
2| Note
22
Citation
23 I Descriptors (Starred First)
*Membrane ultrafiltration, *High Resolution Mass Spectrometry, *Chemical
Identification, Gel Permeation Chromatography, Thin Layer Chromatography, Solvent
Extraction, Secondary Waste Effluent, Molecular Weight Determination
Identifiers (Starred First)
27
Membrane ultrafiltration with a 1000 MW cutoff membrane is an effective means
of separating the high and low molecular weight fractions of effluent and provides use-
ful analytical samples for further study. Solvent extraction techniques and thin layer
Chromatography show promise as appropriate methods for isolating and resolving the low
molecular weight effluent subtraction.
While infrared spectrometry yields good spectra of various effluent fractions,
most are still much too complex to be able to infer much specific information from the
spectra of the complex mixture. The same is true of ultraviolet spectrometry. High
resolution mass spectrometry (HRMS) shows promise for being able to obtain a great
deal of specific chemical information even from still very complex mixtures.
Using HRMS, the following classes of compounds have been identified in a solvent
extract of the low molecular weight effluent fraction:
fatty acids, aromatic acids, alkylphenylethylene oxides, chlorinated phenols,
aliphatic and aromatic amines, and amides.
This report was submitted in fulfillment of Project Number 16020FEN,
Contract 14-12-886, under the sponsorship of the Water Quality Office, Environmental
Protection Agency.
Abstractor
Institution
WR:I02 IREV. JUUY tB»»l
WRSIC
SEND. WITH COPY OF DOCUMENT. TO: WATER RESOURCES SCIENTIFIC INFORMATION CENTER
U.S. DEPARTMENT OF THE INTERIOR
WASHINGTON. D. C. 20240
GPO: 1070 - 4O7 -891
-------
ENVIRONMENTAL PROTECTION AGENCY
Publications Distribution Section
Route 8, Box 116, Hwy. 70, West
Raleigh, North Carolina 27607
Official Business
POSTAGE AND FEES PAID
ENVIRONMENTAL PROTECTION AGENCY
------- |