Integrated Galaxy-wide Initial Mass Function

• IGIMF is a framework that integrates the IMFs of embedded star clusters to derive the overall stellar mass distribution in a galaxy.
• It employs a power-law embedded cluster mass function and deterministic sampling methods to connect star formation rate with variations in the high-mass stellar end.
• The approach informs galaxy chemical evolution and stellar population synthesis by modeling environmental dependencies such as metallicity, density, and cosmic-ray influence.

Searching arXiv for recent and foundational IGIMF papers to ground the article. arXiv search query: "Integrated Galaxy-wide Initial Mass Function IGIMF" Integrated Galaxy-wide Initial Mass Function (IGIMF) denotes the galaxy-wide stellar initial mass function obtained by summing the stellar IMFs of all embedded star-forming clusters formed during a short star-formation epoch, rather than identifying the galaxy-wide IMF with the IMF of a single cluster. In this framework, the birth-mass distribution of a galaxy is a derived quantity controlled by clustered star formation, the embedded cluster mass function (ECMF), the relation between cluster mass and the most massive star that can form, and, in later extensions, metallicity, density, cosmic-ray regulation, and cluster-to-cluster IMF variations (Weidner et al., 2010, Yan et al., 2017).

1. Definition and formal structure

The basic IGIMF construction writes the galaxy-wide IMF as an integral over the cluster population,

$$\xi_{\mathrm{IGIMF}}(m,t)=\int_{M_{\mathrm{ecl,min}}}^{M_{\mathrm{ecl,max}}(\mathrm{SFR}(t))}\xi\!\left(m\le m_{\max}(M_{\mathrm{ecl}})\right)\,\xi_{\mathrm{ecl}}(M_{\mathrm{ecl}})\,dM_{\mathrm{ecl}},$$

where $\xi(m)$ is the stellar IMF inside a single embedded cluster, truncated at the cluster-dependent upper stellar mass $m_{\max}(M_{\mathrm{ecl}})$, and $\xi_{\mathrm{ecl}}(M_{\mathrm{ecl}})$ is the ECMF. The star-formation epoch is typically taken to be $\delta t=10\,\mathrm{Myr}$, so that the stellar mass formed in the episode is $\mathrm{SFR}\times\delta t$ (Weidner et al., 2010, Yan et al., 2017).

In foundational IGIMF work, the cluster-scale IMF is usually represented as a canonical Kroupa-like multi-part power law. One widely used form has slopes $\alpha_0=0.30$ for $0.01\le m/M_\odot<0.08$, $\alpha_1=1.30$ for $0.08\le m/M_\odot<0.50$, and $\xi(m)$0 above $\xi(m)$1, with the high-mass segment extending to the cluster-specific $\xi(m)$2 (Weidner et al., 2010). Other implementations employ simplified two-part forms or tapered power laws, but retain the same structural principle: the galaxy-wide IMF is not imposed a priori, it is produced by integrating over a distribution of star-forming units (Dib, 2022, Weidner et al., 2010).

The ECMF is usually taken as a power law,

$\xi(m)$3

with $\xi(m)$4 near $\xi(m)$5 in many applications. The lower cluster-mass limit is commonly set to $\xi(m)$6, motivated by Taurus–Auriga-like groups, while the upper limit $\xi(m)$7 depends on the instantaneous SFR (Calura et al., 2010, Weidner et al., 2013).

2. Constituent relations and sampling assumptions

The defining structural relations of IGIMF theory couple cluster formation to galaxy-wide star formation. A commonly used calibration links the maximum embedded-cluster mass to the SFR through

$\xi(m)$8

while other implementations write the same scaling in logarithmic form, for example

$\xi(m)$9

These relations imply that high-SFR galaxies populate much more massive clusters than low-SFR systems (Weidner et al., 2010, Fontanot et al., 2016).

A second constitutive ingredient is the empirical relation between the mass of the most massive star and the stellar mass of the host cluster. In IGIMF calculations this is often enforced by the coupled conditions that there is exactly one star above $m_{\max}(M_{\mathrm{ecl}})$0 and that the integral of the truncated IMF equals the cluster mass. In practice, this suppresses very massive stars in low-mass clusters and is one of the main reasons the integrated galaxy-wide IMF steepens at low SFR (Weidner et al., 2010, Yan et al., 2017).

Sampling is not a purely technical detail in this literature. The optimally sampled formulation treats both the ECMF and the cluster IMF deterministically, enforcing exact mass conservation and reproducing the observed $m_{\max}(M_{\mathrm{ecl}})$1–$m_{\max}(M_{\mathrm{ecl}})$2 and $m_{\max}(M_{\mathrm{ecl}})$3–SFR relations with little intrinsic scatter. The same framework predicts that if there were no binaries, galaxies with $m_{\max}(M_{\mathrm{ecl}})$4 should host no Type II supernova events, because the galaxy-wide most massive star remains below $m_{\max}(M_{\mathrm{ecl}})$5 (Yan et al., 2017).

The ECMF itself is not always kept fixed. Some implementations assume a constant $m_{\max}(M_{\mathrm{ecl}})$6, whereas others allow $m_{\max}(M_{\mathrm{ecl}})$7 to flatten with SFR. One family of models adopts

$m_{\max}(M_{\mathrm{ecl}})$8

while some semi-analytic realizations use

$m_{\max}(M_{\mathrm{ecl}})$9

This dispersion of prescriptions is itself part of the current IGIMF landscape (Recchi et al., 2014, Fontanot et al., 2016).

3. SFR, density, and metallicity dependence

The original and most robust IGIMF effect is a suppression of the galaxy-wide massive-star fraction at low SFR. In the review-level treatment of the theory, representative high-mass slopes above $\xi_{\mathrm{ecl}}(M_{\mathrm{ecl}})$0 are $\xi_{\mathrm{ecl}}(M_{\mathrm{ecl}})$1–$\xi_{\mathrm{ecl}}(M_{\mathrm{ecl}})$2 at $\xi_{\mathrm{ecl}}(M_{\mathrm{ecl}})$3, $\xi_{\mathrm{ecl}}(M_{\mathrm{ecl}})$4–$\xi_{\mathrm{ecl}}(M_{\mathrm{ecl}})$5 at $\xi_{\mathrm{ecl}}(M_{\mathrm{ecl}})$6, and $\xi_{\mathrm{ecl}}(M_{\mathrm{ecl}})$7–$\xi_{\mathrm{ecl}}(M_{\mathrm{ecl}})$8 at $\xi_{\mathrm{ecl}}(M_{\mathrm{ecl}})$9. In this regime the IGIMF is top-light in the sense of a steepened high-mass tail (Pflamm-Altenburg et al., 2010).

At the opposite extreme, starbursts can drive the IGIMF in the top-heavy direction. In the starburst formulation of Weidner, Kroupa, and Pflamm-Altenburg, the high-mass cluster-scale slope is kept canonical below $\delta t=10\,\mathrm{Myr}$0, but for more massive clusters is parameterized as

$\delta t=10\,\mathrm{Myr}$1

with an imposed floor $\delta t=10\,\mathrm{Myr}$2. The physical motivation is crowding of pre-stellar cores in compact, massive proto-clusters, together with dynamical evidence from ultra-compact dwarf galaxies and globular clusters (Weidner et al., 2010).

The corresponding galaxy-wide results depend sensitively on the ECMF and on whether low-mass clusters are present during starbursts. For $\delta t=10\,\mathrm{Myr}$3 and $\delta t=10\,\mathrm{Myr}$4, the fitted IGIMF high-mass slope above $\delta t=10\,\mathrm{Myr}$5 is $\delta t=10\,\mathrm{Myr}$6 for $\delta t=10\,\mathrm{Myr}$7. If instead $\delta t=10\,\mathrm{Myr}$8, the same sequence becomes $\delta t=10\,\mathrm{Myr}$9. With a top-heavy ECMF, $\mathrm{SFR}\times\delta t$0 and $\mathrm{SFR}\times\delta t$1, the corresponding values are $\mathrm{SFR}\times\delta t$2 (Weidner et al., 2010).

A closely related branch of the theory incorporates explicit density and metallicity dependence in the high-mass slope. One frequently used prescription sets

$\mathrm{SFR}\times\delta t$3

With the empirical relation $\mathrm{SFR}\times\delta t$4, denser and more metal-poor clusters acquire flatter $\mathrm{SFR}\times\delta t$5, so the IGIMF can respond simultaneously to SFR, density, and metallicity (Recchi et al., 2014, Dabringhausen et al., 2023).

4. Generalizations of the framework

Later developments relaxed the assumption that every cluster shares identical IMF parameters apart from deterministic truncation. One extension treats the cluster IMF as a tapered power law with parameters $\mathrm{SFR}\times\delta t$6 drawn from Gaussian distributions. In that approach the IGIMF is a four-dimensional integral over cluster mass and IMF-parameter distributions. Using Milky Way young-cluster dispersions $\mathrm{SFR}\times\delta t$7, $\mathrm{SFR}\times\delta t$8, and $\mathrm{SFR}\times\delta t$9, increasing $\alpha_0=0.30$0 shifts the IGIMF peak mass from $\alpha_0=0.30$1 for a universal cluster IMF to $\alpha_0=0.30$2 for $\alpha_0=0.30$3 and to $\alpha_0=0.30$4 for $\alpha_0=0.30$5, rendering the IGIMF more bottom heavy (Dib, 2022).

A second major extension is the cosmic-ray regulated IGIMF. There, the cluster-scale IMF keeps a variable break mass $\alpha_0=0.30$6, so the low-mass structure responds to the cosmic-ray energy density while the high-mass slope still responds to cluster density. This construction can produce IGIMFs that are shallower at the high-mass end and steeper at the low-mass end than a Kroupa IMF under appropriate combinations of SFR and $\alpha_0=0.30$7, and it reproduces the observed increase of dwarf-to-giant ratios and IMF-sensitive spectral indices with velocity dispersion, albeit with shallower trends than observed in local early-type galaxies (Fontanot et al., 2018, Fontanot et al., 2023).

A third line of work incorporates the IGIMF directly into stellar population synthesis. The SPS-VarIMF models compute spectra, luminosities, remnant populations, and mass-to-light ratios for time-dependent gwIMFs. In those calculations, late-type galaxies can be distinguished in UV and optical colors under invariant and varying gwIMF assumptions, whereas early-type galaxies often have almost identical colors but gwIMF-dependent $\alpha_0=0.30$8 ratios that differ by up to an order of magnitude. The same models predict that massive present-day elliptical galaxies would have been $\alpha_0=0.30$9 times as bright as at present when they were forming (Zonoozi et al., 5 Feb 2025).

These generalizations imply that “IGIMF” no longer denotes a single rigid implementation. It designates a family of clustered-star-formation constructions whose common core is the integration over embedded clusters, but whose details differ in the treatment of the cluster IMF, environmental dependence, and stochastic versus deterministic structure.

5. Chemical evolution and galaxy-scale applications

One of the earliest systematic applications of the IGIMF was to galaxy chemical evolution. In the analytical formalism with SFR- and metallicity-dependent IGIMFs, the returned fraction $0.01\le m/M_\odot<0.08$0 and net metal yield $0.01\le m/M_\odot<0.08$1 become time-dependent functionals of the IGIMF, so the governing chemical-evolution equations become non-linear and require iterative solution. In this framework the mass–metallicity relation emerges naturally, because low-SFR dwarfs form top-light IGIMFs with smaller yields, whereas high-SFR systems form flatter IGIMFs with larger yields. In the closed-box case, however, the predicted low-mass metallicities remain too high by $0.01\le m/M_\odot<0.08$2–$0.01\le m/M_\odot<0.08$3 dex after $0.01\le m/M_\odot<0.08$4, and modest inflow plus mass-dependent outflow is still required for a good fit to the Lee et al. relation (Recchi et al., 2014).

Semi-analytic galaxy formation models use the same mechanism to address abundance ratios in early-type galaxies. In the SAG implementation of the top-heavy IGIMF, the SFR-dependent high-mass slope steepens the modeled $0.01\le m/M_\odot<0.08$5–stellar-mass relation from the essentially flat universal-IMF value $0.01\le m/M_\odot<0.08$6 to $0.01\le m/M_\odot<0.08$7 for the favored SAGTH5B2 model, compared with the observed $0.01\le m/M_\odot<0.08$8 for the full Thomas et al. sample. The same study found that $0.01\le m/M_\odot<0.08$9 and $\alpha_1=1.30$0 best match the abundance-ratio slope, mass-to-light ratios, and luminosity functions simultaneously (Gargiulo et al., 2014).

In the GAEA semi-analytic model, the IGIMF likewise reproduces the observed rise of $\alpha_1=1.30$1 with stellar mass and predicts that “proper” stellar masses and $\alpha_1=1.30$2 ratios exceed the “apparent” values inferred under a universal Chabrier IMF from photometry. In that framework the IGIMF does not significantly alter the trend toward shorter formation timescales in more massive galaxies; rather, $\alpha_1=1.30$3 becomes a tracer of the highest SFR episodes, because flatter high-mass IGIMFs during intense star formation raise the Type II to Type Ia supernova ratio (Fontanot et al., 2016).

These results anchor the IGIMF in a specific explanatory program. The framework does not merely rescale star-formation indicators; it modifies SN rates, remnant fractions, enrichment timescales, and therefore the interpretation of abundance patterns, mass–metallicity relations, luminosity functions, and stellar masses.

6. Observational diagnostics, tensions, and open questions

Several observational diagnostics recur across the IGIMF literature. At low SFR, the framework explains the H$\alpha_1=1.30$4 deficit relative to FUV in dwarf galaxies, produces an approximately linear SFR–$\alpha_1=1.30$5 relation over about four orders of magnitude, and predicts gas depletion times of about $\alpha_1=1.30$6 across star-forming galaxies (Pflamm-Altenburg et al., 2010). In dwarf spheroidals, detailed chemical-evolution models of Sagittarius find that the IGIMF predicts lower $\alpha_1=1.30$7 ratios than classical IMFs and lower hydrostatic-to-explosive $\alpha_1=1.30$8-element ratios, qualitatively matching the data, while for the ultra-faint dwarf Boötes I a time-evolving gwIMF yields a mildly bottom-light and top-light gwIMF together with quenching about $\alpha_1=1.30$9 after formation (Vincenzo et al., 2015, Yan et al., 2020).

At high masses, the framework is increasingly used to interpret dynamical and stellar-population anomalies in early-type galaxies. In the radial-acceleration analysis of 462 ETGs, canonical-IMF baryonic masses fail at high accelerations under Newtonian and MONDian radial-acceleration relations, whereas using the IGIMF improves agreement because the theory predicts an overabundance of stellar remnants in massive ETGs. Under the standard central $0.08\le m/M_\odot<0.50$0 gradient, the net mass boost in the most massive systems is typically $0.08\le m/M_\odot<0.50$1 (Dabringhausen et al., 2023). In disk galaxies, environment-dependent gwIMF calculations similarly imply that high-mass galaxies have stellar-plus-remnant masses underestimated by factors of order a few when a constant $0.08\le m/M_\odot<0.50$2 is assumed, which shifts the apparent baryonic Tully–Fisher relation away from the MOND slope while leaving the “true” IGIMF-based baryonic masses closer to it (Zonoozi et al., 16 Jul 2025).

The principal controversies are not over the existence of clustered star formation, but over the calibration of the constituent relations. The SFR–$0.08\le m/M_\odot<0.50$3 relation is extrapolated in some starburst applications; the $0.08\le m/M_\odot<0.50$4 or $0.08\le m/M_\odot<0.50$5 prescriptions are physically motivated but not unique; the ECMF slope $0.08\le m/M_\odot<0.50$6 and lower cutoff $0.08\le m/M_\odot<0.50$7 in starbursts remain poorly constrained; and different realizations make different assumptions about metallicity dependence, low-mass slopes, stochasticity, and the treatment of binaries (Weidner et al., 2010, Dib, 2022). A related misconception is that all IGIMF models necessarily predict the same qualitative trend at all masses. In fact, low-SFR systems are generally top-light at the high-mass end, high-SFR starbursts can be top-heavy, and some extensions simultaneously produce bottom-heavy low-mass behavior.

A further open question concerns how much of the observed galaxy-to-galaxy IMF phenomenology can be explained by deterministic clustered-star-formation physics alone, and how much requires additional cluster-to-cluster IMF scatter or other environmental regulators. The 2022 cluster-variation models find that the observed low-mass stellar mass functions of ultra-faint dwarfs can only be reproduced when IMF variations of the same order as those measured in present-day Milky Way clusters are included; models without such variations fail even when metallicity and density dependencies are added (Dib, 2022). This suggests that the long-term future of the subject may lie not in choosing between “universal IMF” and “IGIMF,” but in quantifying which layers of environmental dependence are indispensable and which are merely convenient parameterizations.