Journal of Chromatographic Science, 2015, Vol. 53, No. 10, 1639–1645
doi: 10.1093/chromsci/bmv092
Advance Access Publication Date: 2 August 2015
Article
Article
Determination of Carbon Number Distributions of
Complex Phthalates by Gas Chromatography–
Mass Spectrometry With Ammonia Chemical
Ionization
Frank P. Di Sanzo1,*, Peniel J. Lim2, and Wenning W. Han3
1
ExxonMobil Research and Engineering, Annandale, NJ, USA, 2ExxonMobil Chemicals, Houston, TX, USA,
and 3ExxonMobil Chemicals, Baytown, TX, USA
*Author to whom correspondence should be addressed. Email: [email protected]
Received 31 December 2013; Revised 11 May 2015
Abstract
An assay method for phthalate esters with a complex mixture of isomer of varying carbon numbers,
such as di-isononyl phthalate (DINP) and di-isodecyl phthalate (DIDP), using gas chromatography–
mass spectrometry (GC–MS) positive chemical ionization (PCI) with 5% ammonia in methane is described. GC–MS-PCI-NH3, unlike GC–MS electron ionization (EI) (GC–MS-EI) that produces generally
m/z 149 ion as the main base peak and low intensity M+ peaks, produces higher intensity (M + 1) ions
that allow the determination of total (R + R0 ) carbon number distributions based on the various R and
R0 alkyl groups of the di-esters moiety. The technique allows distinguishing among the various commercial DINP and DIDP plasticizers. The carbon number distributions are determined in the acceptable range of <0.1 mole percent to >85 mole percent (m/m). Several examples of analysis made on
commercial DINP and DIDP are presented. The use of only 5% instead of 100% ammonia simplifies
use of GC–MS-PCI-NH3 but still produces sufficient M + 1 ion intensities that are appropriate for the
assay. In addition, use of low concentrations of ammonia mitigates potential safety aspects related to
use of ammonia and provides less corrosion for the instrument hardware.
Introduction
Phthalates are esters added to polyvinyl chloride to increase workability and flexibility. Commonly used phthalates include single isomeric
type, such as bis(2-ethyl hexyl) phthalate (DEHP), as well as phthalates consisting of complex mixtures of alkyl groups on the di-ester
moiety. Such complex phthalates include di-isononyl phthalate
(DINP) and di-isodecyl phthalate (DIDP). By industry convention,
the prefix “iso,” when applied to alkyl groups R, R0 (Figure 1) containing six or more carbon atoms, signifies only that an alkyl group is
branched. It does not indicate any specific branching pattern. The acronym DINP is thus not a chemical name of a single compound but a
commercial name applied to a phthalate ester with a complex mixture
of isomers of varying carbon numbers. Commercial products sold as
DINP are not single isomer products. Furthermore, DINP is a high purity di-ester (typically >99.9 wt% ester content) containing mainly
C9–C9-branched isomers but depending on the CAS number, it
does also contain C9–C10-branched isomers. The DINP (CAS
68515-48-0) structure arises from the use of isononyl alcohols that
are an isomeric mixture of branched primary alcohols. The isomers
differ by their branching structures and carbon number (C8–C10).
Therefore, the esterification of phthalic anhydride using abovedescribed alcohols produces a complex mixture of isomers and homologs of phthalates consisting of varying R1, R2 carbon numbers, with
each carbon number containing a multitude of branched alkyl isomers. Consequently, the DINP can contain a variety of isomers of
the following phthalate ester homologs: C8/C8, C8/C9, C9/C9, C9/
C10, C8/C10, C10/C10 and possibly C10/C11.
Likewise, commercial products sold as “DIDP” are not single isomer products. Commercial DIDP (CAS 68515-49-1) is a high purity
di-ester (typically >99.9 mass percent ester content) containing mainly
© The Author 2015. Published by Oxford University Press. All rights reserved. For Permissions, please email: [email protected]
1639
1640
Figure 1. Chemical backbone of phthalate di-esters. For simple esters, R = R0 ;
for complex esters as DIDP and DINP, R and R0 each consist of multiple
branched R/R0 isomers with alkyl chain lengths varying from C8 to C11.
C10-branched isomers. Similar to DINP, esterification of phthalic anhydride with commercial C9–C11 alcohols, with C10 predominant,
produces a complex mixture of isomers and homologs of varying carbon numbers. Consequently, the DIDP can also contain a variety of
isomers of the following phthalate ester homologs: C9/C9, C9/C10,
C10/C10, C10/C11 and C11/C11.
Single isomeric phthalates (e.g., di-butyl phthalate) are routinely
analyzed by gas chromatography/mass spectrometry electron ionization (GC–MS-EI) with m/z 149 as the principal fragment ion (1).
The current consensus technique to monitor phthalates for regulatory
reasons is based on GC–MS-EI (1) for the quantification of low-level
phthalates in commercial products like toys. These methods make it
difficult to easily distinguish the various complex (e.g., DIDP and
DINP) phthalates used commercially. In contrast, the complex esters,
when compared with single isomeric esters, are more difficult to differentiate by GC–MS using EI because of lack of molecular ion cannot be
readily distinguished due to random cleavage of the mixed ester R, R0
alkyl groups. However, the use of a GC–MS technique that produces
more intense molecular ions than EI is desirable in obtaining a reasonable quantitative total (R + R0 ) carbon number distribution of such
complex mixtures since the total carbon numbers (R + R0 ) have unique
M + 1 mass ions.
A desirable approach to develop an analysis based on molecular
ions for the esters is to operate the MS in a chemical ionization
mode that increases the abundance of the molecular ion. Cameron
and Prest (2) in an application note demonstrated that use of 100%
ammonia in GC–MS-positive chemical ionization (PCI)-NH3 produced mostly the M + 1 molecular ion of esters, including DIDP and
DINP. Similarly, Butler et al. (3) used GC–MS-PSI-NH3 (100% ammonia) for the analysis of the simple ester DEHP in a culture media.
However, Cameron and Butler did not develop a full assay method.
Guo et al. (4) demonstrated that methane yielded the most fragmentation with molecular M not being the base peak, isobutane yielded some fragmentation but M + 1 was the base peak and ammonia
100% yielded the least fragmentation with M + 1 the base peak. The
degree of M + 1 production is consistent with the proton affinity (PA)
of the three reagent gases: methane (PA = 424), isobutane (PA = 824)
and ammonia (PA = 854) (5). However, it should be noted that the reference (4) only quantified using the molecular ions 419 and 447,
which correspond to only C9/C9 and C10/C10 for DINP and DIDP,
respectively, and the technique was not used to obtain complete
carbon number distributions to distinguish the commercial types of
the complex DINP and DIDP.
David et al. (6, 7) have extensively reviewed the analyses of phthalates. One application (7) reported the use of atmospheric pressure
Di Sanzo et al.
chemical ionization with water and GC/TOF (API-GC–MS-TOF) to
determine the complex DINP and DIDP phthalates in sludge. The
technique produced mostly M + 1 ions. David et al. noted the presence
of several carbon numbers in the mixtures. Although this technique
potentially may also yield reasonable results for the assay of the
complex esters, the use of it for such assay application was not fully
demonstrated. In addition, GC–MS with a conventional CI source is
more prevalent in industrial laboratories than API-GC–MS-TOF.
This article explores the use of a low concentration of ammonia in
methane as a chemical ionization reagent gas for distinguishing the
commercial complex esters by enhancing the molecular M + 1 ion response and yet minimizes use of large quantity of ammonia. The emphasis is on the use of only the M + 1 ions of the carbon number
distributions. Its application to the bulk assay of industrial or commercially available complex 95%+ DINP and DIDP is demonstrated.
Although this article does not address the quantification of low-level
phthalates present in commercial plastic products, e.g., toys, the methodology of GC–MS/PCI-NH3 may be used to further distinguish the
various complex phthalates in such analysis by analyzing the concentrated extracts obtained from the commercial products, e.g., toys. The
latter application of the methodology will be a topic for a future evaluation as well as the potential use of isobutane reagent gas, which also
demonstrates a relative high abundance of M + 1 ions.
Experimental
Reagents and materials
Reagent grade methylene chloride (CH2Cl2) was used as a solvent in
all GC–MS analyses. N-Butyl phthalate (CAS 84-74-2), DINP
Table I. GC–MS/PCI-NH3 Operating Conditions
Inlet configuration
Cool on-column inlet temperature
program: initial temp.
Injection volume
Column flow
Column
Oven temperature program
Initial temp.
Ramped to 200°C at 50°C/min
Ramped to 350°C at 15°C/min
MS temperatures
MS ionization mode
Solvent delay
Tuning and reagent gas
Mass scan range
Major M + 1 ions for reconstructed
TICs (in parenthesis total R + R0 )
C8/C8 (C16)
C8/C9 (C17)
C9/C9 (C18)
C9/C10 (C19)
C10/C10 (C20)
C10/C11 (C21)
C11/C11 (C22)
Cool on-column mode (or split/
splitless also acceptable)
38°C; set to track oven
automatically
0.1 µL
1.2 mL/min at constant flow
mode
HP-5 ms, 30 m long, 0.25 mm
I.D.; 0.25 µm film
35°C hold for 1 min
Hold for 0 min
Hold for 2 min
Source at 250°C, quadrupole at
150°C
Chemical ionization source
4 min
PCI autotune (CH4), reagent
gas: 5% (v/v) NH3 in CH4
m/z 190–500
M + 1 (m/z)
391
405
419
433
447
461
475
Determination of Carbon Number Distributions of Complex Phthalates
(commercial complex phthalates) reference CAS 68515-48-0 and
DIDP (commercial complex phthalates) reference CAS 68515-49-1
were purchased from Sigma–Aldrich (St Louis, MO, USA).
1641
A GC–MS spectrum verification standard (SVS) was synthesized
in-house from reaction of a mixture consisting of 1/1.25/1.25 mole
ratio of n-nonyl alcohol, n-decyl alcohol and phthalic anhydride, respectively. The reaction temperature was ∼200°C. In such reaction, excess alcohols were typically used to complete the reaction while
removing the water that is generated from the reaction. The remaining
excess alcohols are then stripped to obtain the desired phthalate ester
product.
The 5 volume percent of ammonia in methane in cylinder size 6R
with pressure of 8,274 kPa was purchased from Matheson Gases
(Montgomeryville, PA, USA) and installed in a ventilated hood.
Apparatus
Figure 2. Synthetically prepared phthalates from n-nonyl and n-decyl alcohols
used to verify the PCI-MS Spectra. Only three phthalates are produced with the
indicated carbon number R/R0 . The expected ratio from synthesis is ∼1:2:1
which closely reflects the TIC. This figure is available in black and white in
print and in color at JCS online.
An Agilent 7890 GC equipped with Agilent 5975C Mass Spectrometer with PCI capability was used with the GC–MS operating conditions listed in Table I. The vacuum pump vent was directed to a
controlled ventilated exhaust.
The GC–MS was tuned using the Agilent “autotune” software
optimized for CH4 chemical ionization. The presence of methane in
the reagent gas facilitated the tuning. The CI reagent gas flow rate
was from 0.5 to 2 mL/min. A flow rate of 1 mL/min gave the best
ion ratio signal for the Agilent-specified MS mass calibration tuning
compound.
A signal-to-noise ratio of at least five of the M + 1 mass ion was
obtained when a 1 p.p.m. (m/m) or 100 pg on-column concentration
Figure 3. PCI mass spectra of the three compounds in Figure 2. M + 1 are ions produced using 5% (v/v) NH3 in methane. This figure is available in black and white in
print and in color at JCS online.
1642
of di-n-butyl phthalate in methylene chloride was injected under the
optimum operating conditions.
Sample preparation
The commercial samples of DINP and DIDP were diluted as received
in methylene chloride to produce a 2,000 p.p.m. (m/m) (0.2%) solution. Two milliliters of the solution was placed into a GC–MS autosampler vial and analyzed by GC–MS-PCI-NH3 using the conditions
described in Table I.
Results
Verification of chemical ionization performance
The SVS was initially used to verify, in the molecular region of interest
for quantification of DINP and DIDP carbon number distributions,
that molecular M + 1 were produced using the 5% in methane PCI
under the conditions of Table I. Figure 2 represents the GC–MS/
PCI-NH3 total ion chromatogram (TIC) which confirms, as expected,
the presence of the three products of C9/C9, C9/C10 and C10/C10.
Only three isomers are present because the phthalates were synthesized from the single isomeric n-nonyl and n-decyl alcohols and not
the mixture of multibranched alcohols used to produce the commercial DINP and DIDP. The mass spectra of the three isomers (Figure 3)
indeed indicate that a high abundance of the M + 1 ions is produced in
the molecular ion region of interest when the 5% NH3 in methane is
used as the PCI reagent gas. In contrast, the use of EI at 70 eV of the
three compounds revealed undetectable molecular ions with a base
peak, as expected, of m/z 149. In addition, EI does not produce unique
fragmentation ions that could be used to differentiate the various
carbon numbers of the phthalates. For example, using the pure three
SVS compounds, it was observed that the major fragmentation ions
between the m/z 149 base peak and the M + 1 ion region were common
to several of the carbon numbers. The major fragmentations ions, besides m/z 149 base peak, included m/z 167, 289, 307 for C10/C10; m/z
Di Sanzo et al.
167, 293 for C9/C9 and 167, 293, 307 for C9/C10. It is evident that
one or more carbon numbers would overlap in case such fragmentation ions were to be used and would not readily differentiate the commercial DINP and DIDP using EI.
The GC–MS-PCI-NH3 TIC (Figure 4) of a commercial DINP
phthalate indicates that a complex multitude of isomeric distribution
of several R, R0 carbon number combinations are present that not result in a single compound TIC as observed with the SVS but in a signal
response spread over a broadly spaced GC retention time window. A
snapshot mass spectrum (Figure 5) at retention time 11.4 min of the
TIC of Figure 4, for example, reveals that at least four carbon numbers
are present and that they coelute. The overall de-convolution of the
various carbon number phthalates is possible using the M + 1 ions listed in Table I. Figure 6 shows examples from such reconstructions of
the TIC in Figure 4 for three of the carbon numbers indicated.
Figure 7 represents the GC–MS-PCI-NH3 TIC and selected extracted ions of a commercial DIDP plasticizer. Similar to the DINP
TIC in Figure 4, the DIDP is also a complex mixture spread over a
wide retention time band and consisting of several carbon numbers
and many isomers. As an example, two of the carbon number distributions are shown as extracted ions; other carbon number distributions may be obtained using their respective M+1 ions as listed in
Table I. For DIDP, the C10/C10 carbon number is usually the most
intense in signal. Smaller amounts of C11/C11 may also be present
as the esterification of DIDP with C10-branched primary alcohols
may contain small amounts of C11-branched alcohols.
Discussion
Analyses of commercial DINP and DIDP phthalates
To obtain a quantitative relative carbon number distribution for DINP
and DIDP, it is first necessary to confirm that the GC–MS-PCI-NH3
response is linear for the various carbon numbers and the amount
of sample introduced into the GC–MS using the operating conditions
in Table I. The linearity initially was confirmed using di-n-butyl
Figure 4. PCI-TIC chromatogram of a commercial DINP reference sample. This figure is available in black and white in print and in color at JCS online.
Determination of Carbon Number Distributions of Complex Phthalates
1643
Figure 5. Snapshot mass spectrum at retention time 11.4 min of Figure 4 to illustrate the overlapping carbons for the DINP. This figure is available in black and white
in print and in color at JCS online.
Figure 6. PCI-MS extracted ion chromatograms of the TIC in Figure 4 of several carbon number phathalates in DINP. This figure is available in black and white in print
and in color at JCS online.
1644
Di Sanzo et al.
Figure 7. PCI-MS TIC chromatogram and examples of extracted ion chromatograms of two carbon number phthalates in a commercial DIDP reference sample. This
figure is available in black and white in print and in color at JCS online.
Table II. Example of Verification of PCI-NH3 Linearity for DINP in the
Range of 0–5,000 p.p.m. (m/m)
Masses
C9/C9
C9/C10
C10/C10
C10/C11
Extracted mass ion PCI-NH3 linearity equation slopes and
from TIC (M + 1) intercepts and R 2 (total relative integrated
area vs. dilution concentration
calibration)
419
433
447
461
Y = (14,220X) − 1,000,000 (R 2 = 0.9987)
Y = (6,702X) − 706,770 (R 2 = 0.9982)
Y = 925X − 92,653 (R 2 = 0.9987)
Y = 75X + 2,386 (R 2 = 0.9986)
phthalate and subsequently using the commercial mix reference DINP
and DIDP samples. For the DINP and DIDP reference samples, serial
dilutions of 0.05, 0.10, 0.25 and 0.50% (m/m) in methylene chloride
were prepared. In the absence of commercially available DINP and
DIDP reference materials, linearity may be verified with an actual production sample of DIDP and/or DINP by using similar specified dilutions. To obtain the linearity plots, each of the M + 1 ions in Table I
was extracted from the TIC. The signal of the extracted ion chromatogram over the entire retention time window of DINP or DIDP elution
is integrated for each carbon number and plotted vs. dilution concentration. As an example, the linearity equations and correlation coefficients are summarized in Table II for DINP. The equations do not
reflect the actual relative response “factors” of each carbon group
since the actual concentration of each carbon number in the
Table III. Examples of a DINP and DIDP Carbon Number
Distributions of Commercial Reference Samples (mol%)
Structure
DINP CAS 68515-48-0
DIDP CAS 68515-49-1
C8/C8
C8/C9
C9/C9
C9/C10
C10/C10
C10/C11
C11/C11
0.13
4.09
63.23
28.54
3.80
0.19
0.03
0.26
0.04
0.49
1.97
88.05
8.74
0.45
commercial or reference samples may not be known. For the analyses
of the commercial samples, the dilution was maintained as to be within the linearity range of each calibrated carbon number. It should be
noted that the linearity range was sufficient to be able to determine a
total single carbon consisting of multiple isomers distributed over the
specified retention time window for that carbon to <0.1 mol%, which
is more than adequate for the intended application.
The relative mass percentage Cx equivalent for each carbon number cluster from C8/C8 to C11/C11 can be calculated from normalization of the peak area sums for each ion trace.
Table III shows relative distributions for each major carbon numbers for commercial DINP and DIDP reference samples. The relative
distributions reflect the structure and relative quantities of the alcohols used in the esterification of the phthalic anhydride. As expected
Determination of Carbon Number Distributions of Complex Phthalates
DINP’s largest carbon number cluster is centered on C9/C9, whereas
that of DIDP is centered on C10/C10. DIDP shows a greater amount
of C11/C11 since slightly greater amounts of C11 alcohols may
have been present in the reaction vs. alcohols used in the DINP
production.
Conclusion
GC–MS/PCI-NH3 (5% NH3) has significant advantages when used to
assay phthalates and particularly phthalates with a more complex isomeric distribution like DINP and DIDP. It is a highly selective ionization technique and produces a much simpler spectrum of the M + 1 ion
with sufficient ion abundance to be able to determine a single carbon
number to <0.1% mole and allows relative carbon number distributions of complex phthalates to be readily determined. Such analysis
is beneficial in process control and regulatory-related evaluations. Furthermore, the use of 5% NH3 in methane facilitates the safety control
of the reagent gases and possible less degradation of the MS components. As a future development, uses of other reagent gases such as isobutane can be explored for similar application.
1645
References
1. Health Canada; Part B: test methods section, method C-34 determination of
phthalates in polyvinyl chloride consumer product, product safety reference
manual, Book 5—laboratory policies and procedures 2006-12-12.
2. Cameron, G., Prest, H.; Determination of phthalate esters by positive chemical ionization MS with retention-locked GC, Vol. 5, No. 2. LCGC Asia
Pacific, Chester, UK, (2002), p. 27.
3. Butler, J., O’Brien, P., Crain, S., Yargeau, V.; Determination of DEHP in
culture media by GC-MS/MS using PCI ammonia, Application Note:
52282 (2012). Thermo Fisher Scientific, Austin, TX, USA.
4. Guo, Z., Jiang, L., Li, X., Liang, S., Lu, Y., Yang, X., et al.; Chinese Patent
Detection Method of Phthalate Plasticizer CN20101217703, July 5, 2010.
http://www.google.com/patents/CN101858897B?cl=en (accessed July 2014).
5. Gross, J.H.; Mass spectrometry—a textbook. Springer-Verlag, Berlin, Germany (2004), p. 340.
6. David, F., Sandra, P., Tienpont, B., Vanwalleghen, F., Ikonomou, M.;
Phthalate esters. In Staples, C. (ed). The handbook of environmental
chemistry/anthropogenic compounds, Vol. 3, Springer, (2003).
7. David, F., Sandra, P., Hancock, P.; Determining high-molecular-weight
phthalates in sediments using GC-APCI-TOF-MS, LC/GC, Chester, UK,
May 1, 2011.