Theoretical amplitudes and lifetimes of nonradial solarlike oscillations in red giants
Key Words.:
Stars: oscillations – Stars: interiors – ConvectionAbstract
Context:Solarlike oscillations have been observed in numerous red giants from ground and from space. An important question arises: could we expect to detect nonradial modes probing the internal structure of these stars?
Aims:We investigate under what physical circumstances nonradial modes could be observable in red giants; what would be their amplitudes, lifetimes and heights in the power spectrum (PS)?
Methods:Using a nonradial nonadiabatic pulsation code including a nonlocal timedependent treatment of convection, we compute the theoretical lifetimes of radial and nonradial modes in several red giant models. Next, using a stochastic excitation model, we compute the amplitudes of these modes and their heights in the PS.
Results:Distinct cases appear. Case A corresponds to subgiants and stars at the bottom of the ascending giant branch. Our results show that the lifetimes of the modes are mainly proportional to the inertia , which is modulated by the mode trapping. The predicted amplitudes are lower for nonradial modes. But the height of the peaks in the PS are of the same order for radial and nonradial modes as long as they can be resolved. The resulting frequency spectrum is complex. Case B corresponds to intermediate models in the red giant branch. In these models, the radiative damping becomes high enough to destroy the nonradial modes trapped in the core. Hence, only modes trapped in the envelope have significant heights in the PS and could be observed. The resulting frequency spectrum of detectable modes is regular for and , but a little more complex for modes because of less efficient trapping. Case C corresponds to models of even higher luminosity. In these models the radiative damping of nonradial modes is even larger than in the previous case and only radial and nonradial modes completely trapped in the envelope could be observed. The frequency pattern is very regular for these stars. The comparison between the predictions for radial and nonradial modes is very different if we consider the heights in the PS instead of the amplitudes. This is important as the heights (not the amplitudes) are used as detection criterion.
Conclusions:
1 Introduction
Oscillations of high radial order pmodes have been firmly detected in several red giants, from radial velocity data (Frandsen et al., 2002; Bedding & Kjeldsen, 2006) and photometric data from space (Barban et al., 2007; De Ridder et al., 2009). These oscillations are stochastically excited by turbulent motion in the outer layers of these stars. A particularly interesting case is Oph, in which De Ridder et al. (2006) detected equidistant frequencies from groundbased spectroscopic observations. Hekker et al. (2006) analysed the lineprofile variations and advocate the presence of nonradial modes. For the same star, Barban et al. (2007) used data obtained by the MOST satellite. They detected at least 7 equidistant peaks and derived mode lifetimes of about 2.7 days. Kallinger et al. (2008a) reexamined the available data for Oph and claim to detect at least 21 independent modes with longer lifetimes between 10 and 20 days. Kallinger et al. (2008b) claim to have detected nonradial modes with long lifetimes (more than 10 days) in the star HD 20884 observed by MOST. A very different result was found by Stello et al. (2006) who claim to have detected modes with very short lifetimes in Hydrae. We finally mention the recent detection by the CoRoT satellite of solarlike radial and nonradial oscillations in many red giants (De Ridder et al., 2009). Different kinds of power spectra are found in these data of unprecedented quality, some of them showing regular patterns while others are more complex. Modes with very long lifetimes (more than 50 days) are detected.
From a theoretical point of view, an important study was made by Dziembowski et al. (2001) (hereafter D01). They underline the mode trapping phenomenon present in these stars and the role of radiative damping for several models with different luminosities. They predict that nonradial modes not trapped in the envelope would have much smaller amplitudes than radial modes. Houdek & Gough (2002) carried out theoretical computations of the amplitudes and lifetimes of radial modes in the star Hydrae. They found lifetimes of the order of 1520 days with a bump going up to infinite value (unstable mode) around Hz. Finally, ChristensenDalsgaard (2004) also discuss the mode trapping and inertia of nonradial adiabatic modes in red giants and subgiants and the possible consequences for the amplitudes.
In the present study, we compute theoretical lifetimes of radial and nonradial modes of several red giant models, using a fully nonradial nonadiabatic pulsation code including the nonlocal timedependent treatment of convection by Grigahcène et al. (2005) (G05). We discuss how these lifetimes are affected by the mode inertia, the radiative damping in the core and the convective damping in the upper part of the convective envelope. Using a stochastic excitation model (Samadi & Goupil, 2001; Samadi et al., 2003b, a), we also compute the amplitudes of radial and nonradial modes and their theoretical heights in the power spectrum (PS). As we show below, predictions for the heights are very different from the amplitudes. Although nonradial modes in general have smaller amplitudes than radial modes, this is not always the case for the heights. As the heights in the PS are used as a criterion for the detection of frequencies, it is important to make this distinction.
In Sect. 2, we discuss the different energetic processes determining the driving and damping of the modes: the radiative damping in the core, the coherent interaction of convection with oscillations and the stochastic excitation by turbulent motion in the upper part of the convective envelope. In Sect. 3, we present the numerical tools, models and method used to solve the problem. In Sect. 4, we present our results for specific red giant models representative of the different cases that can occur.
2 Energetic aspects of oscillations
Different mechanisms lead to the driving and damping of oscillations. For observed modes, two possibilities exist. Firstly, they could be selfexcited. In such case the amplitudes are expected to grow until a large amplitude limit cycle is reached. The observed amplitudes of the high radial order pmodes in red giants are of several m/s. This is too small for such an interpretation and we reject this possibility here. Secondly, the modes can be damped, but turbulent motions supply them with energy in a stochastic way, allowing them to reach observed amplitudes. We assume here that the high frequency modes of red giants are stochastic excited modes. The damping rate of the modes is given by an integral expression of this type:
(1) 
where we assumed the timedependence ( is the time, is the angular frequency and is the damping rate). is the work performed by the gas during one oscillation cycle. is the displacement vector, is the dimensionless mode inertia:
(2) 
is the total mass and corresponds to the radial displacement at the layer where the oscillations are measured (Rosseland optical depth in our results). In radiative zones, the work is easily modeled. It just results from the heat given to the gas by radiation during each oscillation cycle. In convective zones, the problem is much more complex and different terms can lead to damping: variations of convective flux, turbulent pressure, viscosity, and dissipation of turbulent kinetic energy (G05). Stochastic excitation by turbulent motion also occurs, leading to the observed amplitudes. This is discussed in Sect. 2.3.
2.1 Radiative damping
In the gmode cavity of red giants, the radial wavenumber becomes huge because of the high density contrast between the core and the envelope ( is the BruntVäisälä frequency). Large variations of the temperature gradient ensue, with loss of heat in the hot phase, leading to radiative damping. In the asymptotic limit, a simple expression for this damping can be obtained (Dziembowski, 1977b; Van Hoolst et al., 1998; Godart et al., 2009):
(3) 
where is the radius of the convective core (if present), is the upper radius of the gmode cavity, is a normalization constant obtained by appropriate matching with the envelope solution, and are the adiabatic and real gradients, the gravity, the local luminosity and the pressure. In the central regions of red giants, is very high because of the high density contrast. But at the same time, increases rapidly as we enter in the dense pure He core. As a compromise between these two tendencies, the integrand of Eq. (3) reaches its largest values around the bottom of the Hburning shell. The main radiative damping thus occurs in this region for red giants.
On the other side, the contribution of the gmode cavity to the denominator of Eq. (1) is simply:
(4) 
which also can be significant because of the large ( in the core). Eq. (3) scales like , where is the KelvinHelmholtz time, is the radius of the He core, is the total radius, is the mass of the He core and is the total mass. Equation (4) scales like . As the star climbs the red giant branch, and increase so that the numerator increases more quickly than the denominator and the radiative damping of nonradial modes increases.
2.2 Coherent interaction with convection
A very important region for the driving and damping of the modes is the transition region where the thermal relaxation time is of the same order as the oscillation periods. For solarlike oscillations, this transition region corresponds to the upper part of the convective envelope. Moreover, in this region the timescale of most energetic turbulent eddies is of the same order as the oscillation periods. Hence, the changing physical conditions due to the oscillations lead to periodic variations of the turbulent quantities (convective flux, Reynolds stress, …), which contribute to the work (positive or negative) performed by the modes. This is what we call the coherent interaction between convection and oscillations. It plays a major role in the damping of high radial order pmode oscillations of red giants.
This interaction is difficult to model and only a few theories have been proposed for linear oscillations. Two of them make use of a MixingLength formalism. The first by Gabriel (1996) and further developed by G05 follows the original ideas of Unno (1967), where a turbulent viscous term opposite to the buoyancy is introduced. This treatment implemented in our nonradial nonadiabatic pulsation code is used in this study. The second theory was developed by Gough (1977b) and follows the “kinetic of gas” picture of the MLT. Both theories can include a nonlocal treatment (Spiegel, 1963; Gough, 1977a; Balmforth, 1992; Dupret et al., 2006b). A third formulation, no longer based on a mixinglength approach but on a Reynolds stress one was also proposed by Xiong et al. (1997). These theories encountered some successes, for example they obtain the red edge of the instability strip (Houdek, 2000; Xiong & Deng, 2001; Dupret et al., 2004). The reproduction of the Solar mode lifetime is also possible but not easy (Balmforth, 1992; Xiong et al., 2000; Dupret et al., 2006a).
2.3 Stochastic excitation
For stochastically excited modes, the local squared amplitude of velocity variation at the layer where it is measured (we adopt the optical depth in our results) is given by (e.g., Baudin et al., 2005; Belkacem et al., 2006):
(5) 
is the power stochastically supplied to the modes:
(6) 
where and are the turbulent Reynolds stress and entropy contributions, respectively (see Samadi & Goupil, 2001; Belkacem et al., 2006, for details). Both play a significant role in red giants and are included in our study. and are the damping rate and inertia defined in Eqs. (1) and (2), respectively. To isolate the effect of inertia in Eq. (5), we introduce the product which is, according to Eq. (6), independent of inertia.
We also introduce the maximum height of the mode profile in the PS, which is an observable and will permit us to draw conclusions about the mode detectability (see Sect. 4 for details). To this end, one has to distinguish between two cases, namely resolved and unresolved modes. The resolved modes present a Lorentzian profile in the PS and their heights are given by (see e.g. Chaplin et al., 2005; Belkacem et al., 2006):
(7) 
where is the mode lifetime.
When ( being the duration of observations), the modes are not resolved. In the limit , the heights in the PS tend to behave like (see e.g. Berthomieu et al., 2001; Lochard et al., 2005):
(8) 
In the theoretical predictions of the heights presented in this paper, we choose the very favorable case of the CoRoT long runs with days. We use Eq. (7) when days and Eq. (8) when days, so that is a continuous function of .
Finally, the observed velocity amplitudes are obtained by integration of the projected local velocity over the visible stellar disk (Dziembowski, 1977a). This introduces a visibility factor depending on the inclination angle of the star (through the factor ). We do not include this factor in our theoretical study (the inclination angle is unknown).
3 Numerical tools, models and method
3.1 Structure models
The equilibrium models analyzed in this study were computed with the version updated for asteroseismology of the code ATON (Mazzitelli, 1979). A detailed description of the numerical techniques and implemented physics is given in Ventura et al. (2008) and references therein. For the models in this paper we adopted the classical mixinglength theory of convection (BöhmVitense, 1958), with a value of the mixinglength parameter , and we assume core overshooting that was treated as a diffusive mixing process. For the parameter describing the extension of that extra mixing we took (see Ventura et al., 1998, for details), and for the chemical composition and . The radiative opacities are those from OPAL (Iglesias & Rogers, 1996) for the metal mixture of Grevesse & Noels (1993) completed with Alexander & Ferguson (1994) at low temperatures, and the conductive ones come from Potekhin et al. (1999). Thermodynamic quantities are derived from OPAL (Rogers & Nayfonov, 2002), Saumon et al. (1995) for the pressure ionization regime and Stolzmann & Bloecker (1996) treatment is used for He/C/O mixtures. The nuclear crosssections are from the NACRE compilation (Angulo et al., 1999), and the surface boundary conditions are provided by a grey atmosphere following the Henyey et al. (1965) treatment. Turbulent pressure is not included in our structure models.
We focus our study on five well chosen models representative of the different situations that can occur. Some of their global characteristics are given in Table 1. is the cutoff frequency of an isothermal planeparallel atmosphere (estimated here at the photosphere). Evolutionary tracks with the location of these models in the HR diagram are given in Fig. 1.
(K)  ()  

A  2  5264  1.32  3.26  5.50  315 
B  2  4892  1.80  2.65  11.0  82.1 
C  2  4665  2.10  2.27  17.2  34.6 
D  3  5091  2.00  2.70  12.8  88.6 
E  3  5222  2.00  2.74  12.2  96.3 
3.2 Nonradial nonadiabatic oscillations
We use the nonadiabatic pulsation code MAD (Dupret, 2002) to compute the theoretical lifetimes and inertia of the modes. We use the TDC treatment of G05 for the modelling of the convective flux and turbulent pressure variations entering in the nonadiabatic equations. Different ingredients must be specified. We keep a local treatment for the convective flux. The variations of the turbulent pressure and entropy gradient are treated in a nonlocal way, with values of the nonlocal parameters and according to the definition of Balmforth (1992). Our value of is derived from a fit of the turbulent pressure in the subadiabatic atmospheric layers of a solar hydrodynamic simulation (Dupret et al., 2006b). We recall that our structure models are built without turbulent pressure. We compute it a posteriori in order to be able to determine the turbulent pressure perturbation. This implies that the effect of turbulent pressure on the stratification is not taken into account in our models. However, the impact of this inconsistency on the theoretical lifetimes is expected to be negligible compared to others aspects. The main source of uncertainties in any TDC treatment comes from the perturbation of the closure equations. In our TDC treatment, this uncertainty appears in the form of a free parameter , as introduced by G05. We take by default the value . Adopting this value for the Sun leads to predictions of the mode lifetimes in reasonable agreement with observations. It also gives in all our red giant models stable high radial order pmodes, in agreement with our working hypothesis of stochastic excitation. However, we discuss in Sect. 4.6 the impact of changing this value.
Some changes were required in the MAD code to address specific numerical difficulties associated with red giants. First, we changed the formalism to a Eulerian one in the core. In the gmode cavity of red giants where , we have and thus . Hence, and are not independent functions and they do not allow us to capture the oscillatory behaviour of the eigenfunctions in the cavity. Therefore, a Lagrangian formalism lacks numerical precision in the dense region of evolved stars, which justifies our switch to a Eulerian formalism.
Another significant numerical difficulty comes from the very large number of nodes of the eigenfunctions in the gmode cavity. Van Hoolst et al. (1998) and D01 used a mixed treatment to address this difficulty: numerical solutions of the fully nonadiabatic equations in the envelope are matched with asymptotic solutions in the gmode cavity. Our way to proceed is slightly different. Firstly we solve the fully nonadiabatic equations from the center to the surface, which gives a first estimation of the lifetimes. Second, we separate the integrals in Eqs. (1) and (2) in two parts: region 1 where and region 2 (the remaining). In region 1 we use the results obtained in the asymptotic quasiadiabatic limit (Eqs. (3) and (4)), and in region 2 we use the fully nonadiabatic eigenfunctions obtained numerically in the first step. We then compare the results obtained by the two approaches: in many cases they are very close, showing that our fully numerical solution of the problem is precise enough. In other cases corresponding to high luminosity red giants, the number of nodes becomes huge (), which implies loss of numerical precision. In such cases we are more confident in using results with the asymptotic treatment in the gmode cavity and the fully nonadiabatic solution in the rest of the star.
A last important numerical issue is associated with the mode trapping phenomenon, which is discussed in detail in Sect. 4.1. Algorithms solving the nonadiabatic problem converge towards the closest eigenvalue in the complex plane (the real part corresponds to the angular frequency and the imaginary part to the damping rate ). Some nonradial modes trapped in the envelope have much lower inertia than the others. Hence, their damping rates are much higher than for other nonradial modes, higher than the frequency separation between consecutive nonradial modes: . Hence, with a real initial guess of the eigenvalues (e.g. the adiabatic frequencies), it is impossible to converge towards the modes strongly trapped in the envelope. The only way is to adopt a complex initial value; for example, interpolating the complex nonadiabatic eigenvalues of radial modes gives adequate initial values, allowing the solution to converge towards the nonradial modes strongly trapped in the envelope.
3.3 Ingredients for the stochastic excitation models
The power stochastically supplied to the modes, in Eqs. (5) and (6), is computed as described in Belkacem et al. (2008). The typical convective lengthscales as well as the kinetic energy spectrum are poorly known for red giant stars. Given that the predictions of the stochastic excitation models strongly depend on them, we infer both the kinetic energy spectrum and the injection lengthscale from a representative 3D numerical simulation. To this end, we computed a 3D radiationhydrodynamical model atmosphere with the code COBOLD (Freytag et al., 2002; Wedemeyer et al., 2004). The simulation has a gravity 2.5 and an effective temperature of K. The model has a spatial mesh with grid points, and a physical extent of the computational box of Mm. Using this 3D numerical simulation, we determine the injection lengthscale in the layer where the driving is the largest (see Samadi et al., 2003b, for details). This estimation of the injection scale gives values larger than the mixinglength by a factor of 510 depending on the model. Using the mixinglength as the injection scale would give theoretical amplitudes much lower than typical observations. We assume here that the injection scale is constant in the driving region. In the case of the Sun, such approximation underestimates by about the theoretical amplitudes (Samadi et al., 2003b). The models considered in our study have different gravities. This affects the injection lengthscale and we assume here that it scales as the pressure scaleheight at the photosphere. Such a procedure has been verified in the case of Cen A (Samadi et al., 2008). This allows us to compute the excitation rates for all the models presented in Sect. 3.1. We use a Lorentzian profile for the eddy timecorrelation function (Samadi et al., 2003a). Eventually, concerning the dependency of the kinetic energy spectrum ( is the wavenumber in the Fourrier space of turbulence), we use the Broad Kolmogorov Spectrum (BKS) (Musielak et al., 1994) that reproduces the spectrum inferred from the 3D simulation.
4 Results
4.1 Mode trapping in evolved stars
Mode trapping is a very important phenomenon allowing to understand the behaviour of nonradial modes in evolved stars. D01 and ChristensenDalsgaard (2004) discussed it in the case of red giants. Here we recall basic aspects that are important for the interpretation of our results. The BruntVäisälä () and Lamb () frequencies are huge in the dense core of red giants. Because of these huge values, all nonradial modes with frequencies lower than the cutoff frequency and larger than the fundamental radial mode have a mixed character. In the envelope, and they behave like acoustic modes (as in solarlike main sequence stars). But in the core and the same modes behave there like gravity modes. The presence of an evanescent region between the p and gcavities is at the origin of the mode trapping. Some modes have significant kinetic energy in the gcavity and low ones in the pcavity; they are trapped in the core. Others have low kinetic energy in the gcavity but high ones in the envelope; they are trapped in the envelope.
An illustration of the distribution of energy of a mode trapped in the core and a mode trapped in the envelope are shown in Fig. 2, where we give . Mode trapping directly affects the mode inertia. We give in Fig. 3 the inertia as a function of frequency for modes. We first notice how dense the spectrum of frequencies for nonradial modes is, which is a direct consequence of the huge BruntVäisälä frequency in the core (according to the asymptotic theory, ). Modes trapped in the envelope are the local minima with low inertia in Fig. 3. The modes trapped in the envelope have lower inertia than the modes trapped in the envelope, while it is the contrary for all other modes. In agreement with D01 and ChristensenDalsgaard (2004), the size of the evanescent region and thus the efficiency of trapping increases with the degree of the modes. This is very different from RR Lyrae stars (Van Hoolst et al., 1998) because there the evanescent region is larger for in the observed range of frequency. It is very interesting to notice that when only the modes trapped in the envelope are considered, the frequency spectrum is similar to solartype main sequence stars, with large and small separations appearing clearly. The inertia appears directly in Eqs. (1), (5) and (8). Hence, the mode trapping is expected to strongly affect the amplitudes, lifetimes, and heights of the modes in the PS of red giants. In the next section, we present our main results for different models (Table 1, Fig. 1).
4.2 Model A: bottom of the red giant branch
We first consider the lifetimes of radial modes from the fundamental up to the cutoff frequency. We also give the resolution limit of 75 days (horizontal line). As can be seen, we can separate the modes into two sets. From the fundamental to p (Hz), the lifetimes are very long and such modes could not be resolved. On the other side, for larger frequencies the lifetimes of radial modes are below 75 days and these radial modes have a chance of being resolved. Current observations of red giant solarlike oscillations show that at least some of the modes can be resolved. We note also that Xiong & Deng (2007) predict unstable modes in the lower part of the spectrum. Depending on the parameter used in our TDC model, they can be stable or unstable. We restrict our study to the second set of modes, in which radial modes could be resolved. We propose to call them solarlike modes by analogy with the solar case. Such modes are detected (as shown in the introduction) and their physical interpretation as stochastic modes has been firmly established.
In the top panel of Fig. 5, we give the theoretical lifetimes obtained for the , and modes of model A. According to Eq. (1), the mode lifetime . Moreover, we see from Eq. (7) that the heights in the PS are proportional to for resolved modes. This leads us to consider in the second panel the mode lifetimes divided by the inertia (). As we showed in Fig. 3, the mode trapping leads to an oscillatory behaviour of the inertia with the large frequency separation as periodicity. The oscillatory behaviour of the lifetimes is a direct consequence of the oscillations of . But when the ratio between the two is considered, the oscillatory behaviour disappears and the same results are found for radial and nonradial modes.
In order to interpret this result, we give in Fig. 6 the cumulated work integral for a typical mode trapped in the core (Hz). The surface value is . We see that significant driving and damping only occur in the upper part of the convective envelope, but there is no significant radiative damping in the core. As the radial component of the displacement dominates in the envelope, the numerator of Eq. (1) is the same at a given frequency, whatever the value of . This explains why does not depend on .
In the third panel of Fig. 5, we give the theoretical amplitudes of the modes. The amplitudes of radial modes are higher because they have less inertia. The oscillatory behaviour of the inertia due to mode trapping leads to oscillations of the amplitudes of nonradial modes.
In the bottom panel of Fig. 5, we give the theoretical heights in the PS. We put the resolution limit at days (horizontal line in top panel). If the mode lifetime value is larger than this value, Eq. (8) is used instead of Eq. (7) to obtain the height. For resolved modes (), the same heights are found for radial and nonradial modes, except for the visibility factor (integration over the stellar disk), which is not included in our study. The unresolved nonradial modes with long lifetimes (mainly modes trapped in the core) have smaller heights than the closest radial mode. We emphasize that the predictions for the heights are different from the amplitudes. Although nonradial modes have smaller amplitudes than radial modes, the heights in the PS are often of the same order. Contrary to the amplitudes, the heights are observation dependent. If is smaller, more nonradial modes are unresolved and their heights in the PS are smaller.
Thus, many nonradial modes would be detectable in the PS of stars like model A. Moreover, the interaction between the p and gmode cavities leads to many avoided crossings. Hence, the resulting frequency spectrum is very complex, particularly for modes and its seismic interpretation is not easy.
4.3 Model B: intermediate in the red giant branch
We now present the results obtained for model B that is slightly higher in the red giant branch (Fig. 1, Table 1). In the top panel of Fig. 7, we give the theoretical lifetimes obtained for the , and modes. Again we see the oscillatory behaviour due to mode trapping. But considering now the ratio between the lifetime and the inertia given in the 2nd panel from the top, we find very different results compared to model A. Now, oscillatory behaviour of this ratio is present for nonradial modes, particularly at low frequencies. To understand this, we compare in Fig. 8 the work integral for two modes, one trapped in the envelope (Hz, top panel) and the other trapped in the core (Hz, bottom panel). For the mode trapped in the envelope, the driving and damping occur in the upper part of the convective envelope, as in model A. But for the mode trapped in the core, significant radiative damping occurs around the bottom of the Hburning shell ().
In the third panel of Fig. 7, we give the theoretical amplitudes of radial and nonradial modes of model B. The amplitudes of nonradial modes trapped in the core are very small because of large inertia and radiative damping. The amplitudes of nonradial modes trapped in the envelope (local maxima in the figure) are smaller than radial modes but not negligible because of similar inertia and negligible radiative damping.
In the bottom panel of Fig. 7, we give the heights of the modes in the PS. Nonradial modes trapped in the core have negligible heights and would not be detected. In contrast, nonradial modes trapped in the envelope have similar heights compared to radial modes.
Thus, only radial modes and nonradial modes trapped in the envelope could be observed in the PS of stars like model B. The groups of nonradial modes trapped in the envelope are more or less separated from each other by a constant large separation. Seismic interpretation of the frequency spectrum of such stars would thus be easier than in model A. We notice however that the trapping is not perfect for modes. Hence, a small group of modes is detectable around each local maximum. This would have to be taken into account in any seismic study. Considering observed power spectra, this also means that these groups of modes would have to be not confused with single large linewidths modes.
4.4 Model C: high in the red giant branch
We consider now the results obtained for model C, which is even higher in the red giant branch than the previous ones (Fig. 1). In the top panel of Fig. 9, we give the theoretical lifetimes. They are very different from those of model B. The oscillatory behaviour is no longer present for most and low frequency modes. As can be seen in the figure, the transitions to envelopetrapped modes are very sharp; these modes, known as Strongly Trapped in the Envelope (STE), are nearly perfectly reflected at the internal turning point (near the top of the Hburning shell). We give in Fig. 10 the work integral for an mode with frequency very close to but different from an STE mode. The very high radiative damping illustrated in this figure occurs for all nonradial modes of this model, except the STE mode. Because of this and the large inertia, the gmode cavity dominates the integrals in the numerator and the denominator of Eq. (1) for all nonSTE modes. Hence Eq. (3) and Eq. (4) are good approximations for the full integrals and the constant simplifies when the ratio between the two is considered. This explains why the lifetimes do not show oscillations in the top panel. In the second panel of Fig. 9, we give the theoretical amplitudes. They are very small for all nonradial nonSTE modes, because of the high radiative damping and inertia. Finally, we give in the bottom panel the theoretical heights in the PS. All nonradial nonSTE modes have negligible heights because of the large radiative damping. Thus, only radial modes and nonradial modes completely trapped in the envelope could be detected in stars like model C.
4.5 Models D and E: Hshell versus Hecore burning
We now compare the theoretical lifetimes, amplitudes and heights of the two models D and E with the same luminosity, one before (model D) and the other during (model E) the core helium burning phase.
In Fig. 11, we give the theoretical lifetimes obtained for the modes of these two models. The behaviour is similar for the two models.
In Fig. 12, we give the theoretical heights in the PS for these two models. No significant difference between the two is found except that the heights in the PS are systematically larger for the He burning model, because of its higher effective temperature leading to stronger stochastic driving.
4.6 Uncertainties of the nonadiabatic treatment in the convective envelope
As discussed in Sects. 2.2 and 3.2, the timedependent interaction between convection and oscillations is subject to large uncertainties. In the TDC treatment of G05, a free complex parameter was introduced in the perturbed closure equations. In the previous computations, we adopted . We now redo the computations with . Larger changes of would give unstable modes in the considered range of frequencies, which is not realistic in view of the typical observed amplitudes of red giant high order pmodes.
With , the convective damping in the upper part of the convective envelope is larger. As shown in Fig. 13, this results in smaller lifetimes and amplitudes of the modes trapped in the envelope compared to Fig. 7. In the upper part of the envelope, the eigenfunctions do not depend significantly on the degree of the modes. Hence the uncertainties associated with the TDC treatment affect radial and nonradial modes trapped in the envelope in the same way.
4.7 Nonlinear effects?
Our study is linear, raising the question of whether such an approximation is valid here. Considering first the amplitudes of the displacement at the photosphere, we find typical values about times smaller than the pressure scale height for detectable modes. Hence nonlinear effects in the superficial layers are expected to be small. We consider now the oscillations in the gmode cavity. We can confidently neglect nonlinear terms if . In a cavity with short wavelength oscillations, this condition is equivalent to . In the gmode cavity of envelope trapped modes, is typically about times smaller than . Hence the risk of wavebreaking due to nonlinear effects is negligible in the core.
5 Conclusions
We have determined the theoretical lifetimes, amplitudes and heights in the PS of radial and nonradial modes of several red giant models. The predictions appear to be very different depending on the evolutionary status of the star. What mainly matters is the density contrast between the He core and the envelope. During evolution, it increases considerably as the He core contracts and the envelope expands. Moreover the KelvinHelmholtz time decreases because of the increasing luminosity. The consequence of these two effects is an increasing importance of the radiative damping of nonradial modes as the luminosity of the star increases. We can encounter very different cases depending on the evolutionary status of the star:

Case A: Subgiants and/or low luminosity red giants.
Our model representative of the class is model A at the bottom of the giant branch for . In these stars, the density contrast is not large enough to imply significant radiative damping of nonradial solartype pmodes ( in model A). The amplitudes of nonradial modes are smaller than radial modes because of larger inertia. However, if we consider instead the heights in the PS, they are similar for radial and nonradial modes as long as they can be resolved. As the heights are used as detection criteria in a Fourier analysis, many nonradial modes could be detected in the PS of such stars. The interaction between the p and gmode cavities leads to many avoided crossings. The resulting frequency spectrum is thus very complex. Asteroseismology of such stars would be challenging. 
Case B: Intermediate models in the red giant branch.
Our models representative of this class are model B, D and E with greater luminosity compared to the previous case (). In these stars the radius ratio between the core and the envelope is smaller ( in model B). As a consequence, the radiative damping around the bottom of the Hburning shell becomes large enough to destroy the nonradial modes trapped in the core. Only the nonradial modes trapped in the envelope and the radial modes can survive and have similar heights in the PS. The nonradial modes trapped in the envelope are more or less separated from each other by a constant large separation, allowing us to build echelle diagrams. However, the trapping is not perfect for modes. Hence, a small group of modes is detectable around each local maximum in the PS corresponding to an envelope trapped mode. Such structures would not have to be interpreted as single large linewidth modes. Seismic interpretation of the frequency spectrum of such stars would be easier than in the previous case. 
Case C: High luminosity red giants.
Our model representative of this class is model C with , . This case is not very different from the previous one, except that the density contrast between the core and the envelope is very large ( in model C). As a consequence, the radiative damping is stronger in the core and destroys all nonradial modes except those strongly trapped in the envelope. These modes and the radial modes are the only ones that could be detected in these stars. The trapping is efficient enough in modes and only single envelope trapped modes are detectable. Seismic interpretation of the frequency spectrum of such stars would also be easier.
We also notice that the duration of observation is very important, as it determines the frontier between resolved and unresolved modes. Nonradial modes have always much longer lifetimes () than radial modes because of their greater inertia. If , the modes cannot be resolved and have much smaller heights in the PS. With too short observations, as is typically the case with ground based observations, only radial and strongly trapped nonradial modes have a chance of being detected. On the other hand, future space missions will provide longer observations: e.g. 5 years for Kepler, 3 years for Plato, increasing the ability to detect more nonradial modes.
Asteroseismology of red giants is a very promising field as the theoretical predictions depend strongly on the physical characteristics of both their deep and superficial layers.
Acknowledgements.
The works of K. Belkacem were financially supported throught a postdoctoral fellowship from the “Subside fédéral pour la recherche 2009”, University of Liège. We thank Prof. W. Dziembowski for his useful advices and remarks.References
 Alexander, D. R. & Ferguson, J. W. 1994, ApJ, 437, 879
 Angulo, C., Arnould, M., Rayet, M., et al. 1999, Nuclear Physics A, 656, 3
 Balmforth, N. J. 1992, MNRAS, 255, 603
 Barban, C., Matthews, J. M., de Ridder, J., et al. 2007, A&A, 468, 1033
 Baudin, F., Samadi, R., Goupil, M.J., et al. 2005, A&A, 433, 349
 Bedding, T. & Kjeldsen, H. 2006, in ESA Special Publication, Vol. 624, Proceedings of SOHO 18/GONG 2006/HELAS I, Beyond the spherical Sun
 Belkacem, K., Samadi, R., Goupil, M.J., & Dupret, M.A. 2008, A&A, 478, 163
 Belkacem, K., Samadi, R., Goupil, M. J., Kupka, F., & Baudin, F. 2006, A&A, 460, 183
 Berthomieu, G., Toutain, T., Gonczi, G., et al. 2001, in ESA Special Publication, Vol. 464, SOHO 10/GONG 2000 Workshop: Helio and Asteroseismology at the Dawn of the Millennium, ed. A. Wilson & P. L. Pallé, 411–414
 BöhmVitense, E. 1958, Zeitschrift fur Astrophysik, 46, 108
 Chaplin, W. J., Houdek, G., Elsworth, Y., et al. 2005, MNRAS, 360, 859
 ChristensenDalsgaard, J. 2004, Sol. Phys., 220, 137
 De Ridder, J., Barban, C., Carrier, F., et al. 2006, A&A, 448, 689
 De Ridder et al. 2009, Nature, in press
 Dupret, M.A. 2002, Bull. Soc. Roy. Sc. Liège, 5, 249
 Dupret, M. A., Barban, C., Goupil, M.J., et al. 2006a, in ESA Special Publication, Vol. 624, Proceedings of SOHO 18/GONG 2006/HELAS I, Beyond the spherical Sun
 Dupret, M.A., Grigahcène, A., Garrido, R., Gabriel, M., & Scuflaire, R. 2004, A&A, 414, L17
 Dupret, M.A., Samadi, R., Grigahcene, A., Goupil, M.J., & Gabriel, M. 2006b, Communications in Asteroseismology, 147, 85
 Dziembowski, W. 1977a, Acta Astronomica, 27, 203
 Dziembowski, W. 1977b, Acta Astronomica, 27, 95
 Dziembowski, W. A., Gough, D. O., Houdek, G., & Sienkiewicz, R. 2001, MNRAS, 328, 601
 Frandsen, S., Carrier, F., Aerts, C., et al. 2002, A&A, 394, L5
 Freytag, B., Steffen, M., & Dorch, B. 2002, Astronomische Nachrichten, 323, 213
 Gabriel, M. 1996, Bulletin of the Astronomical Society of India, 24, 233
 Godart, M., Noels, A., Dupret, M.A., & Lebreton, Y. 2009, MNRAS, in press
 Gough, D. O. 1977a, in Lecture Notes in Physics, Vol. 71, Problems of Stellar Convection, ed. E. A. Spiegel & J.P. Zahn (Berlin: Springer Verlag), 15–56
 Gough, D. O. 1977b, ApJ, 214, 196
 Grevesse, N. & Noels, A. 1993, in La formation des éléments chimiques, AVCP, ed. R. D. Hauck B., Paltani S., 205–257
 Grigahcène, A., Dupret, M.A., Gabriel, M., Garrido, R., & Scuflaire, R. 2005, A&A, 434, 1055
 Hekker, S., Aerts, C., de Ridder, J., & Carrier, F. 2006, A&A, 458, 931
 Henyey, L., Vardya, M. S., & Bodenheimer, P. 1965, ApJ, 142, 841
 Houdek, G. 2000, in Astronomical Society of the Pacific Conference Series, Vol. 210, Delta Scuti and Related Stars, ed. M. Breger & M. Montgomery, 454
 Houdek, G. & Gough, D. O. 2002, MNRAS, 336, L65
 Iglesias, C. A. & Rogers, F. J. 1996, ApJ, 464, 943
 Kallinger, T., Guenther, D. B., Matthews, J. M., et al. 2008a, A&A, 478, 497
 Kallinger, T., Guenther, D. B., Weiss, W. W., et al. 2008b, Communications in Asteroseismology, 153, 84
 Lochard, J., Samadi, R., & Goupil, M. J. 2005, A&A, 438, 939
 Mazzitelli, I. 1979, A&A, 79, 251
 Musielak, Z. E., Rosner, R., Stein, R. F., & Ulmschneider, P. 1994, ApJ, 423, 474
 Potekhin, A. Y., Baiko, D. A., Haensel, P., & Yakovlev, D. G. 1999, A&A, 346, 345
 Rogers, F. J. & Nayfonov, A. 2002, ApJ, 576, 1064
 Samadi, R., Belkacem, K., Goupil, M. J., Dupret, M.A., & Kupka, F. 2008, A&A, 489, 291
 Samadi, R. & Goupil, M. . 2001, A&A, 370, 136
 Samadi, R., Nordlund, Å., Stein, R. F., Goupil, M. J., & Roxburgh, I. 2003a, A&A, 404, 1129
 Samadi, R., Nordlund, Å., Stein, R. F., Goupil, M. J., & Roxburgh, I. 2003b, A&A, 403, 303
 Saumon, D., Chabrier, G., & van Horn, H. M. 1995, ApJS, 99, 713
 Spiegel, E. A. 1963, ApJ, 138, 216
 Stello, D., Kjeldsen, H., Bedding, T. R., & Buzasi, D. 2006, A&A, 448, 709
 Stolzmann, W. & Bloecker, T. 1996, A&A, 314, 1024
 Unno, W. 1967, PASJ, 19, 140
 Van Hoolst, T., Dziembowski, W. A., & Kawaler, S. D. 1998, MNRAS, 297, 536
 Ventura, P., D’Antona, F., & Mazzitelli, I. 2008, Ap&SS, 316, 93
 Ventura, P., Zeppieri, A., Mazzitelli, I., & D’Antona, F. 1998, A&A, 334, 953
 Wedemeyer, S., Freytag, B., Steffen, M., Ludwig, H.G., & Holweger, H. 2004, A&A, 414, 1121
 Xiong, D. R., Cheng, Q. L., & Deng, L. 1997, ApJS, 108, 529
 Xiong, D. R., Cheng, Q. L., & Deng, L. 2000, MNRAS, 319, 1079
 Xiong, D. R. & Deng, L. 2001, MNRAS, 324, 243
 Xiong, D. R. & Deng, L. 2007, MNRAS, 378, 1270