`
`
`
`
`
`DrugDevIndPharmDownloadedfrominformahealthcarecombyLibraryCenteron06/24/15For
`
`DRUG DEVELOPMENT AND INDUSTRIAL PHARMACY, 14(14), 1891-1903 (1988)
`
`THE KARL FISCHER TITRATION OF WATER
`
`Kenneth A. Connors
`
`School of Pharmacy
`
`University of Wisconsin
`
`Madison, Wisconsin
`
`53706
`
`Abstract. The Kari Fischer method for the determination of water
`
`is briefiy reviewed.
`
`The chemistry of the reaction of Kari
`
`Fischer reagent with water is discussed, and modifications in the
`
`reagent composition are summarized. Some of these modifications
`
`resuit in more stabie reagents. The visual, spectrophotometric,
`
`and eiectrometric forms of the Kari Fischer titration are
`
`described and compared .
`
`INTRODUCTION
`
`In 1935 Fischer (1) described a titrimetric method specific
`
`for water.
`
`The titrant, which is now known as the Kari Fischer
`
`reagent,
`
`is a soiution of iodine, suifur dioxide, and pyridine in
`
`1891
`
`Copyright © 1988 by Marcel Dekker, Inc.
`
`HIIEHTE If
`
`Janssen Ex. 2012
`
`Lupin Ltd. v. Janssen Sciences Ireland UC
`|PR2015-01030
`
`(Page 1 of 13)
`
`
`
`
`
`personaluseonly.
`
`
`
`
`
`DrugDevIndPharmDownloadedfrominf0rmahea1thcare.c0mbyLibraryCenteron06/24/15For
`
`1 89 2
`
`CONNOR S
`
`methanol.
`
`In the half—century since its introduction,
`
`the Karl
`
`Fischer titration has become a standard procedure for the determi-
`
`nation of water in many types of samples. The subject has been
`
`repeatedly reviewed, most notably in a 1980 book, by Mitchell and
`
`Smith, devoted solely to this technique (2). This exhaustive
`
`treatment by Mitchell and Smith should be consulted for a compre-
`
`hensive survey of the literature on the analytical chemistry and
`
`applications of the Karl Fischer titration.
`
`The present review
`
`has the more limited aim of briefly describing the chemical basis
`
`of the method and of citing some recent work on modifications that
`
`may be of interest to the potential user of the Karl Fischer
`
`titration.
`
`CHEMISTRY
`
`The reaction of water with the Karl Fischer reagent is
`
`commonly written as this 2-step process:
`
`C5H5N’I2 +
`
`+
`
`+ H20 —-—>
`
`2c5H5NH*1‘ + C5H5N-S03
`
`C5H5N-S03 + CH3OH __. C5H5NH+CH3S04‘
`
`(2)
`
`The species C5H5N-I2, C5H5N-S02, and C5H5N-S03 may be charge-
`
`transfer complexes, and C5H5NH+I‘ and C5H5NH+CH3S04" are salts.
`
`The chemistry of the system is certainly more complicated than
`
`HtI‘.3H'I"E
`
`1;
`
`fl;
`
`Janssen Ex. 2012
`
`Lupin Ltd. v. Janssen Sciences Ireland UC
`|PR201501030
`
`(Page 2 of 13)
`
`
`
`
`
`personaluseonly.
`
`
`
`
`
`DrugDevIndPharmDownloadedfrominf0rmahea1thcare.c0mbyLibraryCenteron06/24/15For
`
`THE KARL FISCHER TITRATION OF WATER
`
`1893
`
`indicated by these two reactions, however, as demonstrated by the
`
`observation that the water titer of Karl Fischer reagent gradually
`
`decreases even in the absence of water, presumably due to the
`
`occurrence of side-reactions. Many products have been isolated
`
`from spent (i.e., exhausted) Karl Fischer reagent.
`
`(Surprisingly,
`
`chromatography does not yet seem to have been applied to the
`
`chemistry of this system in a detailed way, and this appears to be
`
`a promising area for further research). Kinetic studies show that
`
`the rate of loss of iodine is first—order each in iodine,
`
`in sul-
`
`fur dioxide, and in water.
`
`The reagent also contains the
`
`triiodide ion 13‘, formed in an equilibrium between iodine and
`
`iodide, and monomethyl sulfite, formed as follows:
`
`CH30H + 302 -T—‘‘ CH3S03H
`
`CH3OH + CH3SO3H :4 CH3SO3‘ + cH3oH2"
`
`(3)
`
`(4)
`
`The pH-rate behavior indicates that CH3S03‘ is the actual reactant
`
`in the redox reaction with water and iodine, as shown in Eq. (5).
`
`H20 + CH3S03‘ + I2 —» CH3S04‘ + 2HI
`
`(5)
`
`It is believed that pyridine serves as a buffer, maintaining the
`
`effective pH in a range so as to generate the reactive conjugate
`
`base, CH3S03-.
`
`Iltlfl H 1'3
`
`11
`
`ft;
`
`Janssen Ex. 2012
`
`Lupin Ltd. v. Janssen Sciences Ireland UC
`|PR2015-01030
`
`(Page 3 of 13)
`
`
`
`
`
`personaluseonly.
`
`
`
`
`
`DrugDevIndPharmDownloadedfrominf0rmahea1thcare.c0mbyLibraryCenteron06/24/15For
`
`1894
`
`CONNORS
`
`COMPOSITION OF THE REAGENT
`
`According to Eqs.
`
`(1) and (2),
`
`the stoichiometric ratios of
`
`the reactants are 1I2:1S02:1CH30H:3C5H5N, but the subsequent
`
`equations, as well as the existence of side reactions, result in
`
`depletion of some reactants, and the reagent is prepared so that
`
`the iodine is the limiting ingredient, a composition of
`
`1I2:3S02:10C5H5N dissolved in excess methanol being typical.
`
`The
`
`titer of Karl Fischer reagent is expressed as milligrams of water
`
`per milliliter of reagent. Typical titers are 3 to 6 mg H20/ml
`
`reagent for macroscale titrations, and a tenth of this for titra-
`
`tions on the microscale. Freshly prepared Karl Fischer reagent
`
`has a strength of about 80% of the theoretical, but this falls
`
`rapidly in the first 24-48 hours, so some authors advise that the
`
`reagent be prepared at least a day prior to its use. Methanol
`
`usually serves as the solvent for the titration sample.
`
`Many modifications of the standard composition have been
`
`proposed, most of these changes being replacements of the pyridine
`
`or the methanol. Karl Fischer titrant prepared with methyl cello-
`
`solve (2-methoxyethanol) is more stable (3) than the usual reagent
`
`containing methanol. Dimethylformamide (DMF)
`
`is also an effective
`
`replacement for methanol (4),
`
`though electrometric end point
`
`detection is required because the color change is not sharp.
`
`Replacement of the pyridine by other bases is an attractive
`
`possibility because of the unpleasant odor of pyridine.
`
`Some of
`
`the resulting reagents are so-called "one-component" reagents,
`
`HtI‘.3H'I"E
`
`1;
`
`fl;
`
`Janssen Ex. 2012
`
`Lupin Ltd. v. Janssen Sciences Ireland UC
`|PR201501030
`
`(Page 4 of 13)
`
`
`
`
`
`personaluseonly.
`
`
`
`
`
`DrugDevIndPharmDownloadedfrominf0rmahea1thcare.c0mbyLibraryCenteron06/24/15For
`
`THE KARL FISCHER TITRATION OF WATER
`
`1895
`
`that is, reagents containing the 12, S02 and base all
`
`in one
`
`titrant solution; "two-component“ formulas typically consist of
`
`the soivent, which is a solution of S02 and the base in methanol
`
`or other medium, and the titrant, consisting of iodine dissolved
`
`in methanol.
`
`The sample is dissolved in the solvent component,
`
`and this is titrated with the titrant component.
`
`For example, a
`
`typical
`
`two—component formula consists of sulfur dioxide and
`
`sodium acetate in methanol as the solvent, and iodine in methanol
`
`as the titrant. The standard composition can be used in the same
`
`way, S02 and pyridine dissolved in methanol being the solvent and
`
`iodine in methanol
`
`the titrant. Scholz has developed both one-
`
`component and two-component reagents in which pyridine is replaced
`
`with diethanolamine (5-7); 2-methoxyethanol
`
`is the solvent
`
`in the
`
`one-component solution, and methanol is the solvent for the two-
`
`component formulas. An unusual reagent is composed of urea,
`
`sodium salicylate, and sulfur dioxide in methanol (the solvent
`
`component) and iodine in methanol (the titrant) (8).
`
`Blomgren and Jenner (9) developed a Karl Fischer reagent with
`
`pyridine in which iodide was added to establish the ratio
`
`I’/I2=3.2. This composition gives a very stable reagent, even
`
`with methanol as the solvent.
`
`Two methods of preparation are
`
`described;
`
`in one of these the iodide is added as pyridinium
`
`iodide, C5H5NH+I', and in the other it is generated in situ by the
`
`addition of a calculated amount of water.
`
`Few analysts will choose to prepare their own Karl Fischer
`
`reagent. Several versions are commercially available,
`
`including
`
`HTIEHTE 1;‘
`
`fl?-
`
`Janssen Ex. 2012
`
`Lupin Ltd. v. Janssen Sciences Ireland UC
`|PR201501030
`
`(Page 5 of 13)
`
`
`
`
`
`personaluseonly.
`
`
`
`
`
`DrugDevIndPharmDownloadedfrominformahealthcarecombyLibraryCenteron06/24/15For
`
`1896
`
`CONNORS
`
`one-component and two-component formulas, pyridine-free reagents,
`
`and stabilized reagents.
`
`The compositions of these products are
`
`not usually specified.
`
`"Stabilized" reagents may be prepared with
`
`2-methoxyethanol
`
`instead of methanol, or they may be made up to
`
`the I‘/I2 ratio given by Blomgren and Jenner (9). The Hydranal®
`
`reagents are based on Scholz's formulas (5-7).
`
`In the coulometric titration method, the iodine is generated
`
`by oxidation of iodide.
`
`21' __. 12 + 2e
`
`(6)
`
`This is done in the presence of the other ingredients of the Karl
`
`Fischer reagent.
`
`The coulometric reagent therefore must contain
`
`iodide but no iodine, and even spent Karl Fischer reagent has been
`
`used. Since one molecule of water corresponds to one molecule of
`
`iodine (Eq. 5), which is equivalent to two electrons, 2F coulombs
`
`corresponds to 18.0 g of water, or 1 mg H20 is equivalent to 10.71
`
`coulombs.
`
`STANDARDIZATION OF THE TITRANT
`
`Liquid water is an obvious primary standard for the standard-
`
`ization of Karl Fischer reagent.
`
`Its disadvantages are its liquid
`
`state and its low molecular weight. Samples of water can be
`
`weighed out, or can be delivered by volume and the weight calcu-
`
`lated from the density.
`
`HtI‘.3H'I"E
`
`1;
`
`fl;
`
`Janssen Ex. 2012
`
`Lupin Ltd. v. Janssen Sciences Ireland UC
`|PR201501030
`
`(Page 6 of 13)
`
`
`
`
`
`personaluseonly.
`
`
`
`
`
`DrugDevIndPharmDownloadedfrominformahealthcarecombyLibraryCenteron06/24/15For
`
`THE KARL FISCHER TITRATION OF WATER
`
`1897
`
`More commonly a solution of water is quantitatively prepared
`
`in methanol, and aliquots of this solution are taken for titra-
`
`tion. This method has the advantage that a substantial sample of
`
`water is weighed (typically 10 to 15 g of H20 per liter of
`
`methanolic solution).
`
`It is essential that the same lot of
`
`methanol be used for the preparation of this standard solution and
`
`for blank titrations of the methanol to correct for traces of
`
`water in the methanol.
`
`n—Propyl alcohol has been recommended as
`
`the solvent because it is less volatile and less hygroscopic than
`
`methanol
`
`(10).
`
`Stable salt hydrates constitute another class of primary
`
`standard. Many of these have been studied, but only two have
`
`found wide use. One of these is sodium acetate trihydrate
`
`(theoretical water content 39.72%). This salt is soluble in
`
`methanol, but questions have been raised about the constancy of
`
`the water content.
`
`By far the most popular primary standard has been sodium
`
`tartrate dihydrate, Na2C4H405-2H20 (theoretical water content
`
`15.66%).
`
`Sodium tartrate dihydrate appears to be non—hygroscopic
`
`and quite stable under ordinary storage conditions.
`
`Its main
`
`drawback as a primary standard is its limited solubility in
`
`methanol, which does not generate confidence that all of its water
`
`of crystallization is released during the titration. Bryan and
`
`Rao (11) have made a careful study of the suitability of sodium
`
`tartrate dihydrate as the primary standard for Karl Fischer
`
`titration.
`
`They studied samples of sodium tartrate dihydrate from
`
`HTIEHTE
`
`1;)
`
`fl;
`
`Janssen Ex. 2012
`
`Lupin Ltd. v. Janssen Sciences Ireland UC
`|PR201501030
`
`(Page 7 of 13)
`
`
`
`
`
`personaluseonly.
`
`
`
`
`
`DrugDevIndPharmDownloadedfrominformahealthcarecombyLibraryCenteron06/24/15For
`
`1898
`
`CONNORS
`
`five commercial sources. obtaining concordant results. Drying the
`
`samples to constant weight yielded an actual average water content
`
`of 15.61%, or about 0.3% less than the theoretical. They make
`
`these recommendations for the most accurate work:
`
`1.
`
`The actual water content should be determined by drying'to
`
`constant weight.
`
`2.
`
`The titration should be carried out slowly so that all of the
`
`water of crystallization is released prior to the end point.
`
`3.
`
`The buret and titration assembly should be made with glass or
`
`stainless steel tubing, because rubber or plastic tubing is
`
`permeable to water vapor.
`
`4.
`
`Titration vessels should be purged with dry carbon dioxide
`
`rather than nitrogen.
`
`5.
`
`The buret and titration assembly should be protected from
`
`atmospheric moisture with phosphorus pentoxide adsorbent
`
`rather than calcium sulfate or magnesium perchlorate.
`
`END POINT‘DETECTIDN
`
`Visual Titration. Active Karl Fischer reagent has the reddish-
`
`brown color characteristic of iodine, whereas spent Karl Fischer
`
`reagent is yellow.
`
`It is therefore possible to carry out Karl
`
`Fischer titrations (in either the direct or the back-titration
`
`mode) with visual detection of the end point, the reagent serving
`
`as the indicator.
`
`These titrations are most successful on the
`
`macro scale with full-strength Karl Fischer reagent.
`
`(Reagent
`
`HTIEHTE
`
`1;
`
`fl;
`
`Janssen Ex. 2012
`
`Lupin Ltd. v. Janssen Sciences Ireland UC
`|PR201501030
`
`(Page 8 of 13)
`
`
`
`
`
`personaluseonly.
`
`
`
`
`
`DrugDevIndPharmDownloadedfrominf0rmahea1thcare.c0mbyLibraryCenteron06/24/15For
`
`THE KARL FISCHER TITRATION OF WATER
`
`1899
`
`prepared with DMF as the solvent tends to yield a poor visual end
`
`point and should not be used in this way).
`
`Because of the simplicity and low cost of the visual titra-
`
`tion method, it is recommended for occasional use or for prelimi-
`
`nary studies to establish feasibility.
`
`The only unusual
`
`precaution to be observed is protection from contamination by
`
`atmospheric moisture. Volumetric flasks are useful titration
`
`vessels because of their long necks, which restrict access by the
`
`atmosphere. Titration precision can be improved by titrating all
`
`solvent blanks, standardizations, and unknown samples to match the
`
`same reference color, which is maintained in a stoppered volu-
`
`metric flask.
`
`Simple directions have been published (12).
`
`Spectrophotometric Titration. Several
`
`instrumental methods make
`
`use of the difference in color of active and spent Karl Fischer
`
`reagent.
`
`Perhaps the simplest of these is a spectrophotometric
`
`titration in which the absorbance of the titration solution is
`
`measured as a function of titrant volume (13).
`
`In a direct titra-
`
`tion with Karl Fischer reagent the absorbance before the end point
`
`is low and nearly constant; after the end point the absorbance
`
`rises in accordance with Beer's law.
`
`The intersection of the
`
`straight lines marks the end point. Since the Karl Fischer
`
`reagent has an absorption spectrum showing rising absorption
`
`toward the shorter wavelength end of the visible region, without
`
`an absorption maximum,
`
`the choice of wavelength is not critical,
`
`and 525 nm was used in the original work.
`
`HTIEHTE 1;‘
`
`ill}-
`
`Janssen Ex. 2012
`
`Lupin Ltd. v. Janssen Sciences Ireland UC
`|PR201501030
`
`(Page 9 of 13)
`
`
`
`
`
`personaluseonly.
`
`
`
`
`
`DrugDevIndPharmDownloadedfrominformahealthcarecombyLibraryCenteron06/24/15For
`
`1 9 O0
`
`CONNORS
`
`The spectrophotometric titration method is applicable on the
`
`microscale, and it has the advantage that the instrumentation is
`
`available in every laboratory.
`
`It is not, however, readily
`
`applicable to large numbers of samples with present instrumenta-
`
`tion,
`
`though development of an efficient photometric Karl Fischer
`
`titrator seems to be feasible.
`
`Electrometric Titration.
`
`By far the most commonly used instru-
`
`mental method of Karl Fischer titration is the "dead-stop"
`
`amperometric technique with two platinum electrodes.
`
`A small
`
`potential
`(10-15 mv)
`is applied across these electrodes. and a
`galvanometer senses the current flow.
`To appreciate the chemistry
`
`taking place, consider a Karl Fischer back-titration,
`
`in which an
`
`excess of active Karl Fischer reagent is added to the sample and
`
`the unreacted reagent is back—titrated with standard water~in—
`
`methanol solution.
`
`Before the end point both iodine and iodide are present, and
`
`at the applied potential electrolysis can take place,
`
`I2 being
`
`reduced at one electrode and I’ being oxidized at the other, with
`
`a resultant flow of current. At
`
`(and after) the end point there
`
`is no I2 present, so there is no reversible couple, no electrol-
`
`ysis can occur, and the current drops to zero.
`
`In a direct Karl Fischer titration the current is zero before
`
`the end point, and it rises sharply after the end point.
`
`Mitchell and Smith (2) have discussed titration assemblies,
`
`electronics, and applications of dead—stop Karl Fischer titration.
`
`Many commercial versions of the apparatus are available.
`
`In their
`
`HIIEHTE 1;‘
`
`fl?-
`
`Janssen Ex. 2012
`
`Lupin Ltd. v. Janssen Sciences Ireland UC
`|PR201501030
`
`(Page 10 of 13)
`
`
`
`
`
`personaluseonly.
`
`
`
`
`
`DrugDevIndPharmDownloadedfrominf0rmahea1thcare.c0mbyLibraryCenteron06/24/15For
`
`THE KARL FISCHER TITRATION OF WATER
`
`1901
`
`simplest and least expensive form these consist of the electronic
`
`and titration vessel components,
`
`the titrant being delivered
`
`manually from a conventional buret. Highly sophisticated
`
`variations are also available in which the titrant addition rate
`
`and end point detection are microprocessor controlled, and the
`
`data are collected, analyzed, and stored by computer.
`
`Potentiometric end point detection with two dissimilar
`
`metallic electrodes operating at a constant current of 10-100 uA
`
`has also been used For Karl Fischer titration.
`
`A commercial pH
`
`meter can be used as the potentiometer.
`
`In coulometric Karl Fischer titration the iodine in the
`
`reagent is generated by electrolysis of iodide, and the extent of
`
`titration is measured in coulombs rather than in volume of
`
`titrant.
`
`The dead-stop amperometric method with two platinum
`
`electrodes is commonly employed as the method of end point
`
`detection.
`
`Because of the accuracy with which iodine can be generated
`
`and measured coulometrically, the coulometric Karl Fischer
`
`titration method may be the best choice for extremely small
`
`amounts of water. The dead-stop amperometric method is a
`
`reliable, rapid, and very widely applicable technique for most
`
`Karl Fischer titrations.
`
`APPLICATIONS
`
`Concentrations of water ranging from parts per million to
`
`100% can be determined by Karl Fischer titration, and amounts from
`
`HtI'.3H'I"E
`
`1;‘
`
`fl?-
`
`Janssen Ex. 2012
`
`Lupin Ltd. v. Janssen Sciences Ireland UC
`|PR201501030
`
`(Page11 of13)
`
`
`
`
`
`personaluseonly.
`
`
`
`
`
`DrugDevIndPharmDownloadedfrominformahealthcarecombyLibraryCenteron06/24/15For
`
`1902
`
`CONNORS
`
`micrograms to many milligrams can be determined. Most titrations
`
`are routine, but some samples may cause problems.
`
`A fairly common
`
`artifact that is observed is the occurrence of false end points or
`
`fading end points.
`
`If the sample is a nearly anhydrous nonpolar
`
`solvent it is not unusual
`
`to obtain a false end point, which may
`
`be due to the insolubility of the (polar) end products in the
`
`nonpolar medium; the solution to the problem is to incorporate
`
`some methanol or pyridine in the sample.
`
`Another problem arises if the sample contains substances that
`
`can react,
`
`in the strongly dehydrating environment of Karl Fischer
`
`reagent, to yield water, resulting in a fading end point as water
`
`is generated.
`
`For example, a carboxylic acid can react with
`
`methanol
`
`to yield an ester plus water; this type of interference
`
`can be overcome by rapid titration or by using one of the modified
`
`reagents.
`
`Yet another kind of problem occurs if the sample contains
`
`substances that can oxidize iodide to iodine, or that can react
`
`with iodine.
`
`For example, thiols are oxidized to disulfides by
`
`iodine.
`
`Somewhat elaborate countermeasures may have to be devised
`
`to circumvent such interferences (2).
`
`REFERENCES
`
`1. K. Fischer, Z. Angew. Chem., gg, 394 (1935).
`
`2.
`
`J. Mitchell, Jr. and D.M. Smith, “Aquametry," 2nd edition,
`
`Part 3, Wiley-Interscience, New York, 1980.
`
`HtI'.3H'I"E
`
`lg‘
`
`ill}-
`
`Janssen Ex. 2012
`
`Lupin Ltd. v. Janssen Sciences Ireland UC
`|PR201501030
`
`(Page 12 of 13)
`
`
`
`
`
`personaluseonly.
`
`
`
`
`
`DrugDevIndPharmDownloadedfrominf0rmahea1thcare.c0mbyLibraryCenteron06/24/15For
`
`THE KARL FISCHER TITRATION OF WATER
`
`1903
`
`E.D. Peters and J.L. Jungnickel, Anal. Chem., Q, 450 (1955).
`
`F.B. Sherman, Talanta, Q, 1067 (1980).
`
`E. Scholz, Am. Lab., August, 1981, p. 89.
`
`E. Scholz, Fresenius' 2. Anal. Chem., 309, 30, 123 (1981).
`
`E. Scholz, Analusis, 19, 491 (1982).
`
`
`M. Bos, Talanta, gr, 553 (1984).
`
`E. Blomgren and H. Jenner, U.S. Pat. 2,780,601 (1957).
`
`Mitchell and Smith, op. cit., p. 116.
`
`w.P. Bryan and P.B. Rao, Anal. Chim. Acta, gg, 149 (1976).
`
`K.A. Connors,
`
`"A Textbook of Pharmaceutical Analysis," 3rd
`
`edition, Wiley-Interscience, New York, 1932, pp. 508-509.
`
`K.A. Connors and T. Higuchi, Chemist—Analyst, 4_8, 91 (1959).
`
`10.
`
`11.
`
`12.
`
`13.
`
`HIIEHTE If
`
`Janssen Ex. 2012
`
`Lupin Ltd. v. Janssen Sciences Ireland UC
`|PR2015-01030
`
`(Page 13 of 13)