Weighing massive neutron star with screening gravity: A look on PSR J0740+6620 and GW190814 secondary component

Rafael C. Nunes,1, ∗ Jaziel G. Coelho,1, 2, † and Jos´eC. N. de Araujo1, ‡ 1Divis˜aode Astrof´ısica, Instituto Nacional de Pesquisas Espaciais, Avenida dos Astronautas 1758, S˜aoJos´edos Campos, 12227-010, SP, Brazil 2Departamento de F´ısica, Universidade Tecnol´ogica Federal do Paran´a,Medianeira, 85884-000, PR, Brazil Neutron stars (NSs) are excellent natural laboratories to constrain gravity on strong field regime and nuclear matter in extreme conditions. Motivated by the recent discovery of a compact object +0.08 with 2.59−0.09M in the binary merger GW190814, if this object was a NS, it serves as a strong con- straint on the NS equation of state (EoS), ruling out several soft EoSs favored by GW170817 event. In this work, we revisit the question of the maximum mass of NSs considering a chameleon screening (thin-shell effect) on the NS mass-radius relation, where the microscopic physics inside the NS is given by realistic soft EoSs. We find that from appropriate and reasonable combination of modified gravity, rotation effects and realistic soft EoSs, that it is possible to achieve high masses and explain GW190814 secondary component, and in return also NSs like PSR J0740+6620 (the most NS mas- sive confirmed to date). It is shown that gravity can play an important role in estimating maximum mass of NSs, and even with soft EoSs, it is possible to generate very high masses. Therefore, in this competition of hydrostatic equilibrium between gravity and pressure (from EoS choice), some soft EoSs, in principle, cannot be completely ruled out without first taking into account gravitational effects.

I. INTRODUCTION bound on the dimensionless spin for the secondary com- panion, using the upper maximum mass constraints from Recently, the LIGO/Virgo scientific collaboration an- the GW170817 event [7, 8], and show that the allowed nounced the discovery of a compact binary coalescence, range in dimensionless spins correspond to rotational fre- quencies much higher than the fastest millisecond pulsars GW190814, involving a 22.2–24.3M black hole (BH) known [9]. In [10] a rigorous upper bounds on maxi- and a compact object with a mass of 2.50–2.67M [1]. The secondary component is either the most massive mum mass under the exclusive assumptions of causality neutron star (NS), or the lightest BH ever discovered and general relativity (GR), showing that the presence in a double compact-object system. The existence of of a NS in GW190814 is not inconsistent with present ultra-massive NSs has been revealed in several studies. observational constraints on the NS EoS. On the other So far, the precise determination of the mass of pulsars hand, in [11] is showed that the stiffening of the EoS re- quired to support ultra-massive NSs is inconsistent with leads to M = 1.97 ± 0.04 M for PSR J1614–2230 [2], PSR J0348+0432 has M = 2.01 ± 0.04 M [3] and the either constraints obtained from the low deformability highly likely the most massive NS observed to date, PSR of medium-mass stars demanded by GW170817 or from energetic heavy-ion collisions. Several other methodol- J0740+6620, with M = 2.14 ± 0.1 M [4]. The discovery of the gravitational wave (GW) binary ogy and speculations about GW190814 are presented in GW190814 triggered intense theoretical efforts about the [12–28]. real nature of its secondary component, in particular, As is well known, the two main observable of a NS, the high mass requires the compact star matter to be i.e., their mass and radius, both depend crucially on the described by a stiff equation of state (EoS). In [5] is ar- choice of EoS and the gravitational theory, where grav- gued that fast rotation is capable to explain the exis- ity will determine the macroscopic hydrostatic equilib- tence of a stable ∼2.6M NS for moderately stiff EoS rium. The general relativity is very well tested and com- but may not be adequate for soft EoSs. In particular, patible in weak-field observations, for instance, from so-

arXiv:2008.10395v2 [astro-ph.HE] 4 Dec 2020 several soft EoSs favored by GW170817 and with max- lar system and terrestrial experiment tests [29–31], but imum spherical masses of ∼2.1M cannot be used to there are theoretical and observational reasons to believe explain this source as a uniformly spinning NS. In [6] that GR should be modified when gravitational fields are the authors infer a lower limit on the maximum mass strong and/or on large scales. On large scales and from Mmax of non-rotating NSs, using arguments based on uni- an observational point of view, the physical mechanism versal relations connecting the masses and spins of uni- responsible for accelerating the Universe at late times is formly rotating NSs. Furthermore, they obtain a lower still an open question, and new degrees of freedom of the gravitational origin are alternatives to explain such an accelerated stage (see, e.g., [32, 33] for review). Theo- ries beyond GR can serve as alternatives to explain the ∗ [email protected] current tension in the Hubble constant that persists in † [email protected] the framework of the standard cosmological model [34– ‡ [email protected] 36]. Also, modified gravity models are motivated to drive 2 the accelerating expansion of the Universe at early times, II. THEORETICAL FRAMEWORK AND known as inflationary era. On astrophysical scales, com- METHODOLOGY pact objects such as BHs and NSs are our best natural laboratories to constrain strong gravity. In these bodies, In this section we review in a nutshell the theoretical gravity prevails over all other interactions and collapse methodology used to obtain our main results. leads to large-curvature and strong-gravity environments [37]. We refer the reader to [38] and references therein for several modified gravity scenarios motivation under A. Screening Mechanisms the regime of strong gravitational field. The study of scalar-tensor theories has been motivated by some cosmological observation, especially in order A consistent way to modify gravity is to introduce to explain the current accelerating expansion stage of some screening mechanism in order to have viable grav- the Universe. One proposed explanation is that grav- ity, i.e., GR, at small distances and scales and possible ity is modified on large scales, but must be suppressed relevant effects on others environmental scales, for in- on small scales, for instance, in the solar system devia- stance, on compact objects or even on cosmological level. tions are constrained to be subdominant by a factor of Recall that the screening mechanisms include chameleons 10−5. Screening gravity circumvent this problem by in- [39, 40], symmetrons [41], dilatons [42], Vainshtein mech- troducing non-linear modifications of the Poisson equa- anism [43], etc. At the heart of screening mechanisms lies tion that dynamically suppress deviations from GR in the fact that there are 29 orders-of-magnitude separating the solar system without the need to fine-tune on the the cosmological and terrestrial densities and 20 orders of scalar mass or the coupling to matter. In short, screen- magnitude separating their distance scales. As a result, ing mechanisms utilize non-linear dynamics to effectively the properties of the new degrees of freedom of the gravi- decouple solar system and cosmological scales. At the tational origin (a scalar field) can vary wildly in different heart of screening mechanisms lies the fact that there environments [44]. Screening gravity have been inten- are 29 orders-of-magnitude separating the cosmological sively investigated in the literature, in the most diverse and terrestrial densities. As a result, the properties of aspects in cosmology and gravitation (see, e.g., [45–58] the scalar field (new degree of freedom of gravitational for a short list). A compilation/list of various works on origin) can vary widely in different environments. compact objects in modified theories of gravity can be In this work, we will consider the well-studied found in [59]. chameleon screening [39, 40]. In chameleon models, the mass of the scalar is an increasing function of the am- bient density. The purpose of this section is review the As regards to NS observations, it seems there is a the- calculation of the parameterized post-Newtonian (PPN) oretical degeneration and is not clear if measurement parameter γ that is relevant for chameleon theories. We of mass and radius constrain possible gravity effects or follow the methodology presented in [62] and references EoSs. As showed in [60], measurement of mass constrains therein. gravity rather than the EoS [see also [61]]. Thus, it is Let us consider a specific subset of the general scalar- difficult to distinguish which of these effects is actually tensor theories, which in the Einstein frame is given by contributing to the actual observed mass and radius val- " # Z √ M 2 ues. In this work, we analyze the NS mass-radius relation 4 pl 1 µ S = dx −g R − ∇µφ∇ φ − V (φ) + Sm[˜g], through a modification of the TOV equations induced by 2 2 the presence of a possible chameleon screening (thin-shell (1) effect), where the microscopic physics inside the NS will where the Jordan frame metricg ˜µν is a Weyl rescaling be modeled by realistic soft EoSs of which it is not possi- of the Einstein frame metric gµν by a conformal factor 2 ble to generate very massive NS with ∼ 2.6 M or even A(φ), i.e.,g ˜µν = A (φ)gµν . The coupling is given by ∼ 2.14 M in GR context. Our main aim is to show that an appropriate combination of modified gravity, rotation d ln A(φ) effects and realistic soft EoSs, can have a joint effect for α(φ) = Mpl . (2) achieve high masses and explain GW190814 secondary dφ component, and in return also NS like PSR J0740+6620. Each particular scenario is set by the choice of A(φ) and V (φ) functions. The PPN metric for a single body, which we will refer to as body A with mass MA, reads This paper is organized as follows. In the next section, we present our theoretical framework and methodology PPN G MA to generate the NS mass-radius relation. In Section III g˜00 = −1 + 2 (3) r we present our main results and in Section IV our final  PPN  G MA comments and perspectives. (The speed of light c is set g˜ij = 1 + 2˜γ δij, (4) equal to unity). r 3

We will refer to the quantity GPPN as the PPN gravi- In a binary system, if we consider the orbital dynam- PPN tational constant. G controls the size of effects com- ics of a body of mass MB orbiting the body sourcing puted using the PPN metric. It is distinct from the grav- this metric, this second body may have its own screening B itational constant G that appears in GR and Newtonian radius rs , so that the chameleon force between the two gravity. We are interested in the regime where some body bodies A and B reads [63] has some degree of screening. In this case, the equation inside of some screening radius is GMAMB  2  2 F = − 1 + 2α Q Q (13) ∇ φ = 0 (5) r2 A B while outside the screening radius reads with the following modified potential 2 ∇ φ = 8πα(φ)GρA r > rs, (6) GM   M(r ) where α is the coupling function. V (r) = 1 + 2α2 1 − s , (14) In order to move on, we need to define α. Here, we r M assume the chameleon field which uses a non-linear po- tential to make the field mass a function of the environ- where we have once again ignored the mass of the scalar. mental density. The equation of motion reads Thus, the physical quantity that is measurable in these theories is 4+n 2 Λ αρA ∇ φ = −n n+1 + , (7) PPN  2  φ Mpl G = G 1 + 2α QAQB , (15) where the effective potential is given by with corresponding PPN parameter γ given by 4+n Λ αφρA Veff = n + . (8) φ Mpl  2 −1 γ = 2 1 + 2α QAQB − 1. (16) The mass-scale Λ can vary over many orders of magni- tude, but it is often compared to the dark energy scale, Notice that other approaches have been developed in since this value is relevant for the present-day cosmic ac- the literature, see for instance [44, 54, 55]. In what fol- celeration. Astrophysically, the chameleon profile of a lows, let us quantify how this approximation presented spherically-symmetric object of mass M and radius R is above can modify spherical compact stars. not sourced by the object’s mass, but rather by the mass inside a shell near the surface, a phenomenon that has been dubbed the thin-shell effect. The reason for this B. Modified TOV is the following: deep inside the object, the field mini- mizes its effective potential corresponding to the ambi- In GR, the hydrostatic equilibrium of a star is ent density but, as one moves away from the center, the described by the Tolman-Oppenheimer-Volkov (TOV) field must eventually evolve asymptotically to the mini- equations (see, e.g., [64]), namely mum at the density of the medium in which the object is immersed (astrophysical or cosmological densities, de- dP GM(r)ρ(r) P (r) 4πr3P (r) 2GM(r)−1 pending on the situation to be investigated). The field = − 2 1+ 1+ 1− can only roll once the density is low enough so that its dr r ρ(r) M(r) r (17) effective mass is light enough. The radius at which this and happens is typically called the screening radius rs. The coupling function α above is a constant for dM = 4πr2ρ(r), (18) chameleons [44, 62]. Ignoring possible scalar mass con- dr tribution, the solution is then where ρ(r) is the energy density, P (r) is the pressure and M(r) is the mass within the radial coordinate r. 2 QAGMA φ = φ0 − 2α , (9) The mass and the radius of a star are the two more r obvious observables. Evidently, these quantities depends where Q is the scalar charge of body A and is given by on the theory of gravitation and microscopic physics, i.e,  M (rA) the EoS of the matter of the star. In view of the for- Q = 1 − A s . (10) A M malism described in the previous section, we can notice A that the chameleon screening can be quantified in the hy- Transforming to the Jordan frame and expanding A(φ) drostatic equilibrium rescaling G by GPPN in Eqs. (17) to first order in GMA/r, one finds Eqs. (3) and (4) with and (18). GPPN is also directly connected with the PPN PPN  2  parameter γ, since α2Q Q can be estimated. G = G 1 + 2α QA and (11) A B 2 In this work, we will restrict to investigating NSs. 1 − 2α QA Thus, we need to assume a EoS that describes the micro- γ˜ = 2 . (12) 1 + 2α QA scopic physics of the matter of the NS to complete the 4 system of equations above. Let us assume a realistic mod- eling of this physics, following the methodology described in [65], where the nuclear matter inside NS are built by joining together different polytropic phases on a sequence 2.8 of different density intervals (piecewise EoS), given by 2.7 ρ > ρ , if for a set of dividing densities ρ < ρ < ρ ..., 0 0 1 2 2.6 the pressure and energy density are everywhere continu- ] 2.5 ous and satisfy M [

M 2.4    1 P (ρ) = K ρΓi , d = −P d , ρ < ρ < ρ , 2.3 i ρ ρ i−1 i (19) 2.2 where  is the energy density and fixed by the first law of 2.1 thermodynamics. The numerical modeling of EoSs is well 0.80 0.85 0.90 0.95 1.00 1.05 1.10 1.15 1.20 , G summarized in [65] (see also [66]). In this work, we will PPN assume the low-density EoSs given by SLy, MS2, GNH3, ALF2, BBB2, H4, PS, WFF3, AP2 and FPS models. FIG. 1. Illustrative example of a neutron star maximum mass In total, we have 10 soft EoSs in our sample. In addi- as a function of PPN parameter γ (blue) and GPPN (orange) tion to several key properties of each ones, it should be assuming EoS SLy. stressed that these EoSs are consistent with constraints from GW170817 (see [67] and references therein). The best fit parametric configuration that describe these EoSs modeling are summarized in Table III in [65]. surface of the star, so the approximation M(rs)i = βiMi On the other hand, if the NS is spinning, its max- serves as a basis to quantify these screening effects within a simple approach, and to investigate possible new effects imum mass can be higher than MTOV due to the addi- tional rotational support against gravitational collapse to on the M-R relation. In what follows, we summarize our a BH. In particular, [68] computes the maximum mass al- main results. lowed by uniform rotation, Mmax, purely in terms of the maximum mass of the non-rotating configuration. The importance of this universal relation, connecting the di- III. RESULTS mensionless spin on the stability line and the maximum dimensionless spin at the mass shedding limit, is that Table I summarizes the maximum mass of a NS tak- it allows us to calculate Mmax sustainable through fast ing into account 4 possible theoretical configurations and uniform rotation, finding that for any EoS it is about 10 input realistic soft EoSs, from which we can quantify 20% larger than the maximum mass supported by the how each ingredient is contributing to Mmax estimates. corresponding non-rotating configuration (see [6, 68] for Taking into account the screening mechanisms, we con- details). Thus, we complement our modeling considering sider a NS with screening radius rs and MA, orbiting a Mmax = ξMTOV, with ξ = 1.203 ± 0.022 [6]. There- generic body B with MB, so given the relation previously PPN fore, our full NS model will be constituted from some mentioned, M(rs)i = βiMi, we have G /G = 1 +α ¯, 2 soft EoSs, rotation effects and chameleon screening. Note withα ¯ = 2α (1 − βB − βA + βBβA), which is a global that this constraints on ξ is obtained by assuming GR. constant in our case. In all estimates in Table I, from the In Section III A, we relax this condition and obtain ξ screening and screening + spinning cases, we assume a within the theoretical framework here investigated. We small and conservative total correction withα ¯ = −0.05. +0.080 will find ξ = 1.14−0.082 in the modified gravity we investi- As we will see below (subsection A), through a statisti- gated. Thus, these values are fully compatible with each cal fit, effective corrections on GPPN are of this order of other and this choice will not biased the main results to magnitude. i.e.,α ¯ ∼ -0.05 at ∼1σ. Correctionsα ¯ > 0 be presented below. tends to decrease the expectation value of maximum For the convenience of analysis, let us consider that mass. Figure 1 shows this using SLy EoS + spinning there is a simple relationship between the expectation effects, quantifying how much γ and GPPN corrections mass inside the screening radius in the form M(rs)i = can influence Mmax prediction. Important to note that PPN βiMi, where i runs over the bodies A and B, in case of the expected corrections for γ and G within screening a binary system. In principle, we could consider more gravity are not the same as measured on earth or solar complicated dependency on the screening mass, for in- system. Without screening mechanisms, we would have n stance, like M(rs)i = βiri Mi, where r is the NS radius to tune γ value to satisfy terrestrial and solar system and n some power law dependency. This type of correc- experiment bounds, but with screening mechanisms this tion will take PPN corrections predict more complicated bound can be automatically satisfied for this screened relation/dependence on NS mass values. On the other region with M(rs) ' M(r) on these scales. Here, we hand, we expect that the screening radius be close to the are relaxing this condition in order to have new γ and 5

EOS Mmax (GR) Mmax (GR with rotation) Mmax (chameleon screening) Mmax (chameleon screening with rotation)

SLy 2.04 M 2.44 M 2.19M 2.63M

MS2 1.80 M 2.16 M 1.84M 2.21M

ALF4 1.94 M 2.32 M 2.06M 2.48M

GNH3 1.96 M 2.35 M 2.03M 2.44M

BBB2 1.91 M 2.29 M 2.07M 2.49M

H4 2.03 M 2.43 M 2.09M 2.52M

PS 1.75 M 2.10 M 1.80M 2.10M

WFF3 1.84 M 2.2 M 1.98M 2.37M

AP2 1.80 M 2.16 M 1.92M 2.30M

FPS 1.79 M 2.14 M 1.93M 2.32M

TABLE I. Summary of the maximum mass values that can be obtained using 10 soft EoSs in different theoretical situations, namely, general relativity, general relativity with rapid uniform rotation effects, chameleon screening and chameleon screening with rapid uniform rotation effects.

3.0 3.0

2.5 2.5

2.0 2.0 ] ] 1.5 M M [ [ 1.5

M

M H4 SLy 1.0 PS MS2 1.0 WFF3 ALF4 AP2 GNH3 0.5 FPS BBB2 0.5 GW190814 GW190814 0.0 J0740+6620 J0740+6620 5.0 7.5 10.0 12.5 15.0 17.5 20.0 22.5 8 10 12 14 16 R [km] R [km]

FIG. 2. Left panel: Neutron star mass-radius relation for different soft EoSs (SLy, MS2, ALF4, GNH3, BBB2) within Spinning + PPN corrections framework, with highlighted regions representing the neutron star mass expectation from the binary system J0740+6620 and GW190814. Right panel: Same as in left panel, but from other soft EoSs (H4, PS, WFF3, AP2 and FPS).

GPPN expected values on a NS under thin-shell effect. tem with a white dwarf companion, and the mass of the Of course, any new γ and GPPN expected values should lighter component of the GW190814 event, 2.6 ± 0.10 not deviate significantly from solar-system tests. In all M , if we interpret this object as being a NS. It is worth our calculations, we are assuming a maximum deviation mentioning that in the results shown in Fig. 2 and Table of 5% when applied to NS structure equations. I, we assume an effectiveα ¯ = −0.05 (corresponding to PPN Fig. 2 shows the NS mass-radius relation from all an effective G /G = 0.95). the EOSs analyzed here, and summarized in Table I The consideration above has an observational sup- for spinning + chameleon screening combination, in di- port. For instance, astrophysical tests on chameleon the- rect comparison with the expected masses of the PSR ories using distance indicators in the nearby universe are J0740+6620, 2.14 ± 0.10 M , a NS in a binary sys- bounds to an effective values of ∆G/G ∈ [0.11 − 0.45] 6

that compact object. The analysis frame change the sit- uation if we look at GW190814 secondary component. cubic fit When considering GR + spinning, all these soft EOSs Prediction 5 can not achieve very high masses, and all them should NS mass sample 68% confidence interval be ruled out. This means that the secondary compact

4 object in GW190814 can not be explained by an EoS like SLy (EoS favored by GW170817). Also, as argued in [5],

x such EoSs like SLy must now be rejected because their a G 3 M m mass-shedding limit is below the lower limit mass of this M object in GW190814. But, we find that modifying the 2 hydrostatic equilibrium equations of NSs, incorporating a possible thin-shell effect (in our study from chameleon 1 screening), we can reach larger masses. We note that chameleon screening can achieve high masses and explain GW190814 with soft EoSs like SLy, BB2, ALF4 and H4 2.10 2.15 2.20 2.25 2.30 2.35 2.40 (see Fig. 2), which in GR context is not possible. The MGR max BBB2, ALF4 and H4 EoSs predictions live in the lower limit of the error bar of GW190814, while the SLy EoS MG GR FIG. 3. Statistical reconstruction in the plan Mmax − Mmax easily archive higher values and can explain GW190814 using Gaussian processes for machine learning between the best fit and more massive objects. Therefore, in presence maximum mass estimates summarized in Table I. The solid of new degrees of freedom of gravitational origin like a blue curve is the Gaussian process mean. Both Mmax are in chameleon screening, it is possible to maintain these soft units of M . EoSs to describe the internal structure of these compact objects. [69]. The observation of GW170817 event imposed strong Note that the same EoS in screening + spinning con- constraints on modified gravity/dark energy scenarios, text can explain both NSs in J0740+6620 and GW190814 inferring that the speed of GWs propagation is equal to at the same time, changing only the radius prediction. the speed of light, so limiting the theoretical parametric For instance, from EoS SLy analysis, one obtains [9.16 space for various theories. Chameleon theories in princi- - 11.27]km and [10.86 - 11.18]km for GW190814 and ple has already this physical characteristic and this con- J0740+6620, respectively. On the other hand, it is inter- straints does not affect in general any aspect in these esting to derive a general relation between the maximum theories. The authors in [70] calculate that the effects NS mass from GR and the modified gravity theory from some EoS sample. We will obtain this relationship based of a plausible variation in GN on the period decay of a local binary system are many orders of magnitude sup- on our 10 soft EoSs sample [65], where the main results are summarized in Table I. pressed with respect to the effects of a change in GN . As −3 a result, G˙ N /GN is constrained to ∼ 10 , while GN is To get this relation, let us use a Gaussian Processes unconstrained by observations of the binary orbital decay (GP). The GP consists of generic supervised learning rate. Here, the notation GN is the gravitational coupling method designed to solve regression and probabilistic between two matter sources. Note that our approach is classification problems, where we can interpolate the ob- linked to the physical variable GN and not G˙ N . In gen- servations and compute empirical confidence intervals eral, all our considerations about GPPN/G, i. e., all range and a prediction in some region of interest [73]. Let us of values assumed to the prior are compatible with the consider that our 10 estimates in Table I (the case with bounds derived from other observations. rotation effects) are data points to be trained between It is important to note that all EoSs assumed here a certain mass range. Note that 10 data points, from share a relatively low maximum mass and thus they are 10 different soft EoSs, give us reasonable information to called soft [65]. Soft EoSs consistent with GW170817 (see get a good prediction within the mass range where soft Table I) are unable to provide enough mass to explain the EoS there due the robustness of this methodology. The secondary in GW190814, even for a NS endowed with GP method is the state-of-the-art to obtain statistical maximum uniform rotation [5]. We noticed that assum- information and model prediction from some previously ing non-rotating mass configuration, it is troublesome to known information or data. predicting high mass in GR, and all these EoSs have dif- Figure 3 show the data reconstruction. The solid blue ficulties in reaching the mass expected by J0740+6620. curve is the GP mean prediction and the shaded areas Thus, in this simplest case, all of these EoSs are ex- is the statistical confidence level (CL) at 68%. Each red cluded. But, adding rotation effects, all these EoSs can point represents the mass information (data point) in the MG GR explain J0740+6620 easily. The authors in [71, 72] were plane Mmax −Mmax, in the case with rotation, which has the first to show that spinning up a NS uniformly can been trained to obtain the reconstruction and prediction increase its mass by up to ∼ 20%. Thus, soft EoSs con- between these mass estimates. During the reconstruc- sistent with rapid uniform rotation is enough to explain tion, we use the most popular covariance functions in the 7

component, if this compact object was a NS, within the screening + rotation framework. For this, we carry out a simple χ2 fit given by

PPN 2 2 (MO − Mmax(G , ξ)) χ = 2 , (21) σO

PPN where Mmax(G , ξ) is the maximum mass expected 20 1. assuming GPPN, ξ as free parameter, where we assume the prior ξ ∈ [1, 1.2]. The range in ξ represent non- 14 1. rotating with ξ = 1 and maximum rotation with ξ = 1.2.

ξ This upper prior on ξ can be interpreted that 20% is 08 1. probably no longer applicable and this correction is an 2 upper limit. The quantities MO, σO, are the mass and 02 the associate error bar at 1σ confidence level (CL) from 1.

96 02 08 02 08 14 20 the GW190814 secondary component observation. 0. 1. 1. 1. 1. 1. 1. PPN Note that in this approach we are obtaining a new ξ G /G ξ MG constraints parametrically through the relation Mmax = MG ξMTOV. Figure 4 shows the parametric space in the plan GPPN/G−ξ at 68% CL and 95% CL, after assuming ξ as FIG. 4. Two-dimensional marginalized distributions in the PPN +0.039 parametric space in the plan GPPN/G − ξ at 68% CL and a free parameter. We find G /G = 0.951−0.041 (with +0.080 95% CL. the corresponding γ = 0.90 ± 0.093) and ξ = 1.14−0.082. During the analysis we assume SLy as input EoS. Note that this value is statistically compatible with the value inferred in GR. Thus, we conclude that the values and literature, the standard Gaussian Squared-Exponential. analysis derived in the previous section are not biased To obtain these results, we make use of some numerical from this perspective. routines available in [74]. Interesting to note that from a On the other hand, other independent measurements, soft EoS sample, in general, we can obtain high mass, like from other observational perspectives, have been car- a 3 solar mass, within modified gravity approach even at +0.19 ried out recently reporting, γ = 0.87−0.17 [75] and 68% CL. |γ − 1| < 0.2(Λ/100)kpc [76], with Λ = 10 − 200kpc. Our In Figure 3, we also show a cubic fit that best fit the constraints are in total agreement with these values. All MG GR plan Mmax − Mmax in the mass range M ∈ [2.10, 2.46] these constraints suggests no deviation from GR. But, M in GR prediction. We find despite the statistical bounds be compatible with GR, note that the gravitational relaxing induced by GPPN is enough to generate NS with M ∼ 2.6 M or more, using MG 2 3 Mmax = −320.38+417.26Mmax−179.96Mmax+25.90Mmax, a soft EoS. In this case, the NS can rotate below of the (20) maximum limit, which in turn will generate an effective PPN where Mmax is the maximum mass in GR. G slightly smaller, but statistically equivalent even Thus, within the theoretical framework investigated at 68% CL when considering the maximum value. From here, Figure 3, summary the most general relations and Eq.(11), if we assume the coupling parameter to be MG GR prediction analysis between Mmax and Mmax within the α ' 1 [44, 62], we find M(rs)/M ' 0.98, showing that soft EoS approach. the screening mass is near to the surface, as expected.

A. Rotation effects in modified gravity IV. FINAL REMARKS The spin of the secondary object is a quantity that has not been constrained by the observations of GW190814. We consider that the gravitational interaction can de- Furthermore, we know that the condition for rapid uni- viate minimally from that predicted from GR, through form rotation contributes significantly to increase the a PPN correction induced by a thin-shell effect via a predicted maximum mass. In the results previously pre- chameleon screening. We analyze its impacts on the NS sented, we assume a rotation value that was obtained in mass-radius relation, motivated to check if it is possi- GR. Once that the gravity theory has been modified, it is ble to generate high NS masses (when compared to GR necessary to re-calculate the rotational effect in the new prediction). Our main conclusion is that from a combina- gravity framework. Let us obtain a new estimate on ξ, in tion of modified gravity, rotation effects and realistic soft the situation we can still explain GW190814 secondary EoSs, it is possible to achieve high masses like M ∼ 2.6 8

M or more, and explain GW190814 secondary compo- possible to obtain new and accurate observational bounds nent. In return, it is also possible to explain the most NS on the free parameters that characterize such scenarios massive confirmed to date, i.e., PSR J0740+6620. Then using possible massive NS observations. one sees here the following interesting aspect: one can not ruled out some soft EoSs without first taking into account effects coming from alternative theories of grav- ACKNOWLEDGMENTS ity. Although our theoretical approach be simple, this consequence is clear (see, e.g., Table I and Figure 3). The authors thank the referee for comments which helped Therefore, gravity can play an important role, without to improve the quality of the manuscript. RCN would the need to invoke very stiff EoSs or unusual aspects in like to thank the agency FAPESP for financial support nuclear matter. under the projects 2018/18036-5 and 2013/26258-4. JGC is likewise grateful to the support of CNPq (421265/2018- It might be interesting to see how more sophisticated 3 and 305369/2018-0). JCNA thanks the partial support gravity theories behave in this perspective, and if it is of CNPq (308367/2019-7).

[1] R. Abbottet et al, GW190814: Gravitational Waves from on the understanding of supranuclear matter, 2007.06057 the Coalescence of a 23 Solar Mass Black Hole with a 2.6 [16] H. Tan, J. N. Hostler, and N. Yunes Kinky neutron stars Solar Mass Compact Object. The Astrophysical Journal in light of GW190814, 2006.16296 896, L44 (2020) 2006.12611 [17] N.B. Zhang and B. A. Li, GW190814’s secondary com- [2] P. Demorest, T. Pennucci, S. Ransom, M. Roberts, J. ponent with mass (2.50-2.67) as a super-fast pulsar, Hessels, A two-solar-mass neutron star measured using 2007.02513 Shapiro delay. Nature 467, 1081–1083 (2010) 1010.5788 [18] Z. Roupas, G. Panotopoulos and I. Lopes, QCD color [3] J. Antoniadis et al A Massive Pulsar in a Compact Rel- superconductivity in compact stars: color-flavor locked ativistic Binary. Science 340, 448 (2013) 1304.6875 quark star candidate for the gravitational-wave signal [4] H. Cromartie et al, Relativistic Shapiro delay measure- GW190814, 2010.11020 ments of an extremely massive millisecond pulsar. Nature [19] A. V. Astashenok, S. Capozziello, S. D. Odintsov and Astronomy 4, 72–76 (2020) 1904.06759 V. K. Oikonomou, Extended Gravity Description for the [5] A. Tsokaros, M. Ruiz and S. L. Shapiro, GW190814: GW190814 Supermassive Neutron Star, 2008.10884 Spin and equation of state of a neutron star companion, [20] A. V. Astashenok and S. D. Odintsov, A. V. Astashenok, (2020). 2007.05526 S. Capozziello and S. D. Odintsov, Extreme neutron stars [6] E. R. Most, L. J. Papenfort, L. R. Weih and L. Rezzolla, from Extended Theories of Gravity, JCAP 1501 (2015) A lower bound on the maximum mass if the secondary 001, arXiv:1408.3856. in GW190814 was once a rapidly spinning neutron star, [21] C. E. Mota, L. C. N. Santos, F. M. da Silva, G. Grams, I. (2020) 2006.14601 P. Lobo and D. P. Menezes, Generalized Rastall’s gravity [7] M. Shibata, E. Zhou, K. Kiuchi, S. Fujibayashi, Con- and its effects on compact objects, 2007.01968 straint on the maximum mass of neutron stars using [22] R. Lobato et al, Neutron stars in f(R,T) gravity using GW170817 event. Physical Review D 100, 023015 (2019) realistic equations of state in the light of massive pulsars 1905.03656 and GW170817, 2009.04696 [8] L. Rezzolla, E. Most, and L. Weih, Using Gravitational- [23] V. I. Danchev and D. D. Doneva, Constraining the equa- wave Observations and Quasi-universal Relations to tion of state in modified gravity via universal relations, Constrain the Maximum Mass of Neutron Stars. The As- 2010.07392 trophysical Journal 852, L25 (2018) 1711.00314 [24] Z. Roupas, Secondary component of gravitational-wave [9] J. Hessels et al, A Radio Pulsar Spinning at 716 Hz. Sci- signal GW190814 as an anisotropic neutron star, ence 311, 1901–1904 (2006) 0601337 2007.10679 [10] A. Godzieba, D. Radice, S. Bernuzzi On the maximum [25] I. Bombaci, A. Drago, D. Logoteta, G. Pagliara and I. mass of neutron stars and GW190814. 2007.10999 Vidana, Was GW190814 a black hole – strange quark [11] F. Fattoyev, J. Horowitz, J. Piekarewicz, and J. Reed, star system?, 2010.01509 GW190814: Impact of a 2.6 solar mass neutron star on [26] V. De Luca, V. Desjacques, G. Franciolini, P. Pani, A. Ri- nucleonic equations of state, 2007.03799 otto, GW190814: Circumstantial Evidence for Up-Down [12] V. Dexheimer, R. Gomes, T. Kl¨ahn, S. Salinas Quark Star, 2009.00942 GW190814 as a massive rapidly-rotating neutron star [27] Z. Cao, L. W. Chen, P. C. Chu and Y. Zhou, The with exotic degrees of freedom. 2007.08493 GW190521 Mass Gap Event and the Primordial Black [13] Y. Lim, A. Bhattacharya, J. Holt, D. Pati, Revisiting Hole Scenario, 2009.01728 constraints on the maximum neutron star mass in light [28] K. Vattis, I. S. Goldstein and S. M. Koushiappas, Could of GW190814 2007.06526 the 2.6M object in GW190814 be a primordial black [14] S.Clesse and J. G. Bellido, GW190425 and GW190814: hole?, 2006.15675 Two candidate mergers of primordial black holes from the [29] C. Will, The Confrontation between General Relativity QCD epoch, 2007.06481 and Experiment, Living Rev. Relativity 17 4 (2014). [15] I. Tews et al, On the nature of GW190814 and its impact 1403.7377 9

[30] J. G. Williams, S. G. Turyshev, D. H. Boggs, Progress in 86 044015, (2012). 2003.12876 lunar laser ranging tests of relativistic gravity, Phys Rev [51] P. Hamilton, M. Jaffe, P. Haslinger, Q. Simmons, H. Lett 93 261101, (2004) . 0411113 M¨uller,J. Khoury, Atom-interferometry constraints on [31] E. G. Adelberger, B. Heckel, S. Hoedl, C. Hoyle, D. Kap- dark energy , Science, 349 (2015). 1502.03888 ner, and A. Upadhye, Particle physics implications of a [52] D. O. Sabulsky, I. Dutta, E. A. Hinds, B. Elder, C. Bur- recent test of the gravitational inverse sqaure law, Phys rage, E. J. Copeland, Experiment to detect dark energy Rev Lett 98 131104, (2007). 0611223 forces using atom interferometry , Phys. Rev. Lett. 123, [32] D. Huterer and D. L. Shafer, Rep. Prog. Phys. 81, 016901 061102 (2019). 1812.08244 (2018), arXiv:1709.01091. [53] C. Renevey, J. Kennedy, and L. Lombriser Parameterised [33] M. Ishak, Living Rev. Rel. 1, 22 (2019), post-Newtonian formalism for the effective field theory of arXiv:1806.10122. dark energy via screened reconstructed Horndeski theories [34] M. Ballardini, M. Braglia, F. Finelli, D. Paoletti and , 2006.09910 A. A. Starobinsky, arXiv:2004.14349. Scalar-tensor the- [54] X. Zhang, W. Zhao, H. Huang, and Y. Cai, Post- ories of gravity, neutrino physics, and the H0 tension, Newtonian parameters and cosmological constant of 2004.14349 screened modified gravity , Phys. Rev. D 93, 124003 [35] H. Desmond, B. Jain and J. Sakstein, A local resolution (2016). 1603.09450 of the Hubble tension: The impact of screened fifth forces [55] A. Hees, A. F¨uzfa, Combined cosmological and solar sys- on the cosmic distance ladder, Phys. Rev. D 100, 043537 tem constraints on chameleon mechanism , Phys. Rev. D (2019). 1907.03778 85, 103005, 2012. 1111.4784 [36] R. D’Agostino and R. C. Nunes, Measurements of H0 in [56] K. Koyama and J. Sakstein, Astrophysical Probes of the modified gravity theories: The role of lensed quasars in Vainshtein Mechanism: Stars and Galaxies, Phys. Rev. the late-time Universe, Phys. Rev. D 101, 103505 (2020). D 91, 124066 (2015). 1502.06872 2002.06381 [57] M. Cerme˜no,J. Carro, A. L. Maroto, M. Angeles´ P´erez- [37] D. Psaltis, Probes and Tests of Strong-Field Gravity with Garc´ıa, Modified Gravity at Astrophysical Scales, Astro- Observations in the Electromagnetic Spectrum, Living physical Journal, 872 130, (2019). 1811.11171 Rev. Relativ. 11 (2008). 0806.1531 [58] B. F. de Aguiar and R. F. P. Mendes, Highly com- [38] E. Berti et al.,, Testing General Relativity with Present pact neutron stars and screening mechanisms. I. Equi- and Future Astrophysical Observations, Class. Quantum librium and stability, Phys. Rev. D 102 024064, (2020). Grav. 32, 243001 (2015). 1501.07274 2006.10080 [39] J. Khoury and A. Weltman, Chameleon Fields: Awaiting [59] G.J. Olmo, D. R. Garcia, and A. Wojnar, Stellar struc- Surprises for Tests of Gravity in Space, Phys. Rev. Lett. ture models in modified theories of gravity: lessons and 93 171104, (2004). 0309300 challenges, 1912.05202 [40] J. Khoury and A. Weltman, Chameleon Cosmology, [60] K. Y. Eksi, C. G¨ung¨orand M. M. T¨urko˘glu, What does Phys. Rev. D 69 044026, (2004). 0309411 a measurement of mass and/or radius of a neutron star [41] K. Hinterbichler and J. Khoury, Symmetron Fields: constrain: Equation of state or gravity? , PRD 89 063003 Screening Long-Range Forces Through Local Symme- (2014). 1402.0488 try Restoration, Phys. Rev. Lett. 104, 231301 (2010). [61] X. T. He, F. J. Fattoyev, B. A. Li, and W. G. New- 1001.4525 ton, Impact of the equation-of-state – gravity degeneracy [42] P. Brax, C. van de Bruck, A. C. Davis, and D. Shaw, on constraining the nuclear symmetry energy from as- The dilaton and modified gravity., Phys Rev D 82 063519 trophysical observables, Phys. Rev. C 91, 015810 (2015). (2010). 1005.3735 1408.0857 [43] A. I. Vainshtein, Phys. Lett. 39B, 393 (1972). [62] J. Sakstein, Tests of Gravity with Future Space-Based Ex- [44] C. Burrage and J. Sakstein, Tests of Chameleon Gravity, periments, Phys. Rev. D 97, 064028 (2018). 1710.03156 Living Rev. Relativity 21 1 (2018). 1709.09071 [63] D. F. Mota and D. J. Shaw, Strongly Coupled Chameleon [45] P. Brax, A. C. Davis, and R. Jha, Neutron Stars in Fields: New Horizons in Scalar Field Theory, Phys. Rev. Screened Modified Gravity: Chameleon vs Dilaton , Phys. Lett. 97 151102, (2006). 0606204 Rev. D 95, 083514 (2017). 1702.02983 [64] Shapiro, S. L. and Teukolsky, S. A. 1983, A Wiley- [46] A. Dima and F. Vernizzi, Vainshtein Screening in Scalar- Interscience Publication, New York: Wiley, 1983. Tensor Theories before and after GW170817: Con- [65] J. S. Read, B. D. Lackey, B. J. Owen and J. L. Fried- straints on Theories beyond Horndeski , Phys. Rev. D man, Constraints on a phenomenologically parameterized 97, 101302 (2018). 1712.04731 neutron-star equation of state, Phys.Rev.D 79 124032, [47] D. F. Mota and D. J. Shaw, Evading Equivalence Prin- (2009). 0812.2163 ciple Violations, Cosmological and other Experimental [66] M. F. O’Boyle, C. Markakis, N. Stergioulas and J. S. Constraints in Scalar Field Theories with a Strong Cou- Read, A Parametrized Equation of State for Neutron Star pling to Matter , Phys. Rev. D 75 063501, (2007). Matter with Continuous Sound Speed, 2008.03342 0608078 [67] B. Margalit and B. Metzger, Constraining the Maximum [48] X. Zhang, T. Liu, and W. Zhao, Gravitational radiation Mass of Neutron Stars from Multi-messenger Observa- from compact binary systems in screened modified gravity tions of GW170817, The Astrophysical Journal 850, L19 , Phys. Rev. D 95, 104027 (2017). 1702.08752 (2017). 1710.05938 [49] P. Brax, A. C.Davis, B. Li, and H. A. Winther, A Unified [68] C. Breu and L. Rezzolla, Maximum mass, moment of in- Description of Screened Modified Gravity , Phys. Rev. D ertia and compactness of relativistic stars. Monthly No- 95 083514, (2017). 1203.4812 tices of the Royal Astronomical Society 459, 646–656 [50] H. Desmond and J. Sakstein, Screened fifth forces lower (2016). 1601.06083 the TRGB-calibrated Hubble constant too , Phys. Rev. D 10

[69] B. Jain, V. Vikram and J. Sakstein Astrophysical Tests of Journal 422, 227 (1994). Modified Gravity: Constraints from Distance Indicators [73] C. E. Rasmussen and C. K. I. Williams, Gaussian Pro- in the Nearby Universe, Astrophys. J. 779 39 (2013). cesses for Machine Learning. The MIT Press (2006). 1204.6044 [74] F. Pedregosa et al, Scikit-learn: Machine Learning in [70] W. J. Wolf and M. Lagos Standard Sirens as a novel probe Python, Journal of Machine Learning Research 12 (2011) of dark energy, Phys. Rev. Lett. 124, 061101 (2020). [75] T. Yang, S. Birrer, and B. Hu, The first simultane- 1910.10580 ous measurement of Hubble constant and post-Newtonian [71] B. Cook, S. Shapiro, and S. Teukolsky, Rapidly Rotating parameter from Time-Delay Strong Lensing, Mon. Not. Neutron Stars in General Relativity: Realistic Equations Roy. Astron. Soc. 497 1 (2020). 2003.03277 of State. The Astrophysical Journal 424, 823 (1994) [76] D. Jyoti, J. B. Munoz, R. R. Caldwell, and M. [72] B. Cook, S. Shapiro, and S. Teukolsky, Rapidly Rotat- Kamionkowski, Cosmic Time Slip: Testing Gravity on ing Polytropes in General Relativity. The Astrophysical Supergalactic Scales with Strong-Lensing Time Delays, Phys. Rev. D 100, 043031 (2019). 1906.06324