^₁₃C MRS is a powerful tool for non-invasive investigation of carbohydrate metabolism, such as natural abundance glycogen in human muscle [1]. In terms of RF-coil design, inherent low sensitivity of ¹³C signals and heteronuclear ¹³C-¹H J coupling require double tuned ¹³C-¹H coils to provide high sensitivity at the ¹³C frequency [2,3]. Moreover, high transmit efficiency at ¹H frequency is needed to ensure proper ¹H decoupling during ¹³C signal acquisition. Double tuned transceiver array with extended spatial coverage were recently designed for X-nucleus spectroscopy [4,5]. We aimed to demonstrate that a 4 channel ¹³C-4 channel ¹H array coil can be successfully used for in vivo ¹³C MRS studies in humans at ultra-high field. Coil performance was evaluated by detecting in vivo ¹H-decoupled ¹³C resonance of Glycogen C₁ in calf muscles at 7T.Coil design: A double tuned 4 channel ¹³C-4 channel ¹H transceiver phased array coil (fig.1) was designed, built and optimized for human studies at 7T [5]. Its dimensions allow MR-measurements in different body part, such as calf muscles or brain. The loop dimensions (¹³C: 88x80mm², ¹H: 120x90mm²) were chosen for efficient B₁⁺ and B₁^- at ¹³C frequency (75MHz) and for high transmit efficiency at ¹H frequency (300MHz). The coil FOV can cover either the back half of the brain or both calf muscles. Bench measurements: Bench measurements were performed using a Network-Analyzer (E5071C; Agilent, USA). The coil, loaded with two calf muscles of a male volunteer, was matched at 50Ω. S_ij parameters, unloaded and loaded quality factors were measured at 75MHz and 300MHz. Simulations: FDTD simulations were performed using Sim4Life (ZMT, Switzerland) on the posable Duke model from “Virtual Family” (ITIS, Switzerland). Legs were placed next to each other in order to load the coil with the two calf muscles. B₁⁺ and SAR_10g,max were measured at 75MHz and 300MHz, with all loops driven in quadrature (0^o,90^o,180^o,270^o), for 1W total input power. MR measurements: MR experiments were performed on a 7T human scanner (Siemens, Erlangen, Germany). Natural abundance ¹³C spectra were acquired on a healthy male volunteer (age=26y, weight=82kg) who gave informed consent according to the procedure approved by the local ethics committee. Coil’s FOV covered both calf muscles of the volunteer in supine position. Glycogen C₁ resonance was acquired using an adiabatic half passage (AHP) pulse for excitation (2ms) and continuous wave (CW) for ¹H decoupling during ¹³C signal acquisition (vector size=2048, TR=3.4s, BW=8 kHz, acquisition time=256ms, decoupling duration=20ms, 1000 averages). Spectra acquired by each loop were combined to further enhance the SNR.Bench measurements: Reflected power of individual loops was lower than -25dB. Isolation between the loops was better than -12dB for the ¹³C array and -15dB for the ¹H array. The cross-coupling between ¹H and ¹³C loops was lower than -20dB at 75MHz and 300MHz. The unloaded to loaded Q-ratio was ~350/90 for the ¹³C loops and ~280/100 for the ¹H loops, demonstrating good coil efficiency. Simulations: The B₁⁺ produced by the ¹³C array in quadrature mode exhibits a large and symmetric spatial coverage on both calf muscles (fig.2). With 1W total input power, it can reach up to 8µT in regions close to the loops, and around 3µT at 3cm depth. With the same input parameters, the ¹H array generates an almost symmetric B₁⁺ over the ¹³C array FOV. Energy deposition (fig.3) is mainly localized in subcutaneous regions close to the coil (SAR_10g,max=1.29[W/kg] at 75MHz and 0.61[W/kg] at 300MHz). MR measurements: Glycogen C1 from the gastrocnemius muscles of the human calves was detected at 100.5ppm, as well as lipids from subcutaneous adipose tissue, such as glycerol at 62ppm and 69.5ppm and unsaturated lipid at 130ppm, indicating high sensitivity of the ¹³C array coil. Moreover, the glycogen C₁ resonance was successfully decoupled, as suggested by the in vivo ¹³C spectrum (fig.4), indicating a good transmit efficiency of the ¹H array coil within the sensitive volume of the ¹³C array coil.It is feasible to perform in vivo ¹H-decoupled ¹³C MRS in the human calves at 7T using double tuned ¹³C-¹H transceiver array coils. The 4 channel ¹³C-4 channel ¹H array coil presented in this work provides larger FOV than single loop surface coils, while exhibiting a strong sensitivity at the ¹³C frequency and high transmit efficiency at the ¹H frequency. These results confirm the potential of array coils to perform further ¹³C MRS in humans at ultra-high field, especially in the brain, where the metabolite concentrations are even lower.