Supporting Information for: Investigation of the Ternary

Phase Diagram of Water—Propan-2-ol—Sodium Chloride:

A Laboratory Experiment
Cory C. Pye,* M. Angelique Imperial, Coltin Elson, Megan L. Himmelman, Jacquelyn A.
5White, Fuhao Lin

Department of Chemistry, Saint Mary’s University, Halifax, NS Canada B3H 3C3

Ideal student results

Sample # Mass NaCl (g) Mass H2O (g) Mass iPrOH (g) Observations
Phase 1
1 10.00 0.00 0.00 (s)
2 0.00 10.00 0.00 (l)
3 0.00 0.00 10.00 (l)
Phase 2
4 5.00 5.00 0.00 (s) + (l)
5 5.00 0.00 5.00 (s) + (l)
6 0.00 5.00 5.00 (l)
Phase 3a
7 2.00 8.00 0.00 (l)
8 4.00 6.00 0.00 (s) + (l)
9 6.00 4.00 0.00 (s) + (l)
10 8.00 2.00 0.00 (s) + (l), caking makes liquid
observation difficult
11 2.00 0.00 8.00 (s) + (l)
12 4.00 0.00 6.00 (s) + (l)
13 6.00 0.00 4.00 (s) + (l)
14 8.00 0.00 2.00 (s) + (l)
15 0.00 2.00 8.00 (l)
16 0.00 4.00 6.00 (l)
17 0.00 6.00 4.00 (l)
18 0.00 8.00 2.00 (l)
Phase 3b
19 2.00 2.00 6.00 (s) + (l) + (l)
20 2.00 4.00 4.00 (s) + (l) + (l)
21 2.00 6.00 2.00 (s) + (l) + (l)
22 4.00 2.00 4.00 (s) + (l) + (l)
23 4.00 4.00 2.00 (s) + (l) + (l)
24 6.00 2.00 2.00 (s) + (l) + (l)
Phase 4
25 3.00 7.00 0.00 (s) + (l)
26 1.00 0.00 9.00 (s) + (l)

Journal of Chemical Education 4/20/20 Page 1 of 7

 27 2.00 7.00 1.00 Did not finish
28 2.00 1.00 7.00 Did not finish

Journal of Chemical Education 4/20/20 Page 2 of 7

 5

Extension using finer grids

10 If the experiment is continued for an entire term, a single student can generate a

phase diagram similar to Figure S1. With more time, including some blunders along

the way, all points along the 10% grid, and selected points along the 5%, 2%, and 1%

grid can more precisely narrow the phase boundaries by this brute-force technique.

The use of propan-1-ol instead of propan-2-ol gives a slightly different diagram

15(Figure S2). This method becomes laborious due to the sheer number of possible

compositions that must be examined. Therefore, alternate methods for precisely

locating phase boundaries were sought.

Journal of Chemical Education 4/20/20 Page 3 of 7

 Figure S1. The ternary phase diagram of sodium chloride—water—propan-2-ol, as determined by this method, by a
20student working for nine weekly three-hour laboratory periods. An analytical balance was used instead of a top-
loading balance, and no temperature control was employed (ambient temperature ranged from 294-297K). The
green lines indicate approximate phase boundaries, whereas the red line correspond to literature values of the
salt solubility in aqueous propan-2-ol.

Journal of Chemical Education 4/20/20 Page 4 of 7

25Figure S2. The ternary phase diagram of sodium chloride—water—propan-1-ol, as determined by this method, by a
student working for nine weekly three-hour laboratory periods. An analytical balance was used instead of a top-
loading balance, and no temperature control was employed. The black lines indicate the three-phase region,
whereas the red line correspond to literature values of the salt solubility in aqueous propan-1-ol.

30Titrations: NaCl (aq) with propan-2-ol

If a stirred concentrated aqueous solution of sodium chloride (22-24% NaCl by

mass) is titrated with propan-2-ol, then once the phase boundary between liquid and

liquid + solid is surpassed, sodium chloride is precipitated out as a fine silt, and if

10Journal of Chemical Education 4/20/20 Page 5 of 7

 continued, then the phase boundary between two and three phases will be crossed.

35Further addition of propan-2-ol results in crossing back into the other liquid + solid

two phase region. Use of 10-20% NaCl solutions would result into crossing into the 2

phase liquid + liquid region. These titrations are slow, however, because one has to

turn off the stirrer in order to observe the phases, and temperature control is a

problem. However, it does make a visually appealing demonstration.

40Titrations: Fajans method for chloride determination

It was possible to successfully determine the chloride content of aliquots by using

Fajans method21. A 1.00 mL aliquot of solution was retrieved by a 1.00 mL syringe

and dispensed to a reaction flask, weighing the syringe before and after dispensing,

in order to determine the density of solution. 5 mL of a pH adjusted 2% dextrin

45solution was added (remade daily due to mold growth), along with 2 drops of

dichlorofluorescein indicator and swirled. The solution was then titrated with ca. 0.1

M silver nitrate (stored in a brown bottle away from sunlight and replaced every two

weeks due to photodegradation). For the chloride rich aqueous phase, to the initial 1

mL was added 9 mL of water, and a 1mL aliquot of this sample was used for the

50titration, otherwise the volume of silver nitrate required was greater than a single

burette. Unfortunately, without also being able to quantify propan-2-ol

simultaneously (see below), the results could not be plotted.

Attempt to determine propan-2-ol content by HPLC and LCMS

Aliquots of 1.5 mL of the analyte were placed in 2.5 mL vials. For HPLC, 20.00 μL

55was injected with a mobile phase 90% H2O/10% MeCN (isocratic elution, 0.5 mL/min

for 8.0 min) on a Zorbax Bonus RP 4.6x150mm (Agilent) at a temperature of 25 °C,

Journal of Chemical Education 4/20/20 Page 6 of 7

 using a diode array UV detector at 200(4) and 210(4) nm. The LC conditions of the

LCMS runs were as above, and the MS conditions were: electrospray ionization, with

50.0 ps nebulizer, 10.0L/min dry gas at 350 °C, scanned between 30-200 m/z. In the

60absence of sodium chloride, a calibration curve could be generated, but its presence,

even at the 1% and 2% levels, affects the calibration curve, making this attempt to

quantify propan-2-ol content unsuccessful.

Temperature and Volume dependence of the phase equilibria

An anonymous reviewer suggested that the temperature could be varied to see

65what the effect on the number of phases would be. For safety reasons, the

temperature range should not exceed 278-313K. If the temperature is too low, the

water may freeze, and the compound NaCl•2H2O may form. If the temperature is too

high, the vapor pressure of the propan-2-ol becomes too high in these capped culture

tubes. In addition, the volumes can be approximately characterized by measuring

70with a ruler, but there is inherent error because the bottom of the culture tube is

spherical and not cylindrical, and the pore space within the solid is not accounted

for.

Journal of Chemical Education 4/20/20 Page 7 of 7