A Study of the Lucas Test

Richard A. Kjonaas
Indiana State University, Terre Haute, IN 47809
Bernie A. Riedford
Mt. Vernon Senior High School, Mt. Vernon, IN 47620
Because of modern instrumentation, most of the classical
wet tests used for organicqualitative analysis no longer have
any value in chemical research and quality control. Some of
them do, however, continue t o serve as valuable tools for
teaching organic functional group reactivity in sophomore
organic and allied health chemistry. One very well known
example, the Lucas test,' distinguishes between primary,
secondary, and tertiary ( l o ,ZO, and 3') alcohols, and demonstrates the relative ease of lo,ZO,and 3' carhocation formation. Even thouah the Lucas test is venr easv to undersand
and perfurm,.st;dentsoften fail t o r e p o ~ a p & i t i v etest with
Z0 alcoholn. This observation holds true when using either a
macro- or microscale procedure. Furthermore, lab manuals
are unclear in defininr! the range of alcohols for which the
test is applicable. Theoriginal paper by Lucas is not of much
help because it is based upon so few examples. In an effort to
optimize its reliability for the teaching laboratory, we have
undertaken a study of the Lucas test and our observations
are reported herein.
T o review the Lucas test briefly, an unknown alcohol is
treated with several volumes of Lucas reagent (ZnCls and
concentrated hydrochloric acid in a 1:l molar ratio). As
shown in the equation below, this reagent transforms the
alcohol into an alkyl chloride.
ROH + HCl-

RCL

+ H,O

Since the alkyl chloride is not soluble in the Lucas reagent,

the test mixture becomes cloudv and eventuallv a senarate
layer appears. The appearance of cloudiness is usually what
the student is instructed to watch for. Hecause of the relative
ease of 3 O carbocation formation, 3 O alcohols react instantly,
whereas loalcohols require much longer than 1 h. For Z0
alcohols, test times such as "3-4 min", "about 5 min", and
"3-10 min" are usually cited by the authors of lab manuals.
Alcohols that are not sufficiently soluble in the reagent
might lead to a false conclusion since their insolubility can
bemistaken for alkyl chloride formation. As a follow-up teat
to confirm 3 O alcohols, the test can be repeated using concentrated HCI rather than the Lucas reagent; 3' aliohole
should react within 5min, whereas loand 2' alcohols should
not react within any reasonable length of time.
For this study we performed the Lucas test using cyclouentanol. cvclohexanol. and all 33 (ienorine stereoisomerism) of the saturated aciclic monofun~ional&ohols having
six or fewer carbonatoms (all C,H~,,~Oalcoholswithn 3 6).
Also, in keeping with the trend toward microscale procedures, drops rather than milliliters were used as the volume
unit. I t should be noted, however, that a certain alcohol/
reagent volume ratio measured in drops might not be precisely the same milliliter ratio since the size of the drops
varies among different compounds.
Presented In palt In the Division of Chemlcal Educatlon Poster
Session at the 24th Mldwest Reglonal Meetlng of the American
Chemical Soclety, November 1-3, 1989.
Lucas. H. J. J. Am. Cham. Soc. 1930,52.802-804.

'

704

Journal of Chemical Education

Experimental
Pasteur pipets C'disposahle pipets") were used for dispensing
drops of the Lucas reagent and all of the alcohols except neopentyl
alcohol, which is a solid at room temperature. The Lucas reagent
wasprepared by dissolving 1.00 mol(136 g, Fisher certified ACS) of
zinc chloride in 1.M) ma1 (100 g or 84.0 mL, Fisher reagent ACS) of
concentrated hydrochloric acid with cooling. In performing the Lucas test, the alcohol (1,2,3, or 4 drops) and reagent (10 drops) were
placed into a 13 mm X 100 mm test tube followed immediately by
vigorous swirling of the test tube for about 3 to 5 s to mix the
contents and then setting it aside. The amount of time between the
initial addition and the first appearance of a definite eloudineas was
taken as the test time. The temperatureof the room and test materiala was closely held at either 20 'C or 25 "C. When necessary to
distinguish between insolubility and a very fast reaction, the cloudy
test mixture was immediately centrifuged,and the organic layer was
withdrawn with a Pasteur pipet and analyzed hy IR and NMR
spectroscopy and hy gas chromatography.

Observatlm and Dlscuoslon

Table 1is a list of the 1"and 3 O alc'ohols used in this studv
along with their cost and boiling points. As expected, the 3;
alcohols reacted instantlv, and the loalcohols failedto eivea
positive test within 1 h . I n fact, with these alcohols-it is
almost impossible to do the test in such a way as to get a n
incorrect result!
Table 1. Cost and Bolllng Polnt of all 22. Nonsacondary
C#)I,*,O ( n 5 6 ) Alcohols
Alwhol

Boillng Polnt. 'C

Cost, @/go

memano1
PRIh4ARYALCOhwLS

ethanol
1-propanol
2-methyl-lpropanol
2,2dime~yl-l-propanol
l-butanol
2-methyl-1-butanol
3-melhyl-l-butanol
22dimelhyl-l-butan01
1-pentanal
3.3dlmeMyC1-butanol
2.3dlmelhyI-1-butand
2-melhyl-l-pamano1
24h+l-bulanol
4-melhyl-l-pentan01
3-melhyl-l-pantand
l-hexanol

78
97
10s
113(mp = 52 'C)
117
129
132

1
1
2
45
1
1
2

ERTIARY ALCOHOLS

2-melhyl-2-pmpanol
2-melhyl-2-butanol
2.3dlmelhyl-2-butand
2-melhyl-2-pentanol
3-methyl-3-pentanol
Ignoring sterecl~somar~sm.

,-"

'eased on me beat 1990 plce in qusmitles of 2 kg a leu d at

marial.

,,

least 97%

pure

Table 2. Lucaa Test TImm ol Cyclownlanol, Cyclohexanol, and All 11. ol the Secondary C , J i 2 d r 0 ( n 5 6) Alcohols wlth 10 Drops ol

Alcohol

Boiling

Order of

Point, "C

Additionb
A to R
R to A

l(20 'C)

3.0

A to R
R to A
A to R
R to A
A to R
R to A
AIoR
R to A
AIoR

0.5

0.5

1.3

Cost.Q/g

1.0
1.0
0.25
0.25
0.25
0.3
0.5
0.75

0.25
0.2

0.1
0.1
0.1
0.1

1.3

0.5
0.7

0.2

0.2

RIoA
RtoA
AtoR
R to A
A to R
R to A
AtoR
R to A

q25 ' C )

15
20

-L

AtoR
R to A
A to R
RtoA

A to R

Test Time.'Minutes
Drops of Alcohol (TempraNre.OC)
l(25 ' C )
2(25 OC)
3(25 'C)

0.2
0.25
1
0.25

0.2

0.3
0.5
0.5
0.15
0.15

'Ign~ring~ t e r B O i m r i S m .

'"A tow means mat the alcohol was added to ma reagent: "R M A" means that the reagent was added to the alcohol.
<mgamount of time between me initial addition and the first appearance of a definite cloudinem.
M the best 1990 orics in ousntities of 2 ka or less of at least 97% owe matwial.
' I many of Vlese #sopropy1abahoi runs, no claudines appeared at all.

Table 2 eives the 2' alcohol test times that were obtained
by alteringthree variables: temperaure, reagentlalcohol volume ratio, and order of addition. The data shows that (1) a 5
OC drop in temperature can give as much as a threefold
increase in the test time, (2) using more alcohol per given
amount of reagent leads to longer test times, and (3) thk test
times are often longer when the reagent is added to the
alcohol rather than when the alcohol is added to the reagent.
A fourth variable that affects the test times but that is not
reflected in the table is the extent of mixing. No matter what
the order of addition, extensive mixing can lead to longer
test times. As would he expected then, swirling a just barely
cloudy test mixture will, in some cases, cause i t to clear up
and then..uDon sittine.
For consistencv
-.become cloudvaeain.
in this study, the test mixtures weremixed by swirlingvigorouslv for about 3 to 5 s and were then olaced in a test tube
rack and allowed to sit until cloudy. A little more initial
swirline. such as 10 to 15 s. usually has verv little effect on
the tesitimes; i t is additional swiriing sometime during the
waitin2 ueriod that, in some cases, substantially delays the
appearance of cloudiness. Continuous swirling "ntil duudy
has little effect on some alcohols hut with others, especially
some of the low-molecular-weight alcohols as well L& those
with long test times, i t can cause the cloudiness virtually
never t o appear. The reason for the increase in test time
when using a smaller reagent/alcohol volume ratio might be
due to the ability of the alcohol to dissolve the alkyl chloride
as i t forms. In fact, a cloudy test mixturecan often becleared
up by adding another drop or twoof the alcohol being tested.
The reasons for the order of addition effect and the mixing
effect are not as easy to e ~ p l a i n . ~
&

In light of all this, one can easily see some reasons why the
test so often fails with 2- alcohols. For example, some students might use "a little extra" alcohol thinking that, if
anything, this should cause the test to work better! Some
instructors may want to caution students about this and
about additional swirling; others may choose rn let the studentv pay the consequences of not followingdirections. Even
if the students do follow the directions, some lab manuals
provide procedures that, because of the reagentlalcohol volume ratio, are likely to fail or, in some cases, certain t o fail.
Also, students might be using larger than desired amounts of
alcohol simnlv because some medicine dronoers deliver lareer drops t h k o t h e r s . The instructor can &imize the nu;her of failed tests bv havine the students use a Lucas test
procedure that giveaveryshGrt test times and then by citing
an expected test time that is long enough to accommodate
The reaaentlalmhol
and the temaereture
effect
were- - - ratio
~
.effect
. -~-~
. .-~~
both very b k i y mentioned in the original paper by Lucas, out the
order 01 addition effect and the extent of mixing effect were not. The
extent of mixing effect might be due, at least in part, to a loss of aikyl
chloride from the test mixture: i.e.. increased agitation might encourage alkyl chloride volatilization. It could also be a ramiflcation of the
reagentlalcohol ratio effect;i.e., areas of low alcohol concentration
that result from inefficientmixina miaht auicklv.~~~~
become cloudv.
, The
order
of addition
effect miaht~~~.
be. to some extent. an examale of
.
-~~
- the
mixing effect. However, that would not explain the observation that
when 4 drops of 3.3dlmethyl-2-butanol are added to 10 drops of
reagent followed by continuous and vigorous swirling until cloudy, the
testtime is increased tooniy about 3 min ratherthan to 8 min, which is
the test time that is obtained when 10 drops of reagent are added to 4
drops of the alcohol with 3 to 5 s of mixing.
~

~~

7 -

~~

Volume 68 Number 8 August 1991

705

most factors that tend t o delay the onset of cloudiness.

These factors include. but are not limited to, temoerature.
quality of the reagent (vide infra), and moderate variations
in the reaeent/alcohol volume ratio. The followina procedure
appears be the best choice. Place 10 drops ofiucas reagent into a small test tube. Add one drop of the alcohol to
be tested. Mix the contents by vigorously swirling the test
tube for about 3 to 5 s, and then d a c e it into a test tube
rack,hnd allow it tostand, withoutadditional mixing, until
cloudy. With 3" alcohols, cloudiness appears instantly
when the alcohol is added to the reagent; 2' alcohols reauire from about 5 s to 5 min, and loalcohols require 1 or
A r e hours.
Because of its very long test time, isopropyl alcohol is a
problem. Some lab books explain that isopropyl chloride is
too volatile to appear as cloudiness. At least one of them
recommends heating, but most just ignore the problem. By
using the short test time procedure suggested above, isopropyl &oh01 consistently gives a positive test in about 15 min
at 25 OC. With isopropyl alcohol, it is especially important
not to suoolv anv additional swirlina or aaitation other
than the Enttk mixing. The 15-min test timecan, of course,
be shortened bv heatine. A simole wav of doine this is to hold
it in the palm of the ha&; this ;educes the tes't time to about
5 min.
Some lab manuals mention that the Lucas reagent "works
best if freshlv oreoared". The tendency of Lucas reagent to
lose its reactivity Is perhaps due to both the loss of H C and
~
the uotake of water from the air. When 100 mL of freshly
prepa.red reagent was allowed tosit in an open 1-Lbeaker f i r
24 h, the volume of the reagent increased to 115 mL, and the
time required for it to gives positive test with 2-butanol (1
drop added to 10 drops of reagent) increased by a factor of
40. b n the other hand, when 160 mL of reagent was stored in
a stoppered 250-mL bottle with occasional use for 7 months,
the test time with 2-butanol increased by less than 50%.
Another batch was made from concentrated HCl that had
been stored for several months in a typical mineral acid
reagent bottle with a glass stopper; the resulting test times
were more than twice as long as those obtained when a fresh
bottle of HC1 was used to make the reagent. The data in
Table 2 can be used a s a measuring stick to check the
quality of a bottle of reagent.
We have also examined the confirmatory test for 3' alcohols. The five 3O alcohols in Table 1do indeed react quickly
(within a minute) with concentrated HCl whereas the 2'
alcohols do not. At least for the limited number of alcohols
used for this study, there appears to be no need for this
confirmation since even the shortest 2' alcohol test time
withLucas reagent ( 5 4 s) is clearly distinguishable from the
instantaneous result that is characteristic of the 3" alcohols.
Furthermore, when using concentrated HC1 instead of the
Lucas reagent, the "cloudiness" is not easy t o recognize. I t
quickly turns to larger droplets or to a difficult-to-see layer
of alkylchloride. Of the five 3' alcohols used in this study, 3methyl-3-pentanol is by far the trickiest when doing the
concentrated HCI test.
Finally, any alcohol that is insoluble in the reagent might
aooear to be eivine a oositive 3O test. For this reason, it is
important forihe &dent to know in advance which alc&ols
~~~~~

~~~

708

Journal of Chemical Education

in the list of unknowns are soluble and can therefore be

relied upon to give a true result. Lab manuals, however, are
inconsistent in telling which ones are soluble. The following
statements are typical (our italics):
(1) "The Lucas test is used to distinguish between l o , ZO,
and 3O monofunctional alcohols having fewer than six
carbon atom^."^
(2) "The Lucas test distinguishes between lo,ZO, and 3'
alcohols having less than eight carbon^."^
Using a 10:l reagent/alcohol volume ratio, 1-octanol gives
an insoluble mixture that does indeed look very much like a
positive test. 1-Heptanol and 2-heptanol, however, are both
ouite soluble in the Lucas reaeent. and manv. mavbe all. of
the seven-carbon CnHzn+20arcohbls might &so de soluble.
Assembline a stock of all of them (39 if stereoisomerism is
ignored) f; testing would be a difficult task. Based on the
results of this study, using 1 drop of alcohol added to 10
drops of reagent, it can be said that, a t the very least, all
saturated acvclic monofunctional alcohols hauing six or
feu.pr carbons are soluble in the Lucas reagent.
That statement may not be true when 3 or 4 drops are used
per 10 drops of reagent. In fact, sometimes when 10 drops of
reagent are added to 3 or 4 drops of some of the alcohols, an
interesting event occurs after ;delay of a couple of minutes
or so. Apparently, some undissolved alcohol that has been
clineine
.--- t o the bottom of the test tube suddenlv breaks loose
and, by streaming to the surface of the very dense Lucas
reagent, gives a momentary illusion of a positive test. The
real cloudiness, alkyl chloride formation, appears later-all
the more reason to use a high reagent/alcohol ratio.

~.~
~

~~

~~~~

~~~

Summan and Conclusion

With a little awareness and cautioll on the part of both the

lab instrudor and the students, the Lucas test can be very
reliable. Resisting the urge to use a little extra alcohol is
es~eciallv
The data in Tables 1 and 2 should be
. im~ortant.
.
useful to the lab instructor when (1) developing a list of
unknowns, (2) deciding whether a particular batch of Lucas
reagent is reliable enough for the students to use, (3) choosing a procedure for the students to use, and (4) describing to
the students the expected results when 2 O alcohols are subjected to the Lucas test. One can reliably expect 2' alcohols
to give a positive test within about 5 s to 5 min when adding 1
drop of alcohol to 10drops of reagent. Isopropyl alcohol is an
exceotion. There is orobablv no need to include the concentra&d HC1 step (the confirmation of 3' alcohols), especially
if the list of unknowns includes onlv those alcohols used in
this study.
Althoueh a comolete list (or rule for eeneratine a list) of all
t does not
alcohols chat are soluble &the ~ u c a b e a g e nGill
exist and orobablv never will, students can at least be told
that all C ; H ~ , + ~ OC.~6 alcohols are soluble.

'Mayo. D. W.; Pike. R. M.; Butcher, S. S. Microscale Organic

~

Laboratom 2nd ed.: Wllev: New York. 1989: D 41 1.

ato on; 'D. C. Liboriixy lnvesii&tions 'in organic Chemistry;
McGraw-Hill: New York. 1989; p 736.