# Last edited on 2017-02-20 10:03:25 by jstolfi ---------------------------------------------------------------------- 2016-??-?? Two pilot experiments were held in 2016, before September, No notes were taken; the following is from memory. In both experiments, a quantity of purple yam was peeled, cut into slices, and boiled in ~1 liter of water until tender. The dark purple liquid was separated with a colander. The boiled and drained yam, still purple, was mashed with a potato ricer, and used in cooking. It was still somewhat sweet. In the first experiment, the liquid was concentrated by boiling, with no sign of loss of color. It obviously had dissolved sugars and/or hydrolyzed starch. Some NaCl was added to prevent fermantation. Ethanol (14 g) was also added. In the second experiment, the cooking liquid (about ~500 ml) was fermented in an attempt to remove the carbohydrates and hopefully the glucose units in the anthocyanin. The liquid was inoculated with dried baker's yeast (Fleischmann's) and placed in a large mason jar (1000 ml). Over a week or two it fermented with abundant foaming, changing color to dark red and abundant solids. The solid were filtered and waashed with a small amount of ethanol to give a dark purple solution. The anthocyanin extracts from these pilot experiments (with water, ethanol, NaCL, and carbohydrates) were merged into ~200(??) ml of a dark purple liquid (0). ---------------------------------------------------------------------- 2016-11-19 A more systematic experiment was performed over several weeks in November 2016. Instead of cooking the yams, it was decided to do cold or warm extraction of the ground tubers. In this way it was hoped to reduce the amount of carbohydrates in the extraction liquid. Also it was decided to try extraction with caninha (Brazilian rhum from sugarcane), which is 39% ethanol by volume according to the label. That because published methods used methanol, so it was assumed that the anthocyanin woudl be more soluble in ethanol than in water. It was also hoped that the ethanol would inhibit extraction of any sugar and other water-soluble substances from the yams. A batch of about 750 g of purple yams were bought at a local supermarket. The quality was a little dubious, may have sit on the shelf for unknown time. The yeams were peeled by hand and diced into 1--2 cm wide pieces, stored under water to avoid oxidation. Some of the pigment was leached and imparted a brownish-purple tinge to the water. The diced yams were divided into 3 batches (1)--(3) of ~230 g each. Peels (not weighted, but definitely less than 230 g) were saved too, as batch (4). The batches were separately blended with different liquid mediums in a kitchen blender at medium-high speed for 5--10 minutes, yielding thick but free-flowing slurries. Batch (1) was blended with ~300 ml of water from a home filter-cooler, at 25 C. Batch (2) was blended with ~300 ml of 39% caninha, also at 25 C. Batch (3) was blended with a mix of ~150 ml of water from the filter-cooler and ~150 ml of caninha. Before blending, the diced yams and the liquid were heated in a microwave owen for ~120 seconds, in 10-20 s increments, until it became uncomfortabl hot to the touch. The temperature was 48 C at the start of the blending and 47 C at the end. Batch (4) was blended with ~150 ml of water at 25 C The slurry from each batch was filtered through dense cotton cloth (?? threads/mm), wringed by hand so as to squeeze as much liquid as possible. The residue left in the cloth, a coarse flour-like moist powder, was not washed. That resulted in solid residues (1r)--(4r) and filtrates (1s)--(4s). The moist residues (1r)--(3r), weighting ~90 g each, were purplish gray. The residue (4r), from the peels, was similar, but smaller, darker, and more brownish. We concluded that water was quite sufficient to extract practically all the anthocyanin. The ethanol in batches (2) and (3) did not make any visible difference in that regard. However it may have inhibited extraction of sugars and other yam compounds. It also inhibited fermentation of the extract, as we learned in the folowing weeks. Residue (4r) was discarded, and residues (1r)--(3r) were combined and partly dried, yielding a free-flowing coarse flour (1+2+3r), still with a small purplish tinge. The residual smell from the caninha masked any smell that the flour could have. The flour could have been used for cooking, but it was left in the fridge for two weeks and was discarded for fear of it having spoiled in that time. The filtrates (1s)--(4s) looked very much alike: purplish brown at first, but they quickly deposited a white ppt, and the liquid above became purple-black. The filtrates (1s)-(3s) measured ~450 ml each, that is, ~150 ml more than the added liquid. We conclude that the yams had at least 65% water by weight, even though they felt only slightly moist when cut. The filtrates (1s)--(4s) were placed in eflasks at room temperature, sealed with doule layer of magipak (PVC food wrap). ---------------------------------------------------------------------- 2016-11-20 After ~30 hours, the filtrates (1s)--(4s) had developed whitish ppts at the bottom, about 2-3 mm thick. The liquids (1s) and (4s), which had been extracted with water only, had became again brownish purple, while (2s) and (3s) were still purple-black. All four filtrates were decanted to give uberliqs (1t)--(4t) with approx 450, 400, 400, and 250 ml, placed in 600, 600, 600, and 250 ml beakers with magipak and rubber bands. The ppts (1u)--(4u) were placed in 150 ml beakers with some of the respective uberliqs. Ppt (1u) was somewhat gooey and with a few embedded bubbles. Ppt (4u) was also gooey but with no bubbles. Ppts (2u) and (3u) were a compact friable mass, as expected from starch in cold water. It seems that, even in only ~30 hours, the water extracts (1s) and (4s) had already begun to ferment; whereas fermentation was inhibited in (2s) and (3s), extracted with caninha, or caninha and water. Ppts (1u) and (4u) were washed with water, while (2u) and (3u) were washed with 46% ethanol. All four were stored in 150 ml beakers under magipak seals. ---------------------------------------------------------------------- 2017-12-10 The sealed beakers with decanted liquids (1t)--(4t) and the once-washed starchy ppts (1u)--(4u) were left on the bench for 20 days due to travel. Liquid (1t), from water extraction, had developed a film on top: ~0.5 mm thick, translucent white with grains ~1 mm across. Under that film was ~120 ml of a colorless liquid, clear but a bit cloudy; and then ~300 ml of a suspension of a reddish brown ppt. Liquid (2t), from caninha extraction, was still ~370 ml of a purple-black liquid, with ~30 ml of witish sediment at the bottom. Liquid (3t), from warm water and caninha extraction, had ~160 ml of a clear light red liquid over ~220 ml of a suspension of a redish-brown ppt, with maybe ~20 ml of a denser suspension or ppt at the bottom. Liquid (4t), from water extraction of peels, looked pretty much like (1t), except that the top clear layer was ligh red instead of colorless. It had only a faint odor of overripe fruit. The thin film on top of liquid (4t) was removed with a spoon to a watchglass (4b), and the rest was separated by decantation into ~130 ml of clear light red uberliq (4a), and the opaque suspension (4c). The latter was filterd in a coffe filter and washed once with ~10 ml of water. The residue left on the coffe filter (4c1) was a reddish brown paste. The filtrate (4c2) was clear light red with a bit of brownish red sefdiment. The pH was ~3.5 according to a PHOX PH014 indicator strip. It was merged with the decanted uberliq (4a) and filtered with 28 µm pore paper. The filtrate (4c3) was ~210 ml of a slightly red clear liquid. A small amount of reddish-brown paste (4c4) was left on the filter. We conjecture that liquids (1t) and (4t) fermented significantly, which created the white film on top and the reddish brown suspended solids. The fermentation acidified the liquid. Some of the anthocyanin went into solution, while some (most of it) remained bound in the suspended solids. Liquid (3t) fermented partially, enough to cause a solid ppt and a similar partitioning of the anthocyanin, but without the white film on top, and with a deeper color in the uberliq. Only liquid (2t) remained unfermented, with all the anthocyanin dissolved. ---------------------------------------------------------------------- 2016-12-11 By the next day, the liquid on the watchglass surrounding the film (4b) had turned grayish. The (4b) sample was then discarded. The reddish brown pastes (4c1) and (4c4) retained in the filters were combined and washed off the paper with ~170 ml of 46% ethanol, yielding a reddish brown suspension (4d). Small clots of the paste remained undispersed even after vigorous stirring. We conclude that the solids created by fermentation contain substances that are solube in water but not so much in ethanol. The (4d) ethanol suspension was placed in a ?? ml baker sealed with magipak and rubber band. The starchy ppts (2u) and (3u), from extractions with caninha and diluted caninha, respectively, and which had been washed with 46% ethanol, looked similar and did not show any signs of fermentation, even after sitting on the bench for 2 weeks. The remaining uberliqs in both (2u) and (3u) were decanted and merged, yielding ~20 ml of dark purplish brown liquid (2+3f), which was merged into the extract (0) from the exploratory experiments. Apparently the anthocyanin in that older extract is soluble in ethanol, although there is sign of a lighter ppt forming. The starchy ppts remaining in (2u) and (3u) were combined too and resuspended in ~100 ml of 46% ethanol. The resulting suspension (2+3e) immediately deposited an off-white ppt again, with a clear light red uberliq. The ppt was washed three more times with 46%, 46%, and 99% ethanol. The uberliqs were clear and reddish-purple but increasingly lighter, the last being colorless. The uberliqs from the washings were discarded. The starchy ppts (1u) and (4u), from water extractions, were discarded with their remaining uberliqs, since the fermantation probably degraded some of the starch. As noted above, liquid (3t), from warm diluted caninha extraction, had developed a reddish clear uberliq and a reddish-brown suspension. Attempts to separate if by filtration failed because the suspension was too fine and went through the coffe paper, while rapidly clogging normal filter paper. In an attempt to coagulate it, first 5 g of diatomite (washed with 46% ethanol), and later a few ml of CaCl2 solution (from a spent Inspira drying pot) were added to (3t). It is not clear whether they helped, but eventually filtration with 28 µm and 12 µm paper, after several hours, yielded finally a reddish brown paste (3c1) and ~300 ml of a clear light red filtrate (3c2) ---------------------------------------------------------------------- 2016-12-12 The reddish brown paste (3c1), including the diatomite, was washed off the filter papers and resuspended in ~200 ml of 96--99% ethanol, yielding ~230 ml of purplish brown and opaque liquid (3d), similar to (4d). ---------------------------------------------------------------------- 2016-12-13 A couple ml of the filtrate (3c2), clear light red, was treated in a test tube (*3) with a slight excess of NaHCO3. A grey ppt was formed and quite a bit of CO2 bubbled out. The ppt must have been mostly CaCO3 formed from the CaCl2. The uberliq too was grayish, indicating that its former reddish hue was due to anthocyanin in some form. The uberliq was decanted to another test tube (*2) and reacidified with citric acid. After some CO2 bubbling the clear reddish color was restored, but much lighter than the original color of (3c2). The ppt left in the test tube (*3) was washed with water but remained gray. Treatment with citric acid released more CO2 but also restored the clear light red color of (3c2), indicating that most (but not all) anthocyanin had precipitated out with the CaCO3 when treated with the NaHCO3. Meanwhile, the suspension (3d) of the brown paste (3c1) in ethanol started to separate again into a clear red uberliq, darker than (3c2) but lighter than the uberliq of (4d); and a reddish brown suspension. The lighter color of the (4d) uberliq may be due only to its lower concentration and its higher water content. ---------------------------------------------------------------------- 2016-12-15 The ethanol from the 5th washin of the starchy ppt (2+3e) was discarded, and the precipitate was placed between filter papers and paper towels to dry. Its non-Newtonian character, typical os starch suspensions, was very evident. ---------------------------------------------------------------------- 2016-12-16 The starchy ppt (2+3e) was dry and solid. It was broken with a spatula and sifted through a ?? mm strainer. Result was 27g of a slightly tan white powder, not airborne. The (2t) liquid (350 ml) was decanted and filtered with grade 201 paper. Filtration slowed down to 1 drop every 5 sec. With a 28µm paper it was fastr at first but slowed down to 1-2 d/s. The filtrate (2c2) was dark purplish brown abd rather opaque, suggesting a very fine suspension rather than a solution. ---------------------------------------------------------------------- 2016-12-17 The filtering of (2t), (2c2), and (2c1) was not finished after 24 h even using 2 funnels in parallel. There was practically no ppt in (2t) but the filter didi retain some. Had to replace the filter paper by 28 µm twice. Took three samples of (2t) in test tubes ((1)), ((2)), ((3)), about 20mm each. In test tubes ((1)) addedand ((2)) added equal amounts of a CaCl2 solution prepared previously. The purple suspension curdled and floated to the top. Prepared some Na2CO3 from NaHCO3 by heating in a pan on a gas kitchen stove. Prepared a solution with ~1 g Na2CO3 in ~10 g H2O. Added ~20mm to tubes ((2)) and ((3)). Tube ((3)) became dark green opaque. Tube ((2)) developed dense ppt (presumablu CaCO3) and became dark green too. Later reverted to brown. Some bubbles developed. Considering the failure of filtering (2t), the filtrate (2c2) and the little solids retained in the filter paper were recombined as (2t) (350 ml). perhaps 20-30 ml of ethanol 46% was used to wash down the filter paper and erlens. Considering the result of test tube ((1)), we added 30 ml of CaCl2 solution to (2t). The solution turned a lighter shade of purple brown. The solids clotted and started to precipitate, making a 1 cm fluffy layer at bottom. Now tackling th emost disgusting batch, (1t). Smell of fermented dough but disagreeable. Decanted ~120 ml of peach-colored milky uberliq (1c4). Left ~300 ml of wine-colored fluffy suspension (1c0) with some white streaks of mold on top. Attempt to filter (1c0) with coffee filter failed. Added ~50 ml of wash water. Added 30 ml of CaCl2 solution, twice The suspendd solids started to separate slowly. Filtered the contents of test tube ((2)) with 20 µm paper. Washed ppt ((2c3)) with water and ethanol 46%. Prepared a citric acid solution with 10 g H3Citr and 10 g H2O. The filtrate ((2c2)) of test tube ((2)) (~30 ml) was treated with a few drops of H3Citr solution. The color changed from clear greenish to clear reddish, both very faint. The ppt ((2c3)) of test tube ((2)) was dried yielding a green-black powder. Placed back in cleaned test tube ((2)) and treated with ~1ml of citric acid solution. The ppt dissolved with much bubbling to give a slightly cloudy light red solution. Test tube ((3)), the sample of (2t) with Na2CO3, was a fine suspension, dark green. It was treated with ~1 ml of H3Citr solution. With much bubbling, the solution turned milky light red, more opaque with ((2)). Encouraged by test ((2)), 5 g of Na2CO3 were dissolved in 15 ml of H2O and added to (1c0). Color became darker but still no significant CaCO3 ppt formed. The dose was repeated. This time the suspension (1c0) became purple-black and with a thick colloidal ppt of CaCO3 (~420 ml). ---------------------------------------------------------------------- 2016-12-18 24 hours later (1c0) was still a fluffy suspension with only 2mm of clear liquid on top. Tried filtering with coffee filter but too slow. Filtrate (1c2) is clear with slight grayish green tinge. Was interrupted leaving some ppt (1c1) on the filter. Boiled the remaining (1c0) to see whether it would filter better. Developed ~100 ml of thick foam (CO2?) on top, would not go away with stirring. Foaming subsided and foam collapsed after a few minutes at 90-95C. Precipitate now dark moss green with very mobile suspension. Chopped the coffe filter with ~1 g of ppt (1c1) and added to (1c0). A bit of foaming resumed, left a few more minutes at 90 C. Also filtered (2t) with W201 paper. The uberliq filters quickly yielding brownish purple clear filtrate (2c2) and solids (2c1). Saved the ppt (2c1) from (2t), with filter paper, in a beaker (2d). Started filtering (1c0) with W201 into ppt (1c1) and filtrate (1c2). The latter is a clear liquid with a light gray-green tinge. In each of test tubes ((1)), ((2)), ((3)) placed 2 ml of (2c2). added: ((1)) + 1 ml CaCl2 solution + 1 ml Na2CO3 ((2)) + 1 ml CaCl2 solution ((3)) + 1 ml Na2CO3 solution Tube ((2)) showed no reaction. Was discarded. Tubes ((1)) and ((3)) formed a dense ppt with a green uberliq. The ppt appeared gray green in ((1)), light green in ((3)); but may have been white, colored by the uberliq. Decanted uberliq of ((3)), light gren clear, into ((2)). Washed ppt of ((3)) with water. Resulting ppt is grey-greenish. Conclusion: precipitation of CaCO3 takes some anthocyanin out of solution, but not all. Useless as a separation technique? ---------------------------------------------------------------------- 2016-12-22 The (1c0) mess was almost impossible to filter. It clogged paper so thoroughly that only ~10 ml went through in 48 hours. ---------------------------------------------------------------------- 2016-12-23 Placed the greenish-black solids decanted/filtered from (1c0) into a beaker and mixed with ~200 ml ethanol 92%, yielding suspension (1d). The uberliq of (1d) was initially clear with slight brownish tinge. The solids fro (2c1) were semi-dried over filter paper resulting in a purple-brown paste. They were mixed with ~200 ml of ethanol 46% yielding suspension (2d). It was hard to resuspend, as many small clots remained even after vigorous stirring. The filtrate (1c2) from (1c0) was greenish black, opaque, like a fine suspension. Added ~5 ml of CaCl2 solution to see whether it would clot. No immediate change. ---------------------------------------------------------------------- 2017-12-25 Filtering all extracts of the resuspended ppts Recall that the ppts (2d) and (4d) were resuspended in ethanol 46%, while (3d) was resuspended in ethanol 92%. Filtered them through W201 paper, yielding filtrates (2e), (3e), (4e) and ppts (2f), (3f), (4f). Filtrate (2e) was very clear reddish. ---------------------------------------------------------------------- 2016-12-27 Refiltered (2c2) with 7 um paper. Merged clear red filtrate with (2e) yielding (2g). Dried the dark purple brown ppt (2f), weighting 4 g dry. Started filtering (3d) with W201 paper into a clear red uberliq (3e) and a dark purple brown ppt (3f). The latter weighted 8g dry. Meanwhile (4c2) became more turbid, and orangish rather than light red. Refiltered (3c2) with 7 um paper. Merged the clear red filtrate with (3e) yielding (3g). Liquid (1c1) was foul smelling and did not seem to contain much anthocyanin. It was discarded. Refiltered (1c2) with 7 um paper, yielding ~350 ml of green black filtrate. ?? (1e). Filtered (1d) with W201 paper, yielding ppt (1f) and clear very light greenish filtrate (1g). Merged green black (1c2) into (1g), which became dark greenish-brown, opaque. Added a few drops of Na2CO3 solution to (4c2). Turned from light orange to cloudy dark green, showing that it contained some anthocyanin. ---------------------------------------------------------------------- 2016-12-28 Filtered (4d) (the ppt from water extract resuspended in ethanol 46%) into clear wine-colored filtrate (4g) and ppt (4f). Combined (4g) with (4c2). The mix is wine-colored, slightly cloudy. ---------------------------------------------------------------------- 2016-12-30 The ppt (1f) from the filtering of (1d), dried, was 12 g, gray-green. The refiltering of (1c2)+(1g) (??) was not complete even after 48h with 2 funnels. Gave up. The dried ppt (4f) was 3 g, dark purplish brown. Essays with acid dissolution of ppts Prepared a solution of 10g of citric acid in 50 g of water. Placed small amounts of (1f), (2f), (3f), and (4f) in test tubes with ~1 ml of solution creating (1f*), (2f*), (3f*), (4f*). (1f*) started bubbling (2f*) no immediate change (3f*) uberliq became clear very light red (4f*) no immediate change. ---------------------------------------------------------------------- 2016-12-31 After 24 h, all four had clear red uberliq. Added 17 g of citric acid to the solution, so it became about ~27 g of H3Cit in ~46 g water. The pH was ~1.5. Added ~1ml of the stronger solution to each tube, stirred with glass rod. (1f*) and (3f*) gave cloudy suspensions. (2f*) and (4f*) remained solid at the bottom, without desintegrating. ---------------------------------------------------------------------- 2017-01-30 Mon Another batch (failed) In 2016-12 had bought ~800 g of purple yams in supermarket. They were kept for ~1 month in the fridge. They were found to be half-rotten, unfit for experiment. Discarded. ---------------------------------------------------------------------- 2017-02-10 Batch 04 Purchased 6 yams at supermarket. Weights unpeeled (g): 170 176 141 171 241 172. Washed, lightly scrubbed with plastic scrubbing sponge. Peeled with potato peeler. 119 g of peels (wet). Weights of peeled yams (g): 151 157 128 159 224 160; total 981 g. Diced into 1-3 cm pieces. Divided into 4 lots of ~252 g each. Covered with majipak and put in fridge for a few minutes.