Extraction of Terpenoids from Pine Needle Biomass Using Dimethyl Ether

Abstract

Pine needles are an industrial feedstock for extracts used in a variety of applications, but conventional extraction methods often result in a degradation of the terpenoid compounds that naturally occur in loblolly pine (Pinus taeda). Separation of these compounds from pine biomass is an energy-intensive operation, typically requiring a significant input of thermal energy. An alternative separation approach with potential energy savings is extraction with a condensable gas, namely, dimethyl ether. Biomass materials are exposed to liquid dimethyl ether under pressure, which mobilizes the organics. The extract is then separated from the insoluble pine matter, and dimethyl ether is volatilized away from the separated organic species. A variety of terpene derivatives were extracted from pine needle biomass using this approach, including monoterpenes, sesquiterpenes, and related oxygenates, which were identified using two-dimensional gas chromatography/mass spectrometry. Additionally, the dimethyl ether-treated needles resemble needles subjected to low-temperature drying, whereas needles treated with a high-temperature drying method appear to have shrunken structures. The results suggest that dimethyl ether extraction has significant potential for separating valuable organics from complex matrices without the application of thermal energy during treatment.

1. Introduction

Pines (Pinus) are a dominant genus found in North American, European, and Asian forests [1]. Historically, pine has been harvested by humankind for use as a building material [1], chemical production (e.g., potash) [2], fuel [1], food [3], and traditional medicine [4]. While pine still serves dominantly in construction materials and as a raw material for pulp and paper industries [5], residues from these industries are considered a low-cost feedstock for biofuel production. Thermochemical conversion pathways for pine are often considered, where fast pyrolysis or catalytic fast pyrolysis can provide high bio-oil yields, which can then be hydrotreated to create fuels in the jet, diesel, and gasoline ranges [6]. However, studies have shown that pine white wood anatomical fractions result in higher pyrolysis performance than their bark and needle fractions, primarily due to higher extractives and inorganic elements [7]. In particular, needles can contain high levels of extractives that are now used in a variety of applications [8]. Pine extracts originate from conifer oleoresin, which serves a biological role in tree defense, forming a chemical barrier against pest insects and fungal spores [9,10,11].

The extracts of pine needles are of particular ongoing research interest on account of their broad utility as specialty chemicals [12], pharmaceuticals [13,14,15], perfumes [12], structural [16,17] antimicrobial [18] biocomposites, bio-oils for combustible fuel [19,20], and food additives [21]. Pine needles are also frequently studied as an indicator of contamination in the atmosphere [22] and the local environment [23]. Relevant in many of these applications, biocompounds found within pine needles are responsible for the resistance of individual pine species to invasive pine beetle species [10,11,24,25]. As such, the separation and identification of pine resin terpenes associated with anti-beetle defensive capabilities has been a research priority.

Extraction of pine-derived biocompounds with organic solvents has been widely applied, primarily for the identification of individual terpenes. Pentane and hexane have been utilized by multiple groups [26,27,28,29], and solvent systems using mixtures of alkane solvents and acetone [30] and diethyl ether [8] have also been applied. These approaches resulted in the separation of terpenes, notably α-pinene, δ-carene, and β-pinene, which, in general, are the most abundant monoterpenes [8]. Solvent extraction is not specific and also results in the separation of fatty acids and esters with comparatively smaller quantities of mono-, sesqui-, and di-terpenoids. Serial extraction of pine needles using hexane, methanol, and, finally, water enables the collection of a wider range of compounds [29].

Hydrodistillation has been employed by several groups for separating organic compounds from pine needles and other fractions. Pine biomass is mixed with water, and the resulting mixture is distilled, resulting in the partitioning of extractable organic compounds within the distillate. Hydrodistillation has been shown to be effective in separating many terpenoid derivatives [24,31,32,33,34,35,36,37,38,39,40]. A simultaneous extraction and distillation process has also been effectively employed to separate terpenes from pine needles using dichloromethane as an extraction solvent [41,42].

Monoterpenes are emitted from pine as volatilized organic compounds and can be collected using adsorption cartridges, enabling the analysis of monoterpenes [43]. Solid-phase microextraction (SPME) effectively samples the headspace over pine samples for the identification of separate mono-terpenoid, sesqui-terpenoid, and aldehyde fractions [25,44,45]. SPME has also been used to separate terpenoids generated from the torrefaction of pine biomass [46], thus separating the mono-terpenes, sesqui-terpenes, acetic acid, and additional products more commonly associated with pyrolysis treatments [19]. Solvent-free microwave extraction is yet another separation method that has shown efficacy in terpenoid recovery from pine biomass [47].

Supercritical carbon dioxide extraction has been employed for the separation of terpenoids [30] and was shown to be effective as a pretreatment for pine biomass pyrolysis [37]. The experiments showed that more terpenes could be extracted from needles compared with bark, cones, and branches [37]. Supercritical fluid extractions require elevated temperatures and pressures (CO₂ becomes supercritical at 72 atm at 30 °C), adding to the energy and equipment costs of a process.

Condensable gases are another class of solvents that have been utilized to recover pine extracts. Bier and coworkers utilized condensable gas extraction with a liquefied butane/propane mixture at ~5 atm and 35 °C to extract terpenes from pine needles, concluding that the approach was more efficient compared with conventional solvent extraction [48]. In this work, pine needles are processed by condensable gas extraction with dimethyl ether (DME) [49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71]. In many ways, DME extraction is like a butane/propane extraction but isolates a different, generally larger fraction of biocompounds than iso-butane, n-butane, or propane extractions. The larger extracted fraction is due to the greater polarity of DME relative to butane/propane, allowing for more polar molecules to be mobilized. DME may also mobilize less polar molecules more effectively by displacing both water and mobile organics; removing water from a porous solid may improve the mobility of less polar organics. This work presents an emerging methodology for the extraction of a more diverse variety of compounds from pine materials than has been previously studied with other solvents. The final extract from DME extraction generally undergoes phase separation into distinct aqueous and organic phases once the DME is removed. These characteristics allow DME to be applied to other aqueous separations [72,73,74,75,76,77,78,79].

In this work, pine needles (leaves) from the loblolly pine (Pinus taeda) were selected as a feedstock for the extraction of terpenoid species with DME. Loblolly pine was selected, as it is a dominant North American pine species; needles were extracted because they feature higher concentrations of terpenoids than woody material [80]. While characterization of terpenes has been performed on the needles of various pine species [80,81,82,83], the exact mixture of terpenoid compounds varies with species, tree habitat, needle age, and season [83].

The extracts were subsequently analyzed using two-dimensional gas chromatography/mass spectrometry (2D-GC/MS), confocal microscopy, scanning electron microscopy (SEM), and energy dispersive X-ray spectroscopy (EDS). The use of two-dimensional GC is advantageous [31]. Given the complexity of the mixtures generated by separation processes, this approach can enable chromatographic separation of closely eluting compounds, which, in turn, can generate unambiguous mass spectra for library matching. Even when using 2D-GC/MS, compound identification using retention index matching is still needed, as noted in multiple earlier studies [31,38,39,45,84].

2. Materials and Methods

2.1. Sample Preparation

A sample (10.2 g) of fresh pine needles, Pinus taeda, was enclosed in a pressure-rated glass extraction cylinder sealed at both ends with Teflon endcaps. DME (Matheson Gas, Irving, TX, USA, 99.5%) was introduced to and evacuated from the reaction chamber via Swagelok fittings and ball valves. The pine needle material had not been size-reduced or dried prior to extraction. Pine needles were received fresh and stored within a vacuum-sealed plastic bag and refrigerated prior to extraction. Furthermore, 3 × ~1 L DME volumes were used to immerse the sample at ambient temperatures (~20 °C) and elevated pressure (DME vapor pressure is ~5 atmospheres at ~20 °C). The solvent was recirculated via a pump for 1 h with each volume of DME. All three volumes were collected in the same collection vessel. When the DME was removed via depressurization, two products were produced as follows: (1) ~2 g of a cloudy aqueous solution and (2) a semisolid resin adhered to the glass vessel. While DME was not captured and recompressed (recycled) in this work, recent publications present methodologies and analyses of DME recycling for industrial separation [78,79,85,86]. A diagram of the extraction apparatus and a more detailed hardware description is presented in Supplementary Materials Figure S1.

The aqueous sample was diluted using methanol (Sigma-Aldrich, St. Louis, MO, USA, anhydrous, 99.8%) prior to analysis using a 2D-GC/MS instrument. The resin was partially dissolved into tetrahydrofuran (THF, Sigma-Aldrich, anhydrous, 99.9%), producing amorphous green semisolids and a solution that was analyzed without further dilution. The THF sample was spiked with a small volume of a solution (AccuStandard, New Haven, CT, USA) containing biphenyl, phenanthrene, and fluoranthene, such that their concentrations were 1 millimole/liter. These compounds served as internal standards and may be used to measure unknown concentrations of terpenes from the calibration curves. The amorphous green semisolids that did not dissolve in THF were analyzed using a micropyrolysis unit interfaced with the 2D-GC/MS.

2.2. Micropyrolysis

The residual green semisolid was analyzed using a micropyrolyzer (CDS Analytical 5250 pyrolyzer, Oxford, PA, USA), which is equipped with a 36-sample carousel that enables analyses of batches of sample tubes. The samples were dropped into the pyrolysis chamber, dried for 4 s at 100 °C, and then held for an additional second at 100 °C. The temperature was then ramped at 50 °C/s to the maximum pyrolysis temperature (T_max) and then held at T_max for 5.00 s. The sample tube was then ejected, and the chamber was cleaned by heating to 1200 °C and holding at this temperature for 10 s. The pyrolysis vapor that was produced was transferred to the 2D-GC/MS via a transfer line maintained at 280 °C.

Samples of the green semisolid were prepared for pyrolysis by loading a pyrolysis tube with a short spacer and a glass wool plug. Approximately 300 µg was then weighed into the pyrolysis tube, to which another glass wool plug was added. Prior to analysis, one µL of solution containing the three internal standard compounds at a concentration of 1 millimole/L was injected into the biomass. The quantity of each internal standard added to the pyrolysis tubes was 1 nanomole.

Data processing involved the subtraction of instrument blanks from the analyses of the pasty solids; this corrected for the presence of paraffinic aldehydes in the instrument from prior analyses.

2.3. 2D-GC Separation

Two-dimensional gas chromatography (2D-GC) was conducted using an Agilent (Santa Clara, CA, USA) 7890 gas chromatograph modified for 2D-GC by a four-jet modulator and a secondary oven, both located within the primary oven. The first chromatographic dimension used a 28 m 0.25 mm inner diameter column with a 0.5 µm Rxi-5ms (Restec, Bellafonte, PA, USA) stationary phase, which is 5% diphenyl/95% dimethyl polysiloxane. The second chromatographic dimension used a 1 m, 0.1 mm inner diameter column with a 0.1 µm Rxi-17 (Restec, Bellefonte, PA, USA) stationary phase, which is 50% diphenyl/50% dimethyl polysiloxane. An additional 21 cm of the Rxi-17 column served as the transfer capillary between the secondary oven and the mass spectrometer. The Rxi-5ms column primarily separates compounds based on the boiling point, while the Rxi-17 column separates on the basis of compound polarity.

The helium carrier gas flow was controlled at 1 mL/min throughout the analysis. The pyrolysis compounds were split within the injector using a split ratio of 20:1. The injector was maintained at 300 °C. Upon initiation of the analysis, the primary column was held at 50 °C for 0.5 min, then ramped at 7.5 °C/min to a target temperature of 260 °C, and finally held at this temperature for an additional 3 min. The transfer capillary was maintained at 280 °C.

The secondary oven and 4-jet modulator were maintained at 5 °C and 15 °C (respectively) above the temperature of the primary oven. A 3.00 s modulation period was used, and hot and cool pulse times were varied to efficiently trap and desorb light compounds early in the run, and heavier compounds later in the run. Before a retention time of 394 s, the hot pulse and cool times were 0.50 and 1.00 s, respectively. This was repeated twice per modulation cycle by the 4-jet modulator. After 394 s, the hot pulse and cool times were 1.00 and 0.50 s, respectively.

Liquid samples were injected using a Gerstel rail sampler. A volume of one µL was injected for analyses of all liquid samples and standards.

2.4. MS Detection

Mass spectrometry detection and analysis were conducted using a Leco Pegasus 4D instrument (St. Joseph, MI, USA), which integrates a time-of-flight mass spectrometer with the Agilent 7890 GC. An acquisition delay of 220 s was employed to allow very light compounds to pass through the MS before analysis was initiated. The instrument was scanned from m/z 43 to 300 at a rate of 200 spectra/s. The fast scan rate enables the deconvolution of closely eluting compounds. Electron ionization via the ion source was conducted at 250 °C with an energy of 70 electron volts (eV).

2.5. Compound Identification

Compounds were identified based on their mass spectra, utilizing library searches. The monoterpenes were identified by a small but notable m/z 136 (M^+•). Significant fragment ions were at m/z 121 (loss of •CH₃) and m/z 107 (loss of •C₂H₅). Frequently, the base peak was m/z 93, formed either by the loss of •C₃H₇ as an integral propyl radical or by serial losses of •CH₃ and then •C₂H₄; this ion tends to be generally diagnostic for monoterpenes. The C₇H₈⁺• ion at m/z 92 was occasionally prominent.

The monoterpenes are all C₁₀H₁₆ isomers and tend to have mass spectra that are similar to one another, which means that their mass spectra are not explicitly diagnostic for compound identification. Frequently, library searches generate matches that have high similarity scores but low probabilities, reflecting a situation where the agreement between an unknown mass spectrum and that of the library match is high, but the library spectrum is not particularly unique, i.e., multiple other compounds have a similar spectrum. On account of this problem, compound identification should ideally involve retention time matching with known identified standards. An alternative approach is to compare the relative retention time order observed for the terpenoid chromatographic peaks with other studies reported in the literature. The latter approach was used in this study, consistent with the objective, which was to qualitatively evaluate the effectiveness of the DME extraction for achieving low-temperature terpenoid extraction.

2.6. Sample Preparation and SEM-EDS Imaging

Needles were either subjected to DME treatment, as described above, or thermal treatment at 105 °C and 40 °C. Treated pine needles were embedded and sectioned using an adapted protocol from Gierlinger et al. [87]. They were rehydrated by soaking in water and slowly infiltrated with polyethylene glycol (PEG) 2000 (Sigma-Aldrich). Rehydrated needles were submerged in 50% PEG 2000 solution in a closed container at 60 °C until they were permeated with the solution. Once the needles sank to the bottom of the container, the lid was removed to allow the water to evaporate from the solution. When the volume of the solution was reduced by half, the needles were submerged in 100% PEG overnight. The needles were removed from the PEG, placed on a microscope slide, and allowed to harden to the slide at room temperature overnight. The needles were sectioned with a scalpel and soaked in water to remove the PEG from the sections. The sections were dried onto a microscope slide at room temperature overnight. The sections were removed from the microscope slides with a scalpel, attached to carbon SEM mounts with double-sided copper tape, and sputter-coated with gold. Sections were imaged using a JEOL JSM-6610LV Scanning electron microscope (SEM, Tokyo, Japan) equipped with an energy-dispersive X-ray spectrometer (EDS).

3. Results

3.1. Structural Changes to Pine Needle Material

Ideally, terpenoid extraction would employ benign conditions, such as a low temperature and pressure, and minimal mechanical grinding. More aggressive extraction conditions would be expected to destroy the pore structure, making the mobilization of terpenoids more difficult. Higher temperatures would also be expected to cause chemical changes in the terpenoids, including isomerization, oxidation, and hydroxylation, modifying the nature of the essential oils that are extracted. For this reason, the microscopic anatomy of the pine needles treated using DME extraction was compared with the needles that were subjected to thermal treatment at 105 °C and 40 °C. Confocal imaging revealed gross anatomical changes in the needles dried at 105 °C, as evidenced by a shrunken appearance and possible pore collapse, as substantiated by SEM imaging in Figure 1. In comparison, the pore structure of the pine needles heated to only 40 °C and the samples subjected to DME extraction displayed intact pore structures. Additionally, the SEM image of the DME-dried sample displays a cuticle (outer layer) that is attached but separated from the needle. In contrast, the cortex layer appears to have delaminated from the 40 °C-dried needle. Further insight into the effect of the DME process is provided by the spatial distribution of silicon and calcium, as shown in the images generated using EDS. The samples treated at 40 °C show that silicon is enriched in the epidermis and mesophyll and that calcium is enriched in the phloem. Figure 2 depicts SEM-EDS mapping of DME-dried, 40 °C-dried, and 105 °C-dried pine needles. Silicon was enriched in the epidermis and mesophyll. Calcium was enriched in the phloem; red = Si and blue = Ca.

In comparison, the image of the DME-treated sample shows that both Si and Ca are depleted, indicating that the process removes the inorganic elements from the vascular structure of the pine needles. Si and Ca are elements that are converted to ash in combustion, and their removal results in a cleaner biomass product [88]. Ash byproducts of biomass are associated with fine particulate emissions, fouling, corrosion, and abrasion [88]. This may suggest that DME extraction could function as a pretreatment ahead of thermal processing by removing inorganics that may catalyze secondary degradation during pyrolysis.

3.2. Analysis of the THF-Soluble Terpenoids

The product of the DME extraction was a dark green semisolid that dissolved into and colored the solution. The analyses of the THF showed an abundant peak corresponding to α-pinene, eluting at ~590 s, together with numerous low-abundance peaks that were also terpenoid in nature. The identification of α-pinene was based on retention time matching with a standard (see Supplementary Materials Appendix S2, Figure S2) and also agreement with the mass spectrum. In the analysis of the standard, concentrations of α-pinene and the internal standard were known, and this enabled the concentration of α-pinene in the sample to be calculated at 1.36 mM/L. The mass spectrum of α-pinene displays a low-abundance molecular ion at m/z 136, monoterpene elemental composition C₁₀H₁₆; Figure 3. The majority of the ion signal is accounted for by m/z 93, 92, 91, 79, and 77. The formation of the base peak at m/z 93 can be reasonably rationalized by a serial elimination of •CH₃ and then C₂H₄; Reaction Scheme 1. The appearance of multiple, high-abundance fragment ions is indicative of the complexity of the fragmentation pathways operating in the α-pinene radical cation; however, this behavior is not unique to α-pinene, as the other monoterpenes display similar behavior and unsurprisingly produce very similar mass spectra.

The abundant m/z 93 and 91 that are seen in the mass spectra of the monoterpenes suggest that these ions could be used to identify other monoterpenes that are not abundant in the extract. An extracted ion chromatogram was generated by summing the signal from the m/z 93 and 91 mass channels, which tended to provide a reliable indication of the presence of both mono- and sesquiterpenes. These could be substantiated by the extracted ion chromatograms of m/z 136 and 204, which correspond to the molecular ions of the monoterpenes and sesquiterpenes, respectively. Figure 4 depicts the extracted ion chromatograms from the analysis of THF-soluble terpenoids.

Standards were not available for matching the mass spectra or retention times. However, the comparison of measured retention indices with values that were published in the literature provided an improved level of accuracy with regard to the compound identifications in Figure 5 and Figure 6. The retention indices measured in the present study were systematically higher compared with the literature values; however, the order of retention indices was consistent when compared with previous publications by Adams [84], Isodorov [45], and Schwob [89], as presented in

Table 1. Retention indices measured for terpenoid compounds identified in the analyses of the THF solution generated from the DME process. Values reported by Isodorov [45], Adams [84], and Schwob [89] are included for comparison.

	Retention Index (s)
Compound	This Study	Isodorov [45]	Adams [84]	Schwob [89]	First Dimension Retention Time	Mass Spectral Similarity Index	Molecular Weight (u)
Tricyclene	958	844	926		577	844	136
α-Pinene	967	936	939	938	589	936	136
Camphene	986	886	953		613	886	136
β-Pinene	1014	911	980	973	649	911	136
δl-Limonene	1062	854	1030	1024	709	854	136
Verbenol	1185	838		1146	859	838	152
trans-β-Caryophyllene	1480	950	1418	1431	1180	950	204
α-Selinene	1515	862	1494		1216	862	204
δ-Cadinene	1573	870	1524	1524	1273	870	204
(+) Spathulenol	1644	836	1576	1574	1339	836	220
(−)-Caryophyllene oxide	1654	774	1581	1576	1348	774	220
epi-α-Cadinol	1700	815	1653		1390	815	222

. The higher retention indices mean that the retention times are longer for the terpenoids with respect to the neighboring compounds in the paraffin series. This is probably the result of the secondary column (a 50% phenyl/50% methyl stationary phase), which separates compounds more on the basis of polarity compared with the methyl siloxane stationary phases used in the literature.

The presence of multiple sesquiterpenes is illustrated by a 3D chromatographic plot of the region eluting from about 1160 to 1460 sec, using the sum of the extracted ion profiles of m/z 93, 91, 204 (expanded by 20×) and 136 (expanded by 20×); Figure 7. β-caryophyllene and δ-cadinene are prominent, eluting on either side of the butylated hydroxytoluene (BHT) preservative, which is an impurity introduced into the process by the THF receiving solvent. It is worthwhile noting that there is another prominent sesquiterpene that is exactly co-eluted with BHT, as evidenced by an m/z 204 extracted ion profile that does not exactly trace that of the BHT-derived ions; however, a mass spectrum could not be deconvoluted. There are minor peaks with a molecular weight of 204 that are seen at later retention times. They furnish mass spectra that are in good agreement with the germacrene derivatives; however, there is a disagreement between their observed elution order and that which is reported in the literature.

3.3. Residual Paste Water Leach and Thermal Desorption/Pyrolysis

The DME extract produced a resin that was partially dissolved in THF; for the preceding analysis, the residual green paste was leached with water to identify compounds that were more extensively functionalized. Residual THF was the only compound significantly above the background when the TIC was plotted without expansion factors; this is shown in Supplementary Materials Figure S3. Several terpene/terpenoid compounds were observed when the extracted ion chromatograms generated from the sum of m/z 91 and 93 were plotted; see Supplementary Materials Figure S4. Specifically, α- and β-pinene and two pinene oxygenates, 2-pinen-4-ol and 2-pinen-4-one, were observed, albeit at low abundance. In this sample, the concentration of α-pinene was measured at 0.07 mM/, indicating that the majority of the α-pinene was recovered by the THF dissolution. There was no evidence of any sesquiterpene derivatives in this analysis of the subsequent water leach.

The post-THF dissolution residuals were also analyzed using micro-pyrolysis/2D-GC/MS. Four different maximum temperatures were employed in the micropyrolysis unit: 300 °C, 400 °C, 500 °C, and 600 °C. These experiments function to both thermally desorb and pyrolyze the residue samples. Generally, 300 °C is inadequate to initiate the pyrolysis of biomass, but it is more than sufficient to volatilize many of the organics present. Pyrolysis of biomass at 400 °C results in the thermal degradation of hemicellulose, which is observed in the form of small organic oxygenates eluting from the 2D-GC at shorter retention times. At 500 °C and 600 °C, significant pyrolysis of the cellulose and lignin fractions occurs; compounds from these processes are observed together with the terpenoids.

Compound identification involved the consideration of matching experimental mass spectra with those found in mass spectral libraries and matching retention indices. Mass spectral library matching is ambiguous because of the large number of terpenoid compounds in the libraries and the similarity of their mass spectra. Compound identification based on retention index matching is similarly difficult due to disagreements in the literature and also because the retention indices in the present study were systematically greater compared to those in the literature. Nevertheless, elution order is largely constant; hence, using elution order together with mass spectral matching afforded compound identification with good accuracy.

Paraffinic aldehydes containing between six and twelve carbon atoms were observed in the thermal desorption/pyrolysis experiments (see Figure 8); these compounds are likely formed from the degradation of the corresponding acids during the thermal desorption/pyrolysis event. No monoterpenes were seen in these analyses, indicating that the THF extraction had effectively removed them from the insoluble green residue generated in the DME extraction. The compounds produced from the experiments conducted at 400, 500, and 600 °C were similar in many cases, and their abundances increased from 300 < 400 < 500 °C but then decreased upon going to 600 °C, suggesting that the highest temperatures caused degradation of some of the sesquiterpenes.

The thermal desorption/pyrolysis experiments contained ten abundant sesquiterpene or sesquiterpenoid compounds in the time window over which the sesquiterpenes elute; see Figure 8. The most abundant peak was a compound identified as (+) spathulenol, which is a hydroxylated sesquiterpene. A small but notable molecular ion was observed at m/z 220 for spathulenol, and the mass spectrum was in excellent agreement with the library entry (Supplementary Materials Figure S18). α-Cadinene is the second most prominent sesquiterpene; this compound was not noted in the analyses of the extracts (Supplementary Materials Figure S14). It is noted that a large number of less abundant sesquiterpenoid compounds were extracted, and the 2D chromatograms in Figure 9 and Figure 10 provide a flavor of the complexity. These compounds were not detected in the analyses of conventional solvent extraction processes, indicating that sesquiterpenes do not have significant solubility in THF, hexane [26], or methylene chloride [90]. However, it is clear that they are mobilized by condensed dimethyl ether. Identifications of compounds within the insoluble residues are presented in

Table 2. Summary of retention indices for compounds identified in the thermal desorption/pyrolysis 2D-GC/MS analyses of the insoluble residues. Values reported by Adams [84] and Isodorov [45] are included for comparison.

	Retention Index (s)
Compound	This Study	Isodorov [45]	Adams [84]	First Dimension Retention Time	Mass Spectral Similarity Index	Molecular Weight (u)
α-Copaene	1374.3	1376	1377	1072	831	204
Cubabene	1405.9	1390	1390	1105	870	204
1,6,10-Dodecatriene, 7,11-dimethyl-3-methylene-, (Z)-	1420.6		1430	1120	710	204
Germacrene D	1485.4		1484	1186	796	204
Germacrene A	1506.1		1508	1207	814	204
α-Muurolene	1530.4	1499	1498	1231	775	204
α-Cadinene	1548.6		1537	1249	868	204
1S,cis-Calamenene	1554.7		1528	1255	898	202
Germacrene B	1566.8		1559	1267	769	204
Cadina-1,4-diene	1563.7		1533	1264	832	204
(−)-Caryophyllene oxide	1591.0	1579	1582	1291	803	220
(+) spathulenol	1620.1	1576	1577	1318	869	220
10-epi-α-Cadinol	1680.1			1372	687	222
trans-Z-α-Bisabolene epoxide	1696.8			1387	809	220

. A plot of the three-dimensional chromatogram generated from the total mass spectral ion current at a 500 °C pyrolysis of THF-insoluble solid is shown in Figure 9. The m/z 204 channel depicting thermal desorption/pyrolysis conducted at 500 °C on a the THF-insoluble solid is shown in Figure 10.

4. Conclusions

Pine needle extracts are widely utilized in diverse applications that include composite materials, chemicals, pharmaceuticals, and food. The dimethyl ether (DME) extraction process presented in this work was effective in separating terpenoid compounds from pine needle biomass. The process has the advantage of being conducted at benign temperatures, which is important, given the propensity of the terpene and sesquiterpene derivatives to undergo isomerization. The DME extraction was able to mobilize a large number of sesquiterpenoid compounds, which were not directly extractable with THF, hexane, or methylene chloride. The variability of terpenoid compounds presents a significant characterization challenge, which was addressed using two-dimensional gas chromatography with mass spectrometry detection. This approach yielded unequivocal separation of the majority of the terpenoid peaks; however, the similarity of the mass spectra of terpenoid isomers rendered identification based on mass spectral library searching ineffective. Improved compound identification was achieved by combining retention index matching with library searching; however, this was complicated by longer retention indices resulting from the polar secondary column used in the two-dimensional GC analysis, which disproportionately lengthens the retention times of the terpenoids compared with the different types of paraffin. Nevertheless, the overall analytical strategy shows good promise for characterizing complex mixtures generated in extractions of pine biomass.