JOIlrt1aI 0/ ChromatograpJ，J'. 87 (1973) 125-136
`
`。ElsevicrScicnlific Publishing Company， Amslcrdom - Printcd in Thc Ncthcrlands
`
`CHROM.6922
`
`THE DETERMINATION OF PHYSOSTlGMINE BY THIN-LAYER CHRO-
`MATOGRAPHY AND ULTRAVIOLET SPECTROPHOTOMETRY.
`
`A. R. ROGERS and G. SMITH
`Dcpartment of Pharll/acJ'. Hcrlot-Watl Ull/vcrsll)'， Edlllbllrgl， EHl 2HJ (Great Brltall/)
`(Rcccivcd Junc 6th. 1973)
`
`SUMMARV
`
`Physostigmine has been separated from its degradation products by thin-Iayer
`chr-omatogrnphy on aluminn， with chloroform-acetone (5 :4) Q S thc solvent for dc-
`velopment. The alkaloid was eluted with methanolic hydrochloric acid and deter-
`mined by ultraviolet spectrophotometry. Two methods were used for the correction
`of irrelevant absorbance: a di町'erentialmethod in which absorbance measurements
`were made at three wavelengths， and a method in which orthogonal functions were
`applied to absorbance measurements at a set of nine wavelcngths
`
`INTRODUCT(ON
`
`Physostigmine (1) inhibi阻 theactivity of cholinesterase and is used in ophthal-
`molog:y as a miotic and to decrease intra-ocular pressure in glaucoma; for this purpose
`it is usually instil1ed into the eye as an aqueous solution containing U1) to 1% ofphy-
`sostigminc salicylate or sulphate. In aqueous so]ution. physostigmine hydrolyses to
`form a colourless phenoJic compound， eseroline (11); this compound is subsequently
`oxidised to rubreserine (111) and other coloured compounds. The anticholinesterase
`activity of the drug resides in the methyJcarbamate side-chain.
`
`Jqoc台:同町
`
`• This 'Yiork forms part of a thcsis submittcd by G. Smith for a Ph.D. dcgrce of thc Hcriot-
`Watt Univcrsity
`
`NOVARTIS EXHIBIT 2039
`Noven v. Novartis and LTS Lohmann
`IPR2014-00550
`Page 1 of 13
`
`

`

`126
`
`A. R. ROGERS， G. SMITH
`
`Bergt separated physostigmine (RF=O.74) from eseroline (RF=O.Sl) and ru-
`breserinc (RF=O.5S) by thin-Iayer chromatography (TLC) on silica gel with chloro-
`form-acetone-33% (w/v) dimethylamine in ethanol (S :4: 1) QS thc solvcnt for dcvcl-
`opment. The physostigmine was eluted wjth 0.1 N sodium hydroxide and the ru・
`breserine formed by hydroJysis and oxidation was determined co1orimetricalty at 480
`nm. Berg attempted to elute the alkaloid with various or且anicsolvents but the re-
`coveries were low.
`The method of Berg was criticised by SmithZ who reported that rubreserine
`reacted with dimethylamine within 10 min to form a yel10w product with an Rp (0.73)
`c10se to that ofphysostigmine. Smith separated physostigmine (RF=O.61-O.67) from
`esero1ine (RF=O.42-O.4S) and rubrescrine (RF=O.33-O.36) by TLC on alumina， with
`chloroform-acetone (S: 4) as the solvent for development At a high or a low relative
`humidity， the Rp value for rubrcscrine was less than 0.1 and “阻i1ing"was appreciabl。
`It was found necessary to store the alumina plates over a saturated solution of sodium
`bromide (relative hurnidity=S80/0) for 3 days before use. After separation of phy-
`閥 抗igminefrom i白 degradationproducts. Smith determined the alkaloid by direct-
`re伺ectancespectrophotometry i the coe冊cientof variation of 2S spots on 5 plates
`was 5.81%.
`The objective of the p問 sentwork was to develop an elution teclmiq田 forthe
`determination of physostigmine from thin・layerchromatograms.
`
`MATERIALS AND EQlJ【PMENT
`
`Materials
`Acetone. 8.S. 509; alumina G (Type E)， Merck; chloroform. AnalaRi hydro-
`chloric acid， laboratory reagent grade; methanol， spectroscopic gTade; physostigmine
`sulphate. B.P.C.; sodium bromide.laboratory reagent grade. Methanolic hydrochloric
`acid. B.P.C.. 1968. Appendix 7
`
`Equipmellt
`Ccntrifuge， Simplex (Martin Chris的 MicrometerSyringe， Agla【Burroughs
`Wel1come). Spectrophotometers. S.P. 500 and S.P. 800 (Pye Unicam). with a matched
`pair of 1・cmsilica cells. Whirlimixer (vortex mixer) (Fisons Scienti自cApparatus).
`Calculations were made with the aid of a desk.top computer， the Programma
`101 (British Olivetti)
`
`EXPERIMENTAL AND RESULTS
`
`Developmem 01 the elutio17 metllOd
`Alumina was spread in 0.25・m mlayers on glass plates (20 x 20 cm) and activated
`for 1 h llt 1100
`: the plates were stored over a saturated solution of sodium bromide
`for 3 days. To each plate. three 20-，u1 samples of a 0.5% aqueous solution of phy-
`sostigmine sulphate were applied as short streaks， by means of a micrometer syringe
`After dcvelopment with chloroform-acetone (5:4) and Jocation of the spots under
`screened ultraviolet (UV) radiation at 366 nm， areas (4 x 2 cm) of the alumina con-
`talnlng the physostigrninc were removed wlth a razor-blade and transferred into 10 ml
`ofthe solvent in a test.tube; the contents orthe tube were mixed for 2 min in a vortex
`
`NOVARTIS EXHIBIT 2039
`Noven v. Novartis and LTS Lohmann
`IPR2014-00550
`Page 2 of 13
`
`

`

`TLC AND UV SPECTROPHOTQMETRY OF PHYSOSTIGMINE
`127
`mixer and clarified by centrifu即 時 ( 11∞xg)forSm問 . The absorbance of the de~
`can臼 d solution was measured from 220 to 360 nm. A spcctrophotometric blank
`soIution wns prepored by removing n bJank nrea (4)( 2 cm) of thc odsorbcnt at a
`location corresponding to the Rp value of physostigmine. and treating this blank
`adsorbent in a similar manner to the adsorbent containing the drug. The di宵erence
`between the ahsorbance at 244 nm (maximun吟 andthat at 267 nm (minimum) was
`taken as a measure of the amount of physostigmine
`1n preliminary experiments. four solvents were used to clutc physostigmine. Of
`these solvents， water (recovery 40-64%)， 0.1 N hydrochloric acid (rccovery 58-87%)，
`and methanol (recovery 61-84%) all gave results t11at were Jow and erratic. The
`solvent of chojce W8S methanolic hydrochloric acid for which the recovcry of physo-
`stigmine W8S 89-103%.
`A po日 ibJesou四 eof variation in the absorbance of elu同 ddrug， especially at
`wavelengths lower than 250 nm. was variation in the absorbance of the blank. In an
`attempt to reduce this irrelevant absorbance. plates were prepared with alumina that
`had been washed盟国twjt}可 chloroform-acetone(5:4)， then witb methanoJJ and finally
`with boiling water to remove traces of the organic solvents. The alumina was thcn
`filtercd through sintered glass and was:hed repeatedly with cold watcr before prepara-
`tion of the pJates. The results of this treatment are given in Table I.
`The treatment of the alumina considerably reduced the absorbance ofthe blank
`but the results were stilJ variable. Jn aD attempt to further reduce the absorbance of
`the blank， various membrane創 tcrswe問 usedto clarify曲 esolution (Spencer and
`Beggs3) but these制 le目 wereunsatisfactory. Millipore GS (cellulose esters)， Celotate
`(ceJJulose acetatc). and Duralon (nyJon) werc aJJ attacked by the methanolic hydro-
`chloric acid， The pore size of Polyvic (polyvinyl chJoride) and MiterσTFE) was 10。
`
`Jarge for the removal of the aJumina particles. Repeated centrifuging of the eluted
`solution did not reduce the absorbance of the blank
`Consideration was then given to the possibiJity that variation in the absorbance
`
`。r曲 ebJanks could be due partly to impurities from the solvent used for development
`
`of the plates. The absorbance of blanks at locations corresponding to the RF of
`physostigmine was determined for由 開edeveloped pIates and for three undeveloped
`plates. The mean absorbances at 221. 244 and 267 nm are given in Table 11
`From the resul
`
`TABLEI
`EFFECTS OF WASHING THE ALUMINA ON THE ABSORBANCE OF BLANK SOLU-
`TIONS
`
`2門'reatment01
`a[umlna
`
`Area 0"
`plale
`
`Absorballce
`
`221 nm
`
`244nm
`
`26711附
`
`間一四一山蜘問仰
`
`のuwnu自unuAUAHV
`
`0.127
`0.086
`0.055
`0.012
`0.043
`0.027
`
`0.085
`0.063
`0.041
`0.008
`0.036
`0，010
`
`123123
`
`Not w8s:hed
`
`Washed
`
`NOVARTIS EXHIBIT 2039
`Noven v. Novartis and LTS Lohmann
`IPR2014-00550
`Page 3 of 13
`
`

`

`NOVARTIS EXHIBIT 2039
`Noven v. Novartis and LTS Lohmann
`IPR2014-00550
`Page 4 of 13
`
`

`

`TLC AND UV SPECTROPHOTOMETRY OF PHYSOSTIGMINE
`
`129
`
`at 244 nm. Measure the absorbancc at 221， 244 and 267 n m of a 0.001% solution of
`physostigmine sulphate in methanolic hydrochloric acid， and calculate the rccovery
`of physost唱 mincfor cnch snmplc
`This method was applied to three 20-，μ1 samples of a 0.5% solution of physo-
`
`TABI，;E III
`
`ABSORBANCE OF PHYSOSTlGMINE SULPHATE DETBRMINED BY ELUTlO N FROM
`THIN-LAYER CHROMATOGRAMS・
`
`PJale
`
`Area
`011 p!atc
`
`Absorbollcc 01 p"ysosllgml"e slIlpllale
`
`A"" AI4‘ nm-Q.S(AUl侃 "，+.1，・7""，)
`
`Rcco阿 ryof
`pltysosl/gmlnc (%)
`
`101.8
`101.2
`J01.5
`97.6
`92.6
`95.3
`98.S
`100.6
`100.0
`100.6
`94.4
`104.4
`97.9
`93.S
`94.4
`9Z.6
`91.2
`91.8
`
`0.346
`0.344
`0.345
`0.332
`0.315
`0.324
`0.3035
`0.342
`0.340
`0.342
`0.321
`0.355
`0.332
`0.318
`0.321
`0.315
`0.310
`0，312
`
`1231231233231231Z3
`
`A
`
`B
`
`C
`
`D
`
`E
`
`F
`
`• Absorbancc of O.田 t% physostigminc sulphatc in mctt附 10lichydrochloric acid = 0.340;
`mcan absorbancc of clutcd physostigminc = 0.330: mcan rccovcry of C1U1Cd physostigmlne =
`リ7.2%:coc町 cicntof variation =-4.2 ラ~ (17 dcgrccs of frccdolT仏
`
`。"'1
`
`0.4
`
`3 割
`o a c
`muEM盟〈
`
`M
`
`_.イ?ーーー
`一--;;，:;;.;~，，::・『包
`""t('J ~dO 盟問 3回
`川畑velength(nm)
`
`32。
`
`340
`
`言否。
`
`Fig. 1. PJ叫 。rabsorbancc against wavclcngth for a 0.001 % solu!ion of physos!igminc sulphatc
`and for two blank solutions.ーーー.Physoulgminc sUlphatc;ーーーI blanks
`
`NOVARTIS EXHIBIT 2039
`Noven v. Novartis and LTS Lohmann
`IPR2014-00550
`Page 5 of 13
`
`

`

`130
`
`A. R. ROGERS. G. SMJTH
`
`stigmine suJphate on each of six plates目 Theresu1ts arc givcn in Table III. The ab-
`sorbance from 210 to 360 nm ofa 0.001% 50lution ofphysostigmine sulphate and that
`of typical blanks arc il1ustrntcd in Fig. 1
`From these results variation in recovery ofthe drug was relatively h唱 hbetween
`plates but lower between areas 00 the same pJate. The elution of similar weights of
`alumina rather than similar areas also gave a relatively high variation in recovery of
`the drug from di町'erentplates， It was decidcd th8t in further work a standard solution
`should bc app1ied to each plate and thc rccovery of drug calculated from the absor-
`bance of this standard 501utioo
`
`COl1firmatioll 01 t/te悶 Udity01 ，he Lamb(!rt-Beel・law/or the elutioll method
`
`The elution method was applied to three 20-，u1 samples of a 0.1% aqueous
`solution of physost叩 叩nesulphate on。 問 platc，together with one 20-，μ1 sample of a
`standard solution containing 0.5% of physostigmine sulphate. The experiment was
`repeated on 0.2， 0.3. 0.4 and 0.5% solutions of physostigmine sulphate together with
`the standard solution. For each sample solution. the mean absorbance， based on 3
`waveJengths， was corrected for recovery of the drug by reference to the absorbancc
`。fthe standard solution applied on the same plate. The results a同 givenin Table IV.
`
`TABLEIV
`ABSORBANCE OF 0.1 TO 0.5% SOLUTIONS OF PHYSOSTlGMINE SULPHATE DE-
`TERMINED BY THE ELUTION METHOD
`
`COl1cellf叩 "011
`of physostigmulc
`sulphatq (%)
`
`ルtfcal1absorbQl1cc Mcan absorba/lcc，
`corrccled for recovcry
`011 ，he plale
`
`0.053
`0.100
`0.188
`0.246
`0.327
`
`0.060
`0.123
`0.200
`0.261
`0.339
`
`12345
`00000
`
`By regression analysis， the slope ofthe calibration graph of absorbance (A) as a
`function of percentage concentration (c) was calculated to be 0.679. The standard
`error of the slope was 0.014. The equation for the regression of absorbance upon
`concentration was represented by:
`A =0.679 c-0.005
`In a further experiment， it was shown by two-dimensional TLC that no decrease
`in content of physostigmine occurred after chromatography in the second dimension.
`
`Correction of lrreleν'an/ absorbance by application of orthogonalfuncllons
`In the method used in the previous experiment.for calculation ofthe absorbance
`due to physostigmine it was assumed that the graph of the absorbance of the blank
`against wavelength was Jjnear. Forthe blanks examined， the graph was approximately
`Jinear but curved very sllght1y at wavelengths below 240 nm
`As aD aJternative method of cor四 ctionfor irrelevant absorbance of the btank，
`
`NOVARTIS EXHIBIT 2039
`Noven v. Novartis and LTS Lohmann
`IPR2014-00550
`Page 6 of 13
`
`

`

`γLC AND UV SPECTROPHOTOMETRY OF PHYSOSTIGMINE
`
`131
`
`attempts were made to apply orthogonal functions (Glenn勺.The aim of preliminary
`cxperiments was to choose the orthogonal polynomial， number and intervals of wave.
`Icngths and thc mcan wavclcngth so that thc contribution of the absorbancc of thc
`physostigmine was a maximum and the contribution of the absorbancc of the blank
`was a minimum
`The absorbance of a 0.001% solution of physostigmine sulphate in water was
`measured at 1・nmintervals from 215 to 350 nm with a manual spectrophotometer.
`The absorbances of blank solutions eluted from alumina were also measured.
`Choice qf polJmomial range atrd IIumber ofwal'elellgths. Since the shapes of the
`absorbance curves of physostigmine sulphate and of the hlank solution were not
`complex. it was considered that a small number of wavelengths would be adequate
`for correction of irrelevant absorbance. From the results of preJiminary experiments，
`nine wavelengths over the range 221-267 nm were chosen
`Calculations of the cocfficients (Po. Pt. pz，. Pl. P... and P!5) of thc orthogonal
`
`functions (polynomials Po・ P1 • P1.' P3. P4 and Ps) were made for a O.∞1% solution
`
`of physostigmine sulphate and for two blank solutions， Tables of orthogonal poly.
`nomials by Fishcr and Yate♂ were used. To iIIustrate the method of calculation
`details ofthe results and calculations are given in Table VI in rcspectofthc coe而cients
`Po and pz， of the physostigmine sulphate solution and one blank solution. The ratios
`(percentages) of the coefficicn白 (Po.PI' pz，. PJ' P4 and pρfor the two blank solutions
`in relation to thc coc冊 目en臼 forthe soh刷 00of physostigmine sulphate are given
`in TabJe V.
`
`TABLE V
`RATIOS (PERCENTAGES) FOR THE COEFFICIENTS OFORTHOOONAL FUNCTIONS
`FOR TWO BLANK SOLUTIONS IN RELATION TO THE COEFFICIENTS FOR A 0.001 %
`SOLtJTlON OF PHYSOSTJGMJNE SULPHATE
`
`Blank
`soluffol1
`
`Pt blank x 100
`Pt plJ)・'sosligmilloslIlpltate
`p，
`p，
`
`p，
`
`p，
`
`p，
`
`p，
`
`“A"
`-ーB'・
`
`30.31
`28.30
`
`26φ98
`25.92
`
`0，04
`0.97
`
`7.74
`9.05
`
`0.03
`0，99
`
`16.31
`11.42
`
`From these results. the contribution of the irrelevant absorbance of the blanks
`is reduced to less than J% if the coefficient pz， or P4 is caJculated for the orthogonal
`function Pz， or P.... Ofthese two functions. Pz was chosen for further work because thc
`calculated coefficlent pz， (6.00) for a 1 % solutlon ofthe drug was hlgher than the coef-
`ficient P4 (2.65)
`
`Choice 01 mean wal'elengJh and h，tervals 01 waveleng，h. The coefficients Pl for
`
`thc 0，001 % solution of physostigmine sulphate and for a blank solu旬 0"were cal.
`cula旬 dfor mean wavelengths of235印 255nm. in sets ofnine wavelengths at intervals
`o( S nm. Convoluted absorbance curves o( thc Pz， coe田cicntagainst mcan wavelength
`are ilIustratcd in Fig. 2
`
`NOVARTIS EXHIBIT 2039
`Noven v. Novartis and LTS Lohmann
`IPR2014-00550
`Page 7 of 13
`
`

`

`NOVARTIS EXHIBIT 2039
`Noven v. Novartis and LTS Lohmann
`IPR2014-00550
`Page 8 of 13
`
`

`

`TLC AND tJV SPECTROPHOTOMETRY OF PHYSOSTIGMINE
`.，
`
`.J:F一一 MeQn wavelef"19\h肺，，'
`
`手
`
`2'
`
`133
`
`羽 田
`bb. 句
`
`70
`
`'
`solution of phν'50S“gminc $ulphatc and for a blnnk solution.ー一一， PhY50stigminc 5ulphatci
`:Fig. 2. Plot of pa coefficicnt aSllinst mCQn wlLvclcngth or thc sct of ninc wavclcngjhs， ror a 0田 1%
`ーーー.blanks.
`Intervllrs。納 得:ve明 gthtnrTリ
`。
`
`2
`
`3
`
`4
`
`S
`
`6
`
`7 8
`
`咽 回 目 | 岨 叩
`
`も官占
`
`-咽
`Fig. 3. Plot of Pt cocfficicnt Qgainst intcrvals of wDvclcngth for thc sct of ninc wQvclcngths， for a
`0.001 % solution of physostigminc sulphatc.一一ー， Physostigminc sulphatc;ーー-， blnnks.
`
`Since the contribution ofthep2 coe田cientofthe blank solution was a minimu田
`for a mean wavelength of 246 nm，出atwavelength was chosen for further work. The
`coefficientp2 was then calculated for intervals of I to 8 nm in sets of nine wavelengths
`of mean 246 nm. A convo:1uted absorbance curve ofthe P2 coefficient against interval
`of waveJength is illustrated in Fig. 3
`From these results. S nm was cbosen as the interval of wavelength since the P2
`coefficient of the solution of physostigmine sulphate was a maximum at曲目 interval
`ofwavelength. For each interval the values of thc Pl coefficients of the two bJank
`solutions were less than 1% ofthe coefficicnts of the physostigmine sulphate solution.
`
`Determination of the ac叩 racyalld reproducfbility of the elution method by application
`%rthogonalfi削 ctiolJS
`Orthogonal functions (P:) were applied to the absorbance at 226. 231.236，241.
`246， 251， 256， 261 and 266 nm of the cluted solutions obtained in the ex.periment for
`determination of the accuracy and問l'roduclbl1ltyof the elutlon method σable 1I1).
`The norma1ised coe慣cients(pz) we問 calculatedand the results are given in
`Table VII
`
`Co，!斤'rmationof the四 Iidilyof thc LQmbert~Beer law for the elll1ion met!tod by appJi.何 -
`tion q( o，.thogona{ functloflS
`Orthogonal functions (P:) were applied to thc absorbances at 226， 231， 236， 241.
`
`NOVARTIS EXHIBIT 2039
`Noven v. Novartis and LTS Lohmann
`IPR2014-00550
`Page 9 of 13
`
`

`

`NOVARTIS EXHIBIT 2039
`Noven v. Novartis and LTS Lohmann
`IPR2014-00550
`Page 10 of 13
`
`

`

`TJ..C AND UV SPECTROPHOTOMETRY OF PHYSOSTIGMINE
`
`IJS
`
`as a function of pe四 entageconcentration (c) was calculatcd to be -J 24.9. The
`standard error of the slope was 2.08. The equation for the regression of the pz coef.
`ficicnt upon concentrn針。nWD5 rcprc5cnted by:
`
`y-ー 124.9c+1.7
`
`DISCUSSION
`
`Orthogonal funct旧 nswere calculated for the absorbance at nine waveJengths
`over the range 226-266 nm of the eluted solutions obtained in the experiment for
`determination of the accuracy and reproducibility of the elution method. Thus the
`absorbances in Table 111 can be compared directly with thepz coefficien白 川 TableVlI.
`There is no significant difference between the mean results for rccovery of physostig“
`mine (97.2% and 96.8%)， and the coefficients of variation of the means are of similar
`magnitude 付.2% and 4.1%).
`The reproducibility ofthe elution method appears to be slightly better than that
`of the direct-reflectance method2 (coefficient of variation. 5.81%) but the technique
`of elution is laborious and slow. Both methods give more reproducible resuIts than
`the gas chromatographic method of Teare and Bors♂ for physostigmine salicylate
`(coefficient of variation， 11.5%)
`The application of 0比hogonalfunctions to thc elution method has provided an
`alternative means for the corrcction of irrelevant absorbance in the eluted solutions
`but the results are no better than thosc ob凶 nedby appJication of the method of
`Ganshirt and Morianzs. Since the graph of the absorbance of the blank against
`wavelength curves very slightly it might bc expec臼 dthat the method of Ganshirt and
`Morianz， in which Iincority is assumcd. would incur grcatcr crro四 Thcrcarc， how-
`ever. several possible sourccs of error in the applica針。nof orthogonal functions to the
`determination of physostigmine sulphate
`A comrnon source of error in all spectrophotometric methods is an overall shift
`in the wavclength-scale of the spectrophotometer. 10 the application of orthogonnl
`functions for correction of irrelevant absorbanco， this error depends upon the slope
`ofthe convoluted absorbance curve of the coefficicnt as a function of the mean wave-
`length of the set. Examination of Fig. 2 shows that a shift of +0.5 nm from the mean
`wavelength of 246 nm問 sultsin a 1.8% decrease in the pz coefficient; a shift of -0.5
`nm resu1ts in a 2.3% increasc in the P2 coefficient
`Errors may also be incurred in setting the wavelength for measurement of ab-
`sorbance， especiaUy where there is a steep slope in the absorbance curve. Thus errors
`are more Jikely to occur in the rneasurement of the absorbance of phys田 tigmine
`sulphate at 236 and 256 nm， where the slopes are steep. than at the peak at 246 nm
`(Fig. 1)
`The prccision ofthe coefficient ofthe orthogonal polynomial may al50 affect the
`results. Jn a contribution by Glenn， quoted by W:
`
`where Nj is the normalising factor.
`
`NOVARTIS EXHIBIT 2039
`Noven v. Novartis and LTS Lohmann
`IPR2014-00550
`Page 11 of 13
`
`

`

`The coe偲cientof variation of the comparative coefficient. c.v. qj is represented
`
`A. R. ROGERS， G. SMITH
`
`136
`
`by:
`
`C.V. 1Q，l "'"一一一Iq，j
`
`100 s(A)
`
`wherc s{A) is the estimated standard deviation of the absorbance values: it is assumed
`to be constant for the 5et of wavelengths
`Glenn has suggcsted a 白・rulcof thumb" that can bc used for cstimat旧 nof the
`precjsion of 3n)' coefficient for any number of wavelengths. Where the mean wave-
`length ofthe set corresponds to a maximum or minimum in the convolutcd absorbance
`curve， s(A) is assumed to be 1.4 x 10→. )f the coefficient of variation is to be less than
`1%. then the comparative coefficient !q，i. must exceed 0.14. Where the mean wavelength
`of the 5et does not correspond with a maximum or minimum. s(A) is assumed to be
`6.3 X 10-3. Under these c∞。ondi比巾ion悶s丸， the c∞。mpar悶a凶tivcc∞。e而 ci怜en叫t丸，Iq，l， mu附 』
`'げfthec∞。efficie町聞nt、首tof variatioo i泊st佃。 beless thao 1%.
`111 the detcrmination ofa 0.001% solution of physostigmine sulphate， the com・
`parative coefficient， !qz!. is 0.32. Since the mean wuvelength of the set (246 nr叫 does
`not correspond with the maximum of the convoluted absorbance curve at 243 nm，
`Iq21 should exceed 0.63 if the coefficieot of variation is to be less than 10/0' If it is
`assun叩 dthat s(A) is 6.3 x 10・3，then thc estimated coe而cientofvariation is about 2%
`111 the present work， Glenn's“rule of tl1umb" appears to provide a realistic
`estimate of the precision of the coefficient of the ortho畠onalpolynomial.
`The apparent complexity of the mathematical theory of orthogonal functions
`has obscured tl可esimplicはYof the calculations used in thcir application for the cor.
`rectioll ofirrelevant absorbance. The method mcri臼 furthertrial in thc spcctrophoto-
`metric anaJysis of drugs eluted from thin・layerchromatograms， especially with the
`increasing use of calculators and desk computers that ennble thc calculations to be
`completcd within a few mInutes
`
`REFBRBNCBS
`
`1 B. H. Bcrg， Acta Phal'lI1. Sm:cica. 3 (1966) 209.
`2 G. Smith. Proc. Soc.ペlIal.Cltcm.. 8 (1971) 66.
`3 R. D. Spcnccr and B. H. Bcggs守 J.Chromafogr.. 21 (目白)52
`4 E. J. Shcllard and M. Z. Alam. J. Cltl'omatog，..， 32 (1968)制 2.
`5 H. Giinshirt and K. Morianz， A，.ch. Pharm. (Ber/ln). 293 (1960) 1065.
`6 A. L. Glcnn， J. Pharm. Pharmacol.， 15【1963)123T.
`7 R. A. Fishcr nnd F. Y副 田，Sfati.rt!ca/ Tables lor slologlcal， Agrlcl/lwra/ al/d Medlcal Rcs四 ，.cll.
`Olivcr & Boyd. Edinburgh.出 hcd.， 1963
`8 F. W. Tca問 andS. 1. B。 附，J. Pha"'". Pllat'macol.， 21 (1969) 277.
`9 A. M. Wahbi. .Ph. D. Th目 is.Univcrsity of London， London， 1967
`
`NOVARTIS EXHIBIT 2039
`Noven v. Novartis and LTS Lohmann
`IPR2014-00550
`Page 12 of 13
`
`

`

`“This Copy of Copyrighted Material was made and delivered under license from Copyright
`Clearance Center, Inc. No further reproduction is permitted.”
`
`NOVARTIS EXHIBIT 2039
`Noven v. Novartis and LTS Lohmann
`IPR2014-00550
`Page 13 of 13
`
`