Arxiv:2007.09714V1 [Gr-Qc] 19 Jul 2020 Tem PSR J0737-3039A/B

Probing Noncommutative Gravity with Gravitational Wave and Binary Pulsar Observations

Leah Jenks,1 Kent Yagi,2 and Stephon Alexander1 1Brown Theoretical Physics Center and Department of Physics, Brown University, 182 Hope Street, Providence, Rhode Island, 02903 2Department of Physics, University of Virginia, P.O. Box 400714, Charlottesville, VA 22904-4714, USA (Dated: July 21, 2020) Noncommutative gravity is a natural method of quantizing spacetime by promoting the spacetime coordinates themselves to operators which do not commute. This approach is motivated, for example, from a quantum gravity perspective, among others. Noncommutative gravity has been tested against the binary black hole merger event GW150914. Here, we extend and improve upon such a previous analysis by (i) relaxing an assumption made on the preferred direction due to noncommutativity, (ii) using posterior samples produced by the LIGO/Virgo Collaborations, (iii) consider other gravitational wave events, namely GW151226, GW170608, GW170814 and GW170817, and (iv) consider binary pulsar observations. Using Kepler’s law that contains the noncommutative effect at second post-Newtonian order, we derive corrections to the gravitational waveform phase and the pericenter precession. Using the gravitational wave and double pulsar binary observations, we find bounds on a space-time noncommutative tensor θ0i in terms of the preferred frame direction with respect to the orientation of each binary. We find that the gravitational wave bounds are stronger than the binary pulsar one by an order of magnitude and the noncommutative tensor normalized by the Planck length and time is constrained to be of order unity.

I. INTRODUCTION of noncommutative gravity stems from these theories. Non-commuting conjugate variables are a cornerstone of Since the advent of gravitational wave astronomy quantum mechanics, and it seems natural that one could with the detection of gravitational waves (GWs) by the apply the same conventions that give rise to, for example, LIGO/Virgo collaboration (LVC), the theory of general the Heisenberg uncertainty principle in quantum mechan- relativity (GR) has been directly testable to greater pre- ics, to a gravitational setting [16, 17]. Noncommutative cision than previously possible. Although no observa- gravity also has string theory implications [22, 23] and tions have yet indicated any compelling deviations from thus we have a wide range of motivations for its study. GR, we are able to study modifications to GR, alternative The version that we will be focused on is characterized theories of gravity and other fundamental physics using by promoting spacetime coordinates to operators which gravitational waves as a probe [1–6]. Particularly, non- satisfy the following canonical commutation relation GR effects are highly constrained by GW observations, [ˆxµ, xˆν ] = iθµν . (1) which can be used to explore many different theories. This has been done, for example for Einstein-Aether the- Here, θµν introduces a new fundamental quantum scale ory [7], Einstein-dilaton-Gauss-Bonet gravity [8,9], dy- which represents the “quantum fuzziness” of spacetime, namical Chern-Simons gravity [8] and others [10, 11]. in analogy to ~ in quantum mechanics. In addition to gravitational waves, pulsar timing ob- Previous work [24] has placed a bound on the time servations are valuable tool in probing modifications to component of the noncommutativity scale θ0i using GR. The system that we will be using to place constraints √ GW150914, and found a constraint Λ . 3.5, at the via pulsar observations is the double pulsar binary sys- order of the Planck scale (see [25, 26] for related works). arXiv:2007.09714v1 [gr-qc] 19 Jul 2020 tem PSR J0737-3039A/B. This system is quite unique, as The authors worked in the post-Newtonian (PN) formal- both neutron stars are radio pulsars, which allows for ex- ism, in which quantities are expanded in powers of (v/c)n tremely precise measurements and provides a rich back- with v representing the relative velocity of the binary con- ground for tests of general relativity and modified theo- stituents, which are considered order n/2PN [27]. Refer- ries of gravity [12, 13]. ence [24] found corrections entering at 2PN in the accel- In this paper, we will employ a combination of gravi- eration and the waveform phase. The authors introduce tational wave and pulsar analysis to introduce two inde- the notation pendent constraints on noncommutative theories. Var- ious noncommutative theories have been proposed pre- θ0i Λθi = , (2) viously, originally introduced as a method of quantizing lptp spacetime [14]. The introduction of noncommutative geometry [15] allowed this idea to be applied more broadly, where θi represent the components of a three-dimensional with a focus on noncommutative quantum field theories unit vector θ, which acts as a preferred direction that [16, 17] as well as multiple formulations of a noncommu- induces precession of the orbital plane. For calculational tative extension to the Standard Model [18–21]. The idea simplicity, the authors assumed that the orientation of 2 the θ is orthogonal to the orbital√ plane as to place an III by computing the 2PN non-commutative correction to approximate upper bound on Λ. the gravitational waveform. We then use posterior samples for two different waveform templates to constrain the In this work we extend and generalize the above anal- √ noncommutativity parameter, Λ, in terms of the quan- ysis by considering the general case for the orientation of tity Lˆ · θ. In SectionIV we then independently constrain the preferred direction θ with respect to the orbital plane √ Λ as a function of Lˆ ·θ by computing the noncommuta- by adopting orbital averaging. We then place constraints tive correction to the GR pericenter precession and using by employing posterior samples from the GWTC-1 cata- the binary pulsar event PSR J0737-3039A/B. Finally, In log following [8], rather than explicitly using the bound sectionV, we summarize our results, and provide some on the non-GR parameter at 2PN order found by LVC concluding remarks as well as directions for future work. [1] as done in [24]. This new approach properly accounts We work in the geometric units c = G = 1. for the uncertainties in the masses.We derive bounds from four different gravitational wave events with rel- atively small masses, namely GW151226, GW170608, II. NONCOMMUTATIVE CORRECTIONS TO GW170814 and GW170817. THE ACCELERATION AND ENERGY We also place constraints on the the time component of the noncommutative tensor using the binary pulsar sys- In GR, one can approximate a binary system as two tem PSR J0737-3039A/B to act as an independent check point masses which have an energy momentum tensor on the gravitational wave constraints. In the binary pul- given by sar system, the noncommutativity induces an additional µν µ ν 3 contribution in the pericenter precession beyond GR at TGR (x, t) = m1γ1(t)v1 (t)v1 (t)δ [x − y1(t)] + 1 ↔ 2. (3) 1PN, due to the preferred direction θ that is induced by Here, mi are the masses of each body, yi the positions µ the inclusion of noncommutative terms. Corrections to and vi the four velocities. γi is given by other observables, such as the mass ratio and Shapiro delay, enter at higher PN orders. Thus, we use the latter to 1 γi = q , (4) determine the masses of the double pulsar binary and use g (g ) (vαvβ/c2) the pericenter precession to constrain the theory (see e.g. i αβ i i i [28] for a related work on constraining noncommutative where g is the metric, g its determinant and i = 1, 2 gravity from the pericenter precession of binary pulsars). µν [27]. It was previously shown in [24] that noncommu- We found that such bounds are slightly weaker than the tative corrections to the expression3 can be found by ones from gravitational wave events. considering that the black holes are sourced by a mas- The structure of the paper is as follows. In SectionII sive real scalar field φ and incorporating the noncom- we derive the lowest order 2PN noncommutative correc- muting operatorsx ˆµ by replacing the product of any tions to the binary system acceleration, beginning from two functions with a Moyal product. It was found that the energy-momentum tensor. We proceed to constrain the energy-momentum tensor, including noncommuta- the noncommutative parameter with LVC data in Section tive corrections, can be written as

m3γ3 T µν (x, t) = mγ (t)vµ(t)vν (t)δ3[x − y (t)] + L vµvν Θkl∂ ∂ δ3[x − y (t)] NC L 1 8 k l 1 2 2 (5) µm νn µν i ~ mγL~ kl 3 +(η η ∂m∂n − η ∂i∂ ) + Θ ∂k∂l δ [x − y1(t)], 4mγL 32

where γL is the Lorentz factor and we define lowest order noncommutative corrections, and thus can 0k 0l 0k pl kp lq make the approximation γL = 1. Then, for a binary sys- kl θ θ θ θ θ θ tem which considers only the lowest order noncommuta- Θ = 2 2 + 2vp 3 + vpvq 4 . (6) lptp lptp lp tive corrections, the energy-momentum tensor simplifies to Here, θ is the noncommutativity parameter defined by Eq. (1) while l and t are the Planck length and time µν µ ν 3 p p TNC (x, t) = m1γ1(t)v1 (t)v1 (t)δ (x − y1(t)) respectively. The second term in Eq. (5) is suppressed 3 2 (7) 2 m1Λ µ ν k l 3 by a factor of ~ and can be neglected. We will con- + v1 (t)v1 (t)θ θ ∂k∂lδ (x − y1(t)) + 1 ↔ 2, sider the contribution from the first term in Eq. (28), 8 which using the convention that a term of order (v/c)n where we have defined a normalization of the noncom- is of order (n/2)PN, enters as a correction at the sec- mutative tensor, Λ as in Eq.2. In analogy to [24], we ond Post-Newtonian order (2PN). We will consider only follow the standard procedure to arrive at the accelera- 3 tion, where we consider only the leading order GR contri- III. GRAVITATIONAL WAVE CONSTRAINTS bution and the lowest order noncommutative correction entering at 2PN: In this section, we study bounds on noncommutative 15M 3(1 − 2ν)Λ2 gravity with gravitational wave observations. We first a = (a ) − θkθlnˆ . (8) i i GR 8r4 ikl derive corrections to the√ gravitational waveform phase. We then find bounds on Λ using posterior samples of 2 Here, M = m1 + m2 is the total mass, ν = m1m2/M is selected gravitational wave events produced by LVC. the symmetric mass ratio while r is the binary separation. We have also introduced the quantities r = |y1 − y2| and n such that n = (y1 − y2)/r to define the quantity A. Gravitational Waveform 1 nˆ = n n n − (δ n + δ n + δ n ). (9) ikl i k l 5 kl i il k ki l From the acceleration and the conserved energy, we From the acceleration we can also determine the correc- can compute noncommutative corrections to the gravita- tion to the GR Lagrangian tional waveform to constrain the theory. We focus on a quasicircular orbit such that r is a constant. Defining the 3M 3µ(1 − 2ν)Λ2 relative position as y(t) = y (t) − y (t), we can rewrite L = L + θkθlnˆ , (10) 1 2 GR 8r3 kl Eq. (14) as and the conserved energy: a = −Ω2y + O(1/c5). (16) 3M 3µ(1 − 2ν)Λ2 E = E − θkθlnˆ . (11) GR 8r3 kl In order to find the leading noncommutatice correction to the waveform, we here keep only the leading GR and Here, µ = m m /M is the reduced mass and 1 2 2PN NC term. The angular velocity Ω is given by δkl nˆ = n n − . (12) 2 kl k l 3 2 M 3(1 − 2ν)Λ ˆ 2 2 Ω = 3 1 + (1 − 3(L · θ) )γ , (17) In these expressions for the acceleration, conserved en- r 16 ergy and Lagrangian, the vector θ acts as a preferred di- where we have defined the quantity rection and will in general induce precession in the orbital plane. Previous work [24] simplified these expressions for M γ = . (18) the acceleration, Lagrangian and conserved energy by as- r suming a constrained case in which the orbital plane is perpendicular to the preferred direction, θ. Given that Similarly, taking into account both the explicit 2PN contribution to the energy as well as the 2PN correction to each binary is expected to be oriented randomly with re- 2 spect to the preferred direction, the chance of the above Ω in the leading order GR contribution, we have assumption being satisfied seems extremely low. To over- µγ 1 come this, we perform an orbital averaging procedure as E = − 1 − (1 − 2ν)Λ2 1 − 3(Lˆ · θ)2 γ2 . (19) is typically done for precessing [29, 30] and magnetized 2 16 [31] binaries. We will consider the following relation as an Then, inverting Eq. (17) and defining the quantity x = orbital average over the unit vector n and the preferred (MΩ)2/3 that corresponds to relative velocity squared, direction θ as follows: we can rewrite the conserved energy in terms of x: 1 (n · θ)(n · θ) = 1 − (Lˆ · θ)2 . (13) 2 µx 1 E = − 1 − (1 − 2ν)Λ2 1 − 3(Lˆ · θ)2 x2 . (20) Here, Lˆ is a unit vector orthogonal to the orbital plane, 2 8 as the projection of the angular momentum of the binary ˆ To determine the lowest order noncommutative correc- system. The case L · θ = 1 corresponds to the limit- tions to the energy radiated by gravitational waves, we ing case in which the preferred direction is perpendicular assume the energy balance equation to the orbital plane. Employing the orbital averaging procedure, we obtain for the acceleration and conserved dE = −L. (21) energy: dt 3 2 15M (1 − 2ν)Λ Here, L is the gravitational wave luminosity, which is ai =(ai)GR − 8r4 given by 2 1 2 × ni(n · θ) − ni − θi(n · θ) , (14) 1 d3I d3I 5 5 L = ij ij + O(1/c2) , (22) 5 dt3 dt3 and M 3µ(1 − 2ν)Λ2 where Iij is the traceless mass quadrupole moment. E = E − 1 − 3(Lˆ · θ)2 . (15) GR 16r3 There are two noncommutative corrections to the 4 quadrupole moment. The explicit 2PN noncommutative which the phase of the waveform in Fourier domain can contribution is time independent, as shown in [24], and be written as will not contribute to the gravitational wave luminosity. π Thus, we only need to consider the Newtonian part of I , ψ(f) = 2πftf − − Φ(tf ). (31) ij 4 which will lead to noncommutative corrections through the acceleration. For the leading order and 2PN non- tf is the time such that dΦ(tf )/dt = f. It can be found commutative corrections to the third derivative of the from (28), and Φ(tf ) is found from (30) to obtain the full quadrupole moment, we find expression for the inspiral phase including the leading order term and explicit 2PN noncommutative correction: 2 ... 8νM yivj + viyj I ij = − r3 2 π 3 −5/3 ψI (f) = 2πftc − φc − + (πMf) " # 4 128ν 15 1 2 2 2 (23) 5 h i × 1 + Λ (1 − 2ν) (n · θ) − γ × 1 − Λ2(1 − 2ν) 35 − 75(Lˆ · θ)2 (πMf)4/3 . 8 5 16 9νM 4 (32) + Λ2(1 − 2ν)(n · θ)(θ v + θ v ). 4r4 i j j i This expression follows the standard PN waveform for- Squaring and keeping only the relevant lowest order mat, terms, we find after orbital averaging and inserting the 4 result into Eq. (22) that the full expression for the lumi- π 3 X (j−5/3) ψ (f) = 2πft − φ − + ϕ (πMf) . nosity is I c c 4 128ν j j=0 32 Λ2(1 − 2ν) (33) L = ν2x5 1 + 23 − 39(Lˆ · θ)2 x2 . 5 32 We will be interested in the ϕ4 coefficient, which is (24) what enters at 2PN. Including our correction in addition It is then straightforward to determine the evolution to the 2PN GR contribution to ϕ4 [33], we have of the orbital phase of the binary system. We define a 15293365 27145 3085 new parameter ϕ = + ν + ν2 4 508032 504 72 ν 5 2 h 2i Θ ≡ (tc − t), (25) − Λ (1 − 2ν) 35 − 75(Lˆ · θ) . (34) 5GM 16 where tc is the coalescence time, such that the energy We can then define the fractional deviation from GR as balance equation can be written as ϕNC δϕNC ≡ 4 dE dx 5M 4 ϕGR = L. (26) 4 dx dΘ ν 158760(1 − 2ν) = This can then be solved order by order to find 4353552ν2 + 5472432ν + 3058673 h i × −7 + 15(Lˆ · θ)2 Λ2. (35) 1 Λ2(1 − 2ν) x = Θ−1/4 1 − [35 − 75(Lˆ · θ)2]Θ−1/2 . 4 1024 We can now employ this result to constrain the quantity (27) h i ˆ 2 2 We then invert this expression for x to find Θ, −7 + 15(L · θ) Λ from gravitational wave events.

1 Λ2(1 − 2ν) h i Θ = 1 − 35 − 75(Lˆ · θ)2 x2 . √ 256x4 16 B. Bounds on Λ (28) Θ is related to the orbital phase by the following Having obtained the expression for the noncommuta- dφ 5 tive correction to ϕ4, it is now straightforward to com- = − x3/2, (29) pute bounds. However, one issue we still face is the pres- dΘ ν ence of ν, the symmetric mass ratio in the expression NC which can easily be solved to ﬁnd δϕ4 . One could simply take the central values given for each binary component mass to compute ν, however x−5/2 5 h i φ = − 1 − Λ2(1 − 2ν) 35 − 75(Lˆ · θ)2 x2 . this method does not take into account the uncertainties 32ν 32 in the mass and will not give as precise of an answer. As (30) an alternative, we will follow the method outlined in [8] Then, as we have assumed that the velocity of each and make use of the LVC posterior samples for multiple binary component is small compared to c, we may use events in order to obtain 90% conﬁdence bounds on the the stationary phase approximation (SPA) [32], under noncommutative parameter. 5

We use posterior samples from the GWTC-1 catalog rection to the pericenter precession, given by for gravitational wave events GW 151226, GW170608 " and GW170814 [5, 34]. Data for events GW150914 and dω 1r p 2 + ecosf¯ = − cos fR + sin f¯S GW170104 are also available, however these two events dt e M 1 + e cos f¯ are characterized by large masses and thus a short in- # (36) spiral period. This makes it difficult to reliably probe sin(ω + f¯) −e cot ι W . non-GR effects through corrections to the waveform [1], 1 + e cos f¯ thus we do not include constraints from these events. We do however also calculate constraints based on the binary The relevant orbital elements here are the eccentricity e, neutron star event GW170817, for which posterior sam- the inclination ι, the nodal angle Ω, the pericenter angle ples are also available [6, 35]. ω and the semilatus rectum p, defined by p = a(1 − e2) Inverting Eq. (35) allows us to obtain an expression for where a is the semi-major axis. Then, φ is the orbital ¯ −7 + 15(Lˆ · θ)2 Λ2 in terms of δϕ , m and m . Then, phase as measured from the ascending node, and f is the 4 1 2 true anomaly, defined by f¯ ≡ φ − ω. Here the noncom- using posterior samples for the two waveform templates mutative correction to the radial, cross-track and out-of- IMRPhenomPv2 (IMRP) and SEOBNRv4 (SEOB), we plane components of the perturbing acceleration, R, S, are able to plot the histograms and probability distribu- and W are given by tion functions (PDFs) for each event, shown in Fig.1. 9M 3(1 − 2ν)Λ2 1 R = − (n · θ)2 − , (37) From the PDFs for each event, we calculate 90% con- NC 8r4 3 straints on Λ2[−7 + 15(Lˆ · θ)2] as an upper and lower bound. We√ can then use these upper and lower bounds to ˆ 3M 3(1 − 2ν)Λ2 constrain Λ as a function of L·θ. These constraints are S = (λ · θ)(n · θ), (38) shown in Fig.2 for both waveform templates. From Fig. NC 4r4 2, we see that there is a region of the Lˆ ·θ plane in which we cannot constrain the noncommutativity parameter, 3 2 p 3M (1 − 2ν)Λ ˆ specifically when Lˆ ·θ = 7/15. However, given that we WNC = (h · θ)(n · θ), (39) 4r4 are considering multiple gravitational wave events and ˆ L · θ varies from one binary to another, statistically the where λ is defined as ∂n/∂φ and hˆ = n × λ. Expand- chance that each of those events would be specifically at ing out these expressions in Cartesian coordinates in the ˆ p L · θ = 7/15 is low, thus we expect we can still place equations of motion yields a complicated expression that meaningful bounds. In total,√ we can see that the 90% can be further simplified as in [37] by introducing the ˆ confidence constraints on Λ as a function of L · θ is in- variables deed constrained to be of order unity, as was previously found in [24]. eP ≡ n|φ=ω = eΩ cos ω + e⊥ sin ω, (40)

eQ ≡ λ|φ=ω = −eΩ sin ω + e⊥ cos ω, (41) IV. BINARY PULSAR CONSTRAINTS ˆ h ≡ eP × eQ = eΩ × e⊥. (42) We now turn to constraints on the noncommutativity tensor from the double pulsar binary system PSR J0737- Here, eP is a unit vector pointing towards the pericenter ˆ 3039A/B [12]. We first derive noncommutative correc- and eQ = h × eP . eΩ is a unit vector which points along ˆ tions√ to the pericenter precession. We then find bounds the ascending node, and e⊥ = h × eΩ. n and λ can be on Λ using the double pulsar system. analogously translated into the P , Q, and h coordinates. Next, we integrate Eq. (36) from 0 to 2π to find the noncommutative correction to ∆ω. We find 3πM 2Λ2(1 − 2ν) ∆ω = − [2 − 3θ2 − 3θ2 + 2(Lˆ · θ) cot ι] A. Pericenter Precession NC 8p P Q

× (θp cos ω + θQ sin ω). (43) Beginning from the acceleration, Eq. (14), we can easily calculate the correction to the pericenter√ precession In this expression there is both explicit ω dependence, as to provide another independent bound on Λ. We will well as implicit ω dependence in θp and θQ. Thus, it is treat the noncommutative correction to the acceleration more enlightening to express everything in terms of eΩ 0 as a perturbing acceleration δa. We can then define the and e⊥. We can then expand ω = ω0 +ω φ and integrate orbital parameters following a standard formulation of over φ. For ω0, it is sufficient to consider the GR con- osculating orbits explained e.g. in [36] and find the cor- tribution only since ∆ωNC above is already proportional 6

IMRP SEOB Probability Probability 0.010

0.012 GW151226 GW151226 0.008 0.010 GW170608 GW170608 GW170814 GW170814 0.008 0.006 GW170817 GW170817

0.006 0.004

0.004

0.002 0.002

0.000 0.000 -100 0 100 200 -200 -100 0 100 200  2  2 Λ2 -7+ 15L · θ  Λ2 -7+ 15L · θ 

FIG. 1. Posterior distributions of Λ2(−7 + 15(Lˆ · θ)2) for various gravitational wave events derived from the posterior samples using the IMRP (left) and SEOB (right) waveform templates.

SEOB and IMRP Constraints ω˙ [38] we obtain Λ −5/3 ( −2/3 4.0 P M 2/3 1 1 P ω˙ =3 b 1 + b GW151226 2π 1 − e2 16 1 − e2 2πM 3.5 GW170608 h io ×Λ2(1 − 2ν) 1 − 3(Lˆ · θ)2 . (46) 3.0 GW170814 GW170817 For completeness, we present the noncommutative cor- 2.5 rections to other orbital elements in AppendixA.

2.0 √ B. Bounds on Λ 1.5

L ·θ √ 0.2 0.4 0.6 0.8 1.0 We now derive constraints on Λ with the double pulsar system PSR J0737-3039A/B. We wish to useω ˙ to FIG. 2. Constraints on the noncommutative parameter Λ for constrain the theory. To do so, we need to determine each gravitational wave event from SEOB (dashed) and IMRP the masses from other observables. Here, we use the (solid) waveform templates. Shapiro delay s, masss ratio R, and the mass functions fA. The noncommutative correction to these observables enter through the Kepler’s law at 2PN or higher (see [39] 2 2 2 ˆ 2 for a similar analysis when the metric is modiﬁed at 1PN to Λ . Noting that θΩ + θ⊥ + (L · θ) = 1 and that for the J0737-3039A/B system, ι ≈ π/2, we obtain the order within the parameterized PN formalism), while the correction to the pericenter precession as one inω ˙ in Eq. (46) enters at 1PN order. This justiﬁes us to use the GR expressions for s, R and fA to determine the masses and useω ˙ to test the noncommutative 3πM 2Λ2(1 − 2ν) h i ˆ 2 gravity. ∆ωNC = 2 1 − 3(L · θ) . (44) 8p Figure3 shows these observables plotted as a function of the pulsar masses. The GR expressions for R, s, and Then, the noncommutative correction to the observable the region for which sin ι (obtained from the mass func- quantity of pericenter precessionω ˙ , which can be found tion measurements) is less than one are plotted. The by dividing ∆ω by the orbital period, Pb is overlapping shaded region corresponds to the allowed mass parameter space from these measurements. Any 4/3 −7/3 correction toω ˙ must remain within the region of over- 3 M Pb 2 h ˆ 2i ω˙ NC = Λ (1−2ν) 1 − 3(L · θ) , lap. The upper and lower bounds forω ˙ correspond to 16 (1 − e2)2 2π h i 2 ˆ 2 (45) variations in the expression Λ 1 − 3(L · θ) such that where we have used Kepler’s law to write p2 in terms of theω ˙ curve marginally passes through the overlapping the orbital period. Adding this to the GR expression for region. The thickness in each of the curves corresponds 7

Binary Pulsar Constraints Λ 30

L ·θ 0.0 0.2 0.4 0.6 0.8 1.0 √ FIG. 4. Bounds on Λ as a function of (Lˆ · θ) for the double pulsar binary.

FIG. 3. Testing noncommutative gravity with the double ing at 2PN in the binary system acceleration. The time pulsar binary. The two masses are determined from the mass component of the noncommutative tensor, θ0i enters as ratio R, the Shapiro delay parameter s and sin ι < 1 using a 2PN correction to the acceleration. When this effect is the GR expressions (since the noncommutative corrections to propagated through, we find that there is a phase shift in these observables enter at higher PN orders than that forω ˙ ), the gravitational waveform again entering at 2PN, shift- with the allowed region shown by the green shade. We then ing the ϕ4 coefficient. Similarly, in the case of binary vary the noncommutative parameter Λ inω ˙ such that it is pulsar dynamics, we find that the correction to the ac- consistent with the green shaded region to determined the bound on Λ. The thickness ofω ˙ in blue corresponds to the celeration leads to a noncommutative contribution to the measurement error onω ˙ . pericenter precession. An updated and more rigorous analysis than in previous work has been performed to use gravitational wave to the uncertainty in theω ˙ measurement and the two events and the binary pulsar system PSR J0737-3039A/B curves correspond to the upper and lower bounds on to constrain the space-time component of the noncom- h i Λ2 1 − 3(Lˆ · θ)2 . We find that the acceptable range mutativity tensor. We find that the gravitational wave for the noncommutative contribution is events including GW151226, GW170608, GW170814 and GW170817 are more constraining than the binary pulsar 2 h ˆ 2i − 15600 . Λ 1 − 3(L · θ) . 1100. (47) event PSR J0363-3039A/B by approximately an order of magnitude. However, the more stringent GW constraints Then, as we did in the gravitational wave analysis,√ tak- are consistent√ with previous results, findng that the the ing this upper and lower bound, we can plot Λ as a quantity Λ is constrained to be of order unity. function of (Lˆ · θ) as in Fig.4. We can see that there A few different avenues exist for future work. For ex- is again a particular value of (Lˆ · θ) = p1/3 that we ample, it would be interesting to constrain the theory are not able to place a constraint as was the case for the from the preferred frame effect [37]. It would also be of GW analysis, however we are still able to place bounds interest to investigate the effects of the spatial compo- ij for the rest of the range. We find that the binary pul- nent of the noncommutative tensor, θ , which enters at sar bounds are actually less stringent than those found 3PN and has potential implications for e.g. string the- from the gravitational wave events by approximately an ory. Additionally, it would be valuable to explore the order of magnitude. However, these constraints remain model dependence of the effects that we have discussed, consistent with the general statement that the noncom- and work towards a more general understanding of how mutativity parameter must be of order unity. noncommutative gravity may come into play with these observables.

V. CONCLUSION ACKNOWLEDGMENTS We have explored noncommutative gravity in light of observations from LVC gravitational wave events as well K.Y. acknowledges support from NSF Award PHY- as the binary pulsar system J0737-3039A/B. We have fo- 1806776, NASA Grant 80NSSC20K0523, a Sloan Foun- cused on the lowest order noncommutative eﬀects enter- dation Research Fellowship and the Ed Owens Fund. 8

K.Y. would like to also acknowledge support by the ∆eNC =0, (A6) COST Action GWverse CA16104 and JSPS KAKENHI 2 2 3πM Λ (1 − 2ν) ˆ Grants No. JP17H06358. ∆ιNC = (L · θ)(θ cos ω − θ sin ω), 4p p Q (A7) Appendix A: Noncommutative Corrections to 3πM 2Λ2(1 − 2ν) ∆Ω = (Lˆ · θ)(θ cos ω + θ sin ω) csc i. Osculating Orbits NC 4p p Q (A8) In addition to the noncommutative correction to the pericenter precession,ω ˙ , the noncommutative correction to the acceleration also induces corrections to the other orbital parameters, p, e, i, and Ω, described in SectionIV. As in ∆ω there is both explicit and implicit ω dependence The “Lagrange planetary equations” for these quantities in these expressions. Using the same expansion method, are [36] we obtain for the noncommutative contributions to the orbital parameters: r dp p3 S =2 , (A1) dt M 1 + e cos f¯ r ¯ 2 ¯ de p ¯ 2 cos f + e + e cos f ∆p =0, (A9) = sin fR + ¯ S , (A2) NC dt M 1 + e cos f ∆e =0, (A10) r NC dι p r 3πM 2Λ2(1 − 2ν) = W cos φ, (A3) ∆ι = (Lˆ · θ)θ , (A11) dt M p NC 4p2 Ω dΩ r p r sin θ = W . (A4) dt M p sin ι 2 2 3πM Λ (1 − 2ν) ˆ ∆ΩNC = (L · θ)θ⊥ csc i. (A12) Plugging in the expressions for S, W, and R obtained in 4p2 SectionIV, it is straightforward to obtain

∆pNC =0, (A5)

[1] B. P. Abbott et al. (LIGO Scientific, Virgo), Tests [8] R. Nair, S. Perkins, H. O. Silva, and N. Yunes, Funda- of general relativity with GW150914, Phys. Rev. mental Physics Implications for Higher-Curvature Theo- Lett. 116, 221101 (2016), [Erratum: Phys. Rev. ries from Binary Black Hole Signals in the LIGO-Virgo Lett.121,no.12,129902(2018)], arXiv:1602.03841 [gr-qc]. Catalog GWTC-1, Phys. Rev. Lett. 123, 191101 (2019), [2] N. Yunes, K. Yagi, and F. Pretorius, Theoretical arXiv:1905.00870 [gr-qc]. Physics Implications of the Binary Black-Hole Mergers [9] K. Yamada, T. Narikawa, and T. Tanaka, Testing GW150914 and GW151226, Phys. Rev. D94, 084002 massive-field modifications of gravity via gravitational (2016), arXiv:1603.08955 [gr-qc]. waves, PTEP 2019, 103E01 (2019), arXiv:1905.11859 [3] E. Berti, K. Yagi, and N. Yunes, Extreme Gravity Tests [gr-qc]. with Gravitational Waves from Compact Binary Coa- [10] Z. Carson and K. Yagi, Parametrized and inspiral- lescences: (I) Inspiral-Merger, Gen. Rel. Grav. 50, 46 merger-ringdown consistency tests of gravity with multi- (2018), arXiv:1801.03208 [gr-qc]. band gravitational wave observations, Phys. Rev. D 101, [4] E. Berti, K. Yagi, H. Yang, and N. Yunes, Extreme Grav- 044047 (2020), arXiv:1911.05258 [gr-qc]. ity Tests with Gravitational Waves from Compact Bi- [11] S. Tahura and K. Yagi, Parameterized Post-Einsteinian nary Coalescences: (II) Ringdown, Gen. Rel. Grav. 50, Gravitational Waveforms in Various Modified Theories 49 (2018), arXiv:1801.03587 [gr-qc]. of Gravity, Phys. Rev. D 98, 084042 (2018), [Erratum: [5] B. Abbott et al. (LIGO Scientific, Virgo), Tests of Gen- Phys.Rev.D 101, 109902 (2020)], arXiv:1809.00259 [gr- eral Relativity with the Binary Black Hole Signals from qc]. the LIGO-Virgo Catalog GWTC-1, Phys. Rev. D 100, [12] M. Kramer et al., Tests of general relativity from timing 104036 (2019), arXiv:1903.04467 [gr-qc]. the double pulsar, Science 314, 97 (2006), arXiv:astro- [6] B. Abbott et al. (LIGO Scientific, Virgo), Tests of Gen- ph/0609417. eral Relativity with GW170817, Phys. Rev. Lett. 123, [13] E. Berti et al., Testing General Relativity with Present 011102 (2019), arXiv:1811.00364 [gr-qc]. and Future Astrophysical Observations, Class. Quant. [7] C. Zhang, X. Zhao, A. Wang, B. Wang, K. Yagi, Grav. 32, 243001 (2015), arXiv:1501.07274 [gr-qc]. N. Yunes, W. Zhao, and T. Zhu, Gravitational waves [14] H. S. Snyder, Quantized space-time, Phys. Rev. 71, 38 from the quasicircular inspiral of compact binaries in (1947). Einstein-aether theory, Phys. Rev. D 101, 044002 (2020), [15] A. Connes, Non-commutative differential geometry, Pub- arXiv:1911.10278 [gr-qc]. lications Mathématiquesde l’Institut des Hautes Etudes´ 9

Scientifiques 62, 41 (1985). (2017). [16] M. R. Douglas and N. A. Nekrasov, Noncommutative [29] L. E. Kidder, C. M. Will, and A. G. Wiseman, Spin effects field theory, Rev. Mod. Phys. 73, 977 (2001), arXiv:hep- in the inspiral of coalescing compact binaries, Phys. Rev. th/0106048. D47, R4183 (1993), arXiv:gr-qc/9211025 [gr-qc]. [17] R. J. Szabo, Quantum field theory on noncommuta- [30] L. E. Kidder, Coalescing binary systems of compact ob- tive spaces, Phys. Rept. 378, 207 (2003), arXiv:hep- jects to postNewtonian 5/2 order. 5. Spin effects, Phys. th/0109162. Rev. D52, 821 (1995), arXiv:gr-qc/9506022 [gr-qc]. [18] M. Chaichian, P. Presnajder, M. Sheikh-Jabbari, and [31] K. Ioka and K. Taniguchi, Gravitational waves from in- A. Tureanu, Noncommutative standard model: Model spiralling compact binaries with magnetic dipole mo- building, Eur. Phys. J. C 29, 413 (2003), arXiv:hep- ments, Astrophys. J. 537, 327 (2000), arXiv:astro- th/0107055. ph/0001218. [19] M. Chaichian, A. Kobakhidze, and A. Tureanu, Spon- [32] K. Yagi, Scientific Potential of DECIGO Pathfinder and taneous reduction of noncommutative gauge symmetry Testing GR with Space-Borne Gravitational Wave Inter- and model building, Eur. Phys. J. C 47, 241 (2006), ferometers, Int. J. Mod. Phys. D 22, 1341013 (2013), arXiv:hep-th/0408065. arXiv:1302.2388 [gr-qc]. [20] X. Calmet, B. Jurco, P. Schupp, J. Wess, and M. Wohlge- [33] T. Damour, B. R. Iyer, and B. Sathyaprakash, A nannt, The Standard model on noncommutative space- Comparison of search templates for gravitational waves time, Eur. Phys. J. C 23, 363 (2002), arXiv:hep- from binary inspiral, Phys. Rev. D 63, 044023 (2001), ph/0111115. [Erratum: Phys.Rev.D 72, 029902 (2005)], arXiv:gr- [21] P. Aschieri, B. Jurco, P. Schupp, and J. Wess, Noncom- qc/0010009. mutative GUTs, standard model and C,P,T, Nucl. Phys. [34] B. Abbott et al. (LIGO Scientific, Virgo), GWTC-1: A B 651, 45 (2003), arXiv:hep-th/0205214. Gravitational-Wave Transient Catalog of Compact Bi- [22] F. Ardalan, H. Arfaei, and M. Sheikh-Jabbari, Noncom- nary Mergers Observed by LIGO and Virgo during the mutative geometry from strings and branes, JHEP 02, First and Second Observing Runs, Phys. Rev. X 9, 016, arXiv:hep-th/9810072. 031040 (2019), arXiv:1811.12907 [astro-ph.HE]. [23] N. Seiberg and E. Witten, String theory and noncommu- [35] B. Abbott et al. (LIGO Scientific, Virgo), GW170817: tative geometry, JHEP 09, 032, arXiv:hep-th/9908142. Observation of Gravitational Waves from a Binary Neu- [24] A. Kobakhidze, C. Lagger, and A. Manning, Constrain- tron Star Inspiral, Phys. Rev. Lett. 119, 161101 (2017), ing noncommutative spacetime from GW150914, Phys. arXiv:1710.05832 [gr-qc]. Rev. D 94, 064033 (2016), arXiv:1607.03776 [gr-qc]. [36] C. M. Will, Theory and Experiment in Gravitational [25] W. Nelson, J. Ochoa, and M. Sakellariadou, Constrain- Physics (Cambridge University Press, 2018). ing the Noncommutative Spectral Action via Astrophys- [37] C. M. Will, Testing general relativity with compact-body ical Observations, Phys. Rev. Lett. 105, 101602 (2010), orbits: a modified Einstein–Infeld–Hoffmann framework, arXiv:1005.4279 [hep-th]. Class. Quant. Grav. 35, 085001 (2018), arXiv:1801.08999 [26] W. Nelson, J. Ochoa, and M. Sakellariadou, Grav- [gr-qc]. itational Waves in the Spectral Action of Noncom- [38] I. H. Stairs, Testing general relativity with pulsar timing, mutative Geometry, Phys. Rev. D82, 085021 (2010), Living Rev. Rel. 6, 5 (2003), arXiv:astro-ph/0307536. arXiv:1005.4276 [hep-th]. [39] L. Sampson, N. Yunes, and N. Cornish, Rosetta stone for [27] L. Blanchet and G. Faye, Lorentzian regularization and parametrized tests of gravity, Phys. Rev. D88, 064056 the problem of point - like particles in general relativity, (2013), [Erratum: Phys. Rev.D88,no.8,089902(2013)], J. Math. Phys. 42, 4391 (2001), arXiv:gr-qc/0006100. arXiv:1307.8144 [gr-qc]. [28] X.-M. Deng, Solar System and stellar tests of noncommutative spectral geometry, Eur. Phys. J. Plus 132, 85