T H E PRODUCTION OF FURFURAL FROM XYLOSE SOLUTIONS BY MEANS OF HYDROCHLORIC ACIDSODIUM CHLORIDE SYSTEMS ELLIS I. FULMER, L. M. CHRISTENSEN, R. M. HIXON,
AND
R. L. FOSTER
Department of Chemistrg, Iowa State College, Ames, Iowa Received Mag $5, 1056 I. INTRODUCTION
One of the outstanding achievements in the utilization of agricultural wastes in the manufacture of industrial chemical9 is the development of the furfural industry. The furfural is produced by heating oat hulls a t about 60 pounds pressure with about 5 per cent sulfuric acid for five or morc’ hours (14, 16, 5). The yield is about 50 to 60 per cent of the theoretical. The studies reported in the present communication have to do with the production of furfural from strong xylose solutions by various combinations of hydrochloric acid and sodium chloride, in order to establish general principles of procedure which could be applied directly to the agricultural material or to the xylose-containing hydrolysates made therefrom. The presence of the salt decreases the solubility of the furfural, ensuring its more rapid and complete removal by the solvent. The salt likewise increases the activity of the acid. This latter phenomenon has been studied with reference to various hydrolytic actions (2, 3, 4, 6 , 7, 8, 9, 10, 11, 20). In general, the hydrolytic action of a strong acid, such as hydrochloric acid, is proportional to the “apparent” hydrogen-ion concentration, which is conveniently expressed in terms of pH. So far as the authors are aware, this principle has not been applied to a dehydration action such as the production of furfural from xylose. Adams and Vorhees ( l ) , Hurd and Isenhour (13, 14), and other$ have employed saltr in this reaction for the purpose of reducing the solubility of the furfural in order to facilitate distillation. 11. GENERAL PROCEDURE
AND PRELIMlNART EXPERIMENTJ
A . General procedure The aqueous xylose solutions plus the dehydrating agents were refluxed with an immiscible solvent in which furfural is very soluble. These solvents included benzene, toluene, and carbon tetrachloride. Toluene is the solvent employed in the studies here reported in detail. The purpose 133
134
FULMER, CHRISTEXSEX, HIXON, AND FOSTER
of the solvent is twofold. It removes the furfural as formed, thus cutting down the opportunity for polymerization of the furfural in the presence of the dehydrating reagents, and it also permits the building up of high concentrations of furfural in the solvent, thus allowing an easy and economical separation by distillation. The concentration of furfural in the toluene was determined by specific gravity measurement with a Westphal chainomatic balance. Quantitative experiments showed the specific gravity of toluene-furfural systems to be a linear function of the concentration of furfural. Toluene was refluxed with equal volumes of the sodium chloride-hydrochloric acid solutions of the strength employed in subsequent experiments ; there was no significant change in specific gravity. The toluene solutions of furfural, obtained by action of the sodium chloride-hydrochloric acid systems upon xylose, wcre shaken with sodium sulfite. The specific gravity was restored to that of the pure toluene. These toluene-furfural systems were subjected to fractional distillation. The distillation curve corresponded to that obtained for known furfural-toluene systems. The furfural so obtained was identical with pure furfural. The xylose was analyzed by three different methods, those of Shaffer and Hartmann (181, Slater and Acree (19), and Kline and Acree (15). Thc purity was found to be 95.5, 96.2, and 95.5 per cent by the respective procedures, with a n average value of 95.7 per cent. Since the theoretical yield of furfural from pure xylose is 64 per cent, the maximum yield from the xylose used would be about 61 per cent.
B. The inJluence of volume ratios u p o n yields of furfural In table 1 are given data showing the influence of volume ratios upon the yield of furfural from 20 per cent xylose refluxed for five hours in the presence of 0.50 N hydrochloric acid-40 per cent sodium chloride. The concentrations of sodium chloride and of xylose throughout this paper are expressed in grams per 100 cc. of acid used. It is evident that the yield of furfural increased slightly u p to a ratio of 67: 100 and was practically constant beyond that point. I n subsequent experiments the toluene and aqueous systems were employed in equal volumes. C . IiiJIzlence of xylose conct~atrationso n yields of furfural In table 2 are givcii data on the yield of furfural from various concentrations of xylose refluxed for five hours in the presence of 0.50 N hydrochloric acid-40 per cent sodium chloride. The results show a decrease in yield with increase in xylose concentration. In subsequent work a xylose concentration of 20 per cent was employed.
D. T h e efect of certain salta Preliminary experiments showed soine interesting effects of salts upon strong xylose solutions. For example, 250 cc. of water wa> added to 75 g.
135
PRODUCTION OF FURFURAL FROM XYLOSE SOLUTIONS
of xylose and 75 g. of ammonium chloride. The solution ww heated on a hot plate; a strong odor of furfural was soon evident. However, the furfural could not be readily extracted from the mixture by means of benzene, toluene, or carbon tetrachloride. On vigorous shaking, the mixture proved to be an excellent emulsifying agent for the solvent. When subjected t o distillation the furfural began to distill over at about 110°C. At 12OOC. the mixture had a strong tendency to foam and the furfural distilled in such amounts that each drop wm diphwic. These results indicate the TABLE 1 Effect of varying volume ratios of toluene with 80 per cent xylose solution with 0.50 N hydrochloric acid and 40 per cent sodium chloride for five hours AQUEOlJB BOLUTION
TOLUENE
cc
.
cc
20 35 50 100 80 125 65 150 50
TOLUENE P E R 100 CC. O F AQUEOUE BOLUTION
YIELD OP FURFURAL P E R 100 0. OF XYLOSE
cc .
oram8
19.4 37.6 53.8 62.9 67.2 67.9 69.9 80 0 100.0
29.8 30.7 30.7 31.8 33.3 33.2 34.0 34.0 33.6
.
103 93 93 159 119 184 93 188 50
TABLE 2 Effect uf varying concentrations ojxylose, using 0.50 N hydrochloric acid and d o p e r cent sodium chloride for jive hours YIELD OF mnW R A L P E R 100 G OF XYLOSE
PER CENT OF THEORETICAL YIELD
42.5 40.0 39.4 36.0 33.6
70 66 65 59
XYLOSE
O F XYLOSE
PER CENT 01THEORETICAL YIELD
29.4 28.0 25.5 22.7
48 46 42 37
per cent
per cenl
4 8 10 15 20
YIELD OB. m n FURAL P E R 100 C
55
30 40 50 60
I
formation of an intermediate product, which is insoluble in the solvents employed but which a t higher temperatures yields furfural readily. Similar results were obtained with ammonium sulfate, ammonium dihydrogen phosphate, and ammonium tartrate. Strong solutions of aluminum sulfate and of zinc chloride also gave furfural on boiling with concentrated solutions of xylose, but in these instances the furfural was readily extracted with the solvents. Strong solutions of sodium chloride, calcium chloride, or of sodium sulfate did not lead to the formation of furfural.
136
FULXlhH, CHRISTENSEN, HIXON, AKD FOSTER
TABLE 3 Yields of f w f u r a l f r o m 20 pel. cent zylose solution with various combinations of hydrochloric acid-sodium chloride __ N O R h l h L l PY
OF
HCI
~~
N
0 26
NaCl
PH
__
__
0
10 15 20 25 30 35 40 45
0 75
1.oo
_ _ _ _ _ _ _ _ _ _ _ _I2
zicr cent
3
0 'VI
-
1.1 1.1 2 . 8 2.2 2 . 2 1.8 2.7 2.7 4.7 3.7 3.7 0.60 0 0.46 1.1 1 . 5 1 . 5 3.1 3 . 1 3 . 1 2.5 3 . 7 3 . 7 5.1 5.0 5.0 0.32 1 . 7 2 . 0 2 . 0 4 . 0 4 . 2 4.2 4.3 5.0 5 . 0 6 . 3 6 , s i . 8 3.7 0.18 L .5 0.04 5.1 -0.10 13.216.3 16.620.1 1.0 -0.2.1 7.4 7 . 1 7.113.6 17.819.6 22.424.6 5.7 -0.38 10.210.010.016.8 12 12.619.3 22.923.8 26.326.4 -0.52 12.4 3.1 20.6 26.926.4 30.930.4 -0.66 0.27 0.13 -0.01 -0.15 -0.29 -0.13 -4.57 -0.71 -0.85 -0.99
1.6 2 . 2 2 . 3 5.2 4.5 4.6 3.4 5.9 5.5 8 . 3 6 . 5 3.9 3.1 3 . 2 7.1 6.2 6 . 2 7 . 3 7.9 7.6 8.7 8 0 4 . 3 4 . 3 4 . 3 8.8 8 . 5 8.511.711.210.513.2123 6.0 5.8 5.8 12.2 11.5 11.0 15.5 13.5 13.5 18.2 16.6 8 . 4 7.9, 7 . 9 15.9 15.5 15.1 19.8 20.4 18.2 22.8 22.9 10.7 i n . 7 i o .7 19.3 20 . o 19.5.21.8 24 .o 28.8 14.8314.814.8 23.2 23.4 28.8 28.8 32.9 17.2 19.027.5 28.232.7 32.433.4 19.3 22,430.4 30.936.1 36.334.0 19.3 25.130.4 32.433.5 37.136.9
7. I
5 10 15 20 25 30 35 40 45
0 5 10 15 20 25 30 35 40 45
0.07 -0.07 -0.21 -0 35 -0.49 -0.63 -0.77 -0.91 -1.05 -1.19
3.7 3.7 3 . 6 5 . 6 6 . 6 6.9 5.0 8 . 5 8.511.510.7 5 . 3 5 . 1 4.9 9.7 9 . 1 9.112.811.811.215.314.E 6 . 9 7.1 6 . 6 13.0 12.6 12.6 16.7 16.7 15.122.4 20.4 9.6 9.5 9.1 18.3 18.2 17.0 21.8 21.9 20.9 26.4 26.2 13.2 13.2 12.6 22.5 23.4 21.4 25.7 25.7 32.4 26.9 31.8 30.9 35.8 18.0 18.2 17.8 26.8 30.233.6 33.937.6 21.431 6 20.2 24.534.4 33.139.0 36.339.8 24.4 34.739.0 38.040.4 24.6 27.534.7 35.540.8 40.740.3 25.7 29.535.6
17 1.5 36 $ 5 1.6 3.1 1.7 53 30
0 5 10 15 20 25 30 35 40
-0.01 -0.15 -0.29 -0.43 -0.57 -0.71 -0.85 -0.99 -1.13 -1.27
2 . 5 4.2 4.2 9.4 8 . 9 6.810.210.010.5 5.0 5 . 6 5.712.712.011.016.415.514.5 7.7 7 . 8 7.817.516.615.120.320.919.1 10.7 10.7 10.521.222.4 20.4 25.8 15.314.414.526.9 25.732.0 19.319.517.431.5 29.534.0 22.934.2 31.635.2 23.0 25.5 26.334.8 34.736;8 26.8 28.838.1 35.537.5 135.5 ,30.937.8 ,36.3,38.4 136.31 26.2
0
46
9.5 3.9
1.6 1.1 -4X 1 ,I
$.!I 5.5
3.0
?.8
.-
137
PRODUCTION O F FURFURAL FROM XYLOSE SOLUTIONS
TABLE 3-Concluded 8 EOURS NORMALITY
N&l
OF
___-N
per cent
1.50
0 5 10 15 20 25 30 35 40 45
-0.24 -0.38 -0.52 -0.66 -0.80 -0.94 -1.08 -1.22 -1.36 -1.50
2.00
0 5 10 15 20 25 30 35 40
-0.45 -0.59 -0.73 -0.87 -1.04 -1.18 -1.32 -1.46
_ _A f
-----_
4.1 7.2 7.113.913.513.519.819.120 9 . 7 . 9 . 8 9.518.918.618.625.425.725 14.513.513.224.025.123.430.3 31 18.018.218.229.0 28.234.4 33 22.9 22.433.0 30.936.0 34 26.2 25.135.7 33.935.1 35 31.0 28.836.9 35.534.0 35 31.6 36.836.8 36.334.2 35 32.8 31.639.2 36.734.9 35 38.9 37.134.7 35
13.4 12.6 11.5 17.4 17.020.4 17.6 17.4 18.9 23.4 23.4 24.6 23.5 23.4 21.9 28.6 28.8 28.3 24.632.4 31.6 31.5 28.233.8 33.9 35.5 30.934.0 35.5 36.0 32.435.0 35.9 36.4 33.934.7 36.3 35.9 35.538.0 36.3 -1.74 36.7 36.036.4 36.7
4R
____
I _ -
f
= exoeriment
tions of sodium chloride and hydrochloric acid. 111. T H E YIELD OF FURFURAL I N THE PRESENCE O F VARIOUS COMBINATIONS OF SODIUM CHLORIDE-HYDROCHLORIC
ACID
In table 3 are given data for the yields of fuifural from 20 per cent xylose in the presence of 0.25, 0.50, 0.75, 1.00, 1.50, and 2.00 N hydrochloric acid with concentrations of sodium chloride from 0 to 45 per cent for various periods of time. The data show the marked effect of the presence of the salt. The highest yield without salt is 20 per cent at 1.50 N hydrochloric acid after six hours. This same yield was given in two hours by approximately the following hydrochloric acid-sodium chloride systems: 0.50N40 per cent; 0.75 N-30 per cent; 1.00 N-25 per cent; and 1.50 N-15 per cent. The highest yields are 35 to 40 per cent, or 57 to 65 per cent of theoretical. IV. THE RELATION O F YIELD O F FURFURAL TO THE pH O F THE SODIUM CHLORIDE-HYDROCHLORIC ACID SYS'I'EMS
In+table 4 are given data for the pH values of 0.25 N hydrochloric acid in the presence of varying concentrations of sodium chloride. These values
138
FULMER, CHRISTENSEN, HIXON, A N D FOSTER
werc determined potentiometrically. No corrections were made for diffusion potential. Data are also presented showing the effect of the sodiuni chloride upon Lhe solubility of the furfural. While the “apparent” hydrogen-ion concentration increased nearly ninefold, the solubility is decreased only to one-third. The relation between pH and concentration of sodium chloride for table 4 is pH = 0.60
- 0.028
X per cent NaCl
(1)
Also, for a given time period, the yield of furfural is a linear function of the “apparpnt” hydrogen-ion concentration, through a major portion of the curve. The deviation from this linear relationship is associated with Effccl O
J S O ~ L Zchloride L ~
TABLE 4 upon the pH of0.25 N hydrochloric acid and upon lhc solubilit?, offurfurnE* at 25°C.
,
“Rm;;
Ih
-I 5 10 15 20 25
0.273 0,346 0.476 0.654 0.860 1.228 1,763 2,331
0 46 0 32
0 18 0 07 -0 09
1 00 1 27 1 74 2 39 3 15 4 50 6 46
2 2 3 3 3 3
90 98 06 10 14 19 3 20
I
0 0 0 0 0 0 0 0
60 52 44 40 36 31 29 25
DISTRIBLTION IlATIO
0 207 0 174 0 144 0 120 0.114 0.097 0 091 0.077 I
a Seven per cent furfural in toluene used, 50 cc. of 0.25 N hydrochloric acid, and 50 cc. of toluene-furfural solution.
decreased yield, due to polymerization of the furfural and other factors which cause a marked darkening of the reaction mixture. During the range for which equation 1 holds, log!
= a
+ 0.028 X per cent NaCl
(2)
that is, the logarithm of the yield of furfural is a linear function of the concentration of salt and also of pH. Equation 2 was applied to all the hydrochloric acid-sodium chloride systems; these values are given asfi in table 4. The agreement is entirely satisfactory through ranges in which the yield of furfural is a linear function of the “apparent” hydrogen-ion concentration. In general the relation is linear up to about 22 per cent yield of furfural. This treatment permits the calculation of yields below 10 per cent, which are somewhat erratic owing to the low concentrations in the toluene.
139
PRODUCTION OF FURFURAL FROM XYLOSE SOLUTIONS
Graphs of pH against furfural yield (f~) for the 0.25 N hydrochloric acidsalt combinations for the 2-, 4-,and 6-hour periods permitted the calculation of the pH of other concentrations of pure acid, on the assumption that the yield of furfural is a function of pH only. Such values are given iii table 5. There are also included data for various concentrations of hydroTABLE 5 T h e p H values of 0.50 N , 0.75 N , 1.00 N , 1.50 N , and 8.00 N hydrochloric acid as calculated on hasis of yields with 0.96 N hydrochloric acid-sodium chloride systems, compared to p H values as calculated f r o m the activity coe,@ient of hydrochloric nrid (ONCENTHATION
OF
HCI
Molar
I
pH FROM FOR
2 hrs
Molal
__0.25 0.25 0.50 0.50 0.75 0.76 1.00 1.02 1.50 1.54 2.00 2.08 -__
QRAPH FURFURAL YIELD
pH
0.25 N HCI-NaCI
4 hrs
6 hrs
Average
lelative value of
y
29 0 27 0 26 0 27 04 0 09 0 09 0 07 02 -0 02 +O 01 -0 01 24 -0 23 -0 24 50 -0 40 -0 45
HCI
,H- pHi
y'm
Y
___________ 0 0 -0 -0 -0
PnoM VALUEI) OF y
0 761 0 757 0 780 0 815 0 901 1 04
___
0.25 0.995 0.50 1.03 0.78 1.07 1.09 1.18 1.82 1.37 2.83
0.60 0.30 0.lP -0.04 -0.26 -0.4.5
1
0 -0 03 -0 04 $0.03 +o 02 0
TABLE 6 Values of p H of various Combinations of s o d i u m chloride and hydrochloric acid*
I PER CENT
0 5 10 15 20 25 30 35 40 45
NaCI/-
o,25
0.60 0.46 0.32 0.50 0.04 -0.10 -0.24 -0.38 -0.52 -0.66
,
NORMALITY O F
0.50
I
0.27 0.13 -0.01 -0.15 -0.29 -0.43 -0.57 -0.71 -0.85 -0,99
0.75
0.07 -0.07 -0.21 -0.35 -0,49 -0.63 -0.77 -0.91 -1.05 -1.19
HCl IO0
150
____
-0 01 -0 15 -0 29 -0 43 -0 57 -0 71 -085 -0 99 -1 13 -1 27
,
-0 -0 -0 -0 -0 -0 -1 -1 -1 -1
24 38 52 66 80 94 08 22 36 50
2.00
-0 -0 -0 -0 -1
-1 -1 -1 -1 -1
45 58 73 87 01 18 32 46 ti0 74
* Data for 0.25 N hydrochloric acid based on experimental data from the relation, pH = 0.60 - 0.028 X per cent NaCl. Data for other concentrations calculated from table 5, using the relation p H = h 0.028 X per cent NaCI.
-
chloric acid calculated from the activity coefficients given by Randall and Young (17). Our value for 0.25 N hydrochloric acid is pH = 0.60 as compared to 0.71 calculated from the activity coefficient. Relative values of the activity coefficients, y', were used to correct our values to the basis of the data of Randall and Young. It is evident that the pH values
140
FULMER, CHRISTENSEN, HIXON, AND FOSTER
calculated from yields of furfural in the presence of the pure acid agree remarkably well with those calculated from the activity coefficients. It was noted above that equation 2 applies, within specified limit,s, t o all combinations of hydrochloric acid and sodium chloride. The assumption seems warranted that the pH of other concentrat'ions of acid decreases at the same rate as obtained for 0.25 N hydrochloric acid, that is, I
pH
=
b
-
0.028 X per cent NaCl
(3)
Values so calculated arc given in table 6. Graphs mere constructed by plotting log j" against p1-I for the 2-, 4,6-, and %hour periods. Each curve included all the data for all combinstions of hydrochloric acid and sodium chloride for the given period. From t'hese graphs the furfural yield for each pH value was read. The data so obtained are given as f2 in table 4. It is evident that,, within experimental accuracy, the yield of furfural from 20 per cent xylose solution in the presence of hydrochloric acid-sodium chloride Combinations is detrrniined by the p H values of the wide variety of combinations employcd.
v.
SUhIhIlRY
St,udies are reported on the production' of furfural froin strong xylose solutions using hydrochloric acid-sodium chloride solutions as dehydrating agents. The xylose-hydrochloric acid-sodium chloride systems were refluxed, a t atmospheric pressure, with toluene. The furfural yield was determined from the specific gravities of the resulting furfural-.toluwr solutions. The concentrations of hydrochloric acid used were 0.25, 0.50. 0.75, 1.00, 1.50, and 2.00 normal, each in the presence of 0, 6 , 10, 15, 20, 25, 30, 35, 40, and 45 per cent sodium chloride. The yield of furfural was twice as great for 4 per cent xylose as for 60 per cent xylose. Det'ailcd experiments were performed for xylose a t 20 per cent concentration for which the yield in furfural is about 20 per cent less than for 4 per cent xylose. The yield of furfural is increased about one-third for each 5 per cent atldition of sodium chloride up to a furfural yield of about 22 per cent,. From this point the increase in yield is less. For all cases, for a given time period, t>heyield of furfural is, within reasonable limits, dependent only upon the pH of the hydrochloric acid-sodium chloride combinations. That is, the yield is dependent upon the thermodynamic degree of dissociation of tlie acid (the activity coefficient), REFERENCES (1) ADANS AND VORHEES:Organic Syntheses, Collective Volume, p. 274. John Wiley and Sons, Fnc., N e w York (1932). (2) BKERLOF:2. physik. Chern. 98, 260 (1921).
PRODUCTION O F FURFURAL FROM XYLOSE SOLUTIONS
141
(3) BOWE:J. Phys. Chem. 31,291 (1927). (4)BRBNBTED: Trans. Faraday SOC.24, 630 (1928). (5) BROWNLEE: Ind. Eng. Chem. 19,422 (1927). AND FULMER: Physiology and Biochemistry of Bacteria, Vol. I, (6) BUCHANAN p. 262 (1928); Vol. 11, p. 234 (1930). The Williams & Wilkins Co., Baltimore. (7) COLINAPTD CHAUDIN: Compt. rend. 192, 1229 (1931). (8) DAWSON:Trans. Faraday SOC.24, 640 (1928). (9) FLOYD:J. Phys. Chem. 36, 2968 (1931). (10) HARNED AND AKERLOF: Trans. Faraday SOC.24, 666 (1928). (11) HUNTER:J. Chem. SOC.1928, 2643. J. Am. Chem. SOC.64,317 (1932). (12) HURDAND ISENHOUR: (13) HURDAND ISENHOUR: J. Am. Chem. SOC. 64,693 (1932). (14) KILLEFER:Ind. Eng. Chem. 18,1217 (1926). (15) KLINEAND ACREE:Bur. Standards J. Research 8, 25 (1930). AND BROWNLEE: Chem. Met. Eng. 27, 299 (1922). (16) MINER,TRICKEY, AND YOUNQ: J. Am. Chem. SOC.60, 989 (1928). (17) RANDALL J. Biol. Chem. 46,365 (1920). (18) SHAFFERAND HARTMANN: (19) SLATER AND ACREE: Ind. Eng. Chem., Anal. Ed. 2,274 (1930). (20) TERRY:J. Am. Chem. SOC.60,1239 (1928).
