Boffis
International Hazard
Posts: 1836
Registered: 1-5-2011
Member Is Offline
Mood: No Mood
|
|
Preparation of furfural from cherry gum and other pentose sources
A flowering cherry tree in our garden has over the past few years begun to weep a clear gelatinous gum like material from its pruning wounds in
spring. The soft gelatinous material hardens and if left on the tree darkens in colour. I collected and dried about 60g of this material and began to
investigate its composition. Being insoluble in water once dried and yet insoluble too in both polar and non-polar organic solvents it is not a resin.
It transpires that it is a complex polysaccharide composed of pentose, hexose and sugar acid units 1). At first I tried to hydrolyse it with very
dilute acid into it constituent sugar units and then separate the sugars by their solubility in methanol etc. Unfortunately this proved very
difficult. I eventually ended up with a liquid that resembled maple syrup but it simply refused to crystallise. The methanol-insoluble solid that was
extremely deliquescent appears to be a starch-like organic material, presumably an incompletely hydrolysed intermediate. From 58g I got 45ml of syrup.
I kept the syrup for over a year without any appreciable deterioration, obvious change or crystallisation. Attempts to dry it out lead to
decomposition. Eventually I decided to try and convert it into furan derivatives by distilling with acid.
To my pleasant surprise this was quite successful so I applied the technique to OTC xylose and ribose with excellent results; recoveries consistently
>50% of theory.
The equipment used for the furfural distillation is basically the same as used by Magpie in his Science madness thread5). I used a 500ml round bottom
flask equipped with a Dean and Stark heavy-type entrainer of 10ml capacity with a discharge stopcock. This was fitted with a Liebig condenser and a
magnetic stir bar. This set-up alleviates the need for a separate steam generator and also the need to extract the aqueous phase with ether or similar
as the volume of aqueous phase left in the entrainer represents a loss of product too small to be of any consequence.
Hydrolysis of Cherry Gum
58.85g of crude, dry cherry tree gum was broken into small piece (<5mm) and placed in a beaker with 180ml of hot water and 20ml of 36% hydrochloric
acid added. The beaker was heated until almost boiling and then maintained at this temperature until all of the gelatinous material had dissolved
leaving only a little woody debris; about 4½ hours, topping up the water level occasionally. After this time 1.2g of decolourizing charcoal were
added and the solution left to cool overnight. The cold solution was vacuum filtered to give a light brown solution.
The solution was evaporated down to about 70ml on a steam bath and left to cool. No crystals formed even after chilling to 4° for 24 hours. So the
thin brown syrup was poured into 300ml of methanol to give a creamy suspension. This was heated almost to boiling and filtered hot using gentle
suction to prevent boiling. The filter cake was air dried at 40-45° on a watch glass but it was found to be extremely deliquescent. When heated on a
spatula to burned away leaving a little char. It was not investigated further but appears to be a product of partial hydrolysis.
The filtrate was distilled to recover most of the methanol and then evaporated on a steam bath to about 45ml and cooled. The brown syrup, resembling
maple syrup, was stored for several months but no crystallization occurred. The syrup did not obviously deteriorate or change either; its SG was about
1.3.
Furfural from Cherry Gum “Syrup”
As no crystalline product could be obtained from the final syrup I decided to try converting it into furfural using a combination of Magpie’s
procedure modified along the lines of Fulmer’s procedure2) using hydrochloric acid and salt. An attempt being made to optimise the concentration of
pentose, hydrochloric acid and salt. The cherry gum “syrup” was estimated to contain about 35g of sugars (or about 25g of pentoses) so I chose to
add 60g of salt (somewhat below the 40% optimum of Fulmer et al. but even this took a good while to dissolve and stirring was impossible until most of
it had), enough 30% hydrochloric acid to give a final concentration of 1M in the flask (27ml of 30% HCl). These ingredients were added to a 500ml
round bottom flask along with a stir bar and the flask placed into a magnetic stirrer heating mantle. A bottom draining “heavy entrainer” type
Dean and Stark trap was added to the flask and a Liebig condenser above. The flask was heated to boiling over about 20 minutes (Note 1) and the flask
was boiled steadily for 5 hours. Furfural began to appear in the condensate after about 30 minutes and appears to have ceased after 5 hours. 4.9ml or
about 5.5g of pale yellow furfural were collected. It is impossible to give a theoretical yield since the composition of the starting syrup was
unknown.
Note 1, there was a lot of foaming with the cherry gum syrup just as it boiled but it then subsided. No excessive foaming was observed with pure
pentose sugars.
 
Furfural from Xylose
The experimental procedure used above was repeated using 25.03g of commercial xylose, 60g of salt and 27ml of 30% hydrochloric acid diluted to 250ml.
Distillation was carried out as before until no further oily droplets of furfural were apparent in the condensate water, this took roughly 7 hours.
The yield was 7.8ml of pale yellow liquid in the Dean and Stark trap. This was run off into a pre-weighed conical flask to give 8.97g which, allowing
for some water in the product, represents about 55% of theoretical yield. I observed that while the distillation was in progress the amount of
furfural in the distillate seems to fall with increasing distillation rate. This is possibly due to the solubility of furfural in hot water causing
the product to be flush back over the weir into the distillation flask.
Furfural from D-Ribose
The previous experiment was repeated with 25.02g of food grade D-ribose in place of the xylose. The other components of the mix were the same. The
flask was heated to boiling as before but this time the heat input was reduced to the point where the system only just refluxed and the condensate
dripped from the condenser at no more than 1 drop per second. This resulted in a significantly faster accumulation of furfural and distillation had
practically ceased after 6 hours. 9.0ml or 10.2g were collected, about 63% of theory.
Discussion
Magpie’s idea of using a Dean and Stark type set up to reduce the amount of aqueous furfural solution to be extracted at the end works well and I
was able to significantly increase the yield. While three experiments hardly represents “optimisation” they do seem to confirm the a fairly dilute
solution gives an excellent yield, the presence of much salt seems to aid distillation perhaps by reducing the solubility of furfural in the reaction
medium and hence assisting in its steam distillation.
The yield of furfural from cherry gum may seem low (15% assuming that it is 100% pentose) but furfural was not the original target of the 1st
hydrolysis and some pentose may have been lost to the methanol insoluble material or degraded before the final hydrolysis. It would be interesting to
try cherry gum directly in the furfural distillation. This type of gum is available from most species of Prunus such as cherries, apricots, peaches
etc. most prolifically in Spring, and for anyone with access to such orchards these fruits trees may represent a good source of furfural. For those in
the southern US states mesquite gum is also rich in pentose bearing polysaccharide that have traditionally been a source of l-arabinose.
The yield of furfural from D-ribose was higher than that of xylose but this may be a result of the lower reflux rate used in the ribose distillation
and a further experiment will have to be run to check this. Either way, if it is available OTC, xylose probably represents the best source of furfural
for the amateur without access to high capacity equipment. At one time xylose was available in Europe from Ebay suppliers but this product seems to
have disappeared. D-Ribose is now readily available from many online suppliers, in the UK at least, and the price is now lower than when Magpie
originally started his thread. While more expensive than xylose it is reasonably priced for amateurs these days and the yield from a single batch in a
1L round-bottom flask (filled to the 600ml level) should be about 23g from 60g of ribose.
Further work
Apart from the further experiment with xylose mentioned above there are several other possible sources to look into. One possible route is from sodium
gluconate which is readily available and cheap. This can be converted into arabinose using a Fenton type oxidation3) and then distilled with
hydrochloric acid without isolation of the arabinose. On a similar vein is the oxidation of glucosamine with sodium hypochlorite to yield arabinose4),
again, it would not be necessary to isolate the sugar as the sodium chloride by-product would simply reduce the amount of salt that need be added.
I also intend to try scaling the reaction up to 1 and 3L capacity flasks. When pure pentoses are used the flasks may be ⅔ filled but due to short
term frothing of the mixture at the start of distillation it is best not to more than ½ fill the flask when using cherry gum.
Another variable to investigate is to increase the salt content to closer to 40% as suggested by Fulmer et al.2) to further improve the yield. This
causes issue with stirring until the salt has dissolved
1) Scerbuchin et al.; Polysaccharides from Cherry tree gum; Chem. Zvesti; vol 22; p248-256 [1968]. This paper shows that gum composition varies but is
typically about 65-70% pentoses, mainly arabinose.
2) Fulmer et al.; The Production of Furfural from Xylose Solution by means of hydrochloric acid-sodium chloride ; JPhysChem; v40 p133-141, [1936]
3) Hocket & Hudson; Improvements in the Preparation of d-Arabinose from Calcium Gluconate; JACS; v56 p1632-1633; [1934]
4) Matsushima Y.; Studies of amino-hexoses. A new method of preparing crystalline d Arabinose; BullCSJapan, v24, pp17-20, [1951]
5) http://www.sciencemadness.org/talk/viewthread.php?tid=15774&...
Most of the papers cited above are already link to this thread.
|
|
|
AvBaeyer
National Hazard
Posts: 644
Registered: 25-2-2014
Location: CA
Member Is Offline
Mood: No Mood
|
|
Boffis,
Really nice piece of work. I have been inspired to order a supply of ribose to make some furfural based on your work. I don't know what I will do with
it yet but it looks like a fun preparation to carry out.
Keep up the great work!
AvB
|
|
|
UC235
National Hazard
Posts: 565
Registered: 28-12-2014
Member Is Offline
Mood: No Mood
|
|
When I attempted to produce furfural from ribose I found that the conversion to furfural seemed to be a lot faster than the steam distillation.
Refluxing acid seemed to carbonize a lot of the product. I never got around to testing it but figured I could probably significantly increase the
yield by slowly adding a dilute solution of ribose to the refluxing acid/salt allowing it to steam distill as it is formed. Of course with a dean
stark, adding a lot of water is a problem unlike the original distillation so it would have to be batch additions of a very concentrated solution
instead I suppose.
|
|
|
Boffis
International Hazard
Posts: 1836
Registered: 1-5-2011
Member Is Offline
Mood: No Mood
|
|
I tried scaling up the preparation to 150g of xylose under the same conditions and obtained the same results in terms of time and yield (53-54%). I
then repeated the experiment with D-ribose and got only 26% (from 60% on a small scale) yield though distillation was rapid. This seems to confirm
U235's results that ribose reacts so rapidly that much is destroyed before it can be distilled out of the reaction mixture. So I thought I would give
U235's idea of a slow addition of ribose solution to the reaction mixture. Yesterday I tried a series of experiments. I set up the larger scale (3L)
flask for distillation as before but only placed into it the acid and part of the water. The ribose (150g) and salt (60g) were dissolved in 250ml of
hot water. The acid was brought to the boil and the ribose solution added through the condenser. However, there was an unforeseen problem, the
furfural that began to distill off but did not sink in the Dean and Stark trap and so was flushed back into the flask. The ribose and salt solution is
denser than the furfural. It was necessary to back-flush the trap with a syringe full of water and a bit of plastic tube. After the addition of only
1/3 of the ribose solution the experiment was abandoned due to this problem, even so the eventual yield was 15.2g of crude furfural which represents
about 46% on the basis of the ribose added. Another experiment was then set up with a 1l 3-neck flask and 2 dropping funnels. The flask size was
probably a little too small but this was a matter of availability.
The 1L flask 3-neck was charged with 70g of salt, 100ml of 30% hydrochloric acid and 150ml of water. A smaller dropping funnel was charged with 100ml
of saturated salt brine while the remaining, roughly 240ml of ribose and salt solution from the previous experiment was placed in the other. The acid
was brought to the boil and the ribose solution added dropwise slowly. Furfural began to pass over into the D&S trap after about 10 minutes and
collected rapidly. When about half of the ribose solution had been added the brine was slowly added at the same time. The addition was complete in 1
hr 40 mins and furfural distillation continued abundantly for a further 20-30 minutes and then tailed off. Distillation was terminated after 3 hours
when no further furfural can be seen in the condenser return. The yield was 31.8g or about 50% of theory, a big improvement over the simple
distillation with all the reactant added at the beginning in spite of the lower overall salt content of the reaction mixture.
It would appear therefor that there is definite merit in U235's idea.There are also published procedures using immiscible solvents such as toluene or
methyl isobutyl ketone as collectors for the furfural, effectively removing it from reaction phase more rapidly and thereby reducing furfural through
decomposition but I have not tried these methods.
Does anyone know what the black residue is? It is slowly and incompletely soluble in dilute sodium hydroxide solution but completely an rapidly
soluble in dilute bleach.
|
|
|
Fery
National Hazard
Posts: 990
Registered: 27-8-2019
Location: Czechoslovakia
Member Is Offline
|
|
Hi Boffis, very nice!!!
If I understood correctly, the NaCl is present there only to increase boiling point? The higher temperature the better yield (from corn cobs) -
industrially this is done using distillation at increased pressure which is unsuitable in lab so the temperature is increased in lab at least slightly
using NaCl?
If this is true you could put the NaCl directly into the flask.
https://www.researchgate.net/publication/230273820_Productio...
Hydrolysis of corncobs for producing furfural was carried out in a pressurized batch reactor using superheated water and diluted sulphuric acid as
catalyst. The range of experimental conditions was T=140–200°C and P=350–1550 kPa. Yields of furfural are reported as a function of reaction
temperature, particle size, acid concentration and liquid/solid ratio.
https://sci-hub.st/10.1002/jctb.280500202
Attachment: riera2007.pdf (339kB)
This file has been downloaded 425 times
|
|
|
Fery
National Hazard
Posts: 990
Registered: 27-8-2019
Location: Czechoslovakia
Member Is Offline
|
|
using methyl isobutylketone (MIBK) to increase yield:
Rapid removal of furfural from the acidic aqueous phase can increase yields by reducing the opportunity for furfural degradation in low pH water to
form polymers and other products loosely termed humins, and immiscible organic solvents can form a separate phase that extracts much of the furfural
following its formation from xylose
Although H2SO4 is a weaker acid catalyst than HCl at the same concentration, it gave higher furfural and similar 5-HMF yields
sulfuric acid was a weak oxidizer and promoted less xylose dehydration than hydrochloric acid
https://sci-hub.st/10.1039/C3RA41857J
Attachment: zhang2013.pdf (1.3MB)
This file has been downloaded 414 times
I must definitely try it but without dean-stark during production phase, avoiding to return water phase back into the reactor but collecting
everything and only then perform second distillation of the distillate using dean-stark but now without acid present in the boiling flask...
I have 5 L of MIBK but we must develop a method suitable for everyone...
[Edited on 7-1-2021 by Fery]
|
|
|
Boffis
International Hazard
Posts: 1836
Registered: 1-5-2011
Member Is Offline
Mood: No Mood
|
|
Hi Fery. My understanding is that for steam distillation to work the compound should be essentially immiscible with water. I think the purpose of the
salt is to increase the the activity of the hydrochloric acid and to reduce the solubility of furfural in water and thereby create an immiscible phase
more quickly. The problem with adding salt is that it dissolves slowly and the amount of salt that will dissolve changes only slightly with increasing
temperature. Furthermore hydrochloric acid in high concentrations reduces the solubility of salt. While emptying out the flasks from some of the
reactions there was a lot of undissolved salt and I think that in the 500ml flask scale about 60-70g is optimum. The problem with the excess salt is
that it makes powerful stirring necessary and not every amateur has an overhead stirrer; magnetic stirrers get stuck and this reduces mixing and I
suspect therefor yield too.
The problem with the straight distillation technique is that while you may get a slightly better yield the extra stages of ether extraction and then
recovering the furfural from ether, recycling the ether etc simply outweighs the small improvement in yield. Xylose seems to be better suited to this
process than ribose so I am going to have a go at oxidizing xylitol to xylose to see if this represents a potential route and there is also a possible
route from arabinose sources, such as sodium gluconate or mesquite gum.
I may have a go at the direct distillation with ribose using a continuous addition and simple distillation to see if I can get a better yield and try
adding salt to the aqueous distillate before extraction with toluene or dichloromethane.
|
|
|
clearly_not_atara
International Hazard
Posts: 2691
Registered: 3-11-2013
Member Is Offline
Mood: Big
|
|
| Quote: | | Furfural from Cherry Gum “Syrup” |
Who would have thought you could do something so cool with a cherry tree? George Washington would be proud.
Based on my understanding of imine stability and a brief Google search, furfural should form particularly stable imines -- not as stable as
dinitrobenzaldehyde, but moreso than benzaldehyde. Have you considered trying to isolate one of these? We almost never see a pure imine...
[Edited on 04-20-1969 by clearly_not_atara]
|
|
|
Fery
National Hazard
Posts: 990
Registered: 27-8-2019
Location: Czechoslovakia
Member Is Offline
|
|
Boffis, my idea avoids using ether to extract product from the first voluminous distillate. Just distill the first distillate for the second time, the
second time using Dean-Stark trap apparatus (the first time only condenser). I did it similarly with minty water when steam distillation of mint
(there is plenty of menthol in the water collected as a distillate so distilling the water for the second time gives extra menthol).
|
|
|
![[*]](images/xpblue/default_icon.gif){kind=icon}
