Category Archives: Technical Trove

Gluten-Free Spirits and Drinks

Gluten-Free Spirits and Drinks

Gluten-Free Spirits and Drinks Alcoholic beverages are those containing more than 0,5% (vol/vol) of alcohol. They can be obtained by various processes (fermentation, addition, distillation, extraction, etc.). As there is no established classification of alcoholic beverages, alcoholic concentration is the most commonly used: 1) fermented alcoholic beverages such as beer, cider and wine, and 2) distilled beverages and spirits (higher in alcoholic concentration).
Antioxidant Activity and Phenolic Content of Apple Cider

Antioxidant Activity and Phenolic Content of Apple Cider

Antioxidant Activity and Phenolic Content of Apple Cider Nilgün Havva Budak1*, Filiz Ozçelik2 , Zeynep Banu Güzel-Seydim3 1 Department of Food Processing, Eğirdir Vocational School, Süleyman Demirel University, 32500 Eğirdir/Isparta, Turkey 2 Department of Food Engineering, Faculty of Engineering, Ankara University, 06100 Ankara, Turkey 3 Department of Food Engineering, Faculty of Engineering, Süleyman Demirel University, 32260 Isparta, Turkey A R T I C L E I N F O A B S T R A C T Fruit and vegetables are an important component of a healthy diet and the main antioxidant suppliers in the human diet. Consumption of foods derived from fruits and vegetables is also essential; fruit juices, ciders, wines, and vinegars also contain significant amounts of polyphenolic compounds. The aim of the study was to determine the effect maceration of antioxidant activity and phenolic content of apple cider. Red delicious apples were used to produce natural apple cider with and without inclusion of maceration. Samples were taken from fresh red apple juice, macerated samples and apple cider. Apple cider (maceration was applied) (CAM) had the highest total phenolic content, chlorogenic acid, ORAC and TEAC levels. Chlorogenic acid was the dominant phenolic substance in apple juice and cider samples and chlorogenic acid was increased with maceration process. Keywords: Apple cider Chlorogenic acid Maceration ORAC TEAC Introduction Fruit and vegetables are an important component of a healthy diet and, if consumed daily in sufficient amounts, could help prevent major diseases such as cardiovascular diseases (CVDs) and certain cancers. Noncommunicable diseases (NCDs), especially cardiovascular diseases (CVDs), cancer, obesity and diabetes, currently kill more people every year than any other cause of death. The recent Joint FAO/WHO Expert Consultation on diet, nutrition and the prevention of chronic diseases, recommended the intake of a minimum of 400g of fruit and vegetables per day (excluding potatoes and other starchy plant) for the prevention of NCDs as well as for the prevention and alleviation of several micronutrient deficiencies, especially in less developed countries (WHO, 2003). Dietary intake of natural antioxidants has recently received increased attention due to the epidemiological evidence that correlates a regular intake of these products with protection against several diseases (Hertog et al., 1995). Fruits and vegetables are the main antioxidant suppliers in the human diet. Among them, apple is important not only for its high antioxidant content, but also for its acceptance among the general consumer population. Vinson et al. (2001) reported that 22% of the fruit phenolics consumed in the United States came from apple. Eberhardt et al. (2000) found that 100 g of fresh apples have an antioxidant activity equivalent to 1500 mg of vitamin C, and more important, that apple phenolic extracts inhibited proliferation of a human cancer cell line. The major antioxidants present in apple are polyphenols, which include phenolic acids (chlorogenic, cinammic, gallic acid, etc.) and flavonoids (catechin, quercetin, quercetin glycosides, etc.). Apple and apple products (juice, cider, vinegar) are commonly consumed worldwide. Apple polyphenols contain mainly polyphenolic acid derivatives and other flavonoids. Generally, these polyphenols are distributed in the whole fruit, with higher concentrations present in the peel rather than in the flesh (Wolfe et al., 2003). The complexation and antioxidant activity of the major apple polyphenols: Chlorogenic Acid (CA), Rutin (Rt) and Quercetin (Qc) with b-cyclodextrin (b-CD) were studied, by fluorescence spectroscopy and Ferric Reducing/Antioxidant Power Assay (FRAP) techniques (Alvarez-Parrilla et al., 2005). Budak et al. (2011) reported that chlorogenic acid is also commonly in apple cider vinegar. The aim of the study was to determine the effect of maceration on antioxidant activity and phenolic content level of ciders derived from red delicious apples during cider production. Red delicious apples were used to produce natural apple cider with and without inclusion of maceration. Material and Methods Material “Red Delicious” apple was harvested in Gelendost, Isparta, and appropriately transported to Fermentation Laboratory in the Süleyman Demirel University Gelendost Vocational School (Isparta, Turkey). Apple Cider Production “Red delicious” apples were used to make natural apple cider to determine the effects of maceration. Flow scheme of apple cider production methods are presented in Figure 1. The samples were named as apple juice (AJ), apple juice sample taken after maceration (ASM), apple juice after fermentation without maceration sample (AS), apple cider (with maceration) (CAM) and apple cider sample (without maceration) (CA). Briefly, after red apples were broken into pieces maceration was carried out during seven days. Addition of 10% pomace was used in the maceration step to increase the polyphenolic contents. Apple cider was obtained after processing apple juice was fermented for two months. Compositional Analysis Total titratable acidity, density and total ash of apple juice and apple cider samples were determined according to AOAC methods (1992). Total sugar in apple juice and maceration samples were analyzed according to the Luff Schoorl methods (AOAC 1990). Water soluble solid (Brix) was measured with Abbe refractometer (Bellingham Stanley Limit 60/70 Refractometer, England). Ethanol content was determined with alcoholmeter (Dujardin-Salleron, France). Total Antioxidant Activity Total phenolic content: Total phenolic contents of the samples were determined according to Folin-Ciocalteu method using gallic acid as a standard (Singleton and Rossi, 1965; Singleton et al., 1999). After addition of Folin-Ciocalteu reagent to the sample solution it was allowed to react for 6 min. Reaction was stopped with using 1.50 mL of 20% sodium carbonate. The extracts were oxidized with Folin-Ciocalteu reagent, and the reaction was neutralized with sodium carbonate. The absorbance of the resulting blue colour was developed in 120 min in a dark place, and the absorbance was determined at 760 nm using a spectrophotometer (Shimadzu Scientific Instruments, Inc., Tokyo, Japan). The measurement was calculated using a standard curve of gallic acid and expressed as milligrams of gallic acid equivalents (GAE) L-1 . 2,2’-azinobis (3-ethlybenzthiazoline)-6-sulfonic acid (ABTS) Assay: 2,2’-azinobis (3-ethlybenzthiazolin-6- sulfonic acid) diammonium salt (ABTS+ ) radical cation was prepared by reacting 7 mM ABTS stock solution with 2.45 mM potassium persulfate (Re et al., 1999). ABTS+ inhibition against Trolox (6-hydroxy-2, 5, 7, 8- tetramethylchroman-2-carboxylic acid) was spectrophotometrically measured (Seeram et al., 2005). Figure 1 Flow chart of apple cider (maceration was applied and was not applied) The concentration of the resulting blue-green ABTS radical solution was adjusted to an absorbance of 0.700 ± 0.020 at 734 nm in a spectrophotometer (Shimadzu Scientific Instruments, Inc., Tokyo, Japan). TEAC values of samples were calculated from the Trolox standard curve and expressed as Trolox equivalents (in μmol/ml of sample). Oxygen Radical Absorbance Capacity (ORAC) Assay: All samples were analysis using the Oxygen Radical Absorbance Capacity (ORAC) (Wu et al., 2008). The samples were appropriately diluted with phosphate buffer (pH 7.4) for ORAC analysis. An aliquot (25 μL) of the diluted sample, blank (phosphate buffer) or Trolox calibration solutions were added to a black, clear-bottom triplicate well in 96 well bottom reading microplate. After the addition of 150 µM flourescein stock solution (0,004 µM) to each well the microplate was incubated at 37 °C for 30 min. Then, 25 µL 2, 2’-Azobis (2-amidinopropane) Harvesting “Red Delicious” apples Appropriate transportation of the apples to the Fermentation Laboratory Breaking into pieces of apples Pressing Obtained Apple Juice (AJ) Resettle and Racked of apple wine Resettle and Racked of apple wine Maturation 2 months at 25°C Maturation 2 months at 25°C Packaging (CA) Packaging (CA) Ethanol Fermentation:Saccharomyc es cerevisiae (3 %) inoculation 7 days at 25 °C (AS) Maceration with pulp and Ethanol Fermentation: Saccharomyces cerevisiae (3 %) inoculation 7 days at 25 °C (ASM) Budak et al./ Turkish Journal of Agriculture – Food Science and Technology, 3(6): 356-360, 2015 358 dihydrochloride (AAPH) solution (153 mM) was added to start the reaction. The microplate reader was programmed to record the fluorescence reading with an excitationemission wavelength of 485 – 520 nm using software Gen 5 TM. Antioxidant activity was kinetically measured with Biotek Synergy™ HT Multi-Detection Microplate Reader (Winooski, Vermont, USA). Quantification of Phenolics by High Performance Liquid Chromatography: Phenolic compounds were evaluated by reversed-phase high performance liquid chromatography (RP-HPLC, Shimadzu Scientific Instruments, Kyoto, Japan). Phenolic compositions of the extracts were determined by a modified method of Schulz et al. (2001). Detection and quantification were carried out with a LC-10ADvp pump, SIL-10ADvp auto sampler, a Diode Array Detector, a CTO-10Avp column heater, SCL-10Avp system controller and DGU-14A degasser (Shimadzu Scientific Instruments, Kyoto, Japan). Separations were conducted at 30 ºC on Agilent® Eclipse XDB C-18 reversed-phase column (250 mm x 4.6 mm length, 5 µm particle size). The mobile phases were A:3.0% acetic acid in distilled water and B: methanol. Flow rate was 0.8 mL/min. Gallic acid, catechin, caffeic acid, chlorogenic acid, p-coumaric acid, ferulic acid, rutin, resveratrol and syringic acid were used as standard. Identification and quantitative analysis were done by comparison with standards. Statistics All data were reported as the mean and standard error. Results analyzed by using SPSS for Windows (version 17.0, SPSS Inc.). Apple cider production was repeated three times. Values represent means of triplicate repetition with parallels. The significance was established at P<0.05. Results and Discussion Composition analysis Total titratable acidity, density, Brix, total ash total sugar and alcohol in apple juices and apple cider samples are reported in Table 1. Total titratable acidity was increased during ethanol fermentation. Especially, total titratable acidity in the sample taken from maceration was significantly higher than the sample that maceration was not applied (P<0.05). Total titratable acidity values in Cashew juice and Golden Delicious fresh apple juice samples were 2.4 g/L and 2.88 g MA/L, respectively (Mohanty et al., 2006; Suárez-Jacobo et al., 2011). Density of the samples varied between 0.9987 -1.0517 g/cm3 . Density and Brix values were significantly decreased during alcohol fermentation due to the conversion of sugar to ethanol (P<0.05). Budak and Güzel-Seydim (2010) explained that total solids and Brix of samples significantly decreased after maceration due to pressing, resettling and racking during grape wine production. Brix of Cashew apple wine was found to be 2.0% (Mohanty et al., 2006). In our study, brix of the apple cider sample was 3.83%. Alcohol contents of apple cider samples were between 5.40-6.10 %. The total sugar contents were 144.24, 85.77, 95.56 g/L in AS, ASM and AS, respectively. Total sugar content also decreased due to the ethanol fermentation by Saccharomyces cerevisae (Budak and Güzel-Seydim, 2010). Total Antioxidant Activity Total phenolic content, TEAC and ORAC results express the total antioxidant activity in the samples. Total phenolic content (mg/L), TEAC (mmol/L) and ORAC(μmol/mL) values of samples are presented in Figures 2 and 3, respectively. AJ sample had the lowest total phenolic content whereas CAM sample had the highest total phenolic content among the samples (P<0.05). Seeram et al. (2008) reported that TEAC and ORAC value of apple juice samples ranged between 2.5- 6.2 μmol of TE/mL and 2.7-4.3 μmol/mL, respectively. Total phenolic contents of AS and CAM samples were 459.31 mg/L and 1026.74 mg/L, respectively (Figure 3). Lachman et al. (2006) determined that total phenolic content of apple juice samples obtained from different varieties were between 760.03- 1343 mg/L. In our study, total phenolic content of ASM and AS samples were 777.83 and 733.61 mg/L. Contents of TEAC, ORAC and total phenolic content in CAM sample was the highest in all samples. TEAC and ORAC values of CAM sample were 13.27 mmol/L and 9.84 µmol TE/mL, respectively. ORAC values of apple cider samples were the highest in all samples. Especially, ORAC values of CAM sample was the highest in all samples (Figure 3). Antioxidant activities of macerated juice samples and ciders were higher than the samples that maceration was not applied. Phenolic Substances Gallic acid, catechin, epicatechin, caffeic acid, chlorogenic acid, and p-coumaric acid were detected in apple juice and apple cider samples (Table 2). Contents of catechin, epicatechin, and chlorogenic acid were identified in all samples. Gallic acid only was detected in apple juice sample. The content of catechin in CAM sample was significantly higher than CA sample (P<0.05). The amount of epicatechin was 4.63 mg/L in CAM sample while CA sample contained 3.33 mg/L (P<0.05). Chlorogenic acid was the dominant phenolic substance in apple juice samples; especially, ASM and CAM samples had the highest content of chlorogenic acid (P<0.05). Chlorogenic acid significantly increased with maceration. p-Coumaric acid contents of apple cider samples ranged between 0.03 and 0.04 mg/L. AlvarezParilla et al. (2005) reported that chlorogenic acid is one of the important apple polyphenols. Chlorogenic acid, catechin, epicatechin, caffeic acid were high concentrations in apple cider that maceration was applied. Therefore, maceration process was important for the concentrations of the polyphenolic compounds. Polyphenolic content (chlorogenic acid, catechin, epicatechin and caffeic acid) of CAM sample had the highest values similar to antioxidant activity of CAM sample (total phenolic content, TEAC and ORAC contents). It has been reported that wine vinegars show an antioxidant capacity that is correlated with their polyphenolic content (Dávalos et al., 2005). In this study, phenolic substances were increased by fermentation. Budak et al./ Turkish Journal of Agriculture – Food Science and Technology, 3(6): 356-360, 2015 359 Table1 Composition Analysis of Samples Samples TTA1Total Density (g/cm3 ) Brix (%) Total sugar (g/L) Total Ash (g/L) Ethanol (v/v) AJ2 1.9±0.04a 1.0517±0.00a 11.67±1.42a 144.24±1.11a 1.9±0.01a – ASM3 2.3±0.03a 0.9999±0.00b 4.58±0.71b 85.77±3.26b 1.7±0.02a 3.10±0.14b AS4 3.7±0.06c 1.0014±0.00c 5.33±0.34b 95.56±0.41b 1.7±0.00b 2.90±0.12b CAM5 2.3±0.03a 0.9987±0.00b 3.83±0.83b – 1.9±0.01a 6.10±0.15a CA6 3.5±0.04c 0.9987±0.00c 3.83±0.13b – 1.8±0.01b 5.40±0.11b 1TTA: Total Titratable Acidity (g/L), 2AJ: Apple juice, 3ASM: Apple juice taken after maceration, 4AS: Apple juice without maceration sample, 5CAM: Apple cider (maceration was applied), 6CA: Apple cider sample (maceration was not applied) Table 2 Phenolic Compounds of Samples Samples Gallic acid (mg/L) Catechin (mg/L) Epicatechin (mg/L) Caffeic acid (mg/L) Chlorogenic acid (mg/L) p-Coumaric acid (mg/L) AJ1 0.43±0.06b 0.50±0.00c 1.60±0.10bc – 12.26±3.37c – ASM2 – 1.47±0.29b 3.50±0.78a 0.46±0.24ab 18.53±4.06a – AS3 – 1.56±0.13b 4.13±0.36a – 17.86±0.60b 0.05±0.01b CAM4 – 2.13±0.28a 4.63±1.20a 0.96±0.08a 24.13±3.46a 0.03±0.01b CA5 – 1.46±0.23b 3.33±0.23ab 0.75±0.05a 16.50±2.27b 0.04±0.00b 1AJ: Apple juice, 2ASM: Apple juice taken after maceration, 3AS: Apple juice without maceration sample, 4CAM: Apple cider (maceration was applied), 5CA: Apple cider sample (maceration was not applied) Figure 2 Total Phenolic Content of Apple Juices and Cider Samples AJ: Apple juice, ASM: Apple juice taken after maceration, AS: Apple juice without maceration sample, CAM: Apple cider (maceration was applied) , CA:Apple cider sample (maceration was not applied) Figure 3 Antioxidant activity of samples by ABTS (TEAC) assay and ORAC assay AJ: Apple juice, ASM: Apple juice taken after maceration, AS: Apple juice without maceration sample, CAM: Apple cider (maceration was applied) , CA: Apple cider sample (maceration was not applied) Conclusion This is the first report confirming that maceration had positive effects on bioactive components of apple cider. Results of this study showed that polyphenolic compounds and antioxidant activity significantly increased in maceration process. Antioxidant activity of apple cider (maceration included) sample was higher than that of apple cider sample (maceration was not applied). Chlorogenic acid was the dominant phenolic substance in apple juice samples while chlorogenic acid increased during maceration. Chlorogenic acid, catechin, epicatechin, caffeic acid contents of macerated apple cider were in high concentrations. Therefore, inclusion of maceration in process would be important for concentration of bioactive compounds. References Alvarez-Parrilla E, Rosa LDL, Torresrivas F, Rodrigo-Garcia J, Gonzalez’Lez-Aguilar GA. 2005. Complexation of Apple Antioxidants: Chlorogenic Acid, Quercetin and Rutin by bCyclodextrin (b-CD). J Incl Phenom Macro 53:121–129. AOAC. 1990. Association of Official Analytical Chemists. Official Methods of Analysis, 13th edition. Washington DC. AOAC. 1992. Association of Official Analytical Chemists. Official Methods of Analysis, 15th edition. Washington DC. Budak HN, Güzel-Seydim Z. 2010. Antioxidant activity and phenolic content of wine vinegars produced by two different techniques. J Sci Food Agric, 90: 2021–2026. Caponia F, Alloggio V, Gomes T. 1999. Phenolic compounds of virgin olive oil: influence of paste preparation techniques. Food Chem 63: 203-209. Eberhardt MV, Lee CY, Liu RH. 2000. Antioxidant activity of fresh apples. Nature. 405: 903-4. Budak et al./ Turkish Journal of Agriculture – Food Science and Technology, 3(6): 356-360, 2015 3
High-Selectivity Electrochemical Conversion of CO2 to Ethanol using a Copper Nanoparticle/N-Doped Graphene Electrode

High-Selectivity Electrochemical Conversion of CO2 to Ethanol using a Copper Nanoparticle/N-Doped Graphene Electrode

Authors Dr. Yang Song, Dr. Rui Peng, Dale K. Hensley, Dr. Peter V. Bonnesen, Dr. Liangbo Liang, Dr. Zili Wu, Dr. Harry M. Meyer III, Dr. Miaofang Chi, Dr. Cheng Ma, Dr. Bobby G. Sumpter, Dr. Adam J. Rondinone First published: 28 September 2016Full publication history   Abstract   Though carbon dioxide is a waste product of combustion, it can also be a potential feedstock for the production of fine and commodity organic chemicals provided that an efficient means to convert it to useful organic synthons can be developed. Herein we report a common element, nanostructured catalyst for the direct electrochemical conversion of CO2 to ethanol with high Faradaic efficiency (63 % at −1.2 V vs RHE) and high selectivity (84 %) that operates in water and at ambient temperature and pressure. Lacking noble metals or other rare or expensive materials, the catalyst is comprised of Cu nanoparticles on a highly textured, N-doped carbon nanospike film. Electrochemical analysis and density functional theory (DFT) calculations suggest a preliminary mechanism in which active sites on the Cu nanoparticles and the carbon nanospikes work in tandem to control the electrochemical reduction of carbon monoxide dimer to alcohol. Introduction Closing the carbon cycle by utilizing CO2 as a feedstock for currently used commodities, in order to displace a fossil feedstock, is an appropriate intermediate step towards a carbon-free future. Direct electrochemical conversion of CO2 to useful products has been under investigation for a few decades. Metal-based catalysts, such as copper,[1] platinum,[2] iron,[3] tin,[4] silver,[5] and gold,[6] along with carbons such as g-C3N4 [7] have been the primary focus for CO2 reduction, with some very high Faradaic efficiencies for methane conversion. Copper is arguably the best-known metal catalyst for electrochemical CO2 reduction,[8] capable of electrochemically converting CO2 into more than 30 different products,[9] including carbon monoxide (CO), formic acid (HCOOH), methane (CH4) and ethylene (C2H4) or ethane (C2H6), but efficiency and selectivity for any product heavier than methane are far too low for practical use.[10] Competing reactions limit the yield of any one liquid product to single-digit percentages.[8]   Polycrystalline Cu foil produces a mixture of compounds in CO2-saturated aqueous solutions that are dominated either by H2 at low overpotential, or by CO and HCOO− at high overpotential, or by hydrocarbons and multi-carbon oxygenates at the most extreme potentials.[9, 11] Theoretical studies predict that graphene-supported Cu nanoparticles would enhance catalytic activity due to the strong Cu – graphene interaction via defect sites,[12] which would stabilize the intermediates from CO2 reduction and improve selectivity towards hydrocarbon products as methane and methanol at lowered overpotential. Early studies revealed that the electrode surface was dominated by adsorbed CO during the CO2 reduction and that CO acted as intermediate in the production of hydrocarbons.[13] Cu produces hydrocarbons and multi-carbon oxygenates when supplied with CO in the absence of CO2, but very negative potentials are still required to promote CO reduction over H2 evolution. Large overpotentials preclude energetically efficient electrolysis and favor hydrocarbons over liquid oxygenates. Recently, high selectivity of CO electroreduction to oxygenates, with ethanol as the major product, was achieved by oxide-derived Cu, in which the surface intermediates were stabilized by the grain boundaries.[14]   We previously reported on a highly textured nitrogen-doped, few-layer graphene electrode that presents with a surface of intense folds and spikes, which we termed carbon nanospikes or CNS. The CNS structure is disordered due to the high nitrogen content which prevents well-ordered stacking. In the current report, a carbon nanospike (CNS) electrode with electronucleated Cu nanoparticles (Cu/CNS) is shown to have much higher selectivity for CO2 electroreduction than H2 evolution, with a subsequent high Faradaic efficiency to produce ethanol. We believe this is achieved both from the high intrinsic CO2 reduction activity of Cu and from the synergistic interaction between Cu and neighboring CNS, which controls reduction to alcohol. The major CO2 reduction product is ethanol, which corresponds to a 12 e− reduction with H2O as the H+ source, display math   display math where E0 is the equilibrium potential. By comparing Cu/CNS to control electrodes comprised of 1) Cu on glassy carbon and 2) bare CNS, we demonstrate that CO2 reduction activity is not a simple consequence of either Cu or CNS. While the reaction mechanism is not yet elucidated, we hypothesize an interaction between adjacent catalytic sites on the Cu and CNS, facilitated by the nanostructured morphology of the catalyst that prevents complete electrochemical reduction to ethylene or ethane, resulting in a high yield of ethanol. Results and discussion   The bare CNS electrode (Supporting Information Figure S1) was characterized in our previous study as a dense nanotextured carbon film terminated by randomly oriented nanospikes approximately 50–80 nm in length, where each nanospike consists of layers of puckered carbon ending in a ∼2 nm wide curled tip.[15] The film is grown by a relatively simple direct-current plasma-enhanced chemical vapor deposition reaction using acetylene and ammonia as reagents.   The CNS film grows quickly and adheres well to the highly-doped silicon wafers that were used for this study. Raman spectra indicate that CNS have a similar structure to disordered, few-layer graphene.[15] The CNS is not crystalline and does not diffract. XPS indicates a nitrogen doping density of 5.1 ± 0.2 % atomic, with proportions of pyridinic, pyrrolic (or piperidinic) and graphitic nitrogens of 26, 25 and 37 % respectively, with the balance being oxidized N. In the current experiment, nanoparticles of Cu were electronucleated from CuSO4 solution directly onto the CNS (Supporting Information, Experimental Methods), and imaged via SEM shown in Figure 1. Electronucleation does not require templating surfactants to control the nanoparticle growth, and leaves the particle surfaces clean. The texture of the CNS promotes nucleation resulting in a large number of smaller particles, in comparison to the glassy carbon control which produced larger particles under identical conditions (Supporting Information Figure S2), with a similar amount of Cu deposited. These well-dispersed Cu particles ranged from about 30 nm to 100 nm with average size of 39.18 nm, with a density ca. 2.21 × 109 particles cm−2 (Figure 1B inset). According to the average particle size, the coverage of Cu on CNS is ca. 14.2 %. TEM measurements (Figure 2 inset) confirm particle size observed via SEM. High-resolution transmission electron microscopy on scraped samples (HR-TEM) shows the Cu/CNS interface (Figure 2 main) and illustrate a close proximity between Cu and CNS. The lattice spacing of this representative Cu nanoparticle was measured as 0.204 nm, which is consistent with Cu (111). Cu2O with lattice spacing ca. 0.235 nm were present on the Cu nanoparticles surface in this image, however due to the negative potential applied for Cu deposition, the oxide likely results from exposure to air during sample preparation and transportation between measurements.[16] The surface area of the textured surface of CNS and the glassy carbon was measured based on the double layer capacitance on both electrodes in 0.1 M KOH. Capacitance was measured by recording anodic-cathodic charging currents (in the potential region where Faradaic processes are absent; see Supporting Information Figure S3). The active surface area of CuNPs was additionally measured by Pb underpotential deposition (Supporting Information) of a representative sample, but could not be measured for each sample without contamination. The Cu nanoparticles typically contribute approximately 8 % to the total electrode ECSA for the CNS. To measure the physical stability of the catalyst, SEM images were collected of the particles and the CNS cross sections before and after a 6-hour reduction experiment (Supporting Information Figures S4, S5).   Figure 1. Figure 1. Open in figure viewerDownload Powerpoint slide Representative SEM images of Cu/CNS electrode with (A) low and (B) high magnification. The average particle size is approximately 39 nm (C) as measured by automated particle sizing of the micrographs.. Figure 2. Figure 2. Open in figure viewerDownload Powerpoint slide HR-TEM of electrodeposited copper nanoparticles on carbon nanospike electrode. Electrodeposited particles are imbedded in N-doped carbon nanospikes providing intimate contact between copper surface and reactive sites in the carbon. CO2 electroreduction activity was first measured by linear sweep voltammetry (LSV) in the potential range of −0.00 to −1.30 V vs. RHE at a sweep rate of 0.05 V s−1 as shown in Figure 3. In the presence of CO2-saturated potassium bicarbonate electrolyte, using the Cu/CNS, significant anodic shifts in the onset potential are observed compared to that under an argon atmosphere; the onset of activity in CO2 saturated electrolyte is ∼ 0.3 V more positive than in argon purged electrolyte. Note that unlike the featureless voltammograms obtained under an argon atmosphere, a subtle current plateau is obtained at ∼ −0.9 V on electrodes with Cu nanoparticle in CO2 saturated electrolyte. But in the case of pristine CNS electrode, no activity towards CO2 reduction is observed except the onset of hydrogen evolution at much more negative potential. Larger current densities were obtained in Cu/CNS than either of the controls.   Figure 3. Figure 3. Open in figure viewerDownload Powerpoint slide LSV curves in potential range of 0.00 to −1.30 V vs. RHE at a sweep rate of 0.05 V s−1 in 0.1 M KHCO3 under (A) argon and (B) CO2 atmosphere on pristine CNS (black), Cu/glassy carbon (red) and Cu/CNS (blue) electrodes. The current density is calculated using the electrochemical surface area (ECSA) of the electrode based on the double layer capacitance on CNS and glassy carbon electrodes in 0.1 M KOH, respectively.. Chronoamperometry (CA) measurements were conducted over a potential range from −0.7 to −1.3 V, which included these two reduction waves (representative data in Figure 4B for Cu/CNS and Supporting Information Figure S6 A for bare CNS and Cu/glassy carbon controls). New electrodes were fabricated for each data point. The gaseous and liquid products of each CA run were analyzed by gas chromatography (GC) and NMR (of headspace and electrolyte, respectively) to calculate overall current density and Faradaic efficiency for CO2 reduction and for each product. The overall sustained current density for CO2 reduction, JCO2 redn, increased with more negative potential (Supporting Information Figure S6B) for all three electrodes, consistent with that shown in LSV curves. The Cu/CNS electrode had a greater propensity for CO2 reduction than either the Cu/glassy carbon or bare CNS electrodes; for instance, JCO2 redn from Cu/CNS was 5-fold higher than for bare CNS and 3-fold higher than for Cu/glassy carbon, at −1.2 V.   Figure 4. Figure 4. Open in figure viewerDownload Powerpoint slide Fractional Faradaic efficiency of electrochemical reduction products at various potentials (A). The distribution of products indicates that up to −0.9 V, only gas phase products are produced. At more negative potentials, the rate of CO production on the copper surface is high enough to allow CO dimerization to occur, producing C2 products and subsequently ethanol. Chronoamperometry on Cu/CNS at −1.2 V (B) indicates that the electrode is stable although the distribution of products does change with time, beginning with a higher rate of H2 production which drops after the first 5000 seconds. Additional information including relative errors is available in Supporting Information Figure S7. The fractional Faradaic efficiency was computed by dividing the total electrons into each product (determined independently by chemical analysis) by the total electrons passed during the amperometry experiment. The fractional Faradaic efficiencies for Cu/CNS plus the controls at a range of potentials are shown in Figure 4 A, and for Cu/CNS at −1.2 V over a 6-hour experiment in Figure 4B (Additional data including relative error for Cu/CNS are available in Supporting Information Figure S7). Due to experimental losses between the anode and cathode, the total fractions are less than 100 %. The Cu/CNS electrode appears to be stable, as the current density and fractional Faradaic efficiencies for each product barely decreased over the 6 h experiment (Figure 4 A). No significant changes in the Cu nanoparticle size and or CNS thickness was observed from SEM (Figures S4, S5), indicating that the Cu/CNS is stable under these experimental conditions.   At −0.9 V vs. RHE and more positive potential, only gas phase products H2, CO and CH4 were obtained from all three electrodes with CH4 as the major product Cu/CNS. In contrast, with bare CNS and Cu/glassy carbon, CO was the major product and the CO / CH4 ratio was almost independent of potential. The higher selectivity towards CH4 in Cu/CNS indicates a higher degree of surface-bound CO hydrogenation, which is a key step in the formation of CH4.[17] At −1.0 V vs. RHE and more negative potential, the current density of CO2 reduction increased and ethanol was produced (as a liquid soluble in the aqueous electrolyte) only from Cu/CNS. In comparison, only CO and CH4 were produced from both control electrodes. At-1.3 V vs. RHE the Cu/glassy carbon also produced trace ethylene (representative GC traces, Supporting Information Figure S8). GC and NMR analysis in search of other products more commonly produced by copper electroreduction, such as methanol or ethane only indicated (representative NMR, Supporting Information Figure S9) occasional trace formate from Cu/CNS.   Examining the breakdown of Faradaic efficiencies for various reactions on Cu/CNS, reveals that at −1.2 V (Figure 4 A), ethanol conversion exhibited the highest efficiency at 63 % (that is, 63 % of the electrons passing through the electrode were stored as ethanol). Also at −1.2 V vs. RHE, the Faradaic efficiency of gas phase products methane and CO dropped to 6.8 % and 5.2 %, respectively. The Faradaic efficiency of CO2 reduction (competing against water reduction) is 75 %. This means that under the best conditions, the overall selectivity of the reduction mechanism for conversion of CO2 to ethanol is 84 %.   The partial current density and Faradaic efficiency of each product from Cu/CNS electrode at various potentials were illustrated in Figure 5. The partial current density and Faradaic efficiency for CO and methane exhibited a volcanic shape dependence to the potentials applied. The maximum total current density and Faradaic efficiency were observed at −1.0 V vs. RHE, and decreased when ethanol generation began. The partial current density for ethanol generation increased dramatically with more negative potential until reaching −1.2 V vs. RHE, where the maximum Faradaic efficiency for ethanol generation was also achieved. Above −1.2 V vs. RHE, the rate of increase for ethanol current density was slower, consistent with CO2 mass transport limitations. Data were not collected above −1.3 V vs. RHE because hydrogen bubbles that evolved from water reduction blocked the electrode. The decline of Faradaic efficiency for ethanol above −1.2 V vs. RHE suggests that the catalyst reached the mass-transport-limited current density for CO2 reduction, and therefore hydrogen evolved via H2O reduction at unoccupied active sites.   Figure 5. Figure 5. Open in figure viewerDownload Powerpoint slide Partial current density (J, red) and Faradaic efficiency (FE, blue) of CO2 reduction products from Cu/CNS electrode at various potentials. Previous reports of CO2 electroreduction on copper have demonstrated a variety of C1 and C2 products, including CO, CH4, CH2O2, ethane, ethylene, ethanol. Heavier hydrocarbons have not been reported as majority products.[9] Concerning the reaction mechanism, initial electron transfer to adsorbed CO2 will form CO2•−ads, which can be further reduced to COads or other C1 intermediates (CHOads or CH2Oads) with additional proton-electron transfer. CO will result from desorption of COads at this stage, or alternatively, further electron transfer to these surface-adsorbed species will lead to CH4.[1b, 13b] CO2 reduction results on the two controls, bare CNS and Cu nanoparticles on glassy carbon, indicate that both Cu metal and CNS are active for electrochemical CO2 reduction. On the Cu surface, stronger adsorption of CO exists than bare CNS, which provides stable intermediates for further reduction to CH4 on Cu/glassy carbon. In contrast, CO was released rather than reduced to CH4 on bare CNS.   Tafel plots (overpotential vs. the log of partial current density) for CO and CH4 are shown in Figure 6. For all three samples the plots are linear at low overpotential range with a slope that is consistent with a rate-determining initial electron transfer to CO2 to form a surface adsorbed CO2•− intermediate (120 mV / dec), a mechanism that is commonly invoked for metal electrocatalysts.[8b] At high overpotential range, steep slopes were obtained, probably indicating control by the combined effects of gas diffusion and ionic mass transport.[18] Comparing Cu/CNS to the control electrodes, a direct and intimate contact was introduced between Cu and CNS (Figure 2). Lim et al. predicted a strong interaction between Cu nanoparticles and carbon, and we expect that to extend to CNS as well.[12] We expect that the strong interaction provides an environment in which a mechanism involving reactive sites on both the Cu surface and on the N-doped CNS may dominate.   Figure 6. Figure 6. Open in figure viewerDownload Powerpoint slide (A) CO and (B) CH4 partial current density Tafel plots. The Cu/CNS catalyst is unusual because it primarily produces ethanol rather than methane or ethylene. Ethanol, as a C2 product, requires carbon-carbon coupling between surface-adsorbed intermediates at some point during the reduction reaction. Recent calculations on C−C coupling on Cu(211) surfaces suggest the kinetic barriers for the coupling are strongly influenced by the degree of the adsorbed CO hydrogenation.[19] These kinetic barriers tend to decrease with increasing degree of the surface bound CO hydrogenation, which can favor the C2 products from CO2 reduction.[20] A high percentage of C2 products would indicate that coupling is preferred to desorption and loss of C1 intermediates, and this preference for adsorption may be due to the nanostructured nature of the surface. Although initial CO2 reduction appears to be the rate-limiting step, the resulting intermediate must be stable enough to persist until a second intermediate is available for C2 coupling. The coupling may be between two surface-bound C1 intermediates, or between a surface-bound C1 intermediate and a nearby C1 intermediate in solution.[17, 21]   The maximum Faradaic efficiency of ethanol for Cu/CNS is reached at −1.2 V vs. RHE. Further increase in overpotential (−1.3 V vs. RHE) increases Jethanol, but results in a lower Faradaic efficiency due to an increase in H2 production. Hence the proton and electron transfers to C1 become more favorable to produce CH4, which provides a competing pathway against C2 coupling. The details of the reaction mechanism are still to be determined at this time, however there are some lessons in the literature that may yield insights into the high selectivity of this catalyst. Ordinarily, on bulk copper the coupled C2 would continue to be reduced to ethylene or ethane so long as the product was in contact with the copper electrode.[8a, 22] In contrast, with this experiment we have not been able to detect any C2 product except ethanol using the Cu/CNS (ethylene was detected in the control sample Cu/glassy carbon), indicating that the dominant reaction mechanism precludes competitive reduction to ethylene or ethane. Kondo, et al. reported that the electronic structure near the Fermi level of graphene is modified in N-doped graphene, where localized π electronic states are reported to form at the neighboring carbon atoms, and propagate anisotropically around the defect due to the perturbation of the π-conjugated system.[23] Due to electron-withdrawing effects in the graphene π-conjugated system, the carbon atoms adjacent to nitrogen are positively polarized. This polarization may provide an active site adjacent to the copper for the C2 intermediates to adsorb, which may inhibit complete electroreduction.[24] Other doped or defected graphenes are well known to be catalytically active for reactions such as dehydrogenation.[25]   First-principles density functional theory (DFT) calculations were carried out to investigate whether the nitrogen dopant or strongly curled morphology in the CNS can help to adsorb the C2 intermediates. As CNS has a similar structure to multilayer graphene, a graphene sheet is adopted to model the interaction between CNS and the C2 intermediates. The dimer of CO, OCCO, was chosen as a C2 intermediate candidate[19] for modeling of the interaction with CNS. For a pristine graphene sheet, our calculations suggest the binding energy between OCCO (through one oxygen atom) and graphene is 0.19 eV with a separation distance ∼2.95 Å (Supporting Information Figure S10 A). For N-doped graphene, the N dopant and adjacent carbon atoms become more active so that the binding energy with OCCO is increased to 0.64 eV with the separation distance shortened to ∼2.70 Å (Supporting Information Figure S10B). The tripling of the binding energy to 0.64 eV clearly indicates that the C2 intermediates, once formed, can be adsorbed by N-doped CNS fairly strongly and may not desorb easily at room temperature. Furthermore, it is important to note that CNS are puckered and curled, indicating local corrugation on the surface. It has been shown previously that local deformation or buckling could enhance the molecular adsorption on carbon nanotubes and graphene.[26] Here we also considered the buckling of pristine and N-doped graphene to investigate the local curvature effect on OCCO adsorption. Upon buckling, the binding energy between OCCO and the concave surface of pristine graphene is increased to 0.34 eV (Supporting Information Figure S10C), while the binding energy between OCCO and the concave surface of N-doped graphene is enhanced to 0.74 eV (Supporting Information Figure S10D). Therefore, the corrugation and curvature naturally embedded into CNS could also help to strengthen the binding between CNS and the C2 intermediates. In addition, we studied the interaction between OCCO and the copper surface, and found that a strong covalent binding is present (when the molecule approaches the Cu surface oriented with one end closer to the surface) with the separation distance reduced to ∼1.98 Å and binding energy increased to ∼1.21 eV (Supporting Information Figure S11). Compared to this relatively strong covalent bond, the binding between OCCO and CNS is weaker, though strong enough to prevent easy desorption of the C2 intermediate. The calculations offer important insights into the observed selective reduction, and we expect that the oxygen atom on one end of the C2 intermediates is covalently bound on reactive copper surface for complete reduction to -CH3, while the oxygen atom on the other end of the C2 intermediates is adsorbed on less reactive CNS and thus protected from complete reduction (hence forming -CH2OH), thereby providing a pathway towards selective reduction to ethanol. The possibility that a synergistic effect between Cu and CNS is responsible for the selectivity is surprising given the large size of the Cu nanoparticles, however there is no doubt that the Cu is necessary for this reaction as bare CNS do not produce the same products. Likewise, Cu nanoparticles nucleated on glassy carbon (or any other copper reported to date) do not produce the same products without CNS. While the Cu particles are relatively large, they are on the scale of the CNS which are around 50 nm in length and tend to be imbedded within the spikes. A more detailed understanding of the reaction pathway of such selective reduction of CO2 to ethanol warrants further study. Conclusion   We report an electrocatalyst which operates at room temperature and in water for the electroreduction of dissolved CO2 with high selectivity for ethanol. The overpotential (which might be lowered with the proper electrolyte, and by separating the hydrogen production to another catalyst) probably precludes economic viability for this catalyst, but the high selectivity for a 12-electron reaction suggests that nanostructured surfaces with multiple reactive sites in close proximity can yield novel reaction mechanisms. This suggests that the synergistic effect from interactions between Cu and CNS presents a novel strategy for designing highly selective electrocatalysts. While the entire reaction mechanism has not yet been elucidated, further details would be revealed from conversion of potential intermediates (e. g. CO, formic acid and acetaldehyde) in future work. Acknowledgements   This research was conducted at the Center for Nanophase Materials Sciences, which is a Department of Energy (DOE) Office of Science User Facility. Ancillary Supporting Information   1 1a Y. Hori, K. Kikuchi, A. Murata, S. Suzuki, Chem. Lett. 1986, 15, 897–898; CrossRef | Web of Science® Times Cited: 199 1b Y. Hori, A. Murata, R. Takahashi, J. Chem. Soc., Faraday Trans. 1 1989, 85, 2309–2326; CrossRef | CAS | Web of Science® Times Cited: 318 1c D. W. DeWulf, T. Jin, A. J. Bard, J. Electrochem. Soc. 1989, 136, 1686–1691. CrossRef | CAS | Web of Science® Times Cited: 129 2 2a Y. Tomita, S. Teruya, O. Koga, Y. Hori, J. Electrochem. Soc. 2000, 147, 4164–4167; CrossRef | CAS | Web of Science® Times Cited: 43 2b M. Z. Ertem, S. J. Konezny, C. M. Araujo, V. S. Batista, J. Phys. Chem. Lett. 2013, 4, 745–748; CrossRef | PubMed | CAS | Web of Science® Times Cited: 56 2c R. Kortlever, I. Peters, S. Koper, M. T. Koper, ACS Catal. 2015. 3 3a C. Costentin, S. Drouet, M. Robert, J.-M. Savéant, Science 2012, 338, 90–94; CrossRef | PubMed | CAS | Web of Science® Times Cited: 166 | ADS 3b C. Costentin, G. Passard, M. Robert, J.-M. Savéant, J. Am. Chem. Soc. 2014, 136, 11821–11829. CrossRef | PubMed | CAS | Web of Science® Times Cited: 31 4 Y. Wang, J. Zhou, W. X. Lv, H. L. Fang, W. Wang, Appl. Surf. Sci. 2016, 362, 394–398. CrossRef | CAS | Web of Science® | ADS 5 5a C. E. Tornow, M. R. Thorson, S. Ma, A. A. Gewirth, P. J. Kenis, J. Am. Chem. Soc. 2012, 134, 19520–19523; CrossRef | PubMed | CAS | Web of Science® Times Cited: 47 5b J. Rosen, G. S. Hutchings, Q. Lu, S. Rivera, Y. Zhou, D. G. Vlachos, F. Jiao, ACS Catal. 2015, 5, 4293–4299; CrossRef | CAS | Web of Science® Times Cited: 9 5c T. Hatsukade, K. P. Kuhl, E. R. Cave, D. N. Abram, T. F. Jaramillo, Phys. Chem. Chem. Phys. 2014, 16, 13814–13819. CrossRef | PubMed | CAS | Web of Science® Times Cited: 46 | ADS 6 6a W. Zhu, Y.-J. Zhang, H. Zhang, H. Lv, Q. Li, R. Michalsky, A. A. Peterson, S. Sun, J. Am. Chem. Soc. 2014, 136, 16132–16135; CrossRef | PubMed | CAS | Web of Science® Times Cited: 52 6b D. Kim, J. Resasco, Y. Yu, A. M. Asiri, P. Yang, Nat. Commun 2014, 5; 6c W. Zhu, R. Michalsky, O. n. Metin, H. Lv, S. Guo, C. J. Wright, X. Sun, A. A. Peterson, S. Sun, J. Am. Chem. Soc. 2013, 135, 16833–16836; CrossRef | PubMed | CAS | Web of Science® Times Cited: 115 6d G. B. Stevens, T. Reda, B. Raguse, J. Electroanal. Chem. 2002, 526, 125–133. CrossRef | CAS | Web of Science® Times Cited: 41 7 S. Ye, R. Wang, M. Z. Wu, Y. P. Yuan, Appl. Surf. Sci. 2015, 358, 15–27. CrossRef | CAS | Web of Science® Times Cited: 21 | ADS 8 8a Y. Hori, I. Takahashi, O. Koga, N. Hoshi, J. Phys. Chem. B 2002, 106, 15–17; CrossRef | CAS | Web of Science® Times Cited: 84 8b M. Gattrell, N. Gupta, A. Co, J. Electroanal. Chem. 2006, 594, 1–19. CrossRef | CAS | Web of Science® Times Cited: 330 9 K. P. Kuhl, E. R. Cave, D. N. Abram, T. F. Jaramillo, Energy Environ. Sci. 2012, 5, 7050–7059. CrossRef | CAS | Web of Science® Times Cited: 250 10 10a A. M. Appel, J. E. Bercaw, A. B. Bocarsly, H. Dobbek, D. L. DuBois, M. Dupuis, J. G. Ferry, E. Fujita, R. Hille, P. J. Kenis, Chem. Rev. 2013, 113, 6621–6658; CrossRef | PubMed | CAS | Web of Science® Times Cited: 274 10b S. Ma, P. J. Kenis, Curr. Opin. Chem. Eng. 2013, 2, 191–199. CrossRef | Web of Science® Times Cited: 69 11 K. Manthiram, B. J. Beberwyck, A. P. Alivisatos, J. Am. Chem. Soc. 2014, 136, 13319–13325. CrossRef | PubMed | CAS | Web of Science® Times Cited: 42 12 D.-H. Lim, J. H. Jo, D. Y. Shin, J. Wilcox, H. C. Ham, S. W. Nam, Nanoscale 2014, 6, 5087–5092. CrossRef | PubMed | CAS | Web of Science® Times Cited: 17 | ADS 13 13a Y. Hori, A. Murata, R. Takahashi, S. Suzuki, J. Am. Chem. Soc. 1987, 109, 5022–5023; CrossRef | CAS | Web of Science® Times Cited: 113 13b J. Kim, D. Summers, K. Frese, J. Electroanal. Chem. Interf. Electrochem. 1988, 245, 223–244. CrossRef | CAS | Web of Science® Times Cited: 94 14 C. W. Li, J. Ciston, M. W. Kanan, Nature 2014, 508, 504–507. CrossRef | PubMed | CAS | Web of Science® Times Cited: 122 | ADS 15 L. B. Sheridan, D. K. Hensley, N. V. Lavrik, S. C. Smith, V. Schwartz, C. Liang, Z. Wu, H. M. Meyer, A. J. Rondinone, J. Electrochem. Soc. 2014, 161, H558-H563. CrossRef | CAS | Web of Science® 16 M. Pourbaix, J. Electroanal. Chem. 1967, 13, 471-&. 17 A. A. Peterson, F. Abild-Pedersen, F. Studt, J. Rossmeisl, J. K. Norskov, Energy Environ. Sci. 2010, 3, 1311–1315. CrossRef | CAS | Web of Science® Times Cited: 283 18 M. L. Perry, J. Newman, E. J. Cairns, J. Electrochem. Soc. 1998, 145, 5–15. CrossRef | CAS | Web of Science® Times Cited: 222 19 J. H. Montoya, A. A. Peterson, J. K. Norskov, ChemCatChem 2013, 5, 737–742. Wiley Online Library | CAS | Web of Science® Times Cited: 47 20 R. Kas, R. Kortlever, A. Milbrat, M. T. M. Koper, G. Mul, J. Baltrusaitis, Phys. Chem. Chem. Phys. 2014, 16, 12194–12201. CrossRef | PubMed | CAS | Web of Science® Times Cited: 47 | ADS 21 F. Calle-Vallejo, M. T. M. Koper, Angew. Chem. Int. Edit. 2013, 52, 7282–7285. Wiley Online Library | CAS | Web of Science® Times Cited: 51 22 22a W. Tang, A. A. Peterson, A. S. Varela, Z. P. Jovanov, L. Bech, W. J. Durand, S. Dahl, J. K. Norskov, I. Chorkendorff, Phys. Chem. Chem. Phys. 2012, 14, 76–81; CrossRef | PubMed | CAS | Web of Science® Times Cited: 108 | ADS 22b I. Takahashi, O. Koga, N. Hoshi, Y. Hori, J. Electroanal. Chem. 2002, 533, 135–143. CrossRef | CAS | Web of Science® Times Cited: 44 23 T. Kondo, Y. Honma, J. Oh, T. Machida, J. Nakamura, Phys. Rev. B 2010, 82, 153414. CrossRef | CAS | Web of Science® Times Cited: 22 | ADS 24 24a K. Gong, F. Du, Z. Xia, M. Durstock, L. Dai, Science 2009, 323, 760–764; CrossRef | PubMed | CAS | Web of Science® Times Cited: 2287 | ADS 24b L. Zhang, J. Niu, L. Dai, Z. Xia, Langmuir 2012, 28, 7542–7550; CrossRef | PubMed | CAS | Web of Science® Times Cited: 93 24c L. Zhang, Z. Xia, J. Phys. Chem. C 2011, 115, 11170–11176. CrossRef | CAS | Web of Science® Times Cited: 403 25 25a Z. W. Tang, H. Chen, X. W. Chen, L. M. Wu, X. B. Yu, J. Am. Chem. Soc. 2012, 134, 5464–5467; CrossRef | PubMed | CAS | Web of Science® Times Cited: 53 25b V. Schwartz, W. J. Fu, Y. T. Tsai, H. M. Meyer, A. J. Rondinone, J. H. Chen, Z. L. Wu, S. H. Overbury, C. D. Liang, ChemSusChem 2013, 6, 840–846; Wiley Online Library | PubMed | CAS | Web of Science® Times Cited: 18 25c G. K. P. Dathar, Y. T. Tsai, K. Gierszal, Y. Xu, C. D. Liang, A. J. Rondinone, S. H. Overbury, V. Schwartz, ChemSusChem 2014, 7, 483–491; Wiley Online Library | PubMed | CAS | Web of Science® Times Cited: 23 25d X. J. Cui, Y. H. Li, S. Bachmann, M. Scalone, A. E. Surkus, K. Junge, C. Topf, M. Beller, J. Am. Chem. Soc. 2015, 137, 10652–10658. PubMed | Web of Science® Times Cited: 12 26 26a L. B. da Silva, S. B. Fagan, R. Mota, Nano Lett. 2004, 4, 65–67; CrossRef | CAS | Web of Science® Times Cited: 65 | ADS 26b V. Tozzini, V. Pellegrini, J. Phys. Chem. C 2011, 115, 25523–25528. CrossRef | CAS | Web of Science® Times Cited: 58

Ethanol, Fruit Ripening, and the Historical Origins of Human Alcoholism in Primate Frugivory

ROBERT DUDLEY Department of Integrative Biology, University of California, Berkeley, Berkeley, California 94720 and Smithsonian Tropical Research Institute, P.O. Box 2072, Balboa, Republic of Panama

SYNOPSIS Ethanol is a naturally occurring substance resulting from the fermentation by yeast of fruit sugars. The association between yeasts and angiosperms dates to the Cretaceous, and dietary exposure of diverse frugivorous taxa to ethanol is similarly ancient. Ethanol plumes can potentially be used to localize ripe fruit, and consumption of low-concentration ethanol within fruit may act as a feeding stimulant. Ripe and over-ripe fruits of the Neotropical palm Astrocaryum standleyanum contained ethanol within the pulp at concentrations averaging 0.9% and 4.5%, respectively. Fruit ripening was associated with significant changes in color, puncture resistance, sugar, and ethanol content. Natural consumption rates of ethanol via frugivory and associated blood levels are not known for any animal taxon. However, behavioral responses to ethanol may have been the target of natural selection for all frugivorous species, including many primates and the hominoid lineages ancestral to modern humans. Pre-existing sensory biases associating this ancient psychoactive compound with nutritional reward might accordingly underlie contemporary patterns of al- cohol consumption and abuse

INTRODUCTION The widespread occurrence of fermentative yeasts in ripening and ripe fruits indicates potential co-option of associated ethanol for use as a behavioral cue by vertebrate frugivores (Dudley, 2000, 2002). In particular, ethanol plumes might serve in localization of these transient nutritional resources, whereas ethanol consumed during the course of frugivory could act as an appetitive stimulant. Because humans are ances-trally derived from frugivorous primates, preference for and excessive consumption of alcohol by modern humans might accordingly result from pre-existing sensory biases associating ethanol with nutritional reward. Little is known, however, about either the natural occurrence of ethanol within fruits or the behavioral responses of frugivorous animals to such cues. As agents of both microbial decay and fermentative activity, yeasts are widespread both on and inside fruits (see Last and Price, 1969; Cipollini and Stiles,1992, 1993b; Spencer and Spencer, 1997). Anaerobic fermentation of sugars by yeasts yields ethanol. This metabolic pathway emerged in yeasts concomitantly with the shift by angiosperms from small wind-dispersed seeds to larger and more fleshy vertebrate-dispersed fruits during the late Cretaceous into the Paleocene (see Erikssonet al., 2000; Benneret al., 2002). Ethanol expression by fermentative yeasts appears to have specifically evolved to inhibit activity of bacterial competitors within ripe fruit (Ingram and Buttke, 1984), and ethanol plumes emanating from ripe fruit might thus have provided useful sensory information, both diurnally and nocturnally, from the very inception of mammalian frugivory. Given this historical presence of yeasts that consume sugars, plants correspondingly express a diversity of antifungal compounds within developing and ripe fruit to impede decomposition (see Janzen, 1977; Borowicz, 1988b; Cipollini and Stiles, 1992; Cipollini and Levey, 1997ab,c).Because microbial decay reduces the likelihood of vertebrate dispersal (Herrera, 1982; Borowicz, 1988a; Cipollini and Stiles, 1993a), the evolutionary pressures for effective antifungal measures that prevent spoilage are substantial.

The phenomenon of ripening, however, involves relaxation of defenses against premature consumptionboth by potential dispersers and by microbial pathogens (Thompson and Willson, 1979; Herrera, 1982; Janzen, 1983). Fruit ripening involves a coordinated series of changes in color, texture, volatile expression, and the conversion of starch to sugars (Brady, 1987; Tucker, 1993). In aggregate, these changes indicate suitability for consumption and dispersal by a vertebrate frugivore. As a consequence, fully ripe fruits are susceptible to microbial decay, an outcome that can interfere with the plant’s evolutionary goal of consumption and dispersal by vertebrates (see Janzen, 1977; Borowicz, 1988a; Cipollini and Stiles, 1993a). Microbes, invertebrate fruit consumers (especially insect larvae), and vertebrate dispersers can thus be viewed as competitors for access to a rich but transient nutritional substrate. In spite of the possible significance of ethanol for frugivore behavior and ecology (Levey and Martı ́nezdel Rio, 2001), knowledge of fermentation for non-domesticated fruits in natural ecosystems is confined to only several examples. Eriksson and Nummi (1982; see also Forsander, 1978) determined fairly low ethanol contents (0.05–0.3% w/w) for rowan berries, rosehips, and hawthorn fruits in autumn and winter conditions in Finland. Such temperate-zone fruits are unlikely, for reasons of low ambient temperatures alone, to be characterized by particularly high ethanol concentrations. Fermentation of fruit crops is instead more pronounced in warm and humid environments that promote both yeast growth and rapid decomposition. Dudley (2002) presented ethanol data for three fruiting taxa in a Neotropical forest, and found that pulp of ripe and very ripe palm fruits (Astrocaryum standleyanum) contained ethanol at concentrations of about 0.5% and 0.6%, respectively. The present study extends existing data on A. standleyanum to include parallel measures of texture and color for quantitative assessment of fruit ripeness, and also compares larger sample sizes of fruits taken both directly from the infructescence and from the ground where a greater range of decompositional conditions is available. A. standleyanum is a common palm species in lowland Panamanian rainforest, and bears large crops of orange fruits that are consumed by red-tailed squirrels, spiny rats, kinkajou, Central American agoutis, collared peccaries, and white-faced capuchin monkeys (Hladik and Hladik, 1969; Croat, 1978; Smythe, 1978, 1989; Hoch and Adler, 1997; Kays, 1999). This palm therefore represents an appropriate target species for evaluating the potential role of ethanol in the sensory and nutritional ecology of mammalian frugivores.Moreover, palm fruits have been proposed (together with figs) to be keystone resources for Neotropical vertebrate frugivores (Terborgh, 1980). Broad biogeographic and taxonomic screening of angiosperm fruits for ethanol content is beyond the scope of the present paper, but demonstration of ethanol-associated ripeness cues in palms may nonetheless be of general relevance for those frugivorous primates predominantly associated with tropical rain forest.

MATERIALS AND METHODS Field collections and laboratory measurements were carried out on Barro Colorado Island (BCI), Republic of Panama, in May and June 2002 during the rainy season (ripe and over-ripe fruits), and in December 2002 (unripe fruits). Fruits were knocked down from infructescences using either a slingshot or thrown rocks (fruits thus obtained are hereafter referred to as hanging fruits), or were collected at the base of fruiting trees ( i.e., fallen fruits). Following transport of fruits within a closed plastic bag to the laboratory on BCI, measurements were made either immediately or within three hours of collection. In the latter case, individually bagged fruits were kept within a cold room at 108C. Fruits were visually categorized as being either ripe or over-ripe; obviously rotting fruits were not used for measurements. Over-ripe fruits were never obtained from infructescences, and thus were only collected from the ground. Spatial heterogeneity of ripeness within an individual fruit was sometimes pronounced, particularly for fallen fruits. In these cases, all measurements were confined to a visually homo geneous region of the fruit. Wet masses of the exocarp (skin), mesocarp (pulp), and endocarp (seed) were also measured on twenty ripe fallen fruits. Color reflectance spectra, puncture resistance, sugar content of the pulp, and ethanol concentration of the pulp were measured on all collected fruits. Reflectance spectra of the skins were determined using a portable spectroradiometer (Colortron II; Lightsource)