Liquid ToF-SIMS revealing the oil, water, and surfactant interface evolution

Yanjie Shena, Jenn Yaoa, Jiyoung Sona, Zihua Zhub and Xiao-Ying Yu*a
aEnergy and Environment Directorate, Pacific Northwest National Laboratory, Richland, WA 99354, USA. E-mail:
bEnvironmental Molecular Sciences Laboratory, Pacific Northwest National Laboratory, Richland, WA 99354, USA

Received 31st January 2020 , Accepted 29th February 2020

First published on 2nd March 2020

Bilge water from ships is regarded as a major pollutant in the marine environment. Bilge water exists in a stable oil-in-water (O/W) emulsion form. However, little is known about the O/W liquid–liquid (l–l) interface. Traditional bulk characterization approaches are not capable of capturing the chemical changes at the O/W l–l interface. Although surfactants are deemed essential in droplet formation, their roles in bilge water stabilization have not been fully revealed. We have utilized novel in situ chemical imaging tools including in situ scanning electron microscopy (SEM) and in situ time-of-flight secondary ion mass spectrometry (ToF-SIMS) to study the evolving O/W interface using a NAVY bilge model for the first time. The droplet size distribution (DSD) does not change significantly without the addition of X-100 surfactants under static or rocking conditions. Both the oil components and the water clusters are shown to evolve over time at the O/W droplet interface by in situ liquid SIMS imaging. Of particular interest to droplet stabilization, the contribution of surfactants to the aged bilge droplets becomes more significant as the droplet size increases. The higher mass surfactant component does not appear on the droplet surface immediately while many lower mass surfactants are solvated inside the droplet. We have provided the first three-dimensional images of the evolving O/W interface and demonstrated that in situ surface chemical mapping is powerful enough to reveal the complex and dynamic l–l interface in the liquid state. Our observational insights suggest that surfactants are important in mediating droplet growth and facilitating effective separation of bilge water emulsion.


Due to rapid development of sea transportation, pollution in the marine environment has caught attention worldwide. Bilge water, a wastewater mainly formed from ships, is regarded as a major pollutant to the marine environment.1–3 Typically, bilge water is a mixture of seawater, oily fluids, detergent fluids, lubricants, particles, and other similar wastes; and it usually exists in a stable oil-in-water (O/W) emulsion form.4–7 It contains hazardous substances and can cause damage to the water quality and endanger the environment. Moreover, it is toxic to aquatic biological systems as well as human health. Several chemical and physical methods have been developed to treat bilge water, such as flotation, separation by centrifugation, filtration, and coagulation.8 However, the treatment of bilge emulsion in the presence of oil, detergent, and particles is difficult using physical methods, especially when the diameter of the emulsified oil droplet is below 20 μm.9 Known methods of bilge water treatment are not efficient universally, including ultrafiltration, wet air oxidation, electrocoagulation, the use of biotechnology, and membrane distillation.8,10–13 Thus, new methods to treat bilge water are urgently needed. To facilitate method development, it is important to understand the properties of bilge water emulsion, including the formation, stability, breaking, and evolution of emulsion droplets at the O/W liquid–liquid (l–l) interface under conditions relevant to onboard ships.14,15

A few recent papers reported experimental studies of O/W emulsion stability. Optical microscopy including confocal laser scanning microscopy, atomic force microscopy, and scanning electron microscopy (SEM) were used to characterize O/W emulsions.16–19 Salinity and pH were shown to be associated with emulsion breaking and separation processes.14,20 The emulsifier dosage, ratio of oil to water, stirring intensity, emulsifying temperature, and mixing time also have an effect on droplet stability using conventional measurements.21–24 However, the evolution of chemical changes at the O/W l–l interface cannot be characterized using known bulk approaches.

Our group developed a novel vacuum compatible microreactor, or System for Analysis at the Liquid Vacuum Interface (SALVI), enabling multimodal imaging of liquids using vacuum instruments.25–27 In situ liquid time-of-flight secondary ion mass spectrometry (ToF-SIMS) has proven to be a powerful tool to study liquid surfaces, air–liquid (a–l), l–l, and solid–liquid interfaces. It provides detailed elemental, molecular, and isotopic mapping of the surface or interface in biology, materials, and the environment. ToF-SIMS itself provides more than one mode of measurement, offering spectral, two-dimensional (2D), and three-dimensional (3D) chemical mapping of surfaces and interfaces.28–30

To further the emulsion interfacial research using in situ imaging and based on our recent dry SIMS surface characterization, we propose the following hypotheses of the evolution of the O/W interface in bilge water emulsion: (1) water plays a critical role in the O/W interface; (2) low mass oil and detergent components (i.e., m/z+ 0–250) and relative high mass surfactant components (i.e., m/z+ 250–500) appear immediately in the fresh bilge water surface and persist in the aged bilge water; (3) high mass surfactant components (i.e., m/z+ 500–800) may take time to migrate to the surface; and (4) surface surfactant composition has an effect on the emulsion DSD. This work is the first in situ chemical imaging study of emulsion using new chemical imaging tools including in situ scanning electron microscopy (SEM) and in situ ToF-SIMS. In situ liquid SEM has been first used to determine the change in droplet size distribution (DSD) of oil droplets between fresh and aged bilge water emulsions. In situ liquid SIMS has been used to study the evolution of O/W emulsions at the l–l interface as the surface ages and given visualization of the surfactant, oil components, and water in the liquid state. Optical microscopy was used to quickly assess the effect of surfactants on the DSD evolution of bilge water with and without X-100 enhancement over the course of six days.


Emulsion sample preparation

Components of the oil mix and detergent mix were provided by our collaborator at the Naval Surface Warfare Center, Carderock Division; and they were used to generate synthetic bilge emulsion. The synthetic emulsion consists of a liquid mixture of 10 mL oil mix (Navy Standard Bilge Mix (NSBM) #4) and 1 mL detergent mix. The detailed procedure has been reported previously.14 Briefly, the oil mix contains 50% diesel fuel marine (MIL-PRF-16884N), 25% 2190 TEP steam lube oil (MIL-PRF-17331K), and 25% 9250 diesel lube oil (MIL-PRF-9000L). The detergent mix contains 50% type 1 general purpose detergent (MIL-D-16791G (1)), 25% commercial detergent Tide Ultra (liquid), and 25% degreasing solvent (MIL-PRF-680C, type III).15 Fresh bilge refers to emulsions prepared following the reported procedure and immediately followed by in situ analysis without sitting or rocking. Aged bilge refers to emulsions analyzed 24 h after preparation. To study the effect of surfactants, a bilge sample was prepared with extra X-100 surfactants compared to the NAVY recipe. This sample is referred to as X100 + fresh bilge. More details of all the liquid emulsion samples are given in Table S1 (ESI).

The prepared bilge liquids are injected into the SALVI microchannel prior to in situ SEM and in situ liquid ToF-SIMS chemical imaging. Optical microscopy was also used to obtain images for DSD determination. To obtain clear images of individual oil droplets, the emulsion sample was diluted using deionized (DI) water before in situ SEM imaging and optical imaging. As a comparison, dried droplets were analysed using static ToF-SIMS to offer better mass accuracy. Additional experimental details of static ToF-SIMS sample preparation were presented previously15 and in the ESI.

SALVI device fabrication

The details of SALVI device fabrication were described in our previous publications.31,32 In general, the SALVI device consists of a polydimethylsiloxane (PDMS) block with a 200 μm (width) × 300 μm (depth) microfluidic channel inside, a 100 nm thick silicon nitride (SiN) membrane (Norcada, Canada) with a window of 1.5 × 1.5 mm2 on a supporting silicon frame (7.5 × 7.5 mm2). In addition, PTFE tubing is used to introduce a liquid to the microfluidic channel. Fig. 1c and d show the schematic of SALVI coupled with in situ liquid SEM and in situ liquid SIMS, respectively.
image file: d0cp00528b-f1.tif
Fig. 1 The schematic showing bilge water and liquid sample preparation for multimodal imaging analysis: (a) liquid bilge water emulsions after mechanical mixing and sonication; (b) a SALVI device filled with emulsion; (c) in situ liquid SEM measurements of bilge water; (d) in situ liquid ToF-SIMS analysis of bilge water; (e) DSD determination using in situ SEM; and (f) a representative ToF-SIMS mass spectrum and 3D image.

In situ liquid SEM

The SiN membrane detection window of the SALVI device was coated with 10 nm thick carbon prior to in situ SEM imaging to reduce the charging effect. The microfluidic device was injected with approximately 100 μL liquid sample (e.g., bilge water emulsion) and sealed with a PEEK union. The loaded SALVI device was mounted on the SEM stage and stabilized with double-sided copper tape to further reduce charging. More details were reported previously.33,34 The backscattered electron (BSE) images were acquired in the high vacuum mode. The vacuum in the chamber was maintained at ∼4 × 10−6 Torr. The accelerating voltage and current were set at 20 kV and 0.11 nA, respectively.

The SEM images of the bilge droplets were pre-processed using ImageJ to enhance the contrast. The image analysis of the DSDs of fresh and aged emulsions was done by fitting the data with a lognormal distribution model in MATLAB (MathWorks, MATLAB 2018b). The lognormal model is often used as a default model for regression analysis of particle size.35 Table S2 (ESI) provides fitting results of the DSD obtained from in situ SEM.

Optical microscopy

Optical microscopy (Nikon Eclipse TE2000-U) was used to study the effect of surfactant on the droplet size growth in bilge water emulsion with and without X-100 over six days under both static and rocking conditions. In static conditions, the emulsion mixture was allowed to sit in the hood for six days and an aliquot was taken from the mixture for DSD determination daily. No additional sonication was used before analysis. During rocking, the emulsion mixture was stationed in a platform shaker (INNOVA 2300, New Brunswick Co., Inc.) to shake at 180 rpm continuously for six days and DSD determination was determined daily. All samples were analysed using 1000 times of magnification and droplet images were collected using a digital CCD camera (Hamamatsu, C4742-95-12HR) on an inverted microscope (Nikon Eclips TE2000-U). The pixel size information was processed and given by the MetaMorph® software with a resolution of ∼92 nm. The camera captured images at 4000 (H) by 2624 (V) pixels. At least six images were acquired for each sample. A Matlab program was used to determine the DSD (Table S3, ESI).


A ToF-SIMS V instrument (IONTOF GmbH, Münster, Germany) was used for in situ liquid SIMS. During liquid ToF-SIMS, the SiN membrane was punched through using a pulsed Bi3+ ion beam (25 KeV, 10 kHz, pulse width 150 ns). As the primary ion beam made a hole of 2 μm in diameter on the SiN membrane, the SIMS image of the liquid surface was collected for 100 s for 2D and 3D image analysis. Then the Bi3+ ion beam continuously sputtered on the liquid surface for 200 s with a reduced pulse width of 80 ns to acquire spectral measurements. Before each analysis, the 1 KeV O2+ beam was used to clean the SiN membrane with a scanning area of 200 μm × 200 μm to remove surface contamination from the atmosphere. At least 4 data points were acquired for each sample in the positive and negative ion mode, respectively. The main chamber pressure was maintained around 4 × 10−7 mbar during analysis. The analysis depth of liquid ToF-SIMS is approximately less than 10 nm.31,36,37 Fig. S3 and S4 (ESI) show the reproducibility of the liquid ToF-SIMS mass spectra. During static ToF-SIMS analysis of dried emulsion samples, a standard procedure was followed.15

The ToF-SIMS mass spectral data was processed using IONTOF Surface Lab 6.3 software. The mass spectra were calibrated using OH (m/z 17), SiC2H5O (m/z 73), Si2C3H9O3 (m/z 149), and Si3C5H15O4 (m/z 223) in the negative mode; and CH3+ (m/z+ 15), (CH3)3Si+ (m/z+ 73), Si2OC5H15+ (m/z+ 147), Si3C5H15O3+ (m/z+ 207), and Si4O4C7H21+ (m/z+ 281) in the positive mode, respectively. Spectral principal component analysis (PCA) was conducted using MATLAB software.

In the first round of PCA, the known interference peaks were removed, for example, m/z 1 H, 12 CH, 16 O, 24 C2, 25 C2H, 28 Si, 29 SiH, 41 SiCH, SiCH2, 44 SiO, 45 SiOH, 58 SiCH2O, 59 SiCH3O,60 SiO2, 73 SiC2H5O, 75 SiCH3O2, 149 Si2C3H9O3, 163 Si2C5H15O2, 223 Si13C5H15O4, and 237 Si3C7H21O3 in the negative mode; and m/z+ 28 Si+, 59 CH3SiO+, 73 (CH3)3Si+, 147 Si2O5H15+, 207 Si3C5H15O3+, 281 Si4O4C7H21+, 369 Si5O4C11H33+, 209 Bi+, 418 Bi2+, and 657 Bi3+ in the positive mode. In the second round of PCA, peaks were selected using the following criteria: (1) the intensity of the peak in the ToF-SIMS mass spectrum was significantly higher than its neighboring peaks (i.e., S/N >3); (2) product peaks were identified using reference spectra of dry samples in the static SIMS analysis;15 (3) known water cluster peaks; and (4) the peaks in the SIMS mass spectrum with m/z+ or m/z > 40. The mass calibrated data were treated by normalization to the total selected ion intensity, square-root transformation, and mean entering before running spectral PCA.25,38

Results and discussions

In situ SEM determination of DSD

Fig. 2 depicts the representative SEM BSE images of the fresh and aged bilge droplets. Fig. S1a and S1b (ESI) show more SEM results. Both fresh and aged bilge have mainly monomodal size distribution. Table S2 (ESI) gives the summary of the DSD determination using in situ liquid SEM. The measurements illustrate that bilge emulsion droplets change slightly in size over time. The SEM images show that the droplet increases from the mean diameter of 1.69 ± 0.43 μm in the fresh bilge to 3.12 ± 0.26 μm in the aged one after one day using the bimodal fitting. When using the lognormal distribution to analyze DSD, the results are slightly different, i.e., 1.98 ± 0.55 μm and 3.35 ± 1.06 μm for fresh and aged bilge emulsion, respectively. The results using bimodal and lognormal fittings are slightly different, however, showing a similar trend. This finding confirms that the droplet does not remain the same once formed and changes over time.
image file: d0cp00528b-f2.tif
Fig. 2 Representative in situ liquid SEM images (a and b) and histograms showing fresh and aged bilge DSD determination using the lognormal (c and d) and bimodal Gaussian distribution (e and f), respectively.

Coagulation does occur even in fresh bilge emulsion. A bimodal Gaussian distribution fit was conducted combining all valid in situ SEM images (Fig. 2). The observations show that only a small percentage of bilge droplets quickly coalesce. The bimodal fitting results suggest that the majorities of drops are in the primary mode, i.e., 1.57 ± 0.51 μm and 3.15 ± 1.49 μm for fresh and one-day aged bilge emulsion (abbreviated as aged hereafter). The secondary mode is 3.81 ± 1.79 μm for fresh and 25.33 ± 1.61 μm for aged bilge emulsion. Thus, the lognormal fit is reasonable to explain most data. Larger droplets are easier to handle in bilge water separation. Fresh and aged bilge water emulsions tend to be less than 4 μm, thus not so easy to treat. We focus on the in situ surface chemical analysis of the primary mode or the smaller droplets with a diameter of one to several micrometers in this work.

Bilge water DSD evolution

Optical microscopy was used to quickly assess DSD evolution of bilge water with and without X-100 addition over the course of six days at static and gently rocking conditions (Fig. S2a and b, ESI). In static conditions, DSD increased from 1.29 ± 0.51 μm from the freshly prepared bilge to 1.87 ± 1.03 μm for that on day 6 without the addition of X-100 (Fig. S2a, ESI). When adding X-100, coagulation was immediately significant. The DSD in the latter case ranged from 4.66 ± 4.72 μm from the freshly prepared emulsion sample to 17.54 ± 50.64 μm on day 6. The X-100 surfactant seems to have a more significant effect on the emulsion DSD growth compared to just time alone. The large variance of DSDs in the X-100 enhanced bilge water sample shows that extremely large droplets frequently form under static conditions. This observation suggests that the X-100 surfactant can promote coagulation of droplets to a large size randomly as time elapses.

In contrast, when gently rocking the bilge water samples with and without X-100 surfactants, the results are different. The DSD does not change significantly for bilge water as expected (Fig S2b, ESI), because shaking helps prevent coagulation and mimics the gentle collisions encountered in flowing water. However, when adding X-100 surfactant to the bilge water, the DSD under rocking conditions shows a growth pattern in six days despite a slight delay compared to that in static conditions. The droplets grow from 5.17 ± 5.76 μm to 19.01 ± 15.60 μm. This finding supports the hypothesis that surfactants like X-100 have a significant effect on the emulsion DSD growth mainly by mediating the surface chemistry. In addition, coagulation is significant in rocking conditions, like in static conditions. In contrast, the DSD grows much larger when adding extra X-100 surfactant to the bilge water, further illustrating the effect of surface chemistry on DSD change. More information on DSD is summarized in Table S3 (ESI). In the following, we focus on the effect of surfactants on the evolved O/W interfacial chemistry.

In situ ToF-SIMS imaging of the O/W interface

Looking into the positive ion mass comparison plots (Fig. 3 and Fig. S5, ESI), oil components are observed in the fresh bilge, aged bilge, and X-100 + fresh bilge in the low mass range (i.e., m/z+ 1–250). Most of the characteristic peaks are identified as hydrocarbon fragments, such as m/z+ 41 C3H5+, 57 C4H9+, 67 C5H7+, 69 C5H9+, 71 C5H11+, 81 C6H9+, 91 C7H7+, 105 C8H9+, and 121 C9H13+. In the higher mass range (i.e., m/z+ 250–500), only a small fraction of high mass peaks from the oil mix are observed in the fresh bilge, aged bilge, and X100 + fresh bilge samples. In the high mass range (i.e., m/z+ 500–800), no significant oil peaks are observed in the bilge samples. We previously studied dried bilge samples using static ToF-SIMS and hypothesized that the bilge surface would evolve over time.15 However, water is lost in dry sample analysis. In this work using liquid SIMS, our results provide a more direct evidence that oil components mainly contribute to the peaks in the low mass range. They appear immediately in the fresh bilge surface and persist in the aged bilge. Lower mass peaks of the detergent components have a similar behavior. In contrast, higher mass detergent components show different behaviors.
image file: d0cp00528b-f3.tif
Fig. 3 In situ liquid ToF-SIMS spectral comparison of synthetic bilge water emulsions and key components in m/z+ 0–250. Green, red, and blue bars represent oil, detergent, and water cluster peaks, respectively.

Spectral composition between static and in situ ToF-SIMS

High mass resolution analysis using dry samples was performed to provide confident reference spectra for peak identification in liquid ToF-SIMS. The details of dry sample preparation, static ToF-SIMS analysis, and peak identification were reported recently.15 Comparisons of the peaks in static dry samples and dynamic liquid samples show that the mass to charge (m/z) ratios are in good agreement. Thus, peak identification using unit mass in liquid SIMS is reasonable, and it is based on reference spectral analysis in static ToF-SIMS with a higher mass accuracy. The key peak identification is presented in Table 1 and more information is shown in Tables S4 and S5 (ESI).
Table 1 Key possible peak identification in the positive and negative mode
m/z+ exact m/z+ obs. Formula Chemical description Ref.
a Based on the molecular weight.b Reference is from the PubChem database.41
41.039 41 C3H5+ Hydrocarbon 39
43.055 43 C3H7+ Hydrocarbon 39
55.055 55 C4H7+ Hydrocarbon 39
57.070 57 C4H9+ Hydrocarbon 39
67.055 67 C5H7+ Hydrocarbon 39
69.070 69 C5H9+ Hydrocarbon 39
71.086 71 C5H11+ Hydrocarbon 39
81.070 81 C6H9+ Hydrocarbon 39
91.055 91 C7H7+ Hydrocarbon 39
105.070 105 C8H9+ Hydrocarbon 39
121.102 121 C9H13+ Hydrocarbon 39
243.121 243 C10H20O5Na+ Fragment of TPGS 15
507.278 507 C22H44O11Na+ Fragment of TPGS 40
551.304 551 C24H48O12Na+ Fragment of TPGS 40
595.330 595 C26H52O13Na+ Fragment of TPGS 40
639.356 639 C28H56O14Na+ Fragment of TPGS 40
683.382 683 C30H60O15Na+ Fragment of TPGS 40

m/zexact m/zobs. Formula Chemical description Ref.
49.008 49 C4H Hydrocarbon a
63.023 63 C5H3 Hydrocarbon a
79.055 79 C6H7 Hydrocarbon a
105.055 105 C4H9O3 Polyethylene glycol b
205.217 205 C12H29O2 Polyethylene glycol b
221.081 221 C12H13O4 Diethyl phthalate b
255.232 255 C16H31O2 n-Hexadecanoic acid b
325.063 325 C18H14O4P Triphenyl phosphate b
337.238 337 C20H33O4 Triton X-41 b
425.290 425 C24H41O6 Triton X-45 b
513.343 513 C28H49O8 Triton X-114 b

Oil and detergent components in bilge water

Fig. 3 and Fig. S5 (ESI) depict the liquid SIMS spectral comparison of the fresh bilge, aged bilge, X-100 + fresh bilge, oil mix, detergent mix, and DI water in the positive ion mode. Detergent mix and oil mix share some common peaks in the low mass range (i.e., m/z+ 1–250); and most of them are identified as hydrocarbon peaks, such as m/z+ 41 C3H5+, 57 C4H9+, 71 C5H11+, 91 C7H7+, 105 C8H9+, and 121 C9H13+. Several high peaks are observed in the detergent mix in the relative high mass range (m/z+ 250–500), such as m/z+ 331, 347, 419, and 463. However, they are not observed in the fresh bilge, aged bilge, or X-100 + fresh bilge sample analysis using in situ liquid SIMS. This finding indicates that the contribution from detergent components is not as significant in the relative high mass range. This result is different from dry emulsion analysis. In dry samples, these detergent peaks were observed in bilge water emulsions, especially in the aged emulsion.15

This discrepancy prompts a postulation that the surface chemistry is different from what was conceived based on dry emulsion surface analysis previously. In addition, it gives the observational evidence that dry emulsion does not fully represent the liquid droplet environment. One disadvantage of dry sample analysis is the loss of water in the O/W interface. Consequently, dry samples cannot offer direct observations of the changing O/W interface in liquid. In contrast, liquid SIMS makes up the traditional SIMS deficiency and gives 2D and 3D images of the dynamic l–l interface. In the O/W liquid environment, some peaks from the detergent mix in the m/z+ 250–500 range disappear in the aged bilge surface in liquid; however, they are observed in dry samples. This interesting finding suggests that these detergent components may prefer to be solvated in liquid and remain in the bulk liquid inside the droplet. This may lead to the impression that the components on the bilge droplet surface are mainly small hydrocarbon chemicals from oil mixtures and surfactants. When the liquid sample becomes dry, these solvated components could appear in the bilge surface due to loss of water.

Observations are different in the high mass range, i.e., m/z+ 500–800. Unlike the oil components that have negligible contributions to the bilge surface composition, the detergent components are the main contributors. Significant detergent peaks are fragments of D-tocopheryl polyethylene glycol succinate (TPGS), an efficient emulsifier,15,42–44 including m/z+ 507 C22H44O11Na+, 551 C24H48O12Na+, 595 C26H52O13Na+, 639 C28H56O14Na+, and 683 C30H60O15Na+. These detergent peaks have dominant appearance in the aged bilge surface; however, they are not observed in the fresh emulsion at all. This finding suggests that these large molecular weight components do not appear immediately in the fresh bilge surface. In addition, these surfactant peaks migrate to the droplet surface over time. This result is similar to what was found in dry bilge samples. The observation of large surfactant peaks in the liquid bilge surface confirms that large molecular weight components take longer to move to the droplet surface, which is postulated based on dry droplet surface analysis. Additionally, the observation of these high molecular weight surfactants in both static and liquid ToF-SIMS may suggest that these surfactants are not situated as deep in the droplet bulk liquid phase, perhaps they exist in the thin film of the l–l interface as the bilge water emulsion forms initially. This may explain why they could be more easily observed in both the static and liquid ToF-SIMS unlike some of the mid-mass range detergent components. The latter may be located deeper in the bulk of the droplet. Molecular dynamic simulations would be helpful to elucidate and verify this postulation in the future.

Water cluster in bilge water

Water clusters are observed in the detergent mix and play a crucial role in forming bilge droplets. Water clusters marked by blue bars in Fig. 3 and Fig. S5 (ESI) are spread in the detergent mix and bilge samples in a wide mass range. In the lower mass range of m/z+ 1–250, the characteristic water clusters are (H2O)nH+, n = 7, 8, 10, 11, 12, or 13. In the relative high mass range (i.e., m/z+ 250–500), representative water clusters are featured as the following (H2O)nH+, n = 14–20. The low intensity of water clusters indicates that these water clusters are not the main composition in the bilge surface. It also confirms that some detergent components (e.g., m/z+ 250–500) dissolved in liquid hardly move to the bilge surface. In the high mass range (i.e., m/z+ 500–800), no obvious water clusters are observed at the bilge surface. This may also be attributed to the much lower ion yields of larger water clusters. The existence of water clusters makes the bilge surface different over time when comparing fresh and aged droplets in liquid, and this is probably due to hydrogen bonding and other types of weak intermolecular interactions. Additionally, water cluster distribution varies over time, contributing to the different bilge surface composition. The negative spectral comparison in Fig. S6 (ESI) shows similar results to those in the positive mode. More information is provided in the ESI.

The l–l interface change confirmed by spectral PCA

To better understand the bilge water surface compositional changes between emulsion samples and their components, selected peak spectral PCA is conducted in the positive and negative mode, respectively. In the negative mode, principal components (e.g., PC1, PC2 and PC3) explain 92.15% of all data. PC1 explains 42.91% of data and mainly separates the DI water control, oil mix, and fresh bilge from aged bilge and detergent mix. It is not surprising that the X-100 + fresh bilge sample crosses over PC1 positive and negative, because it consists of emulsions and an extra amount of X-100. The latter is a popular detergent component and surfactant. PC2 explains 32.25% of data and mainly separates the detergent mix and DI water from the aged bilge, fresh bilge, oil mix, and detergent mix. PC3 explains 16.99% of all data and mainly separates the fresh bilge and detergent mix from the aged emulsion and DI water (Fig. 4).
image file: d0cp00528b-f4.tif
Fig. 4 Selected peak spectral PCA results in the negative mode: (a) Scores plots of PC1 vs. PC2; (b) scores plots of PC1 vs. PC3; and loadings plots of PC1 (c), PC2 (d), and PC3 (e), respectively. Peaks are labeled in their unit masses.

In PC1 negative mode loadings, oil mix, fresh bilge and water cluster are the main contributors. Oil mix peaks, such as m/z 49 C4H, 63 C5H3, 79 C6H7 and water clusters, such as m/z (H2O)nOH, n = 2, 3, 5, 6, 9, and 10, make contributions to the formation of fresh bilge. PC1 positive separates aged bilge and detergent mix from fresh bilge and oil mix. This finding suggests that the bilge surface composition changes over time, and that detergent components contribute more to the aged emulsion. Some detergent peaks, especially in the relative high mass range, identified as m/z 325 C18H14O4P and 337 C20H33O4, only exist in the aged bilge but not fresh bilge. This indicates that these detergent components do not appear immediately in the fresh bilge surface yet move to the aged bilge surface as time elapses. This finding confirms the earlier hypothesis that the chemical surface composition change in O/W bilge water droplets over time using in situ molecular imaging.

The PC2 positive scores show that oil mix, fresh bilge, aged bilge and X-100 + fresh bilge share the same peaks. This finding suggests that some oil components, such as m/z 49 C4H, 63 C5H3, and 79 C6H7, and detergent components, such as m/z 105 C8H9, 205 C12H29O2 and 221 C12H13O4, appear immediately in the fresh bilge and persist in the aged bilge over time. The PC2 negative scores show that the detergent mix shares common peaks with DI water. This makes sense because the detergent mix contains water. In PC3 positive, oil and detergent components have similar peaks, and both contribute to fresh bilge and X100 + fresh bilge. Characteristic peaks include m/z 49 C4H, 205 C12H29O2, 425 C24H41O6 and 513 C28H49O8. PC3 negative separates aged bilge and DI water from fresh bilge and detergent mix, which further confirms that the bilge surface composition changes over time. In addition, water clusters, such as m/z (H2O)nOH, n = 3–6 and 9–10, make contributions to the aged bilge. The X-100 + fresh bilge sample shares common peaks with fresh and aged bilge as anticipated. Overall, the observation of water clusters in both fresh and aged bilge indicates that water clusters play a vital role in the formation of bilge emulsion. Our results demonstrate that the bilge surface composition changes over time. Detergent components, especially those appearing in the aged but not fresh bilge, are in the high mass range, further confirming the change of bilge surface as a result of surface evolution from surfactants. The change in water cluster distributions and diffusion may also play a central role in this surface change. Further research would require theoretical simulation to investigate this phenomenon and provide more fundamental insights.

The scores plots of PC1 vs. PC2, PC1 vs. PC3, and their loading plots in the positive mode are presented in Fig. S7 (ESI). These plots in the positive mode show agreement with those in the negative mode. Detailed information is provided in the ESI.

3D imaging of the evolved O/W interface

3D images are very useful in visualizing the spatial distribution of chemical species.30 Normalized 3D images of selected key peaks are used to study the evolution of chemical species at the O/W l–l interface in this work. Fig. 5 shows reconstructed and normalized 3D images in the positive mode; the color brightness indicates the relative intensity of a particular peak. The hydrocarbon fragments (e.g., m/z+ 43 C3H7+, 67 C5H7+, and 81 C6H9+) appear in fresh and fresh + X-100 bilge. The hydrocarbon peaks have good abundance in the aged bilge, supporting the finding that the hydrocarbon components appear in the fresh bilge immediately and persist in the aged bilge over time.
image file: d0cp00528b-f5.tif
Fig. 5 Comparison of normalized 3D images of key peaks in the positive (a) and negative mode (b). Darker colors indicate higher relative peak intensities and light colors indicate lower intensities.

The detergent components are evenly distributed in the detergent mix, for example, m/z+ 243 C10H20O5Na+, 551 C24H48O12Na+, and 595 C26H52O13Na+. However, they do not contribute significantly in the fresh bilge, especially in the large mass range (i.e., m/z+ 551, 595). In the aged bilge, these larger detergent components have higher relative abundance and thus they are more dominant contributors to the surface chemical makeup. The contrast of detergent components in the fresh and aged bilge indicates that the existence of these large surfactant species is not noticeable immediately on the surface of the fresh bilge droplets. Instead, they migrate to the bilge droplet surface over time. Our results further confirm that the O/W interface evolves chemically in surface composition and physically in size.

Water clusters including but not limited to m/z+ 199 (H2O)11H+, 235 (H2O)13H+, and 325 (H2O)18H+ are well distributed in the detergent mix, fresh, and aged bilge water emulsions. In the detergent mix, a higher abundance of (H2O)11H+ is seen. In comparison, the relative abundances of (H2O)13H+ and (H2O)18H+ are higher in the fresh and aged bilge water. This finding indicates that the water cluster distribution changes and affects the surface water hydrogen bonding environment.

Reconstructed 3D images of selected key peaks in the negative mode in Fig. 5b show similar findings. The hydrocarbon fragments, such as m/z 49 C4H, 63 C5H3, and 79 C6H7, appear in the fresh bilge and fresh + X-100 bilge emulsions. As the droplet surface ages, the hydrocarbon components become evenly distributed in space. This observation supports the hypothesis that the hydrocarbon components appear in the fresh bilge immediately and persist in the aged bilge over time. The detergent components are homogeneously distributed in the detergent mix, for example, m/z 105 C14H9O3, 221 C12H13O4, and 255 C16H31O2. They also are evenly distributed in the fresh bilge, aged bilge, and X-100 + fresh bilge emulsions. When looking into the relative intensity of the same detergent peaks between the fresh bilge and aged bilge, the relative percentages of these components change over time. For example, m/z 221 C12H13O4 has a higher relative intensity in the aged bilge. This finding also supports the hypothesis that the l–l interface has evolved over time. Water clusters are observed in all samples except the oil mix, and they are key components of both fresh bilge and aged bilge emulsions. This observation verifies that water is an essential liquid phase at the O/W interface in emulsion formation.41

New insight into the O/W interface

To better understand the evolution of the O/W interface, we further compare the results obtained in static ToF-SIMS and dynamic liquid ToF-SIMS. Static SIMS (Fig. 3 and Fig. S5, ESI) gives the analysis of dry samples (Fig. S8, ESI) and the results were published previously.15 Liquid SIMS probes the l–l interface and gives direct visualization of the liquid environment.27,33 Fig. 6 is a schematic illustration based on the liquid ToF-SIMS results. As a surface technique, the analysis depth of ToF-SIMS is known to be the top few nm of a material. When using ice as an approximation, liquid ToF-SIMS is probing the top layer of the oil–water interface approximately less than 10 nm (Fig. 6b and d).31,36,37 It is worth noting that the data spots in the 3D images do not represent a fixed location of the ion due to liquid diffusion,31 and they offer a representation of the chemical species spatial distribution of the thin layer of liquid being probed by the primary ion beam. Because of the diffusion of liquid, only the top layer of the liquids can be imaged using in situ liquid ToF-SIMS. At the vacuum–liquid interface, the temperature drop was estimated to be ∼12 K if starting the experiment at room temperature. This temperature change will not induce freezing and has a slight effect on the diffusion rate of ions. The concentration change was estimated to be approximately 1.2–2.2 times of the initial concentration.31 In addition, the diffusion rates of oil and the O/W mixture are different. In the original design of SALVI, we considered the self-diffusion constants of small ions and water to be in the range of 1–3 × 10−5 cm2 s−1.31,45 The diffusion constants of oils are much slower than water46 and the temperature dependence study of oils showed an insignificant effect on the diffusion constants.47 The dynamic viscosity of single droplets of diesel emulsions does not vary much with high water to diesel ratios.48
image file: d0cp00528b-f6.tif
Fig. 6 The schematic showing the evolution of the O/W interface in fresh bilge water (a and c) and aged bilge water (b and d), respectively. (a and c) Depict a single droplet; (b and d) the interfacial region of a bilge water emulsion droplet.

The SIMS spectral comparison indicates that the liquid ToF-SIMS can reveal the l–l interface compared to static SIMS. The lower mass oil fragments and detergent components (e.g., m/z+ 0–250) migrate to the l–l interface immediately and persist in aged bilge water as time goes, as evidenced by liquid ToF-SIMS spectral observations. The relative high mass detergent components (e.g., m/z+ 250–500) are more likely to exist in the bulk liquid in a droplet underneath the surface, because SIMS imaging could not detect them in either fresh or aged droplet surfaces in the liquid state (Fig. S5a, ESI). Interestingly, the high mass detergent components (e.g., m/z+ 507 C22H44O11Na+, 551 C24H48O12Na+, 595 C26H52O13Na+, 639 C28H56O14Na+, and 683 C30H60O15Na+) may exist closer to the top of the liquid, however, likely below the 10 nm of the l–l interface initially. These surfactant components move to the top of the l–l interface after one day and get detected by in situ liquid ToF-SIMS (Fig. 6c and d). Although the static analysis of dry samples gives similar results concerning the low and high mass surfactant peaks, it misses the solvated mid-range mass fragments due to drying and water loss by collapsing the liquid structure (Fig. S8b, ESI). Thus, in situ liquid ToF-SIMS provides a unique tool to observe the l–l interfacial change and gives us new insight into the evolution of the O/W interface, which was previously not possible.


We investigated the O/W l–l interface characterization of synthetic bilge water emulsions based on a Navy model using in situ SEM, optical microscopy, and in situ liquid SIMS. We found that the bilge water DSD is largely monodisperse; however, infrequent coagulation occurs even in freshly prepared bilge emulsions. The mean droplet size changes slightly as time elapses for bilge water without X-100 at static or gentle rocking conditions. However, the DSD becomes much larger for the bilge water with additional X-100 under either static or rocking conditions. This finding indicates time itself is insufficient to promote droplet growth and that the addition of surfactants like X-100 has a stronger influence on the DSD. Furthermore, the O/W l–l interfacial chemical composition evolves over time. We verified some of the hypotheses based on static dry emulsion results obtained in this work. First, our new finding suggests that water clusters play a crucial role in the formation and evolution of bilge water emulsion. Second, both oil and small detergent components appear in the O/W interface as soon as fresh droplets form as expected. Although high mass organics from the surfactant components do not appear immediately at the l–l interface, they migrate to the droplet surface as time passes. Our liquid SIMS observations reveal that some organic components in the mid-mass range (i.e., m/z 250–500) are more likely to remain solvated and stay in the bulk of the liquid phase inside the droplet, therefore, they do not appear on the droplet surface. Such an observation is illuminating to further the surface chemistry of droplet composition and understand the evolving O/W interface.

In situ liquid SIMS offers us a unique chance to visualize the water phase in the O/W emulsion compared to static SIMS. The latter loses the information of water due to emulsion drying. In this work, we have first demonstrated that the bilge water emulsion interface involving water, oil, and surfactants in the liquid phase can be studied using novel in situ multimodal imaging. More systematic studies of factors, such as salinity, pH, surfactants, that affect bilge water emulsion stabilization and weakening using this Navy model would be interesting for future work. Equally important is the theoretical modeling of the dynamic l–l interface that affects the chemistry of the droplet.

Conflicts of interest

The authors declare no competing financial interest.


This work was performed under the support of the SERDP project WP18-1660. The EMSL user proposals 50569 and 51176 were used for instrument access. We thank Danielle Paynter of NAVY CARDROCK for providing the NAVY bilge water model. We thank Michael Perkins of Pacific Northwest National Laboratory (PNNL) for graphic support. We thank William Chrisler for access to the optical microscopy at PNNL. PNNL is operated by Battelle under Contract No. DE-AC05-76RL01830. The processed data is available to readers by sending a request to the corresponding author, Dr Xiao-Ying Yu (


  1. J. Ge, J. Zhang, F. Wang, Z. Li, J. Yu and B. Ding, J. Mater. Chem. A, 2017, 5, 497–502 RSC.
  2. Z. Wang, X. Jiang, X. Cheng, C. H. Lau and L. Shao, ACS Appl. Mater. Interfaces, 2015, 7, 9534–9545 CrossRef CAS PubMed.
  3. Y. Zhu, J. Wang, F. Zhang, S. Gao, A. Wang, W. Fang and J. Jin, Adv. Funct. Mater., 2018, 28, 1804121 CrossRef.
  4. P. A. Alpert, R. Ciuraru, S. Rossignol, M. Passananti, L. Tinel, S. Perrier, Y. Dupart, S. S. Steimer, M. Ammann, D. J. Donaldson and C. George, Sci. Rep., 2017, 7, 12693 CrossRef PubMed.
  5. J. Church, J. G. Lundin, D. Diaz, D. Mercado, M. R. Willner, W. H. Lee and D. M. Paynter, Sci. Total Environ, 2019, 691, 981–995 CrossRef CAS PubMed.
  6. A. A. Umar, I. B. M. Saaid, A. A. Sulaimon and R. B. M. Pilus, J. Pet. Sci. Eng., 2018, 165, 673–690 CrossRef CAS.
  7. J. P. Chaudhary, N. Vadodariya, S. K. Nataraj and R. Meena, ACS Appl. Mater. Interfaces, 2015, 7, 24957–24962 CrossRef CAS PubMed.
  8. K. Ulucan and U. Kurt, J. Electroanal. Chem., 2015, 747, 104–111 CrossRef CAS.
  9. J. A. Caplan, C. Newton and D. Kelemen, Mar. Technol., 2000, 37, 111–115 Search PubMed.
  10. B. K. Körbahti and K. Artut, Desalination, 2010, 258, 219–228 CrossRef.
  11. C. Akarsu, Y. Ozay, N. Dizge, H. Elif Gulsen, H. Ates, B. Gozmen and M. Turabik, Water Sci. Technol., 2016, 74, 564–579 CrossRef CAS PubMed.
  12. X. Mei, H. Wang, D. Hou, F. L. Lobo, D. Xing and Z. J. Ren, Front. Environ. Sci. Eng., 2019, 13, 53 CrossRef.
  13. S. Putatunda, S. Bhattacharya, D. Sen and C. Bhattacharjee, Int. J. Environ. Sci. Technol., 2018, 16, 2525–2536 CrossRef.
  14. J. Church, D. M. Paynter and W. H. Lee, Langmuir, 2017, 33, 9731–9739 CrossRef CAS PubMed.
  15. J. Son, Y. Shen, J. Yao, D. Paynter and X.-Y. Yu, Chemosphere, 2019, 236, 124345 CrossRef CAS PubMed.
  16. M. Ding, T. Zhang, H. Zhang, N. Tao, X. Wang and J. Zhong, Food Hydrocolloids, 2019, 95, 326–335 CrossRef CAS.
  17. M. Ding, T. Zhang, H. Zhang, N. Tao, X. Wang and J. Zhong, Food Chem., 2020, 309, 125642 CrossRef CAS PubMed.
  18. T. Zhang, M. Ding, N. Tao, X. Wang and J. Zhong, LWT, 2020, 117, 108654 CrossRef CAS.
  19. T. Zhang, M. Ding, H. Zhang, N. Tao, X. Wang and J. Zhong, Food Chem., 2020, 308, 125597 CrossRef PubMed.
  20. H. Eskandarloo, M. J. Selig and A. Abbaspourrad, Chem. Eng. J., 2018, 335, 434–442 CrossRef CAS.
  21. G. Chen and D. Tao, Fuel Process. Technol., 2005, 86, 499–508 CrossRef CAS.
  22. K. Abdolmaleki, M. A. Mohammadifar, R. Mohammadi, G. Fadavi and N. M. Meybodi, Carbohydr. Polym., 2016, 140, 342–348 CrossRef CAS PubMed.
  23. C. W. Angle, T. Dabros and H. A. Hamza, Chem. Eng. Sci., 2006, 61, 7309–7324 CrossRef CAS.
  24. I. Vyrides, E. M. Drakou, S. Ioannou, F. Michael, G. Gatidou and A. S. Stasinakis, J. Environ. Manage., 2018, 217, 356–362 CrossRef CAS PubMed.
  25. Y. Ding, Y. Zhou, J. Yao, Y. Xiong, Z. Zhu and X.-Y. Yu, Analyst, 2019, 144, 2498–2503 RSC.
  26. Y. Fu, Y. Zhang, F. Zhang, J. Chen, Z. Zhu and X.-Y. Yu, Atmos. Environ., 2018, 191, 36–45 CrossRef CAS.
  27. X.-Y. Yu, J. Yao, D. B. Lao, D. J. Heldebrant, Z. Zhu, D. Malhotra, M.-T. Nguyen, V.-A. Glezakou and R. Rousseau, J. Phys. Chem. Lett., 2018, 9, 5765–5771 CrossRef CAS PubMed.
  28. X. Sui, Y. Zhou, F. Zhang, J. Chen, Z. Zhu and X.-Y. Yu, Phys. Chem. Chem. Phys., 2017, 19, 20357–20366 RSC.
  29. F. Zhang, X. Yu, J. Chen, Z. Zhu and X.-Y. Yu, npj Clim. Atmos. Sci., 2019, 2, 28 CrossRef.
  30. F. Zhang, X. Yu, X. Sui, J. Chen, Z. Zhu and X.-Y. Yu, Environ. Sci. Technol., 2019, 53, 10236–10245 CrossRef CAS PubMed.
  31. L. Yang, X.-Y. Yu, Z. H. Zhu, T. Thevuthasan and J. P. Cowin, J. Vac. Sci. Technol., A, 2011, 29, 061101 CrossRef.
  32. L. Yang, X.-Y. Yu, Z. Zhu, M. J. Iedema and J. P. Cowin, Lab Chip, 2011, 11, 2481–2484 RSC.
  33. J. Yao, D. B. Lao, X. Sui, Y. Zhou, S. K. Nune, X. Ma, T. P. Troy, M. Ahmed, Z. Zhu, D. J. Heldebrant and X.-Y. Yu, Phys. Chem. Chem. Phys., 2017, 19, 22627–22632 RSC.
  34. X.-Y. Yu, B. Arey, S. Chatterjee and J. Chun, Surf. Interface Anal., 2019, 51, 1325–1331 CrossRef CAS.
  35. X. Yang, J. Lee, D. E. Barker, X. Wang and Y. Zhang, J. Air Waste Manage. Assoc., 2012, 62, 725–735 CrossRef PubMed.
  36. A. M. Piwowar, S. Keskin, M. O. Delgado, K. Shen, J. J. Hue, I. Lanekoff, A. G. Ewing and N. Winograd, Surf. Interface Anal., 2013, 45, 302–304 CrossRef CAS PubMed.
  37. C. Szakal, J. Kozole, M. F. Russo, Jr., B. J. Garrison and N. Winograd, Phys. Rev. Lett., 2006, 96, 216104 CrossRef PubMed.
  38. Y. Ding, Y. Zhou, J. Yao, C. Szymanski, J. Fredrickson, L. Shi, B. Cao, Z. Zhu and X.-Y. Yu, Anal. Chem., 2016, 88, 11244–11252 CrossRef CAS PubMed.
  39. W. Li, H. Li, J. Li, X. Cheng, Z. Zhang, F. Chai, H. Zhang, T. Yang, P. Duan, D. Lu and Y. Chen, J. Environ. Sci., 2018, 69, 61–76 CrossRef PubMed.
  40. J. Wei, A. Bristow, E. McBride, D. Kilgour and P. B. O’Connor, Anal. Chem., 2014, 86, 1567–1574 CrossRef CAS PubMed.
  41. NIH, PubChem,, accessed April 15, 2019.
  42. V. Khare, W. A. Sakarchi, P. N. Gupta, A. D. M. Curtis and C. Hoskins, RSC Adv., 2016, 6, 60126–60137 RSC.
  43. L. Y. Kong, B. G. Su, Z. B. Bao, H. B. Xing, Y. W. Yang and Q. L. Ren, J. Chromatogr. A, 2011, 1218, 8664–8671 CrossRef CAS PubMed.
  44. J. Lu, Y. Huang, W. Zhao, Y. Chen, J. Li, X. Gao, R. Venkataramanan and S. Li, Mol. Pharmaceutics, 2013, 10, 2880–2890 CrossRef CAS PubMed.
  45. K. Tanaka, J. Chem. Soc., Faraday Trans., 1978, 74, 1879–1881 RSC.
  46. E. E. Perez, A. A. Carelli and G. H. Crapiste, J. Food Eng., 2011, 105, 180–185 CrossRef CAS.
  47. S. E. M. Hamam, J. Environ. Sci. Health, Part A: Environ. Sci. Eng., 1987, 22, 445–456 CrossRef.
  48. J. Won, S. W. Baek, H. Kim and H. Lee, Energies, 2019, 12, 1963 CrossRef.


Electronic supplementary information (ESI) available: Additional experimental details and results. See DOI: 10.1039/d0cp00528b
Equal contribution.

This journal is © the Owner Societies 2020