BABAR-PUB-05/12

SLAC-PUB-11137
hep-ex/0504047
Search for b → u transitions in B − → D0 K − and B − → D∗0 K −
B. Aubert, R. Barate, D. Boutigny, F. Couderc, Y. Karyotakis, J. P. Lees, V. Poireau, V. Tisserand, and A. Zghiche
Laboratoire de Physique des Particules, F-74941 Annecy-le-Vieux, France

E. Grauges
IFAE, Universitat Autonoma de Barcelona, E-08193 Bellaterra, Barcelona, Spain

A. Palano, M. Pappagallo, and A. Pompili

Università di Bari, Dipartimento di Fisica and INFN, I-70126 Bari, Italy

J. C. Chen, N. D. Qi, G. Rong, P. Wang, and Y. S. Zhu

Institute of High Energy Physics, Beijing 100039, China

G. Eigen, I. Ofte, and B. Stugu

University of Bergen, Inst. of Physics, N-5007 Bergen, Norway

G. S. Abrams, M. Battaglia, A. W. Borgland, A. B. Breon, D. N. Brown, J. Button-Shafer,

R. N. Cahn, E. Charles, C. T. Day, M. S. Gill, A. V. Gritsan, Y. Groysman, R. G. Jacobsen,
R. W. Kadel, J. Kadyk, L. T. Kerth, Yu. G. Kolomensky, G. Kukartsev, G. Lynch, L. M. Mir,
P. J. Oddone, T. J. Orimoto, M. Pripstein, N. A. Roe, M. T. Ronan, and W. A. Wenzel
Lawrence Berkeley National Laboratory and University of California, Berkeley, California 94720, USA

M. Barrett, K. E. Ford, T. J. Harrison, A. J. Hart, C. M. Hawkes, S. E. Morgan, and A. T. Watson

University of Birmingham, Birmingham, B15 2TT, United Kingdom

M. Fritsch, K. Goetzen, T. Held, H. Koch, B. Lewandowski, M. Pelizaeus, K. Peters, T. Schroeder, and M. Steinke
Ruhr Universität Bochum, Institut für Experimentalphysik 1, D-44780 Bochum, Germany

J. T. Boyd, J. P. Burke, N. Chevalier, W. N. Cottingham, and M. P. Kelly

University of Bristol, Bristol BS8 1TL, United Kingdom

T. Cuhadar-Donszelmann, C. Hearty, N. S. Knecht, T. S. Mattison, and J. A. McKenna

University of British Columbia, Vancouver, British Columbia, Canada V6T 1Z1

A. Khan, P. Kyberd, and L. Teodorescu

Brunel University, Uxbridge, Middlesex UB8 3PH, United Kingdom

A. E. Blinov, V. E. Blinov, A. D. Bukin, V. P. Druzhinin, V. B. Golubev, V. N. Ivanchenko,

E. A. Kravchenko, A. P. Onuchin, S. I. Serednyakov, Yu. I. Skovpen, E. P. Solodov, and A. N. Yushkov
Budker Institute of Nuclear Physics, Novosibirsk 630090, Russia

D. Best, M. Bondioli, M. Bruinsma, M. Chao, I. Eschrich, D. Kirkby,

A. J. Lankford, M. Mandelkern, R. K. Mommsen, W. Roethel, and D. P. Stoker
University of California at Irvine, Irvine, California 92697, USA

C. Buchanan, B. L. Hartfiel, and A. J. R. Weinstein

University of California at Los Angeles, Los Angeles, California 90024, USA

S. D. Foulkes, J. W. Gary, O. Long, B. C. Shen, K. Wang, and L. Zhang

University of California at Riverside, Riverside, California 92521, USA

D. del Re, H. K. Hadavand, E. J. Hill, D. B. MacFarlane, H. P. Paar, S. Rahatlou, and V. Sharma

University of California at San Diego, La Jolla, California 92093, USA

Submitted to Physical Review D

Work supported in part by the Department of Energy contract DE-AC02-76SF00515

 2

J. W. Berryhill, C. Campagnari, A. Cunha, B. Dahmes,

T. M. Hong, A. Lu, M. A. Mazur, J. D. Richman, and W. Verkerke
University of California at Santa Barbara, Santa Barbara, California 93106, USA

T. W. Beck, A. M. Eisner, C. J. Flacco, C. A. Heusch, J. Kroseberg, W. S. Lockman, G. Nesom,

T. Schalk, B. A. Schumm, A. Seiden, P. Spradlin, D. C. Williams, and M. G. Wilson
University of California at Santa Cruz, Institute for Particle Physics, Santa Cruz, California 95064, USA

J. Albert, E. Chen, G. P. Dubois-Felsmann, A. Dvoretskii, D. G. Hitlin,

I. Narsky, T. Piatenko, F. C. Porter, A. Ryd, and A. Samuel
California Institute of Technology, Pasadena, California 91125, USA

R. Andreassen, S. Jayatilleke, G. Mancinelli, B. T. Meadows, and M. D. Sokoloff

University of Cincinnati, Cincinnati, Ohio 45221, USA

F. Blanc, P. Bloom, S. Chen, W. T. Ford, U. Nauenberg, A. Olivas, P. Rankin,

W. O. Ruddick, J. G. Smith, K. A. Ulmer, S. R. Wagner, and J. Zhang
University of Colorado, Boulder, Colorado 80309, USA

A. Chen, E. A. Eckhart, J. L. Harton, A. Soffer, W. H. Toki, R. J. Wilson, and Q. Zeng

Colorado State University, Fort Collins, Colorado 80523, USA

B. Spaan
Universität Dortmund, Institut fur Physik, D-44221 Dortmund, Germany

D. Altenburg, T. Brandt, J. Brose, M. Dickopp, E. Feltresi, A. Hauke, V. Klose, H. M. Lacker, E. Maly,

R. Nogowski, S. Otto, A. Petzold, G. Schott, J. Schubert, K. R. Schubert, R. Schwierz, and J. E. Sundermann
Technische Universität Dresden, Institut für Kern- und Teilchenphysik, D-01062 Dresden, Germany

D. Bernard, G. R. Bonneaud, P. Grenier, S. Schrenk, Ch. Thiebaux, G. Vasileiadis, and M. Verderi

Ecole Polytechnique, LLR, F-91128 Palaiseau, France

D. J. Bard, P. J. Clark, W. Gradl, F. Muheim, S. Playfer, and Y. Xie

University of Edinburgh, Edinburgh EH9 3JZ, United Kingdom

M. Andreotti, V. Azzolini, D. Bettoni, C. Bozzi, R. Calabrese, G. Cibinetto, E. Luppi, M. Negrini, and L. Piemontese
Università di Ferrara, Dipartimento di Fisica and INFN, I-44100 Ferrara, Italy

F. Anulli, R. Baldini-Ferroli, A. Calcaterra, R. de Sangro,

G. Finocchiaro, P. Patteri, I. M. Peruzzi, M. Piccolo, and A. Zallo
Laboratori Nazionali di Frascati dell’INFN, I-00044 Frascati, Italy

A. Buzzo, R. Capra, R. Contri, M. Lo Vetere, M. Macri, M. R. Monge,

S. Passaggio, C. Patrignani, E. Robutti, A. Santroni, and S. Tosi
Università di Genova, Dipartimento di Fisica and INFN, I-16146 Genova, Italy

S. Bailey, G. Brandenburg, K. S. Chaisanguanthum, M. Morii, and E. Won

Harvard University, Cambridge, Massachusetts 02138, USA

R. S. Dubitzky, U. Langenegger, J. Marks, S. Schenk, and U. Uwer

Universität Heidelberg, Physikalisches Institut, Philosophenweg 12, D-69120 Heidelberg, Germany

W. Bhimji, D. A. Bowerman, P. D. Dauncey, U. Egede, R. L. Flack,

J. R. Gaillard, G. W. Morton, J. A. Nash, M. B. Nikolich, and G. P. Taylor
Imperial College London, London, SW7 2AZ, United Kingdom

M. J. Charles, G. J. Grenier, U. Mallik, and A. K. Mohapatra

 3

University of Iowa, Iowa City, Iowa 52242, USA

J. Cochran, H. B. Crawley, V. Eyges, W. T. Meyer, S. Prell, E. I. Rosenberg, A. E. Rubin, and J. Yi

Iowa State University, Ames, Iowa 50011-3160, USA

N. Arnaud, M. Davier, X. Giroux, G. Grosdidier, A. Höcker, F. Le Diberder, V. Lepeltier, A. M. Lutz, A. Oyanguren,

T. C. Petersen, M. Pierini, S. Plaszczynski, S. Rodier, P. Roudeau, M. H. Schune, A. Stocchi, and G. Wormser
Laboratoire de l’Accélérateur Linéaire, F-91898 Orsay, France

C. H. Cheng, D. J. Lange, M. C. Simani, and D. M. Wright

Lawrence Livermore National Laboratory, Livermore, California 94550, USA

A. J. Bevan, C. A. Chavez, J. P. Coleman, I. J. Forster, J. R. Fry, E. Gabathuler,

R. Gamet, K. A. George, D. E. Hutchcroft, R. J. Parry, D. J. Payne, and C. Touramanis
University of Liverpool, Liverpool L69 72E, United Kingdom

C. M. Cormack and F. Di Lodovico

Queen Mary, University of London, E1 4NS, United Kingdom

C. L. Brown, G. Cowan, H. U. Flaecher, M. G. Green, P. S. Jackson, T. R. McMahon, S. Ricciardi, and F. Salvatore

University of London, Royal Holloway and Bedford New College, Egham, Surrey TW20 0EX, United Kingdom

D. Brown and C. L. Davis

University of Louisville, Louisville, Kentucky 40292, USA

J. Allison, N. R. Barlow, R. J. Barlow, M. C. Hodgkinson, G. D. Lafferty, M. T. Naisbit, and J. C. Williams

University of Manchester, Manchester M13 9PL, United Kingdom

C. Chen, A. Farbin, W. D. Hulsbergen, A. Jawahery, D. Kovalskyi, C. K. Lae, V. Lillard, and D. A. Roberts

University of Maryland, College Park, Maryland 20742, USA

G. Blaylock, C. Dallapiccola, S. S. Hertzbach, R. Kofler, V. B. Koptchev,

X. Li, T. B. Moore, S. Saremi, H. Staengle, and S. Willocq
University of Massachusetts, Amherst, Massachusetts 01003, USA

R. Cowan, K. Koeneke, G. Sciolla, S. J. Sekula, F. Taylor, and R. K. Yamamoto

Massachusetts Institute of Technology, Laboratory for Nuclear Science, Cambridge, Massachusetts 02139, USA

H. Kim, P. M. Patel, and S. H. Robertson

McGill University, Montréal, Quebec, Canada H3A 2T8

A. Lazzaro, V. Lombardo, and F. Palombo

Università di Milano, Dipartimento di Fisica and INFN, I-20133 Milano, Italy

J. M. Bauer, L. Cremaldi, V. Eschenburg, R. Godang, R. Kroeger,

J. Reidy, D. A. Sanders, D. J. Summers, and H. W. Zhao
University of Mississippi, University, Mississippi 38677, USA

S. Brunet, D. Côté, P. Taras, and B. Viaud

Université de Montréal, Laboratoire René J. A. Lévesque, Montréal, Quebec, Canada H3C 3J7

H. Nicholson
Mount Holyoke College, South Hadley, Massachusetts 01075, USA

N. Cavallo,∗ G. De Nardo, F. Fabozzi,∗ C. Gatto, L. Lista, D. Monorchio, P. Paolucci, D. Piccolo, and C. Sciacca
Università di Napoli Federico II, Dipartimento di Scienze Fisiche and INFN, I-80126, Napoli, Italy

M. Baak, H. Bulten, G. Raven, H. L. Snoek, and L. Wilden

 4

NIKHEF, National Institute for Nuclear Physics and High Energy Physics, NL-1009 DB Amsterdam, The Netherlands

C. P. Jessop and J. M. LoSecco

University of Notre Dame, Notre Dame, Indiana 46556, USA

T. Allmendinger, G. Benelli, K. K. Gan, K. Honscheid, D. Hufnagel, P. D. Jackson,

H. Kagan, R. Kass, T. Pulliam, A. M. Rahimi, R. Ter-Antonyan, and Q. K. Wong
Ohio State University, Columbus, Ohio 43210, USA

J. Brau, R. Frey, O. Igonkina, M. Lu, C. T. Potter, N. B. Sinev, D. Strom, and E. Torrence

University of Oregon, Eugene, Oregon 97403, USA

F. Colecchia, A. Dorigo, F. Galeazzi, M. Margoni, M. Morandin,

M. Posocco, M. Rotondo, F. Simonetto, R. Stroili, and C. Voci
Università di Padova, Dipartimento di Fisica and INFN, I-35131 Padova, Italy

M. Benayoun, H. Briand, J. Chauveau, P. David, L. Del Buono, Ch. de la Vaissière,

O. Hamon, M. J. J. John, Ph. Leruste, J. Malclès, J. Ocariz, L. Roos, and G. Therin
Universités Paris VI et VII, Laboratoire de Physique Nucléaire et de Hautes Energies, F-75252 Paris, France

P. K. Behera, L. Gladney, Q. H. Guo, and J. Panetta

University of Pennsylvania, Philadelphia, Pennsylvania 19104, USA

M. Biasini, R. Covarelli, S. Pacetti, and M. Pioppi

Università di Perugia, Dipartimento di Fisica and INFN, I-06100 Perugia, Italy

C. Angelini, G. Batignani, S. Bettarini, F. Bucci, G. Calderini, M. Carpinelli, F. Forti, M. A. Giorgi,

A. Lusiani, G. Marchiori, M. Morganti, N. Neri, E. Paoloni, M. Rama, G. Rizzo, G. Simi, and J. Walsh
Università di Pisa, Dipartimento di Fisica, Scuola Normale Superiore and INFN, I-56127 Pisa, Italy

M. Haire, D. Judd, K. Paick, and D. E. Wagoner

Prairie View A&M University, Prairie View, Texas 77446, USA

J. Biesiada, N. Danielson, P. Elmer, Y. P. Lau, C. Lu, J. Olsen, A. J. S. Smith, and A. V. Telnov

Princeton University, Princeton, New Jersey 08544, USA

F. Bellini, G. Cavoto, A. D’Orazio, E. Di Marco, R. Faccini, F. Ferrarotto, F. Ferroni, M. Gaspero,

L. Li Gioi, M. A. Mazzoni, S. Morganti, G. Piredda, F. Polci, F. Safai Tehrani, and C. Voena
Università di Roma La Sapienza, Dipartimento di Fisica and INFN, I-00185 Roma, Italy

S. Christ, H. Schröder, G. Wagner, and R. Waldi

Universität Rostock, D-18051 Rostock, Germany

T. Adye, N. De Groot, B. Franek, G. P. Gopal, E. O. Olaiya, and F. F. Wilson

Rutherford Appleton Laboratory, Chilton, Didcot, Oxon, OX11 0QX, United Kingdom

R. Aleksan, S. Emery, A. Gaidot, S. F. Ganzhur, P.-F. Giraud, G. Graziani, G. Hamel de Monchenault,

W. Kozanecki, M. Legendre, G. W. London, B. Mayer, G. Vasseur, Ch. Yèche, and M. Zito
DSM/Dapnia, CEA/Saclay, F-91191 Gif-sur-Yvette, France

M. V. Purohit, A. W. Weidemann, J. R. Wilson, and F. X. Yumiceva

University of South Carolina, Columbia, South Carolina 29208, USA

T. Abe, M. T. Allen, D. Aston, R. Bartoldus, N. Berger, A. M. Boyarski, O. L. Buchmueller, R. Claus,

M. R. Convery, M. Cristinziani, J. C. Dingfelder, D. Dong, J. Dorfan, D. Dujmic, W. Dunwoodie, S. Fan,
R. C. Field, T. Glanzman, S. J. Gowdy, T. Hadig, V. Halyo, C. Hast, T. Hryn’ova, W. R. Innes, M. H. Kelsey,
P. Kim, M. L. Kocian, D. W. G. S. Leith, J. Libby, S. Luitz, V. Luth, H. L. Lynch, H. Marsiske, R. Messner,
 5

D. R. Muller, C. P. O’Grady, V. E. Ozcan, A. Perazzo, M. Perl, B. N. Ratcliff, A. Roodman, A. A. Salnikov,

R. H. Schindler, J. Schwiening, A. Snyder, J. Stelzer, D. Su, M. K. Sullivan, K. Suzuki, J. M. Thompson,
J. Va’vra, M. Weaver, W. J. Wisniewski, M. Wittgen, D. H. Wright, A. K. Yarritu, K. Yi, and C. C. Young
Stanford Linear Accelerator Center, Stanford, California 94309, USA

J. Strube
University of Oregon, Eugene, Oregon 97403, USA and
Stanford Linear Accelerator Center, Stanford, California 94309, USA

P. R. Burchat, A. J. Edwards, S. A. Majewski, B. A. Petersen, and C. Roat

Stanford University, Stanford, California 94305-4060, USA

M. Ahmed, S. Ahmed, M. S. Alam, J. A. Ernst, M. A. Saeed, M. Saleem, F. R. Wappler, and S. B. Zain

State University of New York, Albany, New York 12222, USA

W. Bugg, M. Krishnamurthy, and S. M. Spanier

University of Tennessee, Knoxville, Tennessee 37996, USA

R. Eckmann, J. L. Ritchie, A. Satpathy, and R. F. Schwitters

University of Texas at Austin, Austin, Texas 78712, USA

J. M. Izen, I. Kitayama, X. C. Lou, and S. Ye

University of Texas at Dallas, Richardson, Texas 75083, USA

F. Bianchi, M. Bona, F. Gallo, and D. Gamba

Università di Torino, Dipartimento di Fisica Sperimentale and INFN, I-10125 Torino, Italy

M. Bomben, L. Bosisio, C. Cartaro, F. Cossutti, G. Della Ricca, S. Dittongo,

S. Grancagnolo, L. Lanceri, P. Poropat,† L. Vitale, and G. Vuagnin
Università di Trieste, Dipartimento di Fisica and INFN, I-34127 Trieste, Italy

F. Martinez-Vidal
IFIC, Universitat de Valencia-CSIC, E-46071 Valencia, Spain

R. S. Panvini†
Vanderbilt University, Nashville, Tennessee 37235, USA

Sw. Banerjee, B. Bhuyan, C. M. Brown, D. Fortin, K. Hamano, R. Kowalewski, J. M. Roney, and R. J. Sobie
University of Victoria, Victoria, British Columbia, Canada V8W 3P6

J. J. Back, P. F. Harrison, T. E. Latham, and G. B. Mohanty

Department of Physics, University of Warwick, Coventry CV4 7AL, United Kingdom

H. R. Band, X. Chen, B. Cheng, S. Dasu, M. Datta, A. M. Eichenbaum, K. T. Flood,

M. Graham, J. J. Hollar, J. R. Johnson, P. E. Kutter, H. Li, R. Liu, B. Mellado, A. Mihalyi,
Y. Pan, R. Prepost, P. Tan, J. H. von Wimmersperg-Toeller, J. Wu, S. L. Wu, and Z. Yu
University of Wisconsin, Madison, Wisconsin 53706, USA

M. G. Greene and H. Neal

Yale University, New Haven, Connecticut 06511, USA

We search for B − → D̃0 K − and B − → D̃∗0 K − , with D̃∗0 → D̃0 π 0 or D̃∗0 → D̃0 γ, and D̃0 → K + π −
(and charge conjugates). These final states can be reached through the b → c transition B − →
D(∗)0 K − followed by the doubly CKM-suppressed D0 → K + π − , or the b → u transition B − →
D(∗)0 K − followed by the CKM-favored D0 → K + π − . The interference of these two amplitudes is
sensitive to the angle γ of the unitarity triangle. Our results are based on 232 million Υ (4S) → BB
decays collected with the BABAR detector at SLAC. We find no significant evidence for these decays.
We set a limit rB ≡ |A(B − → D0 K − )/A(B − → D0 K − )| < 0.23 at 90% C.L. using the most
conservative assumptions on the values of the CKM angle γ and the strong phases in the B and D
 6

decay amplitudes. In the case of the D∗ we set a 90% C.L. limit rB

∗2
≡ |A(B − → D∗0 K − )/A(B − →
∗0 − 2 2
D K )| < (0.16) which is independent of assumptions on γ and strong phases.

PACS numbers: 13.25.Hw, 14.40.Nd

I. INTRODUCTION In the decays of interest, the sign of the bachelor kaon

is opposite to that of the kaon from D̃0 decay. It is
Following the discovery of CP violation in B-meson de- convenient to define ratios of rates between these de-
cays and the measurement of the angle β of the unitar- cays and the similar decays where the two kaons have
ity triangle [1] associated with the Cabibbo-Kobayashi- the same sign. The decays with same-sign kaons have
Maskawa (CKM) quark mixing matrix, focus has turned much higher rate and proceed almost exclusively through
towards the measurements of the other angles α and the CKM-favored and color-favored B transition, fol-
γ. The angle γ is arg(−Vub ∗
Vud /Vcb∗ Vcd ), where Vij are lowed by the CKM-favored D-decay, e.g., B − → D0 K − ,
CKM matrix elements. In the Wolfenstein convention [2], D0 → K − π + . The advantage of taking ratios is that
∗ most theoretical and experimental uncertainties cancel.
γ = arg(Vub ).
Several proposed methods for measuring γ exploit the Thus, ignoring possible small effects due to D mixing and
interference between B − → D(∗)0 K (∗)− and B − → taking into account the effective phase difference of π be-
D(∗)0 K (∗)− (Fig. 1) that occurs when the D(∗)0 and the tween the D∗ decays in Dγ and Dπ 0 [5], we define the
D(∗)0 decay to common final states, as first suggested in charge-specific ratios for D and D∗ as:
Ref. [3].
Γ([K ∓ π ± ]D K ± )
ū R±
Kπ ≡
2
= rB 2
+ rD + 2rB rD cos(±γ + δ),
K (∗)− u Γ([K ± π ∓ ]D K ± )
b D̄(∗)0
W− s (1)
b c B− W− c̄
B− D(∗)0 ū s Γ([K ∓ π ± ]Dπ0 K ± )
R∗±
Kπ,Dπ 0 ≡
ū ū K (∗)− Γ([K ± π ∓ ]Dπ0 K ± )
∗2 2 ∗
= rB + rD + 2rB rD cos(±γ + δ ∗ ), (2)
FIG. 1: Feynman diagrams for B − → D(∗)0 K (∗)− and
D(∗)0 K (∗)− . The latter is CKM- and color-suppressed with and
respect to the former. The CKM-suppression factor is
∗
|Vub Vcs ∗
/Vcb Vus | ≈ 0.4. The naive color-suppression factor is Γ([K ∓ π ± ]Dγ K ± )
1
. R∗±
Kπ,Dγ ≡
3 Γ([K ± π ∓ ]Dγ K ± )
∗2 2 ∗
= rB + rD − 2rB rD cos(±γ + δ ∗ ), (3)

As proposed in Ref. [4], we search for B − → D̃0 K − where

and B − → D̃∗0 K − , D̃∗0 → D̃0 π 0 or D̃∗0 → D̃0 γ, fol-


lowed by D̃0 → K + π − , as well as the charge conjugate
A(B − → D0 K − )



rB ≡

, (4)
(c.c.) sequences. Here the symbol D̃0 indicates the de- A(B − → D0 K − )

cay of a D0 or D0 into K + π − . In these processes, the
favored B decay (B − → D(∗)0 K − ) followed by the dou-


bly CKM-suppressed D decay (D0 → K + π − ) interferes
A(B − → D∗0 K − )

∗
rB ≡


, (5)
with the suppressed B decay (B − → D(∗)0 K − ) followed A(B − → D∗0 K − )

by the CKM-favored D decay (D 0 → K + π − ). The in-
terference of the b → c (B − → D(∗)0 K − ) and b → u


(B − → D (∗)0 K − ) amplitudes is sensitive to the relative
A(D0 → K + π − )



rD ≡

, (6)
weak phase γ. A(D0 → K − π + )

We use the notation B − → [h+ − −
1 h2 ]D h3 (with each
hi = π or K) for the decay chain B → D̃0 h−
−
3 , D̃
0
→ (∗)
+ − ∗0
h1 h2 . For the closely related modes with a D̃ , we use δ (∗) ≡ δB + δD , (7)
the same notation with the subscript D replaced by Dπ 0 (∗)
or Dγ, depending on whether the D̃∗0 decays to D̃0 π 0 or and δB and δD are strong phase differences between the
D̃0 γ. We also refer to h3 as the bachelor π or K. two B and D decay amplitudes, respectively. The value
of rD has been measured to be rD = 0.060 ± 0.002 [6].
We also define the charge-independent ratio

∗ Alsowith Università della Basilicata, Potenza, Italy Γ([K + π − ]D K − ) + Γ([K − π + ]D K + )

† Deceased RKπ ≡ (8)
Γ([K − π + ]D K − ) + Γ([K + π − ]D K + )
 7

and the equivalent ratios for the D∗ modes, It is important to note the different signs of the third
terms in the expressions for R∗Kπ,Dπ0 and R∗Kπ,Dγ in
Γ([K + π − ]Dπ0 K − ) + Γ([K − π + ]Dπ0 K + ) Equations 12 and 13. This relative minus sign is due
RKπ,Dπ0 ≡ ,
Γ([K − π + ]Dπ0 K − ) + Γ([K + π − ]Dπ0 K + ) to the phase of π between the two D∗ decay modes [5].
(9) ∗
It allows for a measurement of rB with no additional in-
and puts since rB ∗2
= 12 (R∗Dπ0 + R∗Dγ ) − rD
2
, and rD is known
Γ([K + π − ]Dγ K − ) + Γ([K − π + ]Dγ K + ) quite precisely (rD = 0.060 ± 0.002). We will use this
RKπ,Dγ ≡ . (10) equation for rB ∗
and our results for R∗Dπ0 and R∗Dγ to set
Γ([K − π + ]Dγ K − ) + Γ([K + π − ]Dγ K + ) ∗
an upper limit on rB with no assumptions.
Then, On the other hand, RKπ depends on the three un-
knowns rB , γ, and δ, see Equation 11; thus, in order to
R+ −
Kπ + RKπ 2 2 extract a limit on rB from the experimental limit on RKπ
RKπ = = rB +rD +2rB rD cos γ cos δ (11)
2 we must make assumptions about γ and δ. As we will
discuss in Section IV, we have chosen to quote an upper
and, similarly for the D∗ modes,
limit on rB based on the most conservative assumptions
R∗Kπ,Dπ0 = rB
∗2 2
+ rD ∗
+ 2rB rD cos γ cos δ ∗ , (12) on γ and δ.
In this paper we report on an update of our previous
search for B − → D̃0 K − [10], and the first attempt to
R∗Kπ,Dγ = rB
∗2 2
+ rD ∗
− 2rB rD cos γ cos δ ∗ . (13) study B − → D̃∗0 K − . The previous analysis was based
on a sample of B-meson decays a factor of 1.9 smaller
Equations 11, 12, and 13 assume no CP violation in the than used here, and resulted in an upper limit RKπ <
normalization modes [K ∓ π ± ]D K ∓ , [K ∓ π ± ]Dπ0 K ∓ , and 0.026 at the 90% C.L. This in turn was translated into
[K ∓ π ± ]Dγ K ∓ . In the following we use the notation R∗Kπ a limit rB < 0.22, also at the 90% confidence level. A
when there is no need to refer specifically to R∗Kπ,Dπ0 or similar analysis by the Belle collaboration [11] gives limits
R∗Kπ,Dγ . RKπ < 0.044 and rB < 0.27 (90% C.L.).
(∗) ∗
As discussed below, the parameter rB is expected to Information on rB , rB , and γ can also be obtained
be of the same order as rD . Thus, CP violation could from studies of B → D̃0 K ± and B ± → D̃∗0 K ± ,
±
manifest itself as a large difference between the charge-
(∗)+ (∗)−
D̃0 → KS π + π − . An analysis by the Belle collabora-
specific ratios RKπ and RKπ . Measurements of these tion [12] finds quite large values rB = 0.31 ± 0.11 and
∗
six ratios can be used to constrain γ. rB = 0.34 ± 0.14, although the uncertainties are large
(∗)
The value of rB determines, in part, the level of in- enough that these results are not inconsistent with the
terference between the diagrams of Fig. 1. In most tech- limits listed above. On the other hand, a similar analy-
(∗) sis by the BABAR collaboration [13] favors smaller values
niques for measuring γ, high values of rB lead to larger
interference and better sensitivity to γ. Thus, rB and for these amplitude ratios: rB < 0.19 at 90% C.L. and
∗
∗
rB are key quantities whose values have a significant im- rB = 0.16 ± 0.08.
pact on the ability to measure the CKM angle γ at the
B-factories and beyond.
(∗) ∗ ∗
In the Standard Model, rB = |Vub Vcs /Vcb Vus | Fcs ≈ II. THE BABAR DATASET
0.4 Fcs . The color-suppression factor Fcs < 1 accounts
for the additional suppression, beyond that due to CKM
The results presented in this paper are based on 232 ×
factors, of B − → D(∗)0 K − relative to B − → D(∗)0 K − .
106 Υ (4S) → BB decays, corresponding to an integrated
Naively, Fcs = 13 , which is the probability for the color
luminosity of 211 fb−1 . The data were collected between
of the quarks from the virtual W in B − → D (∗)0 K − to 1999 and 2004 with the BABAR detector [14] at the PEP-
match that of the other two quarks; see Fig. 1. Early es- II B Factory at SLAC [15]. In addition, a 16 fb−1 off-
timates [7] of Fcs were based on factorization and the ex- resonance data sample, with center-of-mass (CM) energy
perimental information on a number of b → c transitions 40 MeV below the Υ (4S) resonance, is used to study
available at the time. These estimates gave Fcs ≈ 0.22, backgrounds from continuum events, e+ e− → q q̄ (q =
(∗)
leading to rB ≈ 0.09. However, the recent observations u, d, s, or c).
and measurements [8] of color-suppressed b → c decays
(B → D(∗) h0 ; h0 = π 0 , ρ0 , ω, η, η  ) suggest that color-
suppression is not as effective as anticipated, and there-
(∗) III. ANALYSIS METHOD
fore the value of rB could be of order 0.2 [9].
(∗)
As we will describe below, the measured RKπ are con-
(∗) This work is an extension of our analysis from Ref. [10],
sistent with zero in the current analysis. Since RKπ de-
(∗) which resulted in 90% C.L. limits on RKπ < 0.026 and
pend quadratically on rB , we will use our measurements rB < 0.22, as mentioned in Section I. The main changes
(∗)
to set restrictive upper limits on rB . in the analysis are the following:
 8

TABLE I: Notation used in the text for the decay modes that define the data samples used in the analysis.
Abbreviation Mode Comments
DK B − → D0 K − , D0 → K − π + and c.c. normalization
Dπ B − → D0 π − , D0 → K − π + and c.c. control
DK B − → D̃0 K − , D̃0 → K + π − and c.c. signal
D∗ K B − → D∗0 K − , D∗0 → D0 π 0 /γ, D0 → K − π + and c.c. normalization
D∗ π B − → D∗0 π − , D∗0 → D0 π 0 /γ, D0 → K − π + and c.c. control
D∗ K B − → D̃∗0 K − , D̃∗0 → D̃0 π 0 /γ, D̃0 → K + π − and c.c. signal

• The size of the dataset is increased from 120 to within the resolution of about 2.5 and 20 MeV, re-
232 × 106 Υ (4S) → BB decays. spectively (here mB is the known B mass [6]).

• This analysis also includes the B ± → D̃∗0 K ± 5. We then perform simultaneous unbinned maximum
mode. likelihood fits to the final signal samples (DK and
D ∗ K), the normalization samples (DK and D∗ K),
• The event selection criteria have been made more and the control samples (Dπ and D∗ π) to extract
stringent in order to reduce backgrounds further. RKπ and R∗Kπ , see Section III C. The fits are based
• A few of the selection criteria in the previous anal- on the reconstructed values of mES and ∆E in the
ysis resulted in small differences in the efficiencies various event samples.
of the signal mode B ± → [K ∓ π ± ]D K ± and the 6. Throughout the whole analysis chain, care is taken
normalization mode B ± → [K ± π ∓ ]D K ± . These to treat the signal, normalization, and control sam-
selection criteria have now been removed. ples in a consistent manner.
The analysis makes use of several samples from different
decay modes. Throughout the following discussion we
will refer to these modes using abbreviations that are
A. Basic Selection Criteria
summarized in Table I.
The event selection is developed from studies of simu-
lated BB and continuum events, and off-resonance data. Charged kaon and pion candidates in the decay modes
A large on-resonance control sample of Dπ and D∗ π of interest must satisfy K or π identification criteria [16]
events is used to validate several aspects of the simu- that are typically 85% efficient, depending on momentum
lation and analysis procedure. and polar angle. The misidentification rates are at the
The analysis strategy is the following: few percent level.
The invariant mass of the Kπ pair must be within 18.8
1. The goal is to measure or set limits on the charge- MeV (2.5σ) of the mean reconstructed D0 mass (1863.3
independent ratios RKπ and R∗Kπ . MeV). For modes with D̃∗0 → D̃0 π 0 and D̃∗0 → D̃0 γ the
2. The first step consists of the application of a set of mass difference ∆M between the D̃∗0 and the D̃0 must
basic selection criteria to select possible candidate be within 3.5 MeV (3.5σ) and 13 MeV (2σ), respectively,
events, see Section III A. of the expectation for D̃∗0 decays (142.2 MeV).
A major background arises from DK and D∗ K decays
3. After the basic selection criteria, the backgrounds in which the K and π in the D decay are misidentified
are dominantly from continuum. These are signif- as π and K, respectively. When this happens, the decay
icantly reduced with a neural network designed to could be reconstructed as a DK or D∗ K signal event. To
distinguish between BB and continuum events, see eliminate this background, we recompute the invariant
Section III B. mass (Mswitch ) of the h+ h− pair in D̃0 → h+ h− switching
the particle identification assumptions (π vs. K) on the
4. After the neural network requirement, events h+ and the h− . We veto any candidates with Mswitch
are characterized by two kinematical variables within 18 MeV of the known D mass [6]; this requirement
that are customarily used when reconstructing is 90% efficient on signal decays. In the case of D̃0 K, we
B-meson decays at the Υ (4S). These vari- also veto any candidate for which the D̃0 is consistent
ables
 are the energy-substituted mass, mES ≡
with D∗+ → D0 π + or D∗0 → D0 π 0 decay.
( 2s + p
0 · p
B )2 /E02 − p2B and energy difference
√
∆E ≡ EB ∗
− 12 s, where E and p are energy and
momentum, the asterisk denotes the CM frame, the B. Neural Network
subscripts 0 and B refer to the Υ (4S) and B can-
didate, respectively, and s is the square of the CM After these initial selection criteria, backgrounds
energy. For signal events mES = mB and ∆E = 0 are overwhelmingly from continuum events, especially
 9

200 *0
1000 0
(a) D K signal
500 *0
(b) D K(π0 ) signal (c) D K(γ ) signal
800 400 150

600 300 100

400 200
50
200 100
0
0 0
-1 -0.5 0 0.5 1 -1 -0.5 0 0.5 1 -1 -0.5 0 0.5 1
NN output NN’ output NN’ output

0 103 *0
3
(d) D K background *0
(e) D K(π0 ) background (f) D K(γ ) background
10 102
102
2
10
10
10
10

1 1
1
-1 -0.5 0 0.5 1 -1 -0.5 0 0.5 1 -1 -0.5 0 0.5 1
NN output NN’ output NN’ output

FIG. 2: Distributions of the continuum suppression neural network (N N and N N  ) outputs for the three modes. Figs. (a-c)
show the expected distribution from signal events. The dashed line histogram shows the distribution of simulated signal events
and the histogram with error bars shows the distribution of D(∗)0 π control sample events with background subtracted using the
mES sideband (5.20 GeV < mES < 5.27 GeV). Figs. (d-f) show the expected distribution for continuum background events.
The dashed line histogram shows the distribution of simulated continuum events and the histogram with errors shows the
distribution of off-resonance events. The mES and ∆E requirements on the off-resonance and continuum Monte Carlo events
have been kept loose to increase the statistics. Each Monte Carlo histogram is normalized to the area of the corresponding
data histogram.

e+ e− → cc̄, with c̄ → D0 X, D 0 → K + π − and c → DX, 1. A Fisher discriminant

 [17] constructed
 from the
D → K −Y . quantities L0 = i pi and L2 = i pi cos2 θi calcu-
The continuum background is reduced using a neural lated in the CM frame. Here, pi is the momentum
and θi the angle with respect to the thrust axis of
network technique. The neural network algorithms used
for the modes with and without a D∗ are slightly differ- the B candidate of tracks and clusters not used to
ent. First, for all modes we use a common neural network reconstruct the B meson.
(N N ) based on nine quantities, listed below, that distin- 2. | cos θT |, where θT is the angle in the CM frame
guish between continuum and BB events. Then, for the between the thrust axes of the B candidate and
modes with a D∗ we also take advantage of the fact that the detected remainder of the event. The distribu-
the signal is distributed as cos2 θD∗ for D∗ → Dπ or tion of | cos θT | is approximately flat for signal and
sin2 θD∗ for D∗ → Dγ, while the background is roughly strongly peaked at one for continuum background.
independent of cos θD∗ . Here θD∗ is the decay angle of ∗ ∗
the D∗ , i.e., the angle between the direction of the D 3. cos θB , where θB is the polar angle of the B candi-
and the line of flight of the D∗ relative to the parent B, date in the CM frame. In this variable, the signal
evaluated in the D∗ rest frame. Thus, we construct a follows a sin2 θB
∗
distribution, while the background
second neural network, N N  , which takes as inputs the is approximately uniform.
output of N N and the value of cos θD∗ . We then use as K
4. cos θD K
where θD is the decay angle in D̃0 → Kπ.
a selection requirement the output of N N in the D̃0 K
analysis and the output of N N  in the D̃∗0 K analysis.
(∗) (∗)
D
5. cos θB , where θBD
is the decay angle in B →
0 ∗0
The nine variables used in defining N N are the follow- D̃ K or B → D̃ K. In signal events the distri-
K D(∗)
ing: butions of cos θD and cos θB are uniform. On
 10

the other hand, the corresponding distributions in Our final event selection requirement is N N > 0.5
combinatorial background events tend to show ac- for DK and N N  > 0.5 for D∗ K. In addition, to re-
cumulations of events near the extreme values. duce the remaining BB backgrounds, we also require
K
cos θD > −0.75. These final requirements are about 40%
6. The charge difference ∆Q between the sum of the efficient on simulated signal events, and reject 98.5% of
charges of tracks in the D̃0 or D̃∗0 hemisphere and the continuum background.
the sum of the charges of the tracks in the oppo- The overall reconstruction efficiencies, estimated from
site hemisphere excluding those tracks used in the Monte Carlo simulation, are about 14% for DK, 8% for
reconstructed B. The partitioning of the event in D∗ K with D ∗ → Dπ 0 and 7% for D∗ K with D∗ → Dγ.
the two hemispheres is done in the CM frame. For Note that a precise knowledge of the efficiencies is not
signal, ∆Q = 0, whereas for the cc̄ background needed in the analysis. We apply the identical require-
∆Q ≈ 73 × QB , where QB is the charge of the ments to the normalization modes DK and D∗ K. Then,
B candidate. The value of ∆Q in cc̄ events is a in the extraction of RKπ and R∗Kπ , the efficiencies of the
consequence of the correlation between the charge overall selection cancel in the ratio.
of the c (or c̄) in a given hemisphere and the sum
of the charges of all particles in that hemisphere.
Since the ∆Q RMS is 2.4, this variable provides C.
(∗)
Fitting for event yields and RK π
approximately a 1σ separation between signal and
cc̄ background.
The ratios RKπ and R∗Kπ are extracted from the ra-
7. QB · QK , where QK is the sum of the charges of tios of the event yields in the mES distribution for the
all kaons not in the reconstructed B, and QB , as signal modes (DK and D ∗ K) and the normalization
defined above, is the charge of the reconstructed B modes (DK and D∗ K), taking into account potential
candidate. In many signal events, there is a charged differences in efficiencies and backgrounds. All events
kaon among the decay products of the other B in must satisfy the requirements discussed above and have
the event. The charge of this kaon tends to be a ∆E value consistent with zero within the resolution
highly correlated with the charge of the B. Thus, (−52 MeV < ∆E < 44 MeV). Here we discuss the proce-
signal events tend to have QB · QK ≤ −1. On dure to extract RKπ ; the values of R∗Kπ,Dπ0 and R∗Kπ,Dγ
the other hand, most continuum events have no are obtained in the same way.
kaons outside of the reconstructed B, and therefore The mES distributions for DK (signal mode) and DK
QK = 0. (normalization mode) are fitted simultaneously. The fit
parameter RKπ is given by RKπ ≡ c · NDK /NDK , where
8. The distance of closest approach between the bach- NDK and NDK are the fitted yields of DK and DK
elor track and the trajectory of the D̃0 . This is events, and c is a correction factor, determined from
consistent with zero for signal events, but can be Monte Carlo, for the ratio of efficiencies between the
larger in cc̄ events. two modes. We find that this factor c is consistent with
9. A quantity MK defined to be zero if there are no unity within the statistical accuracy of the simulation,
leptons (e or µ) in the event, one if there is a lepton c = 0.98±0.04 (these correction factors are c = 0.97±0.05
in the event and the invariant mass (mK ) of this and c = 0.99 ± 0.05 for D∗ → Dπ 0 and D∗ → Dγ, re-
lepton and the bachelor kaon is less than the mass spectively).
of the D-meson (mD ), and two if mK > mD . This The mES distributions are modeled as the sum of a
quantity differentiates between continuum and sig- threshold combinatorial background function [18] and a
nal events because the probability of finding a lep- Gaussian lineshape centered at mB . The parameters of
ton in a continuum event is smaller than in a BB the background function for the signal mode are con-
event. Furthermore, a large fraction of leptons in strained by a simultaneous fit of the mES distribution
cc̄ events are from D → K ν, where K is recon- for events in the sideband of ∆E (−120 MeV < ∆E <
structed as the bachelor kaon. For these events 200 MeV, excluding the ∆E signal region defined above).
mK < mD , while in signal events the expected The parameters of the Gaussian for the signal and nor-
distribution of mK extends to larger values. malization modes are taken to be identical. The num-
ber of events in the Gaussian is Nsig + Npeak , where
The neural networks (N N and N N  ) are trained with Nsig = NDK or NDK and Npeak is the number of back-
simulated continuum and signal events. The distribu- ground events expected to be distributed in the same way
tions of the N N and N N  outputs for the control sam- as DK or DK in mES (“peaking backgrounds”).
ples (Dπ, D∗ π, and off resonance data) are compared There are two classes of peaking background events:
with expectations from the Monte Carlo simulation in
Fig. 2. The agreement is satisfactory. We have also ex- 1. Charmless B decays, e.g., B − → K + K − π + . These
amined the distributions of all variables used in N N and are indistinguishable from the DK signal if the
N N  , and found good agreement between the simulation K − π + pair happens to be consistent with the D-
and the data control samples. mass.
 11

TABLE II: Summary of fit results.

Mode DK D∗ K, D∗ → Dπ 0 D∗ K, D∗ → Dγ
Ratio of rates, RKπ or R∗Kπ , ×10−3 RKπ = 13+11
−9 R∗Kπ = −2+10
−6 R∗Kπ = 11+18
−13
Number of signal events 5+4
−3 −0.2+1.3
−0.7 1.2+2.1
−1.4
Number of normalization events 368 ± 26 150 ± 17 108 ± 14
Number of peaking charmless events 0.7+1.4
−0.7 0.0+0.3
−0.0 0.1+0.8
−0.1
Number of peaking D(∗) π events in signal sample 0.48 ± 0.05 0.18 ± 0.03 0.01 ± 0.02

90 40 20
Events/(8 MeV)

Events/(8 MeV)

Events/(8 MeV)
(a) (b) (c)
30 15
60
20 10
30
10 5

0 0 0
-0.1 0 0.1 0.2 -0.1 0 0.1 0.2 -0.1 0 0.1 0.2
∆ E (GeV) ∆ E (GeV) ∆ E (GeV)

FIG. 3: ∆E distributions for normalization events (DK and D∗ K) with mES within 3σ of mB with the fit model overlaid.
(a) DK events. (b) D∗ K events with D∗ → Dπ 0 . (c) D∗ K events with D∗ → Dγ. The dashed (dot-dashed) curves are the
contributions from Dπ or D∗ π (DK or D∗ K) events. The dotted curves are the contributions from other backgrounds, and
the solid line is the total.

2. Events of the type B − → D0 π − , where the bach- DK component in the sample of DK candidates.
elor π − is misidentified as a K − . When the D0
decays into K − π + (K + π − ), these events are indis- • The same sample of Dπ events, but reconstructed
tinguishable in mES from DK (DK), since mES is as DK events, is used to constrain the shape of the
independent of particle identification assumptions. Dπ background in the DK sample.

The amount of peaking charmless B background (1) is The Dπ peaking background is much more impor-
estimated directly from the data by performing a simul- tant in the DK (normalization) channel than in the
taneous fit to events in the sideband of the reconstructed DK (signal) channel. This is because in order to con-
D mass. In this fit the number of charmless background tribute to the signal channel, the D0 has to decay into
events is constrained to be ≥ 0. K + π − , and this mode is doubly CKM-suppressed. For
The ∆E distribution of the Dπ background (2) is the DK (signal) sample, the contribution from the resid-
shifted by about +50 MeV due to the misidentification of ual Dπ peaking background in the mES fit is estimated
2
the bachelor π as a K. Since the ∆E resolution is of or- as NDKπ
= rD π
NDK , where rD = 0.060 ± 0.002 is the ra-
der 20 MeV, the ∆E requirement does not eliminate this tio of the doubly CKM-suppressed to the CKM-favored
background completely. The remaining Dπ background D → Kπ amplitudes and NDK π
was defined above.
after the ∆E requirement is estimated relaxing the ∆E The complete procedure simultaneously fits seven dis-
requirement and performing a fit to the ∆E distribution tributions: the mES distributions of DK and DK, the
of the DK candidate sample, as described below. DK distributions in sidebands of ∆E and m(D0 ), the
We fit the ∆E distribution of DK candidates, with ∆E distribution of DK, and the ∆E distributions of Dπ
mES within 3σ of mB , to the sum of a DK component, reconstructed as Dπ and as DK. All fits are unbinned
a Dπ background component, and a combinatorial back- extended maximum likelihood fits. They are configured
ground component, see Fig. 3. From this fit we can es- in such a way that RKπ and R∗Kπ are explicit fit pa-
timate the number of Dπ background events after the rameters. The advantage of this approach is that all un-
∆E requirement, which we denote as NDK π
. In this fit, certainties, including the uncertainties in the PDFs and
the ∆E shapes of the DK and Dπ components are con- the uncertainties in the background subtractions, are cor-
strained from the data as follows: rectly propagated in the statistical uncertainty reported
by the fit.
• The large Dπ sample, with the bachelor track iden- The fit is performed separately for DK, D ∗ K, D∗ →
tified as a pion, is used to constrain the shape of the Dπ 0 , and D ∗ K, D ∗ → Dγ and is identical for all three
 12

200 100 50

Events/(2.5 MeV)

Events/(2.5 MeV)

Events/(2.5 MeV)
(a) (b) (c)
160 80 40
120 60 30
80 40 20
40 20 10
0 0 0
5.2 5.25 5.3 5.2 5.25 5.3 5.2 5.25 5.3
mES (GeV) mES (GeV) mES (GeV)

FIG. 4: mES distributions for normalization events (DK and D∗ K) with ∆E in the signal region with the fit model overlaid. (a)
DK events. (b) D∗ K events with D∗ → Dπ 0 . (c) D∗ K events with D∗ → Dγ. The dashed curves represent the backgrounds;
these are mostly from Dπ or D∗ π, and also peak at the B-mass. As explained in the text, the size of the Dπ and D∗ π
backgrounds is constrained by the simultaneous fits to the distributions of Fig. 3.

10 5 5
Events/(2.5 MeV)

Events/(2.5 MeV)

Events/(2.5 MeV)
(a) (b) (c)
8 4 4
6 3 3
4 2 2
2 1 1
0 0 0
5.2 5.25 5.3 5.2 5.25 5.3 5.2 5.25 5.3
mES (GeV) mES (GeV) mES (GeV)

FIG. 5: mES distributions for DK and D∗ K events with Kπ mass in a sideband of the reconstructed D mass and with ∆E in
the signal region. These events are used to constrain the size of possible peaking backgrounds from charmless B-meson decays,
i.e., decays without a D in the final state. The fit model is overlaid. (a) DK events. (b) D∗ K events with D∗ → Dπ 0 . (c)
D∗ K events with D∗ → Dγ. Note that the Kπ mass range in the sideband selection is a factor of 2.7 larger than in the signal
selection.

modes, except in the choice of parameterization for some for the D ∗ K mode we find R∗Kπ,Dπ0 = (−2+10 −6 ) × 10
−3

signal and background components in the ∆E fits. (for D∗ → Dπ 0 ) and R∗Kπ,Dγ = (11+18 −13 ) × 10
−3
(for
Systematic uncertainties in the detector efficiency can- ∗
D → Dγ).
cel in the ratio. This cancellation has been verified by Next, we use our measurements to extract information
studies of simulated events, with a statistical precision of on rB and rB ∗
. In the case of decays into D̃∗0 we start
a few percent. The likelihood includes a Gaussian uncer- from equations 12 and 13 to derive
tainty term for this cancellation which is set by the statis-
tical accuracy of the simulation. Other systematic uncer-
∗2
R∗Kπ,Dπ0 + R∗Kπ,Dγ 2
tainties, e.g., the uncertainty in the parameter rD used to rB = − rD . (14)
2
estimate the amount of peaking backgrounds from D(∗) π,
are also included in the formulation of the likelihood. We use the relationship given by equation 14 in conjunc-
The fit procedure has been extensively tested on sets tion with rD = 0.060 ± 0.002 and our results for R∗Kπ,Dπ0
of simulated events. It was found to provide an unbiased and R∗Kπ,Dγ to extract information on rB ∗2
with no as-
estimation of the parameters RKπ and R∗Kπ . sumptions on the values of γ and strong phases.
Since the uncertainties in R∗Kπ,Dπ0 and R∗Kπ,Dγ are
non-Gaussian, care must be taken in propagating them
IV. RESULTS 2
into an uncertainty in rB . We convolve the fit likelihoods
∗ ∗
for RKπ,Dπ0 and RKπ,Dγ to obtain a likelihood function
The results of the fits are displayed in Table II and ∗2
for rB . The convolution is performed by generating un-
Figs. 3, 4, 5, and 6. As is apparent from Fig. 6, we see correlated pseudo-experimental results for R∗Kπ,Dπ0 and
no evidence for the D ∗ K and DK modes. R∗Kπ,Dγ picked according to the respective likelihoods.
For the DK mode we find RKπ = (13+11 −9 ) × 10
−3
; The uncertainty in rD is included by choosing rD values
 13

10 6 5

Events/(2.5 MeV)

Events/(2.5 MeV)

Events/(2.5 MeV)
(a) (b) (c)
8 5 4
4
6 3
3
4 2
2
2 1 1
0 0 0
5.2 5.25 5.3 5.2 5.25 5.3 5.2 5.25 5.3
mES (GeV) mES (GeV) mES (GeV)

FIG. 6: mES distributions for candidate signal events with the fit model overlaid. (a) DK events. (b) D∗ K events with
D∗ → Dπ 0 . (c) D∗ K events with D∗ → Dγ.

from a Gaussian distribution. For each pseudo experi- 1

ment R∗Kπ,Dπ0 , R∗Kπ,Dγ , and rD 2
are combined according
0.8
∗2
to equation 14 to give a value for rB . The resulting distri-



0.6
∗2
bution for rB , which we interpret as a likelihood function 0.4
∗2
for rB , is shown in Figure 7. This convolution relies on 0.2
the fact that the measurements of R∗Kπ,Dπ0 and R∗Kπ,Dγ
0

Ê

-0.04 -0.02 0 0.02 0.04 0.06 0.08 0.1
are uncorrelated (the correlation due to the uncertainty
in rD , which was used to extract the Dπ peaking back-
grounds in the two modes, is negligible).
Our result is rB ∗2
= (4+14
−8 ) × 10
−3
. Based on the like- FIG. 8: Likelihood function (arbitrary units) for RKπ as ex-
∗2 tracted from the fit described in the text. The integral of the
lihood for rB shown in Figure 7 we set an upper limit
rB∗2
< (0.16)2 at the 90% C.L. using a Bayesian method likelihood function for 0 < RKπ < 0.029 is nine-tenths of the
∗2 integral for RKπ > 0. The vertical dashed line is drawn at
with a uniform prior for rB > 0.
RKπ = 0.029.

0.8
0.7
0.6
51◦ < γ < 66◦ suggested by global CKM fits [19]. We use




0.5
0.4 the information displayed in this Figure to set an upper
0.3
0.2
limit on rB . The least restrictive limit on rB is computed
0.1 assuming maximal destructive interference between the
0
-0.03 -0.02 -0.01 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 b → c and b → u amplitudes: γ = 0◦ , δ = 180◦ or
 ¾ γ = 180◦ , δ = 0◦ . The limit is rB < 0.23 at 90% C.L.

∗2
FIG. 7: Likelihood function (arbitrary units) for rB obtained TABLE III: Summary of the results of this analysis.
from the results for R∗Kπ,Dπ0 and R∗Kπ,Dγ as described in the
∗2
text. The integral of the likelihood function for 0 < rB < Measured Value 90% C.L. limit
2 ∗2
(0.16) is nine-tenths of the integral for rB > 0. The vertical RKπ 0.013±0.011
0.009 < 0.029
∗2
dashed line is drawn at rB = (0.16)2 . R∗Kπ,Dπ0 −0.002±0.010
0.006 < 0.023
R∗Kπ,Dγ 0.011±0.018
0.013 < 0.045
rB n.a. < 0.23
∗2
rB 0.004±0.014
0.008 < (0.16)2
In the case of decays into a D/D, there is not enough
information to extract the ratio rB without additional
assumptions. Thus, we first extract an upper limit on
the experimentally measured quantity RKπ . This is done
starting from the likelihood as a function of RKπ (see
V. SUMMARY
Fig. 8) using a Bayesian method with a uniform prior
for RKπ > 0. The limit is RKπ < 0.029 at 90%C.L.
Next, in Fig. 9 we show the dependence of RKπ on rB , In summary, we find no significant evidence for the
together with our limit on RKπ . This dependence is decays B ± → [K ∓ π ± ]D K ± and B ± → [K ∓ π ± ]D∗ K ± .
(∗)
shown allowing a ±1σ variation on rD , for the full range We set upper limits on the ratios RKπ of the rates for
0◦ − 180◦ for γ and δ, as well as with the restriction these modes and the favored modes B ± → [K ± π ∓ ]D K ±
 14

∗
0 0.05 0.1 0.15 0.2 0.25 0.3 Our results favor values of rB and rB somewhat smaller
0.07
than the value of 0.2 which can be estimated from the
22.5 measurements of color-suppressed b → c transitions [9].

Number of signal events

∗
0.06
20 If rB and rB are small, the suppression of the b → u
amplitude will make the determination of γ using meth-
0.05 17.5 ods based on the interference of the diagrams in Fig. 1
difficult.
RKπ

15
0.04
12.5

0.03 10 VI. ACKNOWLEDGMENTS

7.5
0.02 We are grateful for the extraordinary contributions of
5 our PEP-II colleagues in achieving the excellent lumi-
0.01 nosity and machine conditions that have made this work
2.5
possible. The success of this project also relies criti-
0 0 cally on the expertise and dedication of the computing
0 0.05 0.1 0.15 0.2 0.25 0.3
organizations that support BABAR. The collaborating
rB institutions wish to thank SLAC for its support and the
kind hospitality extended to them. This work is sup-
FIG. 9: Expectations for RKπ and the number of signal events ported by the US Department of Energy and National
vs. rB . Dark filled-in area: allowed region for any value of Science Foundation, the Natural Sciences and Engineer-
δ, with a ±1σ variation on rD , and 51◦ < γ < 66◦ . Hatched ing Research Council (Canada), Institute of High Energy
area: additional allowed region with no constraint on γ. Note Physics (China), the Commissariat à l’Energie Atom-
that the uncertainty on rD has a very small effect on the ique and Institut National de Physique Nucléaire et de
size of the allowed regions. The horizontal line represents the Physique des Particules (France), the Bundesministerium
90% C.L. limit RKπ < 0.029. The vertical dashed lines are
für Bildung und Forschung and Deutsche Forschungsge-
drawn at rB = 0.203 and rB = 0.233. They represent the
90% C.L. upper limits on rB with and without the constraint meinschaft (Germany), the Istituto Nazionale di Fisica
on γ. The light filled-in area represents the 68% C.L. region Nucleare (Italy), the Foundation for Fundamental Re-
corresponding to RKπ = 0.013±0.011
0.009 .
search on Matter (The Netherlands), the Research Coun-
cil of Norway, the Ministry of Science and Technology of
the Russian Federation, and the Particle Physics and As-
tronomy Research Council (United Kingdom). Individu-
and B ± → [K ± π ∓ ]D∗ K ± . We also use our data to set als have received support from CONACyT (Mexico), the
upper limits on the ratios of b → u and b → c amplitudes A. P. Sloan Foundation, the Research Corporation, and
∗
rB and rB . All of our results are summarized in Table III. the Alexander von Humboldt Foundation.

[1] BABAR Collaboration, B. Aubert et al., Phys. Rev. Lett. [9] M. Gronau, Phys. Lett. B557, 198 (2003).
89, 201802 (2002); Belle Collaboration, K. Abe et al., [10] BABAR Collaboration, B. Aubert et al., Phys. Rev. Lett.
Phys. Rev. D66, 071102 (2002). 93, 131804 (2004).
[2] L. Wolfenstein, Phys. Rev. Lett. 51, 1945 (1983). [11] Belle Collaboration, M. Saigo et al., Phys. Rev. Lett. 94,
[3] M. Gronau and D. Wyler, Phys. Lett. B265, 172 (1991); 091601 (2005).
M. Gronau and D. London, Phys. Lett. B253, 483 [12] Belle Collaboration, A. Poluetkov et al., Phys. Rev D70,
(1991). 072003 (2004).
[4] D. Atwood, I. Dunietz, and A. Soni, Phys. Rev. Lett. 78, [13] BABAR Collaboration, B. Aubert et al., hep-ex/0408088,
3257 (1997); Phys. Rev. D63, 036005 (2001). contributed paper to the ICHEP04 conference.
[5] A. Bondar and T. Gershon, Phys. Rev. D70, 091503 [14] BABAR Collaboration, B. Aubert et al., Nucl. Instr. and
(2004). Methods A479, 1 (2002).
[6] Particle Data Group, S. Eidelman et al., Phys. Lett. [15] PEP-II Conceptual Design Report, SLAC-0418 (1993).
B592, 1 (2004). [16] BABAR Collaboration, B. Aubert et al., Phys. Rev. D66,
[7] See, for example, M. Neubert and B. Stech, in Heavy Fla- 032003 (2002).
vors, 2nd Edition, edited by A.J. Buras and M. Lindner, [17] R. A. Fisher, Ann. of Eugenics √7, 179 (1936).
2
World Scientific, Singapore, 1997. [18] The function is f (m√
ES ) ∝ mES 1 − x exp[−ζ(1 − x )],
2

[8] CLEO Collaboration, T.E. Coan et al., Phys. Rev. Lett. where x = 2mES / s and ζ is a fit parameter; AR-
88, 062001 (2001). Belle Collaboration, K. Abe et al., GUS Collaboration, H. Albrecht et al., Z. Phys. C48,
Phys. Rev. Lett. 88, 052002 (2002); A. Satpathy et al., 543 (1990).
Phys. Lett. B553, 159 (2003). BABAR Collaboration, B. [19] J. Charles et al., hep-ph/0406184; updated results can
Aubert et al., Phys. Rev. D69, 032004 (2004). be found at http://ckmfitter.in2p3.fr.