Synthesis and surface spectroscopy of α-pinene isotopologues and their corresponding secondary organic material

Article abstract of DOI:10.1039/c9sc02399b

Atmospheric aerosol-cloud interactions remain among the least understood processes within the climate system, leaving large uncertainties in the prediction of future climates. In particular, the nature of the surfaces of aerosol particles formed from biogenic terpenes, such as α-pinene, is poorly understood despite the importance of surface phenomena in their formation, growth, radiative properties, and ultimate fate. Herein we report the coupling of a site-specific deuterium labeling strategy with vibrational sum frequency generation (SFG) spectroscopy to probe the surface C-H oscillators in α-pinene-derived secondary organic aerosol material (SOM) generated in an atmospheric flow tube reactor. Three α-pinene isotopologues with methylene bridge, bridgehead methine, allylic, and vinyl deuteration were synthesized and their vapor phase SFG spectra were compared to that of unlabeled α-pinene. Subsequent analysis of the SFG spectra of their corresponding SOM revealed that deuteration of the bridge methylene C-H oscillators present on the cyclobutane ring in α-pinene leads to a considerable signal intensity decrease (ca. 30-40%), meriting speculation that the cyclobutane moiety remains largely intact within the surface bound species present in the SOM formed upon α-pinene oxidation. These insights provide further clues as to the complexity of aerosol particle surfaces, and establish a framework for future investigations of the heterogeneous interactions between precursor terpenes and particle surfaces that lead to aerosol particle growth under dynamically changing conditions in the atmosphere.

Full text of DOI:10.1039/c9sc02399b

Chemical
Science
View Article Online
View Journal
EDGE ARTICLE
Synthesis and surface spectroscopy of a-pinene
isotopologues and their corresponding secondary
organic material†
Cite this: DOI: 10.1039/c9sc02399b
All publication charges for this article
have been paid for by the Royal Society
of Chemistry
Mary Alice Upshur,‡a Marvin M. Vega, ‡a Ariana Gray Be, ‡a Hilary M. Chase,a
´
b
c
c
Yue Zhang,
Aashish Tuladhar,
Zizwe A. Chase,c Li Fu,
Carlena J. Ebben,a
c
bd
a
*
Zheming Wang,
and Regan J. Thomson
Scot T. Martin,
Franz M. Geiger
a
*
Atmospheric aerosol–cloud interactions remain among the least understood processes within the climate
system, leaving large uncertainties in the prediction of future climates. In particular, the nature of the
surfaces of aerosol particles formed from biogenic terpenes, such as a-pinene, is poorly understood
despite the importance of surface phenomena in their formation, growth, radiative properties, and
ultimate fate. Herein we report the coupling of a site-specific deuterium labeling strategy with vibrational
sum frequency generation (SFG) spectroscopy to probe the surface C–H oscillators in a-pinene-derived
secondary organic aerosol material (SOM) generated in an atmospheric flow tube reactor. Three a-
pinene isotopologues with methylene bridge, bridgehead methine, allylic, and vinyl deuteration were
synthesized and their vapor phase SFG spectra were compared to that of unlabeled a-pinene.
Subsequent analysis of the SFG spectra of their corresponding SOM revealed that deuteration of the
bridge methylene C–H oscillators present on the cyclobutane ring in a-pinene leads to a considerable
signal intensity decrease (ca. 30–40%), meriting speculation that the cyclobutane moiety remains largely
intact within the surface bound species present in the SOM formed upon a-pinene oxidation. These
insights provide further clues as to the complexity of aerosol particle surfaces, and establish a framework
for future investigations of the heterogeneous interactions between precursor terpenes and particle
surfaces that lead to aerosol particle growth under dynamically changing conditions in the atmosphere.
Received 17th May 2019
Accepted 21st July 2019
DOI: 10.1039/c9sc02399b
rsc.li/chemical-science
Here, we present the synthesis and characterization of several
deuterium-labeled a-pinene isotopologues and particle-phase
Introduction
SOM prepared from them that enable us to begin elucidating
mechanisms driving gas:particle conversion, cloud activation,
and the ultimate fate of particles containing SOM.
Acquiring mechanistic insight into the formation, growth, and
climate-relevant properties of aerosol particles containing
secondary organic material (SOM) from gas-phase terpenes
presents an ongoing challenge in the atmospheric science and
modeling communities.1–4 While the preparation and use of
SOM precursor isotopologues holds the promise for enabling
detailed molecular studies of aerosol particle chemistry and
physics, labeling these precursors is not yet widely accessible.
The formation, composition, and physicochemical proper-
ties of biogenic SOM-containing particles are reasonably well
known to depend on the structure and reactivity of terpene
volatile organic compound (VOC) precursors, tropospheric
conditions of temperature and relative humidity (RH), and the
extent of particle phase state heterogeneity.4–7 In contrast,
heterogeneous phenomena occurring at particle surfaces
remain elusive, despite their intrinsic importance in key
formation and growth processes, i.e. condensational growth
(gas:particle conversion),8 coagulation (particle:particle
merging),9 and cloud droplet activation (cloud condensation
nucleation activity).10,11 Scientic insights into the surface-
specic molecular origins of SOM-containing particle forma-
tion, growth, and cloud activation are therefore critical to the
study of such constituents in the climate system, yet direct
studies of the particle interface are notoriously challenging to
carry out nondestructively with chemical specicity and under
aDepartment of Chemistry, Northwestern University, Evanston, IL 60208, USA. E-mail:
r-thomson@northwestern.edu; f-geiger@northwestern.edu
bJohn A. Paulson School of Engineering and Applied Sciences, Harvard University,
Cambridge, MA 02138, USA
cWilliam R. Wiley Environmental Molecular Sciences Laboratory, Pacic Northwest
National Laboratory, Richland, WA 99352, USA
dDepartment of Earth and Planetary Sciences, Harvard University, Cambridge, MA
02138, USA
† Electronic supplementary information (ESI) available: Detailed information for
the synthesis of isotopologues, SOM collection, and supplementary SFG data. See
DOI: 10.1039/c9sc02399b
‡ These authors contributed equally to this work.
This journal is © The Royal Society of Chemistry 2019
Chem. Sci.

View Article Online
Chemical Science
Edge Article
tropospherically relevant conditions of temperature, pressure, SFG signal intensity (around 2930 cmÀ1) observed in the unla-
and RH.4,12,13
beled a-pinene spectrum. Following the preparation of SOM
Our previous work has demonstrated the applicability of from all accessible isotopologues, the ring-substituted SOM SFG
vibrational sum frequency generation (SFG) spectroscopy as spectra reveal a considerable, albeit far from complete reduc-
a nondestructive nonlinear optical probe to analyze surface- tion of the 2950 cmÀ1 signal intensity, indicating that individual
localized oscillators of particle-phase SOM.8,13–15 SFG spectros- oscillators on the SOM surface can be chemically accessed and
copy has revealed molecular information regarding the struc- thus identied. These rst attempts toward identifying the
ture, dynamics, and orientations of terpenes and their oxidation surface oscillators in a-pinene-derived SOM will enable future
products at atmospheric model interfaces in the C–H and C]O investigations of adsorption, desorption, and reactive uptake of
vibrational stretching regions.16–19 Notably, the SFG intensity the gas-phase species onto or into aerosol particles, with direct
obtained from a-pinene-derived SOM in the C–H region is implications for elucidating aerosol formation and growth
independent of particle size8,20 and can be readily tracked under mechanisms. Beyond such surface spectroscopic studies, our
varying conditions of RH and temperature in real time.
general synthetic route offers the opportunity to utilize site-
These initial studies have unlocked the prospect of using specic isotopically labeled a-pinene analogues in new mass
SFG spectroscopy to probe the dynamics of VOC uptake onto or spectroscopy-coupled chamber studies on SOM formation,
into the particle phase under varying RH, temperature, and growth, and properties within the greater atmospheric
consequently particle viscosity.13,16 Surface-selectively investi- community.
gating such early-stage heterogeneous interaction events
leading to particle generation would provide critical new
insights useful to solving the outstanding uncertainties within
aerosol particle mechanisms. Unfortunately, the SFG spectra of
Results and discussion
Synthesis of a-pinene isotopologues
Our earlier site-specic deuterium labeling of C8 and C10
methyl C–H oscillators in a-pinene utilized naturally occurring
a-pinene as a convenient starting material, which allowed a top-
down approach to access endo-methyl and/or vinyl-methyl
deuterated analogs 2–4 (Fig. 1A). This approach enabled the
unambiguous spectral assignment of the vinyl C10 methyl
symmetric stretch within the complex vibrational spectrum of
a-pinene,22 but was not broadly applicable to labeling less
accessible sites like the C6 methylene bridge and C1/C5
bridgehead positions. We therefore considered that a more
most terpenes and oxidation products relevant to biogenic SOM
remain complex and congested, with overlapping spectral
lineshapes that make vibrational mode assignments chal-
lenging.16–19 Moreover, the signal frequencies and intensities
obtained from SOM derived from a particular terpene overlap
signicantly with those of the precursor molecule, making the
aforementioned dynamic studies nearly impossible.8,13 For
instance, the highest SFG signal intensity obtained from a-
pinene-derived SOM collected on a lter sample occurs at
2950 cmÀ1, shied only by ꢀ15 cmÀ1 from the otherwise
similarly shaped SFG spectrum of a-pinene itself.8,21–24 Site-
specic deuterium labeling offers the opportunity to address
these limitations by “silencing” signal intensity generating C–H
oscillators in the SFG spectra of both the precursor terpene and
its corresponding SOM, thereby enabling vibrational mode
elucidation of the highly congested SFG spectra. We also envi-
sion probing the interactions between unlabeled VOCs and
labeled particle surfaces using deuterated precursors and
preparing labeled SOM from them to distinguish the unlabeled
gas-phase molecules from those in the particle phase.
With such goals in mind, we present here results from
a detailed isotope-labeling study in which we synthesized novel
a-pinene isotopologues and their corresponding SOM, and
characterized them using SFG spectroscopy. We expand upon
our previous work on methyl C–D a-pinene isotopologues22 by
presenting the synthesis and SFG spectroscopy of three new
analogues with methylene bridge, bridgehead methine, allylic,
and vinyl deuteration. Specically, we report on the develop-
ment of a strategy for the de novo synthesis of a-pinene iso-
topologues adapted from a previous total synthesis completed
by Thomas and Fallis.25,26 We then provide SFG spectra of the
isotopologues at vapor:solid interfaces to investigate contribu-
tions from such C–H stretches on the complex SFG spectrum of
a-pinene in the pursuit of identifying high signal intensity
generating oscillators. Spectra of the ring-substituted iso-
Fig.
numbering throughout this manuscript follows the convention indi-
topologues revealed up to 80% suppression of the dominant cated by compound 1.
1 a-Pinene isotopologues synthesized in this study. Atom
Chem. Sci.
This journal is © The Royal Society of Chemistry 2019

View Article Online
Edge Article
Chemical Science
general approach capable of labeling any position on a-pinene yield by rst conducting a base-promoted aldol condensation
might involve a de novo or bottom-up synthesis, wherein specic with benzaldehyde followed by tosylation of the liberated
steps in the synthetic sequence would enable either hydrogen or primary alcohol. As demonstrated by Thomas and Fallis,
deuterium incorporation and thus allow for the generation of installation of the benzylidene group on the less hindered C3
multiple isotopologues in a unied fashion (Fig. 1B). Within side of the ketone within 11 was necessary to enforce
this context, the previously reported total synthesis of a-pinene production of the desired regioisomer during the subsequent
and b-pinene completed by Thomas and Fallis in 1973 (ref. 25 formation of the cyclobutane ring. Accordingly, exposure of
and 26) represented an ideal starting point for our approach. tosylate 12 to sodium hydride in dimethoxyethane at 80 ꢁC
Their synthesis utilized Hagemann's ester as a starting material, afforded methylene bridge labeled bicycloheptane 13 in 88%
and our analysis of their route indicated that it should be yield. Removal of the benzylidene blocking group was
feasible to incorporate deuterium atoms at several positions accomplished using g-aminobutyric acid (GABA) as a retro–
with high levels of control. The successful development of such aldol catalyst, as described by Thomas and Fallis.25 Somewhat
a bottom-up approach that allows for deuteration at C1 and C3 disappointingly, the best reproducible yield obtained for
to C6, as exemplied by the synthesis of three new analogues 5– nopinone-d2 (14) under optimized conditions was 41% yield
7 with methylene bridge, bridgehead methine, allylic, and vinyl (71% yield based on recovered starting material), in part due to
deuteration, is described below.
product volatility. With advanced material in hand, however,
Synthetic efforts commenced with targeting deuteration of the synthesis of a-pinene-d2 (5) was completed via triate 15,
the C6 methylene bridge of a-pinene (Scheme 1), referred to which underwent methylation with methyl cuprate under the
here as a-pinene-d2 (5). Reduction of Hagemann's ester (8) procedure we had developed previously (51% yield over two
with LiAlD4 generated diol 9 (72% yield), wherein the labeled steps from 14).22
hydroxyl methyl appendage represented incorporation of
Having established the viability of modifying the Thomas
deuterium at the site that would eventually become the C6 and Fallis synthesis to generate methylene bridge labeled
methylene bridge within the target. Selective oxidation of the analogue, a-pinene-d2 (5), we turned our attention to installing
allylic hydroxyl group within 9 by DDQ provided enone 10, deuterium at the two bridgehead positions (Scheme 2). We
which was converted to ketone 11 following acetylation and reasoned that deprotonation of ester 8 to yield an extended
methyl cuprate 1,4-addition (52% yield over three steps from
9). Ketone 11 was then transformed into tosylate 12 in 73%
Scheme 2 Synthesis of deuterated a-pinene-d5 (6) and a-pinene-d7
(7). Reagents and conditions: (i) NaH (4 eq.), THF, 0 ꢁC, 1.5 h, then D2O,
Scheme 1 Synthesis of deuterated a-pinene-d2 (5). Reagents and D2SO4, pH 2, 1 h, 89%; (ii) LiAlH4 (1.2 eq.), Et2O, 0 ꢁC, 3 h, 94%; (iii)
conditions: (i) LiAlD4 (1.1 eq.), Et2O, 0 ꢁC, 0.5 h, 72%; (ii) DDQ (1.3 eq.), LiAlD4 (0.9 eq.), Et2O, 0 ꢁC, 3 h, 83%; (iv) DDQ (1.2 eq.), 1,4-dioxane,
1,4-dioxane, 18 h, 78%; (iii) Ac2O (1.3 eq.), DMAP (5 mol%), DCM, 12 h, 18 h, 91% of 17 or 78% of 18; (v) Ac2O (1.3 eq.), DMAP (5 mol%), DCM,
89%; (iv) MeMgBr (1.5 eq.), CuBr$DMS (1 eq.), THF, À30 ꢁC, 0.5 h, 75%; 12 h, 85%; (vi) MeMgBr (1.5 eq.), CuBr$DMS (1 eq.), THF, À30 ꢁC, 0.5 h,
(v) benzaldehyde (2.2 eq.), 10% NaOH (1.25 eq.), EtOH, 12 h; (vi) TsCl 74%; (vii) benzaldehyde (2.2 eq.), 10% NaOD (1.25 eq.), EtOH, 12 h; (viii)
(1.5 eq.), NEt3 (2 eq.), DCM, 12 h, 73% over 2 steps; (vii) NaH (2 eq.), TsCl (1.5 eq.), NEt3 (2 eq.), DCM, 12 h, 63% over 2 steps; (ix) NaH (2 eq.),
DME, 80 ꢁC, 1 h, 88%; (viii) 7 M KOH (10 eq.), GABA (0.25 eq.), DMSO, DME, 80 ꢁC, 1 h, 93%; (x) 7 M NaOD (7 eq.), d6-DMSO, 90 ꢁC, 0.75 h,
100 ꢁC, 0.75 h, 41% yield (71% brsm); (ix) LDA (1.5 eq.), Comins' reagent 20% yield (41% brsm); (xi) LDA (1.5 eq.), Comins' reagent (2 eq.), THF,
(2 eq.), THF, À78 to 0 ꢁC, 2 h; (x) MeLi (2.7 eq.), CuI (2 eq.), THF, À5 ꢁC, À78 to 0 ꢁC, 2 h; (xii) MeLi (2.7 eq.), CuI (2 eq.), THF, À5 ꢁC, 14 h, 36%
14 h, 51% over 2 steps.
over 2 steps.
This journal is © The Royal Society of Chemistry 2019
Chem. Sci.

View Article Online
Chemical Science
Edge Article
enolate might allow for direct labeling at the desired position(s)
upon quenching with an appropriate source of deuterium. To
that end, Hagemann's ester (8) was treated with excess sodium
hydride to generate a solution of the requisite enolate, which
was quenched by the slow addition of D2O and then carefully
acidied with deuterated sulfuric acid. Under optimized
conditions, we found that ester 16 could be generated in 88%
yield with complete incorporation of two deuterium atoms at
both C1 and C5. We were initially surprised to see labeling of
the enone vinyl position, but this can be rationalized by
considering the highly delocalized nature of the enolate formed
from Hagemann's ester. It was important that the reaction
mixture be kept at a pH of ꢀ2.0 during the quench in order to
ensure complete deuteration at both positions, lest an insepa-
rable mixture of partially labeled species form. From this point
deuterated Hagemann's ester (8) was treated similarly to the
previous modied synthesis presented. Reduction with LiAlH4
followed by allylic oxidation with DDQ provided enone 17 (85%
over two steps). Acetylation of the free hydroxyl and conjugate
addition with the cuprate reagent produced acetate 19. When
benzylidene formation was conducted under the original
conditions, hydrogen–deuterium exchange had occurred adja-
cent to the ketone, removing the C1 deuterium label previously
inserted. The exchange could be easily avoided by using 10%
NaOD during the aldol condensation, allowing ketone 20 to be
isolated with complete deuterium installation at C1. Subse-
quent treatment of ketone 20 with sodium hydride induced
cyclization as before to generate cyclobutane 21 in 89% yield,
which now possessed deuterium labels at both C1 and C5
bridgehead carbons.
Removing the benzylidene group within 21 proved to be
problematic due to unforeseen loss of the bridgehead deute-
rium atoms. GABA was found to facilitate this process, as was
the use of NaOH and regular DMSO. To circumvent this issue,
we conducted the retro–aldol reaction in the absence of GABA
using NaOD and deuterated DMSO as the solvent. Under our
best conditions found, the benzilidene group could be cleaved
without loss of bridgehead deuterium atoms in 20% yield (41%
brsm), however, we now observed additional deuteration a- and
b- to the ketone at C3 and C4, respectively (i.e., 22 was the major
product). While deuteration next to the ketone was not unex-
pected, we were surprised to see deuteration at the b-position.
We reason that deprotonation of 21 at the C4 allylic position
must be facilitating this exchange process before the desired
retro–aldol reaction takes place. Unfortunately, despite much
experimentation, the efficiency of this allylic exchange could
not be rendered 100% effective; 1H NMR spectroscopy of 22
shows 50–70% of double to single deuteration at C4. This
inefficiency, however, was not expected to be problematic in the
planned SOM synthesis and spectroscopic analysis, and so we
pressed forward to generate the desired a-pinene-d5 (6) as in the
prior analogue synthesis. We also prepared a-pinene-d7 (7),
substituting LiAlD4 for the reduction of deuterated Hagemann's
ester (8) to generate alcohol 18, which was subjected to the same
sequence as described for a-pinene-d5 (6, see ESI† for complete
details).
SFG spectroscopy of a-pinene isotopologues
With synthetic isotopologues 5–7 in hand, we expected that
deuteration of any position contributing to the SFG response of
unlabeled a-pinene would change the SFG signal intensity ob-
tained from the deuterated analogues, effectively leading to
intensity “silencing” of the deuterated C–H oscillators within
the SFG spectra. However, we note that SFG signal generation is
coherent in nature, akin to a vector eld of oscillators adding up
in space to produce a nal signal vector arriving at the detector
(i.e., a total signal at a given frequency), and thus, elimination of
one or more vectors by deuteration does not necessarily lead to
an overall total signal reduction. Moreover, as is the case in all
vibrational spectroscopies, we caution that the observed signal
could be produced not only from the coherent superposition of
the electric elds generated from SFG-active oscillators at
a given frequency, but also those coupled to modes oscillating
at proximate frequencies.
Fig. 2 shows the 0.6 cmÀ1 resolution SFG spectra of vapors of
isotopologues 5–7 and of unlabeled a-pinene in contact with
fused silica (Fig. S2† shows that spectra obtained on CaF2 are
comparable, while Fig. S3† shows standard-resolution spectra
obtained at NU). The SFG response of unlabeled a-pinene is
predominantly observed within two vibrational frequency
regions centered around 2930 cmÀ1 and 2880 cmÀ1. Previous
atomistic simulation studies accounting for Fermi resonances
and nonlocal modes have reported that the bridge methylene
group on a-pinene contributes considerably to the dominant
spectral feature around 2930–2940 cmÀ1
.
Specically, static
27
DFT calculations indicate that the 2930–2940 cmÀ1 peak is 77%
composed of the bridgehead methylene symmetric stretch and
a Fermi resonance (FR), with the remaining 23% of the SFG
response at this frequency due to overtone bands of the gem-
dimethyl CH3 symmetric stretches (19%), the asymmetric
stretch of the vinyl CH3 (1%), and other stretches (3%).27 MD
Fig. 2 Sub 1 cmÀ1-resolution ssp-polarized vapor-phase SFG spectra
of (À)-a-pinene (black trace) and isotopologues 5–7 in contact with
fused silica. Intensities are scaled based the highest intensity peak of
the vapor (À)-a-pinene spectrum collected on the same day as each
of the isotopologues.
Chem. Sci.
This journal is © The Royal Society of Chemistry 2019

View Article Online
Edge Article
Chemical Science
simulations indicate that the bridgehead methylene symmetric
stretch and FR contribute 79%, with the remaining contributions
from the overtone band FR of the gem-dimethyl CH3 symmetric
stretches (12%), the gem-dimethyl CH3 asymmetric stretch (3%),
vinyl CH3 asymmetric stretch (1%), bridging methine CHs (4%),
and other modes (1%).27 At 2880 cmÀ1, static DFT calculations
indicate that ꢀ75% of the SFG response is due to the gem-
dimethyl CH3 symmetric stretches and their Fermi resonance,
with the remainder being due to the allylic CH2 symmetric and
asymmetric stretching modes (ꢀ20%) and other modes.27 The
C–H symmetric stretch on the vinyl CH3 group was found to
contribute to the 2880 cmÀ1 peak by a somewhat greater extent
(ca. 2%) in the ensemble-averaged spectrum from MD simula-
tions than in the static DFT calculations.27 Isotopologues 5–7
each exhibit a marked decrease in the SFG response in the
2930 cmÀ1 region relative to unlabeled a-pinene, while their
methyl symmetric stretching modes between ꢀ2850 cmÀ1 and
ꢀ2880 cmÀ1 are largely unchanged, as expected. The latter
outcome suggests that the molecular orientation distributions of
the isotopologues 5–7 are comparable to that of the unlabeled
compound, which is an important result given the dependence of
SFG responses on molecular orientation.
Fig. 3 shows that contributions to the 2930 cmÀ1 signal from
the methyl groups seem to be minor if not negligible, as
deuterating one or two of those groups (isotopologues 2–4) still
produces the highest peak intensity at that frequency.22 The
overall reduction of signal intensity in the 2930 cmÀ1 region
exhibited in the spectra of isotopologues 5 and 7 is consistent
with the previously reported contributions from deuteration of
the bridge methylene group,27 but is rather unexpected for
isotopologue 6, given its bridge methylene group is not
deuterated. The results shown in Fig. 2 and 3 then suggest that
the two bridging methine groups on the ring (deuterated in
isotopologues 6 and 7) also contribute somewhat to the overall
SFG response at ꢀ2930 cmÀ1, perhaps to a larger extent than the
ca. 4% previously reported by MD simulations.27 Similar inter-
actions between C–H oscillators in isotopologues of isoprene
epoxydiol have already been described earlier.28 Additionally,
given the aforementioned coherent nature of the SFG process,
this outcome may reveal how the vector eld addition that
produces the observed SFG signal may be ultimately inuenced
by deuteration pattern. The band centered at 2840 cmÀ1
disappears completely in the spectra of isotopologues 6 and 7,
likely due to silencing of the allylic methylene group, which
agrees with predictions from the discussed computations,27
though we cannot unambiguously assign this band given that
there exists partial deuteration at that position (ca. 50–70% of
double to single deuteration at C4). Synthetic efforts are
underway to selectively deuterate the allylic and vinyl positions
of a-pinene, though such studies are beyond the scope of this
work. Nevertheless, we caution that the unambiguous assign-
ment of the 2930 cmÀ1 peak will continue to be challenging.
Fig. 3 Sub 1 cmÀ1-resolution ssp-polarized vapor-phase SFG spectra
of (À)-a-pinene and isotopologues 2–7 in contact with fused silica,
with all maximum intensities normalized to 1 and offset for clarity.
formation pathways that populate the aerosol gas:particle
interface with surface-active a-pinene oxidation products such
as pinonaldehyde, pinonic acid, and other proposed higher
order products.29–31 This organic pool has already been shown to
be key for understanding droplet growth32,33 as well as cloud
activation potential.11 The known surface-active a-pinene
oxidation products contain the dimethylcyclobutane group,11
which we hypothesize is net-ordered at the particle surface,
given its nonpolar character. Supporting this hypothesis is the
observation that the SFG spectra of a-pinene-derived SOM
deposited onto Teon lters or optical windows are dominated
by an intense peak at ꢀ2945–50 cmÀ1, with a smaller peak at
ꢀ2880 cmÀ1. Moreover, the lineshapes are highly conserved
through the particle growth process, as SFG spectra of size-
resolved particles containing SOM show.8 In general, the SFG
spectra of a-pinene-derived SOM and of a-pinene vapor in
SFG spectroscopy of isotope-labeled a-pinene-derived SOM
contact with optical windows are similar in lineshape but
Given the results presented in the previous section, iso-
topologues 2–7 should serve as test molecules for probing SOM
shied by ꢀ15 cmÀ1
.
Given this information, and the
8,13,23,24
This journal is © The Royal Society of Chemistry 2019
Chem. Sci.

View Article Online
Chemical Science
Edge Article
results presented in the previous section, deuterating all or part 2950 cmÀ1 region for SOM prepared from isotopologues 5 and
of the four-membered ring should then lead to considerable 7, which are the only isotopologues we studied that feature
reductions in SFG signal intensity from a-pinene-derived SOM a deuterated methylene group on the four-membered ring. It
near ꢀ2945–50 cmÀ1
. We note that our previous work appears that this reduction originates from deuterating the CH2
comparing the spectra of a-pinene-derived SOM using standard group on the four-membered ring, since isotope labeling of the
versus high resolution SFG systems indicates that no gain in methine positions (isotopologue 6) on the same ring do not
spectral resolution is observed in the a-pinene-derived SOM result in any changes in the surface SFG spectrum of the cor-
spectra collected at increased spectral resolving power.13 responding SOM. We conclude that the high-intensity peak
Therefore, it was concluded that the Northwestern University seen in the SFG spectra of a-pinene-derived SOM at ꢀ2950 cmÀ1
SFG system is appropriate for analyzing the surfaces of a- contains considerable contributions (ca. 30%) from the meth-
pinene-derived SOM,13 and all SOM spectra presented herein ylene group on the four-membered ring, while the methyl
were collected using this system.
groups do not contribute.
Fig. 4 shows the standard resolution SFG spectra of SOM
Together, our ndings support the hypothesis that the
derived from isotopologues 2–7 compared to that of SOM surface-active constituents of a-pinene-derived SOM conserve
derived from unlabeled a-pinene. As perhaps anticipated, the the cyclobutane functionality, as the bridge methylene group
SFG spectral lineshapes for SOM derived from isotopologues 2– was found to contribute ca. 30% of spectral intensity in the SFG
4 are largely invariant with methyl group deuteration (Fig. 4A). response at ꢀ2950 cmÀ1. While isotope labeling of these posi-
In contrast, and surprisingly to us, none of the complex line- tions in isotopologues 5 and 7 resulted in signicant changes in
shape changes observed for the ring-labeled isotopologues (5–7) the signal intensities and lineshapes within the vapor phase
are observed for the SOM prepared from them (Fig. 4B). Yet, we SFG spectra, it appears as though the bridge methylene oscil-
note a 30 to 40 percent signal intensity reduction in the lators in the SOM do not contribute nearly as much intensity as
initially expected to the ssp-polarized SFG spectra. It is possible
that this outcome may arise from differences in the orientation
distributions of the cyclobutane moieties present in the SOM
surface-active oxidation products compared to those present in
the precursor isotopologues. However, we expect it is more
likely that the remaining signal intensity around 2950 cmÀ1 in
the SFG spectra of SOM prepared from the bridge methylene
deuterated isotopologues 5 and 7 is attributed to C–H oscilla-
tors of a-pinene and its oxidation products that have not yet
been identied through deuteration.
Conclusions
Dually motivated to synthesize a-pinene isotopologues for
probing the interfacial chemistry of atmospheric organic aero-
sol particles by surface-specic vibrational spectroscopy, and
for other MS-coupled chamber studies within the greater
atmospheric community, we have successfully synthesized
three different isotopologues of a-pinene by modifying
a previous synthesis completed by Thomas and Fallis. These
modications allowed for deuterium labeling of the vinylogous
position, both methine protons, the methylene bridgehead, and
the allylic methylene positions of a-pinene. The synthesis and
isolation of a-pinene isotopologues opened the possibility of
using these isotopologues for vibrational mode elucidation of
the high congested SFG spectrum of a-pinene and its corre-
sponding SOM.18,22
With this goal in mind, we provided SFG spectra of the ring-
and methyl-substituted isotopologues at vapor:solid interfaces
to observe the relative contributions from silenced C–H oscil-
lators on the SFG spectrum of a-pinene. The ring-substituted
isotopologues revealed a signicant (50–80%) suppression of
the SFG signal intensity in the broad band, centered around
2930 cmÀ1, that is emblematic of unlabeled a-pinene. In
contrast, the methyl-deuterated isotopologues we previously
examined22 do not show such a signal intensity reduction,
Fig.
4 Standard resolution ssp-polarized SFG spectra of SOM
produced from (À)-a-pinene (black trace) and a-pinene iso-
topologues (A) 2–4 and (B) 5–7 in contact with fused silica. Intensities
are scaled based the highest intensity peak of the (À)-a-pinene SOM
SFG spectrum collected on the same day as each of the isotopologue
SOM samples.
Chem. Sci.
This journal is © The Royal Society of Chemistry 2019

View Article Online
Edge Article
Chemical Science
emphasizing the coherent nature of the SFG signal generation to ensure that ozonolysis products were generated. Excess
process from the various C–H oscillators in the highly strained ozone (53 ppm) was introduced via a ow inlet arranged
bicyclic system of a-pinene, and the need for complementary perpendicular to the VOC inlet to promote rapid mixing passed
experimental and computational studies for further spectral through the reactor (4 SLPM ow rate) and ensure the precursor
deconvolution. Future work will include accessing the C–D was fully reacted. The particles were produced in the absence of
region, carrying out internal heterodyne phase-resolved nucleating seed particles and collected for 10 h on Teon lters
measurements, improving SFG calculations, and synthesizing (0.2–0.45 mm pore size, Sigma-Aldrich Inc.) with a collection
other a-pinene isotopologues.
efficiency of 98%. From the ow rate, collection time, and
By subsequently preparing and collecting SFG spectra of particle mass concentration (obtained from a scanning mobility
SOM derived from the a-pinene isotopologues, we then made particle sizer), the mass of the particles collected on the lters
the rst attempts thus far toward identifying the surface oscil- was estimated in the range of 0.1–1.8 mg (see ESI† for more
lators in a-pinene-derived SOM. From the SOM SFG data, we details on SOM collection). Following particle collection,
speculate the presence of the dimethylcyclobutane ring at the samples were stored in a À20 ꢁC freezer and sealed using Teon
SOM surface, as SOM prepared from the ozone oxidation of two tape and paralm. Samples were warmed to room temperature
isotopologues featuring a deuterated bridge methylene group prior to SFG measurements.
shows up to 40% less signal intensity than SOM prepared from
the methyl-substituted isotopologues to which we had access.
Vibrational sum frequency generation spectroscopy
Yet, given the substantially higher (up to 80%) SFG signal
intensity reductions observed in the SOM precursor iso-
topologues that are bridge-deuterated, we conclude that SOM
formation leads to other SFG active modes in the C–H stretch-
ing region at the particle surface.
SFG spectra of (À)-a-pinene and the synthesized a-pinene iso-
topologues (5–7) were recorded in the C–H stretching region
using standard 10–15 cmÀ1 (Northwestern University, NU)23,43–45
and 0.6 cmÀ1 (Pacic Northwestern National Laboratory,
PNNL)21,22,46,47 spectral resolution broadband SFG laser systems.
A comprehensive description of the sample congurations and
experimental setups has been published previously.8,18,24,45 C–H
SFG spectra of the molecular standards were collected by
exposing a solid fused silica or calcium uoride optical window
to the equilibrium vapor pressure of the molecule of interest
and subsequently measuring the vapor:solid interface. SFG
spectra of synthetic SOM were measured in the C–H stretching
region by pressing the Teon lter containing the collected
material to a solid fused silica optical substrate. Spectra were
measured using the ssp polarization combination and at labo-
ratory ambient temperature and relative humidity. The angles
of incidence for the visible and IR beams were 45ꢁ and 55ꢁ from
surface normal respectively for the high resolution system and
45ꢁ and 60ꢁ from surface normal respectively for the standard
resolution system.
The standard resolution spectra reported are an average of 7
individual spectra each taken for 2 minutes each. For the
standard resolution spectra, the IR energy distribution was
accounted for by normalizing the data to the ppp-polarized
nonresonant SFG signal of gold deposited onto fused silica.
Frequencies were calibrated using a polystyrene lm.13,14,45,48
The high-resolution spectra are reported as an average of 3–5
individual acquisitions each recorded for 5 minutes. The data
points were binned by 5 points, or by 1.73 cmÀ1, in Igor Pro
Version 6.11 (WaveMetrics, Lake Oswego, OR, USA). SFG
intensities in the 0.6 cmÀ1 resolution spectra were normalized
to the ppp-polarized nonresonant SFG response of clean z-cut a-
quartz, and frequencies were calibrated again using
polystyrene.
Ultimately, our synthesis of isotopically labeled SOM and
analysis of their surface-active constituents opens the door to
future investigations of the dynamics of gas-phase VOC
adsorption on and/or uptake into the particle phase under
varying tropospherically realistic conditions.13,16 Beyond the
SFG studies presented here, our syntheses of a-pinene iso-
topologues have been used in mechanistic chamber-based
studies, including the demonstration of particle–particle
chemical exchange using single-particle mass spectrometry,34 as
well as identication of intramolecular hydrogen migration in
autoxidation leading to highly oxygenated multifunctional
organic compounds.35 More broadly speaking, synthetically
accessing isotopologues of terpene hydrocarbons that are not
commercially available offers the scientic community at large
new opportunities to elucidate mechanistic questions in aerosol
formation and growth, terpenoid biosynthesis, sustainable
biofuels, and food and avor chemistry, among other areas of
study.34,36–39
Methods
Synthesis of a-pinene isotopologues
See ESI† for the synthetic procedures that were followed.
Collection of synthetic secondary organic material (SOM)
A complete description of the Harvard ow tube reactor used in
this work has been previously published.8,13,40 In brief, SOM was
produced from (À)-a-pinene (99%, optical purity ee: 97% (GLC),
Sigma Aldrich, Inc.) and the synthesized a-pinene iso-
topologues by introducing a solution of each precursor in 1-
butanol (1 : 625 v/v dilution ratio)8,41,42 at selected injection rates
that altered the gas phase concentration in the range of 125–300
ppb, which subsequently changed the organic particle mass
Conflicts of interest
loading in the ow tube. 1-Butanol was used as an OH scavenger There are no conicts to declare.
This journal is © The Royal Society of Chemistry 2019
Chem. Sci.

View Article Online
Chemical Science
Edge Article
13 M. Shrestha, Y. Zhang, M. A. Upshur, P. Liu, S. L. Blair,
H.-f. Wang, S. A. Nizkorodov, R. J. Thomson, S. T. Martin
and F. M. Geiger, J. Phys. Chem. A, 2015, 119, 4609–4617.
14 C. J. Ebben, B. F. Strick, M. A. Upshur, H. M. Chase,
J. L. Achtyl, R. J. Thomson and F. M. Geiger, Atmos. Chem.
Phys., 2014, 14, 2303–2314.
Acknowledgements
MAU, AGB, and HMC gratefully acknowledge NSF Graduate
Research fellowships. MAU also acknowledges a National
Aeronautics and Space Administration Earth and Space (NASA
ESS) fellowship and support from a PEO. Scholar Award. AT
acknowledges support by the US Department of Energy (DOE),
Office of Science, Office of Basic Energy Sciences, Division of
Materials Science and Engineering. FMG gratefully acknowl-
edges a Friedrich Wilhelm Bessel Prize from the Alexander von
Humboldt Foundation. This work was supported by the
National Science Foundation (NSF) under Grant No. CHE-
1607640 (to FMG, STM, and RJT) and CHE-1212692 (to RJT
and FMG). A portion of the research was performed using EMSL
(User Proposal #49291), a DOE Office of Science User Facility
sponsored by the Office of Biological and Environmental
Research. We thank Prof. Hongfei Wang (Fudan University) for
helpful discussions.
´
15 A. G. Be, H. M. Chase, Y. Liu, M. A. Upshur, Y. Zhang,
A. Tuladhar, Z. A. Chase, A. D. Bellcross, H.-F. Wang,
Z. Wang, V. S. Batista, S. T. Martin, R. J. Thomson and
F. M. Geiger, ACS Earth Space Chem., 2019, 3, 158–169.
16 H. M. Chase, J. Ho, M. A. Upshur, R. J. Thomson, V. S. Batista
and F. M. Geiger, J. Phys. Chem. A, 2017, 121, 3239–3246.
17 H. M. Chase, B. Rudshteyn, B. T. Psciuk, M. A. Upshur,
B. F. Strick, R. J. Thomson, V. S. Batista and F. M. Geiger,
J. Phys. Chem. B, 2016, 120, 1919–1927.
18 A. L. Mifflin, L. Velarde, J. Ho, B. T. Psciuk, C. F. A. Negre,
C. J. Ebben, M. A. Upshur, Z. Lu, B. L. Strick,
R. J. Thomson, V. S. Batista, H.-F. Wang and F. M. Geiger,
J. Phys. Chem. A, 2015, 119, 1292–1302.
19 M. A. Upshur, B. F. Strick, V. F. McNeill, R. J. Thomson and
F. M. Geiger, Atmos. Chem. Phys., 2014, 14, 10731–10740.
20 C. J. Ebben, S. R. Zorn, S.-B. Lee, P. Artaxo, S. T. Martin and
F. M. Geiger, Geophys. Res. Lett., 2011, 38, L16807.
Notes and references
1 J. H. Kroll and J. H. Seinfeld, Atmos. Environ., 2008, 42, 3593– 21 A. L. Mifflin, L. Velarde, J. Ho, B. T. Psciuk, C. F. Negre,
3624.
2 B. Noziere, M. Kalberer, M. Claeys, J. Allan, B. D'Anna,
C. J. Ebben, M. A. Upshur, Z. Lu, B. F. Strick,
R. J. Thomson, V. S. Batista, H.-F. Wang and F. M. Geiger,
J. Phys. Chem. A, 2015, 119, 1292–1302.
`
´
S. Decesari, E. Finessi, M. Glasius, I. Grgic, J. F. Hamilton,
T. Hoffmann, Y. Iinuma, M. Jaoui, A. Kahnt, C. J. Kampf, 22 M. A. Upshur, H. M. Chase, B. F. Strick, C. J. Ebben, L. Fu,
I. Kourtchev, W. Maenhaut, N. Marsden, S. Saarikoski,
J. Schnelle-Kreis, J. D. Surratt, S. Szidat, R. Szmigielski and
A. Wisthaler, Chem. Rev., 2015, 115, 3919–3983.
H. Wang, R. J. Thomson and F. M. Geiger, J. Phys. Chem. A,
2016, 120, 2684–2690.
23 C. J. Ebben, I. S. Martinez, M. Shrestha, A. M. Buchbinder,
¨
¨¨
¨ ¨
3 J. K. Holopainen, M. Kivimaenpaa and S. A. Nizkorodov,
Trends Plant Sci., 2017, 22, 744–753.
A. L. Corrigan, A. Guenther, T. Karl, T. Petaja, W. W. Song,
S. R. Zorn, P. Artaxo, M. Kulmala, S. T. Martin,
L. M. Russell, J. Williams and F. M. Geiger, Atmos. Chem.
Phys., 2011, 11, 10317–10329.
4 A. P. Ault and J. L. Axson, Anal. Chem., 2017, 89, 430–452.
5 M. Kanakidou, J. H. Seinfeld, S. N. Pandis, I. Barnes,
F. J. Dentener, M. C. Facchini, R. Van Dingenen, B. Ervens, 24 C. J. Ebben, M. Shrestha, I. S. Martinez, A. L. Corrigan,
¨ ¨
A. Nenes, C. J. Nielsen, E. Swietlicki, J. P. Putaud,
Y. Balkanski, S. Fuzzi, J. Horth, G. K. Moortgat,
R. Winterhalter, C. E. L. Myhre, K. Tsigaridis, E. Vignati,
E. G. Stephanou and J. Wilson, Atmos. Chem. Phys., 2005, 5,
1053–1123.
6 K. A. Prather, C. D. Hatch and V. H. Grassian, Annu. Rev. Anal.
Chem., 2008, 1, 485–514.
7 Y. Qiu and V. Molinero, J. Am. Chem. Soc., 2015, 137, 10642–
10651.
A. A. Frossard, W. W. Song, D. R. Worton, T. Petaja,
J. Williams, L. M. Russell, M. Kulmala, A. H. Goldstein,
P. Artaxo, S. T. Martin, R. J. Thomson and F. M. Geiger, J.
Phys. Chem. A, 2012, 116, 8271–8290.
25 M. T. Thomas and A. G. Fallis, J. Am. Chem. Soc., 1976, 98,
1227–1231.
26 M. T. Thomas and A. G. Fallis, Tetrahedron Lett., 1973, 14,
4687–4690.
27 J. Ho, B. T. Psciuk, H. M. Chase, B. Rudshteyn, M. A. Upshur,
L. Fu, R. J. Thomson, H.-F. Wang, F. M. Geiger and
V. S. Batista, J. Phys. Chem. C, 2016, 120, 12578–12589.
8 M. Shrestha, Y. Zhang, C. J. Ebben, S. T. Martin and
F. M. Geiger, J. Phys. Chem. A, 2013, 117, 8427–8436.
9 N. Sharma, S. China, J. Bhandari, K. Gorkowski, M. Dubey, 28 H. M. Chase, B. T. Psciuk, B. L. Strick, R. J. Thomson,
R. A. Zaveri and C. Mazzoleni, Geophys. Res. Lett., 2018, 45,
11474–11482.
V. S. Batista and F. M. Geiger, J. Phys. Chem. A, 2015, 119,
3407–3414.
10 A. J. Prenni, M. D. Petters, S. M. Kreidenweis, P. J. DeMott 29 K. Kristensen, K. L. Enggrob, S. M. King, D. R. Worton,
and P. J. Ziemann, J. Geophys. Res.: Atmos., 2007, 112,
D10223.
S. M. Platt, R. Mortensen, T. Rosenoern, J. D. Surratt,
M. Bilde, A. H. Goldstein and M. Glasius, Atmos. Chem.
Phys., 2013, 13, 3763–3776.
´
11 A. G. Be, M. A. Upshur, P. Liu, S. T. Martin, F. M. Geiger and
R. J. Thomson, ACS Cent. Sci., 2017, 3, 715–725.
12 Y. Wu, W. Li, B. Xu, X. Li, H. Wang, V. F. McNeill, Y. Rao and
H.-L. Dai, J. Phys. Chem. Lett., 2016, 7, 2294–2297.
30 J. W. DePalma, A. J. Horan, W. A. Hall IV and M. V. Johnston,
Phys. Chem. Chem. Phys., 2013, 15, 6935–6944.
Chem. Sci.
This journal is © The Royal Society of Chemistry 2019

View Article Online
Edge Article
Chemical Science
31 M. P. Tolocka, M. Jang, J. M. Ginter, F. J. Cox, R. M. Kamens 39 N. Duhamel, D. Martin, R. Larcher, B. Fedrizzi and
and M. V. Johnston, Environ. Sci. Technol., 2004, 38, 1428–
1434.
32 C. R. Ruehl, J. F. Davies and K. R. Wilson, Science, 2016, 351,
1447–1450.
33 N. Sareen, A. N. Schwier, T. L. Lathem, A. Nenes and
D. Barker, Tetrahedron Lett., 2016, 57, 4496–4499.
40 Y. Zhang, P. Liu, Z. Gong, F. M. Geiger and S. T. Martin, J.
Visualized Exp., 2018, e55684, DOI: 10.3791/55684.
41 Y. Gao, W. A. Hall and M. V. Johnston, Environ. Sci. Technol.,
2010, 44, 7897–7902.
V. F. McNeill, Proc. Natl. Acad. Sci. U. S. A., 2013, 110, 42 Y. Zhang, M. S. Sanchez, C. Douet, Y. Wang, A. P. Bateman,
2723–2728.
Z. Gong, M. Kuwata, L. Renbaum-Wolff, B. B. Sato, P. F. Liu,
A. K. Bertram, F. M. Geiger and S. T. Martin, Atmos. Chem.
Phys., 2015, 15, 7819–7829.
34 Q. Ye, M. A. Upshur, E. S. Robinson, F. M. Geiger,
R. C. Sullivan, R. J. Thomson and N. M. Donahue, Chem,
2018, 4, 318–333.
43 A. M. Buchbinder, N. A. Ray, J. Lu, R. P. Van Duyne,
P. C. Stair, E. Weitz and F. M. Geiger, J. Am. Chem. Soc.,
2011, 133, 17816–17823.
35 X. Zhang, A. T. Lambe, M. A. Upshur, W. A. Brooks, A. Gray
´
Be, R. J. Thomson, F. M. Geiger, J. D. Surratt, Z. Zhang,
A. Gold, S. Graf, M. J. Cubison, M. Groessl, J. T. Jayne, 44 O. Esenturk and R. A. Walker, J. Phys. Chem. B, 2004, 108,
D. R. Worsnop and M. R. Canagaratna, Environ. Sci.
Technol., 2017, 51, 5932–5940.
36 J. S. Dickschat, Eur. J. Org. Chem., 2017, 33, 4872–4882.
10631–10635.
45 G. Y. Stokes, A. M. Buchbinder, J. M. Gibbs-Davis,
K. A. Scheidt and F. M. Geiger, J. Phys. Chem. A, 2008, 112,
11688–11698.
¨
37 R. Mewalal, D. K. Rai, D. Kainer, F. Chen, C. Kulheim,
G. F. Peter and G. A. Tuskan, Trends Biotechnol., 2017, 35, 46 L. Velarde, X.-y. Zhang, Z. Lu, A. G. Joly, Z. Wang and
227–240. H.-f. Wang, J. Chem. Phys., 2011, 135, 241102.
38 C. M. Kenseth, Y. Huang, R. Zhao, N. F. Dalleska, 47 L. Fu, S.-L. Chen and H.-F. Wang, J. Phys. Chem. B, 2016, 120,
J. C. Hethcox, B. M. Stoltz and J. H. Seinfeld, Proc. Natl.
Acad. Sci. U. S. A., 2018, 115, 8301–8306.
1579–1589.
48 A. M. Buchbinder, E. Weitz and F. M. Geiger, J. Phys. Chem. B,
2010, 114, 554–566.
This journal is © The Royal Society of Chemistry 2019
Chem. Sci.