1: %\documentclass[preprint,prb,aps,nobibnotes]{revtex4}%endfloats

2: \documentclass[prb,aps,nobibnotes,twocolumn]{revtex4}%endfloats

3: \usepackage{graphicx}%

4: \usepackage{dcolumn}

5: \usepackage{amsmath}   

6: \usepackage{bm}

7: \usepackage{dcolumn}

8: \usepackage{longtable}

9: \usepackage{pifont,color}

10: %\usepackage[dvips]{color}

11: \voffset=1.5cm

12: %\hoffset=0.8cm

13: 

14: \bibliographystyle{achemsol}

15: 

16: \begin{document}

17: 

18: \title{\centering\Large\bf What are the benefits of bound (protonation) states

19:                            for the electron-transfer kinetics? } 

20: \author{Dmitry V.\ Matyushov}

21: \affiliation{Center for Biological Physics, 

22:   Arizona State University, PO Box 871604, Tempe, AZ 85287-1604}

23: \date{\today}

24: \begin{abstract}

25:   We describe a model of electron transfer reactions affected by local

26:   binding to the donor or acceptor sites of a particle in equilibrium

27:   with the solution. The statistics of fluctuations of the

28:   donor-acceptor energy gap caused by binding/unbinding events are

29:   non-Gaussian, and the resulting free energy surfaces of electron

30:   transfer are non-parabolic. The band-width of the charge-transfer

31:   optical transition is predicted to pass through a maximum as a

32:   function of the concentration of binding particles in the solution.

33:   The model is used to rationalize recent observations of

34:   pH-dependence of electron transfer rates involving changes in the

35:   protonation state of the donor-acceptor complex.

36: \end{abstract}

37: \preprint{Submitted to J.\ Phys.\ Chem.\ B}

38: \maketitle

39: 

40: \section{Introduction}

41: \label{sec:1}

42: Many redox reaction in chemistry and biology involve bound states

43: which are weaker than common chemical bonds, but stronger than

44: intermolecular interactions in bulk molecular liquids. A prominent

45: example of such association is binding of water molecules to solutes

46: via hydrogen bonds. Since the strength of such bonds typically varies

47: between different electronic states of the solute, water association

48: can be recorded by optical

49: solvatochromism.\cite{Reichardt:94,DMjpcb:97,SolvPolarity:04,Agmon:05}

50: Another example is the one of proton equilibria involved in biological

51: electron transfer (ET) reactions responsible for photosynthesis and

52: respiration.\cite{McEvoy:06} In Photosystem II, the primary donor

53: P$_{680}^+$ is oxidized by tyrosine which changes its pK$_a$-value

54: from 10 to $-2$ upon oxidation. As a result, it cannot hold onto its

55: phenolic proton in aqueous solution releasing it in what is argued to

56: be a concerted electron-proton transfer mechanism.\cite{Cukier:04}

57: Another related example is photoacidity when photoexcited

58: intramolecular charge transfer lowers pK$_a$ for the release of a

59: proton to the solvent.\cite{Agmon:05} In view of the wide spread of

60: such reactions, often referred to as proton-coupled

61: ET,\cite{Cukier:96,Soudackov:00,Hammes-Schiffer:01,Cukier:04} in

62: biological systems in general and enzymetic reactions in

63: particular,\cite{Ferguson-Miller:96,Kirby:97} one wonders if their

64: occurrence is just a coincidence caused by ubiquity of acid-base

65: equilibria in proteins, or, alternatively, protonation/deprotonation

66: transitions are used by nature to fine-tune the ET energetics. This

67: question, also used for the paper title, is the subject of this study.

68: 

69: It has been argued that coupling of ET to the change in protonation

70: alters the reaction free energy, often turning uphill reactions along

71: the ET coordinate into downhill reactions along a combined

72: electron-proton reaction path.\cite{Mayer:04} In order to make this

73: argument more qualitative, at least one more parameter, the

74: reorganization energy, needs to be

75: specified.\cite{Marcus:93,Borgis:96} The picture of crossing parabolas

76: used to calculate barriers of charge-transfer

77: reactions\cite{Marcus:65} involves in fact three components: the

78: reaction free energy $\Delta G$ representing the vertical shift of the

79: minima, the curvature $(2\lambda)^{-1}$ given in terms of the reorganization

80: energy $\lambda$, and the horizontal shift of the parabolas' minima

81: responsible for the Stokes shift, $\Delta X=X_{01}-X_{02}$ (Figure

82: \ref{fig:1}).

83: 

84: \begin{figure}

85:   \centering

86:   \includegraphics*[width=6.5cm]{fig1}

87:   \caption{Diagram of the free energy surfaces of ET obtained from

88:     three parameters, the reaction free energy $\Delta G$, the free energy

89:     curvature $(2\lambda)^{-1}$, and the Stokes shift (separation between

90:     the minima $\Delta X$). The reaction coordinate $X$ is the energy

91:     separation between the donor and acceptor electronic states. The

92:     activated state is given by the resonance of the donor and

93:     acceptor levels, $X=0$, while $X_{01}$ and $X_{02}$ indicate the

94:     positions of the minima. }

95: \label{fig:1} 

96: \end{figure} 

97: 

98: The reaction coordinate $X$, used to monitor the progress of a

99: charge-transfer reaction, is the instantaneous energy gap between the

100: donor and acceptor electronic energy levels.\cite{King:90} The

101: vertical axis in Figure \ref{fig:1} refers to \textit{free energies}

102: $G_i(X)$ as functions of \textit{energy} $X$. They cross at zero

103: energy gap $X=0$ allowing electronic tunneling ($i=1,2$ refer to

104: reactants and products, respectively). In fact, the Marcus-Hush

105: description of ET\cite{Marcus:89} imposes one more

106: additional constraint requiring the Stokes shift to be identically

107: equal to $2\lambda$,

108: %

109: \begin{equation}

110:   \label{eq:0} 

111:     \Delta X = 2 \lambda = \beta \langle (\delta X)^2 \rangle 

112: \end{equation}

113: where $\beta = 1/(k_{\text{B}}T)$.  Therefore, only two parameters, $\Delta G$

114: and $\lambda$, are required to determine the activation barrier.  The strong

115: link between the Stokes shift and the reorganization energy is a

116: consequence of the Gaussian statistics of $X$.  Once the statistics

117: become non-Gaussian, the free energy surfaces $G_i(X)$ are

118: non-parabolic and eq \ref{eq:0} does not hold.\cite{DMacc:07} Here, we

119: propose a simple model for a charge-transfer reactions triggering the

120: change in the state of binding of some particle dissolved in the

121: solvent.  In this development, we are less concerned with the details

122: of a particular binding mechanisms, but more interested in addressing

123: a more general question of whether the involvement of bound states can

124: potentially provide a new mechanism of tuning the energetics of ET not

125: incorporated in the picture of crossing parabolas. We find that the

126: statistics of fluctuations of the donor-acceptor energy gap, induced

127: by local binding/unbinding events, are non-Gaussian thus violating the

128: link between the Stokes shift and reorganization energy given by eq

129: \ref{eq:0}. The main consequence of this result is more flexibility,

130: compared to the standard picture, in tuning the energetics of ET.

131: 

132: \section{Model} 

133: \label{sec:2}

134: We will consider a somewhat simplified situation in which the bound

135: state is fully thermalized on the time-scale of ET i.e.\ on the time

136: $\tau_{\text{ET}}=k_{\text{ET}}^{-1}$ required for the system to climb

137: the top of the potential barrier from the equilibrium bottom of the

138: free energy surface ($k_{\text{ET}}$ is the ET rate constant). This

139: assumption implies that populations of bound and unbound states follow

140: Boltzmann statistics along the ET reaction coordinate, a situation

141: similar to the treatment of intramolecular vibrations in Sumi-Marcus

142: model.\cite{Sumi:86} This condition can be achieved when the rate of

143: release of the binding particle B, $k_b$, is much higher that the rate

144: of ET

145: %

146: \begin{equation} 

147: \label{eq:1} 

148: k_{\text{ET}} \ll k_b 

149: \end{equation}

150: 

151: When this requirement is satisfied, one can use statistical mechanics

152: to construct a one-dimensional free-energy surface for the ET reaction

153: in which the exchange between localized and delocalized states of the

154: binding particle produce fluctuations of the donor-acceptor energy

155: gap, just as any other solvent mode interacting with the transferred

156: electron.\cite{Stuchebrukhov:03}

157: 

158: 

159: \begin{figure}

160:   \centering

161:   \includegraphics*[width=7cm]{fig2}

162:   \caption{Diagram of particle B exchanging between the bound and

163:     dissolved states with the rate constant $k_b$. The equilibrium

164:     population of the bound state assumes that the unbinding rate is

165:     much faster than the rate of ET ($k_{\text{ET}}\ll k_b$); $\Delta

166:     \epsilon=\epsilon_2-\epsilon_1$ denotes the change in the binding energy caused by the

167:     electronic transition. }

168:   \label{fig:2}

169: \end{figure}

170: 

171: The release of one particle from the bound localized state requires

172: the Gibbs energy $g_i$ balancing the binding energy $\epsilon_i$ with an

173: entropy gain from moving the particle into the bulk. The binding

174: energy $\epsilon_i$ is generally different in the initial, D--A ($i=1$), and

175: final, D$^+$--A$^-$ ($i=2$), configurations of the donor-acceptor

176: complex (Fig.\ \ref{fig:2}). The difference in free energies needed to

177: release particle B to the solution is equal to the difference in

178: binding energies, $\Delta g = \Delta\epsilon = \epsilon_2-\epsilon_1$. Therefore, this quantity can

179: be estimated from the difference of the equilibrium dissociation

180: constants pK$_a^{(i)}$:

181: %

182: \begin{equation}

183:  \label{eq:2}

184:   \Delta \epsilon  = k_{\text{B}}T  \ln 10 \Delta \mathrm{pK}_a

185: \end{equation}

186: 

187: The requirement of non-adiabaticity of ET, which we implicitly assume

188: here, imposes an additional constraint on the rate of bound particle

189: release.  The change of the donor and acceptor energy levels caused by

190: the motion of the bound particle should not break the Landau-Zener

191: non-adiabaticity condition close to donor-acceptor energy resonance:

192: %

193: \begin{equation}

194:  \label{eq:3}

195:   \frac{2\pi}{\hbar} |V_{DA}|^2/ \dot\epsilon \ll 1

196: \end{equation}

197: Here, $V_{DA}$ is the donor-acceptor electronic coupling. Assuming the

198: rate of binding energy change as $\dot \epsilon \simeq k_b \Delta \epsilon $ and $V_{DA}$ of

199: the order of 1 cm$^{-1}$, eq \ref{eq:3} requires the rate of release

200: of B to be faster than 1 ns$^{-1}$.  This threshold rate is comparable

201: to the rate of proton release from bound protonation states of

202: proteins,\cite{Gutman:97} while the time of water exchange between the

203: protein surface and the bulk is faster, in the range of tens of

204: picoseconds.\cite{Halle:04}

205: 

206: Once particle B is released from its bound state, it becomes a part of

207: the bulk, which is the aqueous solution or a solvent mixture for water

208: binding or the ionic atmosphere for ionic (proton) association. The

209: interaction of the bulk with electronic states of the donor and

210: acceptor is relatively well understood. For water binding,

211: fluctuations of the dipolar polarization produce the thermal noise.

212: For electrolytes, the Debye-H{\"u}ckel electric field both shifts the

213: donor-acceptor energy gap and creates its fluctuations thus producing

214: a corresponding reorganization energy.\cite{German:92} This ionic

215: reorganization energy $\simeq \kappa^2 e^2 R / \epsilon_s$ scales quadratically with

216: the inverse Debye-H{\"u}ckel length $\kappa$ and linearly with a solute

217: dimension $R$.  Since it is also inversely proportional to the static

218: dielectric constant of the solvent $\epsilon_s$, the effect of fluctuations

219: of the ionic atmosphere in polar solvents is commonly small relative

220: to the reorganization energy of polarization fluctuations. We will

221: therefore assume that once particle B is released from its bound

222: state, it does not interact any more with the transferred electron and

223: becomes a part of the many-particle electrostatic potential.

224: 

225: This approximation allows us to apply the tight-binding approximation

226: and to use the following expression for the Hamiltonian of the

227: donor-acceptor system in a polar solvent

228: %

229: \begin{equation}

230:   \label{eq:4}

231:   \begin{split}

232:   H_i & = E_{0i} - \left(\mathbf{E}_i^{u} |u\rangle\langle u| + \mathbf{E}_i^{b} |b\rangle\langle b|\right)*\mathbf{P} \\

233:     & - \epsilon_i |b\rangle \langle b|     + \frac{1}{2} \mathbf{P}*\chi^{-1}*\mathbf{P} 

234:   \end{split}

235: \end{equation}

236: Here, $E_{0i}$ incorporates the vacuum energy and the free energy of

237: solvation by the electronic degrees of freedom of the solvent and

238: $\mathbf{E}_i^{u,b}$ denotes the vacuum electric field of the

239: donor-acceptor complex in two ET states ($i=1,2$) and two binding

240: states (u,b).  Also, $|b\rangle \langle b|$ and $|u\rangle \langle u|$ in eq \ref{eq:4}

241: describe the population operators for the bound (b) and unbound (u)

242: states of particle B with the binding energy $\epsilon_i$.  One can expect

243: only a small change in the electric field of the donor-acceptor

244: complex for weak binding of a neutral molecule, $\mathbf{E}_i^b \simeq

245: \mathbf{E}_i^u$. However, the dependence of the electric field on the

246: binding state needs to be incorporated for equilibria of charged

247: particles.

248: 

249: In application to protonation/deprotonation equilibria, the

250: Hamiltonian considered here (eq \ref{eq:4}) is different from the ones

251: used by Cukier\cite{Cukier:96} and Soudackov and

252: Hammes-Schiffer.\cite{Soudackov:00} In their case, the proton was

253: considered to move between two localized states within the

254: donor-acceptor complex, while in the problem considered here the

255: proton is released to the bulk and loses any connection to electronic

256: transitions within the donor-acceptor pair except for the influence of

257: the Debye-H{\"u}ckel field it becomes a part of. The different physics

258: of the problem considered here demands the different Hamiltonian.

259:  

260: The electric field interacts with the (nuclear) dipolar polarization

261: of the solvent $\mathbf{P}$, and the asterisk in eq \ref{eq:4} implies

262: both the tensor contraction and the volume integration.  The

263: statistics of $\mathbf{P}$ are Gaussian with the response function

264: $\mathbf{\chi}$ such that the solvent reorganization energy is

265: %

266: \begin{equation}

267:   \label{eq:5}

268:   \lambda_{u,b} = \frac{1}{2}\mathbf{\Delta E}_{u,b}*\chi*\mathbf{\Delta E}_{u,b}, \quad 

269:                  \Delta \mathbf{E}_{u,b}= \mathbf{E}_2^{u,b} -\mathbf{E}_1^{u,b} 

270: \end{equation}

271: Although $\lambda_{u,b}$ in eq \ref{eq:5} are formally the reorganization

272: energies of the dipolar polarization field of the solvent, we will

273: attach a more general meaning to them as the reorganization energies

274: of the nuclear degrees of freedom of the bulk. The corresponding

275: formal definition is easy to obtain from eq \ref{eq:4} by replacing

276: the Gaussian polarization noise with a sum of statistically

277: independent Gaussian fields.

278: 

279: The free energies of ET are obtained from the Hamiltonian in eq

280: \ref{eq:4} by projecting all possible nuclear motions on the reaction

281: coordinate $X=\Delta H = H_2 - H_1$:

282: %

283: \begin{equation}

284:   \label{eq:6}

285:   e^{-\beta G_i(X)} = \int \mathcal{D} \mathbf{P} 

286:                              \mathrm{Tr}\left[\delta(X - \Delta H) e^{-\beta H_i}\right]

287: \end{equation}

288: Here, $\mathcal{D} \mathbf{P}$ implies functional integration over the

289: field $\mathbf{P}$ and ``Tr'' refers to the sum over bound, $|b\rangle $, and unbound,

290: $|u\rangle$, states of particle B with statistical weights $f_{b,u}$:

291: %

292: \begin{equation}

293:   \label{eq:7}

294:   \mathrm{Tr}[\dots ] = \sum_{a=b,u} \langle a|f_a \dots |a\rangle .    

295: \end{equation}

296: The ratio of the statistical weights gives the entropy of releasing

297: particle B to the bulk, $k_{\text{B}}\ln(f_u/f_b)$.

298: 

299: 

300: Taking the integral and trace in eq \ref{eq:6} results in the following

301: equation for the free energy surfaces of ET

302: %

303: \begin{equation}

304:   \label{eq:8}

305:   \begin{split}

306:   G_i(X)& = G_i^u + \frac{(X-\Delta E_i^u)^2}{4\lambda_u} \\

307:       & - \beta^{-1}\ln\left(1 + 10^{\mathrm{pK}_a^{(i)}-\mathrm{pB}}f_i(X)\right) 

308:   \end{split}

309: \end{equation}

310: In eq \ref{eq:8}, $G_i^u$ is the free energy of the donor-acceptor

311: complex in the unbound state. The solvation free energy entering

312: $G_i^u$ is thus of electrostatic origin and does not include the free

313: energy of binding particle B.  Correspondingly, the vertical energy

314: gap $\Delta E_i^u$ in the second summand is given as $\lambda_u + \Delta G_u$ for the

315: forward transition $1\to 2$ and as $\lambda_u - \Delta G_u$ for the backward

316: transition $2\to 1$. Binding of particle B is expressed by the term

317: under the logarithm where function $f_i(X)$ is the ratio of Boltzmann

318: factors for activation through bound and unbound states:

319: %

320: \begin{equation}

321:   \label{eq:8-1}

322:   f_i(X) = \exp\left[-\beta \frac{(X+\Delta \epsilon - \Delta E_i^b)^2}{4\lambda_b} + \beta \frac{(X- \Delta E_i^u)^2}{4\lambda_u}   \right]

323: \end{equation}

324: Here, $\Delta E_i^b$ are the vertical ET gaps in the bound state of 

325: particle B which are given as $\lambda_b + \Delta G_b$ ($i=1$) and $\lambda_b - \Delta G_b$

326: ($i=2$).  

327: 

328: 

329: In writing eq \ref{eq:8} we have also represented the Boltzmann factor

330: for the release of particle B in terms of pK$_a^{(i)}$ and pB$= -

331: \log[B]$ as follows

332: %

333: \begin{equation}

334:   \label{eq:9}

335:   e^{-\beta g_i - \beta \Delta G_i }= 10^{\mathrm{pB} - \mathrm{pK}_a^{(i)}}

336: \end{equation}

337: In this equation $g_i=\epsilon_i - k_{\text{B}}T\ln(f_u/f_b)$ is the free

338: energy of releasing the bound particle from the molecular fragment in

339: the donor-acceptor complex (e.g.\ tyrosine) and $\Delta G_i = G_i^u -

340: G_i^b$ is the change in the solvation energy of the entire

341: donor-acceptor complex caused by unbinding event. The equilibrium

342: constants $\mathrm{pK}_a^{(i)}$ thus reflects the equilibrium of the

343: entire complex and can be modified compared to the equilibrium

344: constants of the molecular fragment alone.

345: 

346: 

347: The free energies $G_i(X)$ can be used to calculate the first and

348: second moments of the reaction coordinate $X$, e.g.

349: %

350: \begin{equation}

351:   \label{eq:10}

352:   \langle X\rangle_i = Q_i^{-1} \int X e^{-\beta G_i(X)} dX

353: \end{equation}

354: where

355: %

356: \begin{equation}

357:   \label{eq:11}

358:   Q_i = \int e^{-\beta G_i(X)} dX 

359: \end{equation}

360: One gets for the average

361: \begin{equation}

362:   \label{eq:12}

363:    \langle X\rangle_i = (1 - n_i) \Delta E_i^u + n_i (\Delta E_i^b- \Delta\epsilon)

364: \end{equation}

365: and for the variance

366: %

367: \begin{equation}

368:   \label{eq:13}

369:    \begin{split}

370:    \langle (\delta X)^2 \rangle_i &=  2k_{\text{B}}T\left[\lambda_b n_i + \lambda_u(1-n_i)\right]\\

371:             &  +n_i(1-n_i)\left[\Delta E_i^u - \Delta E_i^b + \Delta \epsilon\right]^2

372:    \end{split}

373: \end{equation}

374: In eqs \ref{eq:12} and \ref{eq:13},

375: %

376: \begin{equation}

377:   \label{eq:14}

378:     n_i = \left[1 + 10^{\mathrm{pB}-\mathrm{pK}_a^{(i)}} \right]^{-1} 

379: \end{equation}

380: is the equilibrium population of the bound site. 

381: 

382: \begin{figure}

383:   \centering

384:   \includegraphics*[width=6.5cm]{fig3}

385:   \caption{Free energy surfaces of ET (eq \ref{eq:14-1}) at different

386:     values of pB indicated in the plot assuming pK$_a^{(1)}=10$ and

387:     pK$_a^{(2)}=-2$ (equilibrium constants of tyrosine). The final

388:     charge-transfer state ($i=2$) is always deprotonated while the

389:     initial charge-transfer state ($i=1$) is protonated at

390:     pB$<$pK$_a^{(1)}$ and is deprotonated at pB$>$pK$_a^{(1)}$. The

391:     Stokes shift for the protonated state 1 is $2\lambda_s - \Delta\epsilon = 2.71$ eV

392:     and is equal to $2\lambda_s=2$ eV for the deprotonated state 1. The

393:     solvent bulk component of the reaction Gibbs energy reaction is $\Delta

394:     G = -0.1$ eV. }

395:   \label{fig:3}

396: \end{figure}

397: 

398: 

399: Equations \ref{eq:6}--\ref{eq:14} provide a general description of ET

400: when binding affects both the statistics of the energy gap

401: fluctuations and the electronic density responsible for the electric

402: field.  In order to reduce the number of independent parameters, we

403: first neglect the second effect assuming that the electric field does

404: not change with binding, i.e.\ $\lambda_u=\lambda_b=\lambda_s$, $\Delta E_i^u=\Delta E_i^b=\Delta E_i$,

405: and $G_i^u=G_i^b=G_i$. This situation most closely corresponds to

406: binding of a neutral molecule (water) while the case of a charged

407: particle (protonation) is postponed to the next section.

408: 

409: The free energy surfaces of ET can then be written as follows:

410: %

411: \begin{equation}

412:   \label{eq:14-1}

413:   \begin{split}

414:   &G_i(X) = G_i + \frac{(X-\Delta E_i+\Delta \epsilon)^2}{4\lambda_s} \\

415:   & - \beta^{-1} \ln\left(10^{\mathrm{pK}_a^{(i)} - \mathrm{pB}} + \exp\left[\frac{\beta\Delta\epsilon}{2\lambda_s}(X-\Delta E_i +\Delta\epsilon/2)\right]\right)

416:   \end{split}

417: \end{equation}

418: where $G_i$ refers to the free energy of the donor acceptor complex

419: with particle B released to the solution.  

420: 

421: Figure \ref{fig:3} illustrates the change of the free energy surfaces

422: with pB according to eq \ref{eq:14-1} (pK$_a^{(i)}$ values of

423: tyrosine). For the reaction $1\to 2$, the system needs to climb the

424: activation barrier from the bottom of the free energy well $G_1(X)$ to

425: the crossing point at $X=0$. The bottom of the potential well rises

426: with increasing pB at pB$<$pK$_a^{(1)}$ thus resulting in a smaller

427: activation barrier.  The reaction rate depends on pB. The barrier

428: height stops changing once pB reaches pK$_a^{(1)}$, and the reaction

429: rate is insensitive to pB at pB$>$pK$_a^{(1)}$. Notice that the rate

430: of the reverse transition $2\to 1$ remains unchanged in the whole range

431: of pB values.  This invariance makes the effect of pB on the ET rate

432: distinct from the effect of the driving force which alters the

433: activation barriers for both the forward and backward reactions.

434: 

435: When binding does not affect the electric field of the donor-acceptor

436: complex the equations for the first and second cumulants (eqs

437: \ref{eq:12} and \ref{eq:13}) simplify to

438: %

439: \begin{equation}

440:   \label{eq:15}

441:    \langle X\rangle_i = \Delta E_i - \Delta\epsilon n_i

442: \end{equation}

443: and

444: %

445: \begin{equation}

446:   \label{eq:16}

447:   \langle (\delta X)^2 \rangle_i = 2k_{\text{B}}T \lambda_s + \Delta\epsilon^2 n_i(1-n_i)   

448: \end{equation}

449: Several important observations follow from eqs \ref{eq:15} and

450: \ref{eq:16}.  First, the binding/unbinding fluctuations break the

451: connection (eq \ref{eq:0}) between the reorganization energy from the

452: free energy curvature

453: %

454: \begin{equation}

455:   \label{eq:17}

456:   \lambda_i = (\beta/2) \langle (\delta X)^2 \rangle_i

457: \end{equation}

458: and the Stokes shift

459: %

460: \begin{equation}

461:   \label{eq:18}

462:   \Delta X = 2\lambda_s + \Delta\epsilon \Delta n,\quad \Delta n = n_2 - n_1 

463: \end{equation}

464: This effect is the consequence of the local character of the unbinding

465: events contrasting with the quasi-macroscopic nature of the bulk

466: fluctuations resulting in eq \ref{eq:0}.  Second, the reorganization

467: energy in eqs \ref{eq:16} and \ref{eq:17} arising from binding

468: fluctuations depends on the state of the donor-acceptor complex, i.e.\

469: $\lambda_1 \neq \lambda_2$. This condition implies that the free energy surfaces

470: $G_i(X)$ must be non-parabolic\cite{DMacc:07} to comply with the

471: energy conservation,\cite{King:90} $G_2(X)=G_1(X)+X$.  The effect of thermalized

472: bound states on ET cannot therefore be accounted for within the

473: Marcus-Hush formalism. We also note that the variance of the reaction

474: coordinate in eq \ref{eq:14} does not follow the

475: fluctuation-dissipation theorem\cite{Landau5} requiring $ \langle (\delta X)^2\rangle_i

476: \propto T$ and instead has a more complex temperature dependence arising

477: from the equilibrium dissociation constant entering the equilibrium

478: population in eqs \ref{eq:13} and \ref{eq:16}.

479: 

480: 

481: \begin{figure}

482:   \includegraphics*[width=6.5cm]{fig4}

483:   \caption{Stokes shift, $\Delta X$, and the reaction coordinate variance,

484:     $\langle (\delta X)^2\rangle_1$, vs pB for a bound state with the dissociation

485:     constants of tyrosine (pK$_a^{(1)}=10$ and pK$_a^{(2)}=-2$) and

486:     the solvent reorganization energy $\lambda_s=1$ eV. }

487:   \label{fig:4}

488: \end{figure}

489: 

490: The dependence of both the energy gap variance and the Stokes shift on

491: pB are shown in Figure \ref{fig:4}. The equality of $\beta \langle (\delta X)^2\rangle_1$

492: and $\Delta X$, expected from eqs \ref{eq:0}, is seen when both the initial

493: and final ET states are in the same binding state. They are shifted by

494: the binding energy $\Delta \epsilon$ when the the binding states are different in

495: the two ET states (left corner in Figure \ref{fig:4}). The most

496: interesting region is when pB$\simeq$pK$_a$ and the term proportional to

497: $n_i(1-n_i)$ in eqs \ref{eq:13} and \ref{eq:16} can potentially

498: dominate the energy gap variance. Our model thus makes a simple,

499: experimentally testable prediction that the band-width of a

500: charge-transfer optical transition (absorption for binding to the

501: donor) is expected to pass through a maximum as a function of pB.

502: 

503: 

504: 

505: \section{Protonation affecting electron transfer}

506: \label{sec:3}

507: Measurements of oxidation rates of primary pair P$_{680}^+$ by

508: tyrosine in Photosystem II have produced intriguing

509: results.\cite{Ahlbrink:98} The reaction rate increases with increasing

510: pH until it levels off at pH approximately equal to pK$_a$ of phenolic

511: proton. Similar results were reported by Hammarstr{\"o}m and co-workers

512: for intramolecular ET from tyrosine to Ru(III) covalently connected in

513: a donor-acceptor complex.\cite{Sjodin:00,Lomoth:06} In order to

514: explain the observed pH-dependence, they used the Marcus equation for

515: the activation free energy in which a pH-dependent redox potential was

516: substituted. The pH-dependence of the rate arises in their analysis

517: from the dependence of the tyrosine reduction potential on the

518: concentration of protons in the solution.  

519: 

520: This practice,\cite{Sjodin:00,Carra:03} which had not been anticipated

521: in the original Marcus formulation,\cite{Marcus:65} was criticized by

522: Krishtalik\cite{Krishtalik:03} and, more recently, by Sav{\'e}ant and

523: co-workers.\cite{Costenin:07} Krishtalik argued that the Gibbs free

524: energy of ET reactions cannot possibly depend on the ideal mixing

525: entropy of protons in the bulk, which is the origin of the

526: pH-dependent term in the Nernst equation for the redox

527: potential.\cite{Bockris:70} We can only add to this, absolutely

528: correct, argument that using a pH-dependent $\Delta G$ in the Marcus

529: equation clearly violates the Franck-Condon principle.  The vertical

530: \textit{energy} gap $\Delta G + \lambda $ does not involve entropy change since

531: the nuclei do not move on the time-scale of electronic transitions.

532: When the entropic pH term appears in $\Delta G$ and is not compensated by a

533: corresponding term in $\lambda$, the unphysical entropy of protons mixing

534: appears in the vertical transition energy. The present model allows us

535: to account for the pH-dependence of the ET rate without introducing

536: unphysical approximations.

537: 

538: \begin{figure}

539:   \centering

540:   \includegraphics*[width=6.5cm]{fig5}

541:   \caption{Charge shift rates from tyrosine to Ru(III) from ref

542:     \onlinecite{Sjodin:00} (points) and the fit of the data using eq

543:     \ref{eq:14-1} (solid lines).  The lower curve is obtained using

544:     $\lambda_{u}=2.3$ eV, $\lambda_{b}=1.8$ eV, and the non-ergodicity multiplier

545:     $\alpha=0.27$ ($\Delta G_u=-0.54$ eV and $\Delta G_b=0.2$ eV are taken from ref

546:     \onlinecite{Carra:03}).  The upper curve was obtained with the

547:     same activation parameters by increasing the rate preexponent (see

548:     text for discussion).  The vertical arrow indicates the pK$_a^{(1)}$

549:     value for tyrosine at which the ET rate becomes independent of pH.

550:   }

551:   \label{fig:5}

552: \end{figure}

553: 

554: For protonation affecting ET, B$=$H and pB$=$pH. This problem is,

555: however, potentially more complex than binding/unbinding of neutral

556: molecules. The process of deprotonation proceeds by formation of a

557: geminate pair (on the time-scale of picoseconds for photoexcited

558: states\cite{Agmon:05}) followed by a slower diffusion of the released

559: proton in the Coulomb potential of the deprotonated negative

560: charge.\cite{Agmon:05,Perez:07} This slow process may imply that the

561: unbound proton will not be able to sample all possible states

562: available to a particle in an ideal solution on the time-scale of ET,

563: $\tau_{\text{ET}}$.  A full account of such effects\cite{DMacc:07}

564: presently appears hard to achieve.  We will therefore introduce an

565: empirical non-ergodicity multiplier $\alpha$ to account for the effects of

566: insufficient sampling.  The term pK$_a^{(i)}-$pH in eq \ref{eq:8} is

567: replaced with $\alpha(\mathrm{pK}_a^{(i)}-\mathrm{pH})$, where $\alpha$ is a

568: non-ergodicity coefficient between zero and unity. With this

569: correction, the model qualitatively reproduces the dependence of the

570: ET rate on pH observed by Sj\"odin \textit{et al}.\cite{Sjodin:00} for

571: the oxidation of tyrosine by Ru(III).

572:  

573: Sj\"odin \textit{et al}.\cite{Sjodin:00} have monitored the recovery of

574: Ru(II) from Ru(III) produced by flash photolysis and accelerated by

575: electron transfer from tyrosine covalently linked to the ruthenium

576: complex.  Their observed rates, monitoring the arrival of electrons,

577: mathematically correspond to projecting the complex, potentially

578: multi-coordinate,\cite{Agmon:83,Sumi:86,Soudackov:00} dynamics of the

579: system onto one single ET reaction coordinate, which is exactly the

580: scenario considered here. Since proton is a charged particle, we need

581: the full formulation given by eq \ref{eq:8} to analyze the

582: experimental rates. The reaction Gibbs energies and reorganization

583: energies of ET will potentially be different for the bound and unbound

584: ET pathways and indeed the reaction free energies are $\Delta G_u = - 0.54$

585: eV and $\Delta G_b=0.2$ eV for the unbound and bound proton states,

586: respectively.\cite{Carra:03} The unknown parameters are the rate

587: preexponent, reorganization energies $\lambda_{u,b}$, and the non-ergodicity

588: multiplier $\alpha$.  This latter parameter is expected to be small because

589: of the slow rate of proton dissociation from tyrosine\cite{Sjodin:00}

590: and thus a lower extent of modulation of the donor-acceptor gap by

591: unbinding events than it would be possible for full thermalization.

592: 

593: \begin{figure}

594:   \centering

595:   \includegraphics*[width=6.5cm]{fig6}

596:   \caption{Stokes shift, $\Delta X$, and the variance, $\langle(\delta

597:     X)^2 \rangle_1$, calculated from eqs \ref{eq:12} and \ref{eq:13} using

598:     the reorganization parameters obtained by fitting the experimental

599:     data in Figure \ref{fig:5}.  }

600:   \label{fig:6}

601: \end{figure}

602: 

603: 

604: In order to see if the model can qualitatively account for the

605: experimental observations, we have used the rate preexponent,

606: reorganization energies $\lambda_{u,b}$, and $\alpha$ as free parameters to fit

607: the data from ref \onlinecite{Sjodin:00}. The result is shown in

608: Figure \ref{fig:5}.  The fitting curve ($\lambda_u=2.3$ eV, $\lambda_b=1.8$ eV,

609: and $\alpha=0.26$) of $\log_{10}(k_{\text{ET}})$ vs pH rises linearly with

610: the slope $\alpha$ when pH is below pK$_a^{(1)}$ and levels off after

611: reaching this value.  The reorganization energies $\lambda_{u,b}$ here

612: include reorganization of classical intramolecular vibrations in

613: addition to solvent reorganization.  The parameters extracted from the

614: fit of the rates result in a dramatic violation of eq \ref{eq:0} as is

615: shown in Figure \ref{fig:6}.

616: 

617: 

618: Fitting the experimental jump in the rate at

619: $\mathrm{pH}=\mathrm{pK}_a^{(1)}$ requires a higher value of the rate

620: preexponent as is shown by the fragment of the curve obtained by using

621: the same parameters as in the low-pH fit, but varying the preexponent.

622: The increase of the rate at $\mathrm{pH}=\mathrm{pK}_a^{(1)}$ has been

623: addressed by Carra \textit{et al}.\cite{Carra:03} The low-pH portion

624: of the curve corresponds to electronic transition accompanied by

625: simultaneous proton release (proton-coupled

626: ET\cite{Hammes-Schiffer:01}). The nonadiabatic matrix element in the

627: rate preexponent then includes small Franck-Condon overlap between

628: proton bound and free states.\cite{Cukier:96} This overlap disappears

629: in the flat portion of the plot at $\mathrm{pH}>\mathrm{pK}_a^{(1)}$

630: when tyrosine is deprotonated in both ET states. Only electronic

631: coupling enters the rate preexponent in this regime resulting in a

632: higher rate.  A more quantitative analysis would require extensive

633: calculations.  Since the model presented here does not aim at a

634: detailed quantitative analysis of proton-coupled ET, we limit our

635: discussion to qualitative observations only.

636: 

637: 

638: 

639: 

640: \section{Conclusions}

641: \label{sec:4}

642: %

643: Traditional theories of ET in polar liquids emphasize activation of

644: electronic transitions by long-range, quasi-macroscopic solvent

645: fluctuations.\cite{Marcus:65} Local binding to the donor and acceptor

646: sites provides an additional modulation of the donor-acceptor energy

647: gap. This study addressed the question of whether this additional

648: thermal noise can be accounted for within the standard framework of

649: Gaussian fluctuations of the energy gap, that is by adjusting the

650: magnitudes of the driving force and reorganization energy. 

651: 

652: We have found that the statistics of binding events are non-Gaussian,

653: and the resultant free energy surfaces cannot be reduced to crossing

654: parabolas. The model predicts a regime of a significant dependence of

655: the activation barrier on the concentration of the binding particles

656: in the solution. When the concentration pB is close to the binding

657: equilibrium constant pK$_a$, the variance of the energy gap passes

658: through a maximum which can be observed by spectroscopy of

659: charge-transfer bands.  Despite these new features, the main effect of

660: binding is in shifting the free energy gap of ET, as was suggested

661: previously,\cite{Mayer:04} leaving the reorganization energy mostly

662: unaffected and within the realm of standard models.  Finally, the

663: model helps to rationalize some recent observations of the dependence

664: of rates of intramolecular ET, involving deprotonation of reactants,

665: on the pH of the solution.

666: 

667: \begin{acknowledgments}

668:   This work was supported by the NSF (CHE-0616646).

669: \end{acknowledgments}  

670: 

671: 

672: %\bibliographystyle{apsrev}

673: % \bibliography{/home/dmitry/p/bib/chem_abbr,/home/dmitry/p/bib/photosynthNew,/home/dmitry/p/bib/glass,/home/dmitry/p/bib/et,/home/dmitry/p/bib/dm,/home/dmitry/p/bib/pt,/home/dmitry/p/bib/solvation,/home/dmitry/p/bib/heteroet,/home/dmitry/p/bib/bioet,/home/dmitry/p/bib/protein}

674: 

675: \providecommand{\refin}[1]{\\ \textbf{Referenced in:} #1}

676: \begin{thebibliography}{10}

677: 

678: \bibitem{Reichardt:94}

679: Reichardt,~C. \textit{Chem.\ Rev.} \textbf{1994,} \textsl{94,} 2319-2358.

680: 

681: \bibitem{DMjpcb:97}

682: Matyushov,~D.~V.;\ \ Schmid,~R.;\ \ Ladanyi,~B.~M. \textit{J. Phys. Chem. B}

683:   \textbf{1997,} \textsl{101,} 1035.

684: 

685: \bibitem{SolvPolarity:04}

686: Katritzky,~A.~R.;\ \ Fara,~D.~C.;\ \ Yang,~H.;\ \ T{\"a}mm,~K. \textit{Chem.

687:   Rev.} \textbf{2004,} \textsl{104,} 175.

688: 

689: \bibitem{Agmon:05}

690: Agmon,~N. \textit{J. Phys. Chem. A} \textbf{2005,} \textsl{109,} 13.

691: 

692: \bibitem{McEvoy:06}

693: McEvoy,~J.~P.;\ \ Brudvig,~G.~W. \textit{Chem. Rev.} \textbf{2006,}

694:   \textsl{106,} 4455.

695: 

696: \bibitem{Cukier:04}

697: Cukier,~R.~I. \textit{Biochim. Biophys. Acta} \textbf{2004,} \textsl{1655,}

698:   37-44.

699: 

700: \bibitem{Cukier:96}

701: Cukier,~R. \textit{J.\ Phys.\ Chem.} \textbf{1996,} \textsl{100,} 15428.

702: 

703: \bibitem{Soudackov:00}

704: Soudackov,~A.;\ \ Hammes-Schiffer,~S. \textit{J. Chem. Phys.} \textbf{2000,}

705:   \textsl{113,} 2385.

706: 

707: \bibitem{Hammes-Schiffer:01}

708: Hammes-Schiffer,~S. \textit{Acc. Chem. Res.} \textbf{2001,} \textsl{34,} 273.

709: 

710: \bibitem{Ferguson-Miller:96}

711: Ferguson-Miller,~S.;\ \ Babcock,~G. \textit{Chem. Rev.} \textbf{1996,}

712:   \textsl{96,} 2889.

713: 

714: \bibitem{Kirby:97}

715: Kirby,~A. \textit{Acc. Chem. Res.} \textbf{1997,} \textsl{30,} 290-296.

716: 

717: \bibitem{Mayer:04}

718: Mayer,~J.~M. \textit{Annu. Rev. Phys. Chem.} \textbf{2004,} \textsl{55,} 363.

719: 

720: \bibitem{Marcus:93}

721: Marcus,~R.~A. \textit{Rev. Mod. Phys.} \textbf{1993,} \textsl{65,} 599.

722: 

723: \bibitem{Borgis:96}

724: Borgis,~D.;\ \ Hynes,~J. \textit{J.\ Phys.\ Chem.} \textbf{1996,} \textsl{100,}

725:   1118.

726: 

727: \bibitem{Marcus:65}

728: Marcus,~R.~A. \textit{J. Chem. Phys.} \textbf{1965,} \textsl{43,} 679.

729: 

730: \bibitem{King:90}

731: King,~G.;\ \ Warshel,~A. \textit{J. Chem. Phys.} \textbf{1990,} \textsl{93,}

732:   8682.

733: 

734: \bibitem{Marcus:89}

735: Marcus,~R.~A. \textit{J. Phys. Chem.} \textbf{1989,} \textsl{93,} 3078.

736: 

737: \bibitem{DMacc:07}

738: Matyushov,~D.~V. \textit{Acc. Chem. Res.} \textbf{2007,} \textsl{40,} 294.

739: 

740: \bibitem{Sumi:86}

741: Sumi,~H.;\ \ Marcus,~R.~A. \textit{J. Chem. Phys.} \textbf{1986,} \textsl{84,}

742:   4894.

743: 

744: \bibitem{Stuchebrukhov:03}

745: Stuchebrukhov,~A.~A. \textit{J. Theor. Comp. Chem.} \textbf{2003,} \textsl{2,}

746:   91.

747: 

748: \bibitem{Gutman:97}

749: Gutman,~M.;\ \ Nachliel,~E. \textit{Annu. Rev. Phys. Chem.} \textbf{1997,}

750:   \textsl{48,} 329.

751: 

752: \bibitem{Halle:04}

753: Halle,~B. \textit{Phil. Trans. R. Soc. Lond.} \textbf{2004,} \textsl{359,}

754:   1207.

755: 

756: \bibitem{German:92}

757: German,~E.~D.;\ \ Kuznetsov,~A.~M. \textit{Electrokhimiya} \textbf{1992,}

758:   \textsl{28,} 294.

759: 

760: \bibitem{Landau5}

761: Landau,~L.~D.;\ \ Lifshits,~E.~M. \textit{Statistical Physics;} Pergamon Press:

762:   New York, 1980.

763: 

764: \bibitem{Ahlbrink:98}

765: Ahlbrink,~R.;\ \ Haumann,~M.;\ \ Cherepanov,~D.;\ \ B{\"o}gershausen,~O.;\ \

766:   Mulkidjanian,~A.;\ \ Junge,~W. \textit{Biochemistry} \textbf{1998,}

767:   \textsl{37,} 1131.

768: 

769: \bibitem{Sjodin:00}

770: Sj{\"o}din,~M.;\ \ Styring,~S.;\ \ {\AA}kermark,~B.;\ \ Sun,~L.;\ \

771:   Hammarstr{\"o}m,~L. \textit{J. Am. Chem. Soc.} \textbf{2000,} \textsl{122,}

772:   3932.

773: 

774: \bibitem{Lomoth:06}

775: Lomoth,~R.;\ \ Magnuson,~A.;\ \ Sj{\"o}din,~M.;\ \ Huang,~P.;\ \ Styring,~S.;\

776:   \ Hammarstr{\"o}m,~L. \textit{Photosynth. Res.} \textbf{2006,} \textsl{87,}

777:   25.

778: 

779: \bibitem{Carra:03}

780: Carra,~C.;\ \ Iordanova,~N.;\ \ Hammes-Schiffer,~S. \textit{J. Am. Chem. Soc.}

781:   \textbf{2003,} \textsl{125,} 10429.

782: 

783: \bibitem{Krishtalik:03}

784: Krishtalik,~L.~I. \textit{Biochim. Biophys. Acta} \textbf{2003,} \textsl{1604,}

785:   13.

786: 

787: \bibitem{Costenin:07}

788: Costenin,~C.;\ \ Robert,~M.;\ \ Sav{{\'e}}ant,~J.-M. \textit{J. Am. Chem. Soc.}

789:   \textbf{2007,} \textsl{129,} 5870.

790: 

791: \bibitem{Bockris:70}

792: Bockris,~J.~O. \textit{Modern Electrochemistry;} McDonald: London, 1970.

793: 

794: \bibitem{Perez:07}

795: P{\'e}rez-Lustres,~J.~L.;\ \ Rodriguez-Prieto,~F.;\ \ Mosquera,~M.;\ \

796:   Senyushkina,~T.~A.;\ \ Ernsting,~N.~P.;\ \ Kovalenko,~S.~A. \textit{J. Am.

797:   Chem. Soc.} \textbf{2007,} \textsl{129,} 19.

798: 

799: \bibitem{Agmon:83}

800: Agmon,~N.;\ \ Hopfield,~J.~J. \textit{J. Chem. Phys.} \textbf{1983,}

801:   \textsl{78,} 6947.

802: 

803: \end{thebibliography}

804: 

805: 

806: 

807: 

808: \end{document}

809: 

810: \newpage

811: 

812: \begin{center}

813:   \textbf{FIGURE CAPTIONS}

814: \end{center}

815: 

816: \begin{description}

817: \item[FIGURE 1] Diagram of the free energy surfaces of ET obtained from

818:     three parameters, the reaction free energy $\Delta G$, the free energy

819:     curvature $(2\lambda)^{-1}$, and the Stokes shift (separation between

820:     the minima $\Delta X$). The reaction coordinate $X$ is the energy

821:     separation between the donor and acceptor electronic states. The

822:     activated state is given by the resonance of the donor and

823:     acceptor levels, $X=0$, while $X_{01}$ and $X_{02}$ indicate the

824:     positions of the minima.

825: \item[FIGURE 2] Diagram of particle B exchanging between the bound and

826:     dissolved states with the rate constant $k_b$. The equilibrium

827:     population of the bound state assumes that the unbinding rate is

828:     much faster than the rate of ET ($k_{\text{ET}}\ll k_b$); $\Delta

829:     \epsilon=\epsilon_2-\epsilon_1$ denotes the change in the binding energy caused by the

830:     electronic transition.

831: \item[FIGURE 3] Free energy surfaces of ET (eq \ref{eq:14-1}) at different

832:     values of pB indicated in the plot assuming pK$_a^{(1)}=10$ and

833:     pK$_a^{(2)}=-2$ (equilibrium constants of tyrosine). The final

834:     charge-transfer state ($i=2$) is always deprotonated while the

835:     initial charge-transfer state ($i=1$) is protonated at

836:     pB$<$pK$_a^{(1)}$ and is deprotonated at pB$>$pK$_a^{(1)}$. The

837:     Stokes shift for the protonated state 1 is $2\lambda_s - \Delta\epsilon = 2.71$ eV

838:     and is equal to $2\lambda_s=2$ eV for the deprotonated state 1. The

839:     solvent bulk component of the reaction Gibbs energy reaction is $\Delta

840:     G = -0.1$ eV.

841: \item[FIGURE 4] Stokes shift, $\Delta X$, and the reaction coordinate variance,

842:     $\langle (\delta X)^2\rangle_1$, vs pB for a bound state with the dissociation

843:     constants of tyrosine (pK$_a^{(1)}=10$ and pK$_a^{(2)}=-2$) and

844:     the solvent reorganization energy $\lambda_s=1$ eV.

845: \item[FIGURE 5] Charge shift rates from tyrosine to Ru(III) from ref

846:     \onlinecite{Sjodin:00} (points) and the fit of the data using eq

847:     \ref{eq:14-1} (solid lines).  The lower curve is obtained using

848:     $\lambda_{u}=2.3$ eV, $\lambda_{b}=1.8$ eV, and the non-ergodicity multiplier

849:     $\alpha=0.27$ ($\Delta G_u=-0.54$ eV and $\Delta G_b=0.2$ eV are taken from ref

850:     \onlinecite{Carra:03}).  The upper curve was obtained with the

851:     same activation parameters by increasing the rate preexponent (see

852:     text for discussion).  The vertical arrow indicates the pK$_a^{(1)}$

853:     value for tyrosine at which the ET rate becomes independent of pH.

854: \item[FIGURE 6] Stokes shift, $\Delta X$, and the variance, $\langle(\delta

855:     X)^2 \rangle_1$, calculated from eqs \ref{eq:12} and \ref{eq:13} using

856:     the reorganization parameters obtained by fitting the experimental

857:     data in Figure \ref{fig:5}.

858: \end{description}

859: 

860: \newpage

861:  \includegraphics*[width=9cm]{fig1}

862: \vfill

863: \textbf{Figure 1. Matyushov}

864: \newpage

865:  \includegraphics*[width=9cm]{fig2}

866: \vfill

867: \textbf{Figure 2. Matyushov}

868: \newpage

869:  \includegraphics*[width=9cm]{fig3}

870: \vfill

871: \textbf{Figure 3. Matyushov}

872: 

873: \newpage

874:  \includegraphics*[width=9cm]{fig4}

875: \vfill

876: \textbf{Figure 4. Matyushov}

877: 

878: \newpage

879:  \includegraphics*[width=9cm]{fig5}

880: \vfill

881: \textbf{Figure 5. Matyushov}

882: 

883: \newpage

884:  \includegraphics*[width=9cm]{fig6}

885: \vfill

886: \textbf{Figure 6. Matyushov}

887: 

888: \end{document}

889: