---
title: 'Om Diagnostic: A Null Test for Dark Energy'
url: https://www.emergentmind.com/topics/om-diagnostic
type: topic
---

# Om Diagnostic: A Null Test for Dark Energy

The **Om diagnostic** is a geometrical null test of dark energy constructed from the expansion history alone, typically in a spatially flat FLRW background. In its canonical form,
$$
Om(z)=\frac{h^2(z)-1}{(1+z)^3-1},\qquad h(z)=\frac{H(z)}{H_0},
$$
so that flat $\Lambda$CDM implies
$$
h^2(z)=\Omega_{m0}(1+z)^3+(1-\Omega_{m0}) \;\Rightarrow\; Om(z)=\Omega_{m0}=\mathrm{const}.
$$
This constancy makes $Om$ a null diagnostic of the cosmological constant hypothesis. Because it depends only on the background expansion and only on the first derivative of the scale factor through $H=\dot a/a$, it is generally easier to reconstruct observationally than higher-derivative diagnostics such as the statefinder hierarchy [0807.3548; 1005.0433].

## 1. Formal definition and null-test structure

In the standard flat-universe formulation, $Om(z)$ is obtained directly from $H(z)$ and redshift. The defining property is that flat $\Lambda$CDM yields a redshift-independent constant equal to the present matter density fraction, $\Omega_{m0}$ [2602.09525]. This is the essential reason the diagnostic is called a null test: any statistically significant redshift dependence falsifies the statement that the late-time expansion is exactly that of flat $\Lambda$CDM [0807.3548].

For constant-$w$ dark energy in a flat universe,
$$
H^2(z)=H_0^2\left[\Omega_{m0}(1+z)^3+(1-\Omega_{m0})(1+z)^{3(1+w)}\right],
$$
which gives
$$
Om(z)=\Omega_{m0}+(1-\Omega_{m0})\frac{(1+z)^{3(1+w)}-1}{(1+z)^3-1}.
$$
This expression shows why $Om$ is useful: it converts redshift dependence in the expansion rate into a directly interpretable deviation from the $\Lambda$CDM constant line [1501.04047].

A recurrent interpretation in the literature is slope-based. In the original Sahni-type formulation and in several subsequent applications, a constant $Om(z)$ indicates $\Lambda$CDM, a decreasing $Om(z)$ with redshift is read as quintessence-like behavior, and an increasing $Om(z)$ with redshift is read as phantom-like behavior [2602.09525]. The practical attraction of this criterion is that it does not require an explicit parameterization of $w(z)$ and can often discriminate models even when $\Omega_{m0}$ is not known with high precision [1005.0433].

## 2. Variants and extensions

The original one-point definition was quickly generalized. A widely used two-point form is
$$
Om(z_i;z_j)=\frac{E^2(z_i)-E^2(z_j)}{(1+z_i)^3-(1+z_j)^3},
$$
or, in the rescaled version with $h(z)\equiv H(z)/(100\,\mathrm{km\,s^{-1}\,Mpc^{-1}})$,
$$
Omh^2(z_i,z_j)=\frac{h^2(z_i)-h^2(z_j)}{(1+z_i)^3-(1+z_j)^3}.
$$
In flat $\Lambda$CDM, $Omh^2$ is redshift independent and equals $\Omega_{m0}h_0^2$ [1506.03618].

Shafieloo, Sahni, and Starobinsky introduced **Om3**, a three-point ratio tailored to BAO and SNe data,
$$
Om3(z_1,z_2,z_3)=\frac{Om(z_2;z_1)}{Om(z_3;z_1)},
$$
for which flat $\Lambda$CDM gives $Om3=1$. Its main technical advantage is that it can be constructed from BAO and SNe distance ratios so that prior knowledge of $\Omega_{m0}$, $H_0$, and the sound horizon $r_s$ is not required [1205.2870].

A separate line of development introduced derivative-based diagnostics that generalize $Om$ beyond the flat-background null test. In particular,
$$
O_m^{(2)}(z)=\frac{2\left[(1+z)(1-h^2)+z(2+z)\,hh'\right]}{z^2(1+z)(3+z)},
$$
and
$$
O_k(z)=\frac{3(1+z)^2(h^2-1)-2z(3+3z+z^2)\,hh'}{z^2(1+z)(3+z)},
$$
reduce to $\Omega_{m0}$ and $\Omega_{k0}$, respectively, in $\Lambda$CDM. These forms were used in nonparametric tests of flatness and dark-energy dynamics from $H(z)$ reconstructions [1511.07066].

More recently, the **$w_0$-probe** was proposed as an $Om$-derived estimator of the present dark-energy equation of state,
$$
w_0\text{-probe}(z)\equiv\frac{Om(z)-1}{1-\Omega_{m0}},
$$
with the low-redshift limit $z\to 0$ yielding $w_0$ for any smooth underlying $w(z)$. Its purpose is to extract present-day equation-of-state information from $h(z)$ without differentiating the expansion history [2605.27230].

## 3. Reconstruction from observations

The observational appeal of $Om$ is that it is built from the background expansion alone. Direct inputs include **cosmic chronometers**, for which
$$
H(z)=-\frac{1}{1+z}\frac{dz}{dt},
$$
and Type Ia supernovae, which enter through
$$
\mu(z)=5\log_{10}d_L(z)+25,\qquad
d_L(z)=(1+z)\int_0^z\frac{dz'}{H(z')}.
$$
The same background information can be compressed through BAO observables such as $D_V(z)$, $A(z)$, and $d(z)=r_s/D_V(z)$, which are central to Om3-type constructions [2602.09525; 1205.2870].

Reconstruction strategies span both parametric and nonparametric approaches. Parametric analyses typically fit a chosen $Om(z)$ form or an implied $H(z)$ to combinations of OHD, SNe Ia, BAO, CMB-derived distance priors, and local $H_0$ calibrations, often with MCMC and affine-invariant samplers such as `emcee` [2602.09525]. Nonparametric methods include the **LOESS–SIMEX** pipeline, which reconstructs $H(z)$ and $H'(z)$ locally while correcting for measurement error by simulation–extrapolation; this approach was used to build both $Om$ and derivative-based null diagnostics without imposing a specific cosmological model [1511.07066].

The same general logic also appears in Gaussian-process reconstructions. In the $w_0$-probe analysis, Gaussian-process realizations of $h(z)$ from SNe Ia+BAO+CMB were used to compute both $Om(z)$ and the $w_0$-probe, with additional $\chi^2$-limited samples introduced to mitigate potential over-constraining from GP priors [2605.27230].

A persistent theme in these studies is that $Om$ is observationally simpler than diagnostics involving higher derivatives. Several papers explicitly contrast it with the statefinder $r$, emphasizing that $Om$ depends on the first derivative of the scale factor only, and is therefore easier to reconstruct from data [1005.0433].

## 4. Applications across dark-energy and modified-gravity models

The diagnostic has been used across a wide spectrum of models. In dilaton dark energy, the DDE-specific expression
$$
Om(x)_{DDE}=\frac{(1-\Omega_{m0})x^{3(1+\omega_\sigma)}+\Omega_{m0}e^{-\frac{1}{2}\alpha\sigma}x^3-1}{x^3-1}
$$
was employed to show that the matter–dilaton coupling modifies the $Om$ trajectory, although the effect is very small in the observationally allowed range $\alpha<0.001$; the reconstructed present equation-of-state value was reported as $\omega_{\sigma0}\simeq -0.952$ [1005.0433].

In scalar-field cosmologies, $Om$ was used to separate canonical quintessence, phantom, non-minimally coupled models, and tracker potentials. For power-law potentials $V(\phi)=V_0\phi^{2m}$, the paper on scalar-field models associated negative $Om$ curvature with quintessence and found parameter ranges in which cosmic acceleration slows down near the present epoch, with corresponding ages such as $t_0\simeq13.4825$ Gyr for the fitted $\phi^2$ model [1501.04047]. By contrast, in the purely kinetic DBI-like k-essence realization of Chaplygin gas, both $Om$ and the statefinder were found to fail to distinguish the model from $\Lambda$CDM at $68.3\%$ confidence level for $z\ll1$, with $\Delta Om(x)<0.06$ even up to $z\le 6$ [1003.2786].

Modified-gravity applications are comparably diverse. In viable $f(R)$ gravity, the two-point $Omh^2$ diagnostic was applied to Starobinsky and Hu–Sawicki models, which were reported to reproduce the observed $Omh^2$ values better than a $\Lambda$CDM model normalized to the Planck 2013 value $\Omega_m^0h^2=0.1426\pm0.0025$, while also exhibiting a characteristic signature around $z\sim2$ to $z\sim4$ [1506.03618]. In Hořava–Lifshitz Cardassian cosmologies, all four studied Cardassian variants produced non-constant $Om$ trajectories, interpreted as departures from $\Lambda$CDM [1108.1186]. In loop quantum cosmology, $Om$ trajectories for Chaplygin-gas-like models showed strong loop-quantum deviations at high redshift that fade near the present epoch and into the future [1410.6709]. In Einstein–Aether gravity, reconstructed $Om(z)$ curves classified power-law and future-singularity backgrounds as quintessence-like and emergent/intermediate backgrounds as phantom-like over $0.07\le z\le2.3$ [1509.07027]. In interacting generalized holographic Ricci dark energy, both one-point $Om(z)$ and two-point $Omh^2$ were used; for weak interaction $\lambda=0.0001$, the two-point values near $(0,0.57)$ remained close to the Planck value, indicating near-$\Lambda$CDM expansion [2407.18869].

## 5. Parametric innovations and recent empirical inferences

A notable recent direction is to parameterize $Om(z)$ itself rather than $w(z)$. One example is
$$
Om(z)=\alpha(1+z)^n,
$$
which reduces to flat $\Lambda$CDM when $n=0$ and $\alpha=\Omega_{m0}$, while $n<0$ and $n>0$ encode quintessence-like and phantom-like behavior, respectively [2303.04640]. Using Hubble, Pantheon, and BAO data, this model yielded
$$
H_0=68.4\pm1.3\ \mathrm{km\,s^{-1}\,Mpc^{-1}},\quad
\alpha=0.281^{+0.050}_{-0.046},\quad
n=0.010^{+0.10}_{-0.094},
$$
with derived values $w_0=-0.719^{+0.050}_{-0.046}$, $q_0=-0.579^{+0.075}_{-0.069}$, and $z_{\mathrm{tr}}=0.701^{+0.23}_{-0.15}$, hence a result consistent with a nearly constant $Om(z)$ [2303.04640].

Another proposal uses an exponential form,
$$
Om(z)=\alpha \exp\!\left(\frac{z}{1+z}+\beta\right),
$$
embedded in an $f(T,T_G)$ model. With $31$ CC, $26$ BAO, and $1701$ Pantheon+ data points, the reported constraints were
$$
H_0\in[68.46,77.38]\ \mathrm{km\,s^{-1}\,Mpc^{-1}},\quad
\alpha\in[-0.232,-0.068],\quad
\beta\in[0.218,0.560],
$$
together with $z_{\mathrm{tr}}\approx0.48$–$0.54$, $q_0\approx-0.34$, $\omega_0\approx-0.33$, and a cosmic age of $(13.28$–$13.87)$ Gyr [2506.06399].

The most explicit recent transition model is
$$
Om(z)=\frac{z^\ell}{(1+z)^m},
$$
for which
$$
\frac{d\,Om}{dz}=Om(z)\left(\frac{\ell}{z}-\frac{m}{1+z}\right),\qquad
z_*=\frac{\ell}{m-\ell}\quad (m>\ell).
$$
This analytic zero-crossing permits a finite redshift at which the slope changes sign [2602.09525]. Fitting OHD, Pantheon Plus, and SH0ES gave
$$
H_0=
\begin{cases}
72.12\pm2.10 & \text{(OHD)},\\
72.73\pm0.31 & \text{(OHD+PP)},\\
73.01\pm0.36 & \text{(OHD+PP+SH0ES)},
\end{cases}
$$
with reconstructed slope-transition redshifts
$$
z_*\approx 1.41,\ 0.65,\ 0.33
$$
for the same three dataset combinations, respectively. The paper also reported a deceleration-to-acceleration transition in the interval $z_t\in[0.5,0.8]$, ages of approximately $13.21\pm1.02$, $13.51\pm0.53$, and $13.89\pm0.40$ Gyr, and $\Delta\mathrm{AIC}\approx2$–$4$, $\Delta\mathrm{BIC}\approx2$–$4$ relative to $\Lambda$CDM, indicating statistical competitiveness but not decisive preference for dynamics [2602.09525].

A conceptually distinct extension is the $w_0$-probe. Applied to Gaussian-process reconstructions from SNe Ia+BAO+CMB, it yielded a low-redshift estimate
$$
w_0\simeq -0.62\pm0.03 \quad (95\%~\mathrm{C.L.}),
$$
while both $Om(z)$ and the $w_0$-probe were reported to exclude flat $\Lambda$CDM at the $95\%$ confidence level; in a more conservative $\chi^2$-limited sample, the inferred range was $w_0\in(-0.8,-0.5)$ as $z\to0$ [2605.27230].

## 6. Interpretive caveats, ambiguities, and outlook

Although the null-test logic is straightforward, interpretation requires care. First, the flat form of $Om$ is contaminated by curvature. For small but nonzero $\Omega_k$,
$$
\delta Om(z)=\Omega_k\frac{x+1}{x^2+x+1},\qquad x=1+z,
$$
so curvature can mimic apparent redshift evolution. This is one motivation for curvature-aware constructions such as $O_k(z)$ and for combining $Om$ with external curvature constraints [0807.3548; 1511.07066].

Second, the literature is not fully uniform about qualitative slope assignments. The original Sahni-style usage and several model studies identify decreasing $Om(z)$ with quintessence-like behavior and increasing $Om(z)$ with phantom-like behavior [0807.3548; 2602.09525]. By contrast, some later reconstructions and $q(z)$-based analyses operationalize slope or “curvature” differently and report the opposite mapping for their conventions or model classes [1511.07066; 2304.09749]. This suggests that the sign interpretation should always be read together with the exact definition being used, the redshift variable, and whether the discussion concerns $Om$, $Omh^2$, or a derived quantity such as $O_m^{(1)}$ or $O_m^{(2)}$.

Third, the diagnostic is simpler than derivative-based probes but not free of systematics. It retains sensitivity to the normalization of $H(z)$ through $H_0$, to cosmic-chronometer stellar-population modeling, to BAO sound-horizon assumptions when pairwise cancellations are not used, and to SNe calibration choices [1511.07066]. Pairwise and ratio-based variants such as $Omh^2$ and Om3 were introduced precisely to reduce these dependencies [1205.2870].

Despite these caveats, the diagnostic remains central because it isolates departures from flat $\Lambda$CDM in a compact and observationally tractable form. Across the literature, it supports several distinct roles: a strict null test of $\Lambda$CDM, a low-derivative classifier of effective dark-energy behavior, a scaffold for model-independent reconstructions, and a building block for newer probes such as Om3 and the $w_0$-probe [2605.27230]. The current record is mixed rather than uniform: some analyses find near-constant $Om(z)$ compatible with $\Lambda$CDM, while others report late-time evolution suggestive of non-$\Lambda$ behavior. This suggests that future progress will depend primarily on higher-precision $H(z)$, BAO, SNe Ia, and standard-siren measurements, together with explicit control of curvature and calibration systematics [1205.2870; 2602.09525].

Source: https://www.emergentmind.com/topics/om-diagnostic