Extreme pH Measurements

Joe Lowry, Richard Ross, and Brad Venner

U.S. Environmental Protection Agency, National Enforcement
Investigations Center, Box 25227, Denver, CO 80225, USA

Introduction

Commercial chemical products when abandoned or otherwise disposed are
solid waste and if listed or characteristic are hazardous waste.  Some
DOT corrosive commercial chemical products are H2SO4, HCl, NaOH and KOH.
 Common concentrations of H2SO4 include about 30% for electrolytic
grade, 60% for fertilizer grade, and 98% as concentrated grade.  HCl is
typically 30% for industrial use, 10% for household use and 38% as a
concentrated grade.  Commercial grade NaOH, caustic soda, is typically
50% while KOH, caustic potash, is 45%.  None of the above mentioned
commercial products are listed.  Therefore, if they are to be captured
by RCRA, they must be identified with properties of characteristic
hazardous waste.

H of ≤ 2 or ≥ 12.5 as determined by a pH meter using SW846 Method
9040”.  Method 9040 requires calibration by bracketing sample pH
values with standards that span about 3 pH units.  It also recommends
the use of a pH 2 standard for sample values ≤ 2 and a pH 12 standard
for sample values ≥ 12.  Hence, standards with pH values ≤ -1 and
≥ 15 are needed.  However, the NIST/IUPAC primary pH standards range
from 3.5 to 10 with secondary standards of 1.68 and 12.45.  Standards
commercially available typically range from pH 0.5 to 13.

Nordstrom et al (1) proposed the use of Pitzer equations and
modification thereof in assigning pH values for strongly acidic H2SO4
standards.  Licht (2) proposed the use of the mean activity coefficient
in assigning pH values for strongly basic KOH standards.  In this paper,
we demonstrate that linearity in continuity with the curve of the
primary standards is achieved for some strongly acidic and basic
solutions when using the molal concentration and the activity of water
in assigning pH values. (3-5)

We also demonstrate, contrary to a claim in Method 9040, that the
corrosivity of concentrated acids and bases can be measured by means of
potentiometric pH measurements with a glass electrode.  Various issues
pertinent to RCRA corrosivity characterization via potentiometric pH
measurements are discussed.

Experimental

pH measurements were made with a Corning 345 pH/mV meter, a Corning
Model 96 thermocouple, and an Orion Ross 8172BN combination pH
electrode.  Eight commercial buffers ranging from pH 1.09 to 12.45 were
from Ricca and Radiometer.

H2SO4 (95.8%), HCl (37.8%), HClO4 (69.7%), HNO3 (69.7%), CH3COOH
(99.7%), NaOH (97.7%), NaCl (99%), KCl (99%), KOH (87.8%), glycerin,
2-propanol, polyethylene glycol 400 liquid, and pyridine were from
Baker.  The 60.5% HClO4 was from GFS.  The CH3SO3H (99.5%), (CH3)4NOH
• 5H2O (97%), 1,4-dioxane and mineral oil were from Aldrich.  The
toluene, iso-octane, and hexane were from Burdick and Jackson.  The
methanol was from EMD and the vacuum pump oil 20 was from VWR.  The
kerosene was from Fluka.  These liquids and gravimetrically prepared
solutions of the acids and bases were analyzed.  KOH pellets and
pulverized pellets, and saturated solutions of KOH were also analyzed.

Results and Discussion

EPA on Aqueous

Currently, the EPA defines ‘aqueous’ for the corrosivity
characteristics as meaning in a form amenable to pH measurement.  The
basis for this was stated to be the following in the 1980 regulatory
background document; “A few comments suggested that EPA specifically
define “aqueous” in terms of viscosity or percent water.  The Agency
has not developed a specific definition because of the widely varying
physical and chemical properties which influence the form of wastes. 
Those who generate, treat, store or dispose of a waste can best
determine whether it is in a suitable form for pH measurement.”  Forms
mentioned as suitable were liquids and non-liquids including
suspensions, sols, and gels.  Between 1980 and the current 1998
standpoint, guidance was provided that 60% and later 50% water content
would be aqueous.

In the final rulemaking in 1980 Office Solid Waste (OSW) Method 5.2
replaced the proposed Office of Water Method 150.1 as the required
method.  Method 5.2 was the same as Method 150.1.  In 1982 Method 9040
appeared in the second edition of the OSW SW846 methods manual.  In 1994
Method 9040 in the 1986 third edition of SW846 replaced Method 5.2 at
§40CFR261.22 as the regulatory required method.  Soon thereafter Method
9040A replaced Method 9040 in SW846.   Today Method 9040C appears in
SW846.  Additional information about appropriate forms was given in
Method 9040 and its prodigies.  All the 9040 methods contain; “This
method is used to measure the pH of aqueous wastes and those multiphase
wastes where the aqueous phase constitutes at least 20% of the total
volume of the waste.”  Hence, it would appear a part of the waste
could cause the whole waste to be identified as hazardous and multiphase
waste with < 20% aqueous phase, were provided exclusion.  Personnel
communication with staff provided that it was thought that 20% would
provide a sufficient volume for analysis.  In 1998, OSW provided other
guidance that the 20% means free water.  Free water was not defined.

Starting with Method 9040A, the following appears in the prodigy; “The
corrosivity of concentrated acids and bases, or of concentrated acids
and bases mixed with inert substances, cannot be measured.  The pH
measurement requires some water content.”  Further, the prodigy
require caustic pH samples be measured at 25°C.  This was done to
specifically cause the exclusion of lime treated wastes as at 20°C a
saturated solution has a pH of 12.6 while at 25°C the pH is 12.45. 
This exclusion neglected that at pH 11 and higher metals are mobilized
as hydroxide and carbonate complexes and proper treatment following lime
addition requires pH adjustment to between 8 and 10 to optimize metals
removal.  This modification also essentially precluded field pH
measurements for caustics and made impractical meeting the 15 minute
holding time required to prove a sample does not have the property
associated with a corrosivity hazardous waste.

≥ 5 point); 2) bracketing (≥ 2 point) the sample values; and 3)
single point and assume a slope.  Method 9040 requires bracketing with
standards that span about 3 pH units.  Bracketing is interpolation while
single point and assuming a slope is extrapolation.  Proper application
of Method 9040 would require a pH standard at -1 or less for acidic
samples with the pH 2 standard and one at pH 15 or greater for caustic
samples with the pH 12 standard. The lowest commercial pH standard found
was 0.5 and the highest was 13.  Hence, laboratories would need to
prepare in-house standards in the proper application of the method. 
Presumably, if a sample value can not be bracketed then the sample is
not amenable to measurement by the method.  If a sample value could be
bracketed and was not, then Method 9040 was not followed.  Further,
Method 9040 requires that duplicate pH measurements be made on the
sample and that these measurements agree to within 0.1 pH units.  If
duplicate measurements are greater than 0.1 pH units then the sample is
not amenable to measurement by Method 9040.

Cobbling things together, it appears the OSW meaning of aqueous for the
corrosivity characteristic is any phase of a sample that is at least 20%
of the total volume which contains some water, including physical forms
that are liquids, suspensions, sols, or gels whose chemical form is not
a concentrated acid or base without water, or water saturated with
calcium oxide or its hydrates, and whose pH value measured by pH glass
membrane electrode potentiometry can be bracketed with standards and
replicated within 0.1 pH units for duplicate measurements.

Some Water

The more recent OSW guidance is that aqueous for the corrosivity
characteristic does not mean 50% water as it does for the ignitability
characteristic.  As discussed, aqueous means amenable to pH measurement
by Method 9040 for the corrosivity characteristic.  This guidance
equates one ambiguous un-measurable term, aqueous, with another
ambiguous un-measurable term, amenable.  Along the same vein, Method
9040 seems to equate amenable with the ambiguous phrase ‘some water’
by stating that; “The pH measurement requires some water.”  Although
water might be able to be measured for most samples, ‘some’ is not
defined as a quantity.

Consider, unless a sample is collected, stored and analyzed under an
artificial atmosphere, it will usually contain ‘some water’ because
air has humidity.  Further, the glass electrode will have water adhered
to the membrane and water will leak from the reference electrode.  Water
can aid ionization of solutes or other solvents (H2SO4, CH3SO3H,
CH3COOH, etc.) however self ionization is not unique to water. 
Measurement of pH for non-aqueous systems can readily be found in the
scientific literature.

Generally, a solution is considered aqueous if water is the solvent.  If
two or more miscible solvents are present, than the one at the highest
mole fraction is considered the solvent. If one equates the Method 9040
statement concerning multiphasic samples that the aqueous phase must be
20% of the total volume as 20% water content means amenable or ‘some
water’ than the scientific meaning of aqueous would not be met for
most commercial acids.  Table 1 gives the mole fraction concentrations
for some common commercial concentrated acids.  Note that 20% water does
not give an aqueous solution when the acid has a molecular weight less
than 72 g/mole.  Only phosphoric acid and sulfuric acid have molecular
weights greater than 72 g/mole.  Further note that concentrated
hydrofluoric acid and concentrated hydrochloric acid would be aqueous
solutions.

concentrated solutions or beyond the nominal pH range are lacking. 
There is no scientific basis for a contention that because a material
contains 20% or 50% water that it is amenable to pH measurement by
Method 9040.

Amenable as Reproducible

The NIST/IUPAC standards are assigned pH values with an electrochemical
cell (Harned cell) that has no liquid junctions.  For practical pH
measurements these assigned pH values are used to calibrate an
electrochemical cell involving a glass electrode and a liquid junction
at the reference electrode.  A criterion for a primary standard is a
small liquid junction potential on the glass electrode.  The liquid
junction potential increases with ionic strength and is affected by the
nature of the solute and solvent, colloids, and suspended solids.

Generally for aqueous standardization the pH 7 standard has a potential
of 0 millivolts.  A liquid junction potential manifests itself as a
translational displacement of this intercept term.  It is essentially a
change in the reference electrode potential.  It is potential unrelated
to the hydrogen ion activity and hence an error.

The operational definition of pH is based on the standard buffers and
does not yield a thermodynamic hydrogen ion activity.  Measurements of
pH in concentrated solutions may depart from the scale established with
the dilute buffers.  This could occur as result of changes in the liquid
junction potential.  Drift, slow response and unstable readings are
typically encountered for high liquid junction potentials.  Hence,
irreproducibility can indicate that a sample is not amenable to
measurement by Method 9040.

Potentiometric pH measurements for 13 solvents are presented in Table 2.
 Method 9040 requires that duplicate measurements agree to within 0.1 pH
units (6 mV).  Using the 0.1 pH units criterion, then readings for
methanol, pyridine and water would indicate amenability to measurement. 
Open circuit was only evident for the 1,4-Dioxane measurements.  A
precision requirement of around 0.05 pH units (3 mV) would discern water
from the other solvents.  The average pH values were not near the
corrosivity limits.

Table 2 presents measurements for six concentrated acids and dilutions
of these acids.  Multiple measurements



representing different days are reported for 70% HClO4 and 47% HNO3. 
Using the 0.1 pH units criterion, then the readings for concentrated
HClO4, HCl, HNO3, HOOCCH3, and CH3SO3H and dilutions thereof indicate
they were amenable to measurement by Method 9040.  Concentrated H2SO4
had a standard deviation of 0.17 pH units.  It is interesting to note
that 100% HOOCCH3 and 100% CH3SO3H, which supposedly contained no water,
were amenable to measurement.

Measurements for three concentrated bases and various dilutions are
given in Table 2.  Using the 0.1 pH units criterion, then the readings
for concentrated KOH and (CH3)4NOH and dilutions thereof indicate they
were amenable to measurement by Method 9040.  Concentrated 50% NaOH had
a standard deviation of 0.25 pH units, while 40% and lower met the
criterion.  The direct analysis of pellets of KOH had a standard
deviation of 0.34 pH units.

Dilution

It is common analytical practice for many parameters to dilute a sample
or preparation which contains an analyte in excess of the highest
calibration standard so that the dilution’s signal is within the
calibration range.  The dilution value is multiplied by a factor to
arrive at the sample or preparation concentrations.  Simple
proportionality is assumed.  Dilution would be a convenient means to
meet the bracketing requirements for strong acids and bases.  Common
commercial buffers than could be used.  However, pH can present
difficulties not associated with other parameters.  The logarithmic
scale, dissociation equilibria, activity coefficients and liquid
junction potential are rather unique to pH.  Nonetheless, Method 9040 is
silent on dilution.

Besides liquid junction potential, there is another affect attributed to
the nature of the solvent that is not error.  This is increased activity
of hydrogen ion often attributed to decrease in the dielectric of the
solution.  The liquid junction potential and dielectric effects can be
detected but not necessarily distinguished by dilution.  As a phase is
diluted with water the dielectric will be larger than the concentrated
sample hence the activity coefficient will better approximate that of
the calibration standards.  The liquid junction potential would decrease
with dilution eventually approximating the calibration standards.  For
acids, dilution would generally cause more dissociation.  Nevertheless,
the largest effect would be on the activity coefficient.  Use of
volumetric dilution would cause the activity of the concentrated
solution to be underestimated and hence, pH values to be biased toward
7.  Table 3 provides calculated pH values (10) for H2SO4 showing the
dilution corrected pH values underestimate the concentrated sample pH
value.  As shown, serial volumetric dilution in the standard buffer pH
range would cause about one pH unit change for a strong acid and the
same would be expected for a strong base.  About 0.5 pH unit change
would be expected for a weak acid or base.  The data in Table 3 shows
that the activity coefficient or activity of water approximate one for
the dilutions and the molal and molar concentration are essentially
equal, however neither is the case for the concentrated sample and hence
the reasons for underestimation of the activity.  Determining the
composition of the solution would allow better approximation by use of a
single ion activity coefficient or the activity of water and correct
conversion to a molal concentration.

Figure 1a shows the response of the Ross pH electrode with change in
weight percent for seven acids.  Fair linearity of the signal with
weight percent acid is observed with the strong acids on one line and
acetic acid, a weak acid, off by itself.  The signal for a pH 2 standard
is given as reference.  Figure 1b shows the response with change in
weight percent for three bases.  Again a linear response is observed for
KOH and (CH3)4NOH while the cross sensitivity to sodium ion causes the
curvature for NaOH.  The signal for a pH 12.45 standard is given as
reference.  As shown, all the various concentrated acids, KOH and
(CH3)4NOH and their dilutions gave signals that are discernible from the
corrosivity limits and hence measurable, contrary to the statement in
Method 9040.  As shown, dilution could be used to minimize the cross
sensitivity to sodium for NaOH.  For a saturated solution, the sample
should be filtered prior to dilution to assure that solids are not
dissolved by the dilution.

Assigning pH Values

Besides dilution, bracketing could be achieved by assigning pH values to
various strong acid or strong base secondary standard solutions.  The pH
values of the NIST/IUPAC primary standards have been established using
the primary method involving hydrogen gas electrode potentiometry with
the Harned Cell.  The assigned values show consistency with a
computation approach based on Debye-Huckel theory.  There exists mapping
back and forth between chemical potential and electrical potential. 
Reason for distinction of a primary and secondary standard can be
because of high liquid junction potential or inappropriateness of
Debye-Huckel theory.  Both are likely for concentrated solutions. 
Further, pH values assigned to secondary standards can be derived from
secondary methods such as glass electrode potentiometry in comparison to
the primary standards. (9)  Therefore, the pH value of a secondary
standard might be established empirically using glass electrode
potentiometry and need not be verified with a computation approach.

Figure 2a shows that pH values assigned as the negative logarithm of the
effective molal concentrations (m/aw) for H2SO4, HCl and KOH align well
with the calibration line of the commercial buffers.  The activity of
water (aw) was taken from Stokes (6) for H2SO4, from Licht (2) for KOH,
and calculated from osmotic coefficients (8) for HCl and NaOH.

Figure 2b allows comparison for the H2SO4 standards of the pH assigned
using the effective molal concentration, the molal concentration,
Licht’s method involving the mean activity coefficient and the molal
concentration (log(±)+ log(m)), and by Pitzer equations.  Figure 2c
shows the same information for KOH.

)m, where  is the fractional concentration of sulfate would move
the 0.25 m and 0.5 m pH values only slightly toward lower pH.   The
Pitzer equation and the molal concentration approaches show continuity
with the calibration line and agreement with the effective molal pH
value to about pH -1 and pH 14 or 15 but show more curvature at the
highest concentrations.

If the Pitzer pH values are the negative logarithm of the activity of
the hydrogen ion then it would appear the glass electrode response is
otherwise based.  Competing theories exist attempting to describe the
glass electrode function, including both faradaic (Nernst) and
non-faradaic (Boltzmann) potentiometry premised on ion exchange,
chemisorption, membrane transport, etc.  Dole’s a critical review of
which was given by Pungor [12,13] and more recently by Baucke [14–16],
that aim at describing of glass electrode functioning, and this fact
makes the finding of an exact explanation more difficult. 

sensitivity exist between models of low sodium error electrodes.  Glass
electrode manufacturers typically provide a sodium ion - ionic strength
– temperature nomogram for this correction.  However, these nomograms
are insufficient with regard to ionic strength and sodium molality. 
Further the cross sensitivity can change with use and age.  Figure 4
shows the response of the NaOH standards in comparison to the buffers
and the KOH standards.  The difference in the measured pH and effective
molal computed pH versus sodium concentration is given in Figure 5. 
This provides means to correction for unknowns when sodium is measured. 
Titration of hydroxyl ion can provide confirmation.  Another common
means to overcome interference would be to analysis dilutions of the
sample and compute the sample pH value from the dilution factor.  

Zavitsas (4) states; “In an ionic solution departures from ideality
can occur in two different ways.  Interionic attraction stabilizes an
ion resulting in a decrease of its activity coefficient whereas the
removal of solvent molecules due to solvation of the other ions has
destabilizing effect with a consequent increase of the activity
coefficient.  The ‘salting in’ effect is predominate at low ionic
strengths while at high ionic strengths the ‘salting out’ effect
dominates.”  It is well established that hydrogen ion in aqueous
solution exists in the hydrated form (H+ • nH2O).  Texts often show it
as the hydronium ion (H3O+) but evidence points to H9O4+ and more
recently H14O7+ as predominant.  Ions such as sodium, potassium, and
hydroxide also exist in hydrated forms.  These hydrated water molecules
are not free to interact as the solvent water molecules effectively
lowering the activity of water.  Commonly the activity of water is
measured by the vapor pressure in the presence and absence of solute.

gd1

Œ

Ž

ß

à

€

hß

h

# h1

h1

h1

h1

h

h1

h1

h

hMn

h

h

h¼

h

h

h

 ±)

Literature Cited

(1) Nordstrom, D.; Alpers, C.; Ptacek, C.; Blowes, D.  2000. Environ.
Sci. Technol. 34, 254-258.

(2)  Licht, S.  1985.  Anal. Chem. 57, 514-519.

(3)  Das, K. 1988.  J. of Sol. Chem. 17, 327-336.

(3)  Dole, M.  1932.  J. Am. Chem. Soc.  54, 3095-3105.

(3)  Baucke, F.  1994.  Anal. Chem. 66, 4519-4524.

(4)  Zavitsas, A. 2001.  J. Phys. Chem. B 105, 7805-7817.

(5)  Blandamer, M.; Engberts, J.; Gleeson, P.; Reis, J.  2005, Chem.
Soc. Rev. 34, 440-458.

(6)  Stokes, R.  1947.  J. Am. Chem. Soc.  67, 1686-1689.

(7)  Hansen, R.; Miller, F.; Christian, S. 1955. J. Phys.Chem. 59,
391-395.

(8)  Hamer, W.; Wu, Y.  1972.  J. Phys. Chem. Ref. Data 1, 1047-1099.

(9)  Buck, R.; Rondinini, S.; Covington, A.; Baucke, F.; Brett, C.;
Camoes, M.; Milton, M.; Mussini, T.; Naumann, R.; Pratt, K.; Spitzer,
p.; Wilson, G.  2002, Pure Appl. Chem. 74, 2169-2200.

(10)  Parkhurst, D.L.; Appelo, C.A.J.; 1999. User’s Guide to PHREEQC
(Version 2)—A Computer Program for Speciation, Batch-Reaction,
One-Dimensional Transport, and Inverse Geochemical Calculations. United
States Geological Survey, Water Resources Investigations Report 99-4259,
Washington, DC.

NEIC Applied Research, October 11, 2008

Draft

  PAGE  6  of   NUMPAGES  6 

NEIC Applied Research, October 11, 2008

DRAFT

  PAGE  1  of   NUMPAGES  6 

NEIC Applied Research, August 20, 2008

DRAFT

TABLE 1.  Concentrations of Common Acids

TABLE 2.  Measured Potentials and pH Values for Various Liquids and
Strongly Basic or Acidic Solutions

TABLE 3.  Simulation of the Effect of Dilution on pH for Sulfuric Acid
Bearing Sample

FIGURE 1.  Ross pH electrode signal as a function of concentration, a)
acids and b) bases.

FIGURE 2.  Signal as a function of pH for the standard buffers and for
solutions of strong acid or base with pH assigned, a) based on effective
molal, b) variously for H2SO4, and c) variously for KOH.

TABLE 4.  Comparison of measured pH values with the Standard Hydrogen
Electrode (3) and Calculated pH values. 

