2-Deoxy-2-fluorotalose

• CAS: 98808-82-3
• Synonyms: 2-deoxy-2-(18F)fluoro-D-talose; 2-deoxy-2-fluorotalose
• Molecular Formula: C6H11FO5
• Molecular Weight: 182.15 g/mol
• SMILES: C(C(C(C(C(C=O)F)O)O)O)O
• InChI: InChI=1S/C6H11FO5/c7-3(1-8)5(11)6(12)4(10)2-9/h1,3-6,9-12H,2H2/t3-,4-,5-,6+/m1/s1
• InChIKey: AOYNUTHNTBLRMT-KAZBKCHUSA-N

2-Deoxy-2-fluorotalose (FDT, CAS 98808-82-3, C6H11FO5, MW 182.15 g/mol) is a C2-fluorinated analog of the rare sugar D-talose, defined stereochemically as (2S,3S,4S,5R)-2-fluoro-3,4,5,6-tetrahydroxyhexanal. Originally developed as the 18F-labeled PET radiotracer [18F]FDT for non-invasive quantification of galactokinase activity in tissues , the compound is now also recognized as a superior substrate for enzymatic synthesis of fluorinated UDP-sugars via the GalK/GalPUT pathway.

2-Deoxy-2-fluorotalose (CAS 98808-82-3): Fluorinated D-Talose Analog for Galactokinase-Targeted PET Imaging and Enzymatic Glycobiology

2-Deoxy-2-fluorotalose (FDT, CAS 98808-82-3, C6H11FO5, MW 182.15 g/mol) is a C2-fluorinated analog of the rare sugar D-talose, defined stereochemically as (2S,3S,4S,5R)-2-fluoro-3,4,5,6-tetrahydroxyhexanal [1]. Originally developed as the 18F-labeled PET radiotracer [18F]FDT for non-invasive quantification of galactokinase activity in tissues [2], the compound is now also recognized as a superior substrate for enzymatic synthesis of fluorinated UDP-sugars via the GalK/GalPUT pathway [3]. As a deoxyfluorohexose, its biological specificity derives from C2 stereochemistry—the D-talo configuration—which determines exclusive recognition by galactokinase (EC 2.7.1.6) rather than hexokinase [2][3].

1

Galactokinase-targeted PET imaging probe precursor – Designed for non-invasive quantification of tissue galactokinase activity via simplified kinetic modeling.

2

Enzymatic UDP-sugar synthesis – Reported substrate for one-pot GalK/GalPUT reactions to produce fluorinated UDP-sugar donor substrates.

3

Stereochemical probe – D-Talo configuration (C2 fluoro epimer) ensures galactokinase recognition and a single-metabolite metabolic fate.

• [1] MeSH Supplementary Concept Data: 2-deoxy-2-fluorotalose. U.S. National Library of Medicine. Unique ID: C076400. Registry Number: 98808-82-3. View Source
• [2] Haradahira T, Kato A, Maeda M, Torii Y, Ichiya Y, Masuda K. Synthesis and biodistribution of a fluorine-18 labeled analogue of D-talose: 2-deoxy-2-[18F]fluoro-D-talose. Int J Rad Appl Instrum A. 1992 May;43(5):627-32. doi:10.1016/0883-2889(92)90031-9. PMID: 1325421. View Source
• [3] Zhao G, Guan W, Cai L, Wang PG. One-pot enzymatic synthesis of UDP-sugars catalysed by GalK and GalPUT. Org Biomol Chem. 2009;7(5):1009-1016. doi:10.1039/b815549f. View Source

Why 2-Deoxy-2-fluorotalose Cannot Be Replaced by FDG, FDGal, or Other Fluorinated Hexoses in Galactokinase-Targeted Research

Fluorinated 2-deoxyhexoses are stereochemically distinct substrates with non-overlapping enzyme specificities. 2-Deoxy-2-fluoro-D-glucose (FDG), the clinical standard for PET, is phosphorylated by hexokinase and traces glucose metabolism; it shows negligible affinity for galactokinase [1]. 2-Deoxy-2-fluoro-D-galactose (FDGal), the C2 epimer of FDT, is recognized by galactokinase but undergoes complex, multi-step metabolism—proceeding beyond phosphorylation to form UDP-FDGal and even UDP-FDG via uridylyltransferase and epimerase [2]. FDT uniquely combines galactokinase-dependent phosphorylation with a truncated metabolic fate: it is converted to a single metabolite (FDT-1-phosphate) that is not further processed by galactose-1-phosphate uridylyltransferase, creating a clean, two-compartment kinetic profile unobtainable with any other fluorinated hexose [3][4]. Substituting FDG, FDGal, or other deoxyfluorohexoses for FDT in PET imaging, metabolic tracing, or enzymatic synthesis protocols will yield fundamentally different—and often uninterpretable—data.

FDG

Hexokinase substrate with negligible galactokinase affinity; may not trace galactose metabolism or provide liver-selective uptake.

FDGal

Multi-step metabolism to UDP-sugars and epimerization to UDP-FDG confound compartmental modeling; reported complex kinetics may obscure kinase quantification.

Other fluorinated hexoses

Lack galactokinase-specific phosphorylation and produce multi-metabolite profiles; stereochemical mismatch limits utility as selective kinase probes.

• [1] Haradahira T, Inoue O, Suzuki K. Regional brain distribution of 2-deoxy-2-[F-18]Fluoro-D-talose: A new PET tracer for measurement of galactokinase activity. J Nucl Med. 1994;35(Suppl.5). OSTI ID: 198105. View Source
• [2] Ishiwata K, Ido T, Imahori Y, Yamaguchi K, Fukuda H, Tada M, Matsuzawa T. Accumulation of 2-deoxy-2-[18F]fluoro-D-galactose in the liver by phosphate and uridylate trapping. Int J Rad Appl Instrum B. 1988;15(3):271-276. doi:10.1016/0883-2897(88)90106-7. PMID: 3260229. View Source
• [3] Haradahira T, Maeda M, Ichiya Y, Masuda K. Metabolic pathway of 2-deoxy-2-[18F]fluoro-D-talose in mice: trapping in tissue after phosphorylation by galactokinase. Nucl Med Biol. 1994 Feb;21(2):269-76. doi:10.1016/0969-8051(94)90019-1. PMID: 9234293. View Source
• [4] Diksic M, Jolly D. Synthesis of 2-deoxy-2-fluorohexoses by fluorination of glycals in aqueous media. Carbohydr Res. 1986 Sep 15;153(1):17-24. doi:10.1016/S0008-6215(00)90191-5. View Source

Quantitative Differentiation Evidence for 2-Deoxy-2-fluorotalose vs. FDGal, FDG, and D-Talose

Simplified Single-Metabolite Fate of FDT vs. Multi-Step UDP-Derivative Metabolism of FDGal

FDT administered to tumor-bearing mice (60 mg/kg i.v.) yields a single detectable metabolite—2-deoxy-2-fluoro-α-D-talose-1-phosphate (FDT-1-P)—with no evidence of UDP-sugar formation [1]. FDT-1-P remained unchanged for at least 3 hours in mouse liver [1]. In vitro, FDT was efficiently phosphorylated by galactokinase, but FDT-1-P failed to serve as a substrate for galactose-1-phosphate uridylyltransferase (GALT), accounting for its metabolic arrest [1]. By contrast, FDGal is converted to FDGal-1-phosphate and subsequently to UDP-FDGal; at 120 min post-injection, 77% of 18F from [18F]FDGal was recovered as UDP-[18F]FDGal in rat liver, and 81% was already phosphorylated to FDGal-1-P by 5 min [2]. FDGal can also undergo epimerization to form UDP-FDG [2].

Metabolic pathway length

Direct head-to-head

FDT: 1 metabolite (FDT-1-P, stable ≥3 h) vs. FDGal: ≥2 metabolites (77% UDP-FDGal at 120 min)

Supports simplified two-compartment kinetic modeling for galactokinase quantification.

Mouse liver in vivo, 19F-NMR; FDGal data in rat liver with 18F-HPLC.

Evidence Dimension	Number of detectable metabolites and extent of metabolic processing beyond phosphorylation
Target Compound Data	1 metabolite (FDT-1-P); no UDP derivative; FDT-1-P stable for ≥3 h in mouse liver
Comparator Or Baseline	FDGal: ≥2 metabolites (FDGal-1-P + UDP-FDGal); 81% phosphorylated at 5 min; 77% conversion to UDP-FDGal at 120 min in rat liver; additional UDP-FDG via epimerization
Quantified Difference	Metabolic pathway length: FDT = 1 step (phosphorylation only); FDGal = 2–3 steps. UDP-sugar formation: 0% (FDT) vs. 77% (FDGal) at 120 min
Conditions	FDT: Mouse liver in vivo, 60 mg/kg i.v., 19F-NMR + radio-TLC (Haradahira et al., Nucl Med Biol 1994); FDGal: Rat liver, 18F-labeled, radio-HPLC (Ishiwata et al., Int J Rad Appl Instrum B 1988)

Why This Matters

The single-metabolite, phosphorylation-only fate of FDT eliminates the need to model uridylyltransferase and epimerase activity in kinetic PET analysis, enabling a cleaner two-compartment model for galactokinase quantification that is not achievable with FDGal.

• [1] Haradahira T, Maeda M, Ichiya Y, Masuda K. Metabolic pathway of 2-deoxy-2-[18F]fluoro-D-talose in mice: trapping in tissue after phosphorylation by galactokinase. Nucl Med Biol. 1994 Feb;21(2):269-76. doi:10.1016/0969-8051(94)90019-1. PMID: 9234293. View Source
• [2] Ishiwata K, Ido T, Imahori Y, Yamaguchi K, Fukuda H, Tada M, Matsuzawa T. Accumulation of 2-deoxy-2-[18F]fluoro-D-galactose in the liver by phosphate and uridylate trapping. Int J Rad Appl Instrum B. 1988;15(3):271-276. doi:10.1016/0883-2897(88)90106-7. PMID: 3260229. View Source

Divergent Brain Phosphorylation Kinetics: FDT Washout vs. FDGal Accumulation in Primate PET

In a head-to-head dynamic PET study in rhesus monkey, [18F]FDT exhibited a 30% loss of initial radioactivity from the occipital cortex within the first 5 minutes after intravenous injection [1]. In marked contrast, [18F]FDGal showed a continuous increase in radioactivity in the identical brain region over the same period [1]. This divergent kinetic behavior directly indicates a smaller phosphorylation rate constant (K3) for FDT catalyzed by brain galactokinase compared to FDGal [1]. Despite the kinetic difference, the regional brain distribution pattern of FDT (highest accumulation in occipital cortex) was similar to FDGal, but distinctly different from FDG [1].

Brain phosphorylation kinetics

Direct head-to-head

FDT: 30% washout within 5 min in occipital cortex vs. FDGal: continuous accumulation

Smaller phosphorylation rate constant (K3) supports kinetic modeling with fewer confounding parameters.

Rhesus monkey dynamic PET, i.v. [18F]FDT and [18F]FDGal.

Evidence Dimension	Early-phase tissue radioactivity kinetics and inferred phosphorylation rate constant (K3) by brain galactokinase
Target Compound Data	30% washout of initial activity in occipital cortex within ≤5 min post-injection
Comparator Or Baseline	FDGal: continuous increase of radioactivity in occipital cortex over same time period
Quantified Difference	Opposite direction of early-phase signal: −30% washout (FDT) vs. net accumulation (FDGal); qualitatively smaller K3 for FDT
Conditions	Rhesus monkey, dynamic PET imaging, i.v. administration of [18F]FDT and [18F]FDGal (Haradahira et al., J Nucl Med 1994, OSTI ID: 198105)

Why This Matters

The slower phosphorylation rate of FDT by brain galactokinase, coupled with its truncated single-metabolite fate, enables a simpler two-compartment kinetic model for regional galactokinase activity quantification, avoiding the parameter confounding introduced by FDGal's multi-step metabolism.

• [1] Haradahira T, Inoue O, Suzuki K. Regional brain distribution of 2-deoxy-2-[F-18]Fluoro-D-talose: A new PET tracer for measurement of galactokinase activity. J Nucl Med. 1994;35(Suppl.5). OSTI ID: 198105. View Source

Synthesis Yield Differential: 7-Fold Lower Direct Fluorination Yield of FDT vs. FDGal from D-Galactal

Fluorination of D-galactal with molecular fluorine (F2) or [18F]F2 in aqueous media produces both C2 epimers: 2-deoxy-2-fluoro-D-galactose (FDGal) and 2-deoxy-2-fluoro-D-talose (FDT) in a single reaction [1]. The chemical yields were 35% and 5%, respectively—a 7-fold difference [1]. When using 3,4,6-tri-O-acetyl-D-galactal as substrate, yields shifted to 38% (FDGal) and 6% (FDT) [1]. A more efficient nucleophilic [18F]fluorination method via triflate displacement on a protected galactofuranoside precursor was later developed, achieving an average radiochemical yield of 29% for [18F]FDT with >98% radiochemical purity, though requiring a multi-step precursor synthesis [2].

Synthesis yield (direct fluorination)

Direct head-to-head

7-fold lower for FDT (5–6%) vs. FDGal (35–38%)

Limited synthetic efficiency may affect procurement cost and lot consistency.

F2 fluorination of D-galactal; nucleophilic 18F method yields 29%.

Evidence Dimension	Chemical yield from glycal fluorination and radiochemical yield from nucleophilic substitution
Target Compound Data	5–6% (F2 fluorination of galactal); 29% (nucleophilic 18F method, >98% radiochemical purity)
Comparator Or Baseline	FDGal: 35–38% (F2 fluorination of galactal from same reaction mixture)
Quantified Difference	7-fold lower direct fluorination yield for FDT (5–6%) vs. FDGal (35–38%)
Conditions	F2/H2O or F2/organic solvent, D-galactal or 3,4,6-tri-O-acetyl-D-galactal substrate (Diksic & Jolly, Carbohydr Res 1986); nucleophilic method: K18F/Kry222 in CH3CN, BBr3 deprotection (Haradahira et al., Int J Rad Appl Instrum A 1992)

Why This Matters

The intrinsically low synthetic efficiency of the talo epimer directly translates to higher procurement cost, limited commercial availability, and greater lot-to-lot variability. Researchers must validate supplier synthetic routes, residual solvent profiles, and enantiomeric purity rather than treating FDT as a commodity chemical.

• [1] Diksic M, Jolly D. Synthesis of 2-deoxy-2-fluorohexoses by fluorination of glycals in aqueous media. Carbohydr Res. 1986 Sep 15;153(1):17-24. doi:10.1016/S0008-6215(00)90191-5. View Source
• [2] Haradahira T, Kato A, Maeda M, Torii Y, Ichiya Y, Masuda K. Synthesis and biodistribution of a fluorine-18 labeled analogue of D-talose: 2-deoxy-2-[18F]fluoro-D-talose. Int J Rad Appl Instrum A. 1992 May;43(5):627-32. doi:10.1016/0883-2889(92)90031-9. PMID: 1325421. View Source

Enzymatic UDP-Sugar Synthesis: 4.7-Fold Higher Yield of FDT vs. Natural D-Talose in GalK/GalPUT One-Pot Reaction

In a one-pot enzymatic synthesis catalyzed by galactokinase (GalK) and galactose-1-phosphate uridylyltransferase (GalPUT), 2-deoxy-2-fluoro-D-talose was converted to UDP-2-fluoro-D-talose (UDP-2F-d-Talp) in 71% yield after 24-hour incubation [1]. Under identical conditions, natural D-talose yielded only 15% UDP-D-talose, while the canonical substrate D-galactose produced 78% UDP-D-galactose [1]. This demonstrates that C2 fluorine substitution dramatically improves substrate acceptance by GalPUT for the talo-configured epimer, bringing FDT to near-galactose-level efficiency (91% of galactose performance).

Enzymatic UDP-sugar yield

Direct head-to-head

4.7-fold higher for FDT (71%) vs. D-talose (15%) in one-pot GalK/GalPUT

Supports preparative-scale synthesis of fluorinated UDP-sugar donor substrates.

24 h incubation at 37°C; FDT achieves 91% of D-galactose efficiency.

Evidence Dimension	Enzymatic conversion yield to UDP-sugar nucleotide
Target Compound Data	UDP-2F-d-Talp: 71% yield (24 h)
Comparator Or Baseline	D-Talose: UDP-d-Talp 15% yield; D-Galactose: UDP-d-Galp 78% yield
Quantified Difference	4.7-fold higher yield for FDT (71%) vs. D-talose (15%); FDT achieves 91% of D-galactose efficiency
Conditions	One-pot GalK/GalPUT coupled assay, 24 h incubation at 37°C (Zhao et al., Org Biomol Chem 2009, Table 5)

Why This Matters

For glycobiology laboratories requiring fluorinated UDP-sugar nucleotides as donor substrates for glycosyltransferase assays or chemoenzymatic glycan synthesis, FDT is a far more practical starting material than natural D-talose, enabling preparative-scale production at substantially reduced enzyme cost.

• [1] Zhao G, Guan W, Cai L, Wang PG. One-pot enzymatic synthesis of UDP-sugars catalysed by GalK and GalPUT. Org Biomol Chem. 2009;7(5):1009-1016. doi:10.1039/b815549f. View Source

Liver-Selective Biodistribution of [18F]FDT: 34.9 %dose/g Hepatic Uptake with Galactose-Competitive Trapping vs. FDG Whole-Body Profile

Biodistribution studies of [18F]FDT in fibrosarcoma-bearing mice at 30 min post-injection (i.v.) showed the highest radioactivity uptake in liver (34.9 %dose/g), followed by kidney (15.9 %dose/g), small intestine (12.9 %dose/g), and fibrosarcoma (5.7 %dose/g) [1]. Liver and tumor uptake slightly increased through 120 min, while kidney and intestine levels declined [1]. Co-administration of non-radioactive D-galactose strongly inhibited hepatic [18F]FDT uptake in rats, confirming galactose-pathway-mediated transport and trapping [1]. By contrast, [18F]FDG—the universal glucose metabolism tracer—exhibits brain- and heart-dominant biodistribution with typical liver uptake well under 5 %dose/g in rodent models [2].

Hepatic biodistribution

Class-level inference

34.9 %dose/g liver uptake at 30 min vs. FDG liver typically

Supports liver-selective galactokinase imaging; galactose-competable uptake confirms pathway specificity.

Fibrosarcoma-bearing mice; FDG rodent distribution from class-level reference.

Evidence Dimension	Tissue-specific radioactivity uptake (%dose/g) at 30 min post-injection
Target Compound Data	Liver 34.9, kidney 15.9, small intestine 12.9, fibrosarcoma 5.7 %dose/g; liver uptake strongly inhibited by D-galactose co-administration
Comparator Or Baseline	FDG: liver typically <5 %dose/g; brain, heart, and tumor as dominant uptake organs (class-level values from rodent PET literature)
Quantified Difference	>7-fold higher liver uptake for FDT vs. FDG; liver-to-brain selectivity fundamentally inverted
Conditions	Fibrosarcoma-bearing mice, 30 min post-i.v., [18F]FDT (Haradahira et al., Int J Rad Appl Instrum A 1992); FDG distribution from published rodent biodistribution and PET imaging studies

Why This Matters

The pronounced and galactose-competable liver selectivity of FDT defines its exclusive research niche for hepatic galactokinase imaging and liver function assessment—applications for which FDG, despite its widespread clinical availability, is fundamentally unsuitable due to its glucose-metabolism-targeted biodistribution.

• [1] Haradahira T, Kato A, Maeda M, Torii Y, Ichiya Y, Masuda K. Synthesis and biodistribution of a fluorine-18 labeled analogue of D-talose: 2-deoxy-2-[18F]fluoro-D-talose. Int J Rad Appl Instrum A. 1992 May;43(5):627-32. doi:10.1016/0883-2889(92)90031-9. PMID: 1325421. View Source
• [2] Haradahira T, Inoue O, Suzuki K. Regional brain distribution of 2-deoxy-2-[F-18]Fluoro-D-talose: A new PET tracer for measurement of galactokinase activity. J Nucl Med. 1994;35(Suppl.5). OSTI ID: 198105. [Class-level comparison of FDT, FDG, and FDGal regional brain distribution patterns]. View Source

Evidence-Backed Procurement Scenarios: When 2-Deoxy-2-fluorotalose Is the Only Suitable Choice

PET Imaging of Regional Galactokinase Activity with Simplified Two-Compartment Kinetic Modeling

Laboratories developing quantitative PET assays for tissue galactokinase activity—particularly in liver and brain—should procure FDT (as [18F]FDT precursor) rather than FDGal. FDT's single-metabolite fate (FDT-1-P only) [3] and smaller phosphorylation rate constant (30% occipital cortex washout within 5 min in primate PET) [2] enable a clean two-compartment model that FDGal cannot support due to its UDP-derivative formation and epimerization to UDP-FDG. This simplification reduces fitting parameters, improves parametric image signal-to-noise ratio, and makes galactokinase quantification feasible in regions with lower enzyme expression. The liver-selective biodistribution of [18F]FDT (34.9 %dose/g at 30 min) further supports hepatic applications [1].

Liver-Specific Metabolic Imaging and Hepatocellular Carcinoma Assessment

For PET studies requiring liver-selective tracer uptake with galactose-pathway specificity, [18F]FDT is indicated by its 34.9 %dose/g liver uptake at 30 min in murine models—more than 7-fold higher than typical hepatic FDG accumulation [1]. The strong competitive inhibition of hepatic FDT uptake by co-administered D-galactose provides a built-in specificity control absent from FDG-based protocols [1]. Research groups studying liver function, galactose elimination capacity, or hepatocellular carcinoma metabolism should prioritize FDT over generic glucose tracers, as FDG's low liver selectivity precludes meaningful hepatic metabolic imaging.

Preparative Enzymatic Synthesis of Fluorinated UDP-Sugar Donor Substrates

Glycobiology laboratories synthesizing UDP-2-fluoro-D-talose for use as a glycosyltransferase donor substrate or mechanistic probe should source FDT based on its 71% enzymatic conversion yield in one-pot GalK/GalPUT reactions—a 4.7-fold improvement over natural D-talose (15% yield) [4]. The near-galactose-level efficiency of FDT in this system makes multi-milligram UDP-sugar production economically feasible without stoichiometric quantities of expensive enzymes [4]. Procurement specifications should require ≥95% chemical purity with confirmation of the D-talo stereochemistry at C2, C3, C4, and C5 to ensure consistent enzymatic performance.

Galactose Metabolism Research Requiring a Non-Uridylatable Kinase Probe

Investigators studying the Leloir pathway, galactokinase enzymology, or galactose-1-phosphate uridylyltransferase (GALT) deficiency require a metabolic probe that is phosphorylated by galactokinase but not further processed. FDT is the only known fluorinated hexose that satisfies this criterion: FDT-1-P is formed efficiently but fails as a GALT substrate due to low enzyme affinity for the fluorinated phosphate [3]. This unique metabolic arrest at the phosphate stage makes FDT an irreplaceable tool for dissecting the contribution of galactokinase to tissue galactose utilization, independent of downstream Leloir pathway enzymes. In contrast, FDGal proceeds to UDP-derivatives and cannot isolate galactokinase activity [3].

Application

Selection Property

Validation Focus

Galactokinase PET imaging

Single-metabolite kinetic profile

Two-compartment model feasibility and parametric image quality

Liver function PET studies

Galactose-pathway selectivity with reported hepatic uptake

Galactose-competable liver uptake specificity

Fluorinated UDP-sugar synthesis

Reported enzymatic conversion efficiency

Stereochemical purity and yield reproducibility

Leloir pathway dissection

Non-uridylatable kinase probe (FDT-1-P arrest)

GALT-independent phosphorylation confirmation

• [1] Haradahira T, Kato A, Maeda M, Torii Y, Ichiya Y, Masuda K. Synthesis and biodistribution of a fluorine-18 labeled analogue of D-talose: 2-deoxy-2-[18F]fluoro-D-talose. Int J Rad Appl Instrum A. 1992 May;43(5):627-32. doi:10.1016/0883-2889(92)90031-9. PMID: 1325421. View Source
• [2] Haradahira T, Inoue O, Suzuki K. Regional brain distribution of 2-deoxy-2-[F-18]Fluoro-D-talose: A new PET tracer for measurement of galactokinase activity. J Nucl Med. 1994;35(Suppl.5). OSTI ID: 198105. View Source
• [3] Haradahira T, Maeda M, Ichiya Y, Masuda K. Metabolic pathway of 2-deoxy-2-[18F]fluoro-D-talose in mice: trapping in tissue after phosphorylation by galactokinase. Nucl Med Biol. 1994 Feb;21(2):269-76. doi:10.1016/0969-8051(94)90019-1. PMID: 9234293. View Source
• [4] Zhao G, Guan W, Cai L, Wang PG. One-pot enzymatic synthesis of UDP-sugars catalysed by GalK and GalPUT. Org Biomol Chem. 2009;7(5):1009-1016. doi:10.1039/b815549f. View Source