\begin{filecontents}{CZ4f3.eps} %!PS-Adobe-2.0 EPSF-2.0 %%Title: CZ4f3.eps %%Creator: fig2dev Version 3.2 Patchlevel 1 %%CreationDate: Thu Jun 15 16:02:50 2000 %%For: cachia@cptpc26 (Vincent Cachia thesard Zagrebnov) %%Orientation: Portrait %%BoundingBox: 0 0 362 362 %%Pages: 0 %%BeginSetup %%EndSetup %%Magnification: 1.0000 %%EndComments /$F2psDict 200 dict def $F2psDict begin $F2psDict /mtrx matrix put /col-1 {0 setgray} bind def /col0 {0.000 0.000 0.000 srgb} bind def /col1 {0.000 0.000 1.000 srgb} bind def /col2 {0.000 1.000 0.000 srgb} bind def /col3 {0.000 1.000 1.000 srgb} bind def /col4 {1.000 0.000 0.000 srgb} bind def /col5 {1.000 0.000 1.000 srgb} bind def /col6 {1.000 1.000 0.000 srgb} bind def /col7 {1.000 1.000 1.000 srgb} bind def /col8 {0.000 0.000 0.560 srgb} bind def /col9 {0.000 0.000 0.690 srgb} bind def /col10 {0.000 0.000 0.820 srgb} bind def /col11 {0.530 0.810 1.000 srgb} bind def /col12 {0.000 0.560 0.000 srgb} bind def /col13 {0.000 0.690 0.000 srgb} bind def /col14 {0.000 0.820 0.000 srgb} bind def /col15 {0.000 0.560 0.560 srgb} bind def /col16 {0.000 0.690 0.690 srgb} bind def /col17 {0.000 0.820 0.820 srgb} bind def /col18 {0.560 0.000 0.000 srgb} bind def /col19 {0.690 0.000 0.000 srgb} bind def /col20 {0.820 0.000 0.000 srgb} bind def /col21 {0.560 0.000 0.560 srgb} bind def /col22 {0.690 0.000 0.690 srgb} bind def /col23 {0.820 0.000 0.820 srgb} bind def /col24 {0.500 0.190 0.000 srgb} bind def /col25 {0.630 0.250 0.000 srgb} bind def /col26 {0.750 0.380 0.000 srgb} bind def /col27 {1.000 0.500 0.500 srgb} bind def /col28 {1.000 0.630 0.630 srgb} bind def /col29 {1.000 0.750 0.750 srgb} bind def /col30 {1.000 0.880 0.880 srgb} bind def /col31 {1.000 0.840 0.000 srgb} bind def /col32 {0.733 0.741 0.800 srgb} bind def end save -179.0 469.0 translate 1 -1 scale /cp {closepath} bind def /ef {eofill} bind def /gr {grestore} bind def /gs {gsave} bind def /sa {save} bind def /rs {restore} bind def /l {lineto} bind def /m {moveto} bind def /rm {rmoveto} bind def /n {newpath} bind def /s {stroke} bind def /sh {show} bind def /slc {setlinecap} bind def /slj {setlinejoin} bind def /slw {setlinewidth} bind def /srgb {setrgbcolor} bind def /rot {rotate} bind def /sc {scale} bind def /sd {setdash} bind def /ff {findfont} bind def /sf {setfont} bind def /scf {scalefont} bind def /sw {stringwidth} bind def /tr {translate} bind def /tnt {dup dup currentrgbcolor 4 -2 roll dup 1 exch sub 3 -1 roll mul add 4 -2 roll dup 1 exch sub 3 -1 roll mul add 4 -2 roll dup 1 exch sub 3 -1 roll mul add srgb} bind def /shd {dup dup currentrgbcolor 4 -2 roll mul 4 -2 roll mul 4 -2 roll mul srgb} bind def /reencdict 12 dict def /ReEncode { reencdict begin /newcodesandnames exch def /newfontname exch def /basefontname exch def /basefontdict basefontname findfont def /newfont basefontdict maxlength dict def basefontdict { exch dup /FID ne { dup /Encoding eq { exch dup length array copy newfont 3 1 roll put } { exch newfont 3 1 roll put } ifelse } { pop pop } ifelse } forall newfont /FontName newfontname put newcodesandnames aload pop 128 1 255 { newfont /Encoding get exch /.notdef put } for newcodesandnames length 2 idiv { newfont /Encoding get 3 1 roll put } repeat newfontname newfont definefont pop end } def /isovec [ 8#200 /grave 8#201 /acute 8#202 /circumflex 8#203 /tilde 8#204 /macron 8#205 /breve 8#206 /dotaccent 8#207 /dieresis 8#210 /ring 8#211 /cedilla 8#212 /hungarumlaut 8#213 /ogonek 8#214 /caron 8#220 /dotlessi 8#230 /oe 8#231 /OE 8#240 /space 8#241 /exclamdown 8#242 /cent 8#243 /sterling 8#244 /currency 8#245 /yen 8#246 /brokenbar 8#247 /section 8#250 /dieresis 8#251 /copyright 8#252 /ordfeminine 8#253 /guillemotleft 8#254 /logicalnot 8#255 /endash 8#256 /registered 8#257 /macron 8#260 /degree 8#261 /plusminus 8#262 /twosuperior 8#263 /threesuperior 8#264 /acute 8#265 /mu 8#266 /paragraph 8#267 /periodcentered 8#270 /cedilla 8#271 /onesuperior 8#272 /ordmasculine 8#273 /guillemotright 8#274 /onequarter 8#275 /onehalf 8#276 /threequarters 8#277 /questiondown 8#300 /Agrave 8#301 /Aacute 8#302 /Acircumflex 8#303 /Atilde 8#304 /Adieresis 8#305 /Aring 8#306 /AE 8#307 /Ccedilla 8#310 /Egrave 8#311 /Eacute 8#312 /Ecircumflex 8#313 /Edieresis 8#314 /Igrave 8#315 /Iacute 8#316 /Icircumflex 8#317 /Idieresis 8#320 /Eth 8#321 /Ntilde 8#322 /Ograve 8#323 /Oacute 8#324 /Ocircumflex 8#325 /Otilde 8#326 /Odieresis 8#327 /multiply 8#330 /Oslash 8#331 /Ugrave 8#332 /Uacute 8#333 /Ucircumflex 8#334 /Udieresis 8#335 /Yacute 8#336 /Thorn 8#337 /germandbls 8#340 /agrave 8#341 /aacute 8#342 /acircumflex 8#343 /atilde 8#344 /adieresis 8#345 /aring 8#346 /ae 8#347 /ccedilla 8#350 /egrave 8#351 /eacute 8#352 /ecircumflex 8#353 /edieresis 8#354 /igrave 8#355 /iacute 8#356 /icircumflex 8#357 /idieresis 8#360 /eth 8#361 /ntilde 8#362 /ograve 8#363 /oacute 8#364 /ocircumflex 8#365 /otilde 8#366 /odieresis 8#367 /divide 8#370 /oslash 8#371 /ugrave 8#372 /uacute 8#373 /ucircumflex 8#374 /udieresis 8#375 /yacute 8#376 /thorn 8#377 /ydieresis] def /Times-Roman /Times-Roman-iso isovec ReEncode /Times-Italic /Times-Italic-iso isovec ReEncode /DrawEllipse { /endangle exch def /startangle exch def /yrad exch def /xrad exch def /y exch def /x exch def /savematrix mtrx currentmatrix def x y tr xrad yrad sc 0 0 1 startangle endangle arc closepath savematrix setmatrix } def /$F2psBegin {$F2psDict begin /$F2psEnteredState save def} def /$F2psEnd {$F2psEnteredState restore end} def %%EndProlog $F2psBegin 10 setmiterlimit n -1000 8812 m -1000 -1000 l 10012 -1000 l 10012 8812 l cp clip 0.06000 0.06000 sc % Arc 15.000 slw gs n 6001.3 4800.0 1351.3 -56.4 56.4 arcn gs col7 0.80 shd ef gr gs col-1 s gr gr % Polyline n 6750 3675 m 8400 4800 l 6750 5925 l gs col7 0.80 shd ef gr gs col-1 s gr % Arc 7.500 slw gs n 8107.5 4657.5 551.2 -123.9 165.0 arcn gs col-1 s gr gr % Arc gs n 5983.7 4800.0 1708.7 -44.6 44.6 arcn gs col-1 s gr gr % Arc 15.000 slw gs n 8175.0 4725.0 167.7 -116.6 153.4 arcn gs col-1 s gr gr 7.500 slw % Ellipse n 6000 4800 2400 2400 0 360 DrawEllipse gs col-1 s gr [60] 0 sd % Ellipse n 7200 4800 1200 1200 0 360 DrawEllipse gs col-1 s gr [] 0 sd % Polyline gs clippath 5970 1920 m 6000 1800 l 6030 1920 l 6030 1785 l 5970 1785 l cp clip n 6000 7800 m 6000 1800 l gs col-1 s gr gr % arrowhead n 5970 1920 m 6000 1800 l 6030 1920 l col-1 s % Polyline n 7200 3600 m 8400 4800 l 7200 6000 l gs col-1 s gr % Polyline [30] 0 sd n 6000 4800 m 6750 3675 l gs col-1 s gr [] 0 sd % Polyline [30] 0 sd n 6600 3900 m 6825 4050 l 6975 3825 l gs col-1 s gr [] 0 sd % Polyline [30] 0 sd n 6000 4800 m 7200 6000 l gs col-1 s gr [] 0 sd % Polyline [30] 0 sd n 7050 5850 m 7200 5700 l 7350 5850 l gs col-1 s gr [] 0 sd % Polyline [30] 0 sd gs clippath 7489 3931 m 7425 3825 l 7531 3889 l 7436 3793 l 7393 3836 l cp clip n 8400 4800 m 7425 3825 l gs col-1 s gr gr [] 0 sd % arrowhead n 7489 3931 m 7425 3825 l 7531 3889 l col-1 s % Polyline gs clippath 6704 6327 m 6825 6300 l 6731 6380 l 6852 6320 l 6825 6266 l cp clip n 6675 6375 m 6825 6300 l gs col-1 s gr gr % arrowhead n 6704 6327 m 6825 6300 l 6731 6380 l col-1 s % Polyline gs clippath 5448 3216 m 5325 3225 l 5429 3159 l 5301 3201 l 5320 3258 l cp clip n 5550 3150 m 5325 3225 l gs col-1 s gr gr % arrowhead n 5448 3216 m 5325 3225 l 5429 3159 l col-1 s % Polyline gs clippath 8880 4770 m 9000 4800 l 8880 4830 l 9015 4830 l 9015 4770 l cp clip n 3000 4800 m 9000 4800 l gs col-1 s gr gr % arrowhead n 8880 4770 m 9000 4800 l 8880 4830 l col-1 s /Times-Roman-iso ff 210.00 scf sf 5850 5025 m gs 1 -1 sc (0) col-1 sh gr /Times-Roman-iso ff 180.00 scf sf 4725 5025 m gs 1 -1 sc (-sin) col-1 sh gr /Symbol ff 150.00 scf sf 5250 4350 m gs 1 -1 sc (a) col-1 sh gr /Symbol ff 180.00 scf sf 7425 4500 m gs 1 -1 sc (a) col-1 sh gr /Times-Italic-iso ff 210.00 scf sf 7200 3525 m gs 1 -1 sc (A) col-1 sh gr /Times-Italic-iso ff 210.00 scf sf 6600 5325 m gs 1 -1 sc (a') col-1 sh gr /Times-Roman-iso ff 180.00 scf sf 3825 5025 m gs 1 -1 sc (-sin) col-1 sh gr /Symbol ff 180.00 scf sf 5025 5025 m gs 1 -1 sc (a) col-1 sh gr /Times-Italic-iso ff 210.00 scf sf 5100 4275 m gs 1 -1 sc (D) col-1 sh gr /Times-Italic-iso ff 210.00 scf sf 6450 4350 m gs 1 -1 sc (a) col-1 sh gr /Symbol ff 180.00 scf sf 7875 4725 m gs 1 -1 sc (a) col-1 sh gr /Symbol ff 180.00 scf sf 4125 5025 m gs 1 -1 sc (a) col-1 sh gr /Times-Roman-iso ff 180.00 scf sf 4200 5025 m gs 1 -1 sc (') col-1 sh gr /Times-Roman-iso ff 180.00 scf sf 7500 4500 m gs 1 -1 sc (') col-1 sh gr /Times-Italic-iso ff 210.00 scf sf 7200 6225 m gs 1 -1 sc (B) col-1 sh gr /Times-Roman-iso ff 210.00 scf sf 8475 5025 m gs 1 -1 sc (1) col-1 sh gr /Times-Roman-iso ff 210.00 scf sf 6075 7500 m gs 1 -1 sc (-i) col-1 sh gr /Times-Roman-iso ff 210.00 scf sf 6150 2250 m gs 1 -1 sc (i) col-1 sh gr /Symbol ff 210.00 scf sf 3300 5025 m gs 1 -1 sc (-1) col-1 sh gr $F2psEnd rs \end{filecontents}{CZ4f3.eps} \begin{filecontents}{CZ4f2.eps} %!PS-Adobe-2.0 EPSF-2.0 %%Title: CZ4f2.fig %%Creator: fig2dev Version 3.1 Patchlevel 2 %%CreationDate: Fri Jan 14 16:51:21 2000 %%For: cachia@cptsu6 (Vincent Cachia thesard Zagrebnov) %Magnification: 1.00 %%Orientation: Portrait %%BoundingBox: 0 0 290 290 %%Pages: 0 %%BeginSetup %%IncludeFeature: *PageSize Letter %%EndSetup %%EndComments /MyAppDict 100 dict dup begin def /$F2psDict 200 dict def $F2psDict begin $F2psDict /mtrx matrix put /col-1 {0 setgray} bind def /col0 {0.000 0.000 0.000 srgb} bind def /col1 {0.000 0.000 1.000 srgb} bind def /col2 {0.000 1.000 0.000 srgb} bind def /col3 {0.000 1.000 1.000 srgb} bind def /col4 {1.000 0.000 0.000 srgb} bind def /col5 {1.000 0.000 1.000 srgb} bind def /col6 {1.000 1.000 0.000 srgb} bind def /col7 {1.000 1.000 1.000 srgb} bind def /col8 {0.000 0.000 0.560 srgb} bind def /col9 {0.000 0.000 0.690 srgb} bind def /col10 {0.000 0.000 0.820 srgb} bind def /col11 {0.530 0.810 1.000 srgb} bind def /col12 {0.000 0.560 0.000 srgb} bind def /col13 {0.000 0.690 0.000 srgb} bind def /col14 {0.000 0.820 0.000 srgb} bind def /col15 {0.000 0.560 0.560 srgb} bind def /col16 {0.000 0.690 0.690 srgb} bind def /col17 {0.000 0.820 0.820 srgb} bind def /col18 {0.560 0.000 0.000 srgb} bind def /col19 {0.690 0.000 0.000 srgb} bind def /col20 {0.820 0.000 0.000 srgb} bind def /col21 {0.560 0.000 0.560 srgb} bind def /col22 {0.690 0.000 0.690 srgb} bind def /col23 {0.820 0.000 0.820 srgb} bind def /col24 {0.500 0.190 0.000 srgb} bind def /col25 {0.630 0.250 0.000 srgb} bind def /col26 {0.750 0.380 0.000 srgb} bind def /col27 {1.000 0.500 0.500 srgb} bind def /col28 {1.000 0.630 0.630 srgb} bind def /col29 {1.000 0.750 0.750 srgb} bind def /col30 {1.000 0.880 0.880 srgb} bind def /col31 {1.000 0.840 0.000 srgb} bind def end save -107.0 361.0 translate 1 -1 scale .9 .9 scale % to make patterns same scale as in xfig % This junk string is used by the show operators /PATsstr 1 string def /PATawidthshow { % cx cy cchar rx ry string % Loop over each character in the string { % cx cy cchar rx ry char % Show the character dup % cx cy cchar rx ry char char PATsstr dup 0 4 -1 roll put % cx cy cchar rx ry char (char) false charpath % cx cy cchar rx ry char /clip load PATdraw % Move past the character (charpath modified the % current point) currentpoint % cx cy cchar rx ry char x y newpath moveto % cx cy cchar rx ry char % Reposition by cx,cy if the character in the string is cchar 3 index eq { % cx cy cchar rx ry 4 index 4 index rmoveto } if % Reposition all characters by rx ry 2 copy rmoveto % cx cy cchar rx ry } forall pop pop pop pop pop % - currentpoint newpath moveto } bind def /PATcg { 7 dict dup begin /lw currentlinewidth def /lc currentlinecap def /lj currentlinejoin def /ml currentmiterlimit def /ds [ currentdash ] def /cc [ currentrgbcolor ] def /cm matrix currentmatrix def end } bind def % PATdraw - calculates the boundaries of the object and % fills it with the current pattern /PATdraw { % proc save exch PATpcalc % proc nw nh px py 5 -1 roll exec % nw nh px py newpath PATfill % - restore } bind def % PATfill - performs the tiling for the shape /PATfill { % nw nh px py PATfill - PATDict /CurrentPattern get dup begin setfont % Set the coordinate system to Pattern Space PatternGState PATsg % Set the color for uncolored pattezns PaintType 2 eq { PATDict /PColor get PATsc } if % Create the string for showing 3 index string % nw nh px py str % Loop for each of the pattern sources 0 1 Multi 1 sub { % nw nh px py str source % Move to the starting location 3 index 3 index % nw nh px py str source px py moveto % nw nh px py str source % For multiple sources, set the appropriate color Multi 1 ne { dup PC exch get PATsc } if % Set the appropriate string for the source 0 1 7 index 1 sub { 2 index exch 2 index put } for pop % Loop over the number of vertical cells 3 index % nw nh px py str nh { % nw nh px py str currentpoint % nw nh px py str cx cy 2 index show % nw nh px py str cx cy YStep add moveto % nw nh px py str } repeat % nw nh px py str } for 5 { pop } repeat end } bind def % PATkshow - kshow with the current pattezn /PATkshow { % proc string exch bind % string proc 1 index 0 get % string proc char % Loop over all but the last character in the string 0 1 4 index length 2 sub { % string proc char idx % Find the n+1th character in the string 3 index exch 1 add get % string proe char char+1 exch 2 copy % strinq proc char+1 char char+1 char % Now show the nth character PATsstr dup 0 4 -1 roll put % string proc chr+1 chr chr+1 (chr) false charpath % string proc char+1 char char+1 /clip load PATdraw % Move past the character (charpath modified the current point) currentpoint newpath moveto % Execute the user proc (should consume char and char+1) mark 3 1 roll % string proc char+1 mark char char+1 4 index exec % string proc char+1 mark... cleartomark % string proc char+1 } for % Now display the last character PATsstr dup 0 4 -1 roll put % string proc (char+1) false charpath % string proc /clip load PATdraw neewath pop pop % - } bind def % PATmp - the makepattern equivalent /PATmp { % patdict patmtx PATmp patinstance exch dup length 7 add % We will add 6 new entries plus 1 FID dict copy % Create a new dictionary begin % Matrix to install when painting the pattern TilingType PATtcalc /PatternGState PATcg def PatternGState /cm 3 -1 roll put % Check for multi pattern sources (Level 1 fast color patterns) currentdict /Multi known not { /Multi 1 def } if % Font dictionary definitions /FontType 3 def % Create a dummy encoding vector /Encoding 256 array def 3 string 0 1 255 { Encoding exch dup 3 index cvs cvn put } for pop /FontMatrix matrix def /FontBBox BBox def /BuildChar { mark 3 1 roll % mark dict char exch begin Multi 1 ne {PaintData exch get}{pop} ifelse % mark [paintdata] PaintType 2 eq Multi 1 ne or { XStep 0 FontBBox aload pop setcachedevice } { XStep 0 setcharwidth } ifelse currentdict % mark [paintdata] dict /PaintProc load % mark [paintdata] dict paintproc end gsave false PATredef exec true PATredef grestore cleartomark % - } bind def currentdict end % newdict /foo exch % /foo newlict definefont % newfont } bind def % PATpcalc - calculates the starting point and width/height % of the tile fill for the shape /PATpcalc { % - PATpcalc nw nh px py PATDict /CurrentPattern get begin gsave % Set up the coordinate system to Pattern Space % and lock down pattern PatternGState /cm get setmatrix BBox aload pop pop pop translate % Determine the bounding box of the shape pathbbox % llx lly urx ury grestore % Determine (nw, nh) the # of cells to paint width and height PatHeight div ceiling % llx lly urx qh 4 1 roll % qh llx lly urx PatWidth div ceiling % qh llx lly qw 4 1 roll % qw qh llx lly PatHeight div floor % qw qh llx ph 4 1 roll % ph qw qh llx PatWidth div floor % ph qw qh pw 4 1 roll % pw ph qw qh 2 index sub cvi abs % pw ph qs qh-ph exch 3 index sub cvi abs exch % pw ph nw=qw-pw nh=qh-ph % Determine the starting point of the pattern fill %(px, py) 4 2 roll % nw nh pw ph PatHeight mul % nw nh pw py exch % nw nh py pw PatWidth mul exch % nw nh px py end } bind def % Save the original routines so that we can use them later on /oldfill /fill load def /oldeofill /eofill load def /oldstroke /stroke load def /oldshow /show load def /oldashow /ashow load def /oldwidthshow /widthshow load def /oldawidthshow /awidthshow load def /oldkshow /kshow load def % These defs are necessary so that subsequent procs don't bind in % the originals /fill { oldfill } bind def /eofill { oldeofill } bind def /stroke { oldstroke } bind def /show { oldshow } bind def /ashow { oldashow } bind def /widthshow { oldwidthshow } bind def /awidthshow { oldawidthshow } bind def /kshow { oldkshow } bind def /PATredef { MyAppDict begin { /fill { /clip load PATdraw newpath } bind def /eofill { /eoclip load PATdraw newpath } bind def /stroke { PATstroke } bind def /show { 0 0 null 0 0 6 -1 roll PATawidthshow } bind def /ashow { 0 0 null 6 3 roll PATawidthshow } bind def /widthshow { 0 0 3 -1 roll PATawidthshow } bind def /awidthshow { PATawidthshow } bind def /kshow { PATkshow } bind def } { /fill { oldfill } bind def /eofill { oldeofill } bind def /stroke { oldstroke } bind def /show { oldshow } bind def /ashow { oldashow } bind def /widthshow { oldwidthshow } bind def /awidthshow { oldawidthshow } bind def /kshow { oldkshow } bind def } ifelse end } bind def false PATredef % Conditionally define setcmykcolor if not available /setcmykcolor where { pop } { /setcmykcolor { 1 sub 4 1 roll 3 { 3 index add neg dup 0 lt { pop 0 } if 3 1 roll } repeat setrgbcolor - pop } bind def } ifelse /PATsc { % colorarray aload length % c1 ... cn length dup 1 eq { pop setgray } { 3 eq { setrgbcolor } { setcmykcolor } ifelse } ifelse } bind def /PATsg { % dict begin lw setlinewidth lc setlinecap lj setlinejoin ml setmiterlimit ds aload pop setdash cc aload pop setrgbcolor cm setmatrix end } bind def /PATDict 3 dict def /PATsp { true PATredef PATDict begin /CurrentPattern exch def % If it's an uncolored pattern, save the color CurrentPattern /PaintType get 2 eq { /PColor exch def } if /CColor [ currentrgbcolor ] def end } bind def % PATstroke - stroke with the current pattern /PATstroke { countdictstack save mark { currentpoint strokepath moveto PATpcalc % proc nw nh px py clip newpath PATfill } stopped { (*** PATstroke Warning: Path is too complex, stroking with gray) = cleartomark restore countdictstack exch sub dup 0 gt { { end } repeat } { pop } ifelse gsave 0.5 setgray oldstroke grestore } { pop restore pop } ifelse newpath } bind def /PATtcalc { % modmtx tilingtype PATtcalc tilematrix % Note: tiling types 2 and 3 are not supported gsave exch concat % tilingtype matrix currentmatrix exch % cmtx tilingtype % Tiling type 1 and 3: constant spacing 2 ne { % Distort the pattern so that it occupies % an integral number of device pixels dup 4 get exch dup 5 get exch % tx ty cmtx XStep 0 dtransform round exch round exch % tx ty cmtx dx.x dx.y XStep div exch XStep div exch % tx ty cmtx a b 0 YStep dtransform round exch round exch % tx ty cmtx a b dy.x dy.y YStep div exch YStep div exch % tx ty cmtx a b c d 7 -3 roll astore % { a b c d tx ty } } if grestore } bind def /PATusp { false PATredef PATDict begin CColor PATsc end } bind def % right45 11 dict begin /PaintType 1 def /PatternType 1 def /TilingType 1 def /BBox [0 0 1 1] def /XStep 1 def /YStep 1 def /PatWidth 1 def /PatHeight 1 def /Multi 2 def /PaintData [ { clippath } bind { 32 32 true [ 32 0 0 -32 0 32 ] {<010101010202020204040404080808081010101020202020 404040408080808001010101020202020404040408080808 101010102020202040404040808080800101010102020202 040404040808080810101010202020204040404080808080 010101010202020204040404080808081010101020202020 4040404080808080>} imagemask } bind ] def /PaintProc { pop exec fill } def currentdict end /P5 exch def 1.1111 1.1111 scale %restore scale /cp {closepath} bind def /ef {eofill} bind def /gr {grestore} bind def /gs {gsave} bind def /sa {save} bind def /rs {restore} bind def /l {lineto} bind def /m {moveto} bind def /rm {rmoveto} bind def /n {newpath} bind def /s {stroke} bind def /sh {show} bind def /slc {setlinecap} bind def /slj {setlinejoin} bind def /slw {setlinewidth} bind def /srgb {setrgbcolor} bind def /rot {rotate} bind def /sc {scale} bind def /sd {setdash} bind def /ff {findfont} bind def /sf {setfont} bind def /scf {scalefont} bind def /sw {stringwidth} bind def /tr {translate} bind def /tnt {dup dup currentrgbcolor 4 -2 roll dup 1 exch sub 3 -1 roll mul add 4 -2 roll dup 1 exch sub 3 -1 roll mul add 4 -2 roll dup 1 exch sub 3 -1 roll mul add srgb} bind def /shd {dup dup currentrgbcolor 4 -2 roll mul 4 -2 roll mul 4 -2 roll mul srgb} bind def /$F2psBegin {$F2psDict begin /$F2psEnteredState save def} def /$F2psEnd {$F2psEnteredState restore end} def %%EndProlog $F2psBegin 10 setmiterlimit n 0 792 m 0 0 l 612 0 l 612 792 l cp clip 0.06000 0.06000 sc 7.500 slw % Polyline n 1800 4950 m 4200 3600 l 1800 2325 l gs /PC [[1.00 1.00 1.00] [0.00 0.00 0.00]] def 15.00 15.00 sc P5 [16 0 0 -16 120.00 155.00] PATmp PATsp ef gr PATusp gs col-1 s gr % Arc gs n 3937.5 3525.0 345.7 -139.4 167.5 arcn gs col-1 s gr gr % Arc gs [66.7] 0 sd n 3862.5 3337.5 427.6 -127.9 -164.7 arcn gs col-1 s gr gr [] 0 sd % Polyline gs clippath 6453 3570 m 6573 3600 l 6453 3630 l 6615 3630 l 6615 3570 l cp clip n 1800 3600 m 6600 3600 l gs col-1 s gr gr % arrowhead n 6453 3570 m 6573 3600 l 6453 3630 l col-1 s % Polyline gs clippath 4170 1347 m 4200 1227 l 4230 1347 l 4230 1185 l 4170 1185 l cp clip n 4200 6000 m 4200 1200 l gs col-1 s gr gr % arrowhead n 4170 1347 m 4200 1227 l 4230 1347 l col-1 s % Arc gs n 4187.5 3600.0 1212.5 132.1 -132.1 arcn gs col-1 s gr gr % Polyline [66.7] 0 sd n 3375 4500 m 4200 3600 l 3375 2700 l gs col-1 s gr [] 0 sd % Polyline n 3375 2700 m 2325 1500 l gs col-1 s gr % Polyline n 2250 5700 m 3375 4500 l gs col-1 s gr % Polyline gs clippath 5050 4433 m 5155 4369 l 5092 4475 l 5207 4361 l 5164 4318 l cp clip n 5100 4425 m 5175 4350 l gs col-1 s gr gr % arrowhead n 5050 4433 m 5155 4369 l 5092 4475 l col-1 s % Polyline n 3075 4950 m 3225 4950 l gs col-1 s gr % Polyline gs clippath 3025 4808 m 3130 4744 l 3067 4850 l 3182 4736 l 3139 4693 l cp clip n 3075 4800 m 3150 4725 l gs col-1 s gr gr % arrowhead n 3025 4808 m 3130 4744 l 3067 4850 l col-1 s % Polyline gs clippath 2783 2075 m 2719 1969 l 2825 2033 l 2711 1918 l 2668 1961 l cp clip n 2775 2025 m 2700 1950 l gs col-1 s gr gr % arrowhead n 2783 2075 m 2719 1969 l 2825 2033 l col-1 s /Times-Roman ff 180.00 scf sf 4275 3825 m gs 1 -1 sc (0) col-1 sh gr /Symbol ff 150.00 scf sf 5325 2775 m gs 1 -1 sc (d) col-1 sh gr /Symbol ff 180.00 scf sf 5250 2625 m gs 1 -1 sc (G) col-1 sh gr /Symbol ff 180.00 scf sf 3075 5100 m gs 1 -1 sc (G) col-1 sh gr /Symbol ff 180.00 scf sf 2925 2025 m gs 1 -1 sc (G) col-1 sh gr /Symbol ff 150.00 scf sf 3150 5175 m gs 1 -1 sc (e) col-1 sh gr /Symbol ff 150.00 scf sf 3000 2100 m gs 1 -1 sc (e) col-1 sh gr /Symbol ff 180.00 scf sf 3450 3450 m gs 1 -1 sc (a) col-1 sh gr /Symbol ff 180.00 scf sf 3375 3000 m gs 1 -1 sc (e) col-1 sh gr $F2psEnd rs end \end{filecontents}{CZ4f2.eps} \documentstyle[12pt,amsfonts,epsf,float]{article} \oddsidemargin 0.5cm \topmargin -2.5cm \textheight 24.0cm \textwidth 16.0cm \headheight 0cm \headsep 0cm \footskip 1cm \topmargin 0cm \linespread{1} \newtheorem{theorem}{Theorem} \newtheorem{proposition}{Proposition} \newtheorem{lemma}{Lemma} \newtheorem{corollary}{Corollary} \newtheorem{example}{Example} \newtheorem{definition}{Definition} \newtheorem{remark}{Remark} \def\pitchfork{\hbox{$\cap \hskip -0.18cm \vert$}} \makeatletter \def\@ypreuve[#1]{\@preuve{ #1}} \def\@preuve#1{\begin{trivlist}\item[]{\em Proof#1. }} \newenvironment{proof}{\@ifnextchar[{\@ypreuve}{\@preuve{}}}{\hfill$\Box$\end{trivlist}} \makeatother \begin{document} \font\fifteen=cmbx10 at 15pt \font\twelve=cmbx10 at 12pt \begin{titlepage} \begin{center} \renewcommand{\thefootnote}{\fnsymbol{footnote}} {\twelve Centre de Physique Th\'eorique\footnote{Unit\'e Propre de Recherche 7061 }, CNRS Luminy, Case 907} {\twelve F-13288 Marseille -- Cedex 9} \vspace{2 cm} {\fifteen OPERATOR-NORM APPROXIMATION OF SEMIGROUPS BY SECTORIAL CONTRACTIONS} \vspace{1.5 cm} \setcounter{footnote}{0} \renewcommand{\thefootnote}{\arabic{footnote}} {\bf V. CACHIA\footnote{Allocataire-moniteur at Universit\'e de la M\'editerran\'ee (Aix-Marseille II)\\ Email : cachia@cpt.univ-mrs.fr}, and V.A. ZAGREBNOV\footnote{Universit\'e de la M\'editerran\'ee (Aix-Marseille II) - Email : zagrebnov@cpt.univ-mrs.fr} } \vspace{1.5 cm} {\bf Abstract} \end{center} We extend the Chernoff theory for approximation of contraction semigroups \`a la Trotter. We show that the Trotter-Neveu-Kato convergence theorem holds in operator norm for a family of uniformly m-sectorial generators in a Hilbert space. Then we obtain a Chernoff-type approximation theorem for sectorial contractions on a Hilbert space in the operator norm. As examples we give necessary and sufficient conditions for the operator-norm convergence of Trotter-type product formul\ae. \vspace{1 cm} \bigskip \noindent Keywords : Chernoff theory, Trotter product formula, operator-norm convergence, holomorphic semigroups, sectorial operators. \noindent Mathematics Subject Classification : 47D03, 47B25, 35K22, 41A80 \bigskip \bigskip \noindent anonymous ftp : ftp.cpt.univ-mrs.fr\\ web : www.cpt.univ-mrs.fr\\ May 2000\\ CPT-2000/P.4012 \renewcommand{\thefootnote}{\fnsymbol{footnote}} \end{titlepage} \setcounter{footnote}{0} \renewcommand{\thefootnote}{\arabic{footnote}} \newpage \section{Introduction} In his article \cite{Chernoff}, Chernoff has proved the following proposition: \begin{proposition} Let $F(t)$ be a strongly continuous function from $[0,\infty)$ to the linear contractions on a Banach space $\cal X$ such that $F(0)=I$. Suppose that the closure $C$ of the strong derivative $F'(0)$ is the generator of a contraction semigroup. Then $F(t/n)^n$ converges to $e^{tC}$ in the strong operator topology. \end{proposition} This result describes a method for the approximation of a general contraction semigroup. It leads in particular to a simple proof of the strong convergence for the Trotter product formula (see \cite[\S 3.4]{Davies}). For self-adjoint families of operators on a Hilbert space there exists a generalization of the above result in the operator-norm topology \cite{NZ1}, \cite{NZ3}: \begin{proposition} Let $\{\Phi(s)\}_{s\geq 0}$ be a family of self-adjoint non-negative contractions on a separable Hilbert space $\cal H$, and let $X_0$ be a self-adjoint operator in the closed subspace ${\cal H}_0\subseteq{\cal H}$. Define $X(s)=s^{-1}(I-\Phi(s))$, $s\geq 0$. Then the family $\{X(s)\}_{s\geq 0}$ converges in the uniform resolvent sense to $X_0$ as $s\rightarrow +0$ if and only if the sequence $\{\Phi(t/r)^r\}_{r\geq 1}$, $t\geq 0$, converges in operator-norm to $e^{-tX_0}P_0$ as $r\rightarrow +\infty$, uniformly on any compact $t$-interval in $(0,\infty)$. Here $P_0$ denotes the orthogonal projection from ${\cal H}$ onto ${\cal H}_0$. \end{proposition} The aim of the present paper is to show that the condition of self-adjointness in this result can be replaced by a much weaker condition on the corresponding numerical range. Let $\cal H$ be a separable Hilbert space and let $T$ be a linear operator in $\cal H$. We denote by $\Theta(T)$ the numerical range of $T$ : % $$ \Theta(T) = \{ (Tu,u), u\in{\cal D}(T) \mbox{ and } \|u\|=1 \}. $$ % %The Toeplitz-Hausdorff theorem states that it is a convex set (see e.g. \cite{Halmos} Ch.17). It follows that the complement in $\bb C$ of its closure is connected, except in the special case where $\Theta(T)$ is a strip bounded by two parallel lines. % Below we use some important properties of this set (cf \cite[Ch.V Theorem 3.2]{Kato}): \begin{proposition}\label{propnumrange} If $T$ is a closed operator in $\cal H$, then for any complex number $z\notin \overline{\Theta(T)}$, $T-z$ has closed range, it is injective and its deficiency ${\mathrm def}(T-z)$ is constant (in each connected component). If ${\mathrm def}(T-z) = 0$, then the spectrum of $T$ is a subset of $\overline{\Theta(T)}$ and \begin{equation}\label{estres} \|(T-z)^{-1}\| \leq {1\over dist(z,\Theta(T))}. \end{equation} \end{proposition} For $0 < \alpha \leq \pi$, we denote by $S_\alpha$ the open sector $$ S_\alpha = \{z\in{\mathbb C}\setminus\{0\}: |\arg z|<\alpha \}.$$ \begin{definition} Operator $T$ in $\cal H$ is said to be sectorial with a semi-angle $\alpha\in (0,\pi/2)$ and with vertex at $0$ if $\Theta(T)\subseteq S_\alpha$. If, in addition, $T$ is closed and there exists $z\notin S_\alpha$ that belongs to the resolvent set of $T$, then $T$ is said to be m-sectorial. \end{definition} % We recall the following result (see \cite[Ch.IX Theorem 1.24]{Kato}): % \begin{proposition}\label{P1} Let $T$ be an m-sectorial operator in a Hilbert space $\cal H$, with a semi-angle $\alpha \in (0,\pi/2)$ and with vertex at $0$. Then the semigroup $e^{-tT}$ is holomorphic in $S_{\pi/2-\alpha}$ and it is bounded by $\|e^{-tT}\|\leq 1$ for $t\in S_{\pi/2 - \alpha}$. \end{proposition} We define for $0 \leq \alpha \leq\pi/2$ \begin{equation}\label{Dalpha} D_\alpha = \{z\in{\mathbb C}: |z|\leq \sin\alpha\} \cup \{z\in{\mathbb C}: |\arg(1-z)|\leq \alpha \mbox{ and } |z-1|\leq\cos\alpha\}. \end{equation} \begin{definition}\label{SC} We shall say that the contraction $C$ on a Hilbert space $\cal H$ is sectorial with a semi-angle $0 \leq\alpha <\pi/2$ (and with vertex at 1), if its numerical range $\Theta(C)$ is a subset of $D_\alpha$. Notice that the case $\alpha = 0$ corresponds to a non-negative self-adjoint contraction. \end{definition} Let $\{\Phi(s)\}_{s\geq 0}$ be a family of sectorial contractions on a Hilbert space $\cal H$ such that $\Theta(\Phi(s))\subseteq D_\alpha$ for some $0\leq \alpha<\pi/2$ and for all $s\geq 0$. Let $X(s)=(I-\Phi(s))/s$, and let $X_0$ be a closed operator in a closed subspace ${\cal H}_0\subseteq{\cal H}$ with non-empty resolvent set. Then the main result of this paper (see Theorem \ref{T}) can be formulated as the equivalence between two limits: \begin{eqnarray} \|(\zeta+X(s))^{-1} - (\zeta+X_0)^{-1}P_0\| & \longrightarrow & 0 \mbox{ as } s\rightarrow +0,\ \mbox{Re}\,\zeta>0, \label{lim1}\\ \|\Phi(t/n)^n - e^{-tX_0}P_0\| & \longrightarrow & 0 \mbox{ as } n\rightarrow \infty,\ t> 0.\label{lim2} \end{eqnarray} Our method is an operator-norm generalization of the Chernoff theory \cite{Chernoff}. Definition \ref{SC} will be illustrated in the next section. Then we generalize Lemma 2 of \cite{Chernoff} to the operator-norm topology. In Section \ref{appth} we prove the operator-norm analogue of the Trotter-Neveu-Kato convergence theorem (see \cite[Theorem 5.1]{Trotter1}, or e.g. \cite[Theorem 3.17]{Davies}, \cite[Ch. 1.7]{Goldstein}) and the main result (\ref{lim1}),(\ref{lim2}) formulated above. Finally, we give some applications concerning Trotter-Kato formul\ae. \section{Sectorial contractions}\label{sec1} In this section we give some important examples of operator families corresponding to Definition \ref{SC}, i.e. sectorial contractions. Let $A$ be an m-sectorial operator with a semi-angle $\alpha$ and with vertex at $0$, in a Hilbert space $\cal H$. Then there are two families of sectorial contractions that are generated by $A$ in a natural way. \subsection{Resolvent of a sectorial operator}\label{cfex1} The family of resolvents $F(t) = (I+tA)^{-1}$, $t\geq 0$, has the following properties: \smallskip $(i)$ Since $F(t=0)=I$ and since for $t>0$ \begin{equation} \|F(t)\| \leq {1\over t\,\mbox{dist}(1/t,-S_\alpha)} = 1, \end{equation} \hspace{1.5cm}$F(t)$ is a family of contractions; $(ii)$ $\Theta(F(t))\subseteq S_\alpha$, because for any $u\in\cal H$: \begin{equation} (u,F(t)u) = (v,v) + t(Av,v)\in S_\alpha, \end{equation} \hspace{1.5cm} where $v=(I+tA)^{-1}u$. \smallskip $(iii)$ $\Theta(I-F(t))\subseteq S_\alpha$, since for any $u\in\cal H$ \begin{equation} (u,(I-F(t))u) = ((I+tA)v,tAv) = t(v,Av) + t^2(Av,Av)\in S_\alpha, \end{equation} \hspace{1.5cm} where $v=(I+tA)^{-1}u$. \smallskip Therefore, we have $\Theta(F(t))\subseteq S_\alpha\cap (1-S_\alpha) \subseteq D_\alpha$, i.e. $F(t)$ are sectorial contractions for $t\geq 0$. \subsection{Semigroup generated by a sectorial operator} The semigroup generated by the m-sectorial operator $A$ is holomorphic and contractive in the sector $S_{\pi/2-\alpha}$ (cf Proposition \ref{P1}). We shall show that in fact $\Theta(e^{-tA})\subseteq D_\alpha$ for $t\geq 0$. It is not as simple as for the resolvent. The key statement is the following mapping theorem for the numerical range due to Kato \cite{K1}: \begin{proposition}\label{mapth} Let $f(z)$ be a rational function with $f(\infty)=\infty$. Let $E'$ be a compact convex set in the complex plane, let $E=f^{-1}(E')$ and $K$ be the convex kernel of $E$. If $A$ is an operator with $\Theta(A)\subseteq K$, then $\Theta(f(A))\subseteq E'$. In particular if $D$ is a compact convex subset of $\mathbb C$ with $f(D)\subseteq D$ and $\Theta(A)\subseteq D$, then $\Theta(f(A))\subseteq D$. \end{proposition} \begin{lemma}\label{z^n} Let $f(z) = z^n$ where $n\in{\mathbb N}$, then $f(D_\alpha)\subseteq D_\alpha$. \end{lemma} \begin{proof} If $|z|\leq \sin\alpha$, then $|z^n|\leq \sin\alpha$ and $z^n\in D_\alpha$. Thus it remains the case $z\in D_\alpha$, $|z|>\sin\alpha$. Let us consider the family of the straight lines $z(t)=1-te^{i\beta}$ in $D_\alpha$, parametrized by $0\leq t\leq \cos\alpha$ and $-\alpha\leq \beta \leq\alpha$. We study their images defined by $\zeta(t) = z(t)^n$ to prove that they also lie in $D_\alpha$. They form a regular plane curve, i.e. $\zeta'(t) = -ne^{i\beta}(1-te^{i\beta})^{n-1} \neq 0$. Then for each path $\zeta(t)$ we can define unit tangent vectors $T(t)$: \begin{equation}\label{tgvector} T(t) = \frac{\zeta'(t)}{|\zeta'(t)|} = -e^{i\beta} \left(\frac{1-te^{i\beta}}{|1-te^{i\beta}|} \right)^{n-1}. \end{equation} The curvature $c(t)$ of a regular plane curve is defined by $T'(t) = ic(t)T(t)$, and we find by (\ref{tgvector}): \begin{equation} \hspace{-0.5cm} T'(t) = -e^{i\beta}(n-1)\left(\frac{z}{|z|} \right)^{n-2} \frac{d}{dt}\left(\frac{z}{|z|} \right) = i(n-1) \frac{\sin\beta}{|z|^2} e^{i\beta} \left(\frac{z}{|z|} \right)^{n-1} \!\!\!\!\!\!, \end{equation} here $z=z(t)=1-te^{i\beta}$. Hence the expression for the curvature is $c(t)=-(n-1)\sin\beta /|z(t)|^2$. It does not vanish for any $0\leq t\leq \cos\alpha$ and it has the sign opposite to $\beta$. Therefore, the curve $\zeta(t)$ lies in the half-plane defined by any tangent and containing the corresponding curvature center. Let $t=0$, then the tangent is $z(t)$, and the curvature center is at $1+ie^{i\beta}(n-1)\sin\beta$. Finally, all images $\zeta(t)=z(t)^n$, for $0\leq t\leq \cos\alpha$ and $-\alpha\leq \beta\leq \alpha$, are squeezed between two extreme tangent lines corresponding to $\beta = \pm\alpha$. This finishes the proof. \end{proof} \begin{theorem} If $A$ is an m-sectorial operator with a semi-angle $\alpha$ and with vertex at $0$, then $\Theta(e^{-tA})$ is a subset of $D_\alpha$ for any $t\geq 0$. \end{theorem} \begin{proof} By virtue of example from Section \ref{cfex1} we have that $\Theta((I+tA/n)^{-1}) \subseteq D_\alpha$ for any positive real $t$ and any non-zero integer $n$. According to Proposition \ref{mapth} and to Lemma \ref{z^n} for the rational function $f(z) = z^n$ ($f(\infty) = \infty$), the operator $(I+tA/n)^{-1}$, and the compact convex set $D_\alpha$, we obtain that $\Theta((I+tA/n)^{-n}) \subseteq D_\alpha$ for any $t\geq 0$ and $n\in {\mathbb N}\setminus\{0\}$. On the other hand it is known (see \cite[Ch.IX]{Kato}) that the sequence $(I+tA/n)^{-n}$ converges strongly to the semigroup $e^{-tA}$ as $n\rightarrow \infty$. Therefore, $\lim_{n\rightarrow\infty} ((I+tA/n)^{-n}u,u) = (e^{-tA}u,u) \in D_\alpha$ for any unit vector $u\in {\cal H}$, which proves the assertion. \end{proof} \section{Generalization of the Chernoff lemma} In this section we show that an operator-norm analogue of the main Lemma 2 of \cite{Chernoff} is valid if one supposes that the contraction $C$ is sectorial in the sense of Definition \ref{SC}. \begin{figure}[h] \epsfysize=10cm \setlength{\unitlength}{0.1bp} $$ \begin{picture}(3000,3000)(0,0) \put(1200,2300){\makebox(0,0){\small $\partial D_{\alpha'}$}} \put(1050,2050){\makebox(0,0){\small $\partial D_\alpha$}} \epsfbox{CZ4f3.eps} \end{picture} $$ \caption{Illustration of the set $D_{\alpha}$ (shaded domain) with boundary $\partial D_{\alpha}$, $a = \sin\alpha$, as well as of our choice of the contour $\partial D_{\alpha'}$ in the resolvent set $\rho(C)$, where $a'=\sin\alpha' >a$. The contour consists of two segments of tangent straight lines $(1,A)$ and $(1,B)$ and the arc $(A,B)$ of radius $a'$. The dotted circle corresponds to the set of tangency points for the different values of $\alpha\in[0,\pi/2]$.} \end{figure} \begin{lemma}\label{LC1} If $C$ is a contraction on a Hilbert space $\cal H$, and there exists $0\leq\alpha < \pi/2$ such that the numerical range $\Theta(C)$ is a subset of the domain $D_\alpha$, then \begin{equation}\label{C1} \|C^n (I-C)\|\leq {K\over n+1},\ n\in{\mathbb N}. \end{equation} \end{lemma} \begin{proof} Since the operator $C$ is bounded, its spectrum is a subset of the closure of the numerical range $\overline{\Theta(C)}$. Then taking $\alpha< \alpha'< \pi/2$ we can choose a contour $\partial D_{\alpha'}$ outside $D_\alpha$, but inside the unit circle (see Figure 1), such that by the Dunford-Taylor formula one has: \begin{equation}\label{Cauchy} C^n (I-C) = {1\over 2\pi i} \int_{\partial D_{\alpha'}}{z^n (1-z)\over z-C}dz. \end{equation} By Proposition \ref{propnumrange} (see (\ref{estres})) we obtain: $\|(z-C)^{-1}\|\leq \mbox{dist}(z,D_\alpha)^{-1}$. Thus $\|(z+C)^{-1}\| \leq (\cos\alpha'\sin(\alpha' - \alpha))^{-1}$ for $|\arg z|\geq\pi/2-\alpha'$, and $\|(z+C)^{-1}\| \leq (|1-z|\sin(\alpha'-\alpha))^{-1}$ for $|\arg z|\leq \pi/2-\alpha'$. We consider the following parametrization of $\partial D_{\alpha'}$: for the arc $AB$, we take $z(t)=e^{it}\sin\alpha'$ with $\pi/2-\alpha'\leq t\leq 3\pi/2+\alpha'$, and for the straight lines $(1,A)$, $(1,B)$, we put correspondingly $z_{\mp}(s)=1 - se^{\mp i\alpha'}$ with $0\leq s\leq\cos\alpha'$. Then \begin{eqnarray} \|C^n (I-C)\| & \leq & {1\over 2\pi} \int_{{\pi\over 2}-\alpha'}^{{3\pi\over 2}+\alpha'}\!\!{|\sin\alpha'|^{n+1} |1-e^{it}\sin\alpha'|\over \cos\alpha'\sin(\alpha'-\alpha)}dt + \nonumber\\ & & + {1\over\pi} \int_0^{\cos\alpha'}{|(1-e^{i\alpha'}s)^n e^{i\alpha'}s|\over s(\sin(\alpha'-\alpha))}ds \label{para}\\ &\hspace{-4cm} \leq & \hspace{-2cm} {2(\sin\alpha')^{n+1}\over\cos\alpha'\sin(\alpha'-\alpha)} + \int_0^{\cos\alpha'}{\left((1-s\cos\alpha')^2 + s^2(\sin\alpha')^2\right)^{n/2} \over\pi\sin(\alpha'-\alpha)}ds.\nonumber \end{eqnarray} By convexity (for $0\leq s\leq\cos\alpha'$) one gets that $$ (1-s\cos\alpha')^2 + s^2(\sin\alpha')^2\leq 1-s\cos\alpha', $$ which leads to the inequality: \begin{eqnarray*} \int_0^{\cos\alpha'}\left((1-s\cos\alpha')^2 + s^2(\sin\alpha')^2\right)^{n/2}ds & \leq & \int_0^{\cos\alpha'}(1-s\cos\alpha')^{n/2}ds \\ \int_{(\sin\alpha')^2}^1 u^{n/2} {du\over\cos\alpha'} & \leq & {1-(\sin\alpha')^{n+2}\over (n/2+1)\cos\alpha'}. \end{eqnarray*} Therefore, by (\ref{para}) we get the estimate (\ref{C1}) \begin{equation} \|C^n (I-C)\| \leq {2(\sin\alpha')^{n+1}\over\cos\alpha'\sin(\alpha'-\alpha)} + 2{1-(\sin\alpha')^{n+2}\over \pi(n+2)\cos\alpha'\sin(\alpha'-\alpha)} \leq {K\over n+1}, \end{equation} where \begin{equation}\label{K} K={2\over\cos\alpha'\sin(\alpha'-\alpha)}\left({1\over\pi} - {1\over e\ln(\sin\alpha')}\right). \end{equation} \end{proof} \begin{lemma}\label{LC2} If $C$ is a contraction on a Hilbert space $\cal H$ that satisfies the estimate (\ref{C1}), then: \begin{equation}\label{C2} \|C^n-e^{n(C-I)}\|\leq {2K+2\over n^{1/3}},\ n\in {\mathbb N}\setminus\{0\}. \end{equation} \end{lemma} \begin{proof} Since the operator $C$ is bounded, we have the representation: \begin{equation}\label{somme} C^n-e^{n(C-I)} = e^{-n}\sum_{m=0}^\infty {n^m\over m!}(C^n-C^m). \end{equation} Let $\epsilon_n=n^{2/3}$, $n\geq 1$. We divide the sum (\ref{somme}) into two parts: the central part for $|m-n|\leq\epsilon_n$ and tails for $|m-n|>\epsilon_n$. We estimate tails by using the Tchebychev inequality (see e.g. \cite{Shiryaev}). Let $X$ be a Poisson random variable of parameter $n$, i.e. $\mbox{P}(X=m) = n^m e^{-n}/m!$. Then expectation $\mbox{E}(X)=n$ and variance $\mbox{Var}(X)=n$. Therefore, by the Tchebychev inequality: \begin{equation} \forall \epsilon>0,\ \mbox{P}(|X-n|>\epsilon) \leq {n\over\epsilon^2}, \end{equation} and hence \begin{equation}\label{part1} e^{-n}\sum_{|m-n|>\epsilon_n} {n^m\over m!}\|C^n-C^m\| \leq {2n\over\epsilon_n^2} = {2\over n^{1/3}}. \end{equation} To estimate the central part of the sum (\ref{somme}), where $|m-n|\leq\epsilon_n$, we use (here $[\cdot]$ denotes the integer part): \begin{eqnarray*} \|C^n-C^m\| & = & \left\|C^{n-[\epsilon_n]}\left(C^{[\epsilon_n]}-C^{m-n+[\epsilon_n]}\right) \right\|\\ & \leq & |m-n| \|C^{n-[\epsilon_n]}(I-C)\| \\ & \leq & \epsilon_n\ {K\over n-[\epsilon_n]+1}\leq {2K\over n^{1/3}}, \end{eqnarray*} and we estimate the sum by $1$. Then, together with (\ref{part1}), we obtain (\ref{C2}): \begin{equation} \|C^n-e^{n(C-I)}\| \leq {2K+2\over n^{1/3}} \end{equation} for $n\in {\mathbb N}\setminus\{0\}$. \end{proof} The next statement is an extension of the famous ``$n^{1/2}$-Lemma'' by Chernoff \cite[Lemma 2]{Chernoff}. It follows directly from Lemmata \ref{LC1}, \ref{LC2} and it gives an operator-norm estimate instead of the well-known strong convergence \cite{Chernoff}. \begin{theorem}\label{T1} Let $C$ be a sectorial contraction on $\cal H$ with numerical range $\Theta(C)\subseteq D_\alpha$, $0\leq \alpha <\pi/2$. Then \begin{equation}\label{eqT1} \left\|C^n - e^{n(C-I)}\right\| \leq {M\over n^{1/3}},\ n=1,2,3,\dots \end{equation} where $M=2K+2$ and $K$ is defined by (\ref{K}). \end{theorem} \begin{remark} If no estimate is required, then the operator-norm convergence of the difference (\ref{eqT1}) to zero follows from the Lebesgue dominated convergence theorem for the Cauchy integral for $C^n - e^{n(C-I)}$ on the contour $\partial D_{\alpha'}$, cf (\ref{Cauchy}). \end{remark} \begin{remark} If $C$ is self-adjoint and non-negative (i.e. $\alpha=0$), then directly by the spectral theorem one obtains (cf \cite{NZ1},\cite{NZ3}): \begin{equation} \|C^n (I-C)\|\leq {1\over n+1}\ \mbox{and}\ \|C^n - e^{n(C-I)}\|\leq {1\over n}. \end{equation} \end{remark} \section{Approximation Theorem}\label{appth} In this section we present the proof of our main Theorem. The first step was the Theorem \ref{T1}, the second step is an operator-norm generalization of the Trotter-Neveu-Kato convergence theorem, see e.g. \cite[Ch. 3.3]{Davies}, \cite[Ch. 1.7]{Goldstein}. In the self-adjoint case, such a generalization in the operator-norm topology is obtained in \cite[Lemma 2.1]{NZ3}. For m-sectorial operators we prove it below. \begin{lemma}\label{Trgene} Let $\{X(s)\}_{s>0}$ be a family of m-sectorial operators in a Hilbert space $\cal H$ with $\Theta(X(s))\subseteq S_\alpha$ for some $0< \alpha <\pi/2$ and for all $s>0$. Let $X_0$ be an m-sectorial operator defined in a closed subspace ${\cal H}_0\subseteq{\cal H}$, with $\Theta(X_0)\subseteq S_\alpha$. Then the following conditions are equivalent: \begin{eqnarray} (a) & & \lim_{s\rightarrow +0} \left\|(\zeta+X(s))^{-1} - (\zeta+X_0)^{-1}P_0\right\| = 0, \mbox{ for } {\mathrm Re}\,\zeta>0; \label{(a)}\\ (b) & & \lim_{s\rightarrow +0} \left\|e^{-tX(s)} - e^{-tX_0}P_0\right\| = 0, \mbox{ for } t> 0.\label{(b)} \end{eqnarray} Here $P_0$ denotes the orthogonal projection from $\cal H$ onto ${\cal H}_0$. \end{lemma} \begin{proof} $(a)\Rightarrow(b)$ \begin{figure}[H] \epsfysize=9cm $$\epsfbox{CZ4f2.eps}$$ \caption{Illustration of the path $\Gamma$.} \end{figure} Since $\{X(s)\}_{s>0}$ are m-sectorial with $\Theta(X(s))\subseteq S_\alpha$, the Dunford-Taylor formula \begin{equation} e^{-tX(s)} = {1\over 2\pi i} \int_\Gamma d\zeta {e^{t\zeta}\over\zeta + X(s)} \end{equation} defines a family of holomorphic semigroups $\{e^{-tX(s)}\}_{s>0}$ with $t\in S_{\pi/2-\alpha}$. Here $\Gamma\subset S_{\pi-\alpha}$ is a positively-oriented closed (at infinity) path around $-S_\alpha$, i.e. with the spectrum $\sigma(-X(s))\subseteq\overline{\Theta(-X(s))}$ in its interior (see Figure 2). The same is true for the operator $X_0$: \begin{equation} e^{-tX_0}P_0 = {1\over 2\pi i}\int_\Gamma d\zeta {e^{t\zeta}\over \zeta + X_0}P_0. \end{equation} We set $\Gamma= \Gamma_\epsilon \cup\Gamma_\delta \cup\overline{\Gamma_\epsilon}$, where the arc $\Gamma_\delta = \{z\in{\mathbb C}: |z|=\delta>0, |\arg z|\leq \pi-\alpha-\epsilon\}$ (for $0<\epsilon<\pi/2-\alpha$) and $\Gamma_\epsilon,\overline{\Gamma_\epsilon}$ are two conjugate straight lines, $\Gamma_\epsilon = \{z\in{\mathbb C},\ \arg z = \pi-\alpha-\epsilon,\ |z|\geq \delta\}$. Then for $t>0$ one gets \begin{equation}\label{esa1} \left\| e^{-tX(s)} - e^{-tX_0}P_0\right\| \leq {1\over 2\pi} \int_\Gamma |d\zeta||e^{t\zeta}| \left\|(\zeta+X(s))^{-1} - (\zeta+X_0)^{-1}P_0\right\|. \end{equation} Since operators $X(s)$ and $X_0$ are m-sectorial, by Proposition \ref{propnumrange} one gets the estimate \begin{equation} \left\|(\zeta+X(s))^{-1} - (\zeta+X_0)^{-1}P_0\right\| \leq {2\over\mbox{dist}(\zeta, -S_\alpha)}, \end{equation} which implies \begin{equation}\label{esa2} \sup_{\zeta\in\Gamma} \left\|(\zeta+X(s))^{-1} - (\zeta+X_0)^{-1}P_0\right\| \leq {2\over \delta \sin\epsilon}. \end{equation} Since for the path $\Gamma = \Gamma_\epsilon\cup\Gamma_\delta\cup\overline{\Gamma_\epsilon}$, the integral $I_t = \int_\Gamma |d\zeta| |e^{t\zeta}|$ converges for any $t>0$, then condition $(a)$, estimates (\ref{esa1}), (\ref{esa2}) and the Lebesgue dominated convergence theorem proves $(b)$. $(b)\Rightarrow (a)$\\ By the Laplace transform for semigroups $e^{-tX(s)}$ and $e^{-tX_0}$, one gets for $\mbox{Re}\,\zeta>0$: % \begin{equation}\label{Laplace} (\zeta+X(s))^{-1} - (\zeta + X_0)^{-1}P_0 = \int_0^\infty dt e^{-t\zeta}\left(e^{-tX(s)} - e^{-tX_0}P_0\right). \end{equation} Since the generators are m-sectorial with vertex at $0$, the corresponding semigroups are contractions: $\|e^{-tX(s)}\|\leq 1$, $\|e^{-tX_0}P_0\|\leq 1$ for $t\in S_{\pi/2-\alpha}$, $s>0$. Then the Lebesgue dominated convergence theorem ensures $(a)$ for $\mbox{Re }\zeta>0$, due to the estimate \begin{eqnarray} & & \left\|(\zeta + X(s))^{-1} - (\zeta +X_0)^{-1}P_0\right\| \leq \nonumber\\ & & \hspace{2cm} \int_0^\infty dt e^{-t{\mathrm Re}\,\zeta} \left\|e^{-tX(s)} - e^{-tX_0}P_0\right\| \end{eqnarray} and the pointwise $\|.\|$-convergence $(b)$. \end{proof} \begin{remark} The Vitali theorem concerning the convergence of holomorphic (operator-valued) functions (see e.g. \cite[Theorem 3.14.1]{HP}) allows to extend the pointwise convergences in $(a)$ and in $(b)$ because the corresponding functions are holomorphic and uniformly bounded. In particular, we can replace the conditions $(a)$ and $(b)$ respectively by \begin{eqnarray*} & (\tilde{a}) & \lim_{s\rightarrow +0} \left\|(\zeta+X(s))^{-1} - (\zeta+X_0)^{-1}P_0\right\| = 0, \\ & & \mbox{ uniformly on compacts in } S_{\pi-\alpha};\\ & (\tilde{b}) & \lim_{s\rightarrow +0} \left\|e^{-tX(s)} - e^{-tX_0}P_0\right\| = 0, \\ & & \mbox{ uniformly on compacts in } S_{\pi/2-\alpha}. \end{eqnarray*} \end{remark} In fact, we can prove the equivalence in Lemma \ref{Trgene} under weaker conditions: \begin{corollary}\label{approx} Let $\{X(s)\}_{s>0}$ be a family of m-sectorial operators with $\Theta(X(s))\subseteq S_\alpha$ for some $0< \alpha <\pi/2$ and for all $s>0$. Let $X_0$ be a closed operator in ${\cal H}_0\subseteq{\cal H}$ with non-empty resolvent set. Then the following conditions are equivalent: \begin{eqnarray*} (a') & & \lim_{s\rightarrow +0} \left\|(\zeta+X(s))^{-1} - (\zeta+X_0)^{-1}P_0\right\| = 0, \mbox{ for some } \zeta\in S_{\pi-\alpha};\\ (b') & & \lim_{s\rightarrow +0} \left\|e^{-tX(s)} - e^{-tX_0}P_0\right\| = 0, \mbox{ for t in a subset of } (0,+\infty) \\ \end{eqnarray*} having a limit point, if $X_0$ is generator of a semigroup, i.e. $e^{-tX_0}$ is well defined at least for $t\geq 0$. \end{corollary} \begin{proof} It is sufficient to check that condition $(a')$ implies: $(a)$ holds and the operator $X_0$ is such that $\Theta(X_0)\subseteq S_\alpha$ for some $\alpha\in (0,\pi/2)$, and that condition $(b')$ implies $(b)$ for the same operator $X_0$. By \cite[Ch.IV, Theorem 2.25]{Kato}, the condition $(a')$ is sufficient to have a generalized convergence in the sense of Kato. On the other hand, this generalized convergence implies that $(a')$ holds for any $\zeta$ in the resolvent set of $X_0$. Since it is non-empty and open, it has a limit point. Then we can apply the theorem of Vitali, since the norms $\|(\zeta + X(s))^{-1}\|$ are uniformly bounded in $s$ for $\zeta-\epsilon \in S_{\pi-\alpha}$ (for any fixed $\epsilon>0$). Thus we obtain the pointwise convergence for any $\zeta\in S_{\pi-\alpha}$, and in particular, $(a)$ holds for the operator $X_0$. In order to show that $X_0$ is necessarily m-sectorial with $\Theta(X_0)\subseteq S_\alpha$, we use that the norm-convergence (\ref{(a)}) implies: \begin{equation} \|(\zeta+X_0)^{-1}P_0\| = \lim_{s\rightarrow 0}\|(\zeta+X(s))^{-1}\| \leq {1\over\mbox{dist}(\zeta, -S_\alpha)}. \end{equation} In the case $\mbox{Re}\,\zeta>0$, we obtain ${\|(\zeta+X_0)^{-1}\|}_0\leq 1/\mbox{Re}\,\zeta$, where ${\|\cdot\|}_0$ is the operator-norm in ${\cal H}_0$. This estimate implies that $X_0$ generates a contraction semigroup, and then is m-accretive. By the same way one checks that $e^{i\varphi}X_0$ is m-accretive for any $\alpha-\pi/2 \leq\varphi\leq \pi/2-\alpha$, which proves that $X_0$ is m-sectorial with semi-angle $\alpha$ and with vertex at $0$. The implication $(b')\Rightarrow (b)$ is a direct consequence of the Vitali theorem. By $\|e^{-tX(s)}\|\leq 1$ for any $t\in S_{\pi/2-\alpha}$ and $s>0$, and by $(b')$ one gets that the functions $\{e^{-tX(s)}\}_{s>0}$ converge in operator-norm as $s\rightarrow +0$ for any $t\in S_{\pi/2-\alpha}$. Therefore, it follows that the limit $e^{-tX_0}$ is a holomorphic semigroup with $\|e^{-tX_0}\|_0 = \|e^{-tX_0}P_0\|\leq 1$ for $t\in S_{\pi/2-\alpha}$, and then operator $X_0$ is m-sectorial with semi-angle $\alpha$ and with vertex at $0$, in ${\cal H}_0 = P_0{\cal H}$. \end{proof} It is well-known that the convergence in the strong resolvent sense is well-adapted to the problem of semigroup approximation in the strong topology (see e.g. \cite[Ch. 3.3]{Davies} or \cite[Ch. 1.7]{Goldstein}). Similarly, the above results show that the convergence in the uniform resolvent sense (i.e. the generalized convergence in the sense of Kato \cite[Ch. IV.6]{Kato}) is well-adapted to the semigroup approximation in the operator-norm topology. This topology is natural for holomorphic contraction semigroups, which are continuous in the operator norm outside $0$. We show that the corresponding class of generators (the m-sectorial operators) is in some sense stable for the convergence we consider. Now we can prove our main Theorem: \begin{theorem}\label{T} Let $\{\Phi(s)\}_{s\geq 0}$ be a family of sectorial contractions on a Hilbert space $\cal H$. Let there exists $0<\alpha<\pi/2$ such that $\Theta(\Phi(s)) \subseteq D_\alpha$, for all $s\geq 0$. Let $X(s)=(I-\Phi(s))/s$, and let $X_0$ be a closed operator with non-empty resolvent set, defined in a closed subspace ${\cal H}_0\subseteq{\cal H}$. Then the family $\{X(s)\}_{s>0}$ converges in the uniform resolvent sense to $X_0$ as $s\rightarrow +0$ if and only if \begin{equation}\label{lim} \lim_{n\rightarrow \infty} \left\|\Phi(t/n)^n -e^{-tX_0}P_0\right\| = 0,\ t>0. \end{equation} Here $P_0$ denotes the orthogonal projection onto ${\cal H}_0$.\end{theorem} \begin{proof} Necessity: Notice that $\Theta(\Phi(s))\subseteq D_\alpha$ means that $\Theta(X(s)) \subseteq S_\alpha$ for any $s>0$. Thus, the convergence of $X(s)$ in the uniform resolvent sense coincides with the condition $(a)$ of Lemma \ref{Trgene}. Since \begin{eqnarray} \|\Phi(t/n)^n - e^{-tX_0}P_0\| & \leq & \nonumber\\ & & \hspace{-3cm} \|\Phi(t/n)^n - e^{-n(I-\Phi(t/n))}\| + \|e^{-n(I-\Phi(t/n))} - e^{-tX_0}P_0\|,\label{eqT} \end{eqnarray} the assertion follows from Theorem \ref{T1} for $C=\Phi(t/n)$ and part $(a)\Rightarrow(b)$ Lemma \ref{Trgene}. Sufficiency: We have to estimate the difference \begin{eqnarray} \|e^{-tX(t/n)} - e^{-tX_0}P_0\| & \leq & \nonumber\\ & & \hspace{-3cm} \|e^{-tX(t/n)} - \Phi(t/n)^n\| + \|\Phi(t/n)^n - e^{-tX_0}P_0\|,\ t>0.\label{eq34} \end{eqnarray} The first term in the right-hand side of (\ref{eq34}) is estimated by Theorem \ref{T1} for $C=\Phi(t/n)$. The second term is supposed to tend to zero by (\ref{lim}). Then the assertion follows from the part $(b)\Rightarrow (a)$ of Lemma \ref{Trgene}. \end{proof} \begin{remark} In fact, it is sufficient to have $(a')$, in order to obtain (\ref{lim}). On the other hand, it is sufficient to have (\ref{lim}) for $t$ in a subset of $(0,+\infty)$ having a limit point, in order to obtain the convergence $(a)$. This, as well as the fact that convergence (\ref{lim}) implies that $e^{-tX_0}$ is a holomorphic contraction semigroup for $t\in S_{\pi/2-\alpha}$, follows from the Corollary \ref{approx}. \end{remark} \section{Applications.} \subsection{Error estimate for the Euler approximation of semigroups} Let $A$ be an m-sectorial operator with a semi-angle $\alpha$ and with vertex at $0$, $0< \alpha<\pi/2$. Then the operators $F(t) = (I+tA)^{-1}$, $t\geq 0$ are sectorial contractions, i.e. $\Theta(F(t))\subseteq D_\alpha$ (cf Section \ref{cfex1}). Let $X(s) = (I-F(s))/s$, $s>0$, and $X_0=A$. Then $X(s)$ converges when $s\rightarrow +0$ to $X_0$ in the uniform resolvent sense: \begin{equation}\label{cond1} \|(\zeta+X(s))^{-1} - (\zeta+X_0)^{-1}\| = s\left\| {A\over\zeta+A+\zeta sA}\cdot{A\over\zeta+A}\right\| = O(s),\ \zeta\in S_{\pi-\alpha}, \end{equation} since we have the estimate \begin{equation}\label{cond1'} \left\| {A\over\zeta+A+\zeta sA}\cdot{A\over\zeta+A}\right\| \leq \left(1+{|\zeta|\over \mbox{dist}\left({\zeta (1+s\zeta)^{-1}},-S_\alpha\right)}\right) \left(1+{|\zeta|\over \mbox{dist} (\zeta,-S_\alpha)}\right). \end{equation} Therefore, the family $\{F(t)\}_{t\geq 0}$ satisfies conditions of Theorem \ref{T}. This gives the operator-norm approximation of the exponential function (semigroup for m-sectorial generator) by powers of resolvent (Euler's approximation): \begin{theorem}\label{ex1} If $A$ is an m-sectorial operator in a Hilbert space $\cal H$, with semi-angle $\alpha\in (0,\pi/2)$ and with vertex at $0$, then \begin{equation}\label{cor1.1} \lim_{n\rightarrow\infty}\left\|(I+tA/n)^{-n} - e^{-tA}\right\| = 0,\ t\in S_{\pi/2-\alpha}. \end{equation} Moreover, if $0$ belongs to the resolvent set of $A$, then uniformly in $t\geq 0$ one has the error estimate: \begin{equation}\label{cor1.2} \left\|(I+tA/n)^{-n} - e^{-tA}\right\| \leq O\left(\frac{\ln n}{n}\right). \end{equation} \end{theorem} \begin{proof} The convergence (\ref{cor1.1}) follows directly from Theorem \ref{T}. To obtain (\ref{cor1.2}), we use the representation: \begin{eqnarray} & & (I+tA/n)^{-n} - e^{-tA} = \nonumber\\ & & \hspace{0.5cm} \sum_{m=0}^{n-1} (I+tA/n)^{-(n-m-1)} \left((I+tA/n)^{-1} - e^{-tA/n}\right) e^{-mtA/n}.\label{tele} \end{eqnarray} By Lemmata 1-3 of \cite{CZ1}, we have the estimates $\|A^{-1} \left((I+tA/n)^{-1} - \right.$ $\left. e^{-tA/n}\right)\| \leq 2t/n$ and $\|A^{-1} \left((I+tA/n)^{-1} - e^{-tA/n}\right) A^{-1}\|\leq 3/2 (t/n)^2$. Since by Proposition \ref{P1} the operator $A$ generates a holomorphic semigroup, a standard result (see e.g. \cite[Ch. IX, \S 1.6]{Kato}) says that $\|Ae^{-\tau A}\| \leq C_A /\tau$ for some constant $C_A>0$ and any $\tau>0$. Finally, for any integer $k\geq 1$ we have the identity: \begin{equation} (I+tA/n)^{-k}A = \frac{n}{t}\left(I- (I+tA/n)^{-1}\right)(I+tA/n)^{-k+1}, \end{equation} which leads to the estimate $\|(I+tA/n)^{-k}A\| \leq nK/kt$ by Lemma \ref{LC1} for the sectorial contraction $C= (I+tA/n)^{-1}$. By these estimates and (\ref{tele}) one gets for $n\geq 3$: \begin{eqnarray} \|(I+tA/n)^{-n} - e^{-tA}\| & \leq & \frac{nK}{(n-1)t}(2t/n) + (2t/n) \frac{C_A n}{(n-1)t} \nonumber \\ & & + \sum_{m=1}^{n-2} \frac{nK}{(n-m-1)t} \frac{3t^2}{2n^2} \frac{C_A n}{mt}. \end{eqnarray} Therefore, uniformly in $t\geq 0$ we obtain: \begin{equation} \|(I+tA/n)^{-n} - e^{-tA}\| \leq \frac{4(K+C_A)}{n} + 6 K C_A \frac{\ln n}{n}, \end{equation} which proves (\ref{cor1.2}). \end{proof} \subsection{Arithmetic mean approximation} Let $A_1, \dots, A_k$ be m-sectorial operators in $\cal H$, such that $\Theta(A_j) \subseteq S_\alpha$ for some $\alpha\in (0, \pi/2)$ and for any $1\leq j \leq k$. By $a_1,\dots, a_k$ we denote the corresponding closed sectorial forms. Similar to the Trotter product formula, there are expressions of the type $F_k(t/n)^n$ that approximate the semigroup generated by the form-sum of $A_1,\dots, A_k$. $F_k(t)$ can be taken as a product $f_1(tA_1)\dots f_k(tA_k)$ or as an arithmetic mean $k^{-1}(f_1(ktA_1) + \dots + f_k(ktA_k))$, where $\{f_l\}_{1\leq l\leq k}$ are so-called Kato-functions (see e.g. \cite{K}-\cite{NZ1}). Section \ref{sec1} shows that our condition on the numerical range for sectorial contractions (cf Definition \ref{SC}) is easily satisfied if $F_k(t)$ is the arithmetic mean of resolvents or semigroups generated by $\{A_j\}_{1\leq j\leq k}$. Let us consider these families of contractions (cf \cite{Lapidus}): \begin{eqnarray} F_k^{res}(t) & = & \frac{1}{k}\sum_{l=1}^k (I+ ktA_l)^{-1}, t\geq 0,\label{Fres}\\ F_k^{sem}(t) & = & \frac{1}{k}\sum_{l=1}^k e^{-ktA_l}, t\geq 0.\label{Fsem} \end{eqnarray} Notice that $f(x) = (1+x)^{-1}$ or $f(x) = e^{-x}$ are Kato-functions. Since by conditions on $\{A_j\}_{1\leq j\leq k}$ the operators $\{(I+ktA_l)^{-1}\}_{1\leq j\leq k}$ or $\{e^{-ktA_l}\}_{1\leq j\leq k}$ are sectorial contractions with a semi-angle $\alpha$ (see Section \ref{sec1}), and since $D_\alpha$ is convex, $F_k^{res,sem}(t)$ are families of sectorial contractions with a semi-angle $\alpha$. Moreover, it is shown \cite{Lapidus} that the families $F_k^{res,sem}(t/n)^n$ converge strongly to $e^{-tX_0}P_0$ for any $t\geq 0$ and $k\geq 1$, as $n\rightarrow \infty$, where $X_0$ is the form-sum of $A_1,\dots, A_k$ and $P_0$ is the orthogonal projector on the closure of $\bigcap_{l=1}^k {\cal D}(a_l)$. Therefore, by Theorem \ref{T} we obtain a necessary and sufficient condition for the operator-norm convergence of the Trotter-type formula with arithmetic mean of resolvents (\ref{Fres}) or semigroups (\ref{Fsem}): \begin{theorem}\label{ex2} If $A_1,\dots, A_k$ are m-sectorial operators in $\cal H$ such that for $1\leq j\leq k$, $\Theta(A_j) \subseteq S_\alpha$ for some $\alpha\in(0,\pi/2)$, then $X_k(t) = (I-F_k^{res,sem}(t))/t$ converges (as $t\rightarrow +0$) in the uniform resolvent sense to $X_0 = A_1 \dot{+} \dots \dot{+} A_k$ defined in the closed subspace ${\cal H}_0 = \overline{\bigcap_{l=1}^k {\cal D}(a_l)}$, if and only if \begin{equation} \lim_{n\rightarrow\infty} \left\| F_k^{res,sem}(t/n)^n - e^{-tX_0}P_0 \right\| = 0. \end{equation} Here $P_0$ is the orthogonal projector on ${\cal H}_0$. \end{theorem} If $k=1$, Theorem \ref{ex2} reduces to Theorem \ref{ex1}, where the condition of convergence in the uniform resolvent sense is ensured by (\ref{cond1}). \subsection{Trotter-Kato product formula} Let $A$ be a non-negative self-adjoint operator and let $B$ be m-sectorial with $\Theta(B) \subseteq S_\alpha$ for some angle $0< \alpha< \pi/2$. Then our main Theorem \ref{T} gives necessary and sufficient conditions for the operator-norm convergence of a generalized Trotter product formula. Let us consider the family of operators \begin{equation}\label{Phi} \Phi(s) = f(sA)^{1/2}g(sB)f(sA)^{1/2}, s\geq 0, \end{equation} where $f(t)$ and $g(t)$ are the Kato-functions $(1+t)^{-1}$ or $e^{-t}$. Since $f(sA)^{1/2}$ is self-adjoint and $\|f(sA)^{1/2}\|\leq 1$, by results of Section \ref{sec1} one has: \begin{equation} (u,\Phi(s)u) = (f(sA)^{1/2}u,g(sB)f(sA)^{1/2}u) \in D_\alpha, \end{equation} \begin{eqnarray} & & (u,(I-\Phi(s))u) = (u,(I-f(sA))u) + \nonumber \\ & & \hspace{1cm} + (f(sA)^{1/2}u,(I-g(sB))f(sA)^{1/2}u)\in S_\alpha,\label{48} \end{eqnarray} for any $u\in\cal H$, $\|u\|=1$, which proves that $\Theta(\Phi(s))\subseteq D_\alpha$. Therefore, Theorem \ref{T} gives necessary and sufficient conditions for the operator-norm convergence of $\Phi(t/n)^n$ as $n\rightarrow\infty$. At this point we still not identify the limit $X_0$ via operators $A$ and $B$. But the case is similar to that of \cite[Addendum]{K}. In \cite{K}, Kato proved that for $n\rightarrow \infty$ the generalized Trotter product formula $(f(tA/n)g(Bt/n))^n$ converges strongly to $e^{-tC}P_0$, where $C=A\dot{+}B$ is the form-sum defined in the closure ${\cal H}_0$ of ${\cal D}={\cal D}(A^{1/2})\cap{\cal D}(B^{1/2})$ and $P_0$ is the orthogonal projection onto ${\cal H}_0$, for any non-negative self-adjoint operators $A$ and $B$. One can easily check that this result is also valid for the case of the symmetrized formula $f(tA)^{1/2}g(tB)f(tA)^{1/2}$. Moreover in the Addendum of \cite{K}, it is shown that for $f(t) = g(t) = e^{-t}$ the strong convergence holds also for m-sectorial generators $A$ and $B$. Let $a$ and $b$ be the associated closed m-sectorial forms. Then $A\dot{+}B$ is defined in the closure ${\cal H}_0$ of ${\cal D}(a)\cap{\cal D}(b)$. By the same arguments one checks, that this result is valid for $f(tA)^{1/2}g(tB)f(tA)^{1/2}$, when $f(t)$, $g(t)$ ! are $(1+t)^{-1}$ or $e^{-t}$. Therefore, $\Phi(t/n)^n$ (\ref{Phi}) converges strongly to $e^{-t(A\dot{+}B)}P_0$, where $P_0$ is the orthogonal projector onto ${\cal H}_0$, as $n\rightarrow \infty$. Taking into account these observations and our Theorem \ref{T}, we obtain the following \begin{theorem}\label{ex3} Let $A$ be a non-negative self-adjoint operator, and let $B$ be an m-sectorial operator in $\cal H$. Let the functions $f(t)$ and $g(t)$ be $(1+t)^{-1}$ or $e^{-t}$. For $s\geq 0$ we put $X(s)=s^{-1}(I-f(sA)^{1/2}g(sB)f(sA)^{1/2})$. Let $C=A\dot{+}B$ be the form-sum of $A$ and $B$ defined in the closure ${\cal H}_0$ of ${\cal D}(a)\cap{\cal D}(b)$, and let $P_0$ be the orthogonal projection onto ${\cal H}_0$. Then the following assertions are equivalent: \begin{center} (i) $X(s)$ converges in the uniform resolvent sense to $C$ as $s\rightarrow +0$, \end{center} \begin{equation}\label{sgf} (ii)\ \lim_{n\rightarrow\infty} \left\| \left(f(tA/n)^{1/2}g(tB/n)f(tA/n)^{1/2}\right)^n - e^{-tC}P_0\right\| = 0. \end{equation} \end{theorem} Notice that in contrast to Theorem \ref{ex1}, see (\ref{cor1.2}), Theorem \ref{ex3} gives no error bound estimate for the rate of convergence in (\ref{sgf}). For self-adjoint semigroups the operator-norm convergence of the Trotter-Kato product formula without error bound estimate is established for a general case of Kato-function $f$ and $g$, see \cite{NZ3}. In particular, there it is shown that compactness of $(I+A)^{-1}$ or $(I+A)^{-1}(I+B)^{-1}$ is a sufficient condition of this convergence for non-negative self-adjoint generators $A$ and $B$. Recently \cite{CZ2} we generalized these results to m-sectorial $A$ and $B$ for the Trotter product formula, i.e. for $f(t) = g(t) = e^{-t}$. Another strategy to prove the Trotter-Kato product formula for semigroups is based on operator-norm error-bound estimates. After pioneering papers \cite{R}, \cite{IT}, \cite{NZ1} for couples of self-adjoint generators $A$ and $B$, these results have been recently generalized to nonself-adjoint semigroups. In our paper \cite{CZ1} it is done in Banach and Hilbert spaces for holomorphic contraction semigroups under Ichinose-Tamura conditions on generators, cf \cite{IT}. Another result of this kind is established in \cite{CNZ}. Let $A$ be a non-negative self-adjoint operator. If $B$ is m-accretive, then sufficient conditions for the operator-norm convergence of the Trotter product formula with the error bound $O(\ln n/n)$ are the following: $\|Bu\|\leq a\|Au\|$ and $\|B^*u\|\leq a_* \|Au\|$ for $u\in {\cal D}(A)$, $a<1$ , $a_*<1$. Notice that methods of \cite{CNZ} can be adapted without supplementary conditions to the generalized formula $f(tA/n)^{1/2} g(tB/n) f(tA/n)^{1/2}$ for $! f(t)$ and $g(t)$ be $(1+t)^{-1}$ or $e^{-t}$. Fractional powers conditions on self-adjoint generators ensuring the operator-norm convergence of the Trotter-Kato with error-bound estimates are discussed in \cite{NZ2}. For a generalization to m-sectorial generators see \cite{CZ2}. \section{Conclusion} In the present paper we extend the Chernoff theory to approximation of semigroups in the operator-norm topology in a Hilbert space. The condition for this operator-norm convergence is that the operators $\{\Phi(s)\}_{s\geq 0}$ are not only contractions, but sectorial contractions: they have their numerical ranges $\Theta(\Phi(s))$ in a domain $D_\alpha$, $0\leq\alpha<\pi/2$, uniformly in $s\geq 0$, see (\ref{Dalpha}). This allows us to find the $1/n^{1/3}$-estimate in operator-norm for an analogue of the Chernoff lemma (see Theorem \ref{T1}). Notice that the original Chernoff's estimate in this lemma is not convergent: $\|(C^n - e^{n(C-I)})u\|\leq n^{1/2}\|(C-I)u\|$, for any vector $u$. The next step is the analogue of the Trotter-Neveu-Kato approximation theorem in the operator-norm, see Lemma \ref{Trgene}. We consider uniformly m-sectorial families of generators, which leads to uniformly bounded holomorphic semigroups. The Corollary \ref{approx} shows that the convergence in! the uniform resolvent sense (i.e. in the generalized convergence in the sense of Kato \cite[Ch. IV]{Kato}) is well-adapted to approximate semigroups in the operator-norm topology. On the other hand, this topology is natural for holomorphic contraction semigroups: they are continuous in this topology for $t>0$, and generators of these semigroups are exactly m-sectorial operators with semi-angles $\alpha<\pi/2$. A first consequence of our main Theorem \ref{T} is that the Euler approximation formula for the exponential function converges in operator-norm for any m-sectorial generator (Theorem \ref{ex1}). This operator-norm approximation theory leads to necessary and sufficient conditions for convergence of Trotter-type formul{\ae}. We formulate these conditions for the formula with arithmetic mean of resolvents or semigroups (see Theorem \ref{ex2}) and for the symmetrized product formula $f(tA)^{1/2}g(tB)f(tA)^{1/2}$ where $f(t),g(t) = (1+t)^{-1}$ or $e^{-t}$, see Theorem \ref{ex3}. Our condition on the numerical range $\Theta(\Phi(s))\subseteq D_\alpha$ (sectorial contraction) seems to be naturally related to the m-sectorial operators. In fact we have verified it for the resolvent $(I+tA)^{-1}$, for the semigroup $e^{-tA}$ generated by an m-sectorial operator $A$ (see Section \ref{sec1}), and for the family $\Phi(s) = f(sA)^{1/2}g(sB)f(sA)^{1/2}$, see (\ref{Phi})-(\ref{48}). %% End of article: %% optional % Appendices % Appendix without title: %\appendix{} % Appendix with title: %\appendix{Title} % Appendix with letter: %\appendix{B} % Appendix with letter and title: %\appendix{C} %\appendixtitle{This is an appendix title} %% optional %\begin{acknowledgment} %text... %\end{acknowledgment} %% Not optional, necessary: \begin{thebibliography}{99} \bibitem{CZ1} V. Cachia; V. A. Zagrebnov: ``Operator-norm convergence of the Trotter product formula for holomorphic semigroups'', to appear in {\em J. Operator Theory} (2000). \bibitem{CZ2} V. Cachia; V. A. Zagrebnov: ``Operator-norm convergence of the Trotter product formula for sectorial generators'', to appear in {\em Lett. Math. Phys.} (2000). \bibitem{CNZ} V. Cachia; H. Neidhardt; V. A. Zagrebnov: ``Accretive perturbations and error estimates for the Trotter product formula''. to appear in {\em Integr. Equ. Oper. Theory} (2000). \bibitem{Chernoff} P. R. Chernoff: ``Note on Product Formulas for Operator Semigroups'', {\it J. Functional Analysis} {\bf 2} (1968), 238-242. \bibitem{Davies} E. B. Davies: {\it One parameter semigroups}, Academic Press, London 1980. \bibitem{Goldstein} J. A. Goldstein: {\it Semigroups of linear operators and applications}, Oxford University Press, Oxford 1985. \bibitem{HP} E. Hille; R.S. Phillips: Functional Analysis and Semigroups, {\it Amer. Math. Soc. colloquium publications,} vol. {\bf 31} (1957). \bibitem{IT} T. Ichinose; H. Tamura: ``Error estimate in operator-norm for Trotter-Kato product formula'', {\it Integr. Equ. Oper. Theory} {\bf 27} (1997), 195-207. \bibitem{K1} T. Kato: ``Some mapping theorems for the numerical range'', {\it Proc. Japan Acad.} {\bf 41} (1966), 652-655. \bibitem{Kato} T. Kato: {\it Perturbation theory for linear operators}, Springer Verlag, Berlin 1966. \bibitem{K} T. Kato: ``Trotter's Product Formula for Arbitrary Pair of Self-Adjoint Contraction Semigroups''. {\it Topics in functional analysis Advances in mathematics supplementary studies,} vol. {\bf 3} (1978), 185-195. \bibitem{Lapidus} M. Lapidus: ``G\'en\'eralisation de la formule de Trotter-Lie''. {\it C. R. Acad. Sc. Paris} {\bf 291} A (1980), 497-500. \bibitem{NZ1} H. Neidhardt; V.A. Zagrebnov: ``On error estimates for the Trotter-Kato product formula''. {\it Lett. Math. Phys.} {\bf 44} (1998), 169-186. \bibitem{NZ2} H. Neidhardt; V.A. Zagrebnov: ``Fractional powers of self-adjoint operators and Trotter-Kato product formula'', {\it Integr. Equ. Oper. Theory,} {\bf 35} (1999), 209-231. \bibitem{NZ3} H. Neidhardt; V.A. Zagrebnov: ``Trotter-Kato product formula and operator-norm convergence'', {\it Comm. Math. Phys.} {\bf 205} (1999), 129-159. \bibitem{R} D.L. Rogava: ``Error bounds fo Trotter-type formulas for self-adjoint operators'', {\it Funct. Anal. Application} {\bf 27}, No.3 (1993), 217-219. \bibitem{Shiryaev} A. N. Shiryaev: {\it Probability}, Springer Verlag, Berlin 1984. \bibitem{Trotter1} H. F. Trotter: ``Approximation of semigroups of operators'', {\it Pacific J. Math.} {\bf 8} (1958), 545-551. \bibitem{Trotter} H. F. Trotter: ``On the product of semigroups of operators'', {\it Proc. Amer. Math. Soc.} {\bf 10} (1959), 545-551. \end{thebibliography} \end{document}